ML19327A793

Review of Diablo Canyon Power Plant Diesel Generator Allowed Outage Time Study

ML19327A793
Person / Time
Site:	Diablo Canyon
Issue date:	09/30/1989
From:	Bozoki G, Fitzpatrick R, Sabek M BROOKHAVEN NATIONAL LABORATORY
To:	NRC
Shared Package
ML19327A789	List: ML19327A788 ML19327A793
References
CON-FIN-A-3958 NUDOCS 8910180122
Download: ML19327A793 (91)

Text

{{#Wiki_filter:y.,, -,. - -3. g;;?: ENCLOSURE 4 .,J. l t 703SafRimpMS&WMMsg611MMEI5i M kN d y di @ G g ? ? E G R E 4si g g g g y ( p m u m v4 i i L 1. m is A REVIEW OF THE DIABLO CANYON POWER PLANT DIESEL GENERATOR ALLOWED OUTAGE TIME STUDY. Su G. Bozoki R. Fltzpatrick M. Sabek r September 1989 T h3 Risk Evaluation Group Department of Nuclear. Energy. Brookhaven National uboratory Upton, NY 11973 1 Prepared for L U.S. Nuclear Regulatory Commission H . Washington, DC 20555 Contract No. DE-AC02-76CH00016 FIN A-3968 \\ i . b b !E$ $ $ ! N i b 5 [$$s sh5d $ $ N!!b N aS N N b $ !) E[db8$2 bib ~ l l 1-1- l 8910180122 891004 FDR ADOCK 05000275 PDC l, P h ., - +. -, -,n--

y g l .?

. -^

I'. INTRODUCTION 'I.1 scope nnd Obiectives The scope of the present study is to support the NRC's effort to respond to a request by the Pacific Gas and Electric Co. (PG6E) to. modify the Technical Specifications for Allowed Outage Time (A0T) for the Diesel Generators (DCs) presently operating and an additional one to be installed at its Diablo Canyon Nuclear Power Plant. The direct objectives of this report are: 'I

• To review the approaches applied in a supporting study, attached to.the request, which analyzes the impact of the system and A0T modifications to the core damage frequency (CDF).

f,

• To provide and to compare the results of auditing or validating calculations performed at BNL with those obtained in the PG&E study and to comment on them.

l-I.2 Backaround s l: The PG&E request is supported by detailed analysis of the unavailabilities of system configurations consisting of five and six DGs under various redundancy and A0T conditions as well as an evaluation of the impacts l, of the modified system and A0T conditions to the CDF. The document entitled, "Diablo Canyon Power Plant, Diesel Generator Allowed outage Time Study,.2 contains the description of the approaches used and the results of the calculations. The study extensively uses, "The Diablo Canyon Probabilistic Risk Assessment (DCPRA)"8 presently under review at BNL. Additional information on the AOT study was provided by PG&E in a presentation at'the NRC (June 1989) and in two letters sent to BNL in regard to various 8 review questions. AOT Study September 11, 1989 L -m y ---_,-.

_ _ ~ _ _ _ - i*r .l '. C l4 -( 2.3.Ormanization of the Renort -The first part-(Part 1) of the present report summarizes the results obtained by BNL in reviewing the methodology and calculations described in the A0T stuBy.1 TheLsecond part (Part 2) contains the detailed descriptions of the Diablo' Canyon diesel generator, the diesel fuel transfer systems, and their PRA unavailability models.* This latter part also contains the results of a comprehensive review of the models performed recently at BNL and targeted to A0T-related aspects of the system's analysis. Part 1 is organized as follows: Section 1 presents the proposed relaxation of Technical Specifications and briefly discusses PG6E's methodology and the results of the justification analyses. Section 2 comments on PG&E's approaches and maintenance data used. It compares the results 'of BNL's calculations on diesel unavailabilities (top event split fractions) obtained under.various AOT conditions for both. seismic and nonseismic accident sequences with those given in the A0T study. Section 3 describes the results - of BNL's audit 'and sensitivity calculations performed by scrutinizing the-CDF. . impact and risk ratio analyses of the ACT study. Section 4 summarizes the findings and the main conclusions of the BNL review. Appendix I contains the prior-maintenance duration distribution used in the A0T study and maintenance durttion and failure rate data for various diesel generator subsystems and components, p t Part 2 is essentially represented by Letter Report 07" on the DCPRA review which is entitled, "A Review of System Analysis in the DCPRA: Diesel Generator and Diesel Fuel Transfer Systems." "The description of its organization can be found in its introductory Section I.2, We note that some of the review findings include open issues at this time, however, we do not believe the resolution of these items will have an appreciable effect upon th'e overall PRA results and conclusions. l l A0T Study September 11, 1989 i

pp Riy 2 4

h..

fep 4 I k f' 9 {4 t f: EN...E. M....,_E_Nh' &iW8"WMMM...M....f#iEGWEM. w.EME.....M....N..1W,_<._!#_iU_ MME_E..MP.:is,E,.._*."i]* h.. M my y c. PART 1 REVIEW RESULTS OF THE DG AOT STUDY m o s a w m isWeiew es4gs' PMghra,- m manyss-tegm s hemigage;it s t P 1 t t ) A0T Study September 11, 1989 l 1 9

1* '=.= l 1. JUSTIFICATION ANA1.TSES FOR RELAKED TECHNICAL SPECIFICATIONS FOR DIESEL' -l CENERATORS. For the sake of clear understanding and reader convenience, this section reiterates the Tec'hnical Specification relaxation issues requested by PG&E and provides a'brief summary of the methodology and results of their justification { analyses. 1.1 Procesed Relaxation of Technical Soecifications for Diesel Generators Currently, five DGs constitute the emergency DG system at the DCPP Units 1 and 2: two DGs dedicated to Unit 1, two DGs dedicated to Unit 2, and one, a - " swing diesel

• is shared between the two units. The swing diesel is

[ physically located in Unit 1. i 'In order to increase the flexibility of plant operation and diesel maintenance scheduling sfficiency, PG&E committed itself to install a sixth DG I - by the fourth refueling outage of Unit 2 (scheduled for October 1991). The sixth DG will also be an ALCO type DG like the five existing ones. With the, sixth diesel installed and operable, each Diablo Canyon unit will have three dedicated DGs which will simplify the operation of the plant. The present DCPP Technical Specifications provide a 72-hour AUT when a given unit DG is inoperable with that unit in Modes 1 through 4. When a DG becomes inoperable, the operability of the ac offsite sources must be demonstrated by performing surveillance tests within one hour and at least once per eight hours thereafter. If the DG became inoperable due to causes other than preventive maintenance or testing, the operability of the remaining L DGs must be demonstrated within 24 hours (regardless of when the inoperable DG i is restored to operable status). The inoperable DG aust be restored to operable status within the 72-hour A0T or action must be initiated to place the unit to cold shutdown (Mode 5), where the subject' limiting condition for operation (LCO) is no longer applicable. l-A0T Study September 11, 1989 l

1 v j ~ d F The relaxation of the Technical Specificatiens (No.3.8.1.1 Action : Statement b) proposed by PG4,E is the following: Increase'the AOT fre _pthe h:' ' current three days (72-hours) to seven days (168 hours), so that en:rective j (non scheduled) maintenance, inspection and post maintenance operability ~ testing appropriately and conveniently could be performed. The proposal relates to both diesel configurations, the current five, as j well as the planned six diesel configuration. Preventive (scheduled) . maintenance (overhauls) of the dedicated DGs would be performed, as in the ] past,;during the unit's refueling periods. g. -1.2~ Methodolonv=of the Justification Analyses 1 PG&E claims that the above proposed A0T relaxation.is fully supported by plant experience, training of personnel on advanced diesel maintenance, recent .irprovements of the DGs and by the results of justification analyses described in Reference 1. J 1he justification analyses were directed to the assessments of two main issues: L

a. the appropriateness of a seven day A0T for the purposes of unscheduled maintenance of the present and. planned DG configurations and p

l

b. the safety impact of performing required scheduled maintenance of the C

swing diesel given a seven-day A0T. 1 Two approaches were used for these assessments. The first was based on the DCPRA,a thus Reference'1 and consequently the present report refer to it 4 L as the " risk analysis approach." The second was based on stand alone fault-l tree models of the current and planned DG configurations. Reference 1 as well as the present report refer to this as the " reliability analysis" approach. The annual and relative risks were evaluated by both of the approaches. Relative risk was defined by the ratio of the risk during the A0T to the risk 5 A0T Study September 11, 1989 ---.-m-- ...--_.m,--....,m.

' V ? J !h: ..e.' 4'y L e r,- = during the: time interval when rm DC is in maintenance or test. This " risk ]

ratio" (RR) constrains the duration of the A0T by requiring that the ratio q

. should be less~ than unity. In general, " risk" may refer to system 1 unavailability, core damage frequency or health risks, depending on the- - level" where the effect of the A0T is evaluated. The PG&E A0T study evaluated " risk" at the core damage frequency level. The application of the reliability analysis approach by PG&E was intended to complement'the risk analysis approach. A PRA usually calculates time . averaged risk values; time-dependent effects (like testing) on the availability of the remaining diesels when one DC is'in maintenance or staggered testing are not taken into account. In addition, the unavailability: modelling of the DCs in a PRA does not usually go " deep" enough, so that, failure modes-of the diesel subsystems or support systems are not explicitly indicated in the model. The time-dependent unavailability analysis was performed by PG&E on the fault' tree models of.the" diesel systems by using the FRANTIC ABC PC computer l code. 1.3 Results of the Justification Analyses BothEof the approaches, risk and reliability analyses, were used to calculate the following cases: e Base casec--'SDG configuration, three-day A0T on all DCs to perform unscheduled maintenance. The risk analysis approach addressed also performing scheduled maintenance on the swing diesel durin8 Power operation of one unit with the other unit in refueling. Total scheduled outage was assumed to be ten days (i.e., several three day A0T periods) during a refueling period of 1.5 years. Table 2.1 contains the definitions of the calculations performed by the risk analysis approach, these calculations are denoted by 1A and'2. -, September 11, 1989 A0T Study W

0 ? L. f a e second came. SDG configuration, seven day ACT on all DGs to perform ( unscheduled maintenance. The r*,sk analysis approach also addressed l performin5 scheduled maintenance on the swing diesel. Total scheduled l ~^ outage was seven days (no multiple outages) during a refueling period of 1.5 In Table 2.'1, these calculations are denoted by 1B and 3. l years. e Third Casa. 6DG configuration, seven day ACT on all DCs to perfora l unscheduled maintenance. Since there is no swing diesel, schtduled maintenance can now be performed without affecting the other unit. In Table 2.1 this calculation is denoted by 4 t In order to determine the relative risk, several support calculation 6 were carried out. Those associated with the risk analysis approach are denoted by 5 and 6 in Table 2.1. Calculation 5 analyzed the condition when no j maintenance (scheduled or unscheduled) is allowed on any of the $DCs. Calculation 6 provided the risk ($DG configuration) if the swing diesel were unavailable for the entire year (i.e., calculated the conditional core damage frequency). This calculation assumed seven day A07 for unscheduled maintenance on other DGs. For completeness, the results of both of the analyses, risk and reliability, are reproduced in Table 1.1 from Table 6.1 of Reference 1. Based on the data presented, PG&E concluded that: e The risk ratio criterion is satisued for all cases by both methods of - '!i analysis. e'The effect on risk of changing from a three day to a seven day ACT is insignificant; on the order of 1 to 3 percent of the CDF. e the effect on risk of adding the sixth DG is greater than the effect of changing to a seven. day ACT with an overall decrease of the order of 5' to 15 ptreent in CDF.. Both of the analysis approaches confirmed the appropriateness of a seven day ACT for the purpose of performing unscheduled maintenance for both the five and six DG configurations. r ACT Study September 11, 1989 i l 1

j c, 1

• According to the results of the risk analysis approach,1) there is a t)/

' negligible ' increase in risk associated with a seven-day A0T over a j three day A07 with re5ard to performin5 scheduled maintenance on the i swing diesel and 2) the resulting quantitative benefits of a single seven day A0T far outweighs the risk associated with multiple three day ) l A0Ts. i I I O- ] + t i t I i e b t i t i i a e ~ Y - m. t 1% + b p i j ll I l-ACT Study 8 September 11, 1989 l ': l l.; l' . P ..-.m-

U* o. Table 1.1 Analytical Results for Unplanned and Planned Maintenance Activities 8 + PRA Analysis Reliability Analysis 8 Unelanned (Uncl ar.ne d) Unclanned & Planned Relative Relative Frequency Trequency Ratio Trequency Ratio 8 8 hase Case 3* Day A0T/5 DCs 2.12E 04 2.0BE 04 0.05 LOOP 2.29E 04 0.06 -(10 day Outage)8 LOCA/ LOOP 1.10E 09 0.08 - taan 2 7 Day A07/5 DCs 2.15E 04 2.12E 04 0.08 LOOP 2.35E 04 0.08 (7 day Outage)8 LOCA/ LOOP 1.10E 09 'O.10 came 1

7. Day ACT/6 DCs 2.0?E 04 2.02E 04 0.08 LOOP 2.00E 04 0.05 (0 day)8 LOCA/

100P 7.43E 10 0.13 PRA reflects frequency for Unit 1 only, whereas reliability considers frequency for both units 8Dur: tion of outage for planned maintenance. sA0T Risk Imvel/Non A0T Risk Level. I \\ ACT Study -9 September 11, 1989 l"

y 2. REVIEW OF THE RISK ANALYSIS RESULTS j 2.1 - ceneral I After having. invested some preliminary efforts to review the ACT study, j BNL selected the risk analysis approach and its associated calculations and j t j L results to be the focus of our review efforts, I There were several reasons to choose this particular focused approach, j These are as follows: I

1. As was mentioned in the introduction, the unavailability modelling of i

the DC and diesel fuel transfer systems of the DCPRA were already f under review by PNL (see Part 2) and therefore relevant computer software was already available for further calculations to be carried out in a timely fashion. A substantial in depth review of the results obtained by the reliability, analysis would have required audit calculations of the diesel fault trees practically starting from scratch. l

2. The reliability analysis assumes four hour mission times for the 5DG configuration (the PRA model assumes six hours for nonseismic and 24 hours for seismic events) and two hour mission times for the 6DG configuration. The use of different mission times prevents the direct comparison of the results obtained for SDG and 6DG configurations.

F

3. The reliability analysis did not address seismic effects.

4. The reliability analysis approach did not address or evaluate the risk impact of the scheduled maintenance on the swing diesel.

5. The results ef the reliability analysis, although numerically different from the risk analysis, suppcrts the same conclusions as the risk analysis.

AUT Study 10 September 11, 1989

l r- ..s 1.2 The Review Accroach l As a first phase, the adequacy of the unavailability modelling of the DGs This was done and diesel fuel transfer system in the DCPRA was reviewed. f The detailed partly in the framework of the general review of the DCPRA. Tso main observations which j results are described in Part 2 of this report. f have to be kept in mind, however are reiterated here:

1. The diesel systen analysis in the DCPRA seems to be weak in adequately representins; the potential failure contributicas of diesel subsystems.

(The reliability analysis used a much more detailed diesel model.) 1

2. The unavailability contributions due to the overhauls of the other 1

l unit diesels and the swing diesel when one unit is at power were not taken into account..(In the case of the swing diesel, that is precisely the cause that additional risk calculations had to be performed in the ACT study.) As a second phase (Part 1 of this report), the adequacy of the risk (core damage frequency) impact calculations due to changes in ACT and system redundancy were scrutinized takin6 into account comments 1 and 2 above. This phase consisted of the following steps: J t

a. A review of the quantities which determine the total unavailability of DGs (average total unavailability of DGs due to maintenance duration and maintenance frequency).

b. Review and sensitivity calculations on non seismic and seismic top event split fractions characterizing the unavailabilities of DGs under various boundary and ACT conditions. These top events appear in the support system event tree of the plant core damage frequency model.

c. !.udit and sensitivity calculations on the core damage frequencies.

} These were performed by propagating the audited or newly genera'ted top September 11. 1989 + A0T Study .. ~... ~ .,, - -... ~. -

~ 1 f j i~l* t Table 2.1 lists the ] events through the dominant sequence PRA model. definitions of various core dassge frequency calculations. New i calculations made to study sensitivity or consistency are denoted by "BN1's sensitivity calculation." They will be explaintd later. u i

d. Audit of the risk ratio results.

i 1 t The subsequent subsections and Section 3 describe these steps in detail. l t 2.3 Maintenance Dnavailability of the DCs t In the DCPRA the ACT dependency of the diesel top events appears through & quantity called total diesel maintenance unavailability, Fr. The quantity 4 [ reflects the conditions that due to Technical Specification limitations only one diesel or or e level Control Valve (LCV) of the Fuel 011 Day Tank may be in maintenance at'a time (see also Part II). 3 Thus, Py - Pas + Pm - Pac

• P e,, where Pac is the maintenance t

unavailability of the diesel itself and P m is the maintenance unavailability i f of the 14V.. Furthermore, Poe is defined as. Peo - ZKDCSD*ZMDGST, where ZMLCSD = 10.1 hours is the mean duration of the diesel maintenance; and ZMDGSP - 7.74-4 hr*1 is the mean frequency of diesel maintenance. k Similarly, Pm = EMDGh3D*ZMGNDF, where ZMcN3D - 18.9 hours is the mean duration of the 14V maintenance; and ZMGNDF - 2.03 5 hr*1 is the mean I frequency of LCV saintenance. With these values Pac - 7.817 3 and P e, - t 3.837 4, and Pr - 8.201-3. The above mean maintenance duration and frequency data are A0T. dependent j They were obtained by updating generic maintenance dar.: tion and values. frequency values using plant-specific data. These data were used in *he Bas'e Case" calculations in the ACT study. If one compares this data with those used in the reliability approach, 3 one.can observe some inconsistencies. From Table 5.7 of the A0T study one i September 11, 1989 A0T Study g- ,y- .,_c,-. ,, ~ ~, ...., _ ~ ..r*~.-., -. - -. -.. - - -., -,. - - -, -. ,r4

6 .j l L can easily obtain, by assuming a legnormal distribution, the following data l (without updating any priors): ZMDGSD' Median - 11.63 hours, Mean - 11.90 hours 2MDGSF' Median - 1.04 3 hr*1, Mean - 1.06 3 hr** l Pi,, Median - 1.21 2, Mean - 1.26 2 j r f ~ By using the previous value for P ey, one obtains a new value for the t mean total maintenance unavailability: l i Pp=Pg + Pg - 1.198 2 The main c.tuse of the inconsistency is the diesel maintenance frequency and in a lesser measure the mean maintenance duration. Consider now the generic mean priors: 2MDGSDF --17. hours 2MDGSF '- 1.03 3 hr*1 2MGN3D' - 13 hours 2MGNDf' = 2.7 5 hr*8 L One can observe that the generic mean prior maintenance frequency almost exactly coincides with the plant specific value (w/o update). Its not clear how the DCPRA arrivsd at the updated vr.lue: 2MDGSF - 7.74 4 hr*1

However, l

.the essential problem here is that Pj seems to be the correct total unavailability and this should have been used in the ' Base Case" calculations. & 1, requested additional information from PG&E about the generic prior diesel maintenance duration distribution (ZMDGSD). The distribution and its characteristic parameters are reproduced in Table I.1 of Appendix I from PG&E's ansvar.3 The mean value of that prior is: 2MDGSD' - 10.5 hours, in apparent variance with the value given in the DCPRA (see above), i-In the "Second" and

• Third Case' calculations, i.e., when e 7-day AOT is considered, PG&E increased only the rean maintenance duration of the diesels.

l l p The mean raintenance frequency of the diesel war taken to be the same, as for ACT Study 13 September 11, 4989 L 16

{h t' Jw

• 3 the 3. day A0T. For the increased value of the maan maintenance duration, PG&E l

I took 2MDGSD - 16 hours. The selection of this valueyas supported by several qualitative arguments. Among others the expert opinions of the maintenance j personnel. The arguments were also repeated in Reference 3. The Palisades l diesel outage data ware quoted as experience values, There was no refetence to any other experience data source. i Also from Table -$ 7 of the ACT study, one can easily obtain the Palisades i values-(A07 is 7. days): P . EMDGSD,.s Mean - 11.55 hours, Mean - 11.90 hours } EMDGSF,.t Median - 1.33 3 hr*1, Mean 1.36 3 hr** l P Median - 1.54 2, Mean 1.62 2 l Thus, the mean total maintenance unavailability: Pf"I -Pg+P - 1.66 2 The A0T study ures for the 7-day ACT (ZMDGSD - 16 hours): Pr

• 1 277*2e an underestination of about 30% relative to the value determined based on the i

Palisades data. In order to obtain an independent assessment for a generic mean' diesel maintenance duration, BNL used the diesel subsystem downtimes and f failure rates-collected in a recent EPRI study.' These downtimes are given in i Table 1,2 of Appendix 1 ranked in . reasing order. Based on these data and g by assuming a legnormal maintenance duration distribution, an overall mean maintenance duration value was determined (see Appendix 1). The value obtained by BNL is: 2MDGSG - 20.6 hours, Since it is considerably higher than 16 hours, one can infer that the -above value of Pg, Pr " 1277*2 indeed may underestimate the expected nean sr.intenance unavailability for a 7-day AOT. l-Because of the above ambiguities in the correct values of the mean maintenance duration and frequencie's, it was decided that besides auditing the i A0T Study

• 14-September 11, 1989

/} i

,.e e

1. ?

risk calculations of the A07 stud additional sensitivity calculations would I be performed with a bounding mean diesel maintenance duration of 24 hours and c the original mean maintenance frequency of 7.74 4 hr*3 The corresponding j range of total maintenance unavailability F,, extends from quite low values up to 2.10 a. The exact values are given in Table 2.2. The table also shows that this P range covers a mean maintenance duration range from 0 hours to ( I 17.5 hours, if for the mean saintenance frequency, the reliability analysis value, ZKDC3F' - 1.06 3 hr*8 is taken. In Table 2.1 these calculations are denoted by IC, 3A, 4C, and 6B. hl The sensitivity calculations allowed BNL to determine an unambiguous functional relationship between the total maintenance unavailability and the diesel top event split froctions, i.e., through them the core damage f frequency, i 2.4 Review of Too Event Solit Fractiona l' The DCPRA' defines six top events in the electric part of the Lupport system event tree associated with the unavailability of the diesel generators. The top event definitions, boundary conditions, success criteria, their quantified values for seismic and non seismic accident sequences, the top event split fractions, and the main contributors to the top event split fractions are thoroughly described and discussed in Part 2. For better understanding and convenience, however, the designators of the top events and their relationships with the diesels are also given here: t

• Top Event CF - Diesel Generator 13 (" Swing diesel')
• Top Event CC Diesel Cenerator 12
• Top Event CH Diesel Generator 11
• Top Event 20 Diesel Cenerator 21
• Top Event 2H - Diesel Generator 22
• Top Event Eni Units alignment of swing diesel', -13.

l l A0T Study September 11, 1989 1

't i For the audit calculations of the ACT modified top events and for BNL's 1 s own sensitivity calculations, the same SCTS code models and locally generated PC software were used which had been developed for auditing the DCFRA results. Tables 2.5 and 2.7 of Part 2 show the detailed comparison of the results of the audit calculations with those of the DCPRA and the A0T study in the " Base Case" for (3 day A0T, 10.1 hours mean maintenance duration) nen. seismic and seismic split fractions. l The final results of these calculations are also listed in Tables 2.3.A and 2.3.B of this section for non seismic and seismic split fractions, respectively. These tables also contain the results of the audit calculations i for the 'Second Case * (7 day A07,16 hour mean maintenance duration) and of the BNL's sensitivity calculation (7. day A07, 24 hour mean maintenance duration). For comparison, the tables conveniently also list the values given by PG&E in Table 4.3 of the A0T study.1 In order to check the int ernal consistency of the results obtained, the l various split fractions can te plotted against the Total Maintenance Unavailability, Pr. This functional representation is convenient because it allows us to interpret the results when one e,nsiders a mean diesel l naintenance frequency other than the 7.74 4 hr** offered by PG&E. Figure 2.1 shows such a functional representation for the non seismic top event split fractions GF1, GG3, CH6, 2GA, 2HG. Jihe graph of these split fractions appears to be a straight line. Its extrapolation to Pr - O provides a quite accurate graphica?. checking of the corresponding PG&E value given for

• 2ero Diesel Maintenance" calculations in Table 4.3 of the A0T study.1 (Similar
• graphical" spot checking
• validated" other 'zero maintenance" split fractions as well.)

The split fractions shown in Figure 2.1 essentially represent the unavailability of the individual diesel units in the DCPR/ when all the support systems are av611able (see Part II). U September 11. 1989 A0T Study .-,-n,-- w,- -.., a ~sr - ~. e

~. i pi e J l .e i For comparison, the unavailabilities of DC11 and DC13 determined by the 'more detailed

• reliability analysis calculations are also plotted as a function of the maintenance duration. Its not clear why the PRA f

unavailabilities are larger (about a factor of 2) than those obtained with the reliability acdel, where the support system unavailabilities were not taken to be zero.8 The results of BNt 's audit calculation on those top event split fractions (non seismic and seismic) which had to be completely requantified to account for the condition when the swing diesel is unavailable, are shown in Table 2.4 (More specifically, the unscheduled maintenance duration of the other diesels given the swing diesel is inoperable is set equal to eight hours. This is based on Technical Specification 3.8.1.1 Action Statement f.) For comparison Table 2.4 also indicates the original PG&E values. One can observe that there is a general agreement between the two calculations. Summarizing, (disregarding the discrepancits. previously identified between the results of the risk and reliability approaches in Section 2.1 and the factor of two from just above) one can say that there is an overall The asreament between the SNL audit results and PG&E' split fraction data. small inconsistencies appearing here or there are presumably the consequences of the fact that BNL used point estimates, while PG&E used a Monte Carlo approach in the split fraction quantificction. ) ? 17-September 11, 1989 A07 Study w- -.-c.,_.m

~ 1 i i l r.. 1 1 r. m.i.s. i w o 4. i n p .r.3.o - g. y m, -. s m,. g m .. i i a. a,,r .,- m., p n o i .y .i 4 '..p.

a

( ~~j. ' i, 1 ao. _ l- ,s, e r -o. as i. a,.... 47.. e. g-s is, t r- ,.g ,g,,. q .~- .2 9] { 4 ~1 o 4 ,4 ; ,.u l .r, ,,,,. a 3y, upa - a.. 4 l -1 - e., ,y,,

a

,,. 3 g i;, a.g 7, riu , g, ) ,(. i. gr d . i; - ,. 4 y 1 ' u v u .a.m. t o e , + n. a e .4 a..y + o e - ~ w 6.0 + + ".,c ,,..., 4, f _,,,;,, qq ,, g ,3 . r,,.: c,. _. u _,. p., ,3, y p, .., a.v 4, ww ru o ) qur tvem na me rmo.r.w g,y ,, j m r.. .2 4. 6 i 0 + r,t. a,

1. g

., y,, ,,,,....u.,.. ..q,,mem.ui. ,,, t f ,.#. +, oi, m.a . r,. m y c,. .u,,,., _ tv ,,,., y,; mio,y,...w,m,n n., gg ,n . n r >.h...yiqiitwi wgwo is h' Mi:r. qmi e ci ni qu.- .., iii..f, ,,,,a y,,im..n.i., ..,i, ,2, 4.- j w % 4 q..,. m xx.mn mc s.oe.%an, n, u,#,:,.n,w + 5,0 w m cim e sra i i m w mmew a m j m m"> 4 4 m1 w - eia a3...+ m il ml Fcl6% jr ratt r-mf * '" f_E -. WW t s4 j .. jay,

.r.g.f ng.

g,.. J' 0 pl. e e: s. 4 ,;p -1 af " eP-z ui at." 5.,1 {irjAf sigith- : i j. i-, 3 ,.m g., gp cir

e

.c i i. .i ,[4

npp y

esso a6 ru e mnrTm 4., pp L, n u. s e 53 P jf th % er. 5 i, er i.. - Sr a )J. , it ;: .: l L g e, y ,r. 4 r,.. w .-m u,%c .c r l i ' .i ' p. , '@ T-1 w tiHP;WX.+.f.I]pq i% Y $1 u 8 r n . (N r, Ni EM,,. jr t. ei " .i-ra. 1- ~; ~ 4,, v3. - ib Sh c'"iMT *h % 1ij

r. C

.J r ? o - "r F F r +4a + i "8 =. 40 a ^ mx+m - W }..,

i-p i

3., n., .": e o i-o 'in... Mi-7i t'M b t ' r m + n kii y I.e.si I.% 4

r r

i n i i >> ! l- -l-t- ,1 t- '1r o 6 i 4 i r r, F t r.-vi r o h S '..d i-la-L-b ) r. a ,1 ,.m .. - o e m ~ r i m rm . ~ m e e m w d" M 1 i-e m. P P'* ", p . 4.' -(F h 'i p n i ., p.3,32.r 8' 6 1' T r., s, i.., e r r.+ i r

, v.

c r,., -t. ,, p .c -i-a,

y. :

.um t 1 e i r.r. n 3 i. y:s s 3 T. p i.ir r o j.. ej $ 4. ,j .j-r&- e,.- apr,: byr: -vt Jr .c r.V i-fr 3 's ? w 5 49 ." i - r,d. ..r w i . n att - t" cr- >1 r' v. 4..

1e y.

.T N: ;i 1 - a - c., ,g. y. ..i mi. .a e r .,r. L r- ~. I r. =- a- .i e9,

(

cf. gi 1' i ..c . y. r . j Q {JM m i-rie n m i

r ii n Gm ( '.

e s: r : p. .at - .- j e t .i.: c.s v. a r-r /; j,. y i n i ..n . ic i sp. L ,+ t. si J.t. r ,e ,.y sn r. t >. v. -9 u., r.4 n ? :-- m s.m t :7 i.. i ~

d e

.- J.i .s i-m m .4 g, c,f ri. 2 4- _- 91J e.A n ei r a. i .J .eg r .3 s .c ir

a. -

[. eJ n I i . E'n n, hic i; ji. ic f: en e- . s. i. %di, j f l r f n 65 i 8 Plw .I - ~==e* a 4 e. [., y;;g.W. > 41 4 L .): r t i t'" 9Y r v.cn. e e q...:,. ti .r a-- gh.. M

g. b.

we ...e u.t. , a.. r n ,: r, -q p 7 . r ; v.. t.q r. .. ; - m.. m,. .c pg c.y y a. r,i,. _.. w rm s.m 7 n < + ; r, mm e. .2 m + 1 4 m .m n .a o r r i 4 i; i. .v. yn n.... ,,W,i 4,, u ,q a '.I 4. 4* 5.pt,,, w..i. . n r i. iy .r, E J I'lj g .e 9 1 2 9 l'] -].- 4.. d' ' i f. l., 1 -{ ..g 1,., L t

u n

.y e a . t-u,.. 4, n 6 i u. t

' i r

I c l' _ cli "I- '-i t.b4 big t r r: midaPkh,Wma" RAW T M a n- .e i 4 i.t: 't h i.- 1 p 1ti. 0.00 1/10* 2d' W

i. &

.jt.- f I l 0.0 10.1 16.0 24D (he) HAINTENANCE DURATION Figure 2.1 Diesel generator unavailability (Top Event Split Fractions: GF1, GG3, CH6. 2GA, 2HG in DCPRA) as a function of Total Maintenance The dashed lines (DG11 and.D313) at the bottom of Univailability. the figure and the data point (DG11, DG13) represent the results obtained from the diesel generator reliability study for the present (5 DGs) and the planned (6 DCs) diesel configurations, respectively. 18-September 11, 1989 A07 Study O s --.m m m ,v-w...n,- ..c,...,%- ,-w--

i gj 1 i i 9/ ) Table 2.1 Definition of Calculations - Risk Analysis Approach With One Unit at Power Overhaul I Allowed Mean DG Dverhaul Period of Outage Maintenance Feriod of DGs of the Nudb;r of Number Time, A0T Duration, MKD Swing Diesel Other Unit l Calculation of DGs (Days) (Hrs) (Days) (Days) Notes l 5 5 0 0 0 Audited by BNL. 1A* 5 3 10.1 0 0 Audited by BNL. 1B 5 7 16 0 0 Audited by BNL.- l 10 5 7 24 0 0 BNL's sensitivity calculation. 2 5 3 10.1 10 0 Audited by BNL. i y. 3 5 7 16 7 0 Audited by BNL. 3A 5 7 24 7 0 BNL's sensitivity calculation. 4A 6 O O O BNL's sensitivity calculation. 4B 6 3 10.1 0 0 BNL's sensitivity calculation. 4 6 7 16 0 0 Audited by BNL. 4C 6 7 24 0 0 BNL's sensitivity calculation. i 6A 5 3 10.1 1 year 0 BNL's sensitivity calculation. 6 5 7 f16 1 year 0 Audited by BNL, 68 5 7 24 1 year 0 BNL's sensitivity calculation. CDCPRA assumptions.

• • N) DC maintensnee.

l. j.' l 19 September 11, 1989 ACT Study l l~ ( - -. ~..

c. 7. o -- I t Table 2.2 1 Total Maintenance Unavailability, Pr i Mean DG Mean DG . Maintenance P, With DG Main. Maintenance Pi, With DG Main-l

Duration, tenance Frequency,

Duration, tenance Frequency.

l ZMDGSD, hr ZMDGST = 7.74 4 hr*1 ZKDGSD' hr ZMDGSF' - 1.06 3 hr*1 l 0 3.837-4 0 3.837 4 l 10.1 8.201-3 16' 1.277 2 11.9 1.298 2 24 1.896 2 17.5 1.896 2 i i I t t 'I t i + A0T Study -20 September 11, 1989 a

1 I o. Table 2.3.A Nonseismic conditional Split tractions for the Diesel Generators 10.1 Hours MMD 16 Hours MKD 24 Hours MMD for all Diesels for all Diesels for all Diesels j Top Split Event Traction PG&E BNL PG&E BNL PG6E BNL 5.606 2 GF GF1 4.523 2 4.571 2 4.946 2 5.010 2 5.592 2 CG GG1 4.477 2 4.527 2 4.909 2 4.976 2 5.843 2 CG2 5.561-2 5.474 2 5.682 2 5.649 2 5.606 2-CG3 4.523 2 4.571 2 4.946 2 5.010 2 5.583 2 CH CH1 4.436 2 4.490 2 4.878 2 4.948 2 5.732 2 CH2 5.408-2 5.322 2 5.545 2 5.516 2 7.641 2 CH3 6.265 2 8.097 2 8.063 2 7.873 2 5.592 2 GH4 4.477 2 4.527 2 4.909 2 4.976 2 5.'843 2 CH5 5.561-2 5.474 2 5.682 2 5.649 2 5.606 2 CH6 4.523 2 4.571 2 4.946 2 5.010 2 5.576 2 2G 2G1 4,396 2 4.453 2 4.847 2 4.921 2 5.702 2 2G2 5.364 2 5.271 2 5.507 2 5.475 2 6.211 2 2G3 6.250 2 6.246 2 6.254 2 6.229-2 2.493 1 2G4 2.898 1 2.910 1 2.726 1 2.711-1 5.583 2 2G5 4.436 2 4.490 2 4.878 2 4.948 2 5.732 2 2G6 5.408 2 5.322 2 5.545 2 5.516 2 7.641 2 2G7 8.265 2 8.097 2 8.063-2 7.873 2 2G8 4.477 2 4.527 2 4.909 2 4.'976 2 5,592 2 5.843 2 2G9 5.561 2 5.474 2 5.682 2 5.649 2 5.606 2 2GA 4.52.5 2 4.571 2 4.946 2 5.010 2 5.571-2 2H 2H1 4.356-2 4.417 2 4.817 2 4.894 2 5.675 2 2H2 5.320 2 L.219 2 1.470 2 5.434 2 6.157 2 2H3 6.206 2 6.196 2 6.205-2 6.177-2 7.031 2 2H4 6.922 2 7.003-2 6.996 2 7.017 2 7.883 1 2H5 7.729-1 8.294 1 7.521-1 8.114 1 5.576 2 2H6 4.396 2 4.453-2 4.847 2 4.921 2 5.702 2 2H7 5.364 2 5.271-2 5.507-2 5.475 2 i 6.211 2 2HB 6.250 2 6.246 2 6.254 2 6.229 2 2.493 1 2H9 2.898 1 2.910 1 2.726 1 2.711 1 5.583 2 2HA 4.436 2 4.490 2 4.878 2 4.948 2 5.732 2 2HB 5.408 2 5.322 2 5.545 2 5.516 2 7.641 2 2HG 8.265 2 8.097-2 8.063 2 7.873 2 5.592 2 2HD 4.477 2 4.527 2 4.909 2 4.976-2 5.843 2 2HE 5.561-2 5.474 2 5.682 2 5.649 2 5.606 2 2HG 4.523 2 4.571 2 4.946 2 5.010 2 21-September 11, 1989 A0T Study 9

---

..nn,,

. - - ~, - - - -. - v-

Table 2.3.5 I Seismic conditional Split Fractions for the Diesel Generators i 10.1 Hours MMD 16 Hours MKD 24 Hours MMD t for all Diesels for all Diesels for all Diesels Top Split Event Fraction PG&E BNL PG6E BNL PG&E BNL 9.382 2 GF GF1 8.510 2 8.389-2 8.721 2 8.810 2 9.347 2 GC CG1 8.417 2 8.325 2 8.654 2 8.756 2 9.726 2 CG2 9.502 2 9.085 2 9.428 2 9.375 2 9.382 2 l CG3 8.510 2 8.389 2 8.721 2 8.810 2 9.322 2 CH CH1 8.334-2 8.272 2 8.595 2 8.712 2 9.591 2 CH2 9.329 2 8.913 2 9.275 2 9.219 2 1.098 1 CH3 1.115 1 1.080 1 1.090 1 1.088 1 9.347 2 CH4 8.417 2 8.325 2 8.654 2 8.756 2 9;726 2 l GH5 9.502 2 9.085-2 9.428 2 9.375 2 9.382 2 i CH6 8.510 2 8.389 2 8.721 2 8.810 2 9.300 2 2C 2G1 8.251 2 8.221 2 8.537-2 8.670 2 + 9.532 2 202 9.244-2 8.827-2 9.205 2 9.144 2 1.014 1 2G3 1.016 1 1.794 2 9.964 2 9.956 2 1.777 1 2G4 1.903 1 1.908 1 1.851 1 1.846 1 9.322 2 205 8.334 2 8.272 2 8.595 2 8.712 2 9.591-2 206 9.329 2 8.913 2 9.275-2 9.219 2 1.098 1 2G7 1.115 1 1.080 1 1.090-1 1.088 1 9.347 2 2G8 8.417-2 8.325 2 8.654 2 8:756 2 9.726 2 2G9 9.502 2 9.085 2 9.428 2 9.375 2 9.382-2 20A 8.510 2 8.389 2 8.721 2 8.810 2 9.282 2 .2H 2H1 8.169-2 8.175-2 8.481 2 8,632 2 9.478 2 i~ 2H2 9.162 2 8.744 2 9.138 2 9.072 2 1.005 1 L 2H3 1.005 1 9.688 2 9.863 2 9.857 2 1.094 1 L 2H4 1.112 1 1.077-1 1.087-1 1.085 1 4.937 1 2H5 5.269-1 5.433 1 5.214 1 5.207 1 2H6 8.251 2 8.221 2 8.537 2 8.670 2 9.300 2 9.532-2 2H7 9.244-2 8.827-2 9.205 2 9 144 2 1.00 -1 2H8 1.016 1 1.794-2 9.964 2 9.956 2 1.777-1 2H9 1.903 1 1.908 1 1.851-1 1.846 1 9.322 2 2HA 8.334 2 8.272 2 8.595 2 8.712 2 9.591 2 2HB 9.329 2 8.913 2 9.275 2 9.219 2 1.098 1 2HG 1.115-1 1.080-1 1.090 1 1.088 1 9.347 2 2HD 8.417 2 8.325 2 8.654 2 8.756 2 9.726 2 2HE 9.502 2 9.085-2 C.428 2 9.375 2 9.382 2 2HG 8.510 2 8.389 2 8.721 2 8.810 2 1 l ~- -22 September 11. 1989 ADT Study l. r L

7n, C' 4 '4 ' i .e Table 2.4 Nonseismic and Seismic conditional Split Tractions l Scheduled Haintenance on Diesel 13* n Renamed Nonseismic. Renamed Seismic j i Top Split Split Split Event Traction Fraction PG&E BNL Traction PG&E BNL CF GF1 CFF 1.0 1.0 CFF 1.0 1.0 CC CC1 r CG2 004-4.244 2 4.393 2 GG5 8.114 2 8.218 2 j C03 l r l-CH CH1 CH2 CH7 4.324 2 4.377 2 CHA 8.064 2 8.181 2 CH3 CH8 4.784 2 4.751 2 GHB 8.685 2 8.629 2 l CH4 l CHS CH9 4.344 2 4.393-2 CH6 20 2G1 2G2 203 2GC 4.631 2 4.599-2 2GI 8.531 2 8.471 2 2G4 2G5 i 2G6 2CE 4.324-2 4.377-2 i 2G7 -2G8 209 2GA 2H 2H1 .2H2 2H3 2HI 4.585 2 4.552 2' st 2H4 2HJ 5.573 2 5.560 2 2H5 2H6 2H7 2H8 i 2H9 2HA 2HB 2HC 2HD + 2HE 2HG

• Renamed split fractions were used to evaluate conditional core damage sequences that involved saintenance of the swing DC.

The DG split fractions not listed for this case vere not needed to quantify these sequences. 23-September 11, 1989 ACT Study i

.o' l 1 l 3.. CORE'DAMACE FREQtTENCY AND RISK 1ATIO CA14tTIATIONS 3.1 Esptral i Fifty initiating event categories, including six seismic levels are quantified in the DCPRA. For the A0T study, however, only the leading sequences (contributing approximately 82% of the total cere damage frequency) were selected to be potential subjects of modification due to changss in the diesel-related top event split fractions. This subset of sequences is called the " Dominant Sequence PRA Model' in the ACT study. The omitted sequences are l taken into account by appropriate correction factors. The model consists of two parts:

1) nen seismic sequences and 2) seismic sequences. 420 leading non seismic sequences constitute 'the non seismic part' and 791 leading seismic sequences constitute "the seismic part."

The non seismic and seismic contributions to the total core damage frequency are 83.24 and 16.8%, t respectively. W e 420 non seismic and the top 200 seismic sequences are l listed in the ACT study. Each leading sequence is represented as the algebraic product of the frequency cf a single initiating event and the unavailabilities of the plant safety systems under specific boundary t conditions, or

• top event split fractions." Where appropriate, sequence.

a.pecific recovery actions are also included in the sequence. Normally, the system success probabilities (availabilities) are very close to unity and therefore can be conservatively omitted. For sequences in which this is not the case, the system success probabilities were. included to avoid over-conservatism. The DG success probabilities are included in the non-seismic part. In the seismic part, all the success probabilities are considered. 3.2 Core Damane Frecuencies Without Centribution Dua to Svinn Diesel Ov% haul For core damage frequency calculations in which there is no scheduled maintenance performed on the ving DG while a unit is at power, both non-seismic and seismic sequences -220 and 791 sequences, respectively) were used. l The BNL audit focussed on the non seismic sequences because for the seismic failures the DCPRA treated the DGs as completely correlated and because the ACT Study September 11, 1989 i seismic sequences show a practically negligible (order of -10") dependency on { f the change of the total maintenance unavailability of the diesels, i.e., A07, i f. l t The core damage frequency (according to the terminology of the ACT study, the absolute risk) was evaluated by propagating the top event split fractions determined with various mean diesel maintenance times through the dominant r sequence PRA model. This was done for both diesel configurations; for 5DG and [ t, [ 6DG systems. To represent the 6DG configuration, the swing diesel was ( t modelled as always being aligned to Unit 1. This was accomplished by setting r I the swing diesel alignment top event split fraction SW always to 0. This is an acceptable modelling approach. I In order for BNL to check the internal consistency of the calculations and to express the core damage frequency as a function of the total diesel maintenance unavailability (i.e., A0T), sensitivity and consistency runs were done, in addition to the audit computations, f The results obtained are shown in Table 3.1 along with those obtained by PG&t. The logically connected calculations are grouped together for the 5DG and 6DC configurations. (These are: SDGs Calculations No.5, LA, 1B, and 1C, and 6DGs-Calculations No.4A, 4B, 4, and 40). Figure 3.1 shows the core damage frequency as a function of the total maintenance unavailability F, for the SDG and 6DG configurations. One can observe that the functional correlation between the CDF and the total maintenance unavailability can be fairly approximated by straight lines. The lines for SDC and 6DG configurations run (almost) parallel, showing that under any reasonable A0T condition the 6DG configuration always provides smaller risk than the SDG configuration. A0T Study 25-September 11. 1989 .,, ~,, ,r

na 4 o i l Core baumee Trecuencies Plth Contribution Due to Evine Diesel Overhaul 3.3 For the calculations where maintenance of the swing diesel is considered The (Calculations 2, 3, and 3A) the quantification process is different. calculations are based on the conditional core damage frequency calculations When-the swing diesel is considered to be down for one year; i.e., when top event CF is set to 1.0 (CFF), and the modified and renamed top events of Table 2.4 are used. (The ID numbers of these calculations are: 6A, 6, and 6B.) These latter calculations are rather intricate and complex, especially the Some numerical values and interpretation of the variables were' seismic parts. not provided in the ACT study; BNL received them more recently as supplemental information.' calculations 2, 3, and 3A essentially contain the sum of two terms; the first one is the CDF without scheduled maintenance and the second is the conditional CDF aultiplied by the fraction of time the swing diesel is in scheduled maintenance. The results obtained from the above calculations are also listed in Table 3.1 along with the original PG&E data. The conditional core damage frequency if the swing diesel is down for a year ($DO configuration) is also plotted as a function of the total maintenance unavailability, P,, at the bottom part of Figure 3.1. The curve reflects a strong linear dependency. Comparing the results obtained by PG&E and BNL associated with the swing diesel overhaul (Ct.lculations 2, 3, and 3A) one observes that:

a. By changing the ACT from three to seven days (from Calculations 2 and 3)-PG&E calculated a risk increase of about 1.3%, Vhile BNL obtained a These correspond to a mean diesel maintenance' risk increase of 1.4%.

I frequency of 7.74 4 hr-1,

b. If one takes for the diesel maintenance frequency the value used for the reliability calculation, i.e., 1.06 3 hr*1, and considers the l

26-September 11, 1989 A07 Study l'

4 9 i results of BN1.'s Calculations 3 and 3A which characterize the ACT change, the risk increase would be less than 2.84. 3.4 Risk Ratio Calculat19.D1 This section compares the results of the risk ratios obtained by BNL with j those calculated by PC E. Since the risk ratios are defined differently for unscheduled and scheduled maintenances, they are discussed in the following i I two subsections. l 3.4.1 Risk Ration for t)nacheduled Maintenance h The risk ratio for unscheduled maintenances is defined by the formula: CCDFg3 R,,,,. gg,* c gp CDFON where, MND is the mean maintenance duration of a DG, BP is the base period with no DG maintenance (i.e., average interval between DG outages, CCDF is the conditional core damage frequency when the swing diesel is 33 assumed to be down for a year (in Table 3.*1, Calculations 6A, 6, r and 68), and CDFOM is the core damage frequency when there is no maintenance of any of the DGs (in Table 3.1, calculation 5). 1 The RR values obtained by PG&E for the SDG and 6DG configurations are listed in the column PC&E of Table 3.2.A. These values were obtained by using The base the same base period for both the 5D0 and the 6DG configurations. f' ' period was determined by the DC maintenance frequency, 7.74 4 hr*1 Per unit basis', it was assumed that the frequency of one of three DGs being out for r The maintenance is three times the individual DC maintenance frequency. interval between DG maintenance outages is then the inverse of this value. The ratio CCDF /CDFOM was also treated to be the same for SDG and 6DG 33 configurations. 27 September 11, 1989 1 A07 Study r m --,,c --,,_,e

P t r P By comparing the PG&E RR values with each other, one notices that while l 'there is an increase in the relative risk when the A07 changes from three days j l to seven days for the SDG configuration; the relative risk does not decrease increases the system redundancy from SDGs to if one keeps the A0T the same but In other words, the PG&E calculation does not indicate any advantage of 6DCs. installing the 6th DG. i According to BNL, the cause of this discrepancy is that PG&E used an ' incorrect base period for the $DG configuration. BNL presumed that whenever a' t dedicated diesel is put into unscheduled maintenance at Unit 2, the swing thus from the point of view of Unit 1 diesel will be assigned to thac unit, (Both units are assumed to be operating.) the swing diesel has an outage. Thus, on a per unit basis, the frequency of one of three DGs being out for l maintenance is five times the individual DG maintenance frequency (the sw'ing Of course, in the case of 6DGs (three dedicated DGs per diesel counts thres). unit) the PG&E reasoning is correct. BNL performed two relatI:7 r!sk calculations. In the first one, the DG maintenance frequency was assumed to be 7.74 4 hr*8 corrected by the maintenance frequencies of the LCVs. In the second one, the DG maintenance frequency was calculated by using the Diablo Canyon outage data (Table 5 7 of hly corresponds to a DG maintenance frequency of 2 the A0T ' study ). This rou5 -1.06-3 hr'1 The length of base periods used and the obtained RR results are listed in The results show a risk ratio increase of the columns "BNL" of Table 3.2. A. I about a factor of two higher than the increase obtained by PG&E when the ACT For the same time periods, the BNL changes from three days to seven days. results correctly reflect the expected decrease of the risk ratio when the redundancy of the system increases (5DGs to 6). In other words, the BNL calculations definitely indicate the advantage of the installation of the 6th DG. 28 September 11, 1989 AOT Study ..-m.

't i I 'e i 3.4.2 Risk Ration for Scheduled Maintenantes The risk ratio for scheduled maintenance is defined by the formula: f DF SCHD 33 i ~ E~

• CDF R

s.m. l where, SCHD is the scheduled outage duration (10 days for 3 day ACT and 7 days for 7. day ACT), j RP is the period between scheduled maintenances of the swing diesel (i.e., the refueling period, 1.5 years), l CCDFu is the conditional core damage frequency when the swing diesel is i assumed to be down for a year (in Table 3.1, Calculations 6A, 6, i and 68), and CDF is the core damage frequency calculated with various mean maintenance durations (in Table 3.1, Calculaticns 1A, 1B, and 10). The results of the BNL calculations are shown in Table 3.2.B along with 1 those of PC&E. There is an overall agreement between the two sets of data, f Notice that the risk ratio for the 6DC configuration is zero. There is no scheduled maintenance during operation, hence, by de(inition RR,,,, - O. [ 1' 1 l l A0T Study -29 September 11, 1989 s .m. , ~...

5 l e m 't g !:~~ 9 _I I 1 I t t' l' t1 I I f i

-~4- '--

l no i/ i i si g.1 f i I UnU I e E. s 3 i vf n 5 /1 4 I i jF% 4 r-s i I f 4 i n .r' r s 8 i e i 1 e i f 1 i i , m' d i e 1 4 8 f I e I S J L.31 1 I I a i 1 e T 4 i 2F! I I t f I - i i 6 ~ ) .T, *

• I i

n e i t 6 lL;F8 I i e i i r-i e 1 4 i i f i i i e I 4 J 2F t .l. i 4 f t i 4 4, LP t i -e I f "i l. f l I L t glj t .{ ,F/7 t t j.. 46 a ( 4 ..i - 4 6 f 7 .t ~. L q = p. aj j e; e. g -1 - V 6,.s y e P 4 e e 4c i f i. gh i 4 . i s. i e i 4 .u . i a e g .4 27 . t. j i- -e i - -e f i 4 i e ii. T. n.- ..i. g x v i .r.,. i ~,.f t.i i r i g :. 4 f. ,m .t' 1-1 n. !)r" l _ -1 'jg f ,t' .i f 1 1 2' 1 i e i I .t / i 1 i i i e. f im s g e t i-i i i / / .e t 8 4ar o t a j s i e _r i i i -t F J l. a-a + 1 2 w... i f i , e h_ i f s .w .f i f. nF. i e + t.

__f, 6, f./_

i e i i 4 t a-i f K4 / I ,f "+"P / i t-I e i 4 i e i i .i - / i i $d /4 I e / 1 I i 4 4 i 4 o. i f. i

f i

'e. i i e i f i i r .e ,i 4 e 6-4. i i e 5 ' e e 1 1 i e 4 2 +-i l l -l l 40 l yI'chrthivat riidran)y;)o . 'i. E !!, DG, L' uh #ihOcn t. l Mff! i-t--t-i i = 'r' / Mh -r- . s 4 i. .i / T . T f--t-i 4 4. l. 4 44 X'i- -r ..y i. i i i;A

• i

. g i i i i -?-r c7 =, i r 4-t,9e i.i. V.00 g..I d#

2. = id TOTAL MAfffTENANCE UNAVAllAg!LITY, P -

p Core damage frequency as a function of total maintenance (diesel L - Figure 3.1 plus level control valve) unavailability for the present (5 DGs) and the planned (6 DGs) diesel configurations. Sottom curve: _ conditional core damage frequency for the^present configuration, if the swing diesel is down for a year. I. o (- A0T Study 30 September 11, 1989 i i

• r i

(~ 3 l. l

ecn I Table 3.1 Core Damage Trequencies (Based on the Risk Analysis Approach) l Vith One Unit at Power Overhaul Allowed Mean DG Overhaul Period of Outage Maintenance Period of DGs of the Number of Nun;ber Time, A07 Duration, MMD Swing Diesel other Unit CDF (Yr*h Calculation of DGs (Days) (Hrs) (Days) (Days) PG&E BNL 5 5 0 0 0 2.042 4 1.971 4 1A 5 3 10.1 0 0 2.078 4 2.080 4 la 5 7 16 0 0 2.120 4 2.125 4 i 2.186 4 j ic 5 7 24 0 0 2 5 3 10.1 10 0 2.124 4 2.130 4 3 5 7 16 7 0 2.152 4 2.160 4 J 3A 5 7 24 7 0 2.221 4 1.898 4 4A 6 0 0 0 1.990 4 48 6 3 10.1 0 0 4 6 7 16 0 0 2.017 4 2.027 4 2.078 4 4C 6 7 24 0 0 4.812 4 6A 5 3 10.1 1 year 0 6. 5 7 16 1 year 0 4.650 4. 4.857 4 4.919 4 6B 5 7 24 1 year 0

DG maintenance.

A0T Study 31-September 11, 1989 -. ~ -.. -. - _ _,,,,,,

.e e ^~ e; Table 3.2.A ' Risk Ratio Results - Unscheduled DC Maintenance RR - Risk of Core Damage During Mean Maintenance Duration / Risk of Core Damege During Base Period W Defioition: i Maintenance s J BNL PG&E Mean 14ngth Length Length DC

A0T, Maintenance of Base of Base of Base Configuration Days Duration Hrs Period, Hrs RR,,,,

Period *, Mrs RR,,,, Period. Mrs-RR,,,, 5 Dcs 3 10.1 448.0 .05 245.5 .10 188 .15 l 11,9 4 7 16.0 448.0 .08 245.5 .16 184.9 .24 17.5 245.5 .25 7 24.0 409.2 .06 i 6 DCs 3 10.1 312 .09 l 11,9 ) 7 16.0 448.0 .08 409.2 .09 369.9 .12 17.5

• 409.2

.15 7 24.0 1 I l

• Maintenance frequencies of 14Vs, re included.

a i I 1 'i I i g, '~N +nmi

• 'T" r-ee.,m,

%E ,,7 4

a f l Table 3.2.8 Risk Ratio Results Scheduled DG Maintenance De finitio',i: RR = Risk of Core Damage During Scheduled Outages / Risk of Core Damage Between Refuelings (1.5 Years) DG Maintenance Policy Between Refuelings Scheduled Mean Outage DG Maintenance

Duration, PG6E BNL Configuration A0T, Days Duration, Hrs Days RR,,,, RR...,

Remarks-5 DGs 3 10.1 10 .041 .042 7 16.0 7 .028 .029 .029 7 24.0 7 P 0.0 0.0 No scheduled DG 6 DGs 7 16.0 outage is planned during unit operation. i l i ~ 1 ~. ' l v 1 i ACT Study 33 September 11, 1989

e G \\ i 4. CONCLDs10ns i The review BNL performed a thorough review of the PG&E A0T study. focused on the risk analysis approach. I 4 .The review identified some problematic spots in the analysis: J e The diesel top event split fractions 2G and 2H do not include the unavailability contribution of the overhauls of Unit 2 diesels PG&E performed conditional core damage' performed with Unit 1 at power. calculations when the swing diesel is considered to be down and also (as sensitivity calculations) when the dedicated Unit 1 diesels are There are no calculations as to what is the conditional core down. damage frequency if Unit 2 diesels are down (i.e., when top events 2G 1 or 2H are set to 1). 1

• The ACT analysis as well as the DCPRA are tacit about the coupling of the swing diesel when a dedicated diesel undergoes unscheduled maintenance with both units at power.

For Unit 1, the swing diesel is unavailable if it is coupled to Unit 2 while a dedicated Unit 2 diesel is in maintenance.

• The. risk analysis uses a low value for the maintenance frequency of the diesels. This means that the absolute risks are underestimated at a l ven A0T. With more realistic maintenance frequencies, the correct i

risk values for the present and suggested A0Ts lie around the risk values obtained with the low maintenance frequency and mean maintenance times of 36 hours and 24 hours, respectively. The BNL review found an overall agreement between the top event split i' fraction values obtained by BNL and PG&E, The small inconsistencies appearing sporadically are presumably due to the fact that BNL used point estimates. while PG6E used a Monte Carlo approach in the split fraction quantification. l , There is also an overall agreement between the BNL and 'PG6E core damage There is a slight frequency values (disregarding the 'no maintenance" base). tendency that the BNL CDF values lie somewhat higher than those of PG&E. 34-September 11, 1989 A0T Study e s -,s. ,,.,,., _ _. _ _.. ~. _.,., _ .,_m_..,. ,mm.

os; '2: a f n' Ofo'e ]I- ~ q.- 4 BNL concurs with PG&E's findings that: Le The risk reduction effect of a adding the sixth DC is greater than the effect of changing to a seven day A0T. This is demonstrated by Figure 3.1, which shows that the CDT curve for the 6DG configuration always l - runs below and almost parallel with the CDF curve for 5 DG 'l configuration.

• The effect cm risk of changing from a three day co a seven day A0T. is insignificant, en the order of 2 to 34..

(The curves in Figure 3.1 provide practical tools to evaluate risk changes for any combinstions of diesel maintenance. duration and frequency values.)

• The increase of the risk associated with a seven day A0T over a three day A0T performing scheduled maintenance on the swir.g diesel is also insignificant; less than 2.84.

The risk...los determined by PG6E for. - this' case are in agreement with those obtained by BNL for both 5DG and 6DG confi urations. 5 .BNL found that the risk ratios associated with unscheduled diesel maintenance are higher by a factor of 2 or 3 in absolute value than the values determined _by.PG&E for the SDG configuration for any A0Ts. The risk ratio increase associated with changing the A0T from three days to seven days was also-found to be a factor of two higher' than that of PG&E. -a In contrast with the ff.nding of the A0T study,1 BKL's risk retio - calculations definitely indicate the advantage of the installation of = the sixth ~DG. ii A0T Study September 11, 1989 ,e.

w j a: i =e).- t m FERENCES'

1. Diesel Generator Allowed outage Time study, Pacific Cas & Electric Company, May 1989.

' 2. Final report on the Diablo Canyon ung Ters Seismic Program, Pacific Cas and Electric Co., Diablo Canyon Power Plant, Docket Nos. 50-275 and 50 323, July'198P. - 3. Diesel Generator A' lowed Outage Time Study, Supplemental Information on DG Maintenance Durati m, PC&E l4tter to NRC, July 18, 1989. - 4. Diese1~ Generator Allowed Outage Time Study Supplemental Information Regarding the Equations Presented in Appendix B, p.B 15, September 5, 1989.'

5. Lofgren, E.F. and F. Varcolik, 'Probabilistic Approaches to LCO's and r

Surveillance Requirements for Standby Safety Systems," NUREG/CR 3082, November 1983.

6. Cho,' N.E., et ' a1., ' Analysis of Allowed Outage Times at the Byron Generating Station," NUREG/CR-4404, June 1986.

. 7. Driscoli, C.D., et al., " Surveillance Monitorin5 and Diagnostic Techniques to-Improve Diesel Generator Reliability," EPRI NP 5924, July 1988.

8. Worrel, R.B. and D.W. Stack, "A SETS User's Manual for the Fault Tree Analyst," Sandia National Laboratories, NUREG/CR 0475, November 1978.

4 i -36 September 11, 1969 .A0T Study 6 m-m

.m_-. m_.___ .,_,__g .q, 1 > 3 .4 .c, e c v. k. 4 4 ,e t,- i 1 Ewwamengis qsewwemsegast escwstawwesinw&' Estit; t i i l: APPENDIX I

l. :

DATA ON MAINTENANCE DURATION OF DGs AND DG SUBSYSTEMS ~ pr#1:wmsssWimeven w a: s?@W' < ~empit#MR w#s 's 3 'aganemaewsi'" ; ' me< g g I i i C: ) 1 i .J i 1 i A0T Study September 11, 1989 I e 6

bi, a o- ,4,;' d, r + F Eb Table I.1 Discretized Prior For C neric Maintenance Duration Distribution of DCs Having 72 Hour A0T* t i . Maintenance. r Duration Cumulative (Hours) Probability-Probability t 2.15+0 2.12 1 2.12-1 i l 4.24+0 2.20 1 4.31 1 - 6.48+0 5.63 2 4.88 1 j _7.25+0 2.54-2 5.13-1 7.75+0 2.37-2 5.37 1 8.25+0 2.22-2 5.59 l 8.75+0 2.07 2 5.30 1 9.15+0 1.18 2 5.92-1 9.45+0 1.13 2 6.03-1 ,n 9.80+0~ 1.44-2 6.17 1 1.01+1 1.04-2 6.2'-1 1.04+1 9.97-3

6. 's B 1

- 1.0B+1 1.27-2 6.50 1 1.12+1-1.50 2 6.65 1 1.17+1 1.41-2 6.80 1 i 1.22+1-1.33 2 6,93-1 1.27+1 1.25 2 7.05-1 ? 1.40+1 4.34 2 7.49 1 1.64+1 4.97-2 7.98-1 2.61+1 2.02 1.00+0 y Percentiles titan. -11h 1Q1h .ilith 10.5 hours .507 6.85 23.7 u

• From1* Supplemental Information to Diesel Gcnerator A0T Study," PG&E Letter.8 Data provided to1PG&E by PL&G. Mosleh, A.,

et al., *A Data Base'for Probabilistic Risk Assessment of 1MRs," Pickard, lLowe and Carrick, Inc. ' PLG - ij 0500, 1987. r + >AOT Study September 11, 1989 .--~6

~. ~ 4 -r Table I.2 f Diesel Subsystem Downtimes Ranked in ' Decreasing Order and Subsysten Failure Rates

• Mean Failure Expected Downtime l

Rate x 10*8 Mean Downtime . Rank, _ Major Sub. F (Failures / Per Failure F: xD x 10*8 i i i -i ' system Failure Diesel Mth) D (Hours) Hrs / Diesel Mth l i 1 Engine, Mechanical 1.7. 308 530 i 2 Turbocharger 2.3 82.6 190 j4 3; Coolant Pumps, Motors & Associated Electrical 1.3 58.4 75.9 4 Lubricating 011 Contamin-ation, Clogged Filters. 1.0 50.3 50.3 p 5 Cenerator, Mechanical & Electrical 3.3 43.5 142 'I 6 Air Motor Mechanical 1.4 26.9 37.7 -l 7 Coolant Leakage 3.1 26.8 83.0 8 Exhaust System 1.1 22.0 24.2 t -9 011 Laakage 1.9 20.0 38.0 10 Start Air Leakage 1.6 18.6 29.8 1 11 -Electric Start .54 17.6 9.6 12 ' Control & Instrumente-tion Switches, Relays . and Wiring 3.2 15.1 48.3 l 13 Start Air Signal 1.9 13.5 25.7 14 Covernor Setpoint & .l Synchronizing Errors 1.9 12.4 23.6

15

-Fuel Leakage 1.8 12.0 21.6 16 Voltage Regulator 3.0 10.8 32.4 17-Lubr.'.cating oil Mise,1- -l laneous-1.4 10.8 15.1 ( 18 Protective Trips 2.4 9.5 22.8 19= Start Air - Moisture, Rust & Contam!. nation 2.1 9.4 19.7 e 20 Governor 011 1.8 9.3 16.7 L 21 -Injectors, Engine Fuel 1.4 9.3 13.0 [ 22 'Covernor Sensing & Centrol 3.7 9.2 34.0 L. -23 011 Pumps, Prelube & i Associated Electrical .63 9.0 5.7 L 24 , Fuel Water, Air &. Contamination 1.9 8.5 16.2 25 Tachometer 1.5 8.4 12.6 26 Covernor-Hydraulic / Air Booster, Servomechanism & Linkage 2.5 7.5 18.8 27 Coolant Heat Exchanger, Radiator 1.0 7.4 7.4 28: Load Sequencing Timers 3.8 6.5 23.4 A0T Study September 11, 1989 f R l

y _g e ' l,;- W 8 .q l 3,/ Table 1.2 (Continued); l Mean' Failure Expected Downtime + Rate x 10*8 Mean Downtime

Rank, Major Sub-F (Failures /

Per Failure Fg x Di x 10** 3 1 system Failure -Diesel Mth) D (Hours) Mrs/ Diesel Mth i L. 29-Start Air Valve Electri-cal & Mechanical-2.2 6.5 13.7 30 -Start Air Compressor & Miscellaneous .63 6.0 3.8 31- -Fuel Transfer Pumps 6-Associated Instrumenta-tien & Electrical. '1.3 5.1 6.6 32. Control Electric Power 1.1 4.8 5.3 33 Cooling Miscellaneous 1.1 4.3 4.7 34 . Output Breaker Associated- ' Circuitry 6 Control '1.9 3.1-5.9 sums IF - 63.4 [FDg g - 1607.5 g s 1 .i overall Mean Downtime, Per, Failure. D-[FD/[F - 25.35 3g g 1 i Assumine tornormal' Downtime Distribution Maximum Likelihood p - 2.6203' Maximum-Likelihood a' -.8137 Overall Median Downtime Per Failure 13.74 Overall Mean Downtime Per Fai?ure-20.64 ~Sch Percentile. 3.57 95th Percentile 60.59-

• From Driscoli, G.D., et al., " Surveillance, Monitoring, and Diagnostic -

Techniques to Improve Diesel Generator Reliability," EPRI NP-5924, July 1988. p ~ ACT Study -40 September 11, 1989 i i

, = +, -c- <c.: - p y J t h nj f.g ( 'S j Sjh!'

, e.,h~

? 'Ij h,$. '- " 'b5NEi hhj*Mii[:y.ffh'[.fEfhi ' h I" g f PART 2 r REVIEW RESULTS OF THE SYSTEM ANALYSIS IN THE DCPRA: DIESEL GENERATOR AND DIESEL FUEL TRANSFER SYSTEMS h55#LY$$>$NE$$$$$65S?$$$?$$l?hl$$$$$$bE$$$$$$$$$$$$$$?$$$$$$[$1ll$$$?N;$ 9 i A0T Study September 11, 1989 l 9

e s } ~k I j '~ -] 0,. + ' 1. INTRODUCTION 1 ~ ' 1.1 : Obier.tives The main objective of this letter report is to summarize the results, to .. date, of reviewing the unavailability analysis of the Diesel Generator and Diesel Fuel Transfer Systems described in the DCPRA.1 The review was carried out with special attentiot. to the details of the unavailability modelling of-the maintenance activities on the DGs. (This particular emphasis was ptompted by a request of the Pacific Gas and Electric Co to change the Allowed Outage Time (A0T) of the DGs irom the present outage of three days to seven days, and 2 the fact that the study supporting this request derived data on expected core damage frequency changes based mainly on the DCPRA.) This report reflects BNL's current understanding of the subject systems and as such must be considered interim results. Final results will be pro"ided in the NUREG/CR to be issued at the end of the project. That will reflect, at that time, any additional supporting input submitt.ed by PG&E as well as any direct feedback on these prelimincry findings. 1.2 Ornanization of the Reeort Section 2 provides condensed descriptions about the configurations and functions cf the Diesel Generator and the Diesel Fuel Oil Transfer Systems. It also describes the dependancy of these systems on support equipment, the surveillance and acintenance conditions, the unavailability modelling in the j DCPRA, and the original PRA results. The purpose of'*his approach is to present the reader stand alone documentation to whic the review's findings can be directly compared. Section 3 contains the results of the BNL review and presents the current preliminary find!.ngs. For coepleteness, the ranked cut sets of hardware unavailabilities (both independent and total) obtained by BNL for varfous diesel configurations are given in Appendix A. -1 August 2, 1989 e +. ..m-.-r e-, -.-. - - - -, - .-.--r-- ,-----.e_e.-.

nJ J

t d. 3 4 2* UNAVAILABILITY MODELLING OF THE DIESEL CENERATOR AND DIES"L FUEL OIL TRANSFER SYSTEMS 4 l 2.1 L Riesel Cenerator System Descrietion. Confieurations and Functions The Diesel Cenerator System at the Diablo Canyon plant consists of five dies'el generators: two dedicated to Unit 1, two dedicated to Ur.it 2, and one (a ~

• swing diesel *) shared between the two units. According to the DCFSAR,8 the individual diesel generator units are isolated from each other and from other equipment.- The swing diesel is physically located in Unit 1.

Each diesel generator suppi.ies power to its associated 4.16kV vital bus (H. C, and F - Units-1 and 2). In the event si a les of electrical power from the main generator (due to'a unit trip, a safeguard signal or a loss of voltage on a vital bus) the vital 4.16kV buses are automatic 011y disconnected from the main genarator and transferred to the offsite standby source. (The Unit 1 main generator provides power through auxiliary transformer 12. The standby power is provided through startup transformers 11 and 12.) If this transfer is unsuccessful or the standby power is unavailable, the diesel generators must start and provide power to the affected buses. The diesel generators start on undervcitage signals from their respective buses, load onto those busas (the output breakers are normally open), initiate reloading of the vital loads and continue deliverir.g power at normal frequency to the buses. A safety injection actuation s1 nal (SIS) from either Train A or B of the SSP System 5 will also start the diesels (Train A will start :11 and 13, Train B will start 11 and 12). The swing diesel (13) may supply power to either Unit 1 or Unit 2 vital Bus F. It will start with an under /oltage or an SI signal from either unit (SSPS Train A). Because the output is not shared simultaneourly by the units, L only one of. its two circuit breakers is closed at a time. The breakers have ' ind'vidual sets of control and protection circuits. If one of the units L receives an SI signal (earlier than the other), it is ; bien priority of using l the swing diesel. L l l-2- August 2, 1989 i 'P.

.-. 1 - y s The DCs are 2750 kW, 18 cylinder, vee configuration, ALCO made units, p Each unit consists of a self cont.ined diesel engine directly connectec' to an i alternating current generator,.Each ditsel has dual train electrical atartins circuits - and-air system with turbocharger, ventilation,' fuel oil system self-6 contained radiator cooled jacket cooling water system, lube oil system, and speed control governor system. i e Each independent starting circuit has its own de power source (DG11; de panels 13, 12. DG12; de panels.12, 11. DG13; de panels 11, 13. DG21; dc panels 22, 21. DG22; de panels 23, 22). The operating control circuit is common. Without control power a unit keeps running. A mechanical trip handle, located.in the diesel compartment serves to shut the unit down.

• The air start systei., consists of swo trains.

Each train includes a compressor, a dryer, an air receiver and two air driven motors. Air from receivers is fed through regulator valves and up to the starting air system solenoid valves. Only one motor is needed to start a diesel. Power supply to the comprestor trains.are provided by 480V ac buses: (DG11; Trains A and l B; 1H, 1G. DG12; Trains A and 3;-1G, 3F. DG13; Trains A and B; 1F (backup 2F),1H (backup 2F). DG21;.Traint A and B; 2G, 2F., DG22; Trains A nnd B; 2H, 2G.) One solenoid centrol valve of an air driven motor in each compressor train gets its "open" signal from the normal control, the other solenoid valve. receives signs.1 from the backup control. Upon initiation of a start, the solencid valves open supplying air to the motors. After initiation, pressure switches located on the discharge of the jacket water si pump shuts off the air supply. The air start system supplies air to the Level Control Valvee (LCVs) of the diesel fuel oil day tanks. There is one air supply line per LCV.

• The air start system also includes an air operated turbocharger for quick starting and load pickup. The associated air subsystem consists of one o

turbo air compre er, one starting air receiver tank, and an air cryer. Two f solenoid operate. roff valves, one on each of the two supply lines, g control the air supply to the turbocharger. A solid :: tate speed loss sensor 3-August 2, 1989 E l-m

ys . g-I? ) i s: 1 L ~ ' .controls the turbo assist air supply to prevent a critical loss of speed when a sudden large ~1oad increase occurs. ] Each diesel has also another air system: the combustion air and' exhaust l + system (vencilation), containing the intake and exhaust silencers and the two motor-driven crankcase exhauster fans. ? l The engine fuel oil system involves the fuel oil day tank. Fuel oil is-l supplied by the Diesel Fuel Oil Trar.sfer System (see its description in ' Section 2.2). The fuel oil level in the day tank is controlled by two redundant level control valves (LCVs). Each LCV has two 480V ac control r power sources; a normal supply and a backup supply. The pover sources for LCVs associated with the primary fuel oil transfer pump (Train 02) are: 480V ac buses 10 and 20. Power sources for LCVs associated with the - secondary fuel oil transfer pumps (Trair. 01) are: 480V ac buses, 2M and 1H. The valves may be actuated also manually. l l The cooling of a' diesel unit is provided by a closed loop jacket coolir.g water system. The jacket water pump takes water from the lube oil cooler and the turbocharger aftercooler. There is a 50 gg11on expansion tank connected to the suction side of the pump. The pump discharges water through the' engine block and turbocharger to a common return line. Engine water temperature is maintained-at 170*F by a thermostatically controlled three way valve set. Overheated water is sent to a water radiator, where it is cooled by forced air (engine driven fan) taken from outside the building. h

• - The lubricating oil system consists of an oil reservoir, an engine driven pump and a heat exchanger.

The heat exchanger is cooled by the engine ~ jacket cooling water system. Lubricating oil temperature is thermostatically controlled. The oil is kept in the range of 90'-110'F circulated by a small pre circulation pump even if the generator is idle, to reduce wear during the engine start period. The diesel automatically stops if the oil pressure drops below 40 psig. L .c I l -4 August 2, 1989 !? . a r +w-- a- .,. + v-,--=, .---v-..- -,--,-rw----=v___-_rev-

R3rT n2 a y; +)J '91 ) b

. tc. -,

p( e To contro11 theifuel delivery and ~ therefore, the engine's speed snd generator: t 'l foutput frequency. to a. predetermined.value~ an engine governor speed control ([b 1 unit is used. The governor hes el'ectrical and mechanical controls; both of ~ 1 p;/'

which act through's hydraulic -actuator to control the ; fuel supply.

k tl

q. n The diesels.cannot respond-to a start signal'under the following V'

conditions:'

1. Shutdown relay tripped.

2. Manual-test condition.

3.-Low fuel level in the day tank. 7

4. : Low pressure in both starting air receivers.

5. Loss of de control power.

6. Voltage regulator on manual.

N y Thesaventual problems of the diesels are annunciated by various alarms (14 groups of sigrals) in the control room. L The' loads of the diesels are listed in Table 2.1. Each diesel has enough ' capacity t:a handle some extra startup load. The loading of the diesels during the recirculation phase of a LOCA is under the control of the operator. 1:

Each generator-compe'rtment is provided'with an automatic flooding co, gas li Lsystem for fireJprote

tion.

2.2, Diesel Fuel 011 Transfer Evstem. Confinuration and Function l l .The diesel fuel oil transfer system maintains a supply of fuel oil to each.DC day tank'from two large underground storage tanks (capacity: 40,000 l - gallons per tank). It contains two trains (01 and 02), each having a rotary screw e.ype positive displacement cump. These pumps are self priming. A single pump has enough capacity ;, gpm at 50 psig) to' supply all the five diesels.: (The fuel consumption r, a is about 3.2 gpm per DC). Each pump train has a fuel oil distribution header supplying all five of the DGs. 5-Augusc 2, 1989 t e -e,---=-y,-,-y-, -c.- e, u m

^ ^ ^ a.- 4 3 Q c. r.' j Mancal crosst!e valving between headers allow either transfer pump to deliver to either header. Also, it is possible to pump from either of the storage v tanks. j Local cot.trols for the system are located at each DC. There are two sets of controls; one for punp train 01 and another for pump train 02. These are the LCV switches: a total of 10 (5x2). Each switch starts the transfer pumps and opens the LCV of its respective train. The purep start levels are different: 252 gallons for train 01 and 271 gallons for train 02. Ones a pump is started it will remain running until shut down by the operator. If all the LCVs are closed (the day tanks are full) the feel oil will recirculate back to j the main storage tank. The motors that drive the pumps are powered by 480V vital ac buses (pump-train 01 by either 14s 1H or 2H, from Units 1 and 2 respeccively, and pump -train 02 by either bus 1G or 20), A manual transfer switch determines the alignment, the only criterion for alignment is that the pumps should be powered by different units. The operation of the oil fuel transfer system is,made on a demand basis: when one of the day tanks reaches a low level set point, the fuel transfer pumps start and remain running until all dieseis have been shut down. For the six hour mission time (24 hours for seismic events) of the diesels, the fuel transfer systes must remain functioning to replenish the fuel supply to each running diesel. The minimum total storage in the storage tanks is sufficient for seven days of power generation. The importance of the operability of the fuel oil transfer system for the plant safety is obvious: if the fuel transfer system is unavailable, it results in failure of all the DCs of both units Unit 1 and Unit 2. For . events when both ac powered fuel transfer pumps might become unavails.ble, a dudicated portable fuel oil driven pump is kept at hand. This pump takes suction'directly from the main storage tank and connects to one of the fuel 6-August 2, 1987 w . m-e w --,...., _. ..s-,,. ,w-.

a< j n >: ;4! idelivery headera. F2exible heses are used to make the. appropriate . connections.: i 2.3 Ton Event Definitions. Success criteria Asscciated with the unavailability of the diesel generators, the DCPRA defines six. top events in the electric part of the support system event tree. The designators of these top events and their relationships with the diesels are:

• Top Event GF - Diesel Generator 13 (" swing diesel")

e Top Event GG - Diesel Generator 12

• Top Event GH - Diesel Generator 11 e Top Event 20 - Diesel Generator 21
• Top Event 2H - Diesel Ganerator 22 1
• Top Event SW - Units alignment of the swing diesel 13 If' the offsite grid is available (top event OG in the support system event tree is successful) only the "C" events (GF, GG, CH) ara questioned in

.the support systems event tree. If the offsite grid, fails, all the five top l events are questioned. The boundary conditions of these top events depend on the status of the preceding diesel generators in the event tree. Thus, top 1 h event GF has only one boundary condition (GF1) corresponding to the case when L all support is svallable. GG has three boundary conditions (GG1, when GF ,e succeeded; GG2, when GF failed; and CG3, when GF was bypassed, i.e., not L l demanded). Similarly GH has 6, 2G has 10, and 2H has 15 boundary conditions. Top event SW has four boundary conditions: one for LOCAs; one for LOOPS, when an equal number of diesels are operating at Unit 1 and Unit 2; and two for ~ LOOPS, when an unequal number of diesels are op$ rating at the two units. l only one top event is defined in the DCPRA for the support system event tree associated with the diesel fuel oil transfer system. The designator of It is evaluated for six boundary conditions, depending this top event is: LO. r August 2, 1989 j 1 0

n, n ~'

C*L (j

• .i

on the availability of 480V ac buses at both of the units (i.e.,10, 2H. 20, and 1H).

q t;. The success criteria of the above top events are described in Table 2.1. The Technical Specification requirements with respect to the operability of the' associated systems are 21so indicated. Lonic Model f the Diesels and Diesel Fuel 011 Transfer System Denendenev 2.4 o on Other Suonort Systems The generie reliability block diagram for the diesel generators is shown in Figure 2.1. The diagram is constructed from blocks (supercomponents) of the DC system. The boundaries of the supercomponents (for instance: CH 1, CH-2A, CH 2B) are indicated in Figures 2.2 through 2.9. Notice, that the equipment boundaries for each of the diesels start with the diesel generator and include the output breaker, the fuel oil day tank, the day tank level control valves. and the undervoltage and transfer control relays. The diesel' starting air system was not-modelled separately because it was included as part of the diesel start failure data. The reliability block diagram shows the dependencies on the super-components of the plant (ac and de) electrical systems. The reliability block diagram for the diesel fuel oil transfer system (Top Event, FO) is presented in Figure 7.10. The boundaries of the pump train blocks are indicated in Figure 2.11. The reliability block diagram shows also the system dependencies on other supercomponents of the plant (ac and de) electrical systems. 2.5 Cuantification of Too Event Solit Fractions L The definitions of the boundary conditions and the associated split fractions for top events associated with the DC system are listed in Table p 8-August 2, 1989 i l.

. :$'l [ 1 (4) 1 "ha} O 2*3. Table 2.4 presents a sfeilar IIst for the diesel fuel oil transftr system (Top Event, LD). l Table 2.5 presents the values of diesel generator related top event split fraction values quantified by PG6E. Notice, that to provide better train wisc dependency tracking in the event tree model, the split fractions are expressed in terms of unavailabilities of various diesel state combinations (conditional split fractions, CSF). The arithmetic is explained in the DCPRA, Chapter D.2.1.5. The table presencs also the total unavailability value (TTL) used in the calculation of each CSF, along with the main contributors to the total unavailabilities, such as hardware (HW), maintenance (MN)._ test (TS), and human error (HE). At a given boundary condition the hardware contribution relates to the normal alignment, when no test or maintenance activities are being performed. To provide complete information, the table also indicates the two constituent parts of the hardware contribution to the unavailability: the independent (HWI) and the dependent (HVD)^(i.e., common cause) failures of the supercomponents of the diesels. The maintenance contribution is a significant contributor to the total unavailability. The DCPRA assumes that, due to Tecnnical Specification p limitations, only one diesel or level control valve may be in maintenance at a time. The following relevant quantities are used in the maintenance unavailability quantification: i a Diesel maintenance frequency, ZMDGSF: 7.74 4/hr (Mean Value). Variance - 2.33 8, 5th Percentile - f.25-4, Median - 7.52-4, 95th Percentile - 9.66 4. . Diesel maintenance duration, ZMGSD: 1.01+1 hr (Mean Value). Variance - 3.99, 5th Percentsle - 6.65, Median - 9.74, 95th Percentile - 13.3. Lcvel control valve maintenance frequency, ZMGNDF: 2.03-5/hr (Hean -Value). Variance - 3.52-11, 5th Percentile - 1.14-5, Median - 1.91-5, 95th Percentile - 2.97-5. . August 2, 1989 s t

-e. + I

• 1 Level control valve maintenance duration, ZMCN3D:

1.89+1 hr (Hean Value). Variance - 597.0, 5th Pe,* centile - 1,54, Median = 10.', 95th Percentile = 51.3. -Notice that the total maintenance unavailability of a diesel unit is -determined by_the diesel (as defined in DCPRA) maintenance unavailability plus j the LCV maintenance unavailability. When a diesel is unavailable (not for' reason of preventive maintenance) the other diesels must be surveillance tested once within 24 hours to verify operability. The DCPRA includes the unavailability contribution due to this type of test in the maintenance unavailability (MN). The test contribution to the total unavailability is modelled in the DCPRA as to be due to the weheduled monthly surveillance tests, which include the manual test of the fuel transfer system to the diesels and the quarterly stroke test of the LCVs. There is no explicit human error contribution to the total unavailability, because human errors occurring after maintenances and tests due to leaving diesel components in misalignment are. included in the maintenance and test contributions. Table 2.6 lists the split fraction. values for the various boundary conditions of the FO top event. The table, as the previous one, details the hardware - (independent and dependent components), maintenance test and human - d i error contributions to the total unavailability values. Notice there are no explicit -test or human error contributions. All the tests on fuel oil transfer system can be performed without making the system inoperabla, human errors occurring leaving a fuel oil transfer train in misalignment after maintenance'are included in unavailability values due to maintenance. h l -10 August 2, 1989 l l ~

.j j i.

• ec 2.'6 'Ouantification of Keismie Solit Freetions for DC Too Oygnia

- The basis for detailing the seismic split fraction quantification for the DC top events is to provide insight into how the maintenance unavailability (and through it, the AOT) affects the reismic.too events and consequently the. i seismic contribution to the core damage frequency. (This particular intestigation was done as part of the parallel BNL DG A0T review as discussed in Section 1.) All diesel generator components susceptible to failures by seismic events t contribute to the diesel unavailability. The components considered to be the most vulnerable to seismic effects are the following: j Comoonent Fragility Desienator i DG Control Panel ZDGCPN DG Excitation Panel ZDGEXC E DG Radiator / Water Pump ZDGRUP ? Diesel Generator Itself ZDCSi4N By using the conditional seismic failure probabilities (" fragilities"), the DCPRA combines them into a " seismic term" denoted by SEIST. SEIST has seven values corresponding to the seven seismic levels (i.e., spectral 4 acceleration ranges) defined in the DCPRA. The sever. SEIST values were determined by the mean fragilities of the diesel. components listed in Table 6 L 44 on p.6-175 of Reference 1. l' L In order to calculate seismic split fractions, the DCPRA combines the SEIST values with the total unavailability values (TTL) coming from the K conventional hardware, maintenance, test and human failures. In the case of 1 y . seismic events, however, the DCPRA (correctly and innovatively) treats many l human failures as seismic level-6spendent; that is,.the human factor probabilities are also dependent upon the seismic level. 1 l-( L -11 August 3, 1989 I l 4 i. -.. -.. ~..

r*-c. E,- I c' s To be more specific, the human failure which affects the TTL l unavailabilities is the " failure to reestablish fuel oil transfer to day tanks - by aligning a portable fuel oil transfer pump (see also Figure 2.10) and by controlling the day tank LCVs manually;" its designator is ZHET06. For. l numerical values as a function of seismic level, see Appendix G of the DCPRA i

Table 0.1-2, transmitted recently to BNL by PG&E.'

i I By using the resultant unavailabilities (SEIST + seismic level dependent TTL) the conditional seismic split fractions were determined for each diesel top events according to the rules of the sequential diesel failure model. These split fractions ate listed as a function of the seismic level in Table i 2.7. Each value of the table has a slight A07 dependence through the maintenance contribution to the TTL component of the unavailability. i

8 4

L l' P 1. 12-August 3, 1989 p l l w e e a e ,---n-e

~. ^ t, .., n y,f= -e, l; a ;,-- -* y f ,a-l -Q j \\= a. ' k'. + l l} 483.' AC X La (125V DC1j - lV* "'h 3 l r -- 11 13 ,e 2, 2, ~ Q Q 22 23 LEVEL -- DIESEL CONTROL DIESEL GENERATOR -l GENERATOR X VALVE TRAIN CONTROL POWER s a 4

• THESE ARE THE NORMAL POWER SUPPLIES FOR THE TWO LcV TRAINS; THE BACKUPS ARE 2G AND 1H RESPECTIVELY r

s Figure 2.1.

Reliability block diagram for the diesel generators. d b k. d July 12. 1989 6 a ,-s -..w,--, v-. ..w= .re, -w m

'8 i l' ( R f P .i 't'. 1 f 1:: vi l} I I I l. lll .. t.: :. l l '. c.[.. - ',p l IiIfIIIIItil i... ilil! I 7 .?. i]ighi

l n

elI . :7'.. 1 I

1. .</ -

) i lilil un il. a... :. - 6 L e3 r' /s .I. d...;." ~;' :;. Ijh 3i,4 . q '.....,. m...'y t .F. i s.,;:h. ..,.:i- . i......u.t...,l a a.t l.;.;~. :. 1 '.<.*1.. l ir I It tst r e rr;r = s'p e..::.y.4t.gf..- 4 @ j 1 e; .. s. <. :. - t,i y .q.g.g.a

• 3

,,,Ml.;4;A.;..w.. r ' ee..... C o w :. y.:'; ~. ~L'...., l ". 'I

i

.:s t. <.s, s'

na v;- ._._.. g' I'i-T.../. ht :[... - w.8 a .N.Ke.- k . '. $.a;,.q'w..!. .. g._ v.u,. 9 W w" h. l rP. 't>/. .t .;ya,p p s . e N d.... .(l un i .. ) o . d iti u' I-31 ...m [I ] .'.4 .!.,... g s., .3.. j& g "*3 e g,/< s. 1%. w,...~ c'.r.. ~v...e. u..., -. p 3:,.! ... i;, s,y 1 ..c [,. m y p g.. ' g :,3. p. ,a, i 'ipi$yk:ii%t:' q.r s i 1 i i.,g u

, "Z

.;ef. ; :- t ..u. i i .. e.: 3.a. y,a,. p.<. cy.....: i. e I 1 v:c,y_, 3,.gl ].:.. O p g ? :/..,.:: # w.s. p a ...7.;, n :. +.., s. c.9.. n i n,. . v;t., e...a,.. ..4 .g. 4 4,. g .u s u,..... =9.... ge.4 / Io c.: w. n.s.:,. '.v v .;,. i ,.1 -.u ..,..m... s..

31... 3...e; 3

e i 4 - ...%.9.i.'s. '; 7.$. <. E','n" ilT g-W,..; - .l.. :h. :. V. i / .p .w;:.a..c 1 . m.,s

p,t:. ;;.e,.,:. s,...

^.... w 3:5. <....;.,,y..x. ....j o

Is.

a y, g. ..5,,, u .-. t.. .. y w, g =

t -. *l[

t'Tl. I;i

j il

' ), If: 3 g C4.'. ,M: _, Megs,). : c. u M I! . A. Qsh z...q;W'

s I

n ..o 1 Jg g) lI...f,l F[1 t h I ee O 15 O e I a. g . l 7... s-I..; r 9 g:,... e j i)_i coe,1 .I-1j,._. . e.,.,.g., c.,... i .;;p g i ,o .m i .J g.c.. l ',, 'a :strl I -9 fK , et;' si m E h >.l c

:;, -: _ -s

gI i

k,'i

i( i

!-

.i i illi!.I1., h..k ~ ~ l ~ ii 1- .,t .] gII %;.",. gg'-@10 .,ll f'i ib#.. ii. - i i-q h i it, lT{':;; -l na i.' e (ITJ_iJ l

_- M 3 E.gS _.

at, I 1 jill,!! u .1.I o. e 3

.,.i.t 1,.. O

/o e el ,).t. e

. s,..

i., - ....t li' 1 "...S,., 1 i It J !T m i l C 0 -(' 'l 1. .i. ---? a; tt 1

• CO 2

3 ' 's :. --a +r% t l /. 0 .2 .s x i -.. c l r.. .,.y [ .;- p 1 gig.c. e a m - -].; <.. } g.i.- 91. - ' tg, g. ,u I I r i ir .u%.. i 3-

!

' h h

41 d' 3 i:' 3 ' 1 i.. ; e 2: e %,n

3- + gLptr...itmm.nh. n. )n.Y '."e$ 1 L s )9 1, y 'H, t

p..g.o, m,yt.eo mm e

h), i .u i, i s - k l iS I N MH L. ill k nI ! f:in a m m i m e m m E }i'i bhi. 1;F'i c.1 [ IlOcll, .I

. ".#.c, I I

I i.- !nn! u m...s. m.s., w ii.. w.am en...non "..m w1i

1. i. g ui I

fl g,.. l r ,io ii .. i i m .u.. w. m.,,.....,.....s.o...u.r d h III ti I d l' j dD ( IIMi'4 hl'1Mi.vM. ii.,j@Sii & kll,]lhl.lOh!.F'11 I li k M; pinsLw!'hl,UIfill l>: iMj%gijA,@l@pQ l ql iL& p i,ulti n i e gg l cemetammmm %, iip'Idi a@M ep tSir d.ie!44 !li.' : . n. = 2............... e r Y. y 8 % !I..!:Ih.'M @ . f.. h -tj...;;y:3]$"l; 'f (j h Al o.;.w:; a r , pg.,I <--- .?. 2 . g. ;c a n ...t. g;g j c 1 fe:cha rf i 6 .;.4.g;m 3, ,~ c,! 0.+ b... M. p--,..__ p--.,. p 8 4 .. m. r, --+N, y g., .#.n L:

p. g,..e j..

k i -+l t.w= - p.a[ni r-rpA -l o i m o r -- c sl. p' -p:jgf.f,. .o g]. J..s .c 3 e

._ =

y., ..{ y [. ]i. j' in l j a n.p.o..l.p+,t,e--- 1.r.3;....qG,,l, l ! 4 ... o ,a - :. w.g._- u,. a

f. i

~ ._ Q -;1EiWi.I Qik pc3.:q.a@J* T--

ar ggj i

i :. s E :. a 3.::g p@ ~ t }{g u l s. i I L i. i.} ., - -- la,.e Ylim,..

w..

.*Jf sse-:. >,,--a. i.- W. r >4 :ti.so n.m....p. x I - n w

---

' 7. -:i:.I'i

.,e

-a jh n

O~ U." E fi L-T'T

e. s - -

--m s gi'i Il M .- g-i: .f y p J j ...c.,gm%,att..f. Q,i.

4. h qj+.i= =: =.v7.f.m p. %..

.r: t g - 1..y p in a g ty__ y .M.: e g.;p- ~ili,j g,n 3 m - > i - y i. " i_ .s g w t c.u.. a - M. W ;c ..s s. v=- r.4! I ji 't_

r. -.%v' 6.rl46 g{ 4

. r., - l... s. ,w....,... ' a. = .),. k . k-i - g _c4_.n... J[ 1 i e ss. a._.,;; !. - le...>.. 2"y g:..1. i. %p ;in d,, .5h-i. ?.x..cV. .,. :,t. .w, r,. v.. n .IJ h, ![" O. t, / '. eM,' k . gl %=g==@,. l.,;.h, sic 'i , pg g u. ..- e

g e.

• t [

.,a f--Q g, ~ ~ - - . spy,. i , rli.:..y g r I b -r w,. 1 hi 8 1 t1 rere. ' rcs t , #-o g* n i, ih P.td pii ii.! r"o eile.r +r q MP 1 4 c i l. .' 3'., hfI- ,} ~ - -- "--l,!$x=--ia!=v9 l I '~ 1 1. # * \\~{M.% l L I. I r ,d- - . $,1 i ~ ---~~' t,2. ini 1 1 [' t@it c .s _l t@ ne 3 c,~ "Ci~t'iG,], ' [% - .T n g'

- -7i

@o m a, I w ~ le _" e-a t,,3p) [ r ik(M}. 'e p,,, itll 4 ra,, EIII I l E.1 . [#.1 ,8 U ~. 'b v-i

1. 'I g h tT.-b lt_u #...

U " ::! l-t'![$1 .g. t N..~ 3 F 1: f t .y 1. hni i[~..w..- t_:p:'>b e i y pq ... :. s 4 l-i'h...b-ty.;,,:.,.. .,9, 5~.- e T': ; p:::::=-1 J_p 4.i 1 ..[ ,[,_u--- > a ( s t 3 _ a 2 0 L.'. f. '&, s.s .i,, T'"c

i..

.f:: j; l y. ~ = , a. 4-- l3 ~152 U;o n ( I

- _ =- ~ I I 1 < ; e

.i.. E; [.:v:.>..

1 r.<.. ..s. I. s.. I.../c... rm,- i g.p... T. s r- ... 4:;. :.'...., - ~. .t, m. c.: I. 4 t..u ;. ..;. 5 v. Ig, e 'it t. }L N I n 1 ] 3 a.a..a a -. .c 1,. 1 I y../..r.- ? LO a. c,......,n

t. >.

%,.) w ~ 5 &i3; @4 3- &; 3.i i 8 8 A.8G 8 ~i =- " k @-....-[ @-........! N ) a2 ep I 1 QQ:/....

r.!r$i f, 7".

. %.. c. .e 91. E C 5 .+ e @ t W ,.W ML [N.':$Tf -p ..v l 0i h:Qi E.c S.Y.!d..ED3 l)s,:m..d O;. j a '. ;:a 3 +i .ti a r. hj.-$5$.h.$5i '.% .; q-.) 3 g *, 'N EAd.m l' ~..qt y. e_..-...

• .y......, g;. e..

,..< r... ..w. p...i n... g, 2 4'3.e'h,.Mtf,.;Mt ;d.9 g I3. dJ.3,., " ? y.'. ww 1,

1. .r,s %..

..vg

. :%@......q..m.[.:e,W;.m,.i.e. " a u

..n.a $*...

4.in. T.s 4.f.5. c.s.: r, e ..v.,. ls g . $ M.N u./j r:. &r, E. S ~ c.'<*w

.; i

.?.*2 M;,.,gy/... T ?.'. 8S '..w.v.sv.i.-C.9.Q...z..7..h. .42. - $2.k V, 1n g..v"t: r A ** u m,~. . ;r l

9 V,. 9,rl.,9..,..

:.... ?.,, f.c :.

r N c:,>. 7 Jo<+N a, ': 6 y-u .;q '. -,., e. e..y..;. ; r.... 4s.. .. T. .G.., e. ..% v w.... u: ' u ...:n.,:. s. . v e e ..*.l ......ri', a. v.t "i>.a ;%;. ; il:. m.n'Ta.g. 9.,' '.., :,..

y.,'lc -

o c .... p-m .ns ..**

• 9g..

.r. 4.4,;..: w. '..o ...; n.s, m.A. 3. v.s., ,.o m*. ..:c.4.m. m.$.m.... T,. ,.? av .. s.. .g m.;.,..y....,..... . e.. ... s.c , 3. .. e w w ...,,.:.,,u, m..w..;sp:. !.0[glNU'Y ..-. . w *..m?....9.....-yw,.c. . :. ~- .....,...f.... 8 11 . r.r.., c... Ca 108 .. !*9iW{:.'wI@ , g M!

• 8
• .g f.

.p>4 i %+ g '$. V' v.1 3 r i: .-.-a-- i D "".-..- -e- --3 6.h. .% 2 M. a 'Q:C 4;NEN-EM...,/o s {@ ti .I %:4@i $g

gu. s.. c.y

! g p 1.9 .4. 1 @.;t:: l @m W,,0.. E L . S. 3 I.; e .l 1 J,, vi e.,v.. a.,..a a 3 .x. e..,....q s;. ..t. p.

.-2 3

.,u:< i.. w n y i5 ,. 6,s - 1 .. j. l . 4.t .I. i j g.I (.:..

.s e

s w~ ..p,- l ,.e l .b., - l-

9.,

"r. a l o ', %^ no p 1- '.. b . i, . U 2 "cu.?."c.L c 1:q,

g. r 7.:.t..

i., .4.<.x,.,. m. @.n i, q.,.. t=. g I, .e,w..1: t.. <.. - s., ....c . I g .. u.;.. :. _ c 1 1, ......,s...... f. ~ f,, a; a m...::. l .w.'. 4.'..;i., ;.'s.! : I' f.Il..U.

• .,h.$

I ! I f'......-s. 7, .... :..q;....t,".,.g. ? 4..-. '6 )* rr. ; '-n.... s.r c.....'. ' 3p l ) ,L w ..~, ;.7 - .e .. /. l l n.+ - .e. 4lL. . ; *v'.. ,s,, f ^ a. a c 4 _. e y g g,, :.o.... = ig is ..o 3g 2.,3- [ 8..{w.,, .... g. , *c e w. I c~- r"- gf-u a '..s..... ... p..:. n... n .. c..,. e .....:.e.a. . ; '. t... e l i -1 -n-

-6,-. t .i / s.,.,,.a .,.93'f eyeg.-.

• tf

>' N

n. '... #,

.'.,.?<.. a 9 1I t-i r 3 i

3. e..

s,.n m.. n.. o, '..l .. ~L y .w. o u. N w (..;: .5..d6.n.!.g.=.p 8 4,, i p n .. '.it.#? Y ?- . t.i.. s .:. w .r t. .t . w.-. yt.:y w:..., g' f... o .n /,O.q, 3. 4l.... .o, r ,.y

p..

e; t se- .....s.;. 1 .G : g;* t O g ' - -+ 9h Ih .,.T ik't N* 'f,,S- .I /./.,. m.. 'x.3.. N.:*%.*r

• W
• u w.:

.. g.e ?..

y.'

?..*..

z....,,.. ; ;.s. A _., m:s, . v. egeh.s.m.n.n:i.b.. ,I, ,1, y.,.. r m.... s..s.- d;.w. g e:y. g a,... - ...a...e .u m.kV.t.3.b;.s ;h.. .i... .s.._.j} p. > i. Jt. re :. e. m m.-m__ mm: o w

6...

.g 4_

,..I,t a

I .s..o 3.,j.. ( (* .s m v-n,.. + m.. c..n.: n .. r '.w.m,mc..&.. MW' i .m.. wu s ..,. ~ u.Qm t.* -a:< $w'* int-m,y%. 1. m -.v. . - t r,, s .~i,w ed

a'

. "i. _ ^ !"' 3i.- %.w. 6 a..qr4,s.t.{ ;:. .u.a.A.- t. p.,r C .wt g '.s ,...r.%,.. ;&q..,... i.;. - g n ^t

a.

• '.. !. I :W e

z y s 3: '. 'ryJ$...p,t}f f;S;.2.':*.*t " M.; /..A. f' .t . u

9. '.%.:.e?.o,. 4 4..6 6. it, T,j.g%.~.

g., n..

., G'.'y. 7 y . n.. . V'] .= l -- i, ~ ;,.c. v. n;..,r.n.a(ij p, ' ,..,t:u.s.r < *s.K 'f':ni".,h. r:e r w.f- .t Ae r e ) u o b.s e.w :.y p,,. .,m. - s. u . v.., ...pM.

• a,.

i,... .?. i m.a e 9Zb 17'.N. s,wG C. '.ylt.cd'gj.r;-?s..{.,;f. f; . t' a. h.t.le; T'a; - 4 M. i ,s.f .,' :Il 7, d

e. Qa. :., y,;,. ;h,. *4.g.3i f..gf ya se

.v.- ~, .g ,.ggy

• , f.

g ,urj, .s...... h%;;, v. ,'. ti. : n. .~.lqo.%'.:f.QL. .b m p:. .pm Q ,..nv' '... ..f ' 1,, . y.X,. p..De.n..m...:..: .,. e v. i.<.. i.. a .s. . - m. :.: ~ rtg s p g..uwn.. . > ::a 49.,.. t...~a. ...., o,... . ~. . g,. e.,.a,s x.y.. u. e. ey L

W

!g.

t. 3.

m j$ I. i u m W "-Grigg g... q E .er v.!,..- 1 -4 -m r s,ry.,. ;p 2 x Wp@@.% h. o m. ',... n L. i.:

• Q t"5 BMQ - A.. l.

n d..@W"@-.: gi ?. h a.3'.. g .8 i e e L c7

g. s..c : w

. a. .E 9 .sl i 1 4., e ,,,. g e p w.m~.1}.,,,.. w :.~ >.. nae. 1 ~. M. gI. m u r.

7. - r! '

{ 'i y

4::1 '.- C

c. D.D.w.a 99 f dt'.):-'. *W 9

1 ..**ft .,c,:..

('.

p. r. as-+ A

p..,n n

e.

n ' v..

<Y

• ; *.?. e *

.,T,. a ... d ' (, . }q .b.o ty,s c.G'4s,.yx, e... %.4c. g n 't: .;n. J.:. ( n, i. y l ... g kQl'k,n.e1[.h.. A.t.A& .*,7*... .g 5i @S.h, c.'. c . i;.. .p.? < 5P;- .g q. 7 g Ip k k. 5 ...3.,,M,'. . E.f.,I.,.. *9 i ,T I 4 w 5.c.. e ., 4. t *. 4 a'**' .J n s '8 '.l i q. ',;;.. : .a= .'.p 5 I {5 y .cr msc .m,.aan #.v.ef,d....* '..... g l 1 ~.C= .-{ g .c u. w"'r. / , 5 .f, c.<* - ;,* ?,g'.#.t., g a .g y e .3 a p./e..: t...,.. '. t. . 'r. G. k.

v. g.H: a;'l.5."..:s..y ;.

..n y .!.. /. k

• 'f

. h

• I

!Of.M,..h.! i 1 *.s. l i t.a::. - -{1,;fI. S p s.. <I i y,,. ,4A: *A <...- 3 { s T g i1 ! <y v:*.. t

• l

,1 1. g ..S 4 .p nl ?.. u, n.. y s.",. n . y,.. w.; ;. EI L[ ? A.s f,... e . y' g 1 ? vg. e :'.- t :1. B .n .'.-&' v.'U.i \\.*<&l$0.i l& - ..Vc., i D'.., .is./th. 9. :: n. Q- . f. W g. M..nt. gl.c a?.t.4,'t. l .e 2.. e r. !.. gp g.,r,y.s. s.s.:s,.a e m. a .g e .,.s : g.,t, :.. i < e Q Q' f. ^..'. M,. 'y ..l. - 5 8.h li,~m '* ~l:'n J [ E' - ui, .c .rn.

e $.o.7

c.src

• s g. W

.:s v.J

• gn 4

. :- g t,*d. %:f M:.

g.

j { g, U i% ** " '

i...

a o , _..., ~. ~ ~ ~ -

3... g*. . s l-6' 1 < 9.',. V.,.:i..i/...h , 1.* ;&,0 '. t ' p~.^ .9.. ':"' *. I s' N y*.e W ...c s .o . ~ s.. fv.s.,.....w.. /.:, . M...p2.g. >.i.. :, g c.. - w N j c. @. f 4.t.. z. . e 6< l<. s.,.., .<..m.. e,;, Er &..... o..s.... .,,c + N :.'. O f Lqi.. 4;.v .p&s'a..,.4 +2&:a,:: e . g. wA i ;,.' 4 ' Q.*=.Re.yr.'. ~ o l - * ;

• b:..,

a.. i',

d 1.Y'.1rr?,tt.j ;=,, V .I .. 'i;i Q*>' e .V.:*..

a. n. -.,'. :.

t. tr'..%r * -

i.

t...

?.:A: > s

s..,

.r s,.'.t.. d. \\ .4 4 er e-p, . c,. m.. n. m e.w. s v:,....m... .m. w .n. w... v.m..,...

m. t,.c.. u:.w.

L. g

m...

a U.4.y 4,.,.se:4.,,*..... ,.s,.w.. f,...a.,.'. ' S g.

*

..+.

, '....,,..,.. u.. t mq g_._n.. w:.>s;w. :m,..w,.,:.., O..t s, ..o . m.o 3,W.. u.g _.c:.,a. x.v.,.,,.

a..

r x .n.

u.,,,._.

y m ..y vww,.. 4.. %a.:m! P a .. u. m?$$n.MpW4 n .t y, s +. LI .--' +%q b@ N'i s u/ $ >rt N m.,!.4.r w.. ;,~... a..,2, ]r .s.4 3, = c. w. ..g .. ~,,4.,,.,.. 4. f,o,. s.

u. 4 n..

, m v r... ...t u .s .t .m.,.4 .. u. y w.~.. w%,..u.,.,w..;. s.~.n,.. ~ ...:.s m.<..s..n..a. .c. s. .. r. ~.. m <- .n -s. .y n u.~.. v~. . n.. . f N[ 5 r..,,. f,(.';5 ^f. .h '.I 8 . n..,',. %,. ' #? 4".v'.v#..r.u. m rs s. 3-8 ?.g.. 'y,,t;p a. '*J,'t ,5 1,p' e t' ' nab s' la,aq../ g '.d,;3.. ./t < p. '* g., y - 4.- .. t g* */ M 8 v.,.. d. %,ii

• f/ Wlj.<.=.

%.yi.. ...S f*. Wy; . W,. "4 m;,a g,,,,g.,, p gg,, a. .g e. w j?,. 4 a-6

. a.i.,

e e e c ' bl"i:r:.%} pj$$$';B. 'a.8... ~ 3-p?.' '.R'j.. ?,;,e

9 i

< m#* & 3 .. ~ > ~ y, ...

• s._.yy..,,;

7.,.* h.a : :.n, s... . y,; . T..:. '~. ;,.. Y .s ,?, ., x.r. e s en w o .c..c,;. e so .:s.. .i- ;* k.'T[hy?...... <. %... il',5"*1* *V7 '. 7.(.; 7.6,.@s. m ***".,*,#*5.?,N..*_?"F. F. k. 'U d-e- s. g.. wQ:. v. i. C,,. 'l- .e

• '.... t.'

....t.# O ~ s s G Q:.Y..d j.'M;.; p.' . te.. .).. fr. '.',.h' . g. '. %....b.\\ Vt. D:. E y(. 5p., m.'t,i.f.y,',* g .e ).IJ,?

• g.

%D. .. w//

j. %,..4 J'. r..n,,.,,,;..

p.2, .i,. v. g.q 'd [#.l, / ~ ,f. l'. h,y.'.,. w, o.,.. r ?, L. .a.,<Wp.,,47.s*p.. h...,.'.3.;* r";[' G ,t e. ^.. * ...,,i,c' p s% .8 T %.hl y. f a bd . 's. .st,h*. 't s M '.O.; ;.;, s A (.n ',f.. ', . 3, D , # pas ',,/ s'. .n d,. 4... \\, ]*%,p. a.o t J..,* ""'Jys :.. ..,/ 8 ' t : s, i.n.q.. .. G,..m._ q.,..*g-e..w; s.*,.

• j,',,,.:

$s. .e . e.

t.....

e ....( r 3.,,+....t.,,..

0. l o

sc.:.n ~ -.,. ' *. '~,..- '.t.

r.~ W '...<

-n m.

.s 'p '. :,:g. . ~. o. s. .c - .u - '; e *\\ : .I .* r 3-i W:... 3..:.

• e w*.- '.

x .egi. i..}k..

• ? e.....

'.... la. ~ .typ.,s,.... 4,. g.t ..p s 4 A. =,t..g.' y ' 'A

• /.

g 'i 4.y.,,7 m e t.

v.. ;

-~j.... s

.. ;., a... g.,,..-,,e._ ~: s, h; .,..e. v. s..*/,% , ?.. n .a. .,m. o.. s..t....,,... s n. .. n ,.s:n1 - ;s.. e ... e.. .s. -

• I....

.t* b

.. s.

- ) c. r.g... ', '.'.r .,. ; p......* ' = .a s. .\\. _ j.c ..,.e.. s. ..,. a i t. .. <* ;u .. y y...

.. >t

f.8 :

..r*... y .."u f.co.wi... .:4la,, i i , c.,.

f,..'

** %: '.v

. o '. **. 'ss 1 t ....e..,v.,...e.(..,: > f.',.,.. ':. '. ?.* m. 6..f...,.: +.c .i.j,m,*%. e* \\;. 1 ' c. 4,. . * *f.;s.r!.l.s' e,t %'y .) d..,s..h..m.. l

.m I O.,4 .\\ 3 . 4'. - 1 ls - 1 ' '4 ?.'. .=

n..

- J.N l 'N.'.. '. de " f' .eM.8e.r. 8. 8,*,.*.'.,2* g.*. *

• th,,

'~ h M . " ' y,.'.n se,, '. r or g a.',; 3 c=- t ,( j n; y i e. g. b ' L;.*.6M'2h>.h* R. ) { O O '56%'*WM;u,g ".[4 e I a .O_ > m.th.,='*W:

. *.,, )

w..e.

• • 90W.ber...'..'+ t e. aQ,. r.

Kc r o.A sis i t' .. * '*.h.$j,s;.h.',(*. Nek. / I .h, g$ 4

• ..a'.

.... f= y=. w e t. .' y

• a

.e ~. \\. 4(g Q. ',.,s ',s., 4 O': elf a' c ll! W .t =r-t 'E 6 . 4 e,-),, <, 87.u.n.m.::.,.,:...c..... + ..:.t.

• y.,:. g s

e

• 91 :

-q.u.. w

r. r..

o ,.M. >. dTd %.% lJ27s'd6'i W !i h... ss - ,da.:.'.c...s t 1*

s 6 :s*.t yr'! i;i'.th o' o, v.

• ' N;&

r im s.m d s nW /. "Va. N.";L, .,, O kiffi '. i 1. . g i,3.,,;..1g.. b..,. e..,... tu,g,. 4 n I,E 5.3 I. S,'.h...I,.I,.I I, E,.E. .s. ess . o - n;eg_ t.,.MlP'l, ?.-^1l8. b.4.e 1 (et,.t i MQN

• %'t.h ' J 3

P. - E ';- '. J. So -J. f..o ? + G E E O O" ei 'gr[j .. ; y.., '. .? i U y u c s .y 5 dbtA,b r.<d4h j h 'I [. V jf g g $ I 9 P. 'r E E I g i g -t' 7E S ~- r;..r.s.... g s

• . f

.e 9 i,,..4 o. o g . p. -, o.g . 4,. y-4 y er .,,..p..e;q i o'go i. v, t. ,' q .. +, 3 *= c g i?. V e yg.....,b~e s .t.+., g- . w.. 3, ... u'. ..o.__ 4' i, f 8 .s. s u 1 I .t. 'w - g' g W + e

s. -.. y..

g g g g g 3 5 E a g n s

w. v. m...s.v. z. (',..

a>. g u,, y.,.. v. c,.., w 3 a a a j + g .t..I I I sc. b a,..,; n. "si . !'.a.rt.% 1.. #.; 4 'C .h.- ) k.

• .i o

%. Sh*;f}c a p A. ..v. w y h. i. u n b b b b b b b . q.p..pe.Q g i.g.gpr, gem, [ 1 . _.._.'.. _. '..g%' h _.. .N ~ s r. 4. 1r =.. o

• A D'N 'M*EC+4 T*"D.

I@ . ";if N. 6 4' f.e,d,'ft 26'$f6,,td. t f Y

py(?M..*c.L.*J._. Q

'.

• g;.sn.. v,sc s st

?. 4 i a, ?n.W.' t' ','l*. ' ! ':,'. s e v.y y l. .s **... &. s., e..p t...I,,, ., v. < g::':s 's.i. Q.,.; * ~ : ' ~. y .r. y s .., ~ s {lr T .W.mac;..r\\. e.r s g:n

,d 8

.. e.fi. w.. :.. W .gqji h.a:<: II O I v6.L. : @.i ',e ,.'e,1 ."(C,', 3 ig 4 .,G.jA,7.W.v.,,. 4, wy;.. ' g a. { Y L.. I l ' f g W-'ti. "<y M' '.......6,].. r, '...i,, e;;.v. a.. y I t ?.'. s h.,g QQs,'T 'p e N ao a $. 7. '. - ' a y '. * .s j t.eg. ...,, c: .. ya...,.. . e,. c),%';. %. m.2 w y 9, y'; .. n,'.9.. ;.', o I q.q. ... s si {il': Q m. 8

. ; f,.' ?

.and< ; .x-E 2...: ..: n ?. n n mc + a s;wg.. a y - g['.. '. ' ' '.1./ ' li E 53 d@lM.$ W. C.

• 54-

.p. c-g m. :3..W .. g ? u. h- ,:m.w. &v.. p.v

9. r.

e .,-d%p.i.m.. >.s..:. v-- e ' %... Q.4; 1.'.V p :'t.d,,j. ', 1 s. t 9 H,d." F u.i .i...d,....., M. % W4..

e.. P.

69 .t p 5 @J b .pS6.,,.e ~ E.J..M,. .N a .s . I.r.. \\ 9 r ~ ~. -..... r. wt. ' - r-i 1 ..e...,.-.,n... v....,. .a-u. s". s ~ ~" g~g t. ..;v, p.,t;.+.,; n. en.R:gr gg.. r u s.. .Li .W.dt ';.?;';p, t ;,"*.f.n i fy,c,.f:.;?.'*g..;t. : p LY:.trW

' C.C..

s .s, s % W+......;, I uvs n'ws- ~ m* t.<..i,.3. i.,t. +:: s

...s v,t..;.ke....J..?.

.?. '...-

e i u-5 ie 4 .s... ;...:,. ,E l. ,s x: <e

. r....;

l l

1 v-1 \\.

• • r i

.g. + ,c I &ey 7,* ,.9 } ~, c. .;.~.~'*4,. N s

. ';...t7.' '

u . r4:. '

.s

.n w ..y. .g ;yyi ;a a,. ...ao..t. cy e e + 6d c.t.. .ma cA _ j y, 1.W.... W. *, l -y. ...r,..,-yp'.. .....,.y.- 1..*,...,. g%N t;ny r i ..... es . v..C ; n. 3..,. u.t. O 4: 1 4 n l pe.c on /~ ~ h.e.. 4 o 7..< y

< u.. :. 9. v I:

w s :@{} s. t'W!.*.J %. ef'..f.',.J mpOw I .*/f ,h .cq's 3 r .= M *** v 9 .M.:n:.~VT.,; Q,; e.w '. ...,.. n '*j * > @o '.l,M,,. 'A S:tv\\w.vp. <4 w .i p#r .a. a. 3 deb.h..,.WM, 7 .. ;.c..

• xv. s. g. :,.'QApa. u N.* a

&s'. ,*&t s v.. q w g.n;@v-1

e. r.

6 m. y te. gqyx.:. p_s.**yg..

??g;fk}.h&'.~.P.,".

v ~ e ... t YM. ,, w,

9 :

...e .r..> non** w. i. *..,r:... ;

• 4 e.
• *1

e. q y,",.3,.v

,.4., e ?.sk,.,,,v,%e. **~,. .f. .N

• t

.%. p.,. b '.'. f,. .r?i tW.. I ,.*l, p;,

• gs, 4,. -.'

.-w cp ,..p,

r..4 y

, c.

3 7

,i ,o , y, j,7 g, '?jy y

• *
• Y..

m: E .. ;E&r.4s;r:v.s b.s( \\'n$..*,i.',W5*',i'. #.- $'* R.'j'.'E'*.) U' f,( n.'.k.).a :g.*" k;. o s.. y g.s. #.i = n.ca.>p.%. o s .'p'

• n.N.w.*;'3 e,p &y*,M/.V f v
• k.a','

s*a. t%- a ' ,,;.6 .f'. o..,.n6

• "8 M*VA -

g ,.se.e,. f..;.M, g .. d La.}i. f..,. s ,.4 :t.:w....i.; f p../..:,. ./- 1 w

y,e.. r.,.yr.f

.? .x, f,'., s.\\ o W, v;w. '.. t.,.3.g..,.,p 6,,i, 3,. s., ,.g.. g s 3,. .v. .rk.. s. ' tre !. 5. - o. h( t'.",* i*** ) .r' .* M' r"q'**3{*.l't ve'.*: i .h"s.b @'J,,q t. :f:.t.. > p.%".*. 4; 3.y*i.*l',./,. 6, w E.f w 6

• v. =**h t '

. s 21:.. e. ,.....ar.s. 4 o. \\ k.o \\ e k.:s. '. l s e.a.'v,;*o. G, ';ll g 3, 3 I j.. f.g ?. 4 e i..4 Y. f... m.w..,.,,s * .%.,p+.bN$f h p..... - .h g . t.H gre,. ' .. :.. ' 4.t.. ,.w. A .hi ,r gh "'h.. s. b . i,<'k.,\\ u,h.w - h \\ .Yl .t<..f;w@r. 4. p,l G.,,,:r:., Ag /. a o &.c. y 3..,t y, 3. y:,.w.. o.. /, ' s* 4 .s.e u $..r ~ ? f3 r.b,M'.oi... e e '.~b M.s M d

"h.dt., '

k.a *.2 4 A. u y g 1f. ?. g s.. # Ys

r. *2 7.) d.

C .M e i Jl O 4 t h . ' $f.m 2/,f.,[;. ess**,gsru,%,a.gg.ya-w. . g. g _fi$ Q. $Ag.l%W I(**p,M.Qn. ..s .... {. 3'.. 4&.<, **. 4 9/!N..,.[*6.y'*gl y'~ h

• 4

^*.e y s f e

. ;L, I) _ ~2.-".

d "M..,$a,N. .%..L',-. .e. lI . e y ;,s. g-.b. 3. g. /..n. .p< e. .y. - *n p . a.

p..

a ,v 'u*..*..*,*+*w. ,..'..i.i,'..t. .e l g a 8 4 .e si.p,.:a. -.:...m.e 5* ..f lg&n..& :.

.3 eo J

G ) f;'*0.% .A .f.s..m..s.s s,. - *., t:5,Q. e.. ' = 8 y> g g .n ? a' s s .s. i .; 3 4 r %'q's.-;k 41 %.,. w:,.n...::.... I.ru,m,. u e g .l.a q' . c.. : ..A,...r.. >s. m:. we "t;m:v.e %w.p:.n.ch".. w. s., 27;'~ + W sp 2. .e,..n .Q p.jy

.t.tr...W 2

- ic ..u:s '.,o...*.,.,..c. l-w. m 5

• 2. s..,.

.m. t. - . + n.? p .n.~ s .,.r..< ..*n t l,f ..m.1:,p..;c e.. ..,..~.s.<'. a-g a / v s M. &3 T

11151 !!!!I.. M.:

1 : E 5 i t I i E g 3 .4 ce._ :. {". 2 st l l t [ a t.* I h.' s 1 e s.I i w i, i .U r.;. + g- .: p 4.. g y ss=} s :F.. gy l i I g p.e:gi.4d? = Its T 5 l Te 8* u r g

saa i !

v. g r 1 Els. l t ) g 's. gli s y 1 g n o y I -] <r, v'tte !!!q es: u so g i i 5::q a A 1 ;2 t @V .ta.gg. I.4tt! i:: s 5 8 ? .....tl t. .i ! y I. y l 1 i l l l ^' l i l I - ' * * ~

-. =-. 4' y ah,(q... 4 M rj.-.. Q.;4 3:s ,,s'... f. .I i-i s i.g E 8 II j , v. ,.,.v. I 1 123 tr I; s i J g e .s y N f /,. /. [t s b., (,s.. a s : : g f

• 1 v i..vn c g a

O 1 .. ;a. t n a

4

,..s .d u :.: l-2 g ., s.

.v ng;

} I, gg, 7.# "'.:;;. ', ;. t l aa; j 'l et .ca l.ylg'\\ } f.[ l.4 ,I s I ~ 8, A .[1E!p, 1 l g. .e m ';c '.by.r a' w 2 g is g i [ 4?... ~ ~ i agagna, gg gI [d.1, g I.: g g + r. ; '..{.'f *'...;. *./ c E ll [I g {g k E { jggjj g gg . s. 4~.4....,y gy!!1 ! lI 5E .542 .e ' 4, c,,*.%.; +%..*:*6. tA..i;w. a f I E I EE s.a w =gD' tar.0$kg. _ NP ##*jk,b.,_ m #'3*VqA Mf.EMepr f.. 5 :, NM A _ N!* gs.s.3 . g....- ..( D hg.;y.Mg 'J )* 9...>t..' ' L; 0:..... N. 6,Qu 't,f..a. h 4CJRW ' f,..t. .) l i.'. , H. mgnym 1 5 .. w..w.a.u nr. wu,.,m...),. n. gm.ny.n '. .a m .n.s.g.

a. m.

....r. hLm w+ .z. .m. x.;w:..s.:e m ..e:n ,.~,,. r. p n. n..r & ,_ a..:r.c .m 7.,.9. w.,. - ~ w... m..,, I.. m u ., 4 . i,a y..wm. s..m+ y.,.*v.... y,.n. c. 4.. c i b,...... e n.b ll% U:..,. M..<' ?:, 7.. *l^ :..I n ,:0 ld%R9; &Rlp.ws.:., @

• g hqw

[ p.: ^ !'g ii 1 .a .f. y, or,,g. &.4,,,..r;., o g1. . g.. '... .Y g b.,.. [ y

i. J,..,1 s o.r.0 3,,"p 3,p.g t e

p., j:..r,6. ti ,,~ g,,>. g@ i me== ;,,,,,,. essa.

+ - I. ;. 4,.9-, g,...,,6;f.;m. c.

,,7 < . p.s u }. an ..o......'"...%...,- v., s. ,.7 ,. s. u 4r 4 1 y.ce:p.+sgdh.+.si.*. v4 a ut, o W.T.u a m v..g.5,{1g+4<;.w '< c. u ?. h 4 l T': e. b.i,.:.+. ;..'&. .t.. rJ Qsw4 .n.z.n,? be - .~ ~ c... a. ". y w3f. +

. g.

.u,... - m.na.c.n,. u ,s i i. n 3 G,g4,. 2..W..ll:W~.. 7....W.. F.tr"@v.v...,g m...,. o <, .. x. .e..: .. 1 ; M m,. W.. 4 r ... ec. ~ e.y,:,y. .a..,.# 4., .e-u..., ,.g.. s s., ~ Q .s .y:,.. v,... ,, ~, .. %.,.. q,. :n. t. Q .l .n.. .a

g.-@7;et... g;:.

e.

.s

....,:<t.':::...

t v1 e c.W .o.#, W;. f.c,.9.:n.v.u.. s. e .,,( e,.. y,..g(',.' n e n,

r. o

5 y,s.s (Q -

a,',,. s,.r. o 4 _s 2 .b ~... g.,,. 4. u. ,.t su. v..e. b w N W.. ., m ". .N..'I[h..sf..a.[. Q r ..s ( vv '5 \\.. k..;h.A:. g u L /ks 'h 5o 4. 3 u. t : ,c.,... M je: M.. l . :. o,W . a.:24.m...,.. - u. ..).. e v,Q. G .t d,. *E.e w. ..h ;y .... Su.g.3 .~ [ .'.Y;. ' 9,E.,(<. ' ' ' s'. { k'd.>. $.c.. Wi 6:f.,.(..,..., s. ..:.r .:.a..' - . t; i l +.... e

a. v, w....

s,,a,&e:<!.n.'nk. f. . v..< G..' J. ? Y.': ' .e, . @d.... Ov m.r...-.. r:. y ..,~...:> .4..';'oO':w.Q*.cl .v::. ~.3 wv , p\\<. '. t+?:.iun y ~ s. .i..... i' 9fe. 4,y* * :lp; ;2. e..t.;.'l.si. y.. ' pyr C.

,,-c p

4;. :.

• I g s i

W ll l

• .f..-Q-.

} v Od".M .(.a.n s.~;.. Y..... 's i * . jl p;n h '.l g.l I

s u P4

. [., .:/;,+...g..

e. ;

{yfe ...l sa:1.... '; r:.. .i h... - ?. c '. i> ^ .w.. . >e n... <? ~. *. Je! ' it l 6... t * . c.* .'. s.: Y;.! ".

3.,-.4s

. h.; w.. - .x;m. g s .; - o... .....-L s ?.? J.O .n. g. I . ao..s...s.i.!AW. :. s. o

• ..s.%.m..

.', :,.$.r. n.,v$ i,i',U..,,

} ..,: e f .s. i.*e.,.. ...a.,.

?.it %.'s bA Y

s.b

:.,x.

e. a.c. 9 .:n. rn.. *..... 1,.

• a.' 4'.yA * ";'n '. ' ' ' ' ' * * -

1 n.

_m 1 f =..i . e v.;. t i I: t. t (. t l 1, l t i ) ~ i i 440V AC { l EH OR 1H + r PUMP TAAN 01 F ) i PUMPTRANC2 U. t 480V AC { 1G OR 2G ' PORTABLE FUEL CIL TRANSFER + SYSTEM "

• BACKUP POWER SOURCE

" ThE PORTABLE FUEL CL TRANSFER SYSTEM IS EVALUATED N THE ELECTRC POWER RECOVERY MODEL t B Figure 2.10. Reliability block diagram for the diesel fuel oil transfer system (Top Event, FO). July 12, 1989

ga p s.. r w i

g p

.I'li g g I g g w N li g } } } } 0; j .~ g i =- e bIgi! i ss a, qa:d @-i-.... e a u

j

&La ES: ~ e sn s 8s II M ~ e s C. C. C. C. C. i i 2 r '9 g E N M .M M3 8 8 I I Oi l Ol g el { 01 g I o x 4 a e M h M'g M'l Q E g j 1 g c w E t - l"",; 2 l 1 -r ..a. Z u 2 d ~ o i b g A emt w C T l ~- ---.: j-3.<, g. u. g 0 E l 4 e i au r r w c e hI-h { ~,,$h f f ~2. n. f,s E, h1, a, t i ~i I: -%. i, y {

• E1*N l 5ll u

g =N ,! Nd N f h [ TH j j j Ti.i e ed i s O l'l i : y .r.- !, 3 .--_2 4 e.* g ue g e (h ! I t l (h !jt ll i I l ,, I ,g l t s- ._c si l . I.. u I w l i jg a gN ~ l l 1 g -t+-- p-3 ; 4

)

) } l I ( <T C y l NA, 5 ff k b ={ C -=~ }t g gj

}!

gg,

j

%.s c g i e o s 4 e 9 E? $~ SN F t s = ..,.._.-...w ,.--._,------.,--~,,.,cm, .. ~.., _ _.

F. " ' t e ^ t y a Tab 1w 2.1 j Sus DC Vital Safety Related Loads i T 1 3 (Swing) Centrifugal Charging Pump No.1 Safety injection Pump No.1 Containment Fan Cooler Unit No.2 i Containment ran Cooler Unit No.1 component Cooling Water Pump No.1 [ Auxiliary Saltwater Pump No.1 Auxiliary Teodwater Pump No.3 G 1 2 (2-1) Centrifugal Charging Pump No.2 Residual Heat Removal Pump No.1 Containment Fan Cooler Unit No.3 Containment Fan Cooler Unit No.5 Component Cooling Watar Pump No.2 Auxiliary Sairvater Pump No.2 Containment Spray Pump No.1 s H 1 1 (2 2) Safety Injei cion Pump No.2 l Residual Heat Removal Pump No.2 Containment Tan Cooler Unit No.4 Component Cooling Water Pump No.3 Auxiliary Feedwater Pump No.2 Containment Spray Pump No.2 .t l-1 l 1 I' August 2.1989 24 l l l c

p 1 p: l -+ Table 2.2 l Tcp Event Definittor, and Succass Criteria i Diesel Generator and Diesel Frel Transfer Systems r l Top Event j Designator Top Ev*at Definition Top Event Success Criteria i F,3 Each top event is successful i CF DG13 provides power for 3 CC DG12 6 hours (24 hours 0 if the corresponding DG starts 3 CH DG11 for seismic events) H on undervoltage signal from 3 2C DG21 to 4.16kV ac buses C, its bus, takes bus loads and 2H DC22 Ha continues powering loads for the appropriate mission times. l (Bus index numbers indicate (6 hours or 24 hours). l plant Unit No.) j t SW Swing diesel alignment. The value of SU determines ? g D013 is normally aligned to whether DG13 goes to Unit 2. Unit 1. A value of 0 indicates it does not, a value between 0 and 1 represents the probability that it does. F0 Diesel fuel oil transfer One of two pumps starts on low i system provides fuel oil for day tank level and refills each of the DCs for six hours each day tank for the period [ (24 hours for seismic events), that each diesel operates, i I ~ FSAR Success Criteria: l Any two of three DGs and their associated buses are adequate to serve the vital loads necessary for safe shutdown of a single unit (although one DG may supply power to two vital buses at the same time, no credit is currently given this mode of operation). L The diesel fuel oil transfer system must remain. operable and deliver fuel to 1 each of the DCs for ths time the DGs are required to operate. There must be enough fuel in storage tanks for seven days of power generation. Technical Soecifications: With a single DC inoperable, demonstrate the operability of the remaining ac suurces within 24 hours. Restore the diesel within 72 hours. With two DGs inoperable, demonstrate the operability of the two offsite ac circuits (one 230kV and one 500kV line) within one hour and at least once every eight hours. Restore at least two of the inoperable diesels within two hours. 25 August 2. 1989 i 4 ..n,.. ...,n..,

7,_ r. i Tabic 2.3 Boundary Condition and Split Fraction identifications for Top Events CF, CC, CH, 20, 2H, and SW Top Diesel Conditions Event Case 13 12 11 21 22 Commene.s i CF - Unavailability of DG13 under the following conditions: CF1 All support available. l i CC - Unavailability of DG12 under the following conditions: CG1 0 Offsite grid succeeded, CF succeeded. O CC2 1 Offsite grid succe6ded, CF failed. o CG3 Offsite grid succeeded, CF bypassed (not demanded) CH - Unavailability of Doll under the following conditions: CH1 0 0 Offsite grid succeeded, both CF, CC f succeeded. CH2 0 1 Offsite grid succeeded GF.S/F, CG. F/S (two possible combinations). CH3 1~ 1 offsite grid succeeded, both GF, CG failed. CH4 0 Offsite grid succeeded, CF.S/B, CG. B/S (two possible combinations). CHS 3 Offsite grid succeeded, CF.F/B, CG. [ B/F (two possible combinations). CH6 Offsita grid succeeded, both CF, CG bypassed. 2C - Unavailability of DC21 under the following conditions: 1,. e! L 2G1 0 0 0 offsite grid failed..all CF, 00, and CH succeeded. 2G2 0 0 1 offsite grid failed, two of CF, CG, i and CH succeeded, the third failed (three po sible combinations). 203 0 1 1 Offsite grid failed, two of CF, CG, CH failed, the third succeeded (three possible combinations). 2C4 1 1 1 Offsite grid failed, all CF, CC, CH L failed. 205 0 0 Offsite grid failed, two of CF, 00, CH succeeded, the third bypassed I (three possible combinations). 26 August 2. 1989

g- _. ~ ~ J.. I V i t Table 2.3 (Continued) I i } l Top Diesel Conditions j Event Case 13 12 11 21 22 Comments i 2C6 0 1 Offsite grid failed, one of CF, CC, i CH succeeded, one failes the third i bypassed (six possibls i combinations). 207 1 1 Offsite y*id failed, two of CF, CC, i i CH failed, the third bypassed I (three possible combinations). 2C8 0 Offsite grid failed, one of CF, 00, + CH succeeded, the other two I bypassed (three possible I coubinations). 2C9 1 Offsite grid failed, one of CF, CC, CH failed, the other two bypassed e (three possible combinations).2CA. -Offsite grid failed, all cf CF, CC, CH bypassed. 2H - Unavailability of DC22 under the following conditions: 2H1 0 0 0 0 offsite crid failed, all of CF, CG, i CH, 20 succeeded. 2H2 1 0 0 0 Offsite grid failed, one of CF, CC, CH, 20 failed, the other three succeeded (four possible combinations). 2H3 1 1 0 0 offsite grid failed, two of CF, CC, i CH, 2C failed, the other two succeeded (six possible combinations). 2H4 1 1 1 0 offsite grid failed, three of CF, e?' CC, CH, 20 failed, the fourth succeeded (four possible combinations). 2HS 1 1 1 1 Offsite grid failed, all of CF, CC, k CH, 20 failed. p 2H6 0 0 0 Offsite grid failsd, three of CF, CC, CH, 20 succeeded, the fourth 4 L bypassed (four possible combinations). 2H7 0 0 1 Offsite grid failed, two of CF, CG, CH, 20 succeeded, one failed, the fourth bypassed (12 possible combinations). 2H8 0 1 1 Offsite grid failed, two of CF, CC, l-CH, 20 failed, one succeeded, the i i (; 27 August 2, 1989 4 t

w v! 3 r 'f.c. i.. 4 ' fame 2.3 (Continued) ( . Top Diesel Conditions Event case 13 12 11 21 22 Comments i fourth bypassed (12 possible combinations). Offsite grid failed, three of CF, 2H9 1 1 1 Ca, CH, 20 failed, the fourth bypssed (four possible combinations). Offsite grid failed, two of CF, CC,. 2HA 0 0 CH, 20 succeeded, the other two' bypassed (six possible combinations). 1 0 offsite grid f".ned, two of CF, CC, 2HB CH, 2C bypassed, one failed, the fourth succeeded (12 possible' combinations). 1 1 Offsite grid failed, two of CF, CG, 2HC CH, 2G bypassed, the other two failed (six possible combinations). 0 Offsite gtid failed, three of CF, i 2HD CC, CH, 20 bypassed, the fourth i succeeded (four possible combinations). 1 Offsite grid failed, three of CF, 2HE CC, CH, 20 bypassed, the fourth failed (four possible combinations). I Offsite grid failed, all of CF, CG, i 2HG CH, 2S bypassed. .SW SWO 1DCA, the swing diesel locked to e the Unit 1. SW1 LDSP, with equal chance.for swin5 diesel to operate on each unit. SW2 IDSP, with more DGs aligned to Unit 2 than Unit 1. CW3 1DSP, with more DGs aligned to Unit l l 1 than Unit 2. \\ Notes: 0 - Succeeded 1 - Failed . - Bypassed i. 28-August 2, 1989 9 4 E ~

Q;7 e ?,01 I

s

.,4 l Table 2.4 Dissel Fuel Oil Transfer System Soundary Conditions for Top Event LO Split Fraction ID i F01 All support available. I F02 Support available to one train only. F03 1/2 normal support available; recover support to the other train by realignment to backup support. t F04 2/2 normal support unavailable; recover supports by L realignment to backups. F05 2/2 normal supports unavailable; recover only 1/2 backup support by realignment, j i F06 All support unavailable (guaranteed failure). l I t t t! I l i

Aubust 2, 1989 l

k ? . ~ -

. a ,g 4t . DOI t I Table 2.5 Unavailability values (conditional Split Tra etions) for the Diesel Generator System l Comment TfP Txent Case Calc. CSF TTL HW HWI HWD TS MN HE e .;r. CT1 PC&E 4.523 2 4.554 2 3.703 2 3.689 2 1.393 4 2.950 4 8.217-3 BNL 4.571 2 4.603 2 3.754 2 3.695 2 5.860 4 2.934 4 8.198 3 GG Col PC&E 4.477 2 4.554 2 BNL 4.527 2 4.603 2) as CF1 CO2 PC&E 5.561 2 2.702 3 1.749 3 1.536 3 2.129 4 4.989 5 9.025-4 { BNL 5.474 2 2.540,3 1.581 3 1.366 3 2.149 4 4.980 5 9.089 4 CC3 PC&E 4.523 2 4.554 2 I ** UII BNL 4.571 2 4.603 2 t

t CH1 PC&E 4.436 2 4.554 2) as CFt

+ BNL 4,490 2 4.603 2 i CH2 PC&E 5.408 2 2.702 3 I ** 002 BNL 5.322 2 2.540 3 Cll3 PC&E 8.265 2 2.339 ~ 1.264 4 7.438 5 5.204 5 3.173 5 7.566 5 BNL 8.097 2 2.066 4 1.034 4 5.057 5 5.284 5 3.128 5 7.194 5 i CH4 PC&E 4.477 2 4.554 2 I ** UII BNL A.527 2 4.603 2 CH5 PC&E 5.561 2 2.702 3) as CC2 BNL 5.474 2 2.540 3 CH6 PC&E 4.523 2 4.554 2) as CF1 BNL 4.571-2 4.603 2 PC 2C1 PC&E 4.396 2 4.554 2) as CF1 BNL 4.453 2 4.603 2 202 PC&E 5.364 2 2.702 3) as CC2 BNL 5.271 2 2.540 3 2C3 PC&E 6.250 2 2.339 0) as CH3 BNL 6.246 2 2.066 4 -2C4-PC&E 2.898 1 6.369-5 2.597 5 4.314 6 2.166 5-3.049 5 7.221 6 BNL 2.910 1 5.995 5 2.363-5 1.874 6 2.176 5 3.017 5 6.176 6 t. 30 August 2, 1989 4 4 -.~s..~,. ..~4 , ~,.. ~ ~, - - -

] l G Tabir 2.5'(Continued) i Comment l 7 Top HV NVI HVD TS MN HE w I Event Case Calc. CSF TTL L 2C5' PG&E 4.436 2 4.554 2 4.603 2 ** OII I BNL 4.490 2 l 2C6 PC&E 5.408 2 2.702 3) as 002 BNL 5.322 2 2.540 3 207 PC&E-8.265 2 2.339 4) as CHs. BNL 8.097 2 2.066 4 208-PC&E 4.477 2 4^.554-2) as CF1 BNL 4.527 2 4.603 2 209 PC&E 5.561 2 2.702 3) as 002 BNL 5.474 2 2.540 3 i 2CA. PC&E 4:.523 2 4.554 2) as CP1 BNL 4.571 2 4.603 2 'h < 2H1 PC&E 4.356 2 4.554 2) as cri j BNL 4.417 2 4.603 2 2H2 PC&E -5.320 2 2.702 3) as 002 BNL 5.219-2 2.540 3 2H3 PC&E 6.206 2 2.339 4) as CH3 BNL 6.19u 2 2.066 4 2h4-PC&E 6.922 2 6.369 5) as 204 BNL 7.003 2 5.995 5 c 2115 PC&E 7.729 1 5.034-5 1.951 5 3.020 7 1.820 5 3.039 5 1.436 6 BNL 8.294 1 4.975 5 1.642 5-6.950 8 ?..B36 5 3,006 5 1.272 6 2H6 PCEE 4.396 2 4.554 2 L as CF1 BNL 4.453 2 4,603 2 2H7 PC&E 5.364 2 2.702 3) as CC2 BNL 5.271 2 2.540 3 2116 PC&E 6.250 2 2.339 4) as CH3 o BNL 6.246 2 2.066 4 2h9 PCEE 2.898 1 6.369 5) as 204 BNL-2.910 1 5.995 5 August 2, 1989 3 ~--'~e- --x---.

4. v : gr t

c,. k J,.. .:.= U#' Table 2.5 (Continued) -{ ,u i . ? *P ' ? Comment

c.nt' Case Calc.

CST TTL NW HVI NWD TS MN HE e ~, 2MA PG&E 5.436 2 4.554 ?. BNL' 4.490 2 4.603 2) as CF1 '2H3 PC&E 5.408 2 2.702 3) as CO2 '? BNL 5.322 2 2.540 3 .- s p 2HC PC&E 8.265 2.2.339 4) as CH3 BNL' 8.098 2 2.066 4 e 2HD PC&E 4.477 2 4.554 2) as CT1 BNL 4.527-2 4'.603-2 { i 2HE. PC&E 5.561 2 2.702 3 1 as CC2 BNL 5.474 2 2.540 3 l 2HC PC&E 4.523-2 4.554-2 BNL 4.571 2 4.603 2) as cri $*.'O - PCEE 0.000 i BNL 0.000 f Sul PC&E 5.000 1 BNL 5.000 1 l SU2 PC&E-1.767-3 BNL 1.770 3 SW3 PC&E 9.981 1 BNL-94902-1 i e P -1 I \\ l August 2, 1989 l.

[~I l. t I i Table 2.6 Unavailability Values (Split Tractions) for the l Diesel Fuel Transfer System l 1 i 7.ip Event Case Calc. TTL HW HWI HVD TS HN HE ro foi Pc6E 2.164 4 1.919 4 1.176 5 1.802 4 0.0 2.445 5 0.0 l BKL 2.092 4 1.848 4 8.533 6 1.763 4 0.0 2.447 5 0.0 i T02 PC&E 7.040 3 3.113 3 2.933 3 1.802 4 0.0 3.930 3 0.0 BNL 7.048 3 3.097 3 2.921 3 1.763 4 0.0 3.951 3 0.0 F03 PG&E 3.509 4 1.919 4 1.176 5 1.802 4 0.0 2.445 5 0.0 BNL 3.460 4 1.848-4 8.533 6 1.763 4 0.0 2.447 5 0.0 [ F04 PC&E 2.263 2 1.919 4 1.176 5 1.802 4 0.0 2.445 5 0.0224 BNL 2.250 2 1.848 4 8.533 6 1.763 4 0.0 2.447 5 0.0223 4 F05 PC&E 5.079 2 3.113 3 2.933-3 1.802 4 0.0 3.930 3 0.0224 BNL 2.292 2 3.097 3 2.921-3 1.761 4 - 0.0 3.951 3 0.0223 F0F Pc6E 1.0 BNL 1.0 i t l 9 e j -33 August 2, 1989 l-l j l W-vs-- + * - - w-wyi-- -+e-ew-* -'-yp +---W

e ).4 Table 2.7a Conditional Split Tractions for DC Top Events as a Tunction of Seismic Level TCLE E931t Solomas Level (spectral asselerotina. g ) treetten 18 8.0*8.1 0.2 1.13 1.13 1.?S 1.78 3.0 3.0 2.8 3.S*3.0 3.0 4.0 """"F1'- ~ 8.8160E 02 8.31005 08 8.S660t=02 9.3020E-01 1.0&S01-01 1.70005-01 2.82708-01 G41 8.4k99t 92 8.4170E-03 8.41702 02 8.1670t*01 8.76708 03 1.0660E 01 1.0560t*01 Got.

9. Set 0E 08 9.5020E*02 1.01508 01 1.4310L*01 3.S?t05-01 4.D460E 01 7.3210t*01 i

653 8.3100E*02 8.5100E*t: 8.5660E*02 9.30208 0* 1 0SSDE*01 1.7000t-01 3.8270E*01

48%

0.3340t n2 4.3340E 02 8.3340t*02 8.6700t*. 9.6?00t*02 1.04308 01 1.0430E-01 ' C 48 ' 9.33998 02 9.St90E*02 9.3290t*08 9.7810E*03 9.78101-03 1.16308-01 1.16301 01 i) 1WB S - 1.11808 01 1.1180E 01 1.7450t*01 4.3300E*01 7.1:30t.01 0.7630E*01 t.S?$0E 01 l1 , Ett 4.41708 02 8.4170E-04 4.4170E 02 8.?6708 02 8.7670E*03 1.0660E 01 3.06608 01 - ?AS 9.84308 03

9. Set 0E*03 1.01SCE*61 1.43108 01 8.87001 01 4.84608-01 7.3210t*01

~ l,p. Ea6 8.8100E Da 8.51002-03 8.66605 02 9.3020t*02 1.98$05 01 1.7000E*01 8.8270E-01 ?;st 8.38108 02 8.1810E*03 8.1310E*01 8.8740E 02 8.8740C-02 1.03108 01 1.0310t*01 (St - 9.34408 48 9.3440t*03 9.3440E 02 9.6??0t*08 9.6??05-01 1.1490t*01 1.14908 01 283. - 1.01608 01 1.0160t*01 1 0160t*01 1.07408 01 1.0740t-01 1.36SDE*01 1.3450t*01 fte 1.9930E=01 1.90308 01 S.19508 01 8.S400t*01 9.S??02-01 9.St10E-01 9.94408 01 183 S.33648 08 8.3340E 02 4.33408 02 0.67005 08 8.6700t=0? 1.0430E 01 1.04308 01 tae 9.32908 02 9.8390E-04 9.33901 02 9.7810E-02 9.7410E-02 1.14305 01 1.16302 01 257 1.11508 01 '1.1160E-01 1.74SOE*01 4.2340E-01 7.1720E-01 0.7630E-01 9.S?SCE*01 238 8.41708 02 8.4170E 62 8.4170E*02 8.76708 02 8.76701 03 1.85602 01 1.0560E 01 259 9.80808 02 9.3020E*08 1.01SCE*01 1.4810t 01 2.S?00E*01 4.8460t-C1 7.3210E*01 4*4 4.81998-04 4.S100E*03 8.8660E 03 9.30201 02 1.04SOE-01 1.7000E-01 8.41708 01 441 0.16908 83 8.16908 03 0.16905 02 4.48005 02 8.4000E-03 1.01902-01 1.0190t*01 ' JSA 9.14308 03 9.16305 08 9.16208-02 9.S?t0t*02 9.8740E-02 1.1360E 01 1.1360E*01 l 183 1.90SSE*01 1.00$0E*01

1. 0050t = 0*

1.0600t*01 1.0600E*01 1.2470E-01 1.2470E*01 184' 1.11305 81 1.11305 01 1.1120E 01 1.1990E 01 1 1890E*01 1.39802 01 1.3030t*01 J J3) S.3690E-01 S.3690t*01 8.9720t.01 9.7970t*01 9.94701 01 9.9750t-01 9.9920E 01 i 286 0.28101 92 5.338tE*03 8.1310E 02 8.57401 02 8.8740E 02 1.03105-01 1.0310t*01 237 9.3440E.03 9.34408 02 9.3440t 02 9.6??Ot*01 9.6??0E 02 1 1490E-01 1.1490E 01 488 1.01608 01 1.01608 01 1.0163t*01 1.0740t*01 1.0740t.01 1.34503-01 1.2650E*01 189 1.90308 91 1.90308 01 S.1950t*01 0.84002 01 9.S??OC-01 9.0810E 01 9.9440E 01 Ita 4.33408 03 4.3540E 01 0.3340t*02 8.67005 02 8.6700E-02 1.0430E-01 1.0430E*01 25B 9.33908 02 9.32908-02 s,3390E 02 9.7010E 02 9.?$10E*01 1.1690E*01 1.1630E 01 2BC 1.11508 01 1.11508 01 1.7430E*01 4.33001 01 7.1720E-01 8.7630E-01 9.S?SOE 01 the 8.4170E-03 8.41702-02 8.4170E 02 8.7670E 02 8.76705 02 1.0S60E-01 1.0560t-01 att 9.80805-02 9.50208 02 1 01501 01 1.4810E 01 3.87005-01 4.8460E*01 7.3210E 01 2 85 9.31005-02 8.8100E 03 0.6660E*02 9.3020t.02 1.05508-01 1.7000E-01 3.8270t.01 SWD 8.0D69t*01 8.0000E*01 0.0000E 01 8.0000E-01 0.0000E-01 8.00005 01 0.0000E 01 e SW1 3.0003L*01 S.0000E*01 S.0000E-01 S.0000E*01 S.0000t*01 S.0000E*01 S.0000E-01 awa 3.?S00t*03 1.?S00E-ws 1.?S00E-03 3.0000E*02 1.0000E.02 3.0000E-02 S.0000E-02 SW3 9.9830E 01 9.9870t*G1 9.f320E*01 9.90002-01 9.9000E-01 9.8000E*01 9.S000t=01 I-1 i l l i July 25, 1989 i ,--.-vr- ,e,n--n . - - ~, c

---

s

- - - - - - - - - - - ~ - - - ,e*"

o + Mle 2.4 i conditional split TracEtons for DC Top Events as a Tunction Seisitic Level p ENL sette s.t.u. 1 e.1 c.p..sr.t m 1..it... e 1 re..is 19 6.D.0.3 0.3 1 33 1.2)*1.?S 1.?> d.0 2.0 2.8 2.S*3.0 1.D*4 ....-..--...-.......................--...--.....****----..........0 l Orl= 3."39E*02 0.*29f'-02 J.445E-02 9.19*E-v2 1.044E-01 1.694E-01 2.010E-01 20l* E.000E-02 9.* 0E-02 E<. ".2LE - 02 B.63?E-02 B.6t!?E-02 1.OO4E-01

1. O*.A E 01 C"2e 9.93*C-00 9 007E-02 9.14tE-02 1.419E-01 2.07.1E-01 4.C;1E-01

7.

• 0*ct-01 FC D.0DVE-U2 fl. % 9E-02 S.44EE-U2 9.1T*E-02 1.044E-01

~1.694E*01 2.0101-01 fMle D.070E-02 D.270E-02 8.270E-02 S.61DE-02.Es.41DE-02 1.044E-01 1.044E-01 W1 = 0.917.E-02 E.910E-02 E.910E-02 9.417E-02 9.417E-02 1.14LE-01 1.14LL-01 06t3e 1.0 DOC 01 1.CCOE-01 1.747E-01, 4.006E-01 7.250E-01.3.774E-01 .9.07CE-01 i JH4e E.000E-02 s.020E-02 8.000E-02 9.687E-02 0.607E-02 1.0L4EL01 1.054E-01 ' W.= 9.oDOE-02 9.OS7E-02 9.743E-02 1.419E-01 2.351E-01 4. 0.' 1 E -01. 7.00$C-01

• cme =

C.029[-02 C.*D9E-02 S.44LE-02 9.190E-02 1.044E-01 1.694E-01 2.91tE-01

• U1=

3.201E-02 3.201E-02 3.21 C-0.7 3.t E-02 8.05;E-02 1.0:4E-OL.

1. 0* 4E-01

• • =

8.E 7C-02 8.E 7E-02 8.B 7E-0: 9.306E-02 9.006E-02 1.100E-01 1.1*0E-01 I'J != 9.794E-02 9.794E-02 9.794E-02,, 1. 04 fpE-01,, 1.04EE-01 1.064E-01 1.264E-01

C4a 1.90LE-01 1 929E-01* 5.074E-01 8.614E-01 9.600E-01 9.02*E-01 9.944E-01

• 0 s=

D.*7 E-02 C.272E-02, R.272T. 62 S. 61 E E-02,. 9. 61 BE-02 1.044E-01. 1.044E-01.

• 06 0.910E-02 C.917E-02

6. 910E-(*

9.417E-02 9.417E-02 1 145C-01 1.140E-01 i 727e 1.030E-01 1.CCOE-01 1.747E-01

4. 0 l 'I-01 7.200E-01 C. ? ?4E-0!_P. 573E-01.

• 3=

TJ. 02.NE-02 5.020E-02

8. OLE-02 8.607E-0:

8.607E-02 1.054E-01 1.054E-01 .t7e 9.02 C+02 9.OG7L-02 9.743E-02 1.419E-01 2.051E-01 4.831E-01.. 7 009E-01. + ?.n C 70?E* 02 l'. 209E-02

• 8. 440E-02 9.19'E-02, 1.044E-01

1. 4 94 E-01 2.815E-01 i

i ..tl e D.17it-02: B.175E-02 0.170E-02 E.49 E-02 3.490E-02 1.024E-01. 1.004E-01 !*s2 S.744E-02 B.744E-02 0.744E-0 9.000E-02

9. 200E -02 1 115E-01 1.11tE-01

.:::= 9.433E-02 9.&B9E-02 9.6CEE*00 1.074E-01 1.0'4E-01 1.246E-01 1.246E-01 ? ru4e ,1.077E-01 1.077E-01 1.077E-01 ' 3.166E-01 1.166E-01 1.302E-01 1 3E2E-01 . '. ;* o b.it'E-01 5.495E-01 9.070E-01 9.810E-01 9.952E-01 9.975E-01 9.99 E-01

• H&e 3.001E-02 E.2'll1E-02 B.221E-02 E.550E-02 8.553E-02.

1.034E-01 1.034E-01 . 47a D.2'*7E-02 B. EO7E-02 3.027E-02 9.000E-02 9.006E-02 1.100E-01...1.100E-01 D 'D e S.794E-02 9.794E-02

• 9.794E-02 1.04BC-01 1.048E-01 1.264E-01 1.064E-01

.tdo 1.900E-01 1.929E-01 3.074E-01 3.614E-01

9. 60!E-01.

9. 02*E-01_3.94 4E-01.

t e4Ae C.07 E-02 3.07 E-02' e.072C-02 8.610E-02 8.618E-02 1.044E-01 1.044E-01 L1:be D.967E-02 8.91 E-02 0.9132-02 i

• l sC.

1. 000t-01' '* ~1. 08:E-01~~

• 1.'7 4 7E-01

,9. 417E-02,,,,9. 417E-02,,..1.14 0E*:01 l.14 tE-01 4.*06E-01 7.250E-01 S.774E-01 9.57BE-01 LHD. D.020E-02 3 305E-02 9.0:5E-02 8.687E-02,,9.687E-02 1.054E-01 1.054E-01. s

• e tE m 9.OBLE-02"" 9.007E-02^ 9.74BE 02 1.419E-01 2.U53E-01 4.8*1E-01 7.309E-01 l

E H0s D.0P'PE-02 0.389E-02 9.445L-62 9.19'E-02

1. 0 4 9 E uC1,. 1. 69 4 E-OL.. 2. 910E -01.-

h%.

0.000.-01~*' O.000E-Ol*~ 0.000E-01 0.000E-01 0.000E-01 0.000E-01

• 0. 000E-01 8 '

W % U*s0E-01 5.000E-01 5.000E-01

• 5.000E-01 ~ 3.000E-01 ' 5.000E-01..m.E.000E-01..

7/.4:= 1.770E-00* '1.'770E-03*"'2.770E'03 1.000E-02 1.000E-02 5.000E-02 .5.000E-0 a . w;= 9.900E-ot v.9a:E-01 9.98:E-01

9. 900E-W., 9. 900E-01,9.500E-01._.9.500E-01.

i F July 25, 1989 i t

e 3. RESULTS OF THE BNL REVIEV e s 3.1 ceneral The unsvailability modelling of the Diese\\ Co ierators and the Dissel Fuel Transfer System in the DCPRA were reviewed by LNL with special emphasis because of the following:

a. The DCs are the most important support systems; impseting the s afety of the majority of plant operations, including cold shutdown.

b. As discussed in Section 1, a request for changing the Allowed Outage Time (A0T) of the Dissel Generators was submitted to the NRC by PC&E 8

and the study supporting the request is based mainly on the DCPRA. BNL is reviewing this study in a parellel effort to this review. Therefore, to check the adequacy of tho DCPRA modelling for " system. specific" effects which may also influence granting permission for ADT changes, BNL used the following approach: BNL ompared the vendor specific (ALCO) diesel failure events with those obtained from generic diesel data. This was done to see how well the DCPRA model reflects the vendor + specific L

• experivnee" and to estimate the expected downtime distribution of the diesels. The evaluation was carried out by reviewin5 the failure modes and maintenance unavailabilities involved in the diesel model.

In order to check for calculational inconsistencies, all of the split fractions were recalculated (seismic inclusive). 3.2 Comoarison of ALCO Tvee bc Failures With All Types of DC Failures In order to ces whether the ALCO type DCs used at the Diablo Canyon power plant have some subsy: tem or component specific failure modes (and thus, somb subsystem or component specific expected downtimes) BNL compared the leading follure contributions of rubsystems and components of ALCO diesels with those of all other types of DCu, The data were taken from a recent study performed l l at Battelle on aging of diesel composents.8 Table 3.1 presents the results. 36 August 3, 1989 e

e \\ r s e' = y One can see that the instruments and Control Systen's (and within it the f,,munor's) f ailures are t.he main contributors to the generic failures of ALCO l c w els. Also vith ALCO diesals, the Cooling System and to a lesser extent l the Lubrication System seem to be more prone to failures than in the e'tal i i generic DC population. A positive feature of the ALCO diesels is that the startir.g sjstan appears to be less vulnerable to failure than the generic DG population. Finally, the ALCO fuel system does net seem to be any core prone to failures than the generic one. 3.3 Remarks on the t'navailability Modelline of the Diesels and Fuel 011 4 Innsfer Evstem in the DCPRA i

a. The systein modelling of the DCs in the DCPRA represents an elaborate sequential unavailability analysis of a *five train" system, where one train (the swing diesel) is pla)ing a special role. There is no question that the approach used is mathe:.stically appealing because it ussa the synnetry aspects of the diesel configuration and rendecs the ratults of the analysis very suitable for integration into the DCPRA. The complexity of the calculation, however, for casual readern is difficult and for eventual uses (e.g., change of A0T) is rather cumbersome.

P

b. In contrast with the systems modelling, the unavailability modelling of tae individual diesels (the fault tree modelling) was kept simplistic by using l

the standard " diesel fails to start at." run" failure modes. The diesel starting air system (i.e., air compressors, receivers, etc.) wete not modelled separately because they were considered to be included as part of the diesel start failure data. An attempt was made to display some l cruponents of the diesel subsystems in the model. This effert, however. tended to be inconsistent in that only some support failures were modelled l and inconsequential in that the modelled failures were of such low probability. For example, each supercomponent *2A" and "2B" contains the l failure rates: "DG Air Receiver Rupture During Operation: ZTTK1B - 2.66 8/hr,' " Air Check Valve Transfer Closed During Operation: ZTVCOP - 1,04 Rpt.07/Rev.1 September 11, 1989 j

'r Fa '

o 2

? 8/hr," etc. The failure contribution of the turbocharger, however, with an estimeted' failure probability of 2.73 4/d was neglected. Another example: while the diesel supercomponent boundaries indicate several subsystems as part of the supercomponents (see Tigures 2.6, 2.8, and 2.9), one cannot find any representative component failure rats contributing to tha combined unavailability of those supercomponents. Such subsystems are: the cooling, the lubrication as.d the combuscion air systems. Table 3.1 shows that the cooling system is the second largest contributor to t.he failure of the ALCO type diesels. 2

c. The following remark also has relevance in connection with the ACT study and concerns the expected downtime distribution of the diesel systems. The DCPRA models the maintenance frequency and duration of the LCVs as separate quantities frna those of the dieselr.

If the day tank and other fuel system components are included in the maintenance data of the diesel, it is not clear why the LCV is treated separately. Civen that it is treated separately, the mean and 95th percentile of the " effective" downtime distribution oi~ the diesel system would be determined by the combination of the diesel and the LCV maintenance duratien distributions (the 9sth percentile value of the LCV maintenance durations is $1.3 hours),

d. Tne DCPRA considers only unscheduled maintenances performed on Unit 2 diesels as contributing to the unavailabilities of tbs associated top 4

l. events '2C" and "2H." Unavailabilities due to large overhauls lasting over a protracted period of time performed when Unit 1 is operating and Unit 2 is in refueling (or cold shutdown) (say two times 10 to 16 days each) were not included in.he model.

e. In Table 2.6 the PG6E total split fraction value, P05 seems to be in variance with that obtained by BNL. The probable cause of the discrepancy is that the human error contribution was double counted in the DCPRA. The PG6E value is seemingly also in contradiction with the pC&E seismic values L

l' Rpt.07/Rev.1 38 September 11, 1989

e-l given at the Invest three seisnie levels in Taole 6 46 (p.6182) of [ Reference 1.

f. The detailed analysis of the fuel Oil Transfer System (see rigure D.2.1 3 f

$da st 3 of 4) contains the following item (Item No.12): "In an emergency where it is necessary to get into the fuel oil pump vault to manipulate j valves, it may take several hours to get security to open the vault." This ites renders quastionable the estimates of the human factors (among others a the value of ZHEr06 used in the diesel analysis) considered for recovery of I the ruel 011 Transfer System and through it, the recovery of electrical power. l

g. Among the DC failure related 1.ERs filed by the Diablo Canyon pow.r plant,'

there was one failure in the ruel 011 Transfer System which would affect I all the DCs. This common cause failure involved the degrtdation of the diesel oil in the underground reserve tanks caused by fungi. According to PG&E, the problem does not exist any more. However because of its I peculiarity and importance it is quoted here: 1.ER 88-14 This report is being voluntarily submitted for information purposes only as described in Item 19 of Supplement No.1 l to NUREC+1022. On May 4,1988, during performance of surveillance test procedure (SRP) M 96, ' diesel generator 24 hour load test," the e diesel generator (DC) 1 1 load decreased below the value specified in the SRP acceptance criteria. An investigation showed that a high differential pressure existed across the primary fuel oil filter. After switching to the standby primary fuel oil filter, the load a returned to the required value. An investigation determined dat l the DC day tank contained a fungus and that the first primar-Liter was clogged by fungus. The other DC day tanks also contained a fungus and fungus spores were found in the main storage tanks. The fuel oil in the day tanks waar diocided and filtered u.itil the fuel oil met the criteria of STP M 108, *di,sel fuel oil analysis " for li particulate contamination, flash point, API gravity and viscosity. l The day tanks were drained, inspected and cleaned. The bottom of main storago _ tanks 01 and 0 2 were suctioned out and a biocide was l: added. A biocide program will be developed and icplemented to inhibit the growth of fungus in the DG fuel oil storage system. Also, a sarpling and inspection program for the DC day tanks will be developed. Both will be incorporated into plant procedures, i Rpt.07/Rev.1 39-September 11, 1989 7

- = O l se h..It is not clear how the fire ruppression (CO ) system in the D3 rooms 3 responds to various levels of seismic event. The safety concern is that if an earthquake f ails the diesel units without causing fire, one or more DG rooms might be flooded with CO, and therefore rescue personnel may not be 2 able to recover the DCs within proper time intervals. W 3.4 Audit Calculations i In order to scrutinize the qusntified split fractions chemselves, BNL j performed audit calculations for each of the split fractions associated with each of the boundary conditione. The calculations were extended for both non. ) seismic (n!ssion time: 6 hours) and seismic (sission time: 24 hours) cases. Seismic calculations were not performed for the Fua) Oil Transfer System. In these audit ca'culations the same assumptions, input data, maintenance and j test frequency and duration, as well as mean fragility and human factor values ] were used as in the DCPRA. The SETS code' and locally generated PC software were us9d for the computations. The use of the SETS code allowed the identifiettio., of the most important cut sets contributing to the hardware unavailabilities. These cut sets are not readily accessible for direct review ( in the DCPRA. Appendix A lists the ranked cut sets for single, double, triple, quadruple and quintuple diesel failures. The definition of the basic events appearing in the cut sets are identical to those given in Chapter j i D.2.1.5 of the DCPRA. ? The results obtained by the audit calculations are presented in Tables 2.6 and 2.7 b for the DCs and for the Fuel 011 Transfer System, respectively. They are denoted by *BNL* to be compared with the values given in the DCPRA (denoted by "PC&E*). It has to be emphasized, that if the review of the fragilities would identify incorrect values characterizing diesel components or the use of incorrect human failure rates would be detected during the review of the human factore, complete requantification of the Table 2.7.b split fractions would be necessary. Rpt.07/Rev.1 September 11, 19Po + ~v-r- ~e-,,.a-,-. -w

-x-. 3 0 o i o t F By comparing the PC&E and &NL results one can see dhat there is an f overall agreement between the data. The agreement is even better, if one l takes into account that BNL used point estimates, while PC&E mainly used a l Honte Carlo approach in the split fraction quantification. [ 3.5 SencJ unions ? u The BNL review identified several inconsistencies and neglection of failures of diesel subsystems in the unavailability modelling of diesel j, generators in the DCPRA and the omission of the unavailability contribution from Unit 2 (and swing) diesels overhauls. The combined effect of these neglections may result in underastimation of the associated tap event split fractions and through them the expected ccre damage frequency value of Unit 1. The above remarks made in connection with the DCPRA simultaneously represent preliminary results concerning the verification of the Diesel Generator ACT study. In fact, the results of the audit calculations can also be considered as verification of the ' base case" (i.e., present ACT conditions) in the PC&E diesel ACT study,8 s r Ppc.07/Rev.1 -41* SCPtember 11, 1989

m 2 ;.. ......_ a :,:- ~ wa.,: .-r_ - wu. - - -

___m ; _

..

' o 4 i Table 3.1 -l Systees and Components Contributing Most to Tailures at All Types of DCs and at ALCO Type DCs l t Percent of Percent of Failures Systems and Components All Failures at ALCO DCs Instrument and Controls System 25 26 covernor 10 15 sensors 3 3 Relays 2 1 Startup Components 2 1 t Fuel System 11 10 Piping on Engine 3 1 Injector Pumps 2 1 Fuel 011 Pumps 5 Starting System 10 6 Controls 3 3 Starting Air Valve 2 starting Motors 2 2 Air Compressor 1 1 Switchgear System 10 10 I Breakers 3 4 Relays 5 4 Instrument and Controls 1 1 Cooling System 9 14 Pumps 2 1 ? Heat Exchangers 2 1 l l -Piping' 2 6 Lubrication System-7 8 Heat Exchangers 2 3 Pumps 2 3 Lube 011 1 Other Systems 28 26 Date Base: 1984 failure event recorded between 1974 and 1984 in Re.ference 5. l-Nuclear plants where ALCO Diesel Generators have been used in 198o: Indian Point I and 2 Power Authority of the State of NY Salem 1 and 2. Public Service Electric and Gas Company Palisades, Censumers' Power Company P11Cria 1, Boston Edison Cinna, Rochester Cas and Electric t 42-August 2, 1989 e - _ _. ~., .. ~ -

O t

9.

l t, I R$FERENCES

1. Final report on the Diablo Canyon Long Term Seismic Program, Pacific Cas and Electric Co., Diablo Canyon Power Plant, Docket Nos. 50 275 and 50 323, i

July 1988. r

2. Diesel Generator Allowed Outage Time Study, Pacific Cas & Electric Company, May 1989.

f

3. Units 1 and 2 Diablo Canyon Power Plant.

• Final Safety Analysis Report Update," Pacific Cas and Electric Co., December 1988.

4. PC&E letters to NRC signed by J.D. Shiffer, No. DCL 88 238, October 10, 1988, No. DCL 88 260, October 28, 1988. No. DCL 88 285, Nevember 29,5988, No. DCL 88 297, December 9,1988, No. DCL 89 010, January 16, 1989, and No.

DCL 89 152, June 2,1989 I

5. Hoopingarner, X.R., Vause, J.W., Dingee, D. A., Nesbitt, J.F., " Aging of l

Nuclear Station Diesel Cenerators: Evaluation of Operating and Expert Experience," NUREG/CR 4590, August 1987. i

6. DOE / RECON, Nuclear Safety Information Center (NSIC), 1963 to preseat.

7. Worrel, R.B. and Stack, D.W., 'A SETS User's Nanual for the Fault Tree Analyst," Sandia National Laboratories NUREG/CR 0475, SAND 77-2051, November 1978.

t ? 4 b 43-August 2, 1989 l

.._--g _. t~ n --. ?,' . i I4. .. [ >.-., aw_maa 'n.' swA

..Y

...n.-. z _,,w.w'._'-..-,mma' w,._. APPENDIX A HARDWARE UNAVAILABILITY CUT SETS FOR THE DIESEL GENERA *iORS [W& tegg,gw t?;i&Ph s r: spgg ' qm - ' " " TMs J:.Q?""*"** 7 g L ,l 'S 44 August 2, 1989 i

,m+ e' [ g-( i . Hardware Unavailability Cutsets in case of One DG Total Hardware. HV - Independent Hardware. HWI = 3 3.fd tr.cf rt?tr

• 1

4. 67 3 M. [ 2 F l*Ja r

• 2 2.1318! 06 31V
• r tir gg.

2 2 1 2J*'.';6 .it7 Tr 0,V 3 1 13 F.M ats 12C

• t i.a.1 t 7 '". c.

nea a f _.1 1!?.!!?.t4 DSD

• f 1

1.U.E**,. Dp,e * ? t,. ? ? t ig t e r r,

• 11
• _.,1. t 61 '!M1 t.0) i 9, 1 4 61 U.* 21 19e it t,gs t i ve. ?)

tes

• l 11

.1. 8.*1't: CJ TAR 27 1.6Lt.'t.e8 .'2r

• r 13.,_ 1.0,3,1,*g g e taa
• t o..

J..

• 2 F.6! s.0 wty
• off +

15. I,t17*L:t)

• I y
• mya +

it '..ti!it. ** ?ru * *va

• 1? _ f.ti!EL-23
• Lg
• f v5 -_*

1 86. k.1269tet' ar v

• 5.ya.

12 4.12,1.*L. " 4t.V i.,?.yc + 2 9...i. 126 *? :tf. wty a tyr. e! i July 25, 19S9 1

7 .= m-.-. a i a-t- l l \\ 1 p Hardware Unavailability Cutsets in Case of Two DCs L 1 l [ Total Hardware. NW - Independent Hardware. HVI - 1

• * ~ i* ' 1 3.,9st.es rt Wr
• rinic.

'- - -- T,isisi,.t1 riatr * 'Irli + g r ; 1 1:? !!- 06 034

• t

?. 4 3e tr cl rt;te. ury. F:vic + ...E* CD 3 1 1.*70'I*tt 1 r, e ?q tg. gt ty rtetr. r!rtr, e i ' " ' ~ ~ ~6

1. 6 1 *t. t l T3t

• 1 13e 6t.cb M'y
• rgyle e r;yas t

1 661?I Cf tDA * ~ ti 1.t @ t*0% !DC

• r 7

F.9292!*C6 r!Dir

• NfV
• rfV1C
• t F.439tI*t6 Ntv r!vir
• r131c e i

a 9 5 917 %!* Cl Mtv

• Ove
• 1 it

6. $ s t et.p rt:1'

• DrF
• i 11 6.1 * * *I-t b r1015
• DDC *

[ 12

6. atv er. ct r391r e n gt, e e ott,

1* b.it**I.ca rg;ge t o-

6. 08.7 tt - tt rtCtr.

• Dre e 14 6.0t?'t-3b r!91G
• pnD +

15 ~ 1.13*6 tat 6 M' y

• r!vir
• rivtr..

17 f.120 M. c t r301r

• tri +

L L r .t July 25, 1989

k o s s H"ardware Unavailability Cutsets in Case of Three DCs Total Hardware, rN = Independent Hardware. H'.'I - f ~ 1 6.gr79.*cf ricar e r t:1r. * *1t1=

• 1

6. 9f 7*D t

• r! ate
• e i n i r.
• 1rt e
• t 2.**98r.ct r!D1'
• wrv * '!v1G
• ttD1w *,

2 1 7 7C.5-!5 GDG + 3 1 6617{ ta tna e 3 g.

• t o gg. g

.........rgggr e ritar.

• rielu gr g.

l J 6 6 0(741*C6 F101M

• D0a
• 6 f.*t98 tat' riair
• wry ' r ! al e, rtygu.

+ way

• Ftytr.
• riy1w.?

6 1921?*Cf rte.1r a 8 h.ctFt!*tb FID1'

• 00% +

6 "6.467t!*C6 r i o t t.

• ett.

6 6 19 tat.t* wtv e rev1r

• r ; v a r.
• rgStw.

S t. c tt nt. c' r101w

• 100
• 6 1? tit *tt wrv + *:yar e r:010
• rivin
• t.071E*tt wF V
• riv1r
• riv1G
• r!v1w A

t.0209 eat. F101r

• ttw
• 9 6.Stis# tat' F103r a trG
• it t.St0*!*ta r201G
• T03 *
• ~ Tet +

11 f.03091.t? r!Otu 17 6.0209! 0' F t 61'r7*~ t Ct ~ 11 6.1268 *tt wgy

• iva e

'~ 16 3 ; 13 9 t!'* t'? 9 t I* *'r,W. if 2.479eI*t? Mtv

• rivar e r1C1G
• Fialw
• 16 f.l'9'I*tF r!rar
• wiv
• FIVIG * 'Illw +

1* *

• Y; gt g 6t r e-r f oF
• w tv + r Wit'~~~~f!Vi e

~ e ' "~ia ~ t.1 ? 2 6 c' Mtv

• rio1H
• riv a
• 4 s-k July 25, 1980

c n4- . -~ o 9 '4' l 1 a L i l Hardware Unavailability Cutset.s in Case of Four DCs i Total Hardware. HW - Independent Hardware. HWI. l u a.if93t*Si w Dfr*' i 3. elt g. n r:3gr

• r t a g e.

r:rgw. r ;a t e, r 1ttt171 04 F101F * *-F ID t C-'* r!31 M-*-r 10 tC-p 3, g ry gg. g. pg3ge . wry. t.ygg r131w

• F i v f. *

~3 - V.' 0 3 8 9E *'3 7 i!DFC***T94 6 g g,39 tor.go wry e r;g gr e r;ngc e r3agu e r 3 0 pe.. -"%*"" 17 t3 3 9t-9, -r!Dtv-* ftt 4 g,3riet.t* r 7tr + wrv e r t's;

• t t p :=
• r e v '*, * !

4, i

e. 93 9 9E* 0, Ff 319-* 103"*

1. t tt eract ritte

• wry etage, = ttyg.. r i ptr..

e + " t tit 389t*8) F201r-**fD2 6 1.9 61t*D4 r!3tr

• wry
• rtygr.
• r!a.tw rtver..

a.19 9t r*tT*"" MET ** C's ' r ? g. $3 r. c p r101r

• Mev * *fv10
• r1v1"
• r13?r,.

" Q l'.*b96'3t*07 FID1M* **FIDtt-"00a

• f -' g ; g'!6 F'+ d '"~~~'NU ~*~r'f v t r
• rintG *
• E I;I~'IIv" N ~I

' 9 176 96 3t

• O f-""--f !D 1 F-*-f !??C-*-00 6 v

g g, ggg,g. g, ugy. rgvgr e r;agg a r!ygu F 3 0 tr.-.' e

• i

..=963t-ii i! tit-*-Tf0tC-+-SDt. {g-g, gyfi[ g a wry. rgyg5 rgygg. rg;g. repyc, t -11 iT%963 tat? T 1010-*" TIS 1M-*-0., ' 11

1. 5 2' 'F.709 vie [r

• Arv + ttngr + riv1w
• r:vT57 ;

t --12 in6963 tat? 7501P-*-f!Dtt-*-DD" " 12 f. 2 311r

• 1 ?

Mtv

• riv1' * '1V10 * *1'v1w

~'I DF4 a l r i; . 16 3t

• 17 "--T101'-*-t! Ot w-*-00' i3 f.2.'Iti+12 r !'D i r *.i y t

+ ttvgc

• v3vgw
• F3sts.

""16 s371T5t<sn i!D2*'"*-f!3F M D, ' D

2. F I11' a l1 s# V '7 v a r"'~r'P.'1 G * ' ! V 1" ' i f V N~~.

1

• 19 s;711 ?t*0 s-r!Dic-*-ffDFC-*-TCi,

rivtg gg, g,ggggg.gg w,y. ryggr. rivig

• re stu *

- 1% s. 715 YT* 18--' T!D13"*-f 7 D fif"***D" *

• p y,3gpi.gg ug y. F1v tr
• F1vl',
• r1v3*
• rl'it."'~!

7 1*-" t.** 11 f E

• 8 W----f IP 1w-*-f 10 FC-*-f 30 v

+ '"" Tit t'9 f t =9 7"""'""T! Cr*-**P! S Ifi-*-f DJ - '"""Tr?14 !!* 78 1 !!rt F"' f!D !)f**-T C! v Te .7515 9t

• 9 8---"Mt v-*-f!D is-*-7 v5 v l.

e! i l 1 l July 25,1989 ,~

7. m. O' . a-e. 4,g. ~~g'- Matdware Unavailability Cucsets in Cue of Five DCs F Total Haroware, HW = i t t. ? ? t *f. M r.or. r 3.ttest.t' <tv

• rro +

i J t.593rt-r* ttoir r 3 0t r.

• r1Cl*
• rinr5
• r!D?d
• a 6

r. r :5 rt. t. riotr riorr.. tow. t' r

2. r ts t!. e s rinac*ricts*'er +

-I r.rterr.ca rtrto a r toru * **D+ ? r.rtert.ca r70rr.. rioru a vna. f a

2. r t s t t. at

_rintu

• rioru a tot +

r.rtsrt en riott a r!cra a tro + i to

r. r t? tt.H FIntw *

• torc
• tte +

11

r. rtl?!.04 -

r!oti

• r20tw
• 13r +

12 tertSet.t* rIntr

• r!ota
• 1"!
• 13 r.rt5ft.09 riotr. r!otG
• inJ
• 16

5. 6 89 6!. tg rtoic

• rTorc
• FIOfw
• oot +

19 f.649(t.99 rtoto. riose e r:otu

• D P*
• 9u @u Independent Hardware, HWI =

-t. 6.sattt.c. riotr r t o t t.. r:cta

• rinto. renew r

s. r e s tr. n rinte

• wry a rivic. riosu. rtorc
• rte:re.

3.se66r.ta wtv

• rivtr
• ricic
• riotu. rtn7s
• rteru 6

J. g

• i6t t e r toir
• wrv
• F*rsc
• riosu. rints a rtyra +

r t.ars6t-to riotr usy

• rinto. riotu. rives. r ?crw *

3. t h it.~ti riotr u t v

• ri o tr.
• r i v = 77fn + c r ifri.

5-a r s. au ot. a t riotr = wrv. rivan. rial. crern. r!vre + t. 663 7 11 rtoir. wry

• ricic. rintu. rivrc
• ricrw.

i 2 .J. l ) July 25. 1989 l 4 1 m. a ,i}}