ML20132B689
Text
.
t (115trihutton n
'indeit'fI'1'e'
('
V,il Pdq. ftle e m a,-h e re,
}
fiOV : 0191 e
- ocket No. 50-346
)
i PE'100R10UM F00-Cus C. Lainas, Assistant Director for Opert. ting Reactors, Division of Licensing 4
FROM:
L. S. Rubenstein, Assistant Director for Core and Plant "i
Syster.s. Division of Systems Integration
(;
4
SUBJECT:
Aux!LIARY FEEDWATER RELIABILITY STUOY - DAvlS CESSE UNIT NO. 1 dlp.
g t.'e have reviewed the Itcensee's sutnittal dated Decer ber 31, 1931, which pro.
L j
vid:d the reliability study for the Davis Besse Unit I aust11ary feedwater I
systen.
The enclosed evaluation is based on the licensee's reliability
[
]I analysis of the upgraded systen and was perforned by the Brookhaven National
+
Laboratory.
t!e conclude that the unav exceedsthenumericalcriterionof10*gilabiltgyoftheDavis-BesseAfr$
to 10-per denand i. resented in the r-4 Standard Review Plan.
t Original s3 Fed by J
L S. RubensicIn p
L. S. Rubenstein, Anistant Director for Core and Plant Systens I
Division of Systens Integration
Enclosure:
As Stated cc t/ enclosure:
R. P.attsen L. Rubenstein J. Stolz l {.y O. Pare y
q R. Lobel
- s';
A. CeAgazio J. Fidgely I
d I
Contact:
J. Ridgely T
X29566 4
I g$ 1
[cyc.cccrJ9 ino104 00: u osc00;:e f.'j; P
1 b
N
{
05!:ASC D5 ! M.0 DS!:M1 if3
' JP idqal y /. t/ >
i as j ALobel
' 00Parr (Ru
' ' fi n l
.F l
i l
11/* // 83 l
11Ak33 11/hq3 11//gs 1 j
p, j
j
.-.L..-.
2
l 4
.#. g. ne w,,79, UNITED STAf ts
[
h NUCLE AD RECULATCRY COMMISSION n asa.ia.ct one o e ro+sa g
I
%,,,,,/
February 4,1983 MEMORANDUM FOR:
- 0. Parr, Chief Auxiliary Systems Branch Division of Systems Integration FROM:
A. Thadani, Chief Reliability and Risk Assessment Branch Division of Safety Technology
SUBJECT:
EVALUATION OF THE RELIABILITY STUOY FOR DAv!S BESSE UNil 1 UPCRACED AUXlLIARY FEE 0 WATER SYSTEM 1
Enclosed is our evaluation of the reliability study for the Davis-Besse Unit I auxiliary feedwater system.
This evaluation is based on the review of the licensee's reliability analysis of the upgraded Davis-Besse auxiliary feed-water system performed by Brookhaven National Laboratory (DNL).
The Davis-Besse AFWS is in the low.eliab Qity ra ge and its s. attability exceeds the numerical criterion of 10 to 10 ' failure pr.' ability per demand presented in the Standard Review
'an (SRP).
The addition of an
~ l independent third train will likely ir.;..ove the AFWS unavailability to the
(
accepta01e range of the $RP criterion.
Eliminating the cormen cause arn single point failures, as discussed in t'.w enclosure, will likely improve 1{
the AFWS unavailability by a factor of 2.
If you have any questions, please contact John Tsao (x27422).
I. ' ' 'i. $ ' '
l i
j Ashot Thadant, Chief 1
Reli,tbility and Risk Assessment Branch Division of Safety Technology j
l
Enclosure:
j As stated 4,n J
cc:
T. Speis
.N H. Ernst
'3 J. Ridgely a
R. Lobe!
/
a f
.l C
4 D l..,
')
l 1
,.IS i
.j,.
s.
I Evaluation of the Reliability Study Davis-Besse Unit 1 Upgraded Auxiliary Feedwater System 1.0 Introduction As a result of the Three Mile Island accident, the hRC issued a letter 1 dated requirine the licensees to perform a reliability study on March 20, 1980, The licensee Toledo Edison Company, the Auxiliary Feedwater $ystem (AFWS).
responded by subritting a reliabiltty study 2 for the Upgraded Davis-Besse g
Unit 1 AFWS in January 1982.
Our contractor, Brookhaven National Laboratory (BNL), reviewed the The review licensee's report and perfernd an independent ar41ysis3.
addressed the following three transient conditions, as required in huREG-C6114, for determining the reliability of the AFWS:
1.
LFFW - Loss of 14ain Feedwater 2.
LOOP - Loss of Of fsite Power / Loss of Fiain feedwater l
3.
LOAC - Loss of all AC Power / Loss of liain Feedwater I
Following the guidelines presented in NUREG-0611, the review evaluates mest 1
of AFWS component fatlures but excludes the high energy pipe breaks and the The system boundary external events e.g., earthquake, fire, and floed.
includes the water sources (e.g., condensate storage tanks) and the AFWS nozzles on the steam generators but excludes the AfkS ring header inside of l
Some support systems such as the Service Water System l
the steam generators.
are reviewed qualitatively.
2.0 Otscussion Both pumps are self-The Davis-Besse AFWS has two turbine driven pumps.
cooled and take suction from a cormon header which is normally supplied by
)
t the condensate storage tank (CST) through two manual valves, a check valve The pumps have pressure switches detecting loss and a fine.nesh strainer.
of suction in which the water supply is automatically switched from the CST to the Service Water System.
The Fire Protection Systen is also a backup water source for the AFWS; however, this switchover must be cone manually.
J A
e l
ik y
ad' be**
5 A*85 'M LT S5n4MMHW8W Mv4M.,._%WfMMfQIQDN[th*)}h/sYw 49
'i
,i
- l i
The lir.ensee analyzed four AFWS configurations:
1.
'7re-TM1" configuration:
This is the AFWS that existed in March 1979.
3.
" Post TMl" configuration:
This is the AFWS configuration that contains i
THI-related changes which includes a procedure for providing core cooling t
i using the main feedwater startup pump, reactor coolant systera mke up rt.t.p a#.o tht, ptiot operated relief valve (startup pump with feed and i
bleed).
This written procedure was planned to be implemented in the 1982 refueling outage.
3.
Third Train" configuration:
This configuration uses a larger impeller and higher horsepower motor on the existing Nie feedwater startup pur.p.
{
in an altered discharge mode to provide a back up third, train of aunt-11ery feedwater.
4
" Analysis Based" configuration:
This configuration includes the writter.
procedure for using startup pump with feed and bleed, similar to tht:
case 't above. This configuration also incorporates modifications addressing several dominant failure contributors identifieu in the Itcersee's study.
ShL analyzed a AFWS system that is comparable to the " Analysis Baseo" config-uratton, without constdering the min f eedwater Startup train and feed and bleed operation as proposed by the Itcensee.
The reason is that the startup punt, has about 3C0 GPM capacity which is less than the required 800 GPM for decay heat rerosal and the feed and bleed operation is not allomeo by the NUREG-0611 nethodology.
Therefore, the SNL results should be compared to the licensee's " post-Till" results.
3.0 Results 7able 1 below compares the results of the AFWS unavailability for the three transient cases based on the SNL and applicant analysis.
I
?able 1:
Davis Besse AFWS Unavailability for Three Transient Cases BNL (stimate Licensee Estim te for the Four Configurations Per Demand Per Demand Third Analysis-Pye-TM1 Post-TM!
Train based 1.
LMFW 1.6 10*3 3.3n10-2 6.6x10*4 4.5 10 5 3.3x10-5 I
2.
LOOP 2.8x10~3 4.1x10-2 5.5x10'3 1.4:10~4 9.3x10-5 3.
LOAC 3.4x10-2 8.8n10-2 1.9 10 2 1.9:10-2 1.1 10-2
. 3.1 Dominant Failure Contribution Le t'
Main Feedwater (LMFW) Case 1.
Failure of both turbine driven punps to start.
2.
Maintenance error of valve M5730 3.
Actu. tion logic failure including independent pressure switch failures, spurious AFWS isolation, etc.
4 Failure of steam admission valves to open.
Less o' Offsite Power (LOOP) Case 1.
All the dominant failure contributors in the LMFW case.
2.
Failure of diesel generator 1, which supplies power to the steam semission valves M5106A, MS107, and MS107A.
Loss of All AC Power (LOAC) Case 1.
All dorinant failure cor.tributors in LMFW and LOOP cases.
2.
The TOP 2 train is unavailable because steam admission valves H5106A MS107, and MS107A which depend on AC power, fail to open during a LOAC event.
3.2 Comron Cause/5 ingle Point Failures Because lack of diversity in the AFW pumps, the Davis-Besse AFWS is particularly susceptible to potential common cause failures.
The following failures were obtainee from LERs:
1 1.
Dirt January 1979 - steam admission valves M5106 ard MS106A for punp AFVP-1 failed to operate because of dirt ac:amulation.
The_ potential s
8 common failure is that dirt may also affect pump AFWP-2 and cause both pumps to fail.
2.
Isolation valves closed Novemoer 1977 - motor operated control valves AF599 and AF608 were found to be closed and, thus, isolated both steam generators from the AFW flow path.
Valve AF599 controls AFW flow to steam generator 1 and valve AF608 to steam generator 2.
The reason of the failure is due to omission in the test and maintenance procedures, which failed to alert the operator to reset the valve logic for these valves.
3.
Pressure switch failures f
May 21, 1979 - common cause failure of six AFW pump suction pressure switches due to setpoint drif t and/or malfunction of components.
If l
1
r w
4
[
these switches fall high, pumps will not be able to automatically L
switch over to alternative weter sources in the event of a loss of 4,
suction.
If switches fall low, it will inhibit steam admission to the pump turbines and will fall the entire AFWS.
4.
5 trainers One fine mesh strainer is located on the suction side of each AFW 3
There are no LERs reported on strainer blocking the flow d
pump.
path in Davis-Besss but this failure has occurred in other plants 5.
toss of steam sueoly One potentia 1' common cause failure is loss of steam to both turbines because of steam generator depressurization which could stem f rota two relief valve's failing to'close, or operator error in opening the dump valve.
6.
Improper allonnent of suction October 20, 1977 - it was discovered that both AFW pump suctions were 3
aligned to the deaerator storage tanks which was not an alternative d
water source.
The tank's water temperature is greater than 200*F N
which could damage the turbine bearings in both AFW pumps.
An oper-ater improperly aligned both AFW pumps to deserator tanks thrr) ugh L
valve FW85 instead of FW32.
The Service Water System should be the alternative water source.
Since this incident, valve FW85 has been lock-closed and adders to the valve verification list.
]
4.0 Conclusions The Davis-Besse AFWS unavailability is estimated to be in the low reliability range and outside of the acceptable numerical criterion of the SRP.
The tidition of an independent third AFW train will likely improve the AFWS unavailability to the acceptable range of the SRP criterion.
Eliminating the common cause and single point failures will likely improve the AFWS unavailability oy a factor of 2.
I TABLE 2 Comir. ant Failure Centributors to AFWS Unavailability A
Turtsine-driven puit train 1, f ailures per demand BNL Licensee Fs t +r e te Estinate Description of Components 5x10~4 Human error in aligning mctor operated valves: AF608, AF3870, AF360, ICS388 1x10-4 1.5 10-2 Hardware failure cf motor operated valves:
AF608, AF3870, AF360 1 x 10~ 4 Hardware failure cf check valves:
AF39, AF72 1x10-3 3.6x10-2 TDP-1 fails to start 7x10-3 Actuation logic failure, including independent pressure switch failures, spurious isolation etc.
l 5.8x10~3 5.81x10~3 Pump unavailable due to w intenance 2.1x10~3 Valve maintenance: AF360 AF3870. AF3869 3x10~#
1.94x10~3 Pump unavailable due to test 3a10-3 1x10~4 Failure of steam admissier :
(valve M5106 failure); x 3.9x10-2 (M51064 failure or diesel generator 1 failure) 3.6x10~2 Failure of diesel generter, which supplies AC power to stean adnissicn valves of pump AFWP-1 NOTES We cannot find this failure rate in the licensee's report.
Comment -
A. Contributors to Train 2 are similar to those of Train 1.
- However, l
Train 2 is uravailable in the LOAC event because the steam admis; ion valves that allows steam to the pump turbine AFWP-2 depend on the AC power.
'd I e
a
(
t
ym...._.,
O t
i i
e i
I.
~
t r.
{l 5.0 References 5
1.
"Actiers recuired fron Operating License Applications of Nuclear L
L, Stean Supply Systees designated by Westinghouse and Corbustion h.
4 resulting from the hRC Bulletins ard other task force Engineerica O
review regarding the Three Mile Island Unit 2 Accident," from D. Ross to all pending operating !icense applicatiers, dated L
March 10, 1980.
4:.
Holderness, J.
H., " Davis-Besse Unit No.1 Auxiliary Feedwater F
2.
Sjstem Reliability Analysis Final Report," prepared by EDS Nuclear i
Inc., for Toledo Edison Company, Report No. 02-1040-OR95,
[5 De:e-ter 1981.
3.
Youngolood, R. and Papazoglou " Review of the Davis-Besse Unit a
k-fio.1 Auxiliary Feedwater Systen Reliability Analysis," Ercokhaven 1
haticral Laboratory, flovember 1982.
3 J
4
" Generic Evaluation of Feedwater Transient anc Small Brea. Loss-of-p Coolant Accidents in Westinghouse-Designed Operating Plants,"
{
NUREG-0611, U.S. Nuclear Regulatory Comission, January lE.
i.
- f.,
k' t
2 A
f D,. j P..
e 4
f w
4 q:
eA, 1.-
W.;-
,Ej I
1
)
o f.
a t;
g*
E waymyrwEs.erp5gy.ppMywg.;.gggg3Q47