ML20083D085
| ML20083D085 | |
| Person / Time | |
|---|---|
| Site: | Byron, Braidwood, 05000000 |
| Issue date: | 11/30/1983 |
| From: | Papazoglou I, Youngblood R BROOKHAVEN NATIONAL LABORATORY |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| CON-FIN-A-3393 BNL-NUREG-51633, NUREG-CR-3096, NUDOCS 8312270195 | |
| Download: ML20083D085 (55) | |
Text
.
NUREG/CR-3096 BNL-NUREG-51633
=
Review of the Byron /Braidwood Units 1 and 2 Auxiliary Feedwater System Reliability Analysis Prepared by R. Youngblood, I. A. Papazoglou Brockhaven National Laboratory uclea Regulatory l
DR DOCko!$$$$04 PDR
NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus product or procen disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.
f The views expressed in this report are not necessarily those of the U. S. Nuclear Regulatory Comission.
l l
l i
Available from f GP0 Sales Program Division of Technical Information and Document Contrcl U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Printed copy price: $4.00 and National Technical Information Service Springfield, Virginia 22161
NUREG/CR-3096 BNL-NUREG-51633 Review of the Byron /Braidwood Units 1 and 2 Auxiliary Feedwater System Reliability Ana ysis Manuscript Completed: December 1982 Data Published: November 1983 Prepsred by R. Youngblood, l. A. Papazoglou Brookhaven National Laboratory Upton, NY 11973 Pr:pered for Division of Safety Technology Offics of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Weahington, D.C. 20555 NRC FIN A3393
NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following sources:
- 1. The NRC Public Document Room,1717 H Street, N.W.
V ashington, DC 20555 ,
- 2. The NRC/GPO Sales Program, U.S. Nuclear Regulatory Commission, Washington, DC 20555
- 3. The National Technical information Service, Springfield, VA 22161 Although the hsting that follows represents the majority of documents cited in NHC publications, it is not intended to be exhaustive.
Referenced documents available for inspection and copying for a fee from the NRC Public Docu-ment Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bulletins, circulars, information notices, inspection and investigation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence.
The following documents in the NUREG series are available for purchase from the NRC/GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and )
NRC booklets and brochures. Also available are Regulatory Guides, NRC regulations in the Code of )
l Federal Regulations, and Nuclear Regulatory Commission Issuances.
I Documents available from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.
Documents availaole from public and special technical libraries include all open literature items, I such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and conqressional repoits c2.n usually be obtained from these libraries.
Documents such as :neses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draft reports are available free upon written request to the Division of Tech-nical Information and Document Control, U.S. Nuclear Regulatory Commission, Washington, DC 20555.
Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards Institute,1430 Groadway, New York, NY 10018.
l l
ABSTRACT This report presents the results of a review of the Auxiliary Feedwater System Reliability Analysis for Byron Units 1 and 2/Braidwood Units 1 and 2.
The objective of this report is to estimate the probability that the Auxiliary Feedwater System will fail to perform its mission for each of three different initiators: (1) loss of main feedwater with of fsite power available, (2) loss of of fsite power, (3) loss of all 460 VAC power. The scope, methodology, and failure data are prescribed by NUREG-0611, Appendix III. The results are compared with those obtained in NUREG-0611 for other Westinghouse plants.
v iii
l TABLE OF CONTENTS Page ARSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii L I S T OF T ABL E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES ............................ vii
SUMMARY
AND CONCLUSIONS ........................ viii
- 1. INTRODilCTION . ........................... 1
- 2. SYSTEM MISSION AND DESCRIPTION . .................. 4 2.1 System Mission ........................ 4 2.2 Descri pti on of the B/B AFWS . . . . . . . . . . . . . . . . . . 4 2.2.1 General Description .................. 4 2.2.2 Pumps ......................... 4 2.2.3 Suction ........................ 4 2.2.4 Discharge Paths and Recirculation Flow . . . . . . . . . 5 2.2.5 Support Systems .................... 5 2.2.6 Testing and Maintenance Policy . . . . . . . . . . . . . 5 2.3 Oualitative Assessment .................... 5
- 3. RELIABILITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . 8 3.1 Scope of the Reli ability Analysis . . . . . . . . . . . . . . . 8 3.2 Approach of the B/B Report vs. Approach of BNL Review . .... 8 3.3 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.4 Dominant Failure Modes .................... 9 3.4.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.4.2 Dominant Contributions to Unavailability Given LMFW .. 10 3.4.2.1 Startup Train Unavailability . ........ 11 3.4.3 Dominant Contributors for LOOP . . ........... 11 3.4.4 Domi nant Contributors for LOAC . . . . . . . . . . . . . 12 3.5 Summary of Results ...................... 12 3.5.1 Results . ............. .... ...... 12 3.5.2 Comparison with a Typical Three-Train System . . . . . . 12 3.5.3 Sensitivities ..................... 13 3.6 Comments on Maintenance . . . . . . . . . . . . . . . . . . . . 13 3.7 Difference Between BNL Assessment and Commonwealth Edison Assessment . ...................... 14 v
i-TABLE OF CONTENTS (Cont.)
Page 3.8 Common Cause ......................... 14
- 4. CONCLUSIONS ............................ 25 REFERENCES .............................. 26 APPENDIX A: LOOP Frequency ...................... 27 APPENDIX B: Comment on " Feed-and-Bleed" . . . . . . . . . . . . . . . . 29 APPENDIX C: Inter-Unit Bus Tie .................... 32 APPENDIX D: Commonwealth Edison Memorandum on Inter-Unit Bus Tie . . . 33 i
i l -
l l 1
l l i l
l I
Vi
LIST OF FIGURES Figure Title Page 1.1 Results . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Simplified Diagram of AFWS for Byron /Braidwood Units . . . . . . . . . . . . . . . . . . . . . . . 7 3.1 Simplified Fault Tree: Loss of Main Feedwater .. 16 3.2 Simplified Fault Tree: Loss of Offsite Power . . . 17 3.3 Simplified Fault Tree: Loss of All AC ...... 18 B-la " Simplified Fault Tree" Core Damage Following Extended Loss of Offsite Power .......... 30 B-lb " Simplified Fault Tree" Core Damage Following Extended Loss of Offsite Power (Cont.) ...... 31 i
LIST OF TABLES Table # Title Page 3.1 Dominant Contributors to Train A Unavailability . . 19 3.2 Dominant Contributors to Train B Unavailibility . . 21 3.3 Dominant Contributors to Startup Train Un avail abil i ty .................. 23 3.4 AFWS Unavailabilities . . . . . . . . . . . . . . . 24 vii
i
SUMMARY
AND CONCLUSIONS 1
After ~ the accident at Three Mile Island, a study was performed of the re-liability of the auxiliary Feedwater System (AFWS) of each then-operating Westinghouse plant. The results of this study were presented in NUREG-0611.
Commonwealth Edison has provided NRC with a study of the Byron and Braidwood AFWS, performed using NUREG-0611 as a guideline. BNL has reviewed this study, The BNL conclusions are as follows ("High", " Medium", and " Low" refer to the NUREG-0611 reliability scale).
I
- 1. For an accident initiated by loss of main feedwater with offsite power available:
The reliability of the Byron /Braidwood- Auxiliary Feedwate System plus startuptrainisinthe"High" range (Unavailability =1.2x10g/ demand).
I 2. For an accident initiated by loss of main feedwater coupled with less of offsite power:
The reliability of the Byron /Braidwood Auxiliary Feedwater System is at the high end of the " Medium" range utilized (Unavailability = 2.6x10-d provided that the inter-unit bus tie is
/ demand).
- 3. For an accident initiated by Loss of all AC power:
l The reliability of the Byron /Braidwood Auxiliary Feedwater System is in the " Medium" range (Unavailability = 2.2x10-2/ demand).
l These results are summarized and compared with Commonwealth Edison's re-l sults on the following table. Refer to Table 4.3 of the B/B report and Ap-pendices C and D of this report for explanation of the Commonwealth Edison re-i sults.
l Byron /Braidwood AFWS l
Unavailability per demand BNL Commonwealth Edison LMFW 1.2x10-5 7.1x10-6 i LOOP 2.6x10-4* 9.4x10-5* l LOAC 2.2x10-2 1.7x10-2
- These numbers are calculated assuming an inter-unit bus tie. See Appendices C and D.
viii l
I
i
- 1. INTRODUCTION The purpose of this study is to review and evaluate the " Byron Units 1 a 2 and Braidwogd Units 1 & 2 Auxiliary Feedwater System Reliability Analysis"(21 (hereafter "the B/B report") and perform an independent simplified (A(WS reliability analysis using methodology and data put 'forth in NUREG-0611 ll. The objective of the simplified reliability analysis is to assess the probability of failure of the AFWS on demand under different loss of main feedwater conditions, namely, (a) Loss of Main Feedwater without loss of Offsite Power (LMFW),
(b) Loss of Main Feedwater associated with Loss of Offsite Power (LOOP),
(c) Loss of Main Feedwater associated with Loss of Offsite and Onsite c AC (LOAC).
After the accident at Three Mile Island, a study was performed of the Auxiliary Feedwater Systems ( AFWS) of all then-operating plants. The results obtained for operating Westinghouse-designed plants were presented in NUREG-0611(1) . At that time, the objective was to compare AFWS designs; ac-cordingly, generic failure probabilities were used in the analysis, rather than plant-specific data. Some of these generic data were presented in NUREG-0611. The probability that the AFWS would fail to perform its mission on demand was estimated for three initiating events: LMFW, LOOP, and LOAC.
The results of this study are depicted in Figure 1.1.
Since then, each applicant for an operating license has been required (3) to submit a reliability analysis of the plant's AFWS, carried out in a manner similar to that employed in the NUREG-0611 study. Recentl criterion for AFWS reliability has been defined by NRC(4) in they,New a quantitative Standard Review Flan (SRP).
... An acceptable AFWS should have an unreliability in the range of 10-4 to 104 per demand based on an analysis using methods and data presented in NUREG-0611 and NUREG-0635. Compensating factors such as other methods of accomplishing the safety functions of the AFWS or other reliable methods for cooling the reactor core during abnormal conditions may be considered to justify a larger un-availability of the AFWS."
One goal of this study is to compare the B/B AFWS to the plants studied in NUREG-0611. For this purpose, it is important to follow the NUREG-0611 methodology as closely as possible. Another goal is to evaluate the B with respect to the reliability goal set forth in the new SRP(4), name/B ly AFWS that the AFWS should have an unreliability in the range of 10-4 to 10-5 ,
per demand.
The B/B AFWS design has two trains, but some additional redundancy is available in the startup feedwater system, if offsite power is available. For 1
the LMFW initiator, then, the B/B AFWS design + startup train compares well with three-train designs. For LOAC, the AFWS is typical in having a single AC-independent train. For LOOP, however, the B/B AFWS is not as reliable as most three-train systems.- The primary focus of this review is to elucidate thi s point.
In Section 2, the mission of the AFWS is described, and the B/B AFWS is presented. Salient features of the B/B design are noted and compared to NRC recommendations.
In Section 3, the reliability analysis is presented. The dominant contributors to unavailability are identified. The unavailability of the AFWS for each of the three initiators (LMFW, LOOP, LOAC) is given. The sensitivity of the results to certain assumptions is explained. Differences between BNL's analysis and the B/B report are noted.
In Section 4, BNL's conclusions are presented. The unavailabilities of B/B's AFWS are campared to those of other. Westinghouse plants.
Appendix A discusses the frequency of LOOP. Appendix B sunmarizes the .
logical implications of giving credit for feed-and-bleed. Appendix C ]
discusses the effect of adding an optional link between 4160VAC buses of Units 1 and 2.
I 2
1,ansient Events WFW LMFW/LCOP LMFW/ Loss of A:t AC' l
l Plants low Med E9h Low l f/cd H g5 Low L'e d M52 B/B - _l 0 l C_(m A l l0 i
ll Ibddam Neck O h O l i
San Onofre O O O a "), l Friivie Island ( Q Q Sa'c:n Cs -O ( :) ( l
.n n 7 ion O 3 )
l Yanku nowe O Q h a Trojan O O g Indian l'oint O O h
- (ewanee C ! h
- 11. D. Robinson 9 O ()
Deaver Valfey k O d Ginna O O g)
Pt. Ocach O O b Cook O O O Turkey Pt. O O 0 l
Faricy C O g sor ry C O d ;)
fio. Anna C O C
, l
-> ~ Order of f.*agnitude in Unavailability Repre:ented.
'fJate: Tha scale for this cvent is not the same as that for the Lf.1FW and Lt.'FW/LCOP.
Reliability Characterizaticas for AFWS Designs in Plants Using the Westinghouse NSSS.
Adapted From NUREG-0611 O Assessment based on original report A Assessment based on proposed modification (Refer to Appendix C)
(,
FIGURE 1.1 3 .
- 2. SYSTEM MISSION AND DESCRIPTION 2.1 System Mission The mission of the AFWS is to provide feedwater to the steam generators in the event of LMFW. Core damage will result if decay heat is not removed in sufficient quantity, either by producing steam in the steam generators or by allowing hot primary coolant to escape via the pressurizer while replenishing it with high-pressure injection (" feed-and-bleed"). LOOP and LOAC cause LMFW, and therefore also challenge the AFWS.
At B/B, AFWS mission success is defined as delivery of 160 gpm to each of three steam generators or 240 gpm to each of two steam generators before they boil dry (20 to 30 minuter). Ordinarily, the AFWS is required to operate for about 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> to cool the unit down to 3500F, below which temperature the low pressure residual heat removal system operates.
2.2 Description of the B/B AFWS 2.2.1 General Description The B/B AFWS is sketched in Figure 2.1 of this report (Figure 3-1 of the B/B report). The AFWS consists of two trains, one electric-motor-driven (EMD) and one diesel-engine-driven. In addition, there is the startup train, which is part of the main feedwater system but is available to supply feedwater during some transients. The AFWS is initiated by either a low-low steam generator signal, a safety injection signal, or a loss of offsite power. The startup train requires operator initiation. The following paragraphs outline the most important attributes of the AFWS. The startup train is not analyzed here; its most relevant attributes are summarized in section 3.4.1.1. f 2.2.2 Pumps Both AFWS pumps are 1165 brake horsepower units. The diesel driven AFWS l pump is capable of providing 840 gpm at 3350 feet head, which is nearly twice the capacity required for mission success. This pump is capable of supplying its own cooling and lubrication independently of AC, bi:t when AC is available, backup pumps are provided for oil pressure, water jacket cooling, and room air cooling.
The EMD AFWS pump is capable of providing 890 gpm at 3350 feet of head, which is nearly twice the capacity required for mission success. It requires l AC; therefore, its unavailability given LOOP is dominated by the unavailability of the associated emergency diesel generator.
2.2.3 Suction The primary suction source is the condensate st.orage tank. Redundant flow paths begin at the CST and meet at a header which supplies both pumps (see 1
4
Figure 2.1). No single valve can block the flow. In the event of low suction pressure, there is automatic switchover to service water. There are startup suction filters which are further discussed in Section 3.8, on common cause failures.
2.2.4 Discharge Paths and Recirculation Flow Flow from each pump goes to all four steam generators through independent paths (refer to Figure 2.1). Flow is controlled by manually controlled air operated normally open control valves which fail open on loss of air.
Upstream of the header from which the four discharge paths emerge, there is a valve which is temporarily closed during testing. In the event of a system demand, this valve opens automatically. During the first phase of a pump test, closure of this valve directs flow to recirculation; at the conclusion of a test, the test valve opens, and flow is tested all the way into the steam generators. Thus, no portion of this system stays untested between challenges or shutdowns. There are flow restricting orifices in each path, so that depressurization of a steam generator does not result in excessive diversion of AFW to the faulted steam generator.
The normal recirculation path is back to the CST; if suction switches over to service water, recirculation also switches over.
2.2.5 Support Systems The diesel driven pump is capable of supplying all of its own cooling and lubrication, so that it can operate completely independently of AC. When AC is available, there are external sources of cooling and lubrication which improve the reliability of the unit. The EMD pump requires essential service water for cooling. Here, this has been assumed to be available if emergency AC is available.
2.2.6 Testing and Maintenance Policy Monthly, or after any maintenance ect, a full flow test is conducted.
During a test, a discharge valve is first closed to direct flow to re-circulation (see Figure 2.1). An actual system demand during this period causes this valve to open to permit flow to the steam generators. At the con-clusion of the first phase of the test, this valve is opened to permit testing of full flow to the steam generators. If this test is carried out, a significant class of failures is substantially reduced in probability. For example, the analog of leaving a discharge valve closed in another plant cor-responds here to leaving a discharge valve closed and disabled and failing to carry out plant policy re testing. This testing policy significantly improves the reliability of the system.
2.3 Qualitative Assessment The B/B report deals with the B/B AFWS plus the startup train. The at-5
< tributes of the startup' train are quite different from those of the AFWS. The two-train AFWS no. the following important strengths, which correspond to NRC recommendations made in NUREG-0611:
- 1. Automatic initiation of Trains A and B, even during test (recom-mendation GL-1);
- 2. Redundant flowpaths from the condensate storage tank (CST) to the pumps, so that no single valve can block the flow (recom-medation GL-2);
- 3. The ability of Train B to function independently of AC (recommendation GL-3);
- 4. Automatic switchover from the CST source to the Essential Service Water, in the event that low pump suction pressure is sensed (recommendation GL-4);
- 5. Monthly staggered testing of the entire flowpath fro.a the CST to the steam generators (recommendation GS-6):
On the other hand, the AFWS has only two trains, which for LOOP seriously degrades its reliability. For LMFW, the startup train can be made available, but the NRC recanmendations with respect to automatic initiation, etc., are not met, since the startup train must be manually initiated, l
If even modest credit is given for the startup train, the strengths of the l AFWS are such that the reliability of the combination is high for LMFW initiator. Given LOOP, the redundancy is minimal (2 trains). Given LOAC, B/B is typical in that it has a single AC-independent train; but the train is of a relatively new type, and its reliability is correspondingly uncertain.
i l
- 1 l
l l
l l
l 6
l e .
$b 5I s -t 3 w , ,3,3 h
. = a fe
$ g\ *. - Chh L l m Es= ** 8 1-le Ei f k5 +5 ' O
$2 i IE v "'E w{"kB d K~
M Es ie 15 r,L. l.2 !,
~
- 1g. - -e
<\a* r -4 sigi ~
. E \= -
IX i h g *
<> <> s, g= ig r_, r, a r:
m, r,,
i eg e
,' ee a
=
O*!3 st 1 1 1 1 m e. 6 sb m:)l -$p E ! o o
ges! E'l N1P ,- E-( yf c. , a
- o M 0, w w %
U$ C' O C gl B ,l m < r 5 .- -
g-
$g
- . <e g!
!\
a sSE l
BCD i
e
!S3 !A e
E l
- E t
7fr 55 2
i !CE2 ! !!GS i 5
" X 5 ik $
n: '2 '
X ' : +- .
,m
- i s.
EGRE 5C{$
v z E'r o
= --- : , 3 re
.-. ;;53 --
c:
I$cg m I I I
.O b $b3 <
L A
- L - L -
Y_ _
~ ~ %
L ~
L L L L 1~hhD
. .n o C9
_2 -
i ef9 -
i 9 02 C i .i-2h0 2l5+4 :! -
irEidiszazEadirE*sEim i .i 7-
.L : i f gi i
r s.L e.-
ss 2 =
2 n
4 5 0 L.)
L.
< -l $I s l u- 1 o1 s
=s G
C.
e lG$ lC{ [C{
- - IGE 5-((,'..
= . - -
u
- =I I *
="
=
e 2"EI : N
- Ei! ! ~ t; 8
a- 07' B<CE:
S
=C *%
- c; sCt -
55! !! 5"*!
>- - u
'L .l g- t. n }d M s- 'e i &
kh, .!b1 hb b kb $ EE
.. .i.io.5. -d Y + E 1 1 1 -1 -1 .
I ?* \ [ $; ) I ?, g \ l $5 \
7
- 3. RELIABILITY ANALYSIS -
3.1 Scope of the Reliability Analysis The values given here are point estimates of the unavailability. The fol-lowing points should be borne in mind concerning these results.
- 1. Uncertainty analysis has not been performed. As noted in NUREG-0611, such analysis typically indicatej that the median unavailability is'somewhat higher than the point estimatet 1 (by a factor of two or less), which, for the purposes of NUREG-0611, did not matter.
- 2. Depending on the error factors assigned to individual failure rates, the error factor multiplying the median unavailability can easily be around three for analyses such as this.
- 3. A detailed modelling of common cause failures has not been attempted in this review but it is clear that the estimate would increase somewhat as a result of such a study. (A quclitative discussion of common cause failures is provide in Section 3.8.)
3.2 Approach of the S/B Report vs. Approach of BNL Review The B/B report identified many contributors to the top event " Failure to remove heat from the SG's". It also provided considerable material on common cause failures. As originally written, the report's conclusions did not lend themselves to simple comparison witt)*NUREG-0611 guidelines; revised con-clusions were submitted in a lettert ). However, owing to misunderstanding concerning NRC guidelines, these were calculated using unrealistic as-sumptions. Finally, at a meeting in Bethesda, these misunderstandings were l largely sorted out. A third set of numbers is warranted. These numbers are calculated by relying on the system description in the B/B report to identify dominant contributors, and quantifying them according to NRC guidelines.
After the review was essentially complete, the utility proposed a link be-tween 4160 VAC buses of units 1 and 2 to improve the availability of AC to the EMD AFWS pump given LOOP. The effect of this is discussed in Appendices C and D, and included in Tables 3.1 and 3.4 (results calculated with and withcut crosstie option).
3.3 Assumptions
- 1. Maintenance on more than one AFWS train at a time is not permitted.
- 2. It is assumed here that the B/B assessment of 3 x 10-2 is a real-istic estimate of the probability of the operator failing to supply feedwater from the startup train given LMFW.
(*) Point estimate: unavailability calculated using the medians of the input variables.
8
- 3. Although the B/B report states that an AFWS train can be down for 7 days, the naintenance unavailabilities assessed here are those tabulated in NUREG-0611 as mean maintenance act durations for trains which are allowed to be down for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.
- 4. In the case of the diesel-driven pump, the services are provided cy the pump itsel f. In the case of the EMD pump, Essential Service Water is required for cooling. Here, we have assumed that the Es-sential Service Water availability is the same as that of emergency AC. This may be incorrect given a LOOP affecting both units; this is discussed in Appendix C. Failures associated with lubrication cre assumed to be included in the EMD pump failure probability.
3.4 Dcminant Failure Mcdes Let Vjj = unavailability of the dischange path from pump i's test valve to steam generator j, Pj = unavailability of all components upstream of the test
, valve of pump i (including the test valve itself).
l l (Refer to Figure 2.1). Then the event " failure to provide flow from one AFWS pump to at least two steam generators" may be written PPAB+P(V A BAVBB V BC + V BBVBCV BAVBCVBD VpVBB VBD)BD B +V V
+ P B(VAAV B C+V 4BACAD+V V AA AC AD V
+ AApABAD/
V
+ higher order terms.
Barring substantial common cause failures involving multiple-V events, it is easily seen that the dominant contributors are contained in P PAB -
i Valve maintenance on a flow control valve (e.g., AF005E) would effectively incapacitate an entire train, because it would be necessary to isolate an entire discharge header from its pump. This would be a substantial contribution to maintenance unavailability. Here, it has been assumed that such maintenance acts either will not be performed with the reactor at power or can be combined with pump maintenance.
For present purposes, then, the dominant contributions to the AFWS unavailability can be written Unavailability = AH BH CH + AH BM CH + Ah BH CM Where AH = Train A Hardware failures, AM = Train A Maintenance, BH = Train B Hardware failures, BM = Train B Maintenance, CH = Startup Train Hardware failures + Human Error, CM = Startup Train Maintenance, 3
l 9
a
- , , , , , - . , , - , , - - - , - - - - - , - - . e , . , , , , , m. , , . , , --.. , n -.,-.
and each quantity raust be re-evaluated for each initiator. This expressica is so constructed that no two trains are simultaneously out for maintenance.
Figures 3.1-3.3 show simplified fault trees for each of the three '
initiators consideied. These reflect the considerations leading up to the above fonnula.
Following are descriptions of the important contributors to the un-availability of each train.
3.4.1 Data Ideally, all of the data would be prescribed by NUREG-0611, but there are important contributors whose values were not given'in NUREG-0611. Two -
. instances are diesel generator unavailability and diesel pump unavailability.
Here, the WASH-1400 value for diesel generator failure to start, and i NUREG-0611 prescribed data for diesel maintenance are employed. Diesel pumps I were not contemplated in NUREG-0611; we have adopted the B/B report's value of 1 X 10-2 for the failure to start, and the NUREG-0611 prescription for maintenance unavailability. The BA rao)-:'m rabue is itsed 01 dits oitahqei from Trojan, whose diesel-driven pump was initially much worse than this but appears to have been improved to this level.
3.4.2 Domir ant Contributors to Unavailability Given LMFW Train A: The dominant contributors are described below and given in Table 3.1. They are:
- 1. Maintenance and failure of the EMD pump, including control circuitry;
- 2. Failure of automatic initiation would be u large contributor except j that failures of manual backup initiation within the available time I has a fairly low probability. l Train B: The dominant contributors are described below and given in Table 3.2. They are:
- 1. Maintenance and failure of the diesel-driven pump, including control circuitry;
- 2. Failure of automatic initiation would be a large contributor except that failure of manual backup initiation within the available time has a fairly low probability.
I Startup Train: The dominant contributors are given in Table 3.3. They are:
- 1. Probability of operator failing to initiate flow (several steps are
. necessa ry);
10
- - - , , + -
i
- 2. Probability that the cause of the initiating event (LMFW) will render the startup train unavailable;
- 3. Maintenance or failure of the startup train. (It is doubtful that mainte.1ance would be performed on this train during operation, but it has been assessed.)
3.4.2.1 Startup Train Unavailability The startup train is assessed in the B/B report to have an unavailability of 3.7 ? 10-2 per demand for LMFW initiators and is unavailable when offsite power is unavailable.
The source of water for this train is the condenser hotwell; it is re-quired that at least one of the four condensate / condensate booster pumps be functioning, and that the startup pump be functioning. This train also re-quires that several manual actions be taken before it can supply feedwater following a LMFW. The failure of these manual actions dominates the B/B re-port's 3.7 X 10-2 estimate (10-2 f ailure probability for each of three man-ual acts). Hardware failures were deemed much less important; indeed, if the human error contribution and the maintenance contribution are subtracted from this figure, the availability of the startup train following LMFW is com-parab!e to that of the EMD train of the AFWS. This conclusion has to do with ihe fact that the startup train components are in general already running, and r.eed not be started.
However, the B/B rcport assessed the unavailability of the startup train under normal conditions; the quantity of interest is the unavailability of this train given that a LMFW has just occurred. Some LMFW events cause viola-tion of the requirements listed above; of the 80 LMFW events surveyed in NUREG-0611, 5 would render this train unavailable, or at least come very close to doing so. In one case, a low level in the hotwell caused cavitation of the condensate pumps; in three cases, suction strainers in several pumps were clogged by an upwelling of debris in the hotwell; and one case simply refers to "the loss of the condensate pumps." From the descriptions given in some of these cases, it is not completely clear that all flow paths are completely blocked; but it is clear that there are contributions to the unavailability of this train following LMFW that are comparable to the dominant human error probabilities assessed in the study, and which completely dwarf the hardware failures assessed therein. Pending further study of these events, let us as-sume that half of the above mentioned 5 events correspond to unavailability of
- 2. 5 the startup train; then we must add W ,s,3 X 10-2 to the unavail ability previously calculated, giving 7 X 10-2 as the unavailability of the startup train following LMFW.
3.4.3 Dominant Contributors for LOOP Trair A. Train A depends on emergency AC, so its unavailability is es-sentially the unavailability of the diesel generator. This is the dominant 11
contributor given LOOP. Note that the crosstie option changes this conclusion (Appendix C).
Train B. The dominant contributor is still the diesel-driven pump; the unavailability given LOOP changes only because backup cooling and lubrication pumps are no longer available.
Startup Train. The startup train is unavailable.
3.4.4 Dominant Contributors for LOAC Train A. Train A is unavailable, Train B. The unavailability is still dominated by the diesel-driven pump unavailability.
Startup Train. The startup train is unavailable.
3.5 ' Summary of Results 3.5.1 Results I
The AFWS unavailabilities are calculated in Table 3.4 and plotted in '
Figure 1.1.
For LMFW, the unavailability of the AFWS plus startup train is 1.2X10-5/ demand, which is "High" reliability. For LOOP, the unavailability of the AFWS is 2.6x10-4/ demand, assuming an inter unit crosstie and assuming that LOOP affects only one unit (see Appendix C). For LOOP, the reliability is " Medium". For LOAC, the unavailability is 2.1X10-2/ demand, which is
" Medium".
3.5.2 Comparison with a Typical Three-Train System LMFW. The B/B AFWS and startup train reliability is relatively high for the IRFW transient. This is because credit is given for the startup train, and because the design accords with NRC guidelines which minimize the potential for single-point failures. Given the startup train, B/B essentially has a three-train system. The AFWS alone has a failure probability of 1.8X10-4 given LMFW.
. LOOP: The reliability is reduced from the LMFW result by a factor of 60 for the LOOP initiator. This arises from a multiplicative combination of the fol lowing:
- a. The startup train is unavailable, giving a factor of 14; I
- b. The unavailability of Train A goes up by a factor of 4. l l
12
Recently, the utility has proposed an inter unit connection (see Appendix C) to improve the availability of AC to the EMD pump. This significantly improves the reliability, placing the system at the high end of the " medium" range.
LOAC: The B/B AFWS is typical in having a single AC-independent train, whose unavailability is the AFWS Jnavailability.
3o5.3 Sensitivities
- 1. Failure rate of the diesel pump (Train B). It is argued in the report that while Trojan initially appeared to have a fairly unreliable diesel pump. many of its problems have now been solved. The failure probability used in the report is based on experience at this one plant after corrective action. The value thus obtained is not unreasonable, but it is clear that a value based on the edited experience of one plant is highly uncertain relative to most of the other parameters. One also wonders whether the newness of these devices will lead to relatively lengthy maintenance outages.
- 2. The conclusion is also somewhat sensitive to the value used for diesel generator unavailability. The values presented here are calculated using the WASH-1400 value for failure to start. In any case, the Train A unavailability given LOOP is dominated by this number.
3.6 Comments on Maintenance The B/B AFWS partially depends on emergency AC power in the event of LOOP, but it does not take advantage of the redundancy of the emergency AC source.
Like many plants, B/B has one AFWS train which is operable without emergency AC; but while many plants have two separate electric-motor-driven (EMD) pump trains, each connected to a single emergency AC bus, B/B has a single EMD pump connected to one of the emergency AC buses. Thus, in the event of a LOOP, there is a significant class of double failures which incapacitate the B/B AFWS, and which would not incapacitate a typical three-train design.
Some of these double failures involve maintenance. Note that in the event of LOOP, toth AFWS trains are ultimately diesel-powered: Train A by a diesel generator, Train B by direct diesel drive. Diesel maintenance can be time-consuming, but the Limiting Conditions of Operation appear not to reflect the vulnerabilities of this particular AFWS. For example, if the Train A DG is out for maintenance, the Train B diesel generator must be demonstrated to be operable; but this is no help to the AFWS, which does not use the Train B DG, Under these circumstances, it would be logically consistent to require that Train B of the AFWS be demonstrated to be operable, and similarly, that the Train A DG be demonstrated operable whenever the Train B AFWS diesel is out for maintenance.
l 13
3.7 Differences Between BNL Assessment and Commonwealth Edison Assessment At this point, there is no single document or collection of documents which describes B/B's position as BNL presently understands it. The original report, while useful in itself, was generally considered not to have followed NRC guidelines closely enough in its calculation. The report was followed by a letter (Ref. 5), in which revised estimates of unavailability were given without supporting details. Ultimately, there was clarification of a number of points concerning both the guidelines and the analysis. Thus, a third set of numbers is called for. This is provided in the present work.
Where the guidelines are clear, there is little scope for disagreement.
Where tiie analysis goes beyond the NUREG-0611 guidelines, there are sub-stantive questions. Following are areas in which BNL differs with the utility on points which were covered in the original report in a way which has not been obsoleted by subsequent events.
- 1. Assessment of the startup train. The B/B report analyzes this train as if hardware failures were negligible compared to human error. The analysis is conducted assuming that LMFW events have' negligible impact on the hardware in this train. BNL considers that since this train is part of the main feedwater system, extreme caution is necessary in taking credit for its abil- 1 ity to mitigate failures of the main feedwater system. This is due to the {
fact that there is a relatively high likelihood that failures that cause Loss of Main Feedwater will also cause failure of the startup train. Given this l interdependence of the accident initiator (LMFW) and the mitigating system, BNL assesses the hardware unavailability given LMFW at a much higher value than does the report. (See Section 3.2).
P. Frequency of loss of cffsite power. This was nominally beyond the scope of the review, but some discussion of it is warranted because this initiator frequency varies significantly from one plant to another.
Commonwealth Edison has suggested a figure of 0.017/yr for Loop frequency, without providing full details. This seems overly optimistic; on the other hand, there are grounds for expecting B/B's LOOP frequency to be lower than average. This is discussed in Appendix A.
3.8 Common Cause l
l The B/B report mentions several possibilities for common cause failures: l the steam generators themselves, the automatic start logic, and failure to isolate the blowdown lines, to mention a few. These are beyond the scope of NUREG-0611. Additionally, the B/B report treats hypothetical potential com-monalities parametrically, using the " beta-factor" method. Thi s , too , i s i beyond the scope of NUREG-0611.
(
l One potential area of commonality is that of the valves regulating flow to the steam generators. There are four paths from each pump, so losirig flow from one pump to two SG's due to control valve failures requires three valve 1
14
failures. These are normally open, and fail open, so that prevention of feedeater flow by this mechanism is a relatively insignificant contributor.
The three trains have significantly different character; one of the two AFWS trains is powered by emergency-backed AC, and the other is direct-diesel-driven, while the startup train depends upon of fsite power.
Given a loss of of fsite power, the EMD AFWS pump depends on a diesel generator while the redundant train depends on direct diesel drive. This raises the question of possible coupling (e.g. coupled maintenance errors) between the two diesels.
For LMFW, it could be argued that the system is highly diverse, and there-fore relatively immune to common-cause failures. However, there is one methodological point to be raised which is conceptually related to " common cause" failures: the startup train failure probability must be properly con 2 ditioned on the event that a LMFW has occurred. It is inappropriate to calcu-late the startup train's average undependability, as is done in the B/B re-port. This is explained in the section on Startup Train Unavailability (Section 3.4.2.1).
At some plants, suction is a major potential source of failure of all AFWS trains. At B/B there is some redundancy in the flowpath from the CST to the pumps; and additionally, there is automatic switchever to a backup source in the event of low suction pressure. For LMFW, when the startup train is potentially available, the suction source is the hotwell, which is independent.
Suction strainers do not show up on the simplified system diagram supplied with the B/B report, but it is stated on page F-2 that the AFWS pumps have startup suction filters. These have been a source of common cause failures at other plants. Part of the problem has been failure to remove them on schedule. Since they do not belong in the system after startup, there is an argument against including them in a comparative study of different AFWS de-signs. On the other hand, the record shows that a realistic assessment of AFWS failure probability cannot ignore suction strainers. AFWS failure from this cause is roughly a 10-4 event. This has not been included in the re-sults presented here. The startup train is " diverse" in this sense (it is un-likely to be affected by plugging of AFWS strainers); on the other hand, automatic switchover of suction dces not recover the AFWS, so the startup train is the only hope, and it is available only if offsite power is avail-a bl e.
15
4 r
Figure.3.1 Simplified Fault Tree: Loss of Main Ferdwater S
i LOFW 0FFSITE POWER AvatLABLE I
TRAIN A UlAVAILABLE TRAIN 8 STARTUP TRAIN UNAVAILABLE UMYAI.U3LE ggg EMD I
TRAIN 8 Q DI SEL D
FAILURE OF D
HARDWARE t'Alf:7* P:#7 MAINT. DRIVEN PAfiUAL INITIATION FAILURES FAILS D n PU!iP FAILS ASSOCIATED WITH LOFW o
TRAIN 8 NA!:li.'
$o STARTUP TRalli TRAIN A l'Alf!T.
o TRAIN A l'AINT.
o TRAIN 8 FAINT.
o STARTUP TRAIN STARTUP TRAIN "R
I1A!;1T.
MAINT.
- -o- NOT O TRAIN A TRA N 8 Q STA TUP MAINT. FAINT. TRAIN l'AINT.
16
Figure 3.2: Simplified Fault Tree: Loss of Offsite Power It! SUFFICIENT FW to SG's O
LOOP i i TRAlt! A TRAIN B UNAVAILABLE Ut!AVAILABLE i
m I
TRAIll B TRAIN A MAINT.
PAINT.
O n, otese' enatne FAILS h EMD PUMP UI FAILS a' TRAIN A TPAIN B MAINT. MAINT.
TRAIN A TRAIN B MAINT. MAINT.
TRAIN A EMERGENCY AC UNAVAILABLE NOT m
6 DIESEL 6
DIESEL GEN. A GEN. A PAINT. FAILS 17
Figura 3.3: Simplified Fault Tree: Loss of All AC INSUFFICIENT FLOW TO SG's b
LOAC TRAIN B 4 UllAVAILABLE h
i 4
TRAIN B DIESEL DRIVEN i
MAINTENANCE PUMP FAILS l I i
4 l
4 l l
i I
i l I i
I t
1 1
t IB i I
Table 3.1 Dominant Contributors to Train A Unavailability Contributors for 34FW Value Comments (source is NUREG-0611,unlesr.
Otherwise noted)
S x 10-3 1 x 10-3 pump, Pumpfajluretostart:
4 x10 - control circuit, monthly testing.
9 3 x 10-4 Plugging of any of 2 suction valves &
1 discharge valve (10-4 each) 7 x 10-5 Failure of actuation logic (7 x 10-3) andfailuregfoperatorbackupwithin20 minutes (10- )
E=5.37x10-3.=AH(LMFW) 5.8 x 10-3 Pump maintenance 19 hr. x 0.22/720 2.1 x 10-3 Valve maintenance 7 hr. x 0.22/720 (Discharge valve)
[= 7.9 x 10-3 = AM (LMFW)
ADDITIONAL CONTRIBUTORS FOR LOOP WITHOUT CROSSTIE OPTION 3.6 x 10-2 Failure of Diesel Generator to start or Diesel Generator maintenance (WASH-1400)
E=4.14x10-2=AH(LOOP) 7.9 x 10-3 = AM (LOOP)
ADDITIONAL CCNTRIBUTORS FOR LOAC THIS TRAIN IS UNAVAILABLE GIVEN LOAC AH (LOAC) = 1.
AM (LOAC) = 0.
19
Table 3.1 (Cont.)
! Val ur Comments ADDITIONAL CONTRIBUTOR GIVEN LOOP AND CROSSTIE OPTION 1.72 x 10-3 3.6 x 10-2 x (1 x 10-2 + 2 x 10-3 + 3.6 x 10-2)
Probability of Diesel Generator failure and (operator failure or breaker failure or failure of other unit's diesel) l
}[=7.1x10-3=AH(LOOPW/ CROSSTIE) 7.9 x 10-3 = AM (LOOP W/ CROSSTIE) l l
. l 1
I i
4 I
l i 20 1
- - - , , , , _ _ _ - - . .- , , _ _ _ _ _ _ _ , , , , _ _ _ , _ . _ . , ,_ _m. __._,,,___,____m. . _ _ _ . . _ . , , _ . _ _ _ _ _ , , _ , , , , , , _ . , , , . _ _ _ , , , , , , _ , , _ , _ ,
Table 3.2 DOMINANT CONTRICUTORS TO TRAIN B UNAVAILABILITY CONTRIBUTORS FOR LMFW Value Comments (source is NUREG-0611, unless otherwise noted) 1 x 10-2 Failure probability assessed in B/B report for Diesel-driven pump, reduced from 3 x 10-2 failure probability ob Trojan. This value (1 x 10 gerved
') was at achieved after a period of " burn-in".
l 3 x 10-4 Plugging of any of 2 suction valves or 1 discharge valve (10-4 each) 7 x 10-5 Failure of actuation logic (7 x 10-3) and failure of operator backup within 20 minutes (10-2)
{=1.04x10-2=BH(LMFW) 6.4 x 10-3 Diesel maintenance 21 hr. x 0.22/720 2.1 x 10-3 Discharge valve m&intenance 7 hr. x 0.22/720
{=8.5x10-3=BM(LMFW)
ADDITIONAL CONTRIBUTORS FOR LOOP 3 x 103 Additional contribution to unreliability of diesel-driven pump due to loss of AC -
powered services (B/B report)
[=1.34x10-2=BH(LOOP) 8.5 x 10-3 = BM (LOOP) 21
Table 3.2 (Cont.)
Value Com;nents ADDITIONAL CONTRIBUTORS FOR LOAC NONE BH (LOAC) = 1.34 x 10-2 BM (LOAC) = 8.5 x 10-3 1
l l
l I
I l
l 1
I 22 I
Table 3.3 DOMIf, ANT CONTRIBUTORS TO STARTUP TRAIN UNAVAILABILITY CONTRIBUTORS FOR LMFW Value Comments 3 x 10-2 Human failure to initiate (B/B report) 2 x 10-3 B/B assessment of random hardware failures 3 x 10-2 Additional contribution ot hardware failure probability given that a LMFW has occurred (seetext)
= 6.2 x 10-2 = CH (LMFW) 5.8 x 10-3 Pump maintenance, assessed in B/B report as a " conservatism," it being doubtful that this train can be maintained during operation.
= 5.8 x 10-3 = CM (LMFW)
THIS TRAIN IS UNAVAILABLE GIVEN LOOP OR LOAC CH (LOOP) = 1 = CH (LOAC)
Table 3.4 AFWS UNAVAILABILITIES AH = Hardware & Human Error contributors to Train A unavailability.
BH = Hardware & Human Error contributors to Train B unavailability.
CH = Hardware & Human Error contributors to Startup Train unavailability.
AM = Maintenance contributions to Train A unavailability.
BM = Maintenance contributions to Train B unavailability.
CM - Maintenance contributions to Startup Train unavailability.
AH AM BH BM CH CM LWW 5.4x10-3 7.9x10-3 1.04x10-2 8.5x10-3 6.2x10-2 5.8x10-3 LOOP, NO CROSSTIE ^ 4.14x10-2 7.9x10-3 1.34x10-2 8.5x10-3 1 0 LOOP, CROSSTIE 7.1x10-3 7.9 x1,0-3 1.34x10-2 8.5x10-3 1 0 LOAC 1 0 1.34x10-2 8.5x10-3 1 0 Unavailability = AH BH CH + AM BH CH + AH BM CH + AH BH CM LWW 1.2x10-5 LOOP, NO CROSSTIE 1x10-3 LOOP (one unit) with CROSSTIE 2.6x10-4 LOAC 2.2x10-2 24 I
- 4. CONCLUSIONS The conclusions are plotted in Fig.1.1, and tabulated together with Com-monwealth Edison's results in the " Summary" at the beginning of this report.
Below is a brief discussion of the result obtained for each initiator.
LMFW: Given offsite power, three trains are potentially available to mitigate a loss of main feedwater: the two trains of the AFWS plus the startup train. There are substantial disadvantages associated with the startup train, namely that several operator actions are necessary to initiate it, and it is not truly independent of the main feedwater system. However, it is independent of the AFWS, and therefore enhances the reliability of the combined system. The B/B AFWS plus startup train therefore has "High" re-liability (unavailability 1.2 x 10-5/D), which is typical of many three-train systems.
LOOP: As originally designed, B/B is assessed at the boundary between
" Low" and " Medium" reliability (see Fig.1.1); the redundancy is only that of two trains, because the startup train is not available given LOOP, and the electric-motor-driven pump requires emergency AC. The utility proposes to back up emergency AC with an inter-unit bus tie. This has the effect of substantially reducing the unavailability of emergency AC, so that the system unavailability is essentially that of a two-train system with AC available.
The reliability is at the high end of the " Medium" range (2.6 x 10-4).
There is some argument for expecting a lower than average frequency of loss of offsite power at B/B. This is discussed in Appendix A. Formally, this does not enter the probability per dernand of AFWS failure given LOOP, but LOOP frequency does enter any discussion of expected number of failures.
LOAC: Only the diesel-driven pump train is potentially available to mitigate a loss of all AC. No dependencies which would render this train unavailable given LOAC were identified. Its hardware and maintenance unavailabilities place it in the " Medium" range of system reliability for this initiator (2.2 x 10-2/D), which is typical of single-train systems.
25
I REFERENCES
- 1. Generic Evaluation of Feedwater Transients and Small Break Loss-of-Coolant Accidents in Westinghouse-Designed Operating Plants, NUREG-0511 (Janu-ary 1980).
- 2. " Byron Units 1 & 2 Braidwood Units 1 & 2 Auxiliary Feedwater System Re-liability Analysis", GA-C16444, W. Hannaman, D. Ligon, T. Taniguchi, L.
Bowen, and T. Weis, ( August 1981).
- 3. Letter from D. F. Ross, Jr. to "All Pending Operating License Applicants of Nuclear Steam Supply Systems designed by yestinghouse and Combustion Engineering", dated March 10, 1980.
- 4. USNRC Standard Review Plan, Auxiliary Feedwater System, NUREG-0800, Re- )
vised July 1981.
- 5. Letter from T. R. Tramm to H. R. Denton, October 27, 1981.
- 6. Letter from T. R. Tramm (Commonwealth Edison) to H. R. Denton (NRC) on
" Byron Units 1 and 2 Braidwood Units 1 and 2 Auxiliary Feedwater System Reliability", dated December 15, 1981.
- 7. "A Critique of the Offshore Power Systems Risk Study for the Zion Nuclear Plant," A. J. Buslik and R. A. Bari, BNL Report #28750, (December 1980).
- 8. Loss of Offsite Power Survey Status Report, prepared by Raymond F. Scholl, Jr.(USNRC).
l l
l l
l 26
APPENDIX A LOOP FREQUENCY Although the demand unavailability is of primary interest here, the frequency of LOOP affects the conclusions which are to be drawn. The B/B study presents information concerning LOOP; we offer the following additional i nfonnation.
A survey of nuclear experience with LOOP has been conducted (7). From this survey, the following emerged:
- 1. The average LOOP rate is 0.27/ reactor yr.
- 2. The mean time to partial recovery from LOOP is 2 hout s; the probabil-ity that a given LOOP will last longer than 20 minutes is 1/2 or more (the study is ambiguous).
- 3. Most LOOP events are caused by such things as circuit break trip dur-ing relay testing, improper yard switching operations, improperly set relays, maintenance errors, etc., rather than fortuitous independent multiple outages.
- 4. Some plants have many more outages than others.
The B/B study states that apart f6dm transfonner failures (0.064 per year)
LOOP events are double line outages, which are " expected to be 0.017 per year with an average outage duration of approximately seven hours" (page A-6).
Thus, the B/B study is claiming a much lower LOOP rate than the average.
The figure of 0.017 is not derived in the report, and we cannot therefore comment on the derivation. A proper calculation of multiple line outages should take due account of the simultaneity of climatic assaults on all the lines (e.g., a tornado or lightning storm in the area threatens all the lines at once); a realistic appraisal of LOOP frequency (which is a more general ev-ent) should try to take into account the circuit breaker / yard switching / relay setting events mentioned above. ( An event which was apparently of this general type resulted in a loss of standby power at Zion 1 in 1979). It is not clear whether any of this has been considered; the study simply attributes the predicted reliability of B/B's power supply to the generally high re-liability of the midwestern grid, even tMugh nearby Palisades, Point Beach, Kewaunee and Lacrosse apparently have somewhat higher than average LOOP rates.
Although we are unable to check the B/B claim of 0.017/yr., it may be inap-propriate to assume the NUREG-0611 figure of 0.2/yr., if one wishes to be realistic. There is some empirical basis for giving Commonwealth Edison credit for a better LOOP frequency. Notir-) that Zion had never had a LOOP in its seven years of operation, Buslik and Bari remarked (8) for Zion that "at a 50% confidence level, one can say that the probability of loss of offsite 27
power event . . . i s . . 0.1/ reactor /yr." Apart from a tornado at Dresden, we know of no other LOOP events at Commonwealth Edison sites (Dresden, Zion, Quad Ci ties) . Thus, this utility can claim some su'. cess in this area, and an es-timate in the range of 0.1 is not unreasonable.
In summa.'y: on the basis of what we have been given, the B/B estimate of 0.017/yr. is optimistic. On the other hand, there is evidence that B/B can anticipate a lower than average LOOP frequency. BNL considers 0.1 to be a re-asonable guess.
I 1
28
APPENDIX B Comments on " Feed-and-Bleed" According to the SRP, the AFWS unavailability is to be considered in light of the possibility of other methods of core cooling. One such possibility is
" feed-and-bleed," in which primary coolant is pennitted to escape through a pressurizer relief valve (" bleed"), and is replenished by the high pressure injection system (" feed"). The present discussion is intended to show how credit for feed-and-bleed can qualitatively alter the conclusion that would be drawn from a survey of the AFWS enavailability alone. It is not intended as a comment on the feasibility of feed-and-bleed in general .
It has been emphasized in this review that the two-train B/B AFWS does not take advantage of the redundancy of emergency AC power. Thus, after a LOOP,
. the reliability of the B/B AFWS is relatively low, being dominate by double l failures, some of which involve the unavailability of emergency AC Train A.
The fault tree shown in Fig. B-1 shows how to feed-and-bleed changes the pic-ture. For purposes of this discussion, we have separated out the double event
" failure of all emergency AC given LOOP (which fails " feed") from other events which fail high pressure injection. It is seen that given LOOP, the dominant
- cut set is failure of Diesel Generators A and B and Train B of the AFWS. This triple event is characteristic of the dominant contributors to unavailability of a typical three-train AFWS, in which failure of both DG's fails the two EMD trains, and the third event is the failure of the AC-independent train.
Thus, by using DGB for feed-and-bleed, B/B could argue that their dominant cut sets for core damage given LOOP are the same as those at plants having three-train AFWS's. Such a conclusion is, of course, based on the assumption that failure of emergency AC is the dominant failure mode of feed-and-bleed.
Scrutiny of this assumption is beyond the scope of this review.
29
_ . . - .. . _ ~ - . . _
Figure b-la: Simplified Fault Tree Core Damage Following Extended Loss of Offsite Power
)
l CORE DAMAGE l
r i EXTENDED-LOOP FEED & GLEED AFWS UNAVAILACLE UNAVAILABLE O
T ,
i EMERGENCY TRAIN A TRAIN B
' AC UNAVAIL. UNAV. UNAV.
3 HPI HARDWARE HUMAN ERROR h h m
r AC TRAIN A TRAINi UNAV. MAINT. MAINT, h P b FAILUh5 PUf'
-o- NOT - 6 FAIL 4
l O O O TRAIN B O
TRAIN A TRAINf DGA DGA TRAIN A FAILS MAINT. MAINT. MAINT. fMINT. MAINT I
1 30
.-.. - . . - . . - , . _ . . _ ~ . _ . _ _ . _ . . _ _ . , _ , . . _
Figure B-lb: Simplified Fault Tree Core Damage Following Extended Loss of Offsite Power (Cont.)
EMERGEfiCY
\ AC UNAVAILABLE b
I I I DGA DGB '
UilAV. UllAV.
em n i i j
DGA n DGB O
DGA Mall!T. U DGB MAltlT.
FAILS h FAILS b t t QDGA Q
DGB d
DGA oDGB MAIrlT. t%It1T. MAlflT. MAltlT.
-o- t!0T 31
APPENDIX C INTER-UNIT BUS TIE A major factor in AFWS unavcilability given LOOP is the unavailability of emergency AC power on Train A. Because of this, in a very recent memorandum (attached as Appendix D), the utility has suggested that given a LOOP at Unit 1, say, there should be the option of connecting the 4160 V Bus 141 (Unit 1) to the comparable bus in Unit 2 (the Station Auxiliary Transformer on Unit 2 is sized to handle both buses, and in the event of LOOP on Unit 2 as well, it is possible to shed come loads from both buses in order to carry Unit l's AFW pump on Unit 2's diesel generator).
This raises questions about possible new problems, which are beyond the scope of the present analysis.
Given the feasibility of such an interconnection, the improvement in AFWS reliability is substantial. The unavailability of emergency AC at Train A was formerly " unavailability of diesel generator A" (3.6 x 10-2); given the bus tie, the unavailability of emergency AC is " unavailability of diesel generator A and unavailability of emergency AC from Unit 2." If we assume no LOOP at Unit 2, the unavailability of AC from Unit 2 is essentially due to human error (10-2) and circuit breakers (2 x 10-3). If we assume a LOOP on Unit 2, the unavailability of Unit 2's diesel now contributes as well, so tPat the "
total contribution is 3.6 x 10-2 x (1 x 10-2 + 2 x 10-3 + 3.6 x 19)
= 1.7 x 10-3 (see Table 3.1). .
This assumes that service water is not among the loads shed to permit Unit 2's DG to carry both buses. If service water is not available to the EMD pump, then the benefit of the crosstie is not clear for the case in which LOOP affects both units.
For a LOOP affecting only one ur.it, the varia one calculates a system unavailability of 2.6x10 gle AH is now 7.1x10-3, In perfonning this. calculation, the utility obtains 9 x 10-5, which is a factor of 3 lowe~r than the result given above. This discrepancy arises chiefly because the Train B diesel-driven pump unavailability is taken by the utility in Table in 3.4.
the new Origicalculation nal ly , to be 5 x 10-3 rather than 1 x now, ig;2 the as used utility argues that 5 x 10 ghe utility assessed corresponds thetopump more nearly at mandate the NRC 10-(knowing nothing about the pump, one could infer 5 x 10-3 from a table given in NUREG-0611, which does not explicitly cover pumps of this type). A more natural adaptation of the NRC prescribed data base would be to use diesel generator data, which would lie closer to the 10-2 figure.
32
F APPENDIX D Letter from T. T. Tramm (Commonwealth Edison) to H. R. Denton (NRC) with memorandum.
e I
l 33
/ Commonwealth Edison
[ C ) one First Nabonti Plata. Chicago. Itimois (N C 7 Addr1ss R.; ply to: Post Offica Box 767
/ Chicago,1:linois 60690 December 15, 1981 Mr. Harold R. Denton, Director Of fice of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555
Subject:
Byron Station Units 1 and 2 Braidwood Station Units 1 and 2 Auxiliary Feedwater System Reliability NRC Docket Nos. 50-454/455/456/457 References (a): September 18, 1981, letter from T.R. Tramm to H. R. Denton.
(b): October 27, 1981, letter from
- T.R. Tramm to H. R. Denton, i
(c): October 30, 1981, letter from D. G. Eisenhut to L. O. De1 George.
(d): November 10, 1981, letter from T.R. Tramm to H. R . Denton. l
Dear Mr. Denton:
This is to provide additional information regarding the reliability of th'e Byron /Braidwood Auxiliary Feedwater System ( AFWS) .
References (a), (b) and (d) provided the results of analyses of AFWS reliability for Byron and Braidwood. During the review of these analyses with the NRC staf f, certain recalculations were requested. The recalculations have been completed and are i
summarized in the enclosure to this letter. The reanalysis conforms i
to SRP 10.4.9 by taking An credit for a second auxiliary gower supply unavailability of 9.4 x 10- per demand to the feedwater pump.
has been demonstrated.
Upon receipt of staff concurrence with this approach, the response to FSAR Question 10.53 will be revised to include the results of this reanalysis.
Please address questions regarding this matter to this office. ,
34
H. R. Denton Decembe r 15, 1981 One (1) signed original and fif ty-nine ',59) copies o f this letter and the enclosure are provided for your review and approval.
Very truly yours, hr W T.R. Tramm Nuclear Licensing Administrator
Enclosure:
" Byron /Braidwood Stations, Auxiliary Feedwater System Reliability Analysis, Second Recalculation," December 14, 1981.
cc: 1. Papazoglou, Brookhaven Nt'l Lab #
TRT/lm 3053N 35
t-December 14, 1981
Subject:
Byron /Braidwood Stations Auxiliary Feedwater System Reliability Analysis Second Recalculation
- Summary of Results Unavailability per Demand f
Initiating Report (I) First Second i Event Recalculation (2) Recalculation LOOP 6.4 X 10-4 8.9X10-5 9.4X10-5 l
Assumptions LMFW coincident with LOOP (3) l.
LOOP considered on the unit under study only 4,
- 2. e 3.- Credit taken for manual breaker closure to BUS 141 from Bus 241
- 4. Other assumptions per NUREG-0611 l
Discussion ESF Bus 141, the bus that supplies the 1A motor driven auxiliary f eedw ater pump, is capable of.being supplied from one of three sources:
The system aux transformer (SAT 142-1); the diesel generator (DG 1A); or
- - the Unit 2 ESF Bus 241. (See Figure 1). i In the event of a LOOP event, Bus 141 automatically transfers to DG 1A.
i If DG 1A fails to start, the operator is able to close two breakers from the control room to feed Bus 141 f rom Bus 241, which still has power f rom i the system aux transformer associated with that bus or from diesel l generator 2A. The operator is capable of closing these breakers within the 20 minute steam generator boil dry time assumed in this analysis.
The capability of supplying the Unit 1 ESF busses from the Unit 2 ESF e busses is not a new feature and has existed in the B/B design since its beginning. In fact these crossties are the alternate source of offsite The power to the ESF busses required by our Technical Specifications.
L NRC Staff has reviewed this system and has found it to be acceptable.
36
_ _ _ - - - _ - - _ . . _ _ _ _ . _ _ _ . ~ _ _ _ _ . _ - . . . _ . - - - _ _ - --. _ -
f The crosstie between Bus 141 and Bus 241 is capable of feeding all the .
safety loads of Bus 141. The SAT on Unit 2 is more than adeauately sized to handle the safety loads of both Bus 141 and Bus 241, hence the use of the crosstie does not compromise the safety of Unit 2.
The assumption of loss of off site power to only one unit is supported by the re.ults presented in Table 1. The total f requency of loss of of' ite power to a single unit is 0.605 events / year. The major contribution o the f requency is f ailure of the SAT's, which is 0.3 per transformer w ith a total of 0.6 for the two transformers supplying a single unit. The second contributor to the LOOP per unit is a f ault associated with the line supplying the ring bus section that feeds the SAT's of a unit. This contribution is 0.005 events / year.
As can be seen f rom Table 1, the events that cause loss of off site power to both units is 3.6 X 10-4, the major contributor being tornado damage (Attachment A). This is a small number compared to the overall estimated f requency of LOOP to the unit, and hence, we have concluded that the LOOP initiating event is applicable to only one unit.
How e ve r, under the condition of LOOP to both units, Bus 141 my still be fed f rom Bus 241 when Bus 241 is being supplied f rom diesel / generator
'2A. There are no breaker interlocks to prevent this operation.
The crosstie operation will be conducted under a specific set of l circumstances and controlled by a well defined emergency operating !
procedure. The procedure will cover the following conditions:
(1) LOOP has occurred on Unit 1 and there is no auxiliary f eedwate'r being supplied to the Unit 1 steam generators because the 1A die el/ generator f ailsto start and the IB Auxiliary Feedwater pump faiis to start.
(2) a) LOOP has not occurred on Unit 2. The crosstie breakers may be closed to Bus 141 to pick up AFW pump 1A.
or b) LOOP has occurred on Unit 2. Unit 2 AFWS is operating normally and only normal s hutd ow n loads are present. Both of the Unit 2 emergency diesels have started and are carrying their respective ESF busses. The crosstie breakcrs may be closed to pick up Bus 141 and the 1A AFW pump. Nonessential loads on Bus 241 may be stripped in order to accomodate the 1A AFW pump load. Diesel generator 2A operation must be monitored closely to ensure that the generator is not overloaded.
37
The above operations can be completed within 20 minutes of the initiating
@ vent, which is the steam generator boil dry time assumed in this analysis.
We wish to emphasize that LOOP to both units is not an event that we consider to be credible. Nefertheless the capability exists ta ensure sufficient auxiliary f eedwater flow to the faulted unit.
- The number associated with remote manual closure 6f the two breekers l
required to supply Bus 141 from Bus 241 is 1.2 X 10-2 per demand.
There are two components to this unavailability number: 1X10-2 per aanual action (this is considered to be one manual action) and 1 X 10-3 unavailability per demand for each of the two breakers. The unavajlahijity 3X10 . (3) per demand of diesel generator l A was chosen to be Conclusion (4)
Credit for use of two auxiliary power supplies to Bus 141 leaves us with an AFWS unavailability per demand of 9.4X10-D, which meets the requirements of SRP 10.4.9.
! l l
l LAB:mnh 0984b*
l 38
Notes (1) Table 4.3, under " Automatic Startup with Manual Backup" p 4-14, Byron Units 1 and 2, Braidwood Units 1 and 2, Auxiliary Feedwater System Reliability Analysis, Final Report, dated August, 1981.
(2) October 27, 1981 letter from T.R. Tramm to H.R. Denton.
(3) NUREG-0611 assumed a diesel generator unreliability number of 1X10-2 We have not recalculated our system unavailability using i
the lower diesel generator unreliabilit* , however it stands to reason that the system unavailability would yet be lower than the 9.4X10-5 presented in the results section.
(4) Other approaches to solving our dilemma of adequate AFWS reliability without a second motor driven pump supplied f rom the
,- - other ESF bus were considered. They are discussed in Attachment 2.
t LAB: mph 0984b*
39
PRELIMNf4V
~f~4 big, k SYSTEM PLANNING DEPARTMENT November 23, 1981 Byron Station Loss of System Aux. Power Supply Approx.
Frequency Dur ation (Events / Year) (Hours)
Crld Collapse or Islanding < 10 -8 34 5 kV Switchyard destroyed 3.6 x 10 by tornado 4-345 kV lines outa;;ed due 2.5 x 10~ (1) to independent and/qr common ,
node outages Subtotal-eventE\ isolating 4 l the entire statio i from the 3. 6 x 10 interconnected syitem
'l Cherry Valley line outage causes 0.005 1 LBB isolating SAT (5'i $wre.1)
Subtotal - 34 5 kV supply to SAT 0.0054 SAT outage (2) 0.6 110 Subtotal - events outaging j one SAT 0,4 05 Total - all evmts outaging 0.605 SAT (1) Frequency increases during line maintenance outages rhich average less than 1 day per line per year.
(2) Duration of outage on the faulted transformer. The other transformer
(
can generally be retained to service by switching within 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. Frequency i is twice that for single full-sized transformer.
O e he _ ug 6 ed aw C% men M % b ,
m.h &Wy ay ad.
l l
40
+
s .. i . -a g~g3 =AS-d 3 g -t ., ' t -
- ==:.~
w e % 1.
IC:' I
" 'c'il:= I w
=- Er
- u. - -d-a
=~ --
=_
x
,m...e
. t :==s.
~ r. - r- - ,i> ==.
u:.n .- _. Q.'.6=% --
~ .
a r m= :=s,
. g, ,.. .
- =:::=
j ;:w, .. .
-c
-... _ .. . p.3. ,., 'c.__......_.e
' _ ~n.n . , , > . _ . C. . - - _ ~ _ -
m.. h,..._ ..,e
- ,q> . .,,. ..r e , ...
T ..=_ m. . .
, , r, ..
, , . ,. _, . . ,_,,._,,,,,_,3,3._._ c ifk'k'kthdh,k'k'k'f[fIif'k,'f'f'if'ff'fifif'[' 'ff['fffff'fd'dd'N'f['fhf'
i , il ll
!.i,i i h, :!:I!.i. II !!f I aaseii !;4,N i. li +. ii .I hii 0 1
- J i !, si i !-
a ii
,g . nn et di i ih'.e i:i n n. :iss u i
n a ; h ,:n.:; ,
i s.
h dI!!8i n ba $g z u e s !>i.!il . .
n_. l .
1 1-- .
1~ c- c J 1
T,== T,e . 1
- . . ~
. -r 2 T== 1=.v -er
" _ I.==1=7-r .,. p .g- e a
- pa ,_, =,=. . . . . , ,, ,.. .,, Ir
. .. 7(, fc,,f. .:._
4 T ;re ..,e u) _,_g T == -
b ese , a, '~; * ' ~ > ::* x , % % .- . - ~ - +
r
.1.n= .. ::
= r
. .w. %_,1.' .
uc,;f.?.:
.c, 55y .54.c , >;;=p,:::p-"~
- m :.::gb:.r: j-
- =
. " ~ 1 - m. ..
= =.
np
( ,..,_.,u
_ . .it..
- . . _f a.,h> .
. .... ._.- ,... F,m.
... a:::::
,a =r.-%.. ;;.
t.)
)
'C" CT--a \-" -[+ :..~*
g :0. y . . 1 c:U::
g'v.r* sr.?, e m== sp >>seset h se I m *>6 >s
..,,....s
$es s ,, ~
.. .g. a...a. L.. ' ) { t.o en==I, g 3_ ... __ t"> a=r '=_s
, 6, @,
- e . ..
,IS g
. . . . ..= . "
... p %':.. Jm.g;a o 2....a. 6 ,, 3,je i a.
2 .+ -
.. ~
- M =-a .h hW' ?
e .: $
~ ~ " '
.cna .y, Cui ; g; . d 5 g, $. h h=M i F
= .u.Ipl oi
- , f l 5,, , !! nI* r,1, e ") s n_L
-,h ~ g . ,}C. . 'Q ;;s. y h=
3 .,,,t u. . 2 p ,,i. . -: . , , p-s,
- , fg h.
s 19. - 9?,..=
g7
..aIs..c. 'J"c
. m.r.
r ,,,
y,
.;= . ,, ,
-.- :. ..t 0. -
s n ,,t 21 h ir ;s v.
~~ =t~ ~ __w) y .., .
- .. ', _ . . . . ~1 r 1 r ~ -%
_u
? ('. ,y~' '{, .,. *%.
Y t>w------O";". " i eeucteam."sasarv F" estaneo
- rn =l=. t%> n .
f>-- ,>..-,-p'.." , . . . =*"="a-**
d) ) ) ') > '
4
~
.r3"**"**"
& n.s.
9 R 3
E3 " ' *
"'"'.,'.,*.Y,i) f
.,*..s..e, ' a . e' " " 'gyg' 2 4*C 0s' "* .
. " *>.' ' " *3"' -'{ .
,, , s a 4 $ sM 51 M r f h",,'
1- ,[ $ $i =_": j b ,.,, b b e b; -N_ii-. .
_p g=a'a=' =
,b C 4. '{ ' C, '4, e e . ggggggg g
.,= L ,
H n
yn c,. ,;
ij 3.) ii n y ac..
- 4. c - P"'4 W.:2* '~~a* -
T r:wu -= r..ws y .
y - - -
se .a .. .s e e aa S
- - s-__ y .
, . . . . . .... .c. .< - m m) e
?
+
Y 7 2 -
2=-J . f'12;97 =i.
l ..
r
[ n, y 10!%=:53 1.
- ..,-.. . j c,); __
=
./ . _
=
f r m a_
y x.-,,p.y -
i .-1 r =e- e_
n wa::.: t, ~~1k w=.w=,
r
--- -r
_Ll rn.s 1
tr l l ' =. . ;,,.:. ,;, -- K =G / M,:
1 .=o=
m
'ih rih g,Y g g ,, g - n.w t.-
- $p hywa~ .. .:_ . , .. = 4 = , s=. ., _, ;.e ,c..
- aseJ)
.:.p,I: . _ . . - ,' > ,E p ._ _ - .,= p = > hL.__. ,,
EE E E} A E Ai E)
E E ) h b*E I I) ) f) f) E) ,f) f) f) E) f) E f) b) I) A hI ), ) j
)I f)l}E) 3 k)I f,) 1)II I I!)I k) d")d)k) b kl}A) A)k)l,5)l} } }}tI }) Ik)k)k)!I it h hk)k)k k,)d)b I}. !}) ' b) I i 1
;. ii . ii,.s M 4.[u
!.u if !!i;8s!iO, !r*
S ;i:1J n.t W !!!!II dii n
!! si!! :
..;; ; !! si !!<h .
sin ;; . 11
!!!!; : il n !!u ilis
. !i y!!ildM II a P.!i u:;! b;! :; w!I! N..
- v. /)t: one.a A d- a:: A := L- [
=~,.: $--A-: T :-T,r.=. .=. T,c.-~'(- L: e i2
,T, _==T,..=. . , , r . p, m
L,..-. T q. M.7_g g.n r..- l-
, ==
+
n- .- - ~
$wa
'sCM ..- p .ar 6)= = 4,:::
em >> :: i, ..
+ n
;6 - = = 6
't - ---- Q, ,u cy* me
- ~,v.7 ".,iu l '
C g E ' p"." "** q- N_y fi$ Q
, .1._. . =Q.,
e_._ . ru
- = .a ., . _.
g . . . . , , ,
*'"? T**lI 7m 08
.w s4 r. .e. < r, .
,,. .,,,4,,, ,,, $ ">.s u *>>> > $ $i ', s s$ so $
,.s tj,gn.w-s.%p.=x.. - "I == $ s ) ) ) s. ,$,. . . . ,
,j yL c.
.= g
., ," g n.g L.. . A, m,.m... 3 1 1 g,. ~ ,2 , .,.3,, .qf_., _. ,y,y..,, p%..3 c
!.e: .
,, ,, C,M,
;sI.
. _ , . , s'_ ., 3 .
- V JJ'r
~?.t., g; "h.,.,.h>Jt Q *c"d2) v. fs r 0. 2 m..
I ' ' " " j !TGa,.* _ g' et p !=r 2f .*I J 3DS' '" ' 3 2 2 !*l
/ aupC. i 6-
- 3
, 1- j !.s 3 .3 u.;7JO st
** ::$~g'r g e 3
=2*r= **t ::gr e
o
*;r *a .a 3
'? d .Y.,, < ,,,, 7 I 1 I peUCitat SApEIT Sit ATED p%..
/'s -
*g ta T ii A >--. --------= = p- > -+---p"A g
- 2. /-..p ,h **~** m.~.~g'. C,%--
* = = * * ~ ~
egr.,,., o .., . ._m.,f"f.,
. ,, ry r!2
.g.i
.)
; ,3
,=.1 , ,'
,"3
'm_ ,-zt.' "' $ .'
_rg- . x~ c = a===== l, .'7 3 "
.f' O f, y;-'B;in
"',' J ,,
h,hdM. :.:..- E=="~'1
~
- 1. .cd, y*
.=-
A { n p.,,.,_ M,.:
.c.
o$'d y o" f y .= 1"~ :' ' kd.9: Eni' .'..
- l' i -- F-
,"""'*--*""- m. . ;,, .......,..mu .,
GAT N IUshs 14 2.-l
- e 4., gy
^ susts CHERRY VALLEY L.OG21 BULK <
- d 6 c c.e xv POWER SF BWE N :^ Sust$
Oi - i42-2 SYSTEM wttsow t.. issos us Am.$ <l-
- 6 4'* SuSE5 ic_i GRID G3(JO ,
WEMPLETOWN L.OG24 8 I 2f 4P
&[] N:o
_ o.9 a.v ausas 4
~
@ RED
- 241-1
. E--. +.BUSES i av CHERR.Y VALLEY L. Oco22 "$tf[. h iIj Q
~
- ,e AA ~ 4.9BUSES xv 241-2
- e 4 i gy
^ auses BYRON STATION 34sKv Bus CONFIGURATIOP SYSTEM PLANHtNG DEPARTMEMT NOVEMBER 20,1984
c Ancuer November 18, 1981 EM Pmd.G SEPARTMENT Complete Outage of Byron Station off-Site Power Due to Tornadoes The fr equency of tornadoes outaging all sources of offsite power ! . to Byron Station has beeen estimated. Such outage could occur either from a tornado damaging each of the rights-of-way leading to the station, or from a tornado hitting the switchyard at the station. The present Byron transmission plan utilizes 3 rights-of-way leaving the station site in different direc.tions: to Cherry Valley (21 miles northeast; lines leave site in an easterly direction); Nelson (33 miles southwest; line leaves site in a southerly direction); and Wempletown (30 miles north); line leaves site in a northwesterly direction) . A reasonably straight tornado path cannot intersect all three rights-of-way except by hitting the relatively small area of the station switchyard. The frequency of this event is estimated as once ir 2800 years. If the Wempletown right-of-way were eliminated, tornadocs heading in the prevailing northeasterly direction and passing within five or six miles of the station could outage both of the remaining rights-of-way. We estimate the frequency of this occurrence as once in 350 years, or eight times as fr equent. 1 In this case the lines were modelled to the points where they intersect the ' present Nelson-Cherry Valley right-of-way, as there is a negligible risk of a tornado intersecting two rights-of-way on the same line but separated by several j miles. The results are based on 102 tornado path lengths and compass headings observed in the Chicago area (the " Centennial" data) whose summary statistics are as follows: Standard Mean Deviation Range Path Length (miles) 6.34 10.45 0.52 - 68.4 Compass Heading (degrees) 64.5 19.1 19 - 135 l (00 = North, 90 = East) The station switchyard was assumed to present a 900 foot wide " target" to any tornado heading, allowing for the tornado path width. The analysis method is an improved computerized version of that presented by Mr. John Teles at the 1980 American Power Confererice, with certain analytical refinements relating to the path lengths and compass headings of tornadoes. 7 h '. bf A. W. Schneider, Jr. System Planning Department Approved : -
- 44
l Attachment 2 thar Approaches 3 attempted to show that restoring off site power to Unit I within the team generator boil dry time, or even within the core uncovery time, is credible event. Though we certainly have good f eelings that f ollowing ne transformer failure, the most likely event, the operator can restore he other system aux transformer within a sufficiently short time eriod. Similarly, if a Cherry Valley Line outage occurs, we are casonably certain that station personnel can isolate the f ault and 3: tore power to the system aux transformers. nf ortunately we have not compiled suf ficient data to support our claim. e are continuing to search out data that will provide reasonable eliability numbers f or operator action f ollowing either a transformer allure or a line outage. Until such time we cannot assume that offsite ower is restored. inh 184b* 45
I#'C 0 l'1 ronu D U.S. NUCLEAR HEGULATonY CCMM:S".l ON BIDLIOGRApHIC DATA SliEET NUREG/CR-3096 h 37:1/nt! REG 51633 I TITLE AND & 'itTLE (A cd Volume No.. of mprmenstel 2. lleave blanki Review ot the Byron / Braidwood Units 1 and 2 f Auxiliary redwater System Reliability Analysis 3. REctPIEN CESSION NO. [7.AU W ORISI 5. D ATE rig'P o rtT CC"F LE TE D j R. Youngblood, A. Papazoglou ww xs jno Decfaber 1982
- 9. rs nr cre.uNG ORGANIZA . 'N NAYE AND M AluNG ACOREOS Unctuce Zip Codel, Ql TE PEPORT 105UEO Department of Nucl ar Energy Brookhaven"Na tional .aboratory fem p November j ve ne 1983 Upton, New York 11.'3 6- (Lee e bienas
- 6. (Leave blank)
- 12. SPCNSORING ORG ANIZ ATION NAML %NO M AILING ADDRESO //nc/uce 1,0 Code
. . . 10. PROJE CT*T ASK,wORfC UNIT NO Division of Safety Techno ' y Office Of Nuclear Regulator, Regulation ti. ccNinAcT No.
U. S. Nuclear Regulatory Com ssion Washington, D. C. 20555 A3393
- 13. TYPE OF REPORT PE RIOD CoV E RE D IIne!ssere careJJ
- 15. SUPPLEMENT ARY NOTES 14,(t,,,,er,,aj IG ACSTD ACT (200 words or less)
This report presents the results a review of the Auxiliary Feedwater System Reliability Analysis for Byron Units an 2/Braidwood Units 1 and 2. The objective of this report is to estimate the pr.abil y that the Auxiliary Feedwater System will fail to perform its mission for eac' of thr different initiators: (1) Icss of main feedwater with offsite power avail s.le , (2) 1 ss of offsite power, (3) loss of all 460 VAC power. The scope, methodolog , and failur data are prescribed by NUREG-0611, Appendix III. The results are c . pared with th e obtained in NUREC-0611 for other Westinghouse plants.
~
17 MC Y WCROS A*:0 OC,CUME N T Ars LYSIS 17a D( SCl pq= T o ns Reliability Loss of offsite power Availability Loss of all 460V AC power Auxiliary Feedwater ' stem NUREC-0611 Byron Units 1 & 2 Braidwood Units 1&2 PWR
~
Losa of main feedwater I I" 'C5 Irer F :e pt ,r . M L I E P *A C
'NL,' 3 8 t i T . 5 7 ;,- . t ; r -Ce**r .!~-
. . n I
- m .w l1 *3-; ? ;. . .L 5
. l
. . 00CldMliiffd l N *;. '
* . T',oax, : .s +c.
_ _ . . _ . . .__ _ _ _ _ _ . . . . _m.. ._. . .mm__ . ._ _._ . _ . _ . __
; UNITED STATES Founta ctess san Postaut a f f Es raio NUCLEAR REGULATORY COMMISSION !
WASHINGTON, D.C. 20 lie 6 ,$5l"jc j n a ir . n t OFFICIAL SUSINESS ' PENALTY FOR PRIVATE USE, $300 i 7 i 1 i '1 i l r 1 I i i )i
- 120555078877 1 1AN h
9 ! US NRL ADM-D IV OF TIDC -
, POLICY C PUB MGT BR-PDR NUREG 1 W-501 WASHINGTON DC 20555 e 4
k i' f 1 i l .- I 4 i l l l i 1 l l 1 e l l 1 i i i i i}}