ML20125D428
| ML20125D428 | |
| Person / Time | |
|---|---|
| Issue date: | 12/10/1992 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1944, NUDOCS 9212150192 | |
| Download: ML20125D428 (227) | |
Text
__
S C 8 8 F [ M R G \'A L O OWICIAL TRANSCRT.PT OF PROCEEDINGS i
MCDCy: U. 3. Nt. clear Regulatory Commission Advisory Connittee on Reactor Safeguards T1dC: 392nd ACRS Steetina Docket No.
O LOCATION: Bethesda, Maryland v
DATE: Thursday, De: ember 10, 1992 PAGES. 2J1 - 361 n@o u v o., -D Q flf C
r T3jl
, g y.-
~ " t" " *
^[0L' Ill 8 L u ; " " +
ANN RILEY & ASSOCIATES, DH O' 1612 K St. N.W, Suite 300 y g b, .. #
Washington, D.C 20006 (202) 293-3950 9212150192 921210 (1 PDR ACRS U T-1944 PDR f
.( .g-3 g OFFICIALTRANSCRIFT OF PROCEEDINGS p,
-e. -
lAg@DCy) 1 U_. .
S / Nuclear [Regulato$,) ,Comission-Advisory Comittee- on Reactor. Safeguards
~
~
4 a tTitle:
f392ndIACRSMeeting' '
Docket No'.;
y i, l: a( N , :
,3 :.~
sij j ;!OCATION: . Bethesda,; Maryland /
~
-+ . . .
DATE - fThursdayJDecember:-- 10, 1992! cpfGES:1201 L-3361L
- s. .
2
- 1 '
Ol[
~ '
, , s +
l^
+ p,n > , . . . .
{b' ;,h k,,f,iU$6tshed,ip0I!ihGlCSI3.dII$$
e- ,
c 4
m I, Q>.
' ~ ~
.; .g., 8 m +
, ;..7 .o.
yg4 .
4 e
..ip4 < .
4W MMggggsQg: ..
MN%") , $:w ',
fg .g Q LJ:
L 1612 K St.N.W, Suite 300 L- ,
q
% < cli ,
p Mahington, D.C. 20006 'i l . . -
j;*, '
Y -
,. ((202) 293-3950
@#,,: M PDR
- . ACR'3 9212150192 ,? @. 921210 *.
, '0U M
iu T-1944 PDR
- gl@4Qw,w w f- 1
~.
<, 4 '
4 r.
~ , . . , .m _--2 f. y__ 1 _.s. 1 , _
s ;1; h.;
1 PUBLIC NOTICE BY THE t UNITED STATES NUCLEAR REGULATORY COMMISSION ADVISORY COMMITTEE ON REACTOR SAFEGUARDS DATE: Thursday. December 10, 1992 The contents of this. transcript of the proceedings of the United States Nuclear Regulatory Commisalon's Advisory Committee on Reactor Safeguards, (date) p . . . g;, , p _ g r in- ino, , as Reported herein, are a record of the discus'sions recorded at the meeting held on the above date.
This-transcript has not been reviewed, corrected or edited, and it may contain inaccuracies.
- O --
- ANN RlLEY & ASSOCIATES,-Ltd..
Court Reporters-1612 K. Street, N.W., Suite 300 Washington, D. C.- 20006 (202)' 293-3950-
.. . <=
I l
201 1 UNITED STATES OF AMERICA 2
3 ***
4 5 NUCLEAR REGULATORY COMMISSION 6
7 ***
8 9 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 10 11 ***
12 13 392nd ACRS MEETING 14 ***
15 16 Nuclear Regulatory Commission 17 Conference Room P-110-18 7920 Norfolk Avenue 19 Bethesda, Maryland 20 21 Thursday, December 10, 1992 22 23 24 25
-Q ANN RILEY & ASSOCIATES, Ltd.
Court Reporters
- 1612 K. Street, . N.W., Suite 300 Wminington, D. C. 20006 (202) 293-3950
, ______________---__a
i i
202 )
i 1 PARTICIPANTS:.
- 2 l
3 P. Shewmon, Chairman of the ACRS l 4 J. Carroll, Vice Chairman of the ACRS 5 C. Michelson, ACRS Member l 6 C. Wylie, ACRS Member 7 H. Lewis, ACRS Member 8 I. Catton, ACRS Member
{
l 9 E. Wilkins, ACRS Member
! 10 T. Kress, ACRS Member l 11 W. Lindblad, ACRS Member 12 P. Davis, ACRS Member l
1?J W. Kerr, Consultant 14 R. Fraley, Executive Director, ACRS 15 C. Rourk, NRC/RES 16 W. Minners, NRC/RES l
l.
17 F. Rosa, NRC/RES 18 A. Igne, Designated Federal Official f
i 19 J. Booker, River Bend Station, Gulf States Utilities
- 20 D. Holbert, River Bend Station, Gulf States Utilities 21 J. Burton, River Bend Station, Gulf States Utilities
- 22 B. Christie, Reliability and Performance Associates
- 23 4
1 24-i 25 ANN RILEY & ASSOCIATES, Ltd.
{- Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 .
a -
, s ., . __ ,, y,- y.
203 1 PROCEEDINGS O 2 (8:30 a.m.)
3 MR. SHEWMON: This is the second day of the 392nd 4 meeting of the Advisory Committee on Reactor Safeguards.
5 At today's meeting, we will discuss or hear 6 reports on the River Bend Nuclear Plant Reliability 7 Monitoring Program, protection against lightning and 8 electrical transients for nuclear power plants, election of 9 ACRS officers, Generic Issue 120, on-line testability of 10 protection systems, Subcommittee and member activities, 11 future ACRS activities, and appointment of ACRS members.
12 Portions of today's meeting will be closed to 13 discuss information, release of which would represent a 14 clearly unwarranted invasion of personal privacy.
15 This meeting is being conducted in accordance with 16 the provisions of the Federal Advisory Committee Act. Al 17 Igne, on my right, is the Designated Federal Employee for 18 the initial portion of the meeting.
19 We h:1ve received no written statements or requests 20 for time to make oral statements from members of the public 21 regarding today's sessions. A transcript is being kept. It 22 is requested that each speaker use one of the microphones 23 and identify him or herself and speak with sufficient 24 clarity and volume so that he or she can be readily heard.
25 Are there any general comments before we begin?
g ANN RILEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
l 204 1 (No response.]
2 MR. SHEWMON: Jay, you have River Bend.
i
- 3 MR. CARROLL
- Yes, I do.- Steve Mays made 4 available to me a paper a couple of months ago on the l
5 reliability program that River Bend has been working on for l 6 the last several years, which utilizes the PRA methods to l
l 7 evaluate the safety significance of things that happen on a 8 day-to-day basis at their plant.
9 I read that paper and discussed it with Steve and 4
10 recommended to the Committee a month or so ago that it would 11 be interesting to hear a presentation.oa this subject.
4 12 So with that, I will turn it over to River Bend.
! 13 MR. BOOKER: I'll give a short h:' :roduction. My jp id 14 name is Jim Booker. I'm Manager of Nuclear Industry i 15 Relations for Gulf States Utilities, River Bend Station, St.
i 16 Francisville, Louisiana.
17 Let me assure you that we did not bring this 18 weather with us. You can't blame it on us.
l
. 19 It's been quite a while since we've been before i
20 this Committee. Dr. Kerr is a familiar face down there. We
. .21 were here in the mid-1970s for our construction permit for 22 River Bend and in the mid-1980s for an operating license, i
23 and we had not-had an opportunity to be back before this 24 Committee since then.
25 So we appreciate the opportunity this morning to
)
O ANN RILEY & ASSOCIATES, Ltd.
lU Court Reporters 1612 K. Street, N.W., Suite 300 1
. Washington, D. C. 20006 (202) 293-3950
205 1 be in here to talk to you about the system reliability 2 program at River Bend, and then, also, the second subject, 3 hcw we're using PRA to enhance our reliability program.
4 We have with this morning Oliver Holbert, 5 Supervisor, Reliability Engineering, who Will make the first 6 presentation, and then Joe Burton, Supervisor of PRA, will 7 make the second presentation. Oliver, I'll turn it over to 8 you. Thank you.
9 (Slide.]
10 MR. HoLBERT: First of all, I would like to thank 11 the Committee for inviting Gulf States Utilities to make 12 this presentation. My name is Oliver HolLert and I'm 13 Supervisor, Reliability Engineering, at Gulf States 14 Utilities' River Bend Station plant.
[
15 I joined Gulf States in March of 1988 and became 16 Supervisor-Reliability Engineering in February of 1991.
17 You'll have to excuse me. I'm a little nervous.
18 [ Slide.]
19 MR. HoLBERT: I will make the first of this two-20 part presentation and talk about our reliability program.
21 Joe Burton will make the second part and explain how we use 22 PRA in the day-to-day operations of the plant and also how 23 we integrate the PRA into our reliability program.
24 I would like to start by giving a short history of 25 the program. The program started in 1988 as a one-system
(]-
ANN RlLEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 1
_ _ _ _ _ _ u
l i
4 206 1 pilot project using the feed system. It started out in our 2 Engineering Analysis Group, and that's an important point.
3 Engineering Analysis Group is where our PRA is being 4 developed.
l 5 We used a small company by the n&me of RAPA, 6 Reliability and Performance Associates, with Bob Christie, 4
7 as our consultant. After the feed system was analyzed for l 8 reliability, availability and we did a cost study, 9 management liked what they caw and they added two more 1 10 systems to the pilot program.
11 Those two systems were the circulating water 12 system and normal service water system. In February of l 13 1990, the Reliability Engineering Group was formed and the 14 project was expanded to 23 systems. I will show you a list 15 of those systems a little bit later.
]
16 Let's talk about the program.
! 17 [ Slide.]
18 MR. HOLBERT: Now, we get to the objectives of the-
- 19 program; enhance maintenance effectiveness, improve' system 20 reliability and availability; show quantitatively how j 21 changes effect system reliability and availability; allocate
. 22 resources in a cost-effective manner, and this is mainly by-l 23 optimizing our maintenance program.
- 24 We do the availability and reliability studies and l 25 then we make PM recommendations. In the PM recommendations, Q ANN RILEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite- 300 Washington, D. C. 20006 (202) 293-3950
i 3
i 207 i
1 we add, delete, or make sure that all of the PMs are the PMs 4
O 2 that we really want to be doing.
3 We implement a living program and this living 4 program is one of the unique features of our program. We l j 5 take the information that we gather in our reliability 3
! 6 analysis and we make curves and charts so that we can trend.
7 As we trend, we do a monthly update on each system and feed f 8 that information back into the program. Therefore, we end l
9 up with the living prograa.
10 We provide plant-specific data for the PRA. This f
f 11 is one of the reasons that we're here today. We have 12 integrated our reliability program with the PRA.
13 [ Slide.)
14 MR. HOLBERT: The principal elements of the i 15 program. The first thing we do is we look at to try to l
16 identify what' critical systems and components we have. We j 17 look at failure modes and then we quantify conditional i
!. 18 probability.
t-i 19 We use conditional probability on our safety l
20 systems. As all of you know, our safety systems are-21 normally standby systems and regular standby availability is 22 also used.
l
' 23 As I explained a little' bit earlier, we.look at l
24 our preventive maintenance program, evaluate the adequacy of f 25 existing'PMs. We look at every PM in that system, whether I
lC.
t~
ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 f-Washington, D. C. 20006 (202) 293-3950 L
, 208 1 it's on a critical component or not. We recommend PM 7-s
\'- 2 changes and we implement those changes and then we have the 3 feedback and monitoring which makes this a living program.
4 [ Slide.] l 5 MR. HOLBERT: As I said earlier, I promised to 6 show you all of those systems that we have. These are the 7 23 systems -- 24 systems now that we have in our program.
8 We started out with 23 and we added the turbine system last 9 year around this time.
10 We had some problems with the turbine at River 11 Bend Station, and we looked at our list of systems and 12 realized we didn't put the turbine on.
13 The first ten systems that you see here, we have f( 14 15 completed analysis on those systems.
to go.
So we have 14 systems Out to the right of the name of each system we show 16 the reason that we selected the system. We did studies with 17 -- we looked at our PRA at the time these systems were 18 picked and quite a few systems on here are on here just 19 because of PRA.
20 We also used O&M cost. Back in the 1989 21 timeframe, our NSAG Group, our Nuclear Safety Assessment 22 Group did a study of the 25 most costly systems at River 23 Bend Station. At the top of the list at the time was the 24 feedwater system and that's why that system was chosen as 25 the first system for our pilot program.
O
'# ANN RILEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
i i I
209 I. 1 MR. CARROLL: Costly in the context of 2 unavailability?
i
! 3 MR. HOLBERT: No. Cost in maintenance, the money 4 that we had to spend, the man hours and materials and 5 everything to keep up those systems.
i 6 MR. MICHELSON: When you indicate PRA as the t
- 7 basis, I guess that's because it turns out to be somewhat i
- 8 important to safety as opposed to O&M, where it's purely 4
1 9 availability of the plant itself.
{
10 MR. HOLBERT: That's correct, sir.
1
,! 11 MR. MICHELSON: I'm a little puzzled not to see i
12 the, for instance, containment isolation valves anywhere in 13 these systems. The reason is I look at, for instance, l
"T 14 reactor water cleanup, which is an o&M system only.
!(s_/
15 Clearly, the isolation valves between there and the reactor l 16 are extremely important to safety.
17 How do they get picked up, then?-
18 MR. BURTON: Skip, let me take that one. I'm Joe
20 _that we additionally identified came out of our Level 1 PRA
{
21 model. At the time that we did the initial identification,
! 22 we had not done our containment analysis.
l 23 MR. MICHELSON: Had you done external events?
j 24 MR. BURTON: No. We're in the process of doing j 25 external events now, as well.
i 1
ANN RILEY & ASSOCIATES, Ltd.
!- {s} Court Reporters l
i 1612 K. Street, N.W., Suite 300 !
Washington, D. C. 20006 j (202) 293-3950 1
- -1 L !
210 f- 1 MR. MICHELSON: Until you do external events, you
\~l 2 probably won't recognize the importance of the isolation 3 valve.
4 MR. BURTON: Correct. We're also doing a Level 3 I
5 PRA, which that will become critical. So we expect our list 6 of important systems to grow with time.
7 MR. MICHELSON: But is containment isolation a 8 system in this context or is it going to have the valves -
9 -
10 MR. BURTON: The valves will be associated with 11 individual systems.
12 MR. MICHELSON: So eventually reactor water 13 cleanup might pick up some PRA.
~T 14 MR. BURTON: Yes.
(d 15 MR. HOLBERT: Later on when Joe comes up, he's 16 going to show you the 23 most important systems ranked by 17 PRA. Along with the new PRA numbers and providing for the 18 maintenance rule, we will change this list of systems. It 19 will grow considerably.
20 As I was going to say, this list was made up back 21 in 1989, with the exception of turbine system. As you can 22 see, we've got a good mix of safety systems and balance-of-23 plant systems.
24 [ Slide.]
25 MR. HOLBERT: In our program we produce a O
'/
ANN RILEY & ASSOCIATES, Ltd.
COud RepOners 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
211 l
- 1 reliability / availability analysis and I've got a copy of
\' 2 Volume I of our analysis. As you can see, there ate quite a 3 few pages here. We do availability trending, preventive 4 maintenance recommendations and plant-specific data for our 5 PRA.
6 We've done quite a bit on trying to identify what 7 products come out of every department at River Bend Station.
8 (Slide.]
9 MR. HOLBERT: Let me show you one of the curves 10 that we produce. This is the standby availability for the 11 reactor core and isolation cooling system. As you can see, 12 we went all the way back to 1988 and gathered information so 13 that we came up with monthly -- and the histogram lines are
(~} 14 the monthly standby availability for RCIC.
\m/
15 The curve that you see here is a 12-month sliding 16 average of those monthly figures.
17 MR. LEWIS: Is that the past 12 months or is it a 18 window surrounding the month that it shows?
19 MR. HOLBERT: Each dot is an average of the past 20 12 months.
21 MR. LEWIS: The past 12 months. In that case, I 22 don't understand it because if you look at the thing that's, 23 I guess, in November 1989, it shows slightly over 90 percent 24 and within the past 12 months there are two months that are 25 off the chart and below 70 percent.
/i ANN RILEY & ASSOCIATES, Ltd.
Y Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l l
(202) 293-3950 l
I
.m
212 1 So the lagging 12-month average can't be over 90 0 2 percent. I'm confused, because everything else isn't up at 3 100 percent, anyway.
4 MR. HOLBERT: This area right in here and here and 5 here are the times that we were in outages.
6 MR. LEWIS: Are they counted in the 12-month 7 average?
8 MR. HOLBERT: No, they're not, sir.
9 MR. LEWIS: They're not counted.
10 MR. HOLBERT: No.
11 MR. LEWIS: I see.
12 MR. WILKINS: Is your 12 months really 15 months, 13 then? Calendar months, 12 calendar months?
14 MR. HOLBERT: It would be 12 whole months.
15 MR. WILKINS: Yes. During which time you were 16 only operating nine months or ten months or something.
17 MR. SHEWMON: No. It's 12 months in each one of 18 those and if you come across times when you're not 19 operating, you don't count those.
20 MR. LEWIS: You simply don't count them. So you 21 go another two months back.
22 MR. HOLBERT: Right.
23 MR. LEWIS: To get 12 months.
24 MR. WILKINS: I thought I got a negative answer to 25 that question. That was the first one I asked. So, in Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
. _ _ _ - _ _ _ _ ~ . - _ _ _ _ _ _ _ _ _ _ . _ _ . _ . _.
l l 213 j 1 fact, you're looking at calendar period, which is 2 considerably more than 12 months occasion.
3 MR. BOOKER: Skip, it is 12 months of operation, ,
- i
! 4 isn't it? Excluding downtime.
1 l 5 MR. LEWIS: How do you handle months that-are 6 broken in half? Do you just add up days until you add up to
- - 7 12 months?
l 8 MR. HOLBERT: Each one of these-is the amount of 9 time that you could actually. operate for the month to make 1 10 up your standby availability.
11 MR. LEWIS: I guess I don't understand.
12 MR. HOLBERT: If you had a 30-day month that you l
j 13 operated 29 days, 29 times 24 would be what you could 14 operate. That would be for that month.
I 15 MR. LEWIS: So what you do is you simply go back 16 over the operating time adding up days until you get to a l 17 year, at:d during those days you ask for the availability of j 18 the system.
19 MR. HOLBERT: Yes, sir. Our standby availability I
- 20 is the actual operating time with-the downtime subtracted
- 21 from it. Your downtime would be when you were in 22 maintenance, that the system was not available or if it was
. 23 actually broken, had a failure.
24 MR. LEWIS: I understand that. Then, to-find the 25 12 months,-you simply go back in time as far as you need_to Q ANN. RlLEY & ASSOCIATES, Ltd.
Court . Reporters 1612 K. Street, N.W., Suite _300 Washington, D. C. 20006 (202) 293-3950 -
I J
i 214 1 to add up 12 months of operating time.
j0
. 2 MR. HOLBERT: Yes, sir. j j-j 3 MR. LEWIS: However long it takes.
! 4' MR. HOLBERT: Yes, sir.
5 MR. LEWIS: It could take ten years. It doesn't, f
j 6 but it could take ten years. I'm just trying to understand 7 what you.did.
! 8 MR. HOLBERT: Yes, sir. ;
i
,t 9 MR. BOOKER: Skip, we need to revisit this. I I
j 10 think it's my understanding that we use 12 calendar months.-
1 11 If you're down two months out of that, you only have ten l
l 12 months of data.
1 13 MR. WILKINS: . It's clear that'you gentlemen don't
'14 know the answers to our questions. So we-better stop asking j 15 them. We have our consultant here with us.
16 MR. LEWIS: I think you guys should resolve this i 117 among yourselves on your time.
b 18 MR. CARROLL: In any case, I guess I.'m a little.
[
19 shocked to see the standby availability of RCIC as low as it l- 20 is. I would have expected it be in the high 90siinstead-of 21 in the low 90s.
22 - MR. CATTON: It-looks-like 97.
[- 23 MR.~ HOLBERT: Yes. That's'97 percent up here now, 24 -sir.
[ 25 MR. DAVIS:- That's not' considered-a-. safety system l'
i~
ANN RILEY & ASSOCIATES, Ltd.
L Court Reporters -
L 1612 K. Street, : N.W., . Suite 300 L' Washington, D. C. 20006 .
3 (202) 293-3950
215
,-s 1 for your plant, is it?
'-/ 2 MR. HOLBERT: Yes, it is.
3 MR. DAVIS: It is. Part of the ECCS?
4 MR. HOLBERT: Yes, it is.
5 MR. BURTON: No. No , sir. It's not part of ECCS.
6 It is a safety-related system, yes, sir, but it's not ECCS 7 at River Bend.
8 MR. MICHELSON: It is a Mark VI.
9 MR. BOOKER: It's a BWR-6 Mark III.
10 MR. MICHELSON: Thank you.
11 MR. LEWIS: Incidentally, whatever the answer was, 12 and I'm not going to press it, on the annual, does that also 13 apply to the monthly averages, you go back enough days to r'* 14 add up to a month?
N.]1 15 MR. HOLBERT: No.
16 MR. LEWIS: Or you divide by th* number of days.
17 Whichever you end up with on that one, does it also apply to 18 the monthly averages?
19 MR. HOLBERT: Calendar month, sir.
20 MR. LEWIS: Those are calendar months. So the 21 monthly averages are done differently from the annual.
22 MR. WILKINS: Maybe. You don't know the answer.
23 MR. LEWIS: Maybe. I'm only trying to find out 24 whatever they d.d, which they will resolve on their own 25 time, did they do the same for months as they did for years.
(]'
\
ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
216 1- MR. WILKINS: And you won't be able to answer 2 that.
3 MR. MICHELSON: What is the allowable outage time 4 on RCIC? Limiting condition for operation, how many days on 5 RCIC?
6 MR. HOLBERT: I'm sorry. I don't know the answer 7 to that question. The LCO for --
8 MR. BURTON: On RCIC, I believe it's 14 days, as 9 long as the high pressure core spray system is available.
10 MR. MICHELSON: I wondered because some of these 111 rather low availabilities might have you put you out of 12 commission in some of those months.
13 [ Slide.)
14 MR. HOLBERT: Here is the aval) ability, 12-month 15 sliding average availability for the chillers. We have four 16 100 percent chillers at River Bend Station and we do a 17 combined availability for all four and we have individual 18 availabilities.
9 As you can see, our chillers 1A-and 1C are quite a ,
20 ' bit better availability than 1B and 1D. We've.had quite a -
21 bit of problems with our control circuitry on_these 22 chillers. It seemed before we-did our 2nalysis, there was 21 no program to equalize ours in these chillers. lA-ran quite a bit more than the rest'of them.
'24 25 In fact,- 1A has .got o'ler 2,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> higher
-}' ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suit 0 300 Washington, D. C. 20006 (202) 293-3950
.- . _ . _ =. -- . _ _ . _ _ . _ -
l l
217 1 opcrating time than the others, any of the others. Our 73
2 control circuitry is set up so that we can only run one 3 chiller at a time.
4 You automatically switch to one of the other 5 chillers by --
6 MR. MICHELSON: You have four chillers for the 7 control room?
8 MR. HOLBERT: Four for the control building.
9 MR. MICHELSON: You run one at a time.
10 MR. HOLBERT: Yes, sir. In RF-2, we changed from 11 four 50 percent chillers to four 100 percent chillers. They 12 did studies with the manufacturer and changed the control 13 circuitry around. So now we have more redundancy and better 14 reliability.
15 MR. MICHELSON: What do you use -- this is juut 16 for the control building.
17 MR. HOLBERT: Yes, sir.
18 MR. MICHELSON: What do you do for the reactor 19 building?
20 MR. HOLBERT: There's HVC for the reactor 21 building.
22 MR. MICHELSON: Do you consider it safety-23 related? -
24 MR. HOLBERT: Yes, it is.
25 MR. MICHELSON: And how is the experience there
(]' ANN RlLEY & ASSOCIATES, Ltd.
Coud RepOners 1612 K. Street, N.W., Suite 300 WashingtO'ii, D. C. 20006 (202) 293-3950
218 1 compared with the control building?
2 MR. HOLBERT: We haven't done that system yet, but 3 it's quite a bit better. I can tell you that just from 4 experience.
5 MR. DAVIS: You did a comparison of the turbine 6 building with the control building chillers.
7 MR. HOLBERT: Yes, we did.
8 MR. DAVIS: And the non-safety-related turbine 9 building had a higher unavailability than the safety-related 10 control building.
11 MR. HOLBERT: Yes, sir. That's true. I'll put up 12 the turbine building and show you.
13 [ Slide.)
() 14 15 MR. HOLBERT: The turbine building chillers were quite a bit higher in availability than the safety-related 16 HVK chillers. One of the reasons that we said that the 17 availability is better in the non-safety-related, these are 18 two-out-of-three. We have three of them. It takes two out 19 of three to make 100 percent.
20 These normally run for a-long period. There are 21 not a lot of STPs that have to be done or surveillance test 22 procedures that get run on them. The maintenance people can 23 take these down at their convenience. There is not a lot of 24 switching that goes on. The switching in the HVK is one of 25 the problems that we have found that causes unavailability Q- ANN RILEY & ASSOCIATES, Ltd.
Coud RepOders 1612 K. Street, N.W., -Suite 300 Washington, D. C. 20006 (202) 293-3950
219 1 on those.
'J 2 We have a lot of trouble with our control 3 circuitry in the safety-related HVK chillers. In the much 4 simpler maintain turbine building chillers, we found that 5 they just run and run, don't have very many problems r.t all.
6 MR. MICHELSON: That's the only way chillers like 7 to run is all the time and at fairly constant load.
8 MR. HOLBERT: That's true.
9 MR. MICHELSON: You start fluctuating a load and 10 turning them on or off, then you've got a problem.
11 MR. HOLBERT: That's true, sir. Are there any 12 other questions?
13 MR. LEWIS: Forgive me. I'm just having a lot of
( 14 trouble understanding. If we could look at the previous 15 chart for just one moment, there's one that will resolve my 16 problems if I can get to understand it.
17 If you look at the lower righthand thing on it, 18 the chiller ID 12-month average, then somewhere around the 19 end of 1989, there's this tremendous jump from 96 percent to 20 43 percent, that big first thing, the one to the left where 21 it first goes down.
22 Now, that's preceded by 12 months or running 23 nearly 100 percent as the 12-month lagging average. And in 24 one month, the 12-month lagging average goes down to 40 25 percent. Now, that's not possible.
(7
'l ANN RlLEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300
\Nashington, D. C. 20006 (202) 293-3950 1
__- ._ . .- = _. - . _ - . . .- - .- .-
220 1 MR. WILKINS: A hundred times eleven plus zero 2 times one divided by 12, by God,-is --
3 MR. LEWIS: Is not 40 percent. It is simply 4 impossible.
5 MR. MICHELSON: Would you define what you mean by-6 availability?
7 MR. CHRISTIE: My name is Bob Christie with 8 Reliability and Performance Associates in Knoxville, 9 Tennessee. The situation, as it works, the availability is 10 defined as the uptime over the total time that you want it 11 to be available.
12 In this particular instance, you have four 13
~
chillers. Chiller D does not get as many hours of run-time.
14 As a matter of fact, sometimes it might get one or two hours 15 of run time. Now you have a month where, for instance, you 16 put it in at the beginning of the month and it stays down ,
17 for the whole 720 hours0.00833 days <br />0.2 hours <br />0.00119 weeks <br />2.7396e-4 months <br /> in the month failed. You can have 18 it very low as-far-as the 12-month availability, because the 19 availability is the uptime over the' total time.
20 So we're only counting the uptime on the hours 21 it's run. Since you have four chillers, the most hours
-22 you're going to run every month-is 720, or 744 if you've got 23 a 31-day month, or it could be down.
24 So, again, we're using the true definition of 25 availability and calculating-it on that basis. To go back ANN - RILEY'& ASSOCIATES, Ltd.
- }'-
-Coud Reponers 1612. K. Street, N.W.,' Suite - 300 Washington, D. C. 20006 (202) 293-3950 l
221 1 to your other question of how we were doing it on RCIC, we 2 are using calendar time, in addition to which we're only 3 using the time in Modes 1 and 2.
4 So if you have a month, for instance, where you l 5 might have had a trip and you shut down for half the month 6 for a trip, then the uptime during that month could be a 7 maximum of half the month.
8 So it is calendar time, it is 12 months. You will 9 have operating hours or standby availability hours that are 10 different over a 12-month calendar time, especially if you 11 have the refueling. Instead of having 8,000 or more hours, 12 you might only have 6,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> in that 12 calendar months.
- 13. We always take the last 12 calendar months.
14 MR. LEWIS: Okay. So that's the answer to that 15 one. But let's come back to the chiller. First of all, 16 we're talking about the curve for chiller 1B.
17 MR. CHRISTIE: That is correct.
18 MR. LEWIS: So what you mentioned about the other
, 19 three chillers is irrelevant, right?.
20 MR. CHRISTIE: The others may be running. For 21 instance, if you look at it, up until the' time in November 22 of 1989, the other chillers were probably. carrying most of 23 the load. Chiller D, for the whole' year of 1989, may have 1
24 had, at most, let's say, 500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operating time..
25 If it failed at the beginning of November, it Q ANN RlLEY & - ASSOCIATES, Ltd.
= Court Reporters 1612 K. Street, N.W., Suite 300 Washington,. D. C. 20006 (202) 293-3950 -
, l 4
a 222 !
j 1 would be down the whole 720 hours0.00833 days <br />0.2 hours <br />0.00119 weeks <br />2.7396e-4 months <br />. - So you would have 500
- 2 operating hours for that previous 11 months, and then you i
! 3 would go down for the whole month.
I
- - 4 MR. LEWIS
- So what you're telling me is that its l 5 12 months calculated experience actually'all happened in 6 that month.
t-l 7 MR. CHRISTIE: It could have worked that way.
4 8 MR. WILKINS: . More than half of it, anyway, on his i
- 9 example.
! 10 MR. CHRISTIE: That is exactly right. So you have i
i 11 -to look at the entire report in order to get the operating i
l- 12 hours, and we have the reports available. Steve Mays has 13 the report on the HVAC to see the monthly operating. hours 14 and the year-to-year hours and that, and you can get a
+
15 flavor.
16 But that is the reason why you can.get to less i
17 than 50 percent, for instance.
18 MR. LEWIS: I'm not worried about less than 50
! 19 percent. I'm worried about going down from 100 percent to i
!- 20- 40 percent in one month. What you're telling me is that I 21 should simply ignore the ordinates because they have nothing 4
22 to do with months -- I'm sorry -- the abscissas, they have J
- 23 nothing to do with months..
24 MR.-WILKINS: They have everything to'do with 4
25~ calendar months and very'little to do with hours of
' Q ANN. RlLEY & ASSOCIATES, Ltd.
Coud - Repoders -
1 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006
, (202) 293-3950 r - ,r, -, ,--
4 223 1 operating time. ,
1
- O 2 MR. LEWIS: Exactly.
3 MR. CATTON: Which is what determines the point.
4 MR. HOLBERT: I've got numbers right here. In
- 5 1989, chiller 1D.
6 MR. WILKINS: I'm looking at Page 46 and he
]
7 defines these 12-month averages on our Page 46.
8 MR. LEWIS: Let's find.out if I understood what 9 you said. You, in effect, said that in that particular 1
i 10 instance, the year's experience was more or less l 11 concentrated in that month. That's, in effect, what you 12 said.
j 13 MR. CHRISTIE: That is correct.
14 MR. LEWIS: Fine. Then I understand it.
- 15 MR. WILKINS: In fact, that answers the original i
16 question you asked.
l 17 MR. LEWIS: Yes.
! 18 MR. WILKINS: About whether'12 months is 12 months 19 or 14 months or 15. It's 12 months.
20 MR. LEWIS: That's correct.
21 MR. CHRISTIE: That's what I said, f
22 MR. LEWIS: At-'the very least, if we assume I have 23 every intelligence, these are slightly misleading.
. 24- MR. MICHELSON: In looking at-your operating 25 experience, if chiller ID is on standby for_a few months 4
' C)U ANN -RILEY & ASSOCIATES, Ltd. )
- Court Reporters !
4 1812 K. Street, N.W., Suite 300 l
!- -Washington, D. C. 20006
- (202) 293-3950 l 2
224 j gs 1 time and then you go to crank it up and you find it fails, 2 do you count that as being still available those other 3 months or was that unavailable throughout the time?
4 MR. CHRISTIE: Again, the chiller system is a 5 running system.
6 MR. MICHELSON: No, ID is not necessarily running.
7 MR. CHRISTIE: The chilling system is a running 8 system. The only up hours are counted as the hours where it 9 actually ran. We do have times when we have chillers in 10 tag-out that are having maintenance on them that are not 11 counted against them.
12 MR. MICHELSON: You mean you're running all four 13 chillers at all times.
(i 14 MR. HOLBERT: No.
Gl 15 MR. CHRISTIE: Again, the only up hours that are 16 counted are the hours of the actual chiller running. The 17 downtime is only for the failures that have actually 1
18 occurred where you either had a chiller running and it 19 failed, it failed to run, or you had a chiller where you 20 wanted to switch and l'ou tried to switch to another chiller 21 and it failed to start.
22 MR. MICHELSON: Then you're defining availability 23 as available when needed, not, as opposed, to if needed.
24 MR. CHRISTIE: That's exactly right. That's 25 exactly correct.
(~')
ANN RILEY & ASSOCIATES, Ltd.
'- COud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
225 1 MR. HOLBERT: Let me put those numbers in 2 perspective a little bit. In 1989, on chiller 1D, we had 3 615 run hours, 775 fail hours. For the other three 4 chillers, they were all above 1,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of run time and 5 the most fail time for any one of the other ones was five 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />. So this one was quite a bit worse.
7 MR. SHEWMON: Thank you. Let's go on.
8 MR. MICHELSON: One other small question. Are you 9 all your chillers the same manufacturer?
10 MR. HOLBERT: No, they're not.
j 11 MR. MICHELSON: They're a variety.
12 MR. HOLBERT: We have Trane in the turbine l
13 building.
- , 14 MR. MICHELSON
- Let's say in the control building, 15 1A, B, C and D are all the same manufacturer.
16 MR. HOLBERT: Yes, they are.
17 [ Slide.]
i j 18 MR. HOLBERT: This is how we get our data for our 19 reports. Control room logs supply probably the biggest 20 portion of the information. We also evaluate electrical and j
21 mechanical PMs. We look at all clearances, condition l 22 reports, maintenance work orders for each one of the 23 systems.
24 We look at the data in the Nuclear Plant 25 Reliability Data System, NPRDS, as most of us are familiar l
- Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters
- 1812 K. Street, N.W., Suite 300 l Washington, D.. C. 20006 (202) 293-3950-4 - -
226 1 with, and review all of the failure reports there. That's 2 if the system is covered by NPRDS.
3 In the Emergency Response Information computer, 4 the ERIS computer, we have 100 parameters that we look at on 5 a daily basis. This information is not really included in 6 our reports, but it's gathered for the PRA so that we can 7 have running time on all of our critical components that are i- identified by the PRA.
9 There are approximately 147 components right now 10 that PRA wants monitored by us. Additionally, to get the 11 information that we can't-get from the Emergency Responce 12 System, we have our shift technical advisors to gather daily 13 run times for us on those other systems that we don't cover 14 there.
15 (Slide.)
16 MR. IiOLBERT: This is normalized core damage 17 frequency. The information that we gather on three systems, 18 the RCIC or reactor core isolation cooling, high pressure 19 core spray and the diesel generator unavailability time, is 20 fed back into the PRA so that we can generate this 21 normalized core damage frequency curve.
22 I would ask you to hold yuar questions on this 23 slide until Joe comes up. Joe is going to explain this one 24 quite a bit better.
25 MR. CATTON: Just the normalization factor, what ANN RILEY & ASSOCIATES, Ltd.
O- .Coun Repoders
.1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 1
._ U
i i
l 227 i
1 makes that a one?
2 MR. BURTON: Skip, if you would. Joe Burton, 3 again. The way we normalize our core damage frequency, we 4 come up with a baseline core damage frequency using our j 5 Level 1 PRA model.
6 In there, we input for RCIC, HPCS and the three 9
, 7 emergency diesel generators, the industry goal for 1
j 8 availability on those systems, and then we compute a core i
9 damage frequency. That's our baseline of one.
1 10 Then on a monthly basis we input the maintenance
$ 11 unavailability into the model, the failure to start, and the 12 failure to run numbers which we get from Reliability Systems 13 Group and compute a new core damage frequency, and then we ,
1-14 normalize it against that baseline.
15 MR. CATTON: You're giving me a lot more answer i :
i 16 than I asked for. I just want to 1:now what one unnormalized
- 17 is equal to.
4 18 MR. HOLBERT: It's 1.55-times-ten-to-the-minus-1
- 19 fifth.
20 MR, WILKINS: So if I understood your explanation, 21 Mr. Burton, that means that you could conceivably have a 22 number on this scale that's more than one.
23 MR. BURTON: Yes, sir, you could._ '
24 MR. WILKINS: You don't.
25 MR. BURTON: It's a function of how the plant has Q ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C,. 20006 (202) 293-3950
i a 228 1 actually performed. This just gives us a measure of whether O 2 or not our safety systems are performing the way they 3 should.
4 MR. DAVIS: Does this suggest that the gesteric 5 data may be optimistic or that you're not as good as the 6 generic data?
7 MR. BURTON: No. It's just the opposite. We're 8 better than the generic data.
i 9 MR. DAVIS: You're better.
10 MR. BURTON: Yes. If you go above one, your l 11 worse.
1
] 12 MR. DAVIS: You guys are too quick for me.
13 MR. LEWIS: But that does mean that in the generic
.a 14 data these three systems, which are the only things you've i 15 changed, accounted for more than ten percent of the computer 16 -- I'm supposed to hold my questions on this.
17 MR. BURTON: That will become clearer when I do my i ,
{ 18 presentation.
i 19 MR. WILKINS: If we ever let you get to it.
~
20 MR. LEWIS: There are a lot of questions on this 21 one, but I'm holding them.
f 22 MR. HOLBERT: And I knew this would happen.
23 (Slide.)
i 24 MR. HOLDERT: This is considered by me the second 25 -portion of my system reliability program. The first portion--
i
,O 1
ANN RlLEY & ' ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W.,' Suite 300
- Washington, D. C. 20006 (202) 293-3950
1 i
i 229 1
I 1 is the unavailability and availability that we went through. :
, 2 The second portion is after we make that history study, we i
j 3 go back and make recommendations on the preventive f 4 maintenance of the system.
j 5 PMs are either added, deleted, modified and, by 6 that, I mean extend the frequency or cut the frequency in
, 7 accordance with what we have studied. We use the EPRI RCM i
l 8 workstation right now to do our RCM analysis. We have 1
i 9 looked at probably four different companies that have come 10 in and tried to sell us workstations.
i 11 Because we're with EPRI, we get the workstation j 12 for free and that's why we decided on this one.
I 13 Surveillance test procedures are examined and-IO 14 recommendations are generated, but surveillance test iO 15 l procedure recommendations are not put in immediately because j
16 they have to go through the change process to change the 17 tech spec. They are all being held and when we do a tech i
18 spec upgrade, those recommendations will be considered.
j 19 Equipment qualification PMs --
20 MR. CARROLL: Has quite a list evolved from j 21 looking at things from a risk-based perspective in terms of 1
- 22 changes to the surveillance testing?
4 23 MR. HoLBERT: Yes. When you say quite a list, i 24' though, there are probably not over 35 or 40 in the four 25 systems that we've looked at.
Q ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 -
Washington, D. C. 20006 (202) 293-3950
.h
- - . .n.. - , - . . -- -
-.c . .,, - .. .- . - - . - - -e
k l
230 i 1 The equipment qualification PMs are also hard to j O 2 change, but we have been successful in changing some of 3 these. We have also found some overlaps in the equipment i
4 qualification program and the PM program and we've deleted 5 the overlaps. So our program is showing some profitability 3
i 6 there.
I l 7 MR. CARROLL: Give me an example of an equipment
! 8 qualification PM.
I 9 MR. HOLBERT: EQ PMs are, say, an MOV that's in a 10 critical system that operates quite a bit. It would be
- 11 replaced in ten years, period, without any considerations j 12 for anything else. That's not one that we have tried to i
13 change, but we are working on changing the frequency of the
, 14 recirc motors right now.
- 15 They have a tear-down at ten-year periods and we i 16 feel that condition monitoring should be used where we look 17 at vibration, oil analysis, thermography, and if those
- 18 things are good at nine years, why not extend that
} 19 frequency.
20 GE is doing a study on this right now, also, and 21 we are in a part of that study. '
22 (Slide.)
23 MR. HOLBERT: A lot of the present RCM programs 1
24 have had a lot of problems with implementation. We have 25 been.very fortunate. Changes on eight of our ten systems l
ANN RlLEY & ASSOCIATES, Ltd.
~O-- Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950
231 1 have been implemented. We've made 1,221 recommendations so
@ h far and 80 percent of these have been implemented.
3 MR. DAVIS: Does the NRC get involved in approving i the >?
MR. HOLBERT: No , sir.
~
o MR. DAVIS: They don't.
7 MR. HOLBERT: They don't have to, unless we try to 8 change something that's STP related. We have run into some 9 problems with the diesel because of the Owners' Group matrix 10 and we have recommendations that are being studied right 11 now. That's part of that 20 percent that hasn't been done.
12 MR. DAVIS: Thank you.
13 [ Slide.]
'T 14 MR. HOLBERT: Here is just to show you on our
('d
\
15 residual heat removal system, which is a safety system, wo 16 had no change to 109 PMs. When I'm saying PMs, I'm not 17 talking about STPs. We changed the scope of frequency of 18 six of those. We deleted 28 and we added three PMs.
19 MR. CARROLL: At River Bend, I take it that your 20 group does what, in a lot of plants, system engineers do, is 21 that right?
22 MR. HOLBERT: No, sir. We work very closely with 23 the system engineers. We have system engineers that are 24 assigned to the majority of our systems and as we are going 25 thrnugh the system, the system engineer is kept informed at
(~') ANN RILEY & ASSOCIATES, Ltd.
'/ Coud Rapoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
232 1 all times.
O 2 He helps us with the PM recommendations. He hel's 3 us with the availability and everything.
4 (Slide.)
5 MR. HOLBERT: Benefits of the program. We provide 6 PRA with plant-specific data to replace the generic data in 7 the PRA. Our program gives the system engineer a good 8 indication of the current condition of his or her system and 9 shows if improving or degrading trends are happening.
10 Maintenance supervisors and workers are consulted 11 for their opinion of the effectiveness of the current PM 12 program and their comments are incorporated.
13 conditioning monitoring rather than time-based PMs 14 are recommended, when possible. We have noticed that when 15 the Reliability Engineering starts to do an analysis of the 16 system, a lot of detail, a lot of attention gets paid to the 17 system and the availability normally goes up.
18 We consult with the shift supervisors, the 19 operators and technicians, and the system engineers and the 20 operations people analyze each one of the STPs to make sure 21 that they have cut out as much of the fat of keeping a 22 system down as they can.
23 By that, I.mean if you tag a system out real 1 24 early, but the system is still available, that time is not 25 counted. If we have to pull a breaker out, then the system l l
ANN RILEY & - ASSOCIATES, Ltd.
f] Court Reporters 1612 K. Street, N.W., Suite 300 Washington,- D. . C. 20006
)
(202) 293-3950 1
. 233
, - 1 is unavailable at that point. Nothing easy can be done to
\/ 2 get it back on.
3 Cost. To date, we've expended about 1,600 man 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> per system. System analysis takes us about 13 weeks 5 for a system like the high pressure core spray, up to 32 6 weeks for a system such as residual heat removal, which is J
7 very complicated. Average completion time of analysis is 16 8 weeks at this point.
9 This includes typing and all of the meetings. We 10 normally have two meetings where we have all of the 11 important players for the system in; the system engineer, 12 the design engineer, all of my people in Reliability 4 13 Engineering, the maintenance people that are going to do the
s 14 PM recommendations, and the Implementation Team, which is a (d
15 different group than my group.
16 The monitoring and trending portion of the program
, 17 can be done without completing the PM recommendation part.
18 PM recommendations are not a part of the new maintenance 19 rule, so that portion of our progran can be split.
20 We've conducted cost studies on the first three 21 systems as a part of the pilot program and we found that 22 payback was very small when you look for money saved by 23 deleting man hours. There are quite a few other studies out 24 there that show the same thing.
25 The big payback of a reliability program like ours O
ANN RILEY & ASSOCIATES, Ltd.
Coud Repoders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950
i l
1 234
. 1 is when you -- let's go back. The payback for the program, y
2 we found originally that payback was about three to five 3 years when you look at deleting PMs and like that. But we 4 .found the big payback was increased availability, again, j 5 when we analyze the system, and we calculate that we save 6 approximately $250,000 for each one percent of increase in i
7 availability, and that's in replacement power cost.
4 8 The average increase in availability we found so 9 far in our program, from the time we start up until we 10 finish our analysis, is about two percent.
11 We determined that the optimum staff for this j 12 effort is six persons, five analysts and a dedicated clerk.
13 MR. CATTON: Was that 250,000 per month, per year?
- 14 MR. HOLBERT: Per year. The STA -- we feel that 15 the people in our group, in the group, should have the 16 following backgrounds. We should have a person with STA or 17 SRO experience, an I&C tech, a mechanical engineer, a l 18 maintenance person that can be either electrical or 19 mechanical, and a reliability engineer or a person with the 20 statistical background.
21 We've determined computer programs best for our 22 use and we're currently using the EPRI RCM workstation, as I j 23 said. We are PC-based right now and we use a program to do 24 our charts that was developed by a young man from the 4
25 University of Oklahoma, a coop student that I had in my i
Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293 3950
235 1 program about four semesters ago.
O2 He sat down with Bob Christie and computerized 3 almost everything he did, right down to putting the formulas 4 in Lotus 1, 2, 3 and using Harvard Graphics to do our 5 curves.
6 MR. WILKINS: Oliver, we need to finish up so we 7 can get on with the rest of our program.
8 MR. SHEWMON: Yes. We're running a little behind.
9 MR. HOLBERT: Okay. Future uses of the program.
10 Maintenance rule. We're currently awaiting the results of 11 the V&V program so we can incorporate the results of our 12 findings into our program. We are open to methods that will 13 improve our system.
14 We feel we are on the right track. More work will 15 be needed to comply with the rule. We must set goals using 16 PRA, must identify SSCs under the rule and identify risk-17 significant systems that are under Paragraph Al of the rule, 18 those SSCs that must be monitored.
19 We must select SSCs that are relied upon to 20 mitigate accidents and transients in the EOP items. We feel 21 our system reliability program and our monitoring and 22 -trending process is a good tool that will effectively amount 23 to effectiveness of our maintenance program at River Bend.
24 Close association with the PRA Group is to our-25 advantage in complying with the rule and the use of the core t ]'- ANN RILEY & ASSOC!ATES, Ltd.
Court Reporters
. 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
I 236
! 1 damage frequency curve will help to monitor the performance 1
2 of the program.
3 Are there any questions?
. 4 MR. MICHELSON: I've got one. We got a handout 5 and in there are tables of the failures or the chillers over 6 a period of a couple, three years. Is that a complete
. 7 table for all chiller failures on that?
8 MR. HOLBERT: No. I know what you're talking 9 about. That is not a complete table.
j 10 MR. MICHELSON: Just some typicals.
11 MR. HOLBERT: Yes, sir.
12 MR. MICHELSON: Only one other question, then.
13 When you had a grid disturbance on the 9th of November, all 14 the other chillers functioned properly except for D. Is 15 that correct? Were there any problems with the chillers as 1
- 16 a result of the grid disturbance?
. 17 MR. CHRISTIE: Carl, I'm not exactly familiar with
. 18 that date, but I will go back and look, if we have to.
19 MR. MICHELSON: I just wondered. That's why I l
j 20 asked was this complete, because no others showed up.
)
i 21 MR. CHRISTIE: As far as I know, that indicates 22 all on that date that happened, yes.
23 MR. MICHELSON: Then you had no problem with the-24 other chillers.
25 MR. CHRISTIE: That is correct.
j ANN RILEY & ASSOCIATES, Ltd.
' O' '- COud RepOders 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950
. _ _ _ _. ._ __ _ _ _ __ _ ____ _ ._ _ _ _ _ _ . _. - - - -.-m-
, 237
' 1 MR. DAVIS: Is there time for one quick question?
2 Mr. Holbert, I seem to get the impression from your 3 presentation that the reason that the turbine building 4 chillers were more reliable is because they operate all the i 5 time.
6 But in a paper that we were given presented by Mr.
7 Christie at the ANS meeting in Boston, it says that the 8 reason for the excessive failure rate for the safety-related 9 chillers are excessive maintenance requirements and the 10 requirement to meet the single failure rule.
11 This caught my attention because it seems to imply 12 that NRC regulations are making things less reliable. Is 13 there a way you can straighten this out for me?
14 MR. HoLBERT: I also mentioned when I went through 15 that that the control circuit:Ly and switching were one of 16 the major causes for that unavailability difference.
17 MR. DAVIS: And this paper says that's due to the 18 single failure rule.
19 MR. HOLBERT: Yea sir. Bob, explain.
20 MR. CHRISTIE: Yes. The situation that you have 21 for the control building chillers is safety-related.- To 22 meet all the design basis accidents,-you end up with four 23 redundant chillers, with a fairly complex control system 24 that senses which chillers should be used at any given time 25 and' swings to that chiller depending upon the circumstances -
ANN RlLEY & ASSOCIATES, Ltd.
Q- Court Reporters.
1612 K. Street, N.W., Suite 300 i Washington, D. C. 20006
- (202) 293 3950 ,
l
i I
238 1 that exist at the time it happens.
\
2 Because you have four and you also have other 3 requirements to meet as far as swapping the chillers around j 4 to keep the water from being dormant for chemistry control, 5 etectera, you do end up shifting the chillers quite a bit.
6 It is that shifting of the chillers that results in the 7 continuous failure rate or a higher failure rate for the 8 control building chillers than for the turbine building 3 9 chillers.
10 MR. MICHELSON: But that's not a maintenance.
11 Looking at your tables, you were having a lot of trouble 12 with air-in leakage, for instance, on all the chillers. Of 13 course, you get that if you leave them -- I assume these are J
( 14 shaft-driven compressors.
15 There's a lot of cases where you were down because 16 the air got in and, of course, that will trip it out.
17 MR. SHEWMON: Gentlemen, could we move on, please?
18 [ Slide.]
19 MR. BURTON: Good morning. I'm Joe Burton. I'm 20 the Supervisor of Probablistic Risk Assessment at River Bend 21 Station. I'd like to speak with you a little this morning 22 on how we use our PRA models and methods in conjunction with 23 the Reliability Systems Group to help us enhance performance 24 and reliability at River Bend Station.
25 I've been with Gulf States Utilities for almost 20 ANN RlLEY & ASSOCIATES, Ltd.
, -Q Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
l 239 4 i
! 1 years. The first ten years I spent in our fossil side, then j 2 the last ten have been at River Band Station. I've had five 4
i 3 years experience as a maintenance supervisor in both fossil j 4 and nuclear plants. I've been supervisor of the Independent l 5 Safety Engineering Group, the ISEG, out at River Bend. For
+
6 the last three years, I've run the PRA Group.
7 I've got a senior reactor operator's license or 8 I've held one at River Bend Station. So I've been involved 9 in a lot of different aspects of the plant.
10 [ Slide.)
11 MR. BURTON: This morning I would like to cover 12 several arear. In the interest of time, I will make some of 13 these very brief. I'd like to give a very short 14 introduction to discuss our PRA model just a bit, because 15 some of the things that -- some of the insights we've l
16 received from PRA will help explain some of the information J
- 17 that I want to give you on how we use it in the reliability 18 systems and in the maintenance rule work, and then I will l
- 19 summarize our results.
, 20 MR. LEWIS: Just out of curiosity, I'm always
! 21 curious about where people get their education to become j 22 supervisor of PRA.
23 MR. BURTON: Up until three years ago I couldn't 24 spell PRA.
25 MR. LEWIS: Backwards is easier.
ANN- RlLEY & ASSOCIATES, Ltd.
LQ Court Reporters 1612 K. Street, N.W., Suite 300 Washington,. D. C. 20006 (202) 293 3950
i i !
- 240 i
i 1 MR. BURTON: I have a Bachelor's in Nuclear
! i 4
2 Engineering from Mississippi State University and a Master's !
3 from Louisiana State University.
l 4 MR. LEWIS: I see. And the PRA you picked up on 5 the job.
6 MR. BURTON: I picked it up by doing it.
j 7 MR. LEWIS: Thank you.
8 MR. BURTON: As you'll find if you probe too f
9 deeply into my background.
10 MR. LEWIS: I'm not going to be mean to you. I 11 just wondered.
4 12 MR. BURTON: Okay.
13 (Slide.]
14 MR. BURTON: In terms of using PRA at River Bend 15 over the last three to four years, we've been very active in l 16 trying to integrate the PM program into our decisionmaking 17 at the plant. We've used it to prioritize design changes.
18 We are now starting to use it in our PM l
19 optimization program to help us rank whether or not PMs need
- l
- 20 to be performed. We've done a lot of training with the I
21 operators on PRA results and insights to help improve their 22 awareness of the core damage sequences at River Bend and 23 this type of thing.
1 ,.
24 We have used insights as a way to improve to 25 procedures. During our last refueling outage, we did a 9
,Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300-
, Washington, D. C. 20006 (202) 293-3950 1
241 1 number of safety analyses using the PRA to help determine 2 the level of shutdown risk we'd be facing.
3 We use it typically during operations when we look 4 at the safety significance of tech spec changes, of LCO 5 extensions, justifications of continued operation, because 6 it gives us a way to establish risk significance for events 7 that b ve occurred at the plant or have the potential to 8 occur.
9 (Slide.)
10 MR. BURTON: our PRA Group, we are fortunate, 11 we're kind of your complete severe accident store at River 12- Bend Station. We have not only the responsibility for the 13 individual plant exsmination and the IPEEE and PRA 14 applications that come from that, we also have 15 responsibility for things like station blackout rule 16 analysis, the hydrogen control rule analyses.
17 We provide technical bases for the Emergency 18 operating Procedures that operations uses. We're leading 19 the task force that's putting together the accident 20 management guidelines at River Bend. We do analyses of off-21 site dose and shielding requirements following transients 22 and accidents.
23 MR. DAVIS: Excuse me. Have you submitted your 24 IPE?
25 MR. BURToH our IPE submittal schedule, and ANN RlLEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 )
l 4
242 i
i that's not in the package, but I thought the question might 1.
O 2 come up.
j 3 [ Slide.)
j 4 MR. BURTON: Our submittal for the IPE is the 3
5 first of February. Our IPEEE submittal is June of 1994.
I 6 How, in August of next year, we plan to make a docketed l 7 submittal of our Level 1 and Level 2 PRA methods to NRR so 8 that we can get a safety evaluation report on our methods 9 and actually have something we can use in licensing space.
j 10 MR. DAVIS: Will your IPE submittal be something 11 other than your PRA?
12 MR. BUP. TON: It will be a lot less detailed here 13 than what's submitted here. Here, we will submit all of our 14 system notebooks,'our calculations, our computer software i 15 qualification packages, all sorts of things that we would l
16 not normally include in the generic letter submittal.
l 17 MR. DAVIS: 'lhank you.
18 MR. SHEWMON: I'm not. familiar with getting a PRA E
4 19 docketed. Could you, for 50 words, tell me what you hope 20 that will get you that you don't have now?
21 MR. BURTON: Well, we met with the staff in May.
l 22 All the BWR-6s have an Owners' Group Subcommittee that does -
i 23 IPE work. At that time, talking with both Research and NRR, 24 we came to the realization that the review that we're going I 25 to get for the generic letter and the documentation that I(] ANN RlLEY &- ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington,. D. C. 20006 (202) 293-3950
1 243 1
) I we're going to get back from the staff is going to be a ,
I
- 2 staff evaluation report, which we cannot use as a basis for
] 3 using our PRA methods ove' in licensing space.
l 4 We need a safety evaluation report to do that.
- 5 Well, to get that, we have to give them a lot more detail 6 and that's where this docketed submittal comes in.
7 MR. SHEWMON: Thank you.
} 8 MR. CARROLL: Have they given you any indication i
j 9 as to how long it will take for them to review your 10 submittal?
l 11 MR. BURTON: No, sir.
1 12 MR. CARROLL: Do you have any predictions?
13 MR. BURTON: None whatsoever.
(} 14 15 (Slide.)
MR. BURTON: Now, the PRA, as it was originally 16 conceived at River Bend, we used a focused approach which 17 reduced the size of the model that we were developing. It 18 also reduced the cost of development. But we think it still 19 provides us with a useful decisionmaking tool.
20 MR. KRESS: I'm not sure I understand what.that 21 means.
22 MR. BURTON: A focused scope?
23 MR. KRESS: Does that mean you group events 24_ together?
25 MR. BURTON: No. It's strictly the level of
]
ANN RILEY & ASSOCIATES, Ltd.
COud- Reponers 1612 K. Street, N.W., Suite 300 -
Washington, D. C. 20006 (202) 293-3950
. - . _ - . . -. . _ _ _ ~ - -- - - -.- . .- . _ .-____- . - -
1 J
244 1 detail you go into on the systems. I guess the best example g-w)
# 2 I could give you would be if we're looking at a motor-driven l
i j 3 pump, we've got to get power to the pump, so we have a bus 3
4 out in the plant.
i 5 We know that the pump reliability is probably not 6 as high as the motor reliability. The motor reliability is 7 probably not as high as the circuit breaker reliability, j 8 just on a generic basis. So we cut out those two interim 9 components. We use the pump reliability and the bus i 10 reliability.
11 So there are fewer components. It's a more gross I
12 or a coarser model than some plants have used, but we think 13 it still gives us the information we need for
( 14 decisionmaking.
}
i 15 Now, as we find more and more applications for it, i 16 we intend to improve the fineness of the structure in the
- 17 model. We've included an internal flooding analysis, a very 4 18 detailed analysis of ATWS-events, looking at both human 19 reliability analysis for operator actions and all the e 20 components in our reactor protection system.
21 We've done an interfacing system LoCA analysis.
22 We're doing an analysis of USIs and GIs, and, of course, a 23 containment performance improvement analysis.
24 .)GR. KRESS: Does the ATWS dominate your core.
. 25 damage frequency?
- h ANN RlLEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, -N.W., Suite 300
, Washington, D. C. 20CXXi (202). 293-3950
. --- - . _ _ -- .. . _ - . -... -. - -= .- . _ . . - ..
i 4
1 245 j 1 MR. BURTON: It truncates out.
v 2 '4R . KRESS: It truncates.
! 3 MR. BURTON: Yes, I will get to that in just a 4 moment.
?
- 5 MR. MICHELSON
- What is the USI/GI analysis?
4
! 6 MR. BURTON: That's looking at unresolved safety 7 issues and generic issues that are out there. There are i 8 some that are spelled out in the generic letter, such as A-
!- 9 45 on decay beat removal.
d j 10 K2- MICHELSON: How does this all relate to your 1 11 PRA?
i 12 MR. BURTON: We use the PRA to assess the i"
13 significance of those USIs and GIs at our particular plant.
A 14 MR. MICHELSON: Thank you.
U 15 MR. BURTON: We used a link event tree, a large
- 16 fault tree model. We've got 14 front-line systems and six 3
17 support systems. We looked at 16 initiating events for 18 River Bend. We examined over 250 procedures to identify 19 important human actions. We wound up modeling 81 of those 20 actions in the PRA.
21 We have about 210 basic events in the model that 1
22 use either generic or River Bend-specific' data. Again, this 23 is just --
24 MR.-MICHELSON: You indicate you have 200 -- well, 25 What do you use for input on your motor-operated valve Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950.
._. _ _ _. _ -_ _ __ _.~ _ _ _ . _ ._ _ _ _ _ _ . _._ _ _ _ _ _-_ _
i 246 1 reliability in your PRA?
O 2 MR. BURTON: Right now that's typically generic 3 data out of IEEE.
4 MR. MICHELSON: Because that's the non-loaded 5 data, so to speak.
6 MR. BURTON: Right.
7 MR. MICHELSON: That does not reflect the true 8 operating conditions for the --
9 MR. BURTON: He have our Generic Letter 89-10 MOV 10 program that is accumulating data. As that database builds, 11 we will try to integrate that River Bend-specific data.
- 17. MR. MICHELSON: of-course, even that does not 13 reflect, for instance, the ability to isolate a break.
14 MR. BURTON: Yes, sir.
15 MR. MICHELSON: And there's a very small amount of 16 data in this world on that particular point. But, yet, the 17 PRA, then,_might be misleading you as to the importance of 18 certain valves.
19 MR. BURTON: It might be. Yes, sir.
20 (Slide.]
21 MR. BURTON: Now, our PRJ,model, as it will be 22 submitted in February, gives these results for River Bend.
23 The reason I show these is_because, again, that takes us to 24 _w here we want to go as far as system importance, as we'll-25 get into in just a moment.
-ANN RlLEY & ASSOCIATES, Ltd.
Q Court Reporters 1812 K. Street,- N.W., Suite 300 Washington, D C. 20006 (202) 293-3950
- _ - - _ _ _ - . . . . _ - _ . - _ _ -- . - . - - -. =-. . _ . _ _ _ .- _ - _ _ _
l i 247 l 1 Our core damage frequency is 1.55-E-to-the-minus-2 five per year and it's dominated by station blackout.
3 MR. SHEWMON: Is it E-to-the-minus-five or ten-4 to-the-minus -- okay. That kind of E.
l 5 MR. BURTON: That kind of E, ten-to-the-minus-6 five. We have 86 percent station blackout. About eight 7 percent is transients, another four percent are loss of off-8 site power that did not lead to a station blackout. Lastly, 9 we have a small IDCA contributor, I4CA following a transient 10 type event.
4 11 MR. KRESS: How does that compare with the Grand 12 Gulf?
13 MR. BURTON: Grand Gulf is about 90 percent-14 station blackout and then they have about a ten percent 15 ATWS. In our submittal we will have a detailed comparison 16 of Grand Gulf and River Bend, the Grand Gulf 4550 analysis, 17 and where we differ as far as what we analyzed versus what 18 they analyzed.
19 (Slide.]
20 MR. BURTON: Since station blackout is so 21 important to River Bend, it is pretty obvious what our 22 important systems are going to be. They're systems that
. 23 either get you inte a station blackout or mitigate it once 24 you're there. So AC power in the form of off-site power and
^
25 the diesel generators dominate us in terms of system i
' ANN RlLEY & ASSOCIATES, Ltd.
{- Court Reporters-
, 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
1 I.
248 I 1 importance.
O 2 Next, we have standby service water, which is the i
- 3 primary support for the diesels. Following that, we have i
4 our RCIC system, which is a turbine-driven injection source.
- 5 It's the only thing you've got for high pressure injection l~
6 following a station blackout and it needs DC power.
f
! 7 So when we see the importance ranking like this, i
- 8 it makes sense to us based on the type of core damage ,
! 9 contributions we're seeing.
i 1
10 MR. LINDBLAD: Does that mean that you influenced 11 your line maintenance for off-site power?
1 12 MR. BURTON: We control the maintenance of off-13 site power in some ways. In other ways we don't. By that, 14 I mean Gulf States Utilities has a Transmission Distribution j {}
l 15 Department which is separate from our Nuclear Division.
16 They have PM' programs and maintenance programs that they
! 17 apply to the transmission lines coming into our switchyard, j 18 They also do work in our switchyard.
19 That work is -- access to the switchyard is 20 controlled by our control room, such that our operators are 21 aware of what is going on out in the switchyard at a given 22 time. In the future we may have to have more involvement lLn 23 the --
24 10R. LINDBLAD: So that's a growing influence.
25 MR. BURTON: That's a growing influence. I might ANN RlLEY & ASSOCIATES, Ltd.
-Q Court Reporters 1612 K.- Street, N.W., Suite 300 Washington,' D. C. 20006 (202) 293 3950
249 1 add that this is our first -- the results I just showed you O 2 are our first revision to our PRA model. The initial cut at 3 the PRA model was the plant before the major modifications 4 we did in the last outage.
5 In that model, station blackout was only about 25 l 6 or 30 percent, whereas transients was over 50 percent. Now, 7 our core damage frequency has gone down by a factor of four, ,
i 8 but it's all been in the transient area. So station I i
9 blackout has grown to fill that void. l l
10 MR. MICHELSON: How long do you postulate the 11 station blackout?
12 MR. BURTON: The station blackouts that we have, 13 we have up to a 24-hour emission time in our PRA model.
14 MR. MICHELSCN: What is your survival time?
)
15 MR. BURTON: Well, our coping duration for the 16 station blackout rule is four hours.
17 MR. MICHELSON: Four hours.
18 MR. BURTON: Beyond four hours, we lose DC, we 19 lose RCIC, and we start having problems, but up to four
- 20 hours we're fine.
2 21 MR. KRESS: Would you really lose DC after four i
22 hours?
l 23 MR. BURTON: Our batteries -- the battery-sizing 24 cales we did presume that you have the loss of off-site 25 power, you have to flash the fields on the diesel ~
b ld ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters l 1612 Ki Street, N.W., Suite 300 l- Washington, D. C. 20006
( (202) 293-3950
250 1 generators, you have to carry the RCIC MOV loads, the 0' 2 instrumentation loads for the four-hour duration, and still 3 have enough battery capacity to flash the diesel fields at 4 the end of four hours to get the diesels back on.
5 So that's the sizing cales. We have some margin 6 beyond that. If we were talking about just carrying nothing 7 but the RCIC loads --
8 MR. KRESS: You could probably go many hours.
9 MR. BURTON: It might. We just never run the 10 numbers. We took four hours as being our threshold since 11 that's what we've reported for the station blackout rule.
12 So we may be a little conservative there.
13 MR. MICHELSON: In the case of standby switchgear 14 room cooling, that's done by chilled water, again, I assume.
)
15 MR. BURTON: -That's done by the HVK chillers.
16 MR. MICHELSON: Are you keeping some of the 17 chillers running during this four-hour period?
18 MR. BURTON: They're not available. We don't have 19 AC power to those chillers. The reason the chillers show 20 up, this is where we start seeing some involvement from the 21 loss of off-site power cases, from the transient cases.
22 One our station's specific transients is loss of 23 HVK chillers, which means loss of control building, which 24 gives us a loss of our safety-related switchgear, which 25 takes out our diesels. So it leads you into a station
. f)'
ANN RILEY. & ASSOCIATES, Ltd.
COud RepOders 1612 K. . Street, N.W., Suite - 300 Washington, D. C. 20006 >
, (202) 293-3950
4 251
- Oj 2 1 blackout condition.
MR. MICHELSON: Have you looked at the 1 3 availability of the chillers after four hours of no 4 operations?
5 MR. BURTON: No, we have not.
1 6 MR. DAVIS: Do you consider the predominance of 1
7 the station blackout on core damage frequency a j 8 vulnerability in the context of the IPE definition?
9 MR. BURTON: No, we don't. We consider it i
i 10 insight.
i i 11 MR. DAVIS: You consider it a what?-
12 MR. BURTON: An insight.
l 13 KR. DAVIS: Thank you, t
( 14 MR. CARROLL: Switchyard. A year or two ago,
)
j 15 there was an incident at a plant, I think I remember where 16 it was, but I'm not sure, where a whole bunch of solid-17 state relays got zapped and they lost connection with the 18 grid. The issue came up as to how to restore breakers and 19 there was real confusion between people that ran the 20 transmission system and people.that ran the plant.
21 The people that ran the transmission system wanted 22 to restore breakers in a sequence-that minimized disruption 23 of service to customers. The people that ran the power l
24 plant wanted to get off-site power back.
25 MR. BURTON: Exactly.
P
(
i ANN RlLEY & ASSOCIATES, Ltd.
Coud Rep 0ders 1612 K. Street, N.W., Suite 300 Washington, D.~ C. 20006 (202) 293-3950
252 1 MR. CARROLL: Who wins in a case like that at O 2 Grand Gulf? )
3 MR. BURTON: At River Bend, in a case --
4 MR. CARROLL: I'm sorry.
5 MR. BURTON: We'll be part of energy someday 6 maybe, but it's still river Band. At River Bend, one of the 7 things that we did as part of our station blackout rule 8 compliance was to go into the corporate procedures, the ones 9 that the system dispatcher uses, the ones that our local 10 dispatchers use in Baton Rouge, and include verbiage in 11 there to the effect that River Bend Station has the highest 12 priority for restoration of power, that it has to be done in 13 accordance with our methods at the plant.
14 So the direction is there in their procedures to 15 do it our way.
16 MR. CARROLL: Very good.
17 (Slide.)
18 MR. BURTON: Now, based on what we've seen at 19 River Bend, there is no single system, component or operator
- 20. action that dominates our core damage frequency. That may 21 seem odd given what we've talked about on station blackout, 22 but, if you remember, to get to a station blackout at River -
23 Bend, we've got to have loss of off-site power and we have 24 to lose-three separate = diesel systems.
25 So it's not a single system. If you called it AC ANN RlLEY & ASSOCIATES, Ltd.
l] Court Reporters.
1612 K. Street, N.W., Suite 300 Washington, D.- C 20006 (202) 293-3950
253 1 power, yes, that's a single system, but it is actually 2 multiple systems involved in that.
3 MR. DAVIS: But you could lose them all with loss 4 of service water.
5 MR. BURTON: Well, loss of standby service water.
6 We would also have to lose normal service water, which would 7 take a loss of off-site power. It still takes more than one 8 individual system to give us a problem. It may take one 9 initiating event, loss of off-site power, to get you into 10 that, but you still have to lose multiple systems.
11 (Slide.)
12 MR. BURTON: Now, in terms of component ranking 13 for PMs, this is something new that we've started as part of W
/^ 14 our PM optimization program. We use the CAFTA software to b) 15 manipulate our Level 1 model. It's good at providing for 16 structure, system and components probability numbers and 17 various importance ranking, whether Fussell-Vesely or risk 18 reduction, risk achievement.
1 i 19 There are about six different importance rankings 20 that it will give you. We use Fussell-Vesely for no other 21 reason than it's pretty much a standard one that people use 22 out in the industry.
23 But we look at the importance for failure to 24 start, failure to run, maintenance unavailability; for
. 25 valves, failure to open, failure to close; for other i
ANN RILEY & ASSOCIATES, Ltd.
.]
Court Reporters 1612 K. Street, N.W., Suite 300
- Washington, D. -C. 20006 (202) 293-3950 1
4 254 1 systems, loss of function and common cause failure. ,
2 We look at all of the importance rankings for 4 3 those to come up .ith our total ranking. Then we go back 4 and look through the entire system and to make it a little i
5 more user-friendly for the operators and for the system 4
- 6 engineers, we take out all the E-to-the-minus-whatever type t
7 numbers and just make it one-to-2,000, some kind of scale a
8 like that. We normalize it on the lowest number, such that 9 the lowest number becomes one. We provide those rankings to j 10 our system engineers.
11 Now, that's the first step in a PM optimization
- 12 program that's still in the development stage, bu' we think 13 it will look something like this when it's finally
/'T 14 complated.
V' 1 15 (Slide.).
16 MR. BURTON: We start out with a PM screening 17 where we will look at risk. We'll look at the effects on 18 plant output of the PM. We'll look at regulatory compliance 19 impacts of the PM or, in this case, the STP. And we'll look 23 at the cost of that particular PM activity.
21 Any PMs that need to be examined that are not 22 screened out from these categories will go into the focused 23 PM program and we'll apply root cause analysis techniques
- 24 and predictive maintenance techniques, like thermography and 25 oil analysis, vibration analysis, to test the effectiveness-ANN RILEY & ASSOCIATES, Ltd.
, Coud Repoders
, 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950
1 l I
- - 255 l
' l 1 of those particular PMs, the ones that didn't get screened !
[O i
2 out.
i-3 Then, of course, our Reliability Systems Group 4 will do the monitoring and keep feeding that back into the 2
i 5 screening process.
6 MR. CARROLL: On the importance, on your preceding i 7 slide, you don't need to put it up, is there a category of
! 8 things that deals with unwanted actions? It fails and i
. 9 creates an unwanted action, as a control system might do, i 10 MR. BURTON: That's sort of lumped into loss of j 11 function.
l 12 MR. CARROLL: Loss of function means it doesn't 13 work.
14 MR. BURTON: In the manner that we wanted it to.
15 MR. CARROLL: In the manner you wanted it to.
f 16 MR. LINDBLAD: A spurious action, a-safety valve l 17 or something like that.
h 18 MR. BURTON: There is some spurious. signal type 19 information embedded in the various fault trees, j- 20 [ Slide.]
21 MR. BURTON: One of the areas that we got involved 22- in early on was looking at availability versus core damage-23 frequency. This came out originally as.a result of an INPO 24 assistance visit in 1991 and it involved a set of parameters 25 widch they call the Safety System Performance Indicator
/ ANN RILEY & ASSOCIATES, Ltd.
'] _
. Court Reporters 1612 K. Street, N.W.,- Suite 300 Washington, D. C. 20006 (202) 293-3950 e
~
1 256 1 Program.
2 There are industry-wide goals that have been 3 established for the SSPI Program and we wanted to look at i
l 4 the effect on core damage frequency of varying system l 5 availability.
6 The reason we wanted to do this is because in 7 dealing even with the other BWR-6 Mark IIIs, we found that 8 each plant has unique contributors to core damage due to j 9 their plant configurations, they have different architect
- 10 engineers, different designers, different philosophies in
! 11 operating equipment. All those things can contribute to e
i 12 what's important and what isn't at the plant.
i 13 We think that the PRA methods provide the best way 14 of looking at those system interactions and it gives you a 15 measure of overall risk in terms of core damage frequency.
l
- 16 We think
- for River Bend, our PRA methods are the i
i 17 best way to determine safety system performance indicator 18 goals. In fact, later, the maintentnce rule will ask us to
[ 19 do something along these lines to establish maintenance
- 20 goals.
21 (Slide.]
- r. 22 -MR. BURTON: Just as a for instance, and I hope 2
f 23 this is not too light, this is our high' pressure core spray
- 24 system. Our industry goal is about 98 percent and if you 25 run all of the numbers through the HPCS fault tree,'you get l
ANN RlLEY & ASSOCIATES, Ltd.
lQ
- Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
i l
257 1 an availability for the system that's about-98 percent, 2 alsa.
l 3 Well, that's down in here at our 1.5-E-to-the-4 minus-five core damage frequency. The key thing to look at
{
! 5 here is, for instance, if you doubled, you would only double 6 the core damage frequency if that system degraded all the
. 7 way back to 70 percent. Now, that's not saying that we
- 8 would ever let HPCS degrade to 70 percent.
I l
9 The key here is that it gives you a method whereby 10 you can look at changes in the,_say, 95 to 100 percent-or 90 11 to 100 percent range and assess how effective those changes ,
12 are in terms of risk. You can also see if those changes are 13 even cost-beneficial or not.
14 MR. KRESS: I'm a little surprised that that is 15 all the way linear back to zero. At some point, you've got 16 the other systems that would take over.
17 MR. BURTON: Well you do. We see linear behavior
- 18 in all the systems except our residual heat removal system,
- 19 and that's because it serves several different functions, l 20 from injection all the way to containment heat removal. In 21 that particular curve, you see more of an exponential, but 22 all the other systems-that have a single function and the-23 support systems, like the diesel, we see this-linear l 24 response.
25 MR. MICHELSON: Do you see the same on the water
,(~}
ANN RILEY & ASSOCIATES, Ltd.
Court Reporters -
1612 K. Street, N.W., . Suite 300 1 i
Washington, D. C. 20006 ;
(202) 293-3950
=. _ - .
258 1- 1 system?
2 MR. BURTON: Yes, sir.
3 MR. MICHELSON: The essential water.
. 4 MR. BURTON: Yes, sir.
i 5 (Slide.)
6 MR. BURTON: Now, on a monthly basis, as Skip 7 talked about earlier, we do risk trending on RCIC, HPCS and 8 our A, B and C emergency diesels. On both a monthly and a 9 12-month sliding average, we take numbers from reliability 10 systems for_ standby unavailability, failure to start, 11 failure to run. We incorporate those numbers into our PRA 12 model and we compare those back to a baseline.
13 [ Slide.)
- 14. MR. BURTON: And that gives us this curve, again.
15 One thing that Skip mentioned, and I will clarify that now,
- 16 this baseline of one at this point, the number that is there 17 right now is six-E-to-the-minus-five, not 1.5-E-to-t-_a-18 minus-five.
19 That's because from this point back we were using j 20 our first, our original PRA model. The current PRA Model 21 includes the major modifications that were done during this 22- six-month outage. Now, we are' recalculating these numbers 23 after the outage and they will be done using the new model.
24 When they are, you'll. basically see -- you may see 25 a step discontinuity from here back and here forward where -
' ] ANN RlLEY & ASSOCIATES, Ltd.
Coud Reponers
. 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
259 i
1 we change our baseline. When you have a living PRA process 773
- \ 2 like we have at River Bend, you're going to see those step d '
j 3 changes if you make a major change in the model, because
, 4 you'll make a change in your baseline number.
1 5 MR. LEWIS: Are you doing the recalculation 6 inhouse?
! 7 MR. BURTON: Yes, sir, we are.
8 MR. DAVIS: Let me see if I understand that.
9 After you've made the changes, now one becomes 1.5, but s
i 10 you're still well above that. .
11 MR. BURTON: Well, one was -- this was six.
12 MR. DAVIS: Right.
13 MR. BURTON: From here back it was six-E-to-the-i 14 minus-five. From here forward it's going to be 1.5.
}
15 MR. DAVIS: That's right, but you're above 1.5 now 16 for September of 1992.
i 17 MR. BURTON: No. We haven't made the change yet.
l 18 MR.-DAVIS: I thought that represented your i
- 19 change.
L 20 MR. BURTON: No. We're going to have to basically 21 have a break in the chart from here to here.
l 22 MR. DAVIS: Thank you.
23 MR. BURTON: It's just not in-there yet. We 24 haven't had time to get it done, with working on the 25- submittal.
O ^"" ai'ev
- Associates. 'id-COud RepOders 1612 K. Street, N.W., Suite 300 Washington,- D. C. 20006 (202) 293-3950
l l
i 260 t
s 1 MR. CARROLL: Did I understand that during the 2 outage you did some things that affected the --
3 MR. BURTON: The major change we had during the
- 4 outage was we changed our normal service water system from 5 an open-loop system to a closed-loop system. That had a 6 significant impact on our transient response at the plant, 1
7 the transient being loss of normal service water.
8 That's one of the things that dropped our core i 9 damage frequency from six-E-to-the-minus-five to 1.5-E-to-i 10 the-minus-five. As I say, with a'living process, we will
- 11 see those types of discontinuities as we make major model 12 changes.
13 (Slide.]
I\ 14 MR. BURTON: Now, in the maintenance rule area,
!V 15 the draft reg guide came out in March of this year. Some of l
16 the things that the reg guide recommended were that we use l
- 17 PRA to estimate core damage frequency for systems, 18 structures and' components. We can do that with the. ranking l 19 system that we've developed to use with our PM optimization i 20 program.
I 1 21 The reg guide asks.that we use PRA to establish 22 reliability and availability goals. .We've been 23 experimenting with that and using core damage frequency as 1
~
24 our measure of availability.
25 It asks that you examine the PRA-cut sets to 4
O ^"" a"Y * ^SSoc'AT5S' d-Coun. Reponers 1612 K. Street, N.W.~, Suite 300 Washington, D. C. 20006 (202) 293-3950 1
l 261 i 1 identify any single failures. At this point, we have none h- 2 that we've identified. And common cause failures.. We-have- -
! 3 common cause failures built into our individual system fault j
l 4 trees and we can readily identify those.for maintenance and 5 for operations.
1 l
6 (Slide ]
i j 7 MR. BURTON: We also were asked in the reg guide 8 to identify risk due to initiating-events. I didn't bring f
f 9 it-with me, but we've got a pie chart that shows-initiators i
10 and based on what we've seen up here, like 90 percent of our l 11 initiating events are loss of off-site power, which leads to i
12 core damage. So that's readily available to us'through the l
l 13 PRA.
14 The-last item that the NUREG asks for is the hard-I 15 part'from a risk point of view,.and that's doing an
- 16 evaluation of maintenance work in terms of risk. There are-I
- 17 a couple ways suggested. .one is an on-line evaluation of each maintenance work order.
~
i 18 We don't:know what: form that i
- j. '19 would take yet.
I
- 20 -Some of the things that.are-recommended
- are-to 21 control plant configurations, try.to eliminate any high-
- 22. consequence, high risk
- combinations.of outages on different-l 23 components.- There may be. guidelines.
i 24 Some plants have: developed! lists of high risk ;
25 -combinations'that they give to.the shift supervisors and the. l y
ANN: RlLEY & L ASSOCIATES, Ltd.
]:--- -Court Reporters-c 1612 K. Street, N.W., Suite 300 4 Washington, D. C. 20006
~
-(202).293-3950:
2
4 262 l 4 1 maintenance planners to make sure that they don't anter into 2 those combinations. Another way'to go that's recommended in 3 the NUREG is to use more root cause analyses, MORT charts or 4 vent and causal factor charting to try to identify precursor
- 5 events and potential initiators from a particular
- 6 maintenance activity.
7 Of course, the last one and the most rigorous one 8 would be to do a detailed PRA analysis of each maintenance
, 9 activity. We don't know how much that would take in terms t
10 of resources at this time.
) 12 MR. BURTON: We have done -- for this past
! 13 refueling outage, we did several things. One, there were a j 14 number of options that the Outage Department proposed to do
- 15 the outage. We evaluated each one of those and, on the r
16 basis of risk, made them throw out several of those options.
l 17 We also developed a-tool'that we call the master 1
[
i 18 plant logic diagram, which is basically.an event tree tied 19 to a huge dependency matrix for all the systems and 20 components that we'd need during the outage. We used that 21 to track changes in the outage schedule.
22 So we have tried to get into shutdown risk, but-3 23 we're taking our initial steps there.
24 [ slide.]
25 MR. BURTON: In conclusion, we think that our PRA ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202).293-3950 4
m - - ---- - . -. r,- , , , - w
263
.gg 1 model at River Bend gives us a valuable tool to be able to 2 rank importance, whether we're doing'PM optimization, to be 3 able to set risk-based goals for availability, and to be 4 able to trend the risk significance of changes in i
- 5 availability.
~
6 Doing these things gives us a way to develop the j 7 type of goals and priorities that the maintenance rule
. 8 requires and it gives us a tool we can use to evaluate risk 9 significance when we change plant configurations during 10 maintenance.
t 11 As I said before, we want to use our PRA to help 12 resolve regulatory issues. We need concurrence and support
, 13 from the NRC staff to be able to do that. There are a I[
}
14 number of issues out there where we feel like PRA could make 15 a significant impact in either reducing the scope of it or 16 perhaps focusing the resources a little differently.
17 The most important one of those to River Bend 18 right now is using PRA to assess the risk significance of
- 19 thermo-lag installations in our_ plant. We are the lead i
i 20 plant in an EPRI tailored collaboration to develop fire PRA 21 methodology for the IPEEE.
22 As part of that, one of the items in the IPEEE is 1
i 23 to assess issues that came out of the fire risk scoping 24 study. One of'those issues is degraded fire barrier 25 effectiveness. Well, thermo-lag is a fire-barrior. So
- ANN RILEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300
- : Washington, D. C. 20006 (202) 293-3950-
l 264 j is f 1 we're being told to evaluate it, to help answer Generic 2 Letter 88-20. We want to be able to use those results when 4
3 we get over into the question of which thermo-lag are we 4 going to fix, how are we going to fix it, how long does it
- 5 need to function.
6 But at this point, thermo-lag is being approached 7 as a compliance issue rather than something that's amenable 8 to risk-based regulation. So that's one area that we would
{ 9 like to see PRA techniques married into the resolution.
l l 10 MR. MICHELSON: Have you actually done your fire-i
{_ 11 PRA?
4 12 MR. BURTON: We're working on it. We expect to j 13 have some preliminary results during the first quarter of
() 14 15 1993.
MR. MICHELSON: Because it will have to be 3.
! 16 fairly sophisticated PRA before-you can start making thermo-17 lag decisions with it.
18 MR. BURTON: Well, yes and no. In the modeling-of 19 fire effects in a particular area, we can go back, based on.
20 our component importance, and determine, first of all,.does
- 21 this fire area have any components that are risk
, 22 significant. If it does not, then from a core damage point i; 23 of view, we're not too concerned about the thermo-lag in s
24 that area.
25 MR. MICHELSON: There probably isn't much-thermo-i,
[] ANN RILEY & ASSOCIATES, Ltd.
j Court Reporters 4 1612 K. Street, N.W., Suite 300
{ Washington, D. -C. 20006 (202) 293-3950'
i
- 265-i 1 lag -- thermo-lag was used for divisional barriers in most 4
O 2 cases, i
3 MR. BURTON: But it was used for safety-related 1
l 4 equipment. There's a lot of safety-related equipment out i
j 5 there that's really not risk significant from a core damage
- 6 point of view, i
? MR. MICHELSON: And still protected by thermo-8 lag.
- 9 MR. BURTON: And still protected, yes, sir.
j 10 MR. MICHELSON: What was the basis for that?
11 MR. BURTON: Appendix R.
12 MR. MICHELSON: Yes, but Appendix R addressed 1
?
13 safety-related divisions.
lT 14 MR. BURTON: Yes, sir.
LJ 15 MR. MICHELSON: Okay, i
16 MR. SHEWMON: Another day.
{
17 MR. BURTON: It's another battle for another day, 18 yes.
19 MR. MICHELSON: Have you done your external events 20 or are you working on other external events besides fire?
21 MR. BURTON: For River Bend, we're a reduced scope 22 plant for seismic, which means that from a seismic point of 23 view, we basically have to do a walkdown to determine if 24 there are any_ outliers. We've done that. We should have 25 the results --
ANN RILEY & ASSOCIATES, Ltd.
Q' Court Reporters-1612 - K. Street, N.W., Suite 300 Washington, D. C. 20006 (202)-293-3950
l 266' s 1 MR. MICHELSON: I'm more interested in pipe break.
\ l 2 MR. BURTON: Pipe break, we have assessed that as 1 1
3 part of our internal flooding analysis fr. the IPE.
4 MR. MICHELSON: So you already know the 5 significance of certain of your isolation valves.
{
6 MR. BURTON: Yes, sir.
7 MR. MICHELSON: How did you determine what their l
8 reliability was in doing your PRA for it?
, 9 MR. BURTON: We had to use the generic numbers 4
10 that are available.
j 11 MR. MICHELSON: You do not yet know that's wrong.
12 MR. BURTON: Yes, sir. I agree. There will be i
13 changes. However, we use the best available information.
- /i 14 MR. WILKINS: Do you know which way it's wrong?
- (m,/
l 15 MR. MICHELSON: Yes. It's very non-conservative.
16 MR. KERR: In your efforts to use PRA in licensing i
17 space, as you put it, have you concluded that this will
. 18 require a change in existing regu3ations,.which, at least on 4 19 the face them, one would assume are not quantitatively _ risk 20 related, at least.
21 MR. BURTON: To some extent, yes, sir. For 22 instance, the area of thermo-lag, that's one that.I feel 23 very strongly about. We've got Generic Letter 88-20 that
- 24 gives us guidance on using PRA techniques to evaluate fire 25 barriers. But you read the draft generic letter on thermo-(]- ANN RILEY & ASSOCIATES, Ltd.
COud RepOders-1612 K. Street, N.W., Suite 300 3 Washington, D. C. 20006 (202) 293-3950
1 I
267 1 lag, it makes absolutely no mention of using PRA methods.
>O
- 2 It acts as though they don't exist, breically.
3 MR. KERR: Well, the existing body of regulations 4 acts as if it doesn't exist.
5 MR. BURTON: Yes, sir. And that's the type of f
6 support we need from the staff to be able to try to inject 7 some PRA methodology into regulatory space.
4 8 MR. LINDBLAD: Excuse me. I think the question is 9 interesting. From your viewpoint, do you think the staff l 10 can do it without changing regulations or that they need to 11 change regulations in order for the staff to give you what 12 you want?
13 MR. BURTON: I would think -- and, of course, I f 14 may have to defer to Jim Booker since he's had a lot of 15 years in.our licensing side of the house on this, but I 16 would think that we do need some additional flexibility in 17 the regulations. What the best way to get there is I don't 18 know.
19 MR. KERR: You mentioned that you had done at 20 least some looking at shutdown risk. There are studies
' 21 that, I'm sure you know, would indicate shutdown risk is
- 22 equal to or perhaps greater than-operating risk.
23 Has your work given you any insight into whether 24 this is the case for River Bend or not?
- 25 MR. BURTON:- Well, in some of the work that we did 4
ANN RlLEY & ASSOCIATES, Ltd.
4 -
Coud Repodens:
1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 f
268 gS 1 for this last refueling outage -- let me just describe b 2 briefly what we did. During this outage, we had to do a 3 full core off-load because we had to do some nozzle work on 4 the vessel.
5 The full core off-load was being done at the same 6 time that we were doing our normal diesel maintenance. It 7 was done at the same time that we were dcing the closure of 8 normal service water and the chemical cleaning of our 9 standby service water system.
10 So there were a lot of important systems that were 11 coming up and down during that outage while we had a full 12 core off-load, which is something we had never done before.
13 With that in mind, we wanted to make sure that we assessed
(~~) 14 the risk of the outage as thoroughly as we could.
'w) 15 one of the things that we did was to look at, once 16 we got the full core off-load, the amount of time the 17 operators would have to mitigate, say, a loss of cooling to 18 our spent fuel pool. We found that we were dealing with a 19 minimum of eignt or nine hours before we got to a point 20 where we would start to boil the spent fuel pool.
21 We were dealing with 90 to 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> before we 22 would actually expose fuel. It made about a four or five 23 orders of magnitude decrease in the risk when you looked at 24 it from the point of view of the time the operators would 25 have to mitigate the loss of cooling and this type of thing, (l
ANN RILEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
l 269 i
1 before you actually did any damage to the fuel.
- 2 So we think that the designs we've got out there-
- 3 at River Bend are very forgiving for shutdown situations.
i 4 However, when we're doing major evolutions like that, we
! 5 feel like it does need as thorough an examination aus we can 6 give it.
! 7 MR. SHEWMON: Why don't we stop at this point.
t 8 MR. CARROLL: Yes. I guess we are running out of 9 time. Ua will have a break, so that people who have 10 additional questions can talk to the River Bend people
! 11 during the break, i
12 We do want to thank River Bend folks for making l
j 13 this presentation. I think it's been very interesting to
() 14 15 all of us. I would only say keep up the good work.
sounds like the sort of thing all utilities ought to be
.This 16 doing.
f 17 MR. CATTON: It might even make a believer out of i
l 18 some of us who are a little bit on the edge with respect to i
19 PRA.
20 MR. BOOKER: Let me closelreal quick. Again, let 21 me say we appreciate the opportunity to meet with you all 22 and-share with you all.the experience we're gaining in this 23 system reliability and monitoring and maintenance program 24 and using PRA.-
25 We think it's a very useful toollto identify the-
~O ^a" ai'ev * ^SSoc'ATES' d-
_ Coud Reponers
-1612 K. Street,- N.W., Suite -300
- Washington, D C. 20006
.(202) 293-3950
l l
270' 1 things we need to pay attention to. We need to direct our 1 2 resources to systems and things that will give us the 3 greatest bang for the buck for both safety and reliability.
l 4 We encourage you all to support this program.
5 Thank you.
j 6 MR. SHEWMON: Let's recess.
) 7 (Recess.)
8 MR. SHEWMON: The next-item on the agenda has to l
9 do with protection against lightning and electrical 10 transients for nuclear power plants. Charlie, do you want 11 to introduce things?
12 MR. WYLIE: Thank you, Mr. Chairman. This portion 1-23 of our meeting is to review the NRC staff's proposal i
14 regarding a petition for rulemaking to amend the NRC rules 15 and regulations to add lightning-induced and other '
16 electrical transients to the list of phenomena that licensed 17 nuclear power plants must safely accommodate.
18 The material for this meeting is under Tab-6 in
, 19 your books.
l
- 20 (Slide.)
21 MR. ROURK: My name is Chris Rourk. I work in the 22 office of-Research. I'll be presenting the staff's position-23 on the petition for rulemaking, 50.56, which is sometimes 24 referred to as the petition on lightning protection.
I 25- (Slide.]
- ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
271 1 MR. ROURK: The actual concern -- the petition 1
'- 2 regarded all electrical transients. The petition requested 3 rulemaking to reqaire licensees to analyze all electrical 4 systems in the plant, including balance-of-plant systems for 5 the effects of electrical transients. Some of the sources 6 mentioned were nuclear EMP, lightning and electromagnetic 7 interference.
8 (Slide.]
9 MR. ROURK: The petition also requested that the j 10 results of this analysis of the effects of electrical 11 transients be used in a PRA analysis to determine what the 12 probablistic type effects on the plant might be, also.
13 [ Slide.]
14 MR. ROURK: The staff published a summary of the
.(}
15 petition in the Federal Register on December 23, 1991, 16 requested public comments by February 21, 1992. The staff 17 completed a technical study in September of this year and 18 schedular guidance by the EDO requests.that response to the 19 petition be published by the end of this month.
20 (Slide.]
21 MR. ROURK: The staff proposes to deny the 22 petition in total.
23 [ Slide.]
24 MR. ROURK: Four public comments were received on 25 the petition. Two were from private citizens, one was from f) ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
l 272 g 1 a utility and one was from a citizens group. Three comments j 2 favored the petition, one opposed. The four major points 3 from the public comments were that the increased likelihood 4 of a nuclear weapon attack requires hardening of nuclear 5 plants from the effects of nuclear EMP.
! 6 MR. DAVIS: What was the basis for the conclusion
. 7 that there is an increased likelihood of attack?
! 8 MR. ROURK: There wasn't a basis stated. I think 9 it was just the breakup of the Soviet Union was referred-to.
, 10 MR. DAVIS: I thought that went the other wey, i
- 11 (Slide.)
12 MR. ROURK: Several of the comments referred to
! 13 the delicate nature of electronic and solid-state components 1
'O 14 and asserted that they required regulation and protection.
b 15 One of the comments stated that the IPEEE program should 16 require all licensees to consider effects on their plant, l 17 besides loss of off-site power from lightning.
I 18 The fourth comment from the utility said that NRC 19- chould develop guidance before requiring any actions or 20 actions in the petition would have required utilities to 21 take action before the NRC developed guidance.
22 MR. CARROLL: What_does the IPEEE program require 23 -in the way of lightning at the-present time?
24 MR. ROURK: The IPEEE program -- I'll get to that 4- 25 later, if I can wait.
ANN RILEY & ASSOCIATES, Ltd.
<'(]-
Court Reporters-1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 ,
(202) 293-3950
. 273 4
1 MR. CARROLL: Sure.
l --
2 MR. COE: If I may, you can look at on Page 95 in 4
3 the briefing book and there's a couple of excerpts from the f 4 generic letter that governe the IPEEE.
I 1 5 MR. CARROLL: Okay.
) 6 (Slide.)
7 MR. ROURKs We first reviFfed the GDC and i
j 8 determined that they currently have provisions which would l 9 allow the staff to require licensees to perform any type of i
l 10 analysis in the area of electrical transients.
11 GDC-2 requires that SSCs be designed'to withstand j 12 the effects of natural phenomena, such as lightning and geo-13 magnetically-induced currents.
.[ 14 MR. WYLIE: Are you going to tell us what the 15 staff does to review those two ---the actions by the
- 16 utilities on meeting those GDCs?
I 17 MR. ROURK: Yes. That's covered later.
18 MR. WYLIE: Okay.
19 (Slide.)
20 MR. ROURK: .Also, GDC-4 requires _SSCs to be 21 compatible with-the operating environment. It's noted that 22 IEEE 1050 and other standards-refer to the operating 4
23 electromagnetic environment for designing electronic i
24 components.
25 .(Slide.)
4 ANN RlLEY & ASSOCIATES, Ltd.
COud RepOders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
h 274 .
1 j 1 MR. ROURK: The NRC has also previously studied i.
2 the effect of nuclear EMP on nuclear power plants and l-
! 3 determined that it would not prevent the safe shutdown of j 4 nuclear power plants. The draft reg guide on lightning 5 protection was previously presented to the ACRS, actually, 1
6 and issued for comment. A cost benefit and risk analysis on
- l. 7 the draft guide indicated it didn't have-a significant i
j 8 safety benefit.
t l 9 Part of the problem was the surge arrester sizing.
10 (Slide.]
11 MR. ROURK: The staff's licensing reviews l 12 considered industry standards for plant structural 13 protection, switchyard shield wires, grounding of equipment, 14 and other relevant' standards. I believe I'm right in saying
(
15 there were no explicit requirements in the SRP-for 16 protection against lightning.
L' 17 One of the things that's noted is, like the i 18 national electric code is incorporated by state statutes,
- 19 building standards and many of the IEEE standards'are used 20 by utilities for transmission lines, switchyards, plant 21 grounding.
22 So there is already stuff-out there.
23 MR. WYLIE: You say the staff reviews -- delved 24 into all of these.
- 25 MR. ROURK: Yes. From what I understand, it
' h' ANN RILEY & ASSOCIATES, Ltd.
' Coud Reponers
- 1612 K.- Street,' N.W., -Suite 300 Washington, D. C. 20006 (202) 293-3950 l
- b i
i i 275
- 1 involved consideration of all these existing standards.
- 2 MR. ROSA
- Faust Rosa, NRR. Historically, the 3 staff has relied on industry to design their plants in 4 accordance with established standards. In our prelicensing
- 5 site visits, the staff did go through the plants and note i
! 6 overhead ground wire shielding in the switchyards, the 7 installation of surge arresters, and also the fact that l 8 there was a grounding system and that the instrumentation j 9 grounds and the equipment grounds were separately conducted i . .
10 through the plant.
l i 11 Beyond that, we simply relied on the industry 12 design standards and practices having been implemented by
{
13 utilities.
14 MR. WYLIE: Is that a documented review?
(
- 15 MR. ROSA
- Usually not. It could be documented in l 16 SERs on some plants. It all depended on the reviewer.
17 MR. WYLIE: It is basically not a documented
, 18 review.
19 MR. ROSA: Right.
20 (Slide.]
21 MR. ROURK: As Mr. Rosa has indicated, industry
{
22 standards and design practices were used.
23 'MR. WYLIE: Of course, the lightning protection
. 24 for generating stations, the standard didn't exist until 25 just a few-years ago and most of the plants were already
~
4 O
ANN RlLEY & ASSOCIATES, Ltd.
Coud. Repoders 1612 K. Street, N.W., Suite 300 t Washington, D. C. 20006 (202) 293-3950
276 1
gs 1 designed and built by that time. 1
2 MR. ROSA: There were a sufficient number of 3 papers published by the IEEE that addressed both grounding l 4 and protection against electromagnetic interference in 5 electronic equipment and also lightning. protection for 6 switchyards.
! 7 MR. WYLIE: But to the extent those were used 8 nobody knows.
9 MR. ROSA: That's correct, except the cursory 10 review that I've just described.
]
- 11 [ Slide.]
- 12 MR. ROURK
- We also had the fact that the 13 equipment and systems that were used in nuclear plants were 14 similar or identical to the equipment and systems used in t
15 conventional power plants and they had performed 16 satisfactorily.
I 17 Additional qualification testing was required of 18 systems which relied on solid-state components,-such as the 19 Eagle 21.
20 ER. CARROLL: On that point, I read the excerpt 4
- 21 from the Zion Eagle 21 backfit, which you will find on Pages 22 96 and 97 of the briefing book. Gee, I thought that was
- 23 pretty superficial. It basically said that Westinghouse and 24 -Commonwealth appeared to do the things that ought to be done-25 and we're happy.
f
- . (] ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders i 1612 K.- Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
4 277 l 1 I didn't get any serse that you required any l 2 additional testing or anything with respect to grounding.
i- 3 MR. ROURK: I didn't get a copy of that. I'd have 4 to look into that to answer the question specifically,
+
5 unless you know anything right off the top of your head.
6 I'd have to check into that and see if anything 7 additional was done or what was done.
i 8 MR. ROSA: I might add a bit of information. I do 1
j 9 know specifically that in our reviews of the initial solid-10 state systems, like the Westinghouse solid-state logic-i l 11 protection system and the analog instrumentation systems, l 12 they did undergo-prototype testing to-establish their i-l_ 13 capability for performing their function in the l 14 electromagnetic environment that.you would expect in a
- 15 nuclear power plant.
j 16 MR. WYLIE: Is that a documented report?
r 17 MR. ROSA: I believe it is. I couldn't put my 18 hand on it right now, but I'm almost positive that's the
- 19 case.
20 MR. WYLIE: I think we should get a copy of that.-
!- 21 [ Slide.]
22 MR. ROURK:- The current design-factors of i
23 practices and standards _we've been referring to. It's a
l 24 basically noted tha't-they were developed over 50, maybe 100.
25 years,-as electric power developed and events occurred,-
y-ANN RILEY & ASSOCIATES, Ltd.
Q- Court Reporters 1612 K. Street, N.W., - Suite 300 Washington, D. C. 20006 (202) 293-3950
278 f~g i lessons were learned, such as transformer failures and
- ' 2 generator failures from switching or lightning surges.
3 So a lot of these things were' incorporated just 2 4 into the basic design practices. It might be in the I
5 national electric code or local standards for structural 4
6 protection.
7 (Slide.]
8 MR. ROURK: Some examples of this are the 9 transmission line shield wires. The main station i
10 transformers have capacitive grading on the windings. That 11 was documented in the early 1970s and has been applied to 12 prevent the transmission of surge over-voltages from the 4 13 high voltage to the low voltage side.
14 Coaxial cable was developed to --
- 15 MR. CARROLL
- Physically, how do you do capacitive i
16 grading in a transformer winding?
17 MR. ROURK: Well, you have to do an electric field 18 analysis of the winding design and see what the capacitive 19 coupling is between the two windings, and yod minimize it so 20 that the primary capacitive couplings between-the high 21 voltage winding is not between the high and the low.
22 So that when the surge comes in, the' capacitive l
23 voltage division between the high. voltage' terminals and low 24 voltage terminals minimizes the coupling between the two 25 terminals. There's winding-to-winding-capacitances and ,
ANN RILEY & ASSOCIATES, Ltd.
l Q' Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
279 1 winding to high voltage and low voltage winding
- \' 2 capacitances.
3 MR. LEWIS: You can't really measure the winding-I 4 to-winding capacitance. You can only -- once you have the i
5 transformer, all you can measure is between the terminals, 6 right?
}
7 MR. ROURK: That's true. But you can design --
8 you can use finite element analysis and before computers, l
i
- 9 they had things like teledeltas paper that they would do 10 electro -- the electric field design with.
11 MR. LEWIS: You can take the geometry of the 12 winding, you mean.
13 MR. ROURK: Yes.
! (' ; 14 MR. IEWIS: But then it depends very much-on how
%/
15 good the coating on the wires is and how tightly they're-
, 16 wound and that kind of stuff. You really don't know that, j 17 .do you?
- 18 MR. ROURK
- Well, measurements are done, BIL 19 measurements are, transformers. I don't know for sure, but 20 I'm pretty sure there would be a standard measurement to 21 just; check that they work.
22 If the insulation broke down, you'd have a 23 catastrophic failure of the transformer, which would 24 probably.be the most likely result from changing insulation e
25 strength.
i
-(] ANN RILEY & ' ASSOCIATES, Ltd.
Coud Repoders 1612 K Street, N.W., Suite 300 Washington, . D. C. 20006 (202) 293-3950
\
i 280 gg 1 MR. LEWIS: But that's what I'm addrecaing. If
~'
2 you're worried about the vulnerability to a surge, then the 3 vulnerability depends on what tne maximum field across some 4 weakest link in the insulation inside of it is, and I don't 5 see how you could measure that from the outside.
i i 6 MR. ROURK: No, you couldn't. Sometimes they have
- 7 taps and you can measure it in sections of the transformer.
! 8 But from a practical standpoint, if the transformer fails,
- 9 most of the surge energy has dissipated at that point.
4
- 10 You'd just have a ground fault and then the unit-
~
11 is tripped off by protective relay.
12 MR. WYLIE: 1What you're really speaking of in 4
13 capacitive grading-of the transformer winding is the actual
- 14 design of the transformer and construction of it at the '
lf~)h
\~
15 factor.
4 16 MR. ROURK: That's correct.
17 MR. WYLIE: And how they've taken into the design
- 18 the capacity grading across each layer of the winding, not 19 that this is measured later.
20 MR. LEWIS: I misunderstood.
21 MR. WYLIE: This is the inherent design of the 22 transformer.
23 MR. LEWIS: Okay. I know how to do it.
24- MR. MICHELSON: Doesn't that depend on what your 25 -specs are-in buying that?-
ANN RlLEY & ASSOCIATES, Ltd.
Q* Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 -
-~ .-~ -
j 281 2
j f-s - 1 MR. WYLIE: The basic insulation level, as he j- 2 mentioned, sets it. In the manufactured designs for the 3 wave fronts that the transformer must be able to withstand,
! 4 and he designs that into the transformer by spacing of the l 5 windings and any semiconducting material that he wants to j 6 put in there in order to accomplish that.
-7 But over and above that, then you protect the 8 transformer with the lightning arresters, i
I 9 MR. RoURK: Yes. We had several comments
- 10 regarding everyday experience with surges taking out
{ 11 computers or residential or industrial equipment and it's
, 12 just noted that most residential-and industrial service 13 transformers aren't designed in this way or to minimize the l
14 surge, the transmission of surge voltages.
{)
15 They don't have the other protective feature, such 16 as a grounding grid or distribution lines don't have shield 17 wires. It's typical practice only.to shield transmission 18 level voltages.
19 MR. CARROLL: My impression is that there is a 20 tremendous variation between individual utilities and even a
j 21 on a given utility system you find soil conditions.that make
- 22 it difficult to get or to put in an effective grounding
, 23 grid.
24 Does the Staff ever look at those issues when an 25 applicant comes in for a plant or when an applicant comes in ANN RILEY & ASSOCIATES, Ltd.
GA ~ Coud Reponers
. 1612 K. Street, N.W., Suite 300
! ' Washington, D. C. 20006 i- (202)-293-3950 a
, w --we a
282 s 1 for a backfit of a digital system like Eagle 21?
\' 2 MR. ROSA: To date we have not, sir.
3 MR. CARROLL: Okay.
4 MR. CATTON: Are you going to in the future?
5 MR. ROSA: We have a request, a user's need 6 request for development of a Reg Guide that addresses 7 lightening protection including grounding.
8 It's sort of being held up pe:4 ding the completion 9 of this meeting actually. It was prepared a few days ago.
10 MR. CATTON: Thank you.
11 (Slids.)
12 MR. RUARK: The use of a Reg Guide for existing 13 plants would require backfit analysis and part of the basis 14 for the Staff's position on this petition was events, review 15 of operating eventa -- 152 LERs were identified containireg 16 177 events.
17 Ten of these were also considered and studied in 18 the Accident Sequence Precursors Program. The conditional 19 core damage frequencies ranged from 90 to the minus 4 to 70 20 to the minus 8, but the events which had highar conditional 21 core damage frequencies had aggravating circumstateces such 22 as moisture in the instrument air lines or some equipment 23 that was unavailable -- you know, features that were ~ duo 24 to lightening.
25 I guess the Staff's position is that up till now fl ANN RILEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950
283 1 the only thing we have seen lightening, the only effect from 2 lightening has been to cause the reactor to trip and not to 3 damage any additional equipment.
4 MR. SHEWMON: Has there ever been a station 5 blackout from lightening?
6 MR. RUARK: Yes, there has.
7 MR. SHEWMON: Why isn't that in the LER?
8 MR. RUARK: Yes, station blackout. Well, I have 9 figures on that later.
10 MR. MINNERS: No, no, no. Do you mean loss of 11 power? Not station blackout?
12 MR. ROSA: He said the only thing that was 13 there --
14 MR. SHEWMON: -- were trips.
(
15 MR. RUARK: I have to correct myself on that.
16 That was the most prevalent thing. There have been total 17 losses of offsite power because of lightening.
i 18 [ Slide.)
i l 19 MR. RUARK: Besides loss of offsite power, though,.
t 20 there were also events where fires were caused such as one l
o l 21 of the oil circuit breakers blew up and spread burning oil 22 everywhere and there's been recurring cases of damage to 23 sensors, process sensors, pressure indicators, temperature 24 indicators, but the LERs indicated though.that the
- 25 lightening protection was in' place and functional.
ANN RlLEY & ASSOCIATES, Ltd.
l
-C Court Reporters 1612 ~ K.- Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
_ _ ~
. - . . . . - . -. - - -,- - -- _ - - _ _ - ~ - - - . .- - . .- -
4 i 284 ,
1 There were some LERs that stated that improvements '
2 were being made to the lightening protection but you would 3 have events where the same effects occurred and the 4 structural and power line protection was in order.
] ,
j 5 No events resulted in misoperation or failure of a l 6 safety-related system
! 7 (Slide.)
8 MR. RUARK: The IPEEE review, the guidance in i
9 NUREG-1407 states that the Staff considers the primary i
j 10 effect of lightening to be loss of offsite power but if a
(
11 plant has experienced site-specific events, specifically l
12 ones that are recurring, the Staff would expect them to i 13 review those events.
14 MR. WYLIE: Let me back up a minute, Dick.
15 Your statement "No event resulted in misoperation 16 or failures of safety-related systems," you don't mean to 17 imply that some were not actuated because --
l 18 MR. RUARK: Oh, no. No, they did actuate but they-i 19 actuated correctly.
! 20 MR. WYLIE: Yes.
21 MR. RUARK: Such as when the pressure indicators l
22 blew out --
23 MR. WYLIE: Well., I am talking about from the 24 lightening strokes?
25 MR. RUARK: That too.. They would actuate because l _
Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporteis
- 1612. K. Street, N.W., Suite 300 Washington, D. C. 20006 i- (202) 293-3950 i
285 l
1 of the effect of the lightening but they operated as t o2 designed.
l 3 MR. WYLIE: Or -- well --
i 4 MR. RUARK: You mean they failed safe.
5 MR. WYLIE: There wara some falso actuations, l 6 aren't there, in the LERs where they actuated falsely 7 because of lightening?
8 MR. RUARK: Well, maybe due to loss of offsite 1 9 power, maybe due to a reactor trip --
i 10 MR. WYLIE: I'm talking about the magnetic effects
- 11 or the electrostatic effects of the lightening stroke.
12 MR. RUARK: Well, the only effect that caused them
]
f 13 to operate was because the sensors were damaged. I think a t
14 couple events had instrument amplifiers damaged but when 15 they lost these signals, these systems themselves actuated 16 in response to the loss of signal.
17 MR. KERR Well, now, there certainly are sensors 18 that deliver a signal that's an electrical signal and j 19 there's no particular reason why lightening couldn't 20 generate an electrical signal that would look like neutron l
j 21 flux or whatever, so I can't believe that that is 22 impossible.
23 MR. RUARK:- It is possible but what-has happened 24- _is it looks like they had._ grounding problems or-similar 25 problems that just took out the pressure transmitters and ANN RILEY & ASSOCIATES, Ltd.
O- _
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
1 286 1
l 1 completely damaged them and didn't generate any spurious ;
i 2 signals.
3 The lightening itself is a very fast transient and -
-I ,
j 4 most systems don't respond that fast so that if it did --
5 there have been recurring problems with APRM, neutron flux
+
1
- 6 monitors and things like that.
7 MR. SHEWMON:
8 MR. KERR: Lightening is a fast transient but a i 9 fast transient has low frequency components and the fact i
10 that it is a transient is sort of irrelevant.
i 11 MR. RUARK: This is true. There have never been l 12 any system misoperations though. The high pressure 13 injection system didn't spuriously actuate. It actuated 14 because pressure sensors were taking out and never 15 misoperated because -- and didn't suffer concurrent loss of 16 sensor failure.
5 17 (Slide.]
18 MR. ROURK: The IPEEE review team has been 19 notified of the results of our findings. The plants that 20 seem to have recurrent failures have been brought to their 21 attention. It's expected that they will consider this when
! 22 they're reviewing _the plant submittals.
23 (Slide.)
4 24 MR. ROURK: The NRC staff is also pursuing other 25 things independent of the petition. It's developing a Reg i
i
(]
ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 4 1612 K. Street, N.W.,- Suite 300 Washington, D. - C. 20006
.(202) 293 3950--
1 287 1 Guide on the EMI protection which will endorse other IEEE O 2 1050, or military standard, or something similar. I think 3 the staff was in here briefing the ACRS on that recently.
4 We're also, as Faust mentioned, reconsidering the 5 development of a Reg Guide on structural and transmission 6 protection and augmenting guidance to the staff.
7 MR. ROSA: As of right now, our draft user's memo 8 recommends endorsement of NPFA-78, and also ANSI IEEE 665, 9 which on the generator station grounding. We have in the 10 back of our minds also possible endorsement of ANSI C62 11 standards on the application of surge arrestors.
l
- 12 But beyond that, I don't think we should do much
! 13 in the Reg Guide beyond adding a limited amount of
- 14 additional guidance to specifically address application of l 15 these standards to nuclear power plants.
I
- 16 MR. CARROLL
- Are those standards universally used
, 17 by the utilities, or their architect engineers, Faust?
I 18 MR. ROSA:- To my knowledge, they are.
19 MR. WYLIE: With the exception of the one on 20 generator station grounding has only been out about, what, 21 four years?
22 MR. ROSA: I think it was an '87 standard.
[ 23 MR. WYLIE: So they really haven't designed any
- 24 plants since then?
25 MR. ROSA: There were other, like the Green book, s
ANN RlLEY & ASSOCIATES, Ltd.
IQ
( Court Reporters l
1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006
-(202) 293-3950
i i
4 j 288 4 .
1 IEEE 142, that includes some design information for 1
2 grounding systems. As I said before, there are many papers 3 published by the IEEE that address this area.
{
4 MR. LINDBLAD: While we're calling this protection 5 from lightoning, it really seems to me that we're really
]
l 6 talking about design of the grounding system. The grounding 2
7 system turns out to be the key.
8 MR. WYLIE: Grounding and shielding.
9 MR. LINDBLAD: All right.
10 MR. CATToN: There has to be some incentive to do j 11 it.
12 MR. KERR: one more vote for shielded cables.
13 MR. CATToN: I know that UCLA chose to do their 14 engineering building based on an economic analysis.
15 MR. LINDBLAD: I would say that most of the long-
! 16 term standards that have been around for many years, the 17 NFPA standards, basically were concerned with protection of 18 life and elimination of fire in structures. Fire was the 19 big issue for lightening.
l 20 With a power plant, we're really talking today
- 21 about' loss of equipment by raising the potential of 22 individual equipment grounding levels and. separation of 23 grounding between lightoning, mitigation, and the plant 24 equipment grounding.
25 That, I think, has in the past been done'by -
t Q ANN RlLEY_ & _ ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202); 293 3950 d
f .. . . - _ , , , m.. , , , . , y _ . , - . .,,4
289 1 electrical engineers using company standards or AE i 2 standards, more so than relying on NPFA. But the IEEE 3 standard that Faust is referring to, brought that all into a 4 national type standard in whatever it was, '87 or '88.
5 But I believe that many of the AEs have been using 6 standards like that for many years. But we were calling 7 them grounding design rather than calling then lightening 8 protection.
9 im ROSA: We obtained from the Duront company 10 their lightening protection standard back in the '70s which, 11 in effect, reproduces, well, in fact, all of NFPA 78, j 12 including the portion that addresses grounding.
13 Now, the basis for our staff position in reviewing 14 lightening protection in the past really has been that there
! 15 is no more economic incentive that can be applied to 16 utilities than that that would accrue from adequate l 17 lightening protection.
18 Also, they were the people that have been involved 19 in designing and operating plants that require lightening
{
I-20 protection from the inception of-the electric power <
21 generation.
22 MR. ROURK: One of the problems we've seen have l
23 also been due to local effe:ts that were not relatedto the
- 24 lightening protection. In other words, the utilities had 25 adequate structural and power line protection which was the
( ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300
- Washington,- D. C. 20006 (202) 293-3950 -
L
i 290 1 only protection of the standards at the time.
d 2 But because of some local cffect of lightening, 1
3 perhaps, ground potential rod -- spurious signals are 4 induced into the ables. So, there's an additional problem 5 there which I don't think the industry recognized at the j 6 time probably because the electromechanical components that ,
7 they used responded too slow to be affected by these 8 transients.
9 MR. WYLIE: I would have to go back, now, and look
- 10 at the IEEE standard. I haven't looked at it in a few
! 11 weeks. I don't believe they address that GPR, do they?
12 MR. ROURK: Which IEEE standard?
13 MR. WYLIE: Well, it's really, I guess, the ANSI 14 665, IEEE, whatever it is. The one on generator station
! 15 grounding. Lightening protection.
l 16 MR. ROSA: I guess I didn't catch the question.
l 17 MR. WYLIE: I said I don't believe they address 18 the ground potential rise problem.
19 MR. ROSA: I believe that standard does address 20 it. I've got a copy of the standard right now.
21 MR. WYLIE: I do, too.
- 22 MR. ROSA
- "ut that's quite a large number of 23 pages. I believe it does address it, yes.
j 24 MR. ROURK: I believe 10.50 addresses the ground 25 potential rise also.
I O ANN RILEY & ASSOCIATES, Ltd.
Court Reporters
- 1612 K.- Street, N.W., Suite 300 l
Washington, D, C. 20006
, (202) 293-3950 l
291
.l 1 MR. MICHELSON: How much ground potential rise do 2 you need if you have a solid state circuit control system 1
3 instead of the old relay type system?
.! 4 MR. ROURK: It would depend on the capacitive 1
j 5 coupling of the cable to the ground.
1 j 6 MR. MICHELSON: I don't have no idea how many 1
4 7 volts it would rise. It is my understanding that a fraction f 8 of a volt would be all you need for solid state devices, i
9 where as in relays there is a whole lot bigger change before 10 you have enough current to worry about. Also, these are
, 11 very fast response systems now as opposed to relays, which 12 are millisecond response, multi-millisecond response. These j 13 are microsecond responses.
14 MR. ROSA: I couldn't answer that question, but i 15 the whole design ties together. Overhead ground wire 16 shielding reduces the magnitude of direct strokes to the
- 17 power lines, and --
18 MR. MICHELSON: Let me clarify. I am not so
[ 19 concerned about lightning strokes as I am about electrical 20 faults within the building that probably get around before ,
4 21 they get adequately grounded, break faults, that sort of.
4 22 thing, which we have seen some of in LERs, and they do tend 23 to cause things to circulate around a little bit, and the i
24 ground references to change a little bit, and that is all 25 fixed on this new equipment.
O O
ANN RlLEY & ASSOCIATES, Ltd.
Couit Reporters 1612 K. Street, N.W., Suite 300 Washington, D. - C. 20006 (202) 293-3950
292 1 MR. ROSA: For that type of protection, I think 2 you would have rely on the design of the electronic 3 equipment itself, plus whatever surge arresting features 4 that are incorporated.
5 MR. MICHELSON: But the ground reference is the 6 worry, you can't design for that very well here. You are 7 assuming a certain ground referenced condition, and if that 8 ground reference changes because a link has been lost around 9 the building somewhere --
10 MR. ROSA: That's true.
11 MR. MICHELSON: -- it will rise a volt or so, and 12 you can't design for that with the equipment itself. You 13 have to make sure it doesn't happen.
14 MR. ROSA: You can design to minimize that.
15 MR. MICHELSON: You can design to minimize it, but
. 16 when it becomes gross, which it does in the case of linkage i
17 burnouts, and that sort of thing, I don't know what you do.
I 18 You just have to adequately design.
, 19- You are focusing, though, on lightning, I gather, 20 and not really on internally generated electromagnetic
! 21 interference?
l 22 MR. ROSA: That is correct. This other Reg Guide 23 that is underway, and that will address that.
24 MR. MICHELSON: So this is the wrong place to 25 worry about it. Thank-you.
ANN RlLEY & ASSOCIATES, Ltd.
, i] Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
j 293 g 1 MR. ROURK: Well, yes and no. Part of the 2 concerns are being resolved by this additional action that 3 is being done on EMI. It is being addressed. If it wasn't,
) 4 we would probably recommend that as some sort of additional 5 action, i
- 6 MR. WYLIE
- Certainly the LERs, they are LERs that 7 show the influence of lightning strokes that have gotten 8 into the control room, right?
9 MR. ROSA: Yes. There has been, I believe, one or 10 two instances of that having occurred.
11 MR. WYLIE: They have gotten in there?
12 MR. ROSA: Right.
13 MR. WYLIE: The one that bothers me more than the 14 rest is the lightning stroke in the switchyard or at the 15 meteorological tower, and then those cables come back into 16 the plant, into the cable spread room, and even into the 17 control room, and you get this GPR rise in the switchyard, 18 and that potential is transmitted back into the control 19 room.
20 MR. ROURK: That's correct.
21 MR. WYLIE: I know from experience, though not 22 nuclear plants, that on fossil and hydro plants that that 23 has happened, and even to the extent that there were 24 flashovers on the boards.
25 Now I think the mechanisms and means are available Q ANN RlLEY & ASSOCIATES, Ltd.
Coud _ Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 !
I l
l 294 l l
r~s 1 to prevent that from happening, and they are very simple, l
.( )
2 and it doesn't cost a lot of money to do it. You just do a 3 good job of snielding, and a good job of grounding between 4 the switchyards and plants, and you can prevent that from 5 happening.
6 I think that is the missing link between the plant 7 and the switchyard. The switchyards are covered, and the 8 plants are covered, but it is that missing lini between the 9 two, and the meteorological towers.
10 Today, certainly, with fiber optics, what we are 11 able to do is, we can completely divorce that influence.
12 MR. LINDBLAD: For instrumentation systems.
13 MR. WYLIE: And control, yes.
) 14 MR. LINDBLAD: It has been my experience that
(
15 grounding is a very critical denign item on electric power 16 plants, generating plants, whether they are nuclear or not.
17 I guess, it is underrepresented in the NRC reviews that I 18 have seen, but I think it gets done irrespective of the 19 amount of NRC reviews there have been. Probably it comes 20 back to the issue of safety risk, what is the risk of 21 lightning or weaknesses in the grounding design, and if that 22 risk is high enough, it probably should be greater 23 represented in the NRC review, but while there is lots of 24 impact on the utility, I am not so sure there is that much 25 great impact on the public safety.
( i ANN RILEY a ASSOCIATES, Ltd.
t
'~
Coud Repoders 1612 K. Street, N.W., Suite 300 Wash lngton, D. C. 20006 (202) 293-3950
1 i
i 295
- 1 MR. WYLIE Well, I tend to agree with you on J
- 2 that.
1 !
- 3 [ Slide.)
! 4 MR. ROURK: That is basically the staff's 5 conclusion too.
1 6 MR. LINDBLAD: Could I ask, is there any reference 7 in Appendix R, fire protection, to lightning protection? ,
i j 8 MR. ROURK: I don't know off the top of my head.
l 9 MR. ROSA: I don't believe that there is.
10 MR. LINDBLAD: That's surprising. It covers 11 almost everything else.
! 12 (Laughter.)
13 MR. COE: One last question, if I may. You noted 14 that core damage frequency ranged on the ten events that you 15 selected from 9E to the minus 4 to 7E to the minus 8.
I 16 That's a very, very large range. What were the -- I'd just 17 be interested to know what the events were at the high end 18 of that range. What kind of events? Did they include the i
19 loss of off-site power events, which the staff has deemed to 20 be the biggest impact, or were there others?
l 21 MR. ROURK: Definitely not loss -- well, some of
! 22 them were loss of off-site power, not reactor trip. The 23 ones that resulted in reactor trip were low conditional core i
24 damage probability. Loss of off-site power -- the event 25' with the highest conditional core damage frequency from loss e
i
! h ANN RiLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300-Washington, D. C. 20006 (202) 293 3950 t
296 1 of off-site power was Yankee Rowe, which I think was around O 2 -- was the high 10 to the minus 4's, and it also had a fuse 3 blow in the invertor, which I'm not sure if that was due to 4 the lightning itself or the ground potential rise just 5 affected the -- the voltage was depressed from the fault in 6 the switch yard and the invertor might have been trying to 7 feed this depressed voltage and had an overcurrent.
8 But that was relatively high, but it was slightly 9 higher than just the loss of off-site power scenario for 10 that station.
11 MR. WILKINS: Do you remember off-hand what the 12 next highest one was? The reason I ask is because I think 13 Yankee Rowe is no longer a problem.
() 14 15 MR. RoURK:
MR. WYLIE; No, it's not, thank goodness.
I think that most plants, because of 16 experiences with lightning, have gone back and beefed up 17 their lightning protection and grounding systems. There 18 were several plants that had a predominant number of 19 problems, and I believe they have taken measdres to correct 20 those problems, and I think the new plants, as you say, are 21 being designed to the IEEE -- not the IEEE -- the EPRI 22 guidelines, which reference the IEEE standards and others, 23 as far as lightning protection and grounding and so forth 24 are concerned.
25 So I think that the staff's conclusion, from my C ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950
297 1 standpoint, for this particular petition is correct. The O 2 question is, I guess, what the committee feels like.
3 MR. KRESS: Don't you normally start getting 4 concerned when you see nine times ten to the minus four 5 conditional CDF in the ASP program? And initiating for 6 lightning, that's getting close to one per year for some 7 plants.
8 MR. KERR: That_was at Yankee Rowe, which is no 9 longer operating.
10 MR. WILKINS: That's why I asked for the next one.
11 MR. KRESS: Yes. The next one may be close to 12 that, though.
13 MR. WILKINS: or it may be an order of magnitude 14 away. I don't know.
15 MR. LINDBLAD: It seems to me that the loss of 16 off-site power is influenced also with the transmission line 17 design and the lightning protection of substations outside 18 the plant property just in terms of extensive lightning 19 storms over broad areas, and we remember that the New 20 England blackout, the North American blackout was 21 precipitated by lightning as well. So those things that are 22 beyond the plant description also influence loss of off-23 site power.
24 I would have to say without any great amount of 25 study that I think that the grounding design and lightning Q ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
- =. - . _. -. - -.- - - - - . . - _ . _ _ - ._ ._ - __ . _
i
! }
l
' 298 I 1 protection of power plants probably is paid greater amount
- 2 of attention by designers in building the plant, designing
. 3 and building the plant, than some issues that the NRC l 4 Electrical Branch routinely pays attention to, like 3
- 5 environmental qualification of electrical equipment. I 6 think that the staff pays a great amount of attention to 7 that beyond. And so if we're balancing, have we given as l
! 8 much lightning protection oversight as we have EQ oversight?
i l 9 No, I don't think the NRC does. But-in terms of the risk l 10 significance and safety significance, I think that the 11 conclusion the staff makes on'this line is appropriate and I 12 agree with Charlie.
l 13 MR. CARROLL: WEll, do we have a concern with 1 14 regard to the new generation plants, though, that are going 15 to end up extensively using digital control and protection l 16 systems? Should the staff be doing more because of that l 17 consideration?
I
{ 18 MR. LINDBLAD: I think we do, and I think you'll 19 see more and more optical coupling. So when we talk about 20 digital systems, it's not just semiconductors, but it's
]
21 optical coupling on transmission loops that are going to_be
, 22 important. The utilities have been seeing that since much
. 23 of their switch yard and transmission relaying is done with 24 semiconductors now, and they have solved some.of the early 25 problems they had with digital relays with optical coupling.
1 ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters-1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
i 299 1 MR. KRESS: The optical coupling requires the 2 connector that converts from electrical to optical and then 3 back again on the other side, and that device is susceptible 4 to lightning strikes, I think.
5 MR. WYLIE The transmission from meteorological 6 towers and switch yards would not be transmitted.
7 MR. KRESS: But you'd lose the coupling device.
8 You would attenuate the signals.
9 MR. CATTON: That would depend on the design of 10 your optical coupling.
11 MR. WYLIE: I think the real threat is getting the 12 signals into the control room.
13 MR. LEWIS: There has been a major change in the
, 14 whole world over the last seven or eight years in terms of 15 the efficiency of optical coupling. There has been a 16 genuine technological breakthrough and the whole world is 17 switching to optical communication, but not to transmit 18 power, but to transmit signals.
19 MR. LINDBLAD: I would also point out that I think 20 the one area that I would -- I have worried about in the 21 past but haven't really seen a problem arising yet is an 22 electrical penetration in the containment being hit with a 23 large power surge from a lightning stroke and whether the 24 containment penetration will survive that. But apparently 25 that has never happened because of the strong grounding f(() ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
4 l 300 l 1 system in the plant, j 2 MR. WYLIE: Yes. The building itself, though, i
3 acts as a shield to that.
! 4 MR. LINDBIAD: Yes. Yes.
f
. 5 MR. WYLIE: I can't speak about all penetrations, j 6 but those that do have a big metal box around them, too.
7 MR. CARROLL: Well, what I was raising was the
]
! 8 question of whether we want to say something in a letter .
1 9 about the staff more aggressively pursuing this issue for l 10 the future plants.
i 11 MR. WYLIE: 11aybe that ought to be a separate j 12 letter. This particular one concerns a petition on existing i
- 13 plants, and I would suggest maybe a separate letter after i
14 review, further review may be by the subcommittee or what _
j 15 have you, on what the staff plans to do for the future i
16 plants might be appropriate.
4 17 MR.-WILKINS: Is that a sort of Son of 016 --
{
j 18 MR. WYLIE: No. _It's just mainly to -- oh, you
! 19 mean a policy issue?
20 MR. WILKINS: Yes. Maybe it's the grandson. I
- 21 don ' t - -
MR. WYLIE: But I would-suggest we do a separate i
1 23 letter not tied into this one.
! 24 MR -ROSA: _If I might interject, however-it's 25 done, I think you ought to keep in mind that we still have ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 4' Washington, D. C. 20006
_(202) 293 3950
1 l
301 1 to go through CRGR on this proposed reg guide, and any 2 support we could get for getting it through there would be f
- 3 appropriate.
4 MR. LINDBLAD: I think in a previous presentation 5 this morning, we have seen PRAs done that show a q 6 predominance of dependence on loss of off-site power, and if 7 future plants have all their problems solved except for loss 8 of off-site powar, probably lightning protection, both on l 9 site and off site, will become of greater and greater j
j 10 interest as it goes on. So if the levels of risk associated
, 11 with loss of off-site power becomes significant to people, 12 that will be an issue, I think.
13 MR. SHEWMON: There are such things as safety 14 goals and how safe is safe enough, but I'm sure you'll keep
, 15 that in mind.
4 j 16 MR. LINDBLAD: Yes.
17 MR. WYLIE: I will write basically favorable j 18 letter.
i 19 MR. SHEWMON: Fine. Gentlemen, that finishes l 20 that, then. Thank you.
21 (Whereupon, at 11:17 a.m., the hearing recessed 22 for lunch, to reconvene at 2:00 p.m.)
-23 24 25 O ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, - D. C. 20006 (202) 293-3950
. - - - ~ . - - . -, -. . - . ._
- 302 1 AFTERHOON SESSION
,O l 2 (2:02 p.m.)
I f 3 MR. SHEWMoN: The next item on the Agenda is 3
4 Generic Item 120, online Testability of Protection Systems, 5 Charlie Wylie responsible.
l f 6 MR. WYLIE Okay. I'm going to be very brief.
7 This as the title suggests is the Generic Issue 120 Online
]
j 0 Testability of the Protection Systems and T.Y. Chang of the
~
9 Staff is going to make this presentation.
10 (Slide.)
. 11 (Pause.)
12 MR. CHANG My name is T.Y. Chang. I am the Task i
13 Manager with Engineering Issues Branch, Division of Safety 14 Issue Resolution.
~
15 The issue that we are going to talk about today is 16 Generic Issue 120, online testability of protection systems.
4 17 By online we mean at power and protection system is a system .,
18 that provides functions that will enable the plant to 19 prevent unsafe operation conditions which may lead into I
20 accident conditions or in case an accident does happen, then
! 21 you provide the capability of the plant to mitigate the i 22 consequences.
23 Basically there are two systems involved here.
24 One is the reactor protection system. The other one is the 25 engineored safety features.
ANN RlLEY & ASSOCIATES, Ltd.
>]
Court Reporters 1612 K. Street, N.W., Suite 300 Washington,- D. C. 20006 (202) 293-3950
303 1 That's on the second slide.
O 2 (Slide.)
3 MR. CHANG: General design criteria 21 of 10 CFR 4 50, Appendix A requires that each plant should have the 5 capability to test the protection system functions online, 6 meaning at power.
7 The purpose of this is that if you do the periodic .
8 online testing often enough then you'll be able to discover 9 the existing faults in the system, thereby to increase the 10 reliability of the system.
11 MR. DAVIS: Excuse me. When these systems are
)
12 tested, are they taken out of service for the test or is 13 there any requirement --
{
14 MR. CHANG: That's what online means.
15 MR. DAVIS: oh, I thought you said online means at
- 16 power.
17 MR. CHANG: At power,_yes, so they should have the 18 capability to test each channel independently so when you 19 test a particular channel, that channel is taken out from 20 the service.
21 MR. DAVIS: Thank you.
22 MR. CARROLL: Or put in a trip' condition.
23 MR. CHANG: Pardon?
24 MR.-CARROLL: Or put in a trip condition, half-25 scram or something like that. ,
O ANN RlLEY- & ASSOCIATES, Ltd.
Court Reporters i
1612 K. Street, N W., Suite 300 Washington, D. - C. 20006 (202) 293 3950.
lj
304 e 1 MR. DAVIS: Yes, if you have two out of three, 2 that gives you one already.
3 MR. CHANG: Yes, I think so. Originally two out -
s 4 of four systems and you could switch to two out of three, so 5 that particular channel can be tested.
6 MR. LEWIS: Does GDC-21 define the word " test" or 7 is that open to the beholder?
8 MR. CRANG: It does mention test.
9 MR. LEWIS: But does it define what is meant by 10 test? Test can mean a lot of different things.
11 MR. CRANG: All it says is it should have the 12 capability to be tested periodically at power but it doesn't 13 say what does test mean.
14 MR. LEWIS:
( Okay.
15 MR. CHANG: It's a very general requirement.
16 MR. LEWIS: No, I understand. It's just that test 17 can mean many different things, you know, and this leaves it 18 open to a --
19 MR. CHANG: It does say functional tests, okay?
20 MR. LEWIS: I suppose that should help me.
21 MR. CRANG: That's the function of it.
22 (Laughter.)
23 MR. DAVIS: ;There is a long history of --
24 MR. CARROLL: Relax and be happy.
25 MR. WYLIE: I believe IPEEE-279 though does define Q ANN RILEY & ASSOCIATES, Ltd.-
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. - C. 20006 (202) 293-3950
305 1 what is required.
O 2 MR. CHANG: Right. I am going to be looking at 3 other documents there. GDC-21 is a general blanket 4 requirement and there are more specific guidelines like Reg 5 Guide 1.118, which basically endorses IPEEE standards; 338-6 1977 which gives the detailed test criteria for all of the 7 safety-related system including the protection system.
8 MR. LEWIS: I understand that. I am just being a 9 little bit legalistic. A Reg Guide isn't a requirement, of 10 course, and so the GDC is a requirement and therefore it is 11 open for somebody to take a minimalist approach to this GDC 12 requirement and argue that going beyond it is going beyond 4 13 the regulations. I am just trying to think what the -- if I 14 were a sneaky person, what could I get away with here, and 15 the answer seems to be "a lot."
16 MR. CHANG: Well, okay. Those being the 17 requirements, the Staff recognized that there are cases that l 18 the portions of the, some of the so-called actuated 19 equipment, that's the-last equipment that is going to 20 perform the function of the generic safety feature, like 21 safety injection pumps or isolation valves.
22 There are cases that those actuated e'quipment l 23 cannot be tested online without introducing. adverse effects 24 to the safe operability of the_ plant, so Safety Guide 22 25 recognized that and says that in those cases if you can ANN RlLEY & ASSOCIATES, Ltd.
'{
I-Court Reporters 1612 K. Street, N.W., Suite 300
- Washington, D. C. 20006
- j. (202)-293-3950
306 1 prove the reactor protection system itself, the reliability 2 is acceptable, and also if this actuated equipment can 3 routinely be testing during the outage, then on a case by 4 case basis it should be acceptable.
5 Of course this has to be reviewed by the Staff.
6 MR. CARROLL: That does raise an interesting 7 philosophical point. We have a regulation that requires, 8 without exception, testability and then we write a Reg Guide 9 that says you don't have to follow the law.
10 MR. LEWIS: But the Reg Guide isn't enforceable.
11 MR. CARROLL: The Reg Guide isn't enforceable, but 12 as a practical matter it is being used, or has historically 13 been used by the industry and, in effect, or has the effect 14 of allowing people to break the law.
15 MR. LEWIS: I understand that.- It allows them to 16 break the law.
17 MR. CARROLL: The GDC is, indeed, the law. It is 18- Appendix A to Part 50.
19 MR. LEWIS: It allows them to have a plant which 20 is not testable, which simply means the GDC was written 21 carelessly.
22 MR CARROLL: It isn't today. It is very clear.
23 MR. LEWIS: That sometimes happens. But the sense 24 of my persistent-questioning was to try to-unscramble what 25 is actually required here from what is the custom around O ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., - Suite 300 Washington, D. C. 20006 (202) 293-3950
307 1 here, and what is required seems to be the GDC, and when, 2 for example, you mention acceptably reliable, I think you t 3 used those terms, it begs the question, "What determines 4 acceptability?"
5 MR. DAVIS: And how do you demonstrate it?
6 MR. LEWIS: How do you demonstrate it, and how do 7 you even define it before you try to demonstrate it, and if 8 you did define it, then it would probably not be 9 demonstrable.
10 MR. SHEWMoN: I think if we let him talk some 11 more, he might answer.some of these questions.
12 MR. LEWIS: I am just giving him'the tone of my 13 concerns. He answered the question, which is --
14 (Slide.)
15 MR. CHANG: The scope of GSI-120, in reality, is 16 no more than what the title of the issue implies. It really 17 is about slave relays in the ESFAS system. There are two 18 cases.-
19 one is if the slave relays are hard wired to the 20 actuated equipment, then whenever you try to test the slave 21 relay, the equipment down the stream is going to actuate.
22 So if, for that equipment, it is not desireable to be tested 23 on-line, then those slave relays cannot be tested either.
24 That is one case.
25 The other one is, in certain cases it is a pretty
'C ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950 o
J
' 308 1 complicated coordination between the test technician and the O 2 control room operator to do an on-line test, other than our ,
3 considerations, things like human factors, things like that i
i 4 which could come in which might trigger inadvertent safety 5 related trip actuation or reactor trip.
t 6 So in both cases, even though the purpose of the-7 test is trying to assure the reliability of the system, but, 8 in this case, the risk of the test to trip the reactor and 9 actuate the equipment may outweigh the benefits you can 10 reach by doing the periodic on-line testing.
11 MR. LEWIS: You know, I shouldn't pick on you for 12 this one, but let me anyway. You used the term " assure-the 13 reliability of the system." Are you going to argue that one 14 successful test assures the reliability of the system?
15 MR. CHANG: This is a periodic test.
16 MR. LEWIS: Periodic meaning how long?
17 MR. CHANG: Periodic for the case of slave relays, 18 generally it is a quarterly. So you have it four times a 19 year.
20 MR. LEWIS: Four times a year?
21 MR. CHANG: Right.
22 MR. LEWIS: But after each test, it is considered 23 sufficiently reliable to go on to the next test?
24 MR. CHANG: Yes,_that is the basic assumption.
25 MR. LEWIS: I just wanted to --
l-Q ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 -
Washington, D. C. 20006 (202) 293 3950 p- e-- y -
yI' r*w ,v <,m e n y p-- w se e _my p w * --g,eq
309 1 MR. CHANG: That constitutes an acceptable. We U 2 have confidence that if you do that test, then --
3 MR. LEWIS: I wish you hadn't used the word 4 " confidence."
5 MR. CHANG: A high probability.
6 MR. LEWIS: I wish you hadn't used the word "high 7 probability."
8 I will stop this line, but the only point I am 9 trying to make by pestering you is that nothing is defined 10 in this business apparently.
11 MR. CARROLL: When you talked about quarterly 12 tests, that ce::tainly isn't true of the first bullet 13 situation where it is hard wired?
/ 14 MR. CHANG: That's right.
N))
15 MR. CARROLL: Those are --
16 MR. CHANG: Except those hard wired slave relays, 17 yes.
18 MR. CARROLL: But, isn't it true that the 19 designers of these systems had the opportunity to put in 20 features that would allow periodic testing?
21 MR. CHANG: Yes.
22 MR. CARROLL: I can remember when a utility I used 23 to work for bought a plant from Westinghouse, I had grown up 24 on GE plants, and I was appalled at the design of the solid 25 state protection system because you couldn't do this, and
()
ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
310 f- 1 they said, "No, we will convince the NRC it is okay." I 2 said, "Well, you are never going to convince zo it i- okay, 3 because I think you could have made some very simplu design 4 changes that would have allowed you to test these relays, 5 and I would have felt a lot better.
6 MR. CHANG: Right. In reality, if someone really 7 wants to have the on-line testability of all those systems, 8 it can be designed into the plant, but I think for the older s
9 ones, that is those plants may be overlooked.
10 MR. CARROLL: How about for the new generation of 11 plants, are you going to allow them to do the same kind of 12 things?
13 MR. CHANG: Let me go on, then, and I think you
(~') 14 can see a better picture of it current standards.
L) 15 MR. CARROLL: All right. I will be quiet.
16 (Slide.)
17 MR. CHANG: When we started working on this issue 18 last year, we were not too sure what kind of thing we were 19 getting into, what the concerns really are, and how 20 pervasive are the concerns. In other words, how many plants 21 were to be impacted by these concerns.
22 :2o the first step we did is trying to make a 23 survey of p34u;6 rith different vintages and designs. We 24 looked at 17 dicterent units that covers a wide spectrum of 25 vintages and designs.
(3
\' ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 1 _ - - - - _ - - - -
.. . .. . . - - - -_ . ~ . . .-
]
311 1 The information we gathered included updated 2 FSARs, technical specifications, SERs, and in certain cases 3 the contractor talks to the plant personnel as well to 4 gather operations information.
5 MR. CARROLL: Let's see. I do not learn much from
- 6 your second grouping of seven post-SEP plants. What's the 1 7 breakdown between BWRs and BWRs there?
8 MR. CHANG: The SEP plants, after their review, l 9 they concluded it's okay. So, we don't include the SEP
< 10 plants. That's why we have the SEP plants there.
11 MR. CARROLL: ilo, what I'm asking is of those f
- 12 seven plants that you did look at, how many were Bs and how
- 13 many were Ps?
14 MR. CHANG: Oh, okay. I can give you the
(
4
- 15 information. We have -- those seven plants are units.
- 16 Actually, it's four plants with seven units -- one BNW, one l
17 CE, one GE, and one Westinghouse. So it covers the whole
! ]; wide spectrum of different designs.
j a4 MR. CARROLL: Did that add up to seven?
- P MR. CHANG: Four plants; seven units.
21 MR. CARROLL:- Oh, oh, okay.
I 22 MR. CHANG: I have unitt here. That's what we j - 2 3. 7ean by plants here.
24 MR. CARROLL:- The one above was eight units?
]
25- MR. CRANG: . Right, that's correct.
-O ^"" aev * ^SSoc'ATES' d-Coud RepOders 1612 K. Street, N.W., Suite 300
. Washington, D. C. 20006 i (202) 293-3950
i 312
- 1- MR. CARROLL
- Okay. '
2 2 (S)ide.]
3 MR. CHANG: The INEL survey follows. For all the 4 units reviewed, exemptions were specified by the licensees'.
5 They list out by component member what really was not tested 6 on-line and what equipment was not tested on line. In each 7 case, the submittal was reviewed by the staff and concluded
- 8 that it's acceptable.
l 9 INEL recommends that for future plants, this 10 procedure should be followed, that is, be reviewed on a i
11 case-by-case basis when there's an exemption requested.
12 MR. LEWIS: This_ emphasizes Jay's point earlier l 13 that when institutionalizing a system by which-the staff-t-
14 can, on a case-by-case basis, according to its judgement, 15 grant exceptions from the law. If we write a letter, we 16 should_certainly point that out.
- 17 MR. CARROLL
- The other thing that's a little j 18 disappointing to me is that you haven't taken a position for f 19 the future plants, particularly with the kind of electronics 20 we have available today. .It just seems to me that it's very 21 easy to design systems that allow full-testing.
i l-j 22 MR. CHANG: Right. We certainly encourage them to
- 23 have that capability. Also, in the review of future plants, 24- like water reactors, we think that when we review the 25 requirements document, we did mention that-there is a l
l
!(3 ANN _ RlLEY & ASSOCIATES, Ltd.
Court Reporters
- 1612 K. Street, N.W., Suite 300
! Washington, D. C. 20006 l (202) 293-3950 L
313 1 possibility as far as we understand that more and more
'd 2 digital competence will be designed in the plant in the I&C 3 systems.
4 In place of slave relays, they may want to use 5 some microprocessors and with some fir.a1 actuating contacts.
6 So, in that case the concern may be .:nanged from the slave 7 relays to those components. They have to prove to us that, 8 first of all, they can be designed into on-line testing. If 9 not, then those things shall be reviewed on a case-by-case 10 bacis as well.
11 MR. LINDBLAD: Are you saying that sooner or later 12 you can have an auxiliary relay in digital technology?
13 Slave relay it what I would call an auxiliary relay; isn't
,dfl 14 it?
15 MR. CHANG: Yeah, or you can call it control relay 16 or whatever. As I understand that, they have a 17 microprocessor and a contact --
18 MR. MICHELSON: You would hve that at full power.
19 MR. CHANG: -- to put together to take the place 20 of --
21 MR. MICHELSON: It's just a solid state based
, 22 contactor.
23 MR. LINDBLAD: If we test the contactor at some 24 point, we will be turning on the power.
25 MR. MICHELSON: You've still got the problem. You IO V
ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 -K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
314 6
1, can't test the contactors because you would actuate the
- 2 device unless you put another contactor in and actuate one 3 at a time.
4 MR. CARROLL: That's what I'm saying.
l
! 5 MR. MICHELSON:' That gets kind of expensive on big i 6 power equipment.
i 7 MR. CARROLL: That's what I'm saying. It's 8 possible.
' 9 MR. MICHELSON: Oh, it's possible, yeah. That's 10 how we do the scram relays and scram breakers. That's why 11 you've got four breakers in there. You really need two.
i l- 12 MR. CARROLL: The question, I guess, is, should j 13 the staff take the position consistent with GDC 21 that i 14 that's'the way future plants ought to be built?
15 MR. CHANG: You can_ disregard sarety guide 22, 16 MR. WYLIE: GDC 22?
17 MR. MICHELSON: At this point, I don't know. It 4
- 18 is a-document in the public domain.
. 19 MR. WYLIE: What they're trying to do-is to be 20 practical here. The protection systems are channelized.
i 21 -They are able to test those at power by channel, four l
.22 trains, and~so forth. Take one out and test it. It's the 23 output-relay that we're talking about. -It's the slave
- 24 relay. Just that slave relay. What they've proven is that 25 the reliability of t hat _ relay is so good - that- they can do
]
2]
t ANN RlLEY _ & ASSOCIATES, Ltd.
Court Reporters i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 1-(202) 293'3950
. , - . -,-#c .
, - w,-,m-=, - , . . . , v
j l i
315 1 this, that they can delay this testing until it's i
O 2 appropriate to test it.
3 MR. MICHELSON: In nearly all cases it's only once I
4 in ever so many cases.
l j 5 MR. WYLIE: Relays are good for millions of i
6 operations.
l
- 7 MR. LINDBLAD
- T.t seems to me the other thing i ,
j 8 you're testing, though, is circuit integrity; are you not, 9 and whether a jumper has been left in place?
I 10 MR. CARROLL: Or a coil had burned up.
A j 11 MR. LINDBLAD: I think the circuit integrity issue i
j 12 in some periods --
i 13 MR. WYLIE:
What has been the experience?
14 MR. CHANG: Circuit integrity is part of the whole
! 15 testing package. We have to do continuity tests, for 16 instance, across circuit components and so forth, just to l
17 make sure the circuit is intact.
18 In the old days I think they did use jumpers quite 19 often, but that's discouraged nowadays. I think a lot of 20 times they have~ operating procedures just to double check to i
21 make sure that after the test those jumpers are taken away, j -22 bypass jumpers, i
- 23 MR. LINDBLAD
- That's-by inspection, and not by j 24 test.
25 MR. CHANG: That's by inspection.-
4
{ ANN RILEY &- ASSOCIATES, Ltd.
Court ' Reporters 1612 K. Street, N.W., Suite' 300
. Washington, D. C. 20006 l (202)l 293-3950 .
_ .,, . . . . . ~ , - -
l 1.
. 316 i
I l' MR. LINDBLAD: And administrative control? l lO 2 MR. CHANG: Yes. !
- 3 MR. MICHELSON: What are the protection system i
4 functions that you think you have in mind in this particular j 5 case?
6 What?
I 7 MR. CHANG: Whatever the system is designed to
. 8 function -- to do the function.
i- .
9 MR. MICHELSON: No. What systom?
l
, 10 MR. CHANG: We are talking about ESF.
11 MR. MICHELSON: ESF? Some people don't say a j 12 protection system and ESF are the same thing. It depends on 4
l 13 who you talk to and who they are and their thinking. The 14 GDC came about when the only thing you ever had to worry
- j. 15 about was getting rods in, and, yes, this was written 16 correctly for that, and, yes, they put in four breakers so i 17 they could test each one independently. But that was --
18 protection system then meant only one thing: A system that
}
19 got the rods in. Then it expanded out eventually.
20 We have dreamed up these things that'we've-got to
- 21 get'some water in and so forth. Some people call those 22 protection systems and some call them ESF to-keep these two 23 separated, and I don't know what we mean here.
- 24 MR. CHANG
- Well, protection system is general 25 terminology and that usually covers two different systems, O ANN RILEY & ASSOCIATES, Ltd.
C Court Reporters
"~
1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950 3
.- , , _ , . - --, . , ~ -
i 317 j
< - 1 the reactor protection system and the ESF system.
' 2 MR. MICHELSON: Okay. Well, then you've got a 3 real big bag of-worms.
J.
i 4 MR. CHANG: Well, it turns out that over the 5 years, based on the indication, the concern really is on 6 this slave. relay, the ESF system.
1 7 MR. MICHELSON: Yes, but I think there are
- 8 probably hundreds of those in the ESF systems, literally, 9 because every actuated device has got some kind of a l 10 contractor that turns the power on.
t j 11 MR. CHANG: Yes. Particularly, I think it's about i
, 12 100 or so.
l 13 MR. MICHELSON: Yes. Easily.- You think that was a
14 what was meant when they wrote GDC 21?
- 15 MR. CHANG
- Well, GDC 21, all they say is for the 16 protection system, you should have the capability to perform 17 the function to mitigate any consequences in case an l 18 accident happens, and also that you have to have on-line
- 19 testability capability.
i 20 MR. MICHELSON: Okay. Let me ask another
+
! 21 question. When you talk about ESF systems, are you also 22 talking about all the essential supporting auxiliaries that 23- without which the ESF doesn't work as a part of this idea?
24 You know, water systems and so forth. The ESF is useless if 25 the cooling water doesn't work to the motor or diesel ANN RlLEY & ASSOCIATES, Ltd.
l}
l Coud RepOners 1
-1612 K. Street, N.W., Suite 300 l i Washington, D. C. 20006 (202) 293-3950 l
4 4
318 4
1 generator,_whatever.
73
- 2 MR. KADAMBI: Excuse me, Mr. Chairman. My name is .
l 3 Prasad Kadambi. I wonder if I can just step in here and try H 4 to indicate, at least by way of background, how we approach 5 this. You know, maybe there are other ways of doing it, but 6 we in a sense approached it from the other direction. That i 7 is, we looked at the operation of experience so far and also j 8 we defined the generic safety issue, as T-y mentioned, more l
9 narrowly than, you know, one might define it in terms of if 10 you just looked at on-line testability of protection system.
11 12 Also in one of the earlier slides, he mentioned 13 that, you know, the way we looked at it is it included the 4
14 protection system as well as the ESPAS systems.
l 15 So what we did is we took these -- this universe 16 of slave relays and tried to understand what does the i 17 operational experience tell us about this, and what it tells
. 18 us is that if one insists on on-line testing,_then there is 19 a significant contribution to inadvertent actuation of l 20 safety equipment and inadvertent reactor trips.
21 MR. MICHELSON: I understand all that. .I was only
. 22 trying to look at your definition of protection systems.
23 It's a rather broad-scope definition'and I was_just trying 24 to figure.out how broad it was. Does it includeLthe water 25 systems, for instance?
i O' ^"" ai'ev * ^SSoc'ATes.'id-Coud Reponers 1612 K. Street, N.W., Suite 300 -
] Washington, D. C. 20006 (202) 293-3950
319
,cx 1 MR. KADAMBI: No.
I i 2 MR. WYLIE: It includes the functions.
3 MR. MICHELSON : Well, why doesn't it? That's a 4 part of what's needed to make an ESF system work, is the 5 cooling water.
6 MR. LINDBLAD: They are normally operating.
4 7 MR. MICHELSON: Beg pardon?
8 MR. LINDBLAD: They are normally operating.
9 MR. MICHELSON: Well, that doesn't really -- some 10 are cnd some aren't.
11 MR. KADAMBI: Well, the point I was making was for 12 on-line testability, whether the water system is included in 13 the scope of this generic issue or not depends on where that 14 slave relay -- what will happen if that slave relay is
(~))
15 tested at in-power operation.
16 Am I right, T-Y?
17 MR. CHANG: Yes.
18 MR. MICHELSON: I am not sure we appreciate the 19 scope that you're defining here. For instance, every motor-20 operated valve has to have a contractor. That's what makes 21 the motor work. And every one of them is essentially a slave 22 relay. Now, you can do solid states all the way up to 23 there, but now you have to go to a mechanical contractor to
, 24 supply the power to the device, and every one of those is in 25 your scope, then, and you can't test any of those in the fl "
ANN RILEY & ASSOCIATES, Ltd.
Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
i 320 4
1 same sense.
2 MR. LINDBLAD: But isn't it true that for some of 4
3 those water support systems which are already operating, 4 should the relay actuate and should power be put on the
! 5 motor, it can be tolerated; whereas the systems they're l 6 talking about, they don't wish to tolerate --
- 7 MR. MICHELSON
- Well, I'm thinking of where we 8 have dual water supplies, for instance, and you've got to
{
9 cut off the normal service water, set up valves --
10 MR. LINDBLAD: Like the chilled water at River
! 11 Bend.
12 MR. MICHELSON:_ Whatever. You start talking about 13 motor-operated valves, you're talking about hundreds of f
i
( 14 contractors that are all a part of the engineered safety
}
15 features systems.
i 16 MR. WYLIE: Well, I don't think he's talking about i
17 these. What he's talking about are the output relays from 18 the protection systems, whether it be --
t 19 MR. MICHELSON: Well, that's what'I-asked, and 20 they said-they.are including the ESF.
l 21 MR. WYLIE: Well,'it's like high pressure safety
~
22 injection, that function on a train basis. Isn't that 23 right?
- 24 MR. CHANG: Yes.
. 25 MR. WYLIE: It's the output relay from high ANN RILEY & ASSOCIATES, Ltd.
f] Coud Repoders-1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
321
- 1 pressure injection or low pressure injection, whatever it
, 2 happens to be, whatever --
1 3 MR. CHANG: Yes.
! 4 MR. WYLIE: It's not everything downstream of j S that; it's just the --
6 MR. CARROLL: In a Westinghouse plant' it's the S 7 relay and nothing more.
8 MR. WYLIE: Nothing beyond.
9 MR. CARROLL: For instance.
- 10 MR. MICHELSON
- That's why I was asking what the
- 11 scope of this thing was.
12 MR. KADAMBI: Well, I guess according to the tech
- l. 13 specs, all the relays need to be tested that assure the 14 safety function of the equipment, and what we are talking
[ )
l 15 about is a subset of the tech spec relays, that if it is i 16 actuated on line, can cause adverse effects.
i
( 17 MR. MICHELSON: You are just talking, then, about l
l 18 the control system relays only, none of the ESF systems.
I
! 19 Like RHR has got a whole lot of stuff in it too, but the
! 20- actuation of RHR comes from the' control system through a
! 21 bunch of relay logic and solid state logic --
22 MR. KADAMBI: And there may not be.an adverse 23 effect like a reactor trip if the. control system, you know, 24 actuates inadvertently, and therefore there may not be any l 25 real problem with testing it on-line.
I ANN RlLEY & ASSOCIATES, Ltd.
\] Court Reporters 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950 t
l 322 1 MR. CHANG: Those are the concerns. Those are 2 .really the things that could lead to_ unwanted actuation of l 3 aafety equipment for reactor trips. Those are really the-i 4 concerns.
- 5 MR. MICHELSON: The. reactor trip I understand, the l 6 inadvertent actuation of a valve, for instance, might be a 7 little harder, because now you are talking a lot of logic,
- 8 when you are talking about getting down to the valve logic.-
f 9 As long as you stick to SCRAM only, insertion of i
j 10 rods, no problem, and I thought that is what GDC was 11 addressing.
i
, 12 MR. CARROLL: It.is broader than-that.
i 13 MR. MICHELSON: I am trying to figure out how much l\
(~' 14 broader.
l 15 MR. KADAMBI: Why don't you go on to the next one.
16 [S.l ide. ]
l 17 MR. CHANG: On the other hand, because the staff l 18 did some studies on their own of the operating _ experience of 19 slave relays and comparing it with the eqdipment that is 4
20 . going to actuate downstream, first_of all,_the staff found
,i 21 that, as I mentioned earlier, online-testing of certain 22 relays and triggering inadvertent actuation of safety
. :2 3 equipment and the reactor trips, it_'is.a significant 24_ contributor-to those trips of those actuations.-
25 Also,. based on test data and operating experience,
'.(\T' ANN _ RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. - Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
- - - . . . .. . - . . - -. ~ -._.- . - .
h I
4 1
j 323 s 1 it appears that general slave relay reliability is, in .1
- 2 general, pretty good, and as compared-to the actuate )
3 equipment, such as motor operated valves, we looked into the 4 utility submittals on an IE bulletin, and we made some 4
) 5 calculations, and it appears that the reliability of the 6 slave relays are one order higher than other operating l 7 valves.
4 j 8 MR. MICHELSON: What are you saying, what did you i
9 say?
f 10 MR. CHANG: The failure rate we estimated for i
11 slave relays is one order lower than the failure rates we 1
l 12 estimated for motor operated valves.
13 MR. MICHELSON: What do you think the motor 14 operated valve rating was?
15 M1 . CHANG: Do you want some numbers?
l 16 MR. MICHELSON: You should give me a feel.
17 MR.' CHANG: Based-on the submittals, we think it i
l 18 is around five failures per 1 million hours. Usually, I L
19 think, the failure rate --
20 MR. MICHELSON: Is that for the motor-operated 21 valves?
l 22 MR. CHANG: That is the failures per demand, and.
i 23 it is hard to compare that with relays because relays are 24 usually estimated in terms-of duration of time. So we have 25 based.it on.the service duration.of those plants, and h ANN RlLEY & ASSOCIATES, Ltd.
Coud_ Reponers l 1612 K. Street, N.W., Suite 300 i- Washington, D. C. 20006 l_ -(202)-293-3950
4 324 i
g
-~ 1 converted into similar units so that we could compare.
~# 2 MR. LINDBLAD: Is this valve actuations, or was it 3 just valve flow?
4 MR. CHANG: Whether it functionc, whether it is 5 going to open or not. If it doesn't'open, or if it is i 6 supposed to close, it doesn't close, then it constitutes a
- 7 failure.
f 8 MR. MICHELSON: A relay conductor only operates as 9 frequently as the valve, no more and no less frequently.
10 They operate in unison, so to speak, so you can't have a 11 different rate on one than the other.
12 MR. DAVIS: If they don't fail.
13 MR. MICHELSON: If they fail, that is another
!(} 14 15 question I am going to ask, how do you know whether the valve failed, or the relay failed. If you have a detailed
( 16 enough database, T guess you could figure that out, but the i
17 valve database usually says something went wrong and the L
18 train, such as the valve, didn't open when we asked it to.
19 MR. CHANG: This is based on operating experience, l 20 so what you say is true. Whenever you want to actuate the i
l 21 equipment, the slave relay has to function.
22 (Slide.).
23 (Pause.]
-24 MR. CHANG: The Technical Specification Branch, I 25 think they made a presentation yesterday --
,O ^"" aev a ^SsociaTes' d-Coud Repoderer 1612 K. Street, "N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950-
. 325 1'
1 MR. CARROLL: They did.-
2 MR. CHANG: -- on the tech specs improvement
. 3 program, they are in the process of proposing a NUREG-1366 4 and the associated generic letter. We have Bob Tjader here.
5 He is the author of those documents.
6 Based on what we have learned, that online testing 7 of certain relays may lead to unwanted inadvertent actuation 1
8 of safety equipment or reactor trip, and also the fact'that 9 relay reliability seems to be generally good, they propose l
10 to change the test frequency of the slave relays online.
l 11 The proposed changes that instead of testing 12 online all of the slave relays quarterly, of course with the 13 exception of those hard wired ones. They propose to test it l
14 quarterly but not on every one. It would be on a sampled
{}
15 basis and then this quarterly test would be on a staggered 16 basis over a refueling cycle.
1 17 Also, secondly --
l 18 MR. CARROLL: What is the rationale for that?
19 What is that a good thing to do from a' safety point of view 20 and if it is, has a probabilistic risk assessment been done 21 to show that it is cost beneficial?
22 MR. KADAMBI: Bob Tjader from NRR.
23 MR. TJADER: Good afternoon. NUREG-1366 was 24 initiated as part of the tech spec improvement program and 25 it was initiated to analyze surveillance requirement and the Q- ANN _ R1 LEY & - ASSOCIATES,-Ltd.
Coud Reponers 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006
, (202) 293-3950
1 326 a
1 surveillance requirements were looked at and some -- it was
\' 2 done on a qualitative type basis. In other words, we looked l 3 at LERs -- License Events' Reports, NPRDS failure rates.
4 We went out and discussed the surveillance 5 requirements with the utilities -- five units in particular.
6 As a result of that analysis, and in particular 7 with regard to the slave relay, which is one element of 8 that, we found that the reliability of the slave relays was 9 very high. In fact, of the many slave relays in the plant, l
l 10 there's only approximately two failures per plant per year.
11 At the current time, the surveillance requirements 12 on the slave relay and the tech specs are done on a monthly 13 type basis and this was deemed to be-too frequent.
, ("' 14 Now there were some criteria by which we used --
\-
) 15 in other words, we took a look at inadvertent actuations 16 including plant trips and also actuation of equipment, wear
. 17 on equipment, radiation exposure -- now this doesn't 18 necessarily involve this one here but radiation exposure 1
19 with some burden on the plant as far as the gain and risk
! 20 that was involved per the-surveillance.
21 As a result of that analysis, it seems that we are
! 22 testing the slave relays far too frequently and shifting it l- 23 to a quarterly basis is more than satisfactory.
24 MR.-CARROLL: My question had to do with the
- 25 second item, the staggered testing.
l .]
t ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K.-- Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
4
, 327 f 1 MR. TJADER: The staggered testing?.-
2 MR. CARROLL: The rationale for it.
) 3 MR. TJADER: Well, it is tied in with-the
- 4 inadvertent-actuations and the fact that-we get incidents,-
5 not incidents -- license event reports every=now and then of 6 inadvertent actuations of systems that in effect_the slave 7 relay worked as anticipated and has for over the history of
{
- 8 time and has not-failed, if I am answering the question, t
j 9 MR. KADAMBI: I think the answer is that the
) 10 staggered basis testing is_ expected to decrease-the e
11 inadvertent actuation.
l 12 MR. CARROLL: Why would that be?-
13 MR.'TJADER: Well, you are not testing everything 14 on a monthly basis. You're testing samples from the various 15 systems on a quarterly' basis, using that as an' indicator 16 that the system'is still operable and ready to function as i 17 designed.
18 MR. CARROLL:. But screwing up'one slave relay test I 19 can cause trouble. --
l
- 20 MR. TJADER
- Yes.
21 MR. CARROLL:= On a particular system, and if that 22 system has two slave' relays, why does_it make any difference
- j. 23- whether I do it at:the same_ time or on a' staggered basis?
j- 24 MR.-TJADER: _ If you are t'alking about different
! 25 trains of;a system or something like-that, you'could do it; F(]
ANN RlLEY;& ASSOCIATES,'Ltd.
Court Reporters i 1612 K. Street, N.W., Suite ~ 300 l . Washington, D. _ C. 20006 (202) 293-3950 g
- , _ - . . - . _ . _ . . , _ , . - - _ - . . ~
j l
4 328 4
1 at one train a quarter and then the other train the'next 2 quarter. I think you are looking at it on a series type 3 basis but if it is a parallel type test --
4 MR. MICHELSON: But you are not changing the l 5 frequency of the testing at all?
4 6 MR. TJADER: Well, the NUREG-1366 proposes from 7 going from a monthly to a quarterly frequency in the
{
8 surveillance requirements.
9 MR. MICHELSON: Unless you change the frequency, l 10 you haven't done anything. Whether it is staggered or not 11 makes no difference. I think that's all we are questioning.
- 12 If you haven't changed the frequency, why have you 13 accomplished something by staggering?
! 14 MR. CHANG: Well, actually, I think there is a 15 difference there because now we don't require that for each 16 quarterly test you test all the relays, so you select one
- 17 subgroup of the relays that you are going to test in this 18 quarter and in the next quarter you may select another 19 subgroup, so in reality it would-be just testing one relay 20 in the refueling outage. That is an extreme but the-21 submittal is going to be reviewed on a case by case basis by 22 the Staff.
23 MR. TJADER: Yes. We are reducing the frequency.
24 MR. CHANG: The Staff always has.the final say.
- . 25 MR. CARROLL: We heard'a presentation last month-3
-'() ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202)-293-3950
329 1 in which AEOD talked about their. review of what they very 2 loosely termed common mode failures, and they were talking 3 about staggered testing as a way to avoid some of those. In 4 that instance, the argument was that by doing staggered 5 testing if some common mode failure occurred, you would 6 catch it sooner, because if you are' going to do it on a 7 quarterly basis, if you do them all in the same quarter, you 8 are going to have a problem the day after the last test, and 9 it takes the whole quarter to catch it. It takes only half 10 a quarter to catch it, if you stagger it.
11 The other argument, which I didn't find at all, 12 was that you are less likely to create a common mode failure 13 if you do it on a staggered basis because you wouldn't be 14 using the same people to perform the test, and if you leave 15 the lead jumpers longer, and whatever.
16 Is that part of this?
17 MR. TJADER: It is tied in with it. What we are 18 doing is, every month there will be slave relay testing on 19 the same basis, and over the period of a quarter, it will 20 have covered the slave relays.
21 MR. CARROLL: I am asking what is your rationale 22 for doing that, because that is probably not an efficient 23 way to do work. If I am the maintenance manager in the 24 plant, I would like to do all the. relays in one fell swoop, 25 and turn my people loose on other jobs.
O ANN RlLEY &- ASSOCIATES, Ltd.
Court Reporters 1612 K. Street,' N.W., Suite 300 '
~ Washington, D. C. 20006 (202) 293-3950
-f 330 g 1 MR. TJADER: I don't quite see that argument 2 because we are going from doing everything on a monthly 3 basis to-doing sampling on a monthly basis. In other words, 3
4 shifting it primarily from -- we are_ extending the frequency j 5 from a monthly to a quarterly basis, j 6 MR. CARROLL: Once you have done that, then you 7 have a choice. You could either say it is okay to do it all 8 in one fell swoop, or to stagger it.
9 MR. TJADER: And we are staggering it.
j 10 MR. CARROLL: If I was the maintenance manager, 11 once I got the concession that said we can do it quarterly,
- 12 I would say, "Well, I want to do it all at once." There are
] 13 learning curve problems.
14 MR. TJADER: But it doesn't reduce the burden on a
- 15 monthly basis.
16 MR. CHANG: The proposal in NUREG 1366 is just
, 17 something that is acceptable to the staff, and they can 18 propose that voluntarily on their own. If they don't chose 19 to go that route, they can always stick with-the present 20 requirements.
21 MR. KADAMBI:- I think Mr. Carroll's question is, j 22 what do we feel is the benefit of the staggered testing, and 23 perhaps we will take that on as a question that we will ,
i 24 answer for you_ separately. l l
25 MR. CARROLL: All right. l O ANN - RILEY:& ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
331 a
1 MR. WYLIE: I would think it would be of more-2 value to the utilities.
3 MR. CARROLL: To do it staggered?
l 4 MR. WYLIE: Yes. I mean it is a matter of time j 5 and getting all this stuff done.
6 MR. WILKINS: You would have disagreement on that.
i 7 MR. WYLIE: I am not disagreeing. I am just 8 Maying that I would expect this, j 9 MR. LINDBLAD: I think the point is, if there is
- 10 no particular merit, then it ought to be at the utility's I 11 discretion, i 12 MR. CHANG
- So they have another choice, that is 13 what we are saying.
14 MR. WYLIE:
I am looking to the language, and I 15 don't see the language that'says what they have to do.
16 MR. CARROLL: Once I get a crew assembled, and all 17 their test equipment, and everything ready to go, let's go i
3 18 get it done.
i 19 MR. CHANG: Certainly they can choose that.
20 MR. CARROLL: You are saying if they choose it, 21 then all that means is they get to do all of them every 4
22 month.
2 23 MR. KADAMBI: The other point-is, the committee 24 will have the opportunity to look at the generic letter and 25- NUREG 1366, I believe. So this is one part.of the Generic ANN RILEY & ASSOCIATES, Ltd.
Q' Court Reporters 1612 K. Street, N.W., Suite 300 Washington, - D. C. 20006
- (202) 293-3950
._ __ .- -. - .- - - = - _ . . . . . . - - - _ _ _ _ _
332 1 Issue 120 resolution.
2 MR. CARROLL: All right.
! 3 MR. LINDBLAD: Could I ask a question, I don't 4 know if the staff representatives followed the EPRI 5 requirements for future advanced reactors, is this an issue 6 in the EPRI document, or do our people who work on that know 7 when we say what will the application be on future plants?
! 8 MR. CHANG: In the Advanced Light Water Reactor
- 9 Evolutionary Decign Requirements document, there is a 10 section about ongoing safety issues, how they are going to I
11 address those issues, and on this, I believe, because the
- 12 status at that time when they prepared the document was not 13 clear yet, so they didn't make any commitment on this issue.
14 But we, as the staff, when we reviewed their-i 15 submittal, we did make the comment that on this issue this ,
16 remains an open issue. If they choose to take an exemption 17 to the requirements, it still has to be reviewed on case-18 by-case basis.
4 19 The recommendation is that they should design-the 20 capability of online testing into the plans at this point l 21 instead of trying to seek the exemptions later on, t
22 MR. MICHELSON: Is that the position you took on 23 ABWR also?
i 24' MR. CHANG: We-have not reviewed that document 25 yet, but I think it is going to be the same approach.
ANNL RlLEY & ASSOCIATES, Ltd.
- >O' Coud Repoders.
1612 K. Street, N.W., Suite 300
. Washington, . D. - C. 20006 J
(202) 293-3950
k I
333
. 1 MR. MICHELSON: I am still trying to sort out the i
2 difference. I mean, I understand this whole argument of a
l 3 staggering testing, reducing the amount of testing and all 4 that, but what does that have to do with'the basic generic l 5 issues, which is they need not to have the testing i
6 capability itself? Have you relaxed that need to have the 7 testing capability?
8 MR. CHANG: The concern at the beginning was that
! 9 at that time, because of the lack of data, there was a I
- 10 concern, as I'see it, among the' staff that how-reliable are 11 those slave relays, okay? If those relays cannot be tested 12 on line periodically, what are the impacts on safety?
13 That's how this issue came up.
l[ )
14 MR. MICHELSON: Okay. Now, what is your final j 15 proposed resolution relative to the~need for testing i
16 capability?
- 17 MR. CHANG
- Now, we have found that, first of all,
{ 18 some of the testing on-line of the slave relays may. lead to 4
I 19 unwanted trips _and actuation of equipment, and also-that 20 inadvertent actuation due to this on-line testing of slave
- 21 relay.is a' big contribution, and also that the relays are i
- 22 fairly reliable.
- 23 .So based on all those facts and the fact that the l 24 staff is allowing _ relaxation on the test frequencies and so 25 'forth on.the slave relays, so we --
C J ANN RILEY- &- ASSOCIATES, Ltd.
4 Coud Reponers 1612 K. Street, N.W., - Suite 300 s' Washington, D. C. 20006 (202) 293-3950
334 g 1 MR. MICHELSON: Reducing the test frequency i
2 doesn't remove the need to have an arrangement that's 3 testable unless you reduce it to once every refueling 4 outage, then you could have an arrangement that might be 5 testable.
6 MR. CHANG: Right. See, this bullet, okay -- for 7 those high-risk relays, meaning that those may lead to 8 inadvertent reactor trip, then the staff would allow the on-9 line testing to be moved to the outage. So this is a major 10 difference from what's in the Safety Guide 1.22.
11 MR. WILKINS: Does that require you at some point 12 to go back and change GDC-21?
13 MR. CHANG: No. No.
14 MR. MICHELSON: GDC-21 is real good for some Q( 'N 15 systems; it's just not so good for others.
16 MR. CHANG: GDC-21 is like a general blanket type 17 of document, and you have those --
18 MR. MICHELSON: It was written for a different 19 situation, and there it was very important, and I don't 20 think you're relaxing that at all. But I'm not sure it was 21 -- you know, it didn't even have -- ESF systems weren't even 22 in existence when GDC-21 was formulated. You're talking 23 research reactor days.
24 MR. CHANG: It has changed to Safety Guide 22. j 25 The Safety Guide 22, the consent there is about the hard-j (l#
' ANN RILEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 i
s 1
335 1 wired slave relays with downstream equipment. The. concern 2 at that time was not on the inadvertent. trip caused by 3 testing of slave relay, but over the years, because of the
- 4 experience gathered, it turned out that on-line testing of 5 slave relays, the biggest concern probably is the i
) 6 inadvertent actuation of reactor trip. So that thing gets 7 factored into this NUREG-1366, to reduce the risk of 8 triggering the plant.
9 MR. WILKINS: Well, have you determined that --
10 and I'm trying to read some-language here -- on-line 11 testability of ESF systems is no longer an issue of concern?
12 MR. CHANG: That's what we are recommending on the 13 next slide.
l 14 MR. WILKINS: okay.
, 15 (Slide.)
- 16 MR. CHANG
- Based on the discussion I have given, 17- we recommend that Generic Safety Issue 120 can be considered
- 18 technically resolved. This is mainly based on the first two 19 bullets.
20 The risk of on-line testing of some of the slave 21 relays outweighs the benefit you can reach, and it's really l
22 a balance of the two that-you have-to consider, and also 23 that based on tests and operating experience, slave relays 24 seem in general pretty reliable, but slave relays here, I 25- think in general, we are talking about 120 volts, small l
Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300'
-Washington, D.l C. 20006
- l. (202) 293-3950.
, , ,. . . ~ , - n
I ,
l i- 336 1 relays. Whether it's AC or DC, we're oniv talking about
!-- 2 that group-of control relays. Based on test data and 4
3 operating experience, they seem to be highly reliable.
. 4 MR. MICHELSON: Are you talking control re)ap , or j 5 are you talking such things as motor contractors, because j 6 you kept talking about motor-operated valves and stuff, and 7 that's a motor contractor there. Some people call them i
j 8 relays. They really aren't; they're contractors.
i 9 MR. CHANG: We are talking abelt the small 10 equipment-that may trigger a circuit breaker downstream and i 11 in turn actuates equipment.
1 12 MR. MICHELSON: You are just talking about control i 13 relays, then, in the control. cabinets themselves, very small 14 devices.
) 15 MR. CHANG: That's right.
3 16 MR. WYLIE: -As Lindblad calls them, auxiliary l
l 17 relays.
l 18 MR. MICHELSON: Yes, auxiliary -- well, some 1- 19 people even reserve that for another category-of relays.
l 1- 20 _These are control relays..
4 l 21 MR. CARROLL:- And if that indeed is what the scope
- 22 is, then it isn't very expensive,_the design and circuitry 23 that would allow:the testing.
24' MR..MICHELSON: That's right. You put in test I
-25 positions. :
e ANN RlLEY- & ASSOCIATES, Ltd.
![k}-
1 Court Reporters
- 1612 'K. : Street,- N.W., Suite 300
[ _ Washington, D. C. 20006 l (202) 293-3950-E d
i 337 i
j 1 MR. WYLIE: We're only talking about a limited j 2 number of plents here now; we're not talking about all i
{ 3 plants.
! 4 MR. CARROLL: What do you mean?
j 5 MR. WYLIE: On this particular thing, you're only i 6 talking about a limited number of plants, right?
I 7 MR. CARROLL: No , it's --
l 8 MR. CHANG: For those plants that seek the
!. 9 exemptions on their on-line testing _of slave relays, then 10 each one has to go through this type of review.
11 MR. WYLIE: How many plants are we talking about?
1 l 12 'MR . LINDBLAD: As many'as they ask for.
13 -MR. MICHELSON: Everybody uses the relay logic.
!O 14 MR. WYLIE: Oh,-sure. But, I mean, how many are
!O j 15 we talking about?
l l 16 MR. CHANG: I think there are very few plants that l
17 don't request any exemptions at all. . It is a degree of how many exemptions ~they're looking for.
18 19' MR.--CARROLL: You are talking virtually the whole
!_ 20 population.
21 MR. CHANG: For instance, if it's a boiling water
. 22 reactor,-I don't think you can test ADS actuation valves.
23 MR. SHEWMON:- Have we reached the end?
[ 24 MR.-CHANG:- One more thing I want to mention is 25- that we sent a memo-to-the-EDO office, and'we-recommend-that
(- - ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters
. 1612_ K.' Street, - N.W., Suite 300 -
, -Washington, D. C. 20006
- (202) 293-3950
1 i
I 338 1 ,
j 1 we consider this can be technically resolved. That was in 2 September.
3 MR. LINDBLAD In the review of exemption request, l
j 4 do you consider whether the unit is on a one-year or a two-5 year fuel cycle, or does that not matter to you?
- 6 MR. CHANG
- Well, the tech spec I think, when they
} 7 talk about a cycle, could be either one-year or two-year d cycle.
I 9 MR. MICHELSON Or it could be 18 months.
10 MR. CHANG: Yes, 18 months or 24 months.
- 11 MR. SHEWMON
- How are we doing, Charlie? Do you
, 12 have everything you need?
13 What does the committee l A . WYLIE: I'm not sure.
i 14 want to do? Do they want to write a letter in favor of i 15 this? Op, e'it? What do you want to do?
4 16 MR. MICHELSONt or remain silent.
i l 17 MR. WYLIE Remain silent, or --
i
- 18 MR. LEWIS I must have missed an important point.
! 19 Somewhere along the line, this switched from the general 20 question that's shown in the title to the slave relays. Was l 21 that t~ansition clear?
l 22 MR. WYLIE: Well, it was to me.
j 23 MR. WILKINS: I noticed when it was made.
[ 24 MR. LEWIS Because it started out to be an. issue 3
, 25 of global significance and it --
O ANN RILEY & ASSOCIATES, Ltd.
Court Reporters
, 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 i
.(202)_293-3950 1
- _ . . - . , _ , , , .,_,...__,,..a..,,.~._
_ . _,.-.....L...._.-.-,.
I i 339
] 1 MR. CARROLL: He very early on explained that.
l
) 2 MR. CHANG: I think at the beginning, we said we l l
3 weren't sure about the scope of the problem. That's why we l
1 4 chose the more general title.
1 S KR. SHEWMON: It is the GSI that's under review,
- 6 not the GDC.
j 7 MR. LEWIS: No, I understand that, but the GSI has 8 this generic title, and it's -- the point is if we're l
' 9 talking about a resolution of the GSI, is the GSI as
{ 10 narrowly constricted to slave relays or has one concluded 11 that there really was no basis for making such a general i
i 12 GSI? I know you said it, you slipped through it.
13 MR. CHANG: I think what we found was that the
]
14 scope really should be just on slave relays for ESF.
(
15 MR. LEWIS But this is relevant to what we write,
~
16 because we write --
t 17 MR. SHEWMON: Yes. Do we want to agree with him.
18 MR. LEWIS: On that point. Okay. I was only 19 reacting to the possibility that we might only express 20 ourselves on the slave relays and let the general issue go 21 by. That's all. And I would hate to see that.
22 MR. WILKINS: I would have been happier if the i 23 notebook had had a copy of something that was called GSI-24 120.
25 MR. LEWIS: Yes. I've been looking for it and I.
i I O
ANN RlLEY & ASSOCIATES, Ltd.
, Court Reporters 1612 K. Street, N.W., Suite 300 3 Washington, D. C. 20006 (202) 293-3950 I
-v - - , - - , .- - - .. .--
340 1 don't find it.
2 MR. WILKINS: It's not here.
3 MR. LEWIS: Why don't we have it?
4 MR. SHEWMON: The staff never comes up with a 5 statement of these things in my experience of late. It's 6 always the paraphrase of what they choose to be looking at, 7 not the historical document. I guess that's a general 8 complaint we have.
9 MR. LEWIS: But in this case, where the title in 10 so different from the proposed resolution, it would be nice 11 to know.
12 MR. SHEWMON: In all cases it would be nice to 13 know historically what it was.
14 MR. LEWIS: A little knowledge is a dangerous 15 thing. .
16 ML CARROLL: I'm sure Medhat can get us a copy.
17 MR. WYLIE: I guess the only concern that's been 18 expressed by the committee is the staggered testing. The ,
19 staff --
20 MR. CARROLL: Well, no. I had a second concern, 21 Charlie.
' 22 MR. WYLIE: Okay.
23 MR. CARROLL: I want to know what their -- I 24 didn't get a clear understanding as to what their position 25 was going to be in terms of future plants.
] ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters
- 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 a
k 341 4
i 1 MR. LINDBLAD: But I thought their comments on thu
>O i
1 2 EPRI document were enlightening on that.
3 MR. CARROLL: To some degree, except I don't i
4 remember what the EPRI document says exactly.
J j 5 MR. WYLIE: Well, what does it say?
! 6 MR. WILKINS: That's the ALWR requirements.
l 7 MR. WYLIE: Yes.
8 MR. CARROLL: Can somebody get-the right volume i 9 out?
4 10 Mn. CHANG: I think it just says they have to 11 follow the requirements stated in those documents.
6 12 MR. CARROLL: You are saying that you endorse
{
l 13 whatever the EPRI requirement document says? Is that what I 1
14 understood?
l 15 MR. CHANG: No. On this issue, what I said is 16 that they still have to look into this requirement on the 17 resolution of this issue. In other words, if they seek 18 exemption, they still have to submit the exemption and it 19 has to be reviewed by the staff on a case-by-case basis 4
20 according to 1366 and the generic letter.
i j 21 MR. CARROLL: One of the things EPRI advertises
- 22 that they are doing is negotiating these issues with the 23 staff for the future-plants, getting the staft' to agree with 24 them so that you don't have to do what you're describing.
25 MR. CHANG: In the last issue of the document _it ANN RlLEY & ASSOCIATES, Ltd.
!] Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
.. . . .. . , _ = -- . - -. . -. -
342 2
1 was indicated that there was nothing specific on this issue.
8 I 2 So we gave them the feedback that this is the way that we're 3 going to approach this issue. They have to follow this in j 4 the future.
j 5 MR. CARROLL: All right.
t 4 6 MR. WYLIE Let's see if I'm clear. The 1
! 7 recommendation for staggered testing is over a fuel cycle, 8 right?
j 9 MR. CARROLLt Quarterly.
1 l I
, 10 MR. WYLIE: Quarterly over a fuel cycle.
J l
11 MR. TJADER: Every quarter, everything should be l 12 covered. That, in effect, extends the current surveillance 13 frequency from monthly, which it is now, to quarterly.
14 MR. WYLIE: But you stagger it so that you're not j 15 checking the same channels at the same period?
16 MR. TJADER: That's right.
17 MR. WYLIE You're doing that for common mode
- 18 failure?
i 19 MR. TJADER: It reduces the likelihood of common 20 mode' failure.
I
- 21 MR. WYLIE That is the rationale.
22 MR TJADER: It also reduces equipment available.
\
23 It does have disadvantages, apparently. It might have the
- 24 effect of increasing workload to some extent, but to me I 1
i 25 don't-quite see it at this time when-it's monthly, f' ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300
- Washington, D. C. 20006
- (202) 293 3950
. 343
, 1 currently, for everything, and you're now going to a f 2 staggered test basis. In essence, you could reduce the 4 l
3 workload even further by doing everything on a quarterly 4 basis. But that wasn't a consideration. l l
5 MR. WYLIE: It really is a relaxation.
6 MR. CARROLL: The point is you need a i
7 justification. If you're willing to relax it quarterly, but l
l 8 staggered, you should be able to say why.
j 9 MR. WYLIE: Why isn't quarterly good enough?
i 10 MR. TJADER: I think we have. I think we've 11 stated that the slave relays fall so infrequently i 12 historically.
13 MR. WILKIraf E ore not addressing his issue.
14 MR. WYLIE: kley iM flipt it six months?
l 15 MR. WILKINS: Tikc is not the issue either.
16 MR. SHEWMoN: Do you want.to try to ask the 17 question again?
18 MR. WILKINS: I'm not sure I can. I thought Jay's 19 question was quite clear. If you want each relay tested
{
20 every three months, or four times a year, why do you 21 require, or insist, or at least strongly urge that the 22 utility only test one-third of them each month instead of 23 testing three-thirds of them evety three months, a
24 MR.- TJADER: The' advantage of staggered testing-is i 25 it reduces the likelihood of common load failure. It also 4
ANN RlLEY & ASSOCIATES, Ltd.
..Q Court Reporters 1612 K. Street, N.W.,' Suite 300
- 3 Washington, D. C. 20006 l (202)-293-3950
,,. ~ --- . , - . . _ __
_-____ - . . - _ _ _ . - . - . _ . - - - - _ . - - - _ _ _ _ ._.... -. ____. _ _ _ . .. .~__
1 344 j 1 reduces system unavailability -- no, it does not, not over j 2 the whole time. Extending it to quarterly will reduce that.
4 3 MR. CARROLL: That's right, so the answer is --
l
] 4 MR. TJADER: It reduces the likelihood of common 5 load failure.
6 MR. CARROLL: I'm guessing the same procedures of 7 the same people doing the testing. How does it reduce i 8 common load failure?
i 9 MR. WILKINS: Because they won't make the same
- 10 mistake a month apart that they might have made a half an
! 11 hour apart.
4 12 MR. CARROLL: Do you have any evidence to show i
j 13 that?
14 MR. WILKINS: Only anecdotal.
- 15 MR. LEWIS
- He is being an mathematician.
3 16 MR. WILKINS: When I type something or make an
- i 17 algebraic error, or something like that, a month later I 18 will not make the same typing error or make the same 19 algebraic error.
! 20 MR. LEWIS: In fact, the period should be 21 incommensurable with your biological cycles, right?
l 22 MR. WILKINS: Incommensurable -- just long, 23 compared to it.
24 MR. LEWIS: Well, incommensurable is the only l 25 important thing. Long compared to it is one way_to put it.
,O ANN RILEY & ASSOCIATES, Ltd.
Court- Reporters 1612 K. Street, N.W., Suite 300
. Washington, D. C. 20006 (202) 293 3950
4 i
I
! 345 i
1 1 Never mind. You and I will discuss that later.
' 2 MR. LEWIS: I'm afraid that's true.
l 3 MR. MICHELSON: I'm still having trouble with the i
i 4 scope. I've gone back and read your CR 5916. It outlines l 5 all these requirements. Then it has a sentence which throws
! 6 me. ,
I l
- 7 "These requirements and guidance are intended to '
l 8 ensure that protection systems, including logic, actuation 4
j 9 devices, and associated actuated equipment will be designed 10 to permit testing while the plant is in operation of power l
11 .without adversely affecting the plant's operation."
{
12 You've just addressed one little piece of
}
i
- 13 equipment of the slave relays. How about all the rest of Q
14 the equipment in this train? How does that come into the
)
15 scope? That is what the scope of the generic issue was, I I 16 thought.
i 17 MR.. CHANG: That was the intent at the beginning a
f 18 of the issue. But we did look into operating experience
- 19 over the years. We narrowed down our concern because there 20 is no evidence that you have constraints on the other 23 remaining protection systems.
1
- 22 MR. MICHELSON
- You made all the arguments about i
J- 23 the reliability and the relay compared to the reliability of ,
24- the actuating device, I guess. How about the other parts of <
25- this train? There is more to this train than just the 4-i E
i Q ANN.- RlLEY & ASSOCIATES, Ltd.
.Coud Repuders i 1612 K. Street, N.W., Suite 300
. Washington,- D.. C. 20006 (202) 293-3950
i 346 i 1 relay. There's a number of other components between the I
2 sensor that says that you've got an adverse condition and
]
3 the final addressing of that condition all the way through l 4 the train and actuating the device to address it.
l 5 Now that's what I thought this generic-issue was 5 6 talking about, that you've got to be able to test that 7 train. Admittedly, some parts are more difficult to test 8 than others, but somehow all of a sudden this narrows down l
- 9 to the slave relay as being the whole problem.
j 10 Since it's a non-problem, then the whole issue 11 goes away. I just don't buy that approach. I don't 12 disagree with it.
> 13 MR. CHANG: It is based on the experience of 14 operating plants and the tests. There seems to be no 15 evidence of constraining other areas, whereas the slave 16 relay problem comes up again and again.
l 17 MR. MICHELSON:- Wait a minute. You convinced me
! 18 that the slave relay was a very reliable device compared to 19 the actuated device. I guess you are arguing that all of 20 the other' devices are even more reliable'than the slave 21 relay, which is more reliable than the actuated device.
22 MR. CHANG: Slave relays seem to be a problem area j 23 at~the beginning. There is no indication on other things.
- 24 MR. MICHELSON
- There are all kinds of indications-25 that instruments are problems, instrument contacts are
[~' ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters
- 1612 K. Street, N.W., Suite 300
- - Washington, D. C. 20006
-(202) 293-3950 l j
i 347 1 problems, switch contacts -- all the way through that train O 2 are various potential problems. The relay is just one 3 element in that whole daisy train from the instrument that 4 decides to tell you an adverse condition exists until you 5 address through an actuation device.
6 After we've gone through that all that train, you 7 said, well, the only unreliable device is the slave relay 8 and it's highly reliable compared with the actuated device.
9 I would have understand that. Maybe thrd's what you did and 10 I didn't catch it.
11 MR. CHANG: This is not what we are saying. What 12 we are saying is that we are talking about the on-line 13 testability. It is not whether the equipment or the device 14 is reliable or not.
15 MR. MICHELSoN: You're talking about a system 16 required to perform a particular protective function. That 17 includes everything from the instrument that decides that a 18 function is needed on through to the actuation of the 19 device.
20 MR. SHEWHON: He is talking about the testability 21 of it.- Everything but this one relay he can test on-line, 2? That's what I understood him to say.
23 MR. CHANG:- Basically that is because this is the 24 device right before the equipment is. going to actuate. At
- 25. the-beginning you have a sensor from, say, a transmitter, to ANN RlLEY & ASSOCIATES, Ltd.
Q'-- _
Coud Repoders l
1612 K. Street, N.W., Suite 300 Washington, D. O,- 20006 (202) 293 'l;50 1
1
1 I
1 348 l 1 an analog cabinet. So, you have bi-stable there. Then when 2 it's compared to the set point, you have the digital signal !
! 3 coming out of there into the logic cabinets.
j 4 Then whether it's two out of four or two out of 5 three you then get the reactor trip and it goes into the 6 master relay. Then it turns to the slave relay. You have 7 that piece of equipment whether it is a pump or a value.
j 8 What we are saying here is that normally you can 9 test the remaining portion. Sometimes it is an overlapping s
j 10 test. You can sometimes test the whole thing in one shot, a
11 but-in a lot of cases you can't.
12 MR. MICHELSON: No, not one shot. You can lift 13 leads and so forth to test various elements.
~
14 MR. CHANG: If the actuation of that particular j 15 equipment is a safety concern, it is not going to trigger 16 the reactor trip. In some cases, we can test the whole 17 thing.
18 MR. MICHELSON: You can also test the output relay
(
, 19 also, but you need to list leads to do it. You can't do it i
- 20 by leaving them all connected up. You can do it. There is
- 21 no question of that, I don't believe. The question is 22 whether-you want to go through the effort, and how
- 23 frequently you want to go through the effort, and I don't 24 disagree that maybe the frequency is too high, but-that is 25 an
- entirely different generic issue, I think.
' Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.)V., Suke 300 i
Washington, D. C. 20006
- (202) 293-3950 l
.,,9. sy .gg. -cg-e--y- -my,+ tw--
349 1 MR. CHANG: There are two things here. One is, if l
2 it is hard wired to the equipment, or because of the design 3 of certain circuits --
1 4 MR. MICHELSON: Hard wired only means that you 5 have some wires running between devices. I think you can 6 lift the leads on it. Sometimes you have to lift leads to 7 test in places. You can lift the leads on these, too, if ;
8 you have to. It is a little more work, and I think that is 9 what they do on some of these that insist on being tested.
10 MR. KADAMBI: All I can say is, this issue has a 11 long history, and I am not sure-exactly-what all 12 considerations went into it at all the different points in 13 the history. This is the way we defined the issue at the 14 point of resolution, and the background on it is in the
)
15 documentation that we have developed, and we have gone 16 forward, at this point, to the EDO and recommended that it 17 be considered resolved.
18 MR. WILKINS: Carl, there is a section in the INEL 19 report.
20 MR. MICHELSON: Which section?
21 MR. WILKINS: It is the section called Conclusions 22 and Recommendations starting on.page 75 in the lower right-23 hand corner. They at least --
24 MR. MICHELSON: Wait a minute. You are not I 25 talking about NUREG 9916 are you?
fl ANN RlLEY & ASSOCIATES, Ltd.
'~#- Coud Repoders
_ 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l
350 1 MR. WILKINS: It is page 49 of that document.
I O 2 They at least address these issues. This states 3 their conclusions. It doesn't give all the arguments and 4 all the logic and what-not that leads up to them. It at 5 least addresses them.
6 We never get these things far enough in advance to 7 study them, but at least I have thumbed through it for the
! 8 last several minutes here, and it seems to me that they have 9 made a credible case.
10 MR. MICHELSoN: I was only looking at what the 11 scope was supposed to be of a generic issue, and then 12 whether or not that scope had been resolved. I wasn't 13 looking at the --
/"'N 14 MR. WILKINS: A number of these paragraphs end up V 15 with language-like this. INEL concluded that GSI-120 is not 16 an issue of concern for something or another, and then for 17 something else or another.
18 In the paragraph that you see-that is that final 19 sentence, they talk about all the functions, ESF components, 20 and instrumentation.
21 MR. MICHELSON: You are on page 49?
22 MR. WILKINS: Yes. The control functions, and so 23 on.
24 Then there is generally a separate sentence for 25 relays, and components with staff approved exemptions for -
Q ANN RlLEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 j
)
351 I
- l 1 testing during pc9er operations.
]
l 2 MR. MICHE1 SON: Can you tell me what paragraph you l 3 are looking at?
- 4 MR. WILKINS: That is the very first paragraph on
- I
- l 5 page 4975, and it is towards the middle of the paragraph.
6 This is the paragraph that deals with CE and Werszinghouse l
)
7 PWRs, and there is another paragraph later on that page that 8 deals with GE BWRs, and some more CE and Westinghouse PWRs l
9 in this category, whatever this category is, I have l 10 forgotten. It explains that, too, at the beginning of the 11 paragraph.
12 They do put in the words that address the issues l
13 that you are raising. Now, I suppose you need to read the l
1 14 whole thing-in order to understand the logic behind those
[}
- 15 words.
, 16 MR. LEWIS: Well, we are not reading the Koran and l
17 trying to interpret it, or the Bible, it is their
- 18 responsibility to make the argument clear on how they have 1
j 19 made this chain of connection. I have missed it. I may be r 20 stupid, and we are now arguing about it as if we were trying q 21 to interpret a manifesto. We should not be in that.
22 position, and there are other things when we finish this l 23 issue, i
24 MR. CHANG: Let me say something here. What INEL
- 25 did is to, as I mentioned earlier, is to do a survey of 17
- l
{ ANN RILEY & ASSOCIATES, Ltd.
Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
1 J
352 t l
i i units of different vintages and designs. They did look into 2 the whole reactor protection system, not just slave relay,
- 3 and they ESF actuation system --
i
! 4 MR. LEWIS: Who is they?
5 MR. CHANG: INEL.
t 6 The survey did include the whole scope of 7 protection systems.
l 8 MR. LEWIS: We have to believe what you say INEL 9 reported.
1 10 MR. KADAMBI: I think we could give you a better 11 answer to the question, what-is the justification for 12 narrowing the scope. I think we should bs able to give you 13 a better answer than we have done this afternoon, yes. We 24 will do that.
15 MR. CARROLL: There seems to me that another 16 question is, what is the scope, what do you mean by a slave f 17 relay?
l 18 -I guess I am confused on that point, and I would 19 be helped a great deal by some examples of that.
20 MR. WILKINS: That is in your area of expertise, 21 why don't you tell us?
22 MR. WYLIE: I put relays into a protection system 23 that actuates the devices. The relays actuate the function, 24 like the high pressure injection. These are the relays that 25 open the valves and start the motors and what-have-you.
Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washinfon, D. C. 20006 (102) 293 3950 t
._ _ . _ , . , _ . , . _ _ . _ _ _ . , . ~ . . . . . . . - .
i j
353 1
- j. 1 MR. LEWIS: What makes them slaves?
i l l 2 MR. WYLIE: Well, that is a terminology that has i
i 3 been attached to them. They are auxiliary relays, e
4 basically, built into protection systems.
l
! 5 MR. LEWIS: Does it mean all relays, Charlie?
6 I am just asking you for information. Does it 7 mean all relays?
l
- 8 MR. WYLIE: It is all relays that actuate the 9 devices-that are necessary to,_say, create high pressure 10 injection, or low pressure injection.
1 11 MR. LEWIS: Are-there any non-slave relays in a-l 12 reactor?
13 MR. CHANG: Why they are called a slave relay is 14 because there is a master relay upstream. That is why it is
- 15 call slave relays, and the master relay usually is the relay
! 16 that takes the signal from the logic cabinet and, therefore,
- 17 for that particular trip, the master relay may_ control
- 18 several relays that are called slave relays.
l 19 MR. LEWIS: My thought was that a slave relay is.
l 20 called 3 slave relay because it is operated by another relay 21 but you didn't give that definition. _He did, but Charlie 22 didn't.
23 MR.-CARROLL:- I considered it part of the 24 protection system --
25 MR. MICHELSON: That's a different set,_ Charlie.
- ANN RlLEYL & ASSOCIATES, Ltd.
Q: - Court Reporters 1612 K. Street, N.W., Suite- 300 Washington, D. C. 20006 (202)L 293 3950
354 '
1 MR. LEWIS: Well, it would be nice to know.
2 MR. CHANG: As I mentioned earlier, a slave relay 3 as we are referring to it here are the small relays, usually 4 I think it's 120 volt, AC or DC.
5 It's not the big circuit breaker that actually 6 triggers the motor or pump.
7 MR. MICHELSON: Master relays can also be small 8 relays --
9 MR. SHEWMON: Gentlemen, what do we want to write 10 for a letter?
11 MR. MICHELSON: This should have been a 12 Subcommittee discussion.
13 MR. CHANG: Let me say something.
14 MR. SHEWMON: You said your thing. Let us discuss
(}
15 what we want to do, would you please?
16 Go ahead, Charlie or -- do we want to approve it 17 this way?
18 Do we want to ask them back again?
19 MR. WYLIE: What is your timetable?
20 MR. KADAMBI We have already sent the letter to 21 the EDO saying that this should be technically resolved, 22 considered technically resolved. If there are specific 23 questions or problems that you see in the Staff pcsition, we 24 will address them. As I understand it, I have a list of 25 specific questions you have asked.
O ^a" ai'ev a Associates' 'id-Coud Repodens 1812 K. Street, N.W., Suite- 300 '
Washington, D. C. 20006 (202) 293-3950
i 355 1 One is what is the benefit of staggered testing.
2 The second is the EPRI requirements document and 1
j 3 how the future plans are affected by the current requirement
] 4 in GDC-21, as I understand it. Tell me if I am correct.
l 5 MR. CARROLL: That's correct and I did read the 6 requirements document for the evolutionary plants and it's a J
! 7 lot of Motherhood. It basically is the position you 4
8 describe.
1 9 MR. KADAMBI: Right. The third one is what is the 10 justifica1r for narrowing the scope of the definition of 11 GSI-120 ft'.. the title "online Testability" to what we 12 actually did, which is look at the slave relays.
13 MR. MICHELSON: I think it was a little bigger 14 scope than that. Mainly, what was the intention of the 15 original generic issue 7 What was the issue -- which I don't 16 think you really ever told us today, what the original issue l
17 was.
18 They gave us a title but that wasn't quite enough, 19 and then from that you apparently narrowed the issue down 20 with some logic as to why it could be narrowed down. I 21 didn't get that logic real well.
22 MR. CHANG: Can I attempt to answer that question 23 and yours also?
24 MR. MICHELSON: -Okay.
( 25 MR. CHANG: I think I mentioned earlier I think LQ ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters-1612 K. Street, N.W., Suite 300 Washington, D C. 20006
-(202) 293 3950-i
i I
j 356 I that at the beginning the scope was kind of vague to us and
- 2 even though indications are that slave relays in a lot of 3 cases cannot be tested online such that they would not lead
! 4 to adverse effects, we don't know if there are any other 4
S things that would have the same concern in the reactor 6 protection system --
l
{ 7 MR. MICHELSON: Slave relays were -- go ahead.
8 MR. CHANG: That's why at the beginning we i
i 9 labelled it as the protection system online testability.
- 10 MR. MICHELSON
- You were given the issue. You i 11 didn't " label" it. You were given an issue, Generic 120, j 12 and what is that issue? Then how did you narrow it down to 13 just slave relays, because the issue wasn't slave relays to 14 begin with. It was testability. In fact, Mr. Minners'
- 15 letter points out that it's all the way through to the 16 actuated device you are supposed to be able to test, and 17 suddenly that all converged to " slave relays."
f 18 Now maybe there was a good justification but I 19 didn't get it.
20 MR. SHEWMON: We would like additional comments on j 21 it. Fine. You have answered all you can.
l 22 MR. CHANG: No, I haven't answered that one yet.
l 23 MR. SHEWMON: Well, please let's not try. There's 24 another question here.
25 MR. LEWIS: Yes, I have a completely different O ANN RILEY & ASSOCIATES, Ltd.
Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. . C. .'U. X)6
- (202).293-3950
357 1 subject because I think we have beaten this into the ground.
O 2 THE REPORTER: Mikel 3 MR. LEWIS I'm sorry, forgive me.
4 We've beaten this scope one into the ground and I 5 think you understand what our problem is.
6 All through this thing, ranging from the beginning 7 through the middle to the end and to the recommendations, 8 there have been comments about reliability.
9 You say operating experience shows that the slave 10 relays are very good and what you said was that they failed 11 once or twice or year in an average plant. That's 12 quantitative at least.
13 It says that they are generally more reliability 14 than the equipment they actuate. There have been all sorts
{
15 of things like that said through this. Is there any 16 quantitative backup to all these generalizations or are they 17 all " gut calls?"
18 MR. CHANG: That is all the information we have.
19 MR. LEWIS: So as far as you are concerned, 20 somebody else --
21 MR. CHANG: We do not have any additional 22 quantitative information.
23 MR. LEWIS: You do not or your contractors do not?
24 or you don't know?
25 MR. CHANG: Well, we didn't get into the 7 ANN RILEY & ASSOCIATES, Ltd.
(O Court Reporters 1812 K. Street, N.W., Suits 300
. Washington, D. C. 20006 (202) 293-3950
4 358 1 quantification part. Because of what we have learned we 2 decided not to pursue it.
i 3 MR. LEWIS: So these general statements that you
! 4 have made are your conclusions based on things you have 1
1 5 heard but as far as you know, there is no quantitative basis 6 for them? 71 that right?
7 Mk CHANG: That is a fair statement.
8 MR. LEWIS: Okay, but there are some fairly l 9 quantitative recommendations having to deal with quarterly 4
10 versus monthly, you know, rather quantitative..They have no 11 basis, you are saying? No quantitative basis that you know i 12 about?
13 MR. CHANG: Bob?
)
14 MR. TJADER: The two failures per year per unit.
j 15 That is a number that we have investigated and come upon, i 16 MR. LEWIS: But that is a numerator. To be 17 meaningful there has to be a denominator and something to 18 compare it to, 19 MR. WYLIE: But Hal --
, 20 MR. LEWIS: I am just wondering how far one is 21 going without ---
22 MR. WYLIE: Well, I would suspect the manufacturer 23 has --
R24: MR. DAVIS:' There'are databases used in PRA's and 25 if you look at those you'll discover that. relays are more O ANN RlLEY & _ ASSOCIATES, Ltd.
Court Reporters 1612 K. Street,- N.W., Suite 300 i Washington, D. C. 20006 (202) 293-3950
l 1
359 1
1 reliable than motors and valves.
j O 2 MR. MICHELSON: I don't think anybody would argue 3 with that -- if that were the issue.
i 4 MR. DAVIS: It's been that way for years.
1 l 5 MR. WYLIE: Yes, but that's what's being asked, l 6 what is the quantitative number.
4 1 7 MR. MICHELSON: Yes, but they are not necessarily 8 the most reliable devices.
j 9 MR. WYLIE: Well, I mean if that is the concern, 1
- 10 we should ask.
11 MR. SHEWMON: Do we want to have written comments i 12 on this or do we want to them come back again or what is the e
13 view of the Committee?
14 MR. WYLIE Well, we could have a Subcommittee
{ 15 meeting.
- 16 MR. MICHELSON
- I thought we just did.
- 17 MR. WYLIE
- We didn't think-we needed one -- if
, 18 this thing had been narrowed down just to the slave relays.
l l 19 MR. CARROLL: Except I don't even know what a i
20 - slave relay is-exactly yet.
21 MR. MICHELSON:. No,'I don't know what it is.
22 MR. LEWIS: There apparently was a concern by
. 23 - somebody at some. point that the intent or actually the 24- written words of GDC-21 which mentioned the reactor 4
25 protection system and.the engineered safety features were I ANN RILEY & ASSOCIATES, Ltd.
!] ' Court Reporters i 1612 K. Street, N.W., Suite 300
. Washington, D. C. 20006 (202) 293-3950 e--- , -+-e ,--m, m.,,-s,,. ,,w- ,,se , - , ,----,,m,,,r-~,~ ~ = m ,,- --<
l i 360 l
1 not being honored, that was something was happening in-the ,
2 licensing of new reactors in which this was being l 3 overlooked.
4 That has led to this can of worms in which we are t
5 wasting --
6 MR. SHEWMON: So what do we want to do now, and j 7 then we'll go off of this?
i 8 MR. LEWIS: I do not want to hear the same pitch 9 again. I could see some point to a Subcommittee meeting in i
10 which the people who generated the numbers that led to these 11 conclusions were brought in, but I'm not even so sure I want
, 12 that.
j 13 I think if we are on the hook to say that GSI-14 120, when this resolution is adequate, I could give an 15 answer right now. I would say no, you know, because we
- 16 haven't been given a rationale for the resolution of the GSI i 17 as it was stated and then it is'their job to go home and i 18- bring a better argument when they are ready, i
19 MR. WYLIE: I would suggest a Subcommittee meeting l 20 in January to flesh these things out.
21 MR. SHEWMON - Okay, fine. Good, let's take a ten 22 or fifteen minute break and then go back.
l 23 THE REPORTER: Dr. Shewmon, am I correct that that 24 is all that is reported for today?
25 MR. SHEWMON: As far as I can see, yes.
Q ANN RlLEY & ASSOCIATES, Ltd.
Court Reporters 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950
1 1
l 361 t 1 (Whereupon, at 3:32 o' clock p.m., the reported y
O 2 portion of the meeting was concluded.)
I 3 4
4 1
4
,1
! 5 i
I 6 4
l 7
, 8 1
l 9 i 10 i
l 11
) 12 i i 13 lO" 15 16 l
i 17 ,
s l 18
( 19 l 20 f 21
- 22 l
23
- 24 25 lO l-
^"" ai'ev * ^SSociaTes. d-Court Reporters i 1612 K.- Street, - N.W.,- Suite - 300 l Washington, D. C. 20006
( (202)-293-3950 p.
I
, ~ - . . . - , . . , , . - . . - , _ _ . , _, -
, - , _ . _ , _ . . . _ ,. __. . , . , , - , , , _ ., ~,.,.._..r., , _ , ,
i REPORTER'S CERTIFICATE t
[
l l
This is to certify that the attached proceedings before i the United states Nuclear Regulatory Commission 1
In the Matter oft l NAME OF PROCEEDINGt 392nd ACRS 11ecting i
4
! DOCKET NUMBER:
PLACE OF PROCEEDINGt Be thesda , 11aryland i
wt 1 held as herein appears, and that this is the original transcript thereof for the file of the United 3 States Nuclear Regulatory Commission taken by me and thereafter reduced to typewriting by me or under the i direction of the court reporting: company, and that the
- ! transcript is a true and accurate record of the foregoing proceedings.
Am_s Sh i / '
} Official Reporter i
Ann Riley & Associates, Ltd.
i s
f
.p , , - - , .. .r , ~ , . . _ _ _ , , _ _ , . . . . , , , , . . _ - . , , ,7 .. ,-.
O. O O RES STAFF PRESENTATION TO THE
- ' ACRS FULL COMMITTEE l
i
SUBJECT:
- GENERIC SAFETY ISSUE 120, ONLINE TESTABILITY OF
- PROTECTION SYSTEMS
- i. '
DATE: DECEMBER 10,1992
! PRESENTER: T.Y. CHANG PRESENTER'S TITLE / BRANCH /DIV: TASK MANAGER, ElB/DSIR L i'
', PRESENTER'S NRC TEL. NO.: 492-3922 j SUBCOMMITTEE: CONTROL AND ELECTRICAL POWER SYSTEMS
o
~
O .o ~!
GSI-120 EXISTING REGULATORY-REQUIREMENTS AND GUIDANCE e GDC 21 REQUIRES AT-POWER (ONLINE) TESTABILITY OF PROTECTION SYSTEM FUNCTIONS, lNCLUDING CAPABILITY TO TEST CHANNELS INDEPENDENTLY e , PROTECTION. FUNCTIONS / SYSTEMS
.TO AUTOMATICALLY INITIATE REACTIVITY . CONTROL / REACTOR
-PROTECTION SYSTEM (RPS)
TO SENSE ACCIDENT CONDITIONS AND INITIATE ENGINEERED SAFETY FEATURES /ESF SYSTEM l e SUPPLEMENTARY GUIDANCE PROVIDED lN: -
SAFETYcGUIDANCE 22 (REG. GUIDE 1.22)
REG. GUIDE 1.118 i
.lEEE STANDARD 279-1971/338-1977 i t
_1_
N- -
~
OL O. O SCOPE OF GSI-120 SCOPE OF'.GSI-120 IS THE OPERABILITY OF ESFAS SLAVE RELAYS FOR WHICH ONLINE TESTING COULD HAVE ADVERSE IMPACTS SLAVE RELAY: OPERATION THAT CAUSES A TRIP OR SYSTEM ACTUATION e INAPPROPRIATELY DUE TO HARDWARE FACTORS e INADVERTENTLY- DUE TO OTHER FACTORS lY
~ .
llNEL STUDY AND FINDINGS-e SURVEYED 17 PLANTS (UNITS) OF DIFFERENT VINTAGES AND DESIGNS:
8 RELATIVELY'NEW PWR (OL- DATE AFTER 1982) 7 POST-SEP PLANTS WITH OL.DATE BEFORE 1975-SRP 1 -PWR AND 1 BWR NOT IN ABOVE GROUPS e ' REVIEWED PLANT-SPECIFIC INFORMATION FROM:
- UFSARs TSs SERs DISCUSSIONS AND INFORMAL ' CORRESPONDENCE WITH PLANT PERSONNEL-
_ __=__ =____
O O O!
'INEL STUDY AND FINDINGS (CONTINUED)
- e: INEL. FINDINGS IN NUREG/CR-5916:
[ - FOR. ALL 17. PLANTS -REVIEWED, EXEMPTION REQUESTS FOR ONLINE TESTING WERE REVIE"'ED AND GRANTED BY NRC STAFF ON A CASE-BY-CASE BASIS PER SAFETY GUIDE 22 l EXEMPTIONS TO.ONLINE TESTING NEED TO BE ADDRESSED AND i REVIEWED CASE-BY-CASE FOR FUTURE PLANTS .
.j l
4 4
i.
j.
I i
[ I i
i l
l t I l I
LO O O STAFF STILDY AND FINDINGS
- FINDINGS:
ONLINE TESTING OF CERTAIN SLAVE RELAYS IN PROTECTION SYSTEMS IS SIGNIFICANT CONTRIBUTOR TO INADVERTENT ACTUATION OF SAFETY EQUIPMENT AND REACiDR TRIPS OPERATING EXPERIENCE INDICATES RELAY REllABILITY IS
-GENERALLY GOOD i
SLAVE RELAY RELIABILITY IS' ESTIMATED TO' BE ONE ORDER OF-MAGNITUDE HIGHER THAN MOV RELIABILITY
~
- O O O-STAFF. STUDY AND FINDINGS (CONTINUED) t 4
- PROPOSED NUREG-1366 AND GENERIC LETTER:
CHANGE ONLINETESTING FREQUENCY OF SLAVE RELAYS TO:
- 1. QUARTERLY, AND 2.'ON A STAGGERED BASIS OVER A FUEL CYCLE i -
LEAVE TESTS CARRYING THE HIGHEST RISK-TO A REFUELING 1 OUTAGE OR COLD SHUTDOWN i 1 PLANT-SPECIFIC ANALYSIS REQUIRED TO IDENTIFY SLAVE RELAYS THAT SHOULD BE' TESTED 'DURING REFUELING OUTAGE / COLD l SHUTDOWN -
}
l
. EXEMPTION WILL BE' ALLOWED ONLY WHEN PLANT-SPECIFIC g RELAY EXPERIENCE IS CONSISTENT WITH RELAY EXPERIENCE IN .
[ NUREG-1366 .
F -
. LICENSEES / APPLICANTS CAN VOLUNTARILY REQUEST TS CHANGES
- l. ;
1
~ ~
.O O O .
RATIONALE FOR RECOMMENDING TECHNICAL RESOLUTION TO GSI-120
- e. OPERATING EXPERIENCE INDICATES THAT SLAVE RELAYS ARE GENERALLY MORE' RELIABLE THAN THE EQUIPMENT THEY ACTUATE- 1 i
e SAFETY BENEFITS'OF ONLINE TESTING OF SLAVE RELAYS SHOULD BE BALANCED AGAINST RISK ASSOCIATED WITH INADVERTENT ;
ACTUATION OF SAFETYf EQUIPMENT.OR REACTOR TRIPS e- STAFF RECOMMENDED THAT FREQUENCY OF ONLINE TESTING OF SLAVE RELAYS.CAN:BE' REDUCED.UNDER SPECIFIED CONDITIONS r i i
' i I
1 j !
4 . _ _
lO APPLYING THE
, SYSTEM RELIABILITY PROGRAM TO IMPROVE RELIABILITY AVAILABILITY
, AND
- PERFORMANCE
, AT RIVER BEND STATION 1
!O I
BY l O. A. Holbert Supervisor Reliability Engineering GULFSTATES UTILITIES RIMR BEND STATION ST FRANCIS MLLE, LOUISIAN/
i O PRESENTATION TO THE ACRS DECEMBER 10,1992 WASHINGTON, DC
4
!O System Reliability l Program.
i i Outline i
i
- Introduction l
- Objectives of the System Reliability Program i
- Overview of the Program O
l
- Benefits of the Program l
l
- Lessons Learned 1
- Cost
- - Staffing l - Computerization
- Future Uses of the Program l
1 l
T m - e+.e<y w -e-- .e- r. -. , mw+w-- , v w v
o l
System Reliability O l Program
- Objectives i
f l
Enhance Maintenance Effectiveness t
j Improve System Reliability and Availability lo j Show quantitatively how changes effect System
! Reliability and Availability Allocate resources in a cost effective manner i
j Imalement a Living Program
- Provide plant specific data for PRA iO
-r- , ,. ,., ..,., , y-,.-- ~ , e. y , -,n.,,
[g System Reliability
- Program PrincipalElements
- Identification of Critical Systems / Components l
l
- Identification of Significant Failure Modes I
l
- Quantification of Conditional Probability O
l
- Component and System Availability
- Preventive Maintenance Program Review:
- Evaluate Adequacy of Existing Practices
- Recommendations for PM Changes l
I
- Imalementation of PM Changes l
C
- Feedback and Monitoring l
i O SYSTEM RELIABILITY i
PROGRAM i
j Products Produced l
l
- Avaif a Jility Analysis
\ '
l
- Availability Trending lO
- Preventive Maintenance Recommendations l
- Plant Specific data for PRA e
i i
I O
4
- - , --r,. . -- .. .
System Reliability i
Program
- Systenj Scope:
- 1. Feedwater System -- (PRA/O&M/ Availability)
- 2. Normal Service Water -- (PRA/O&M)
- 3. HVAC System -- (PRA/O&M)
- 4. Reactor Water Cleanup System -- (O&M)
- 5. Circulating Water System -- (O&M)
- 6. Reactor Core Isblation Cooling -- (PRA)
- 7. High Pressure Core $' pray --(PRA)
- 8. Diesel Generators -- (PRA/O&M)
- 9. Residual Heat Removal-- (PRA)
O 1. Fire Protection System -- (PRA)
- 11. Closed Cooling Water System Reactor -- (PRA)
- 12. Closed Cooling Water System - Turbine -- (PRA/O&M)
- 13. Condensate System --(PRA)
- 15. Heater Drain System -- (O&M)
- 16. Instrument Air System --(PRA/O&M)
- 17. Low Pressure Core Spray -- (PRA)
- 18. Main Steam System & Safety Relief Valves --
- (PRA/ Availability)
- 19. Main Turbine -- (PRA/ Availability)
- 20. Radwaste System -- (O&M)
- 21. Reactor Recirculation System -- (O&M)
- 22. Rod Control And Information System --(O&M)
O 23. Standby Liquid Control System -- (PRA)
- 24. Standby Service Water System -- (PRA)
4
!' SYSTEM RELIABILITY o
PROGRAM 9
RCIC '
- STANDBY AVAILABILITY 100% r p 3 & ;
q ,!
[
t
- h. l' h; *nd !
h 95% jh .
- b* *f Y *l
~ '
90 % - 1l.***((L *g ;i
~ ' '
l 3: -
! f:.
- g ... (. b i j c
l [ .
][' r :
j- -p I fj [ hc I I
~
- 85% 1 -l L
- i;;
I' b Q ] l f , f ;
f:- [- j 80% i-l-l ;.
I'
- F j;
~il.r. -[; l: N:
- l ;.
t ,
E
- i
- t. 'h. 2 .
r: ;
i
! i! "j (['
~
.p GMONTHLY AVERAGE *12 MONTH AVERAGE l' 1 O f; -
[ [ ; Ik dI E h . {) f; 70 %
' i S'YSE*Y'II'lSJ2 ' t #' I * ' 4 i ' '
DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND l
88 89 l 90 l 91 l 92 l l INCLUDES PLANNED & FORCED MAINTENANCE HOURS l Began Analysis i
l O
l 5
E___ .,y , . , , , . . . _ _ , . . . . . . _ _ _ . - , - , . , . ,
. 6m. .% k 5e.- A4 4
- 4 ik E. +e.r'*a4.. Ww D 4+-w. ..i sah + A '- 4- 4 4A 1. -u Em.est,1 45.b4--J4 h- - 4*e%a at Rau-- S k O O O O CONTROL BUILDING CHILLERS CHILLER 1A 12-M0 AVG AVAILABILIW CHILLER 1B 12-MO AVG AVAILABILITY ims MAAAAAAAAAAAAAA>hn'a AA imsfgmAA! ' '-
paggy
/
3
,a -
m .
W-y sa sa
.m.-...
m -
m a ft.
m auanxuunnauuuuututu2utu LuuLLa m a us
- n * ' c# '
- a LU1A UMD t 1
- i e s 1 Unllu1Hulu, DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND og 89 l 90 l 91 l 92 84 89 l 90 91 92 l l CHILLER 1C 12-MO AVG AVAILABILITY CHILLER 1D 12-MO AVG AVAILABILITY ims
= '1;J Ju % g i=NJng.
m 9
% = *
/
n .- -l\
,. ,. l\.a l *
- 8. . .
__g8 :_ '
- m .
so,.
/
_ _ s - a-y -
m .1 LLLL L L.LLLLL Lk il d L L.1 LL LLLiiLM LL11 LLLLL.LULLlL. m ALLk1DD1LLLlLL A LllL LL1 L L A LA LisJLL1 Ll LL Ll.11 LJ L DJFMAMJJASONOJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND 84 89 l 90 l 91 l 92 84 89 l 90 l 91 l 92
A .c & h .,.m.... t. AAE 4 4& hL+- .AM.-4A. 6,u -+%-i% &A-a a 4 W4. + a , 4 a_mA.. a L.mu.h_, Ag. m. .A4.a & 4 a _m L e g4.,
O O O O TURBINE BUILDING CHILLERS TB-CHILLER 1A 12-MO AVG AVAILABILITY TB - CHILLER 1812 MO-AVG AVAILABILITY seJ uva;;;uuuua;;;ua;;;uu;a;a;;;a- )- 'Om wwg ,A :::RL :
9a =
MkAkkkkk k k 80% 80% --
70 % 70%
60% 60% ,
50 % 50%
(%
40 % 40%
g m tiiae! ieeairiii+ sea, .:i! i e1 i i i i # et 11.Llu22,, m 2 eia e i LL L '
- I i's t a I t LLLA. 8 a
- i i iii****f n i n.1.L.111 th DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND of 89 l 90 l 91 l 92 M 89 l 90 l 91 l 92 TB CHILLER 1C 12-MO AVG AVAILABILITY sw% mgp -Ja-90,., w
- MI 8.
70%
60% -
50% -
4og . . _ _ _ . . . _ _ _ . _ .
t y .LilLLLAL LLL L L.L. L.LLL L L A .LL L.Ll.1.1.L.L.. L.ll L1 M L .L L l L Al. L.11 i DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND 80 89 l 90 { 91 l 92
SYSTEM RELIABILITY
- o l PROGRAM l How data is obtained l
l
- Control Room Logs are reviewed daily.
l Reliability Engineering evaluates all Electrical and Mechanical PMs, all Condition Reaorts, all Clearances, and all Maintenance Work
- Orders (MWOs) for each analyzed system.
- O l
- Nuclear Plant Reliability Data System l (NPRDS) failure reports are reviewed.
- Emergency Response Information System (ERIS) Computer information, consisting of 100 selected plant parameters that are monitored during plant operation.
Additional data is gathered by Shift Technical O Advisors on a daily basis during power
.: operation.
1 l
O NORMALIZED CORE DAMAGE FREQUENCY J
l 1.1 j
- 12 MONTH SLlDING OMONTHLY VALUES !
l 1
I G
~
~
l' 0.9 W[ F-
- . .4 O .r-(****i.c-~~'r , -
!L r'
l-N' '.;
F I
i O.8 r: ff
)
s ( ,
i i
! i r
I I ' ' ' ' ' ' ' ' ' l ' I L ' ! I i ' ' ' ' ' i t i i i t r i i i i i i>
03 ' i r ?1 i DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
- 88 89 l 90 l 91 l 92
, BASED ON PLANTVALUES FOR RCIC, HPCS AND i DIESEL GENERATORS; OTHER VALUES ARE GENERIC.
Approved &
O s"nervieo7ea^
V:tHGDATALHGCHTiCOF. PRS
]
a
.- . ., , . . . ~ . - - a -
.r, ,r, ,
4 !
o System Reliability
! Program i
- Recommendations for PM Chanees
- PMs are either added, modified, or deleted, based on the Critical Component Analysis and l component evaluation using the EPRI RCM
- Workstation.
iO l
Surveillance Test Procedures (STPs) are examined and recommendations are l generated.
! Equipment Qualification (EQ) PMs are i
reviewed and recommendations are generated.
4 IO
l0 SYSTEM RELIABILITY
- PROGRAM RESULTS i
PROGRAM IMPLEMENTA TION BY MAINTENANCE i
A Maintenance Planners have implemented i changes on 8 of the 10 completed systems j
- Reliability Engineering has made 1221 O recommendations to the PM Program
- Of these,80% have been implemented O
PM Recommendation o
. Summary l l ResidualHeat Removal System 4
i Type of -
. Total Recommendatioir O No Change 109 PMs l
Change in Scope 6PMs and/or Frequency 4
^O J
1
- o System Reliability i
Program l Benefits of the Program l Lessons Learned
- Cost
- Staffing jO - Computerization Future Uses of the Program i
4 1
J
I-
!O r
! l i
USING PRA METHODS FOR !
- ENHANCED RELIABILITY AT l l
RIVER BEND STATION i
i r l
i BY O
j J. L. BURTON, P.E.
j SUPERVISOR - PRA GULF STATES UTILITIES RIVER BEND STATION ST. FRANCISVILLE, LOUISIANA PRESENTATION TO THE ACRS DECEMBER 10,1992 1 WASHINGTON, DC O l 1
l-10 4 USING PRA METHODS FOR ENHANCED RELIABILITY AT l
RIVER BEND STATION i
l
- Introduction l -
Uses of PRAat RBS l
Integration of PRA with Severe Accident Issues RBS PRA Summary l.'
i O
- Use of PRA for Reliability / Maintenance Rule l
l l
Important Systems, Components, and Operator l Actions Critical Component Ranking for PMs Availability vs. Core Damage Frequency -
j -
Monthly Trending of Risk j -
PRA and the Maintenance Rule i
- Conclusions i -
Summarize Need for NRC Support of Risk-Based Resolution of Issues
-, . , . . - . . . ~ - . . - . . - . - - - , ~ - - - - , , . , . _ . . ._ . . - ,. . . . . . . ~ .
4 l
lO l
- Improve decision making
- Prioritize flesigD-change requirements
'
- Prioritize PM requirements O
i
- Enhance safety awareness of utility personnel t
Improve normal and emergency operating and maintenance procedures
- Maintain safety margin during shutdown l
i
- Support Waivers of Compliance and Technical l Specification changes O
- Establish risk significance of events i'
e + r- ,, -w - aw -- .---%w -msi,, s + m +m un~- m a _ . ,m__ .u_m__.
4 O
l INTEGRATION 0-F PRA WITH l SEVERE ACCIDENT ISSLES i
River Bend PRA Group is responsible for:
f -
Individual Plant Examination (IPE)
IPE for Extern #hl Events (IPEEE) i PRA Applications
!O -
Station Blackout (SBO) Analysis i
i Hydrogen Control Owner's Group (HCOG)
Analyses Accident Management Guidelines (AMGs)
Emergency Operating Procedure (EOP)
,. Technical Bases Analyzing Radiological Consequences of Transients / Accidents Shielding Analysis O l l
1 i
l I
!O
SUMMARY
i i
i j
- Focused Approach Reduces Model Size and Cost,
- but Still Provides Useful Decision-Making Tool
- Internal Floodfng Analysis
- Detailed ATWS Analysis
>O i
I
- Interfacing System LOCA Analysis i
.
- USI/GI Analysis i
- Containment Performance Improvement Analysis I
O I
,- ~ --
4 i
!O 1
i
'
- Linked Event Tree - Fault Tree Method l
14 Front-Line gystem Fault Trees
- ,.
- 6 Support System Fault Trees
!O f
- 16 Initiating Events i
i 250 Operating & Maintenance procedures were j reviewed to identify human actions
- 81 Human Actions included in PRA l,,
- 210 values for data input (generic and RBS-specific)
- O
+
w4,-w. a ---e- .s.. =a-.e- - -
y4 y a = b
i O O O -
i i
l RIVER BEND CORE DAMAGE FREQUENCY BASED ON ACCIDENT CLASS ,
4
_e-rq ~
f l 'N-:b_55
- I 15E;asm .
.i -
l 1 r l [__
,, ]Il{
! i,
) i 55 l
i l { xx- gg --
)l R
==
i l$
l k .
l l
l l
I -
E suO - 86.4%
$!! Transients - 7.9%
L_l LOSP - 4.2%
l l Iransient induced LOCA 1.5%
fhver Bend CDI - 1.SSE-5/yr !
i
l 4
iO l IMPORTA:NT SYSTEMS, l COMPONE:NTS, AND l OPERATOR ACTIONS i
i
- The most important systems affecting the core -
damage frdquency are:
Offsite Power l
0 -
oiesei generators Standby Service Water Reactor Core Isolation Cooling DC Power I
Standby Switchgear Room Cooling High Pressure Core Spray Safety Relief Valves
- O i
l IMPORTANT SYSTEMS,
! COMPONENTS, AND l OPERATOR ACTIONS l
n l
- For River Bend, there is no single system, component or operator action which dominates O the core damage frequency l
i
! l l
l l
l O 1 f
w, - - e., ,- , , - , . . , - - - , , ,e - , - - - -, a_-,-.-,,,,,e,,,,-.m,.,- ,m ww ,e- w- ~m----.-~-w- < - , - - -
l.
IO CRITICAL COMPONENT RANKING
! FOR PMS 1
i
- Use CAFTA to provide data on SSCs:
Probability
) -
Importance l (Fussell-Vesely, Risk Reduction, etc) i i t Develop Importance (F-V) for each SSC:
j' -
Fails to Start i -
Fails to Run
- O .
Maintenance unavaiiabiiity j -
Fails to Open Fails to Close Loss of Function l
Common Cause Failure j Rank on Total F-V Importance Normalize rankings on lowest F-V Importance (makes rankings more user-friendly)
- Provide normalized rankings to System Engineers (1st step in new PM Optimization Program)
,-.----+w-- , , , - -
o PM OPTIMlZATION PROGRAM PM Screening
- Risk
- ~ ~ ~ ~
e Output !
'e Compliance i I e 60st i l
l 1
l l 1
O i
< l i
1 r l Focused PM i Program i -
i Feedback
- Predictive Maintenance i e Root Cause Analysis i ,
I I
I I
I I
3 r l l
Reliability I O Monitoring
.. ... m-__-... .._-.. ._ _____
4 O
AVAILA'B ILITY VS CORE DAMAGE FREQUENCY BASK
- Each plant is unique. Core damage contributors are specific toindividual plant configurations and operating practices 1O
- PRA provides the best available method to quantify and evaluate:
the interactions of critical plant systems the impact of safety system unavailabilities on overall risk (CDF)
t l
y - -,e ,,,..w- .--,aw, , -,,,--,,,,--,,,+w-,- -
-, e- , y- ,-
~
'l ~. - . .
O: O O-CORE DAMAGE FREQUENCY VS. HPCS AVAILABILITY
' 7.OOE-05 g NN ..
. 6.OOE-05 \u N, .
PHA Level 1 Availat).hty = 37.5%
N I
T
^ 5.OOE-05 ---
"s' ' N Nx
$ i i moS "x. '
a
_e..
w ~. s e
'a 3 - -
a-g . 3.OOE N,' x ,
s.,
2.OOE-05 "..,""e
,"o e
1.OOE-05 -' l : -- - - l - - - - l - - - - --l - '-I--- .i-- i i --I 0% 10% 20 % 30% 40% 50 % 60% . 70% - 80 % 90 % 100 %
. HPCS Avadatulety
-- - - =---- - -
___.__.--_..-m.;,. . . . -
i' l
4
!O i
MONTHLY TRENDING OF RISK l
1
, Five Safety Systems are used:
i j -
l -
Emergency Diesel-Generators A, B, and C l
l,-
lO M nthly and 12-Month Sliding Averages are used:
l Standby Unavailability l -
Failure to Start i
Failure to Run i
i
.. All other data is based on current Level 1 PRA
, revision
, l
4 O
NORMALIZED CORE DAMAGE FREQUENCY 1.1
- 12 MONTH SUDING CMONTHLY VALUES ;
1
\ -
E I - .
~
/'
o,g L. _ _ _pM** -
~.
\
o ...
%*;, . . . .r
~
l ,,
- ** b t .
l .
~
I -
I - . .
i . .
0.8 b- ,,, j , q I H i i aq ! ) *
'- 1 l'
n ;
Q ..
4
!! t 1 i 6 b !
l i i
, ! l a , ,
l 0.7
'Il ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND 88 89 l 90 l 91 l 92 BASED ON PLANTVALUES FOR RC4C, HPCS AND 01ESEL GENERATOR 8; OTHER VALUES ARE GENERIC.
Approved Gdr'
(<]' Supervisor'YRA WNGOATAWGQmCOP.Pfle
1 O
PRA AND THE MAINTENANCE RELE DRAFT NRC REGULATORY GUIDE (3/24/92) ON MAINTENANCE RULE RECOMMENDS:
.e
- 1) Use PRA to estimate core damage frequency (CDF) due to systems, structures, and components (SSCs)
VJ e SSC Importance Ranking
- 2) Use PRA to establish reliability / availability goals for accident mitigation systems
- CDF vs. Availability
- 3) Examine PRA cutsets for:
e Single failures e Common cause failures O
e*
4
l:
lo
- 4) Identify plant risk due to initiating events e RBS PRA Results
- 5) Evaluate risk due to maintenance on SSCs e On-line MWO evaluation l
l e Control pfant configuration to eliminate high-l risk combinations of SSCs using:
a) Guidelines List high-risk equipment combinations for Shift Supervisors, Maint. Planners, etc.
b) Root Cause Analysis Use MORT, events and causal factors, etc. to identify potential precursors,
, initiating events, etc.
c) PRA Perform detailed risk analysis O
A ~- ~ * - +,,,,w,, -
O CONCLUSIONS River Bend's PRA model provides a useful tool for:
Ranking SSC importance for risk-based PAI optimization Setting risk-based goals for SSC availability Trending risk significance for SSC availability Q -
Developing Maintenance Rule goals and priorities Evaluating risk significance of changes in plant configuration River Bend plans to use PRA to help resolve regulatory issues
- NRC concurrence with risk-based regulation methods is needed on issues such as:
Tech. Spec. Improvements Thermo-Lag Risk Significance
O RES STAFF PRESENTATION TO THE ACRS
SUBJECT:
PETITION FOR RULEMAKING PRM 50-56 DATE: DECEMBER 10, 1992 PRESENTER: CHRIs ROURK O AFFIL.: TASK MANAGER, RES/EIB PHONE: 492-3938 e
O
i O PETITION FOR RULEMAKING i
PRM 50-56 I
i i
! O Petition requests rulemaking ;
i to require licensees to -
i
- determine the effect on all L
electrical systems (including i
i O BOP) from every source of l
i l
! electrical transient.
l l
l o Sources mentioned include l nuclear EMP, lightning, electromagnetic interference lO (EMI).
2 I
1
~ . _ _ . . - , . . _ . . , _ . . _ _ . . , . . _ . _ _ _ . - - _
O o Petition requests a comprehensive study applying PRA methods to a sample of plants.
O i
O 3
O STAEF'S ACILQNS o Federal Register Notice on the PRM published on December 23,1991.
o Public comments requested by O February 21,1992.
o Staff completed technical study September 1992.
o Schedular guidance from EDO requires resolution by O December 1992.
4
i i
O o Staff proposes to issue a i
i letter denying the PRM, i
but will identify on-going l
I actions in this area.
i i
lO 1
i l
l i
,t O.
0 s
i i
- 3 lO EURLI_C CORME_NTS o Four comments received; two individuals, one consumer group, and one utility.
l o Three comments favored rulemaking and one opposed.
O l
o Main points from Public Comments:
1 Increased likelihood of nuclear weapon, attack requires I hardening of power plants.
O.
6
(--___--_- - -.-._- _ - - -_ - __ _ _ __
l l
!O 2 Delicate nature of electronic l
I components requires regulation of protection by the NRC.
i 3 IPEEE program should cover lightning protection for all i
i O l plants.
i 4 The NRC should develop guidance before any utility actions are required.
l O
7
l O CUBBENT REDUIREMENTS o GDC contain requirements to i
h consider the effects of electrical l
i transients of the type mentioned l
i l in the petition.
l
!O t /
I o GDC 2 requires that structures, systems and components (SSCs) important to safety must be designed to withstand the effects of natural phenomena.
O 8
i i
Oo GDC 4 requires SSCs to be I
compatible with the operating l
l environment.
l l
o Industry Standards require i
! electronic equipment to be
\ '
!O
! compatible with the intended operating electromagnetic environment.
't O
9
l i
, O pgsy. NRC ACTIONS o NRC previously studied the effect of EMP on nuclear power plants.
L o NRC previously developed draft
!. Regulatory Guide RS 705-4 on ;
i lO lightning protection. The ACRS requested a risk assessment on l
i l
February 3,1981, which indicated i l l I no significant safety benefit.
l l 3
l l l
lO
! 10 .
4 1
O o The staff's ticensing reviews included consideration of industry standards for plant structural protection, switchyard shield wires, grounding of equipment, shield wire grounding for control O system wiring, etc.
O 11 1
. - - --- -.-- -_- - _ - __----_ - .=
i i
i i 3
!O o NRC licensing review of operating I
plants for conformance to GDC 2 and 4 in regard to electrical l transients was based upon:
i I
1 Use of established industry i
O standards and practices in the i
l l plant design, 1
i 4
a
!O 12 v- -, - --7---w --r-,--,-,-,---,--,----,.-,-w- , w-,,----- -
~.-w.-, - -.-eu,,.,m-..- . _,- _ _ _ _ ..__ m ,,- =c-.._ w a . - - - - + + - - - , --.i< * - * * -- * - - = - = =
i l
!O 2 satisfactory performance of 1
l electrical systems and components in similar electromagnetic i
environment and industry i
application, and i IO 3 Additional qualification testing l
s
- required of solid state components
! and systems (e.g. Eagle 21).
l i
O 13
lO 1
RUBRENT PRACTJ_CE 1
o Lightning and line switching i
! transients have caused 1 transformer and generator failure in the past. Lessons were learned l from these events before use of t
!O protection became widespread.
l o Current design practice provides for mitigation of electrical transients.
O 14
l
'O o Examples include transmission l
line shield wires, surge arresters, capacitive grading for transformer h
i windings, instrumentation and control cable shielding, coaxial cable, station ground grids.
- O l 0 Electrical transients still cause i
(
l problems in residential and i
l industrial use, where widespread i ,
\
l protective measures might not be
- .s l
economically justifiable. i i
!O 15 i
1 lO i'
BEVIEW OF OPERATI_N_G EVENIS l
INVOLVING LIGHTRINQ -
1 a
i i
o 152 LERs containing 177 events for l
the period 1980-1991 were reviewed (some LERs reported IO multiple events).
l o 10 of these events were considered in the Accident Sequence Precursors Program. Conditional Core Damage Frequency ranged from 9.0E-4 to 7.0E-8.
O 16
c.
l j
O o The most significant events
! involved fire or damage to sensors causing reactor trips.
) o The LERs did not indicate that I
4
)
i these events would have been iO prevented by different structural l
l l or power line protection.
l l
! o No events resulted in misoperation or failure of a safety-related m
system.
lO s
{ 17
.,,_-_,..m.,. , , . . . , _ _ . . . _ .
. , , . . . . _ . _ . , . . . . . . . . _ . _ _ . , , _ , _ . _ _ ~ . . . _ , _ . , _ _ _ _ _ _ . . _ _ _ _ , _ . _ , _ _ . . . _ . . . . -
l l
O THE IPEEE REVIEWS i
i o Staff position is that the primary impact of lightning is the loss of l
offsite power (NUREG-1407, April, i
1991).
O l
l o Site-specific consideration may be l
warranted where, based on l
! operating experience, lightning strikes are likely to cause more than loss of offsite power.
O 18
-*w wee, r +-w~ e-, ...m---,-.. , . . - - -m.,-m__.mvm, , - -m--. , , __.__________._mm_
q
! . l 8
]
4 1
i l i
O o The IPEEE review team has been I
notified of the findings in the staff report.
1 !
O I
t' l O 19
l0
_ CURRENT STAFF ACTIONS l o The NRC is: L
- l l 1 developing a regulatory guide on protection from EMI, l
a I
l 2 reconsidering development of a 1
l regulatory guide on lightning O protection and reviewing industry standard NFPA-78 on lightning 1
protection, and 3 augmenting guidance to the staff for review of futura plant designs.
O l 20 l
l
i i
l i
! o s
.C. ONCLUSIONS l
0 Rulemaking is not needed at this time.
l l
! o Currently-operating analog and c
l digital systems are able to
- withstand electrical transients.
I
.-(
i l0 21
. - . . ~ . . . _ _ . - . . , . . . . . . _ . _ _ _ . _ _ . _ . . . _ _ . . . . _ . . . . . _ . _ . . . , _ _ _ _ . . . _ . _ . _ _ . _ . . _ _ _ . _ , . . . . . . - . _ .