ML20155K450
| ML20155K450 | |
| Person / Time | |
|---|---|
| Issue date: | 06/14/1988 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1675, NUDOCS 8806210222 | |
| Download: ML20155K450 (414) | |
Text
'
nce9=M73 D i a_\rA ._
UNrrED STATES O '
NUCLEAR REGULATORY COMMISSION ADVISORY CO.24ITTEE ON REACTOR SAFEGUARDS
)
In the 14atter of: )
)
SUBCO:U4ITTEE ON RELIABILITY ASSURANCE )
)
I l
l O
/
I l . r1 .
x u 4:J 2,'. Au . . a.s 1
I DATE: June 14, 1988 LOCATION: Washington, D.C.
PAGES: 1 through 292 1
l l
l HERITAGE REPORTING CORPORATION l O wn-1224 L StMet, N.W.,2 600 I
WaeWagton D.C.20005
) egg 62joggase0614 (282) 628-4888 T-1675 PDR
1 1 UNITED STATES NUCLEAR REGULATORY COMMISSION 2 ADVISORY COMMMITTEE ON REACTOR SAFEGUARDS 3 )
In the Matter of: )
4 )
)
5 SUBCOMMITTEE ON RELIABILITY )
ASSURANCE )
6 )
7 Tuesday, June 14, 1988 8
Room 1046 9 1717 H Street, N.W.
Washington, D.C. 20555 10 The above-entitled matter came on for hearing, 11 pursuant to notice, at 8:35 a.m.
12 BEFORE: MR. CHARLES J. WYLIE 13 Retired Chief Engineer 0 14 Electrical Division Duke Power Company Charlotte, North Carolina 15 ACRS MEMBERS PRESENT:
16 MR. CARLYLE MICHELSON 17 Retired Principal Nuclear Engineer Tennessee Valley Authority 18 Knoxville, Tennessee and Retired Director, Office for Analysis 19 and Evaluation of Operational Data U.S. Nuclear Regulatory Commission 20 Washington, D.C.
21 DR. CHESTER P. SIESS l Professor Emeritus of Civil Engineering 22 University of Illinois Urbana, Illinois 23 ACRS COGNIZANT STAFF MEMBER:
l 24 Richard Major 25
/~ <
N ;b HERITAGE REPORTING CORPORATION -- (202)628-4888
7s, 1 -PUBLIC NOTICE BY THE
' (f 2 UNITED STATES NUCLEAR REGULATORY COHMISSION'S 4
3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 4
5 6
7 The contents of this stenographic transcript of the 8 proceedings of the United States Nuclear Regulatory 9 Commission's Advisory Committee on Reactor Safeguards (ACRS),
10 as reported herein, is an uncorrected record of the discussions 11 recorded at the meeting held on the above date.
12 No member or the ACRS Staff and no participant at 13 this meeting accepts any responsibility for errors or~
14 inaccuracles of statement or data contained in this transcript.
k'.")T 15 16 17 18 19 20 21 22 23 24 25 Heritage Reporting Corporation (202) 628-4888
a I
n 2
t 1 NRC STAFF PRESENTERS:
2 T.Y. Chang N. Anderson i 3
4 i 5
i 6
7 I
8 i
9 10 .
11
! 12
, 13 0 14 ,
15 ,
r l' 16
\~
17 18 t 19 20 1 21 22 ,
t 23 ,
24 ,
25 O .
HERITAGE REPORTING CORPORATION -- (202)628-4888 {
l
3 1 PROCEEDINGS.
() 2 CHAIRMAN WYLIE: The meeting will come to order.
3 This is a meeting of the ACRS Subcommittee on the Reliability 4 Assurance. I am Charlie Wylie, Chairman of the ACRS S Subcommittee on Reliability Assurance. The other ACRS members 6 in attendance today are Carlyle Michelson, Chet Siess.
7 The Subcommittee will review the final outcome of 8 tha equipment qualification risk scoping study. An update on 9 the implementation of the resolution-of USI A-46, seismic 10 qual:.fication of equipment in operating nuclear power plants, 11 will also be given.
12 Mr. Richard Major is the cognizant ACRS staff member r- 13 for today's meeting. The rules for participation in today's 14 , meeting have been announced as part of the notice of this 15 meeting that was published in the Federal Register on June 1, 16 1988. This meeting is being conducted in accordance with the 17 provisions of the Federal Advisory Committee Act and the 18 Government and Sunshine Act.
19 We have received no written or oral statements from 20 members of the public. There is a request for each speaker to 21 first identify himself or herself and speak with sufficient 22 clarity and volume so that he or she can be readily heard.
23 I remind the Subcommittee the ACRS has been j 24 requested to comment on the EQ risk scoping study approach 25 determined in the risk significance of EQ issues and the i
l HERITAGE REPORTING CORPORATION -- (202)628-4888 1
4 1 staff's conclusions and recommendations of NRR in this area.
2 I would like the Subcommittee to bear in mind to the end of
{}
3 today we will consider what to bring to the Full Committee and 4 what recommendations that we care to make, so I would like to 5 consider the results of the EQ risk scoping study and the 6 staff's conclusions and their recommendations as we go 7 throughout the day.
8 I like the approach that has been taken in this 9 study. I think it is a logical way to determine what should 10 be the priorities in EQ research, in the approach for 11 qualification of that equipment. It is not clear to me at 12 this point, and maybe the staff can help us today, as to what rm 13 the staff intends to do with this, and what recommendations 14 the staff is prepared to make with regard to EQ activities and 15 research in this area. I would like to, now to ask whether 16 any other members of the Subcommittee have any comments?
17 Carl?
18 MR. HICHELSON: A couple of small comments--I agree 19 with you. I would also like to know what they are going to do 20 with it. I would also like if they could spend just a few 21 minutes kind of getting this particular piece of work into 22 context in terms of what else is being done on EQ, and how 23 does this fit in whatever else 's being done?
24 I guess my memory is not too good. I just don't O 25 remember what all the program is anymore or is this all the HERITAGE REPORTING CORPORATION -- (202)628-4888
-. - = _ - . _ - -. . - - . -. ,.
5 1 program we are doing at the present time? I just couldn't get
{} 2 that, that feeling, so if you could spend some time during 3 your introductory remarks, maybe just fitting this into the 4 context of whatever else you are planning on doing, I think it 5 would be very helpful.
6 In looking at the scoping document itself, I had one 7 small problem which maybe you can somehow help with the 8 discussion during the day, and that is you go in and you look 9 at various components in terms of perhaps what might happen to 10 them, and what the reliability might be, but I find absent any 11 discussion of the effect of the environment other than simple 12 failure of the component.
13 To say it differently, we know that if you heat 14 certain types of components because of adverse environment, 15 they don't always fail gracefully. They may fail giving you 16 some unwe.nted actions and the unwanted actions aspect just 17 doesn't seem to come through in your study. Either you get 18 the function even though it is hot, or you don't get the 19 function because it got hot, but you don't tell me what else I 20 might get that might even be potentially more serious. It is 21 a systems interaction question folded into this EQ study, and 22 that seems to me to be absent most of the time in the 23 discussion.
24 Another item which didn't come through from tin e to O 25 time as well as I had hoped, and that is a discussion of what HERITAGE REPORTING CORPORATION -- (202)628-4888
6 1 happens when we have adverse environments outside of
() 2 containment which we can get from various reasons. It is 3 touched on here and there. A little bit of the high energy 4 line is discussed, but it really doesn't get into it, but when 5 I read the inside of containment and I see the problems you 6 have are humidity and so forth, I am saying well, gee, the 7 equipment outside of containment isn't even qualified in many 8 cases, and we are potentially going to expose it to the same 9 high humidity-environments or water, or whatever, because cf 10 small line breaks.
11 Let me introduce one other question. You don't see 12 the small line breaks outside of containment. You did discuss 13 a wee bit medium and large breaks in terms of what the 14 probability is considered to be, and maybe what their 15 consequential effect might be, but small line breaks in the 16 right places could get you in deep trouble potentially, 17 depending on whether there is sensitive electronic equipment 18 nearby or that sort of thing such as electrical inverters, for 19 instance, and you don't seem to touch on the small break 20 outside of containment questions, and it certainly is,
- 21 probability is much higher than the large break, and so I l
l 22 found that also to be kind of another, comments that come out, i
l 23 but those were some of the things that seem to jump out during l
l l 24 the reading of it.
(:) 25 DR. SIESS: One thing that confused me was the use i
HERITAGE REPORTING CORPORATION -- (202)628-4888
7 1 of the word "accident." It was fairly clear that PRAs are
'~T 2 based on severe accident conditions in general. By severe (G
3 accident,.we haven't got the right terminology now because 4 beyond the design basis accident I guess, and most of the 5 equipment qualifications is based on design basis accidents, 6 and yet there are a number of places in this discussion where 7 an accident is used, and I am not sure whether you are 8 referring to DB or severe accident, and that may be an 9 editorial matter you can straighten out if the question comes 10 up. I think it is a distinction. PRAs are usually looking at 11 severe accidents. Much of this is devoted to severe 12 accidents, but equipment qualification is not.
13 MR. MICHELSON: One other small thing--I had one
<-)
%-)
14 other comment I forgot to add which I found to be interesting, 15 and important, and that is I guess maybe for the first time I 16 read a document in which we are beginning to appreciate that 17 these severe environments that are contained during an 18 accident condition DB type accident condition are, challenging 19 equipment which are, as the term you use, have really only 20 been, we only have good statistical information on normal or 21 nominal operation. We don't--
22 DR. SIESS: Let's hope it stays that way.
23 MR. MICHELSON: Well, we know, we know the accident 24 is in many cases challenging equipment to operate in their O 25 design basis format which they have never been tested to other HERITAGE REPORTING CORPORATION -- (202)628-4888
8 1 than maybe back at the laboratory because during normal f')
v 2 operation they just don't get challenged either 3 environmentally or functionally, and generally, this is kind 4 of glossed over lightly or not at all in most documents you 5 read.
6 This one seems to, to stress the fact that there is 7 a big difference.
8 DR. SIESS: There is an important point there 9 because we have relied on qualification in the laboratory for 10 safety-related equipment as opposed to harsh environment, but 11 the PRA people pay no attention to that. They use statistical 12 historical data based reliability data when they need to put 13 relations in. There is no way you are going to get a f-)
NJ 14 statistical basis out of a laboratory test. You can say okay, 15 the test has been made, probability is one that it will 16 survive that environment, probably isn't true, but they have 17 got to deal in probabilities and qualification is not 18 probabalistic, and you know, it is just two different worlds.
19 MR. MICHELSON: They are dcing the best they can, 20 but we have to recognize that that's a serious shortcoming of 21 PRA.
22 DR. SIESS: How serious it is, we don't know.
23 MR. MICHELSON: Well, that's partly what we find 24 out.
25 DR. SIESS: We are never going to find out.
HERITAGE REPORTING CORPORATION -- (202)628-4888 I
' - - -- _ , ~ . _._, _ . . _ . , __ , _ _ _ , _ _ _ _ _ . . . - . , __ , ,, ,_ _ _ _
9 1 CHAIRMAN WYLIE: I think that is one of the things 2 that the report says, that PRAs normally don't look that
{J}
3 detailed.
4 DR. SIESS: There is no way you are ever going to 5 find out the probability that a piece of equipment has been 6 qualified in the laboratory will survive the same conditions 7 in an accident. You can qualify to somewhat higher levels and 8 that should increase your confidence, but in terms of a 9 statistically based probability, you are not going to get it.
10 MR. MICHELSON: You have to test a lot of components 11 to get it.
12 DR. SIESS: Well, yes. We don't want to have enough 13 accidents to find out, 14 MR. MICHELSON: Yes.
15 DR. SIESS: We don't want to test a hundred pieces 16 of equipment because that wouldn't work. I don't know what 17 the EQ requirements would be. If you decided I want to test a 18 hundred pieces of equipment, would the staff be satisfied if 19 99 of them pass?
20 CHAIRMAN WYLIE: You couldn't afford it.
21 DR. SIESS: You could afford it. You could put a 22 hundred relays in a steam chamber, do them all, you know, but 23 what would be acceptable? Ninety-nine passing? Probably not.
24 You know, one failed, you test another one, what are the O 25 requirements? If you test something and it fails, can you HERITAGE REPORTING CORPORATION -- (202)628-4888 <
10 1 test another one and another one and another one like you do, 2 you know, testing welds? Sampling? This is a sampling
{}
3 process, but it is just two different worlds.
4 MR. MICHELSON: Then you only are testing in the 5 laboratory, not what is installed in the plant where it now 6 runs into the sealing problems around the electrical entries.
7 DR. SIESS: Deterministically based set of 8 regulations, and then we have got PRA, and all they wera 9 trying to do here was to see what looking at PRA could tell 10 you about what we are doing deterministically, and some places 11 it is not much relation between them. Insights, what are the 12 insights from PRA?
13 MR. MICHELSON: I found the report to be extremely 14 helpful in better understanding what they really are dealing 15 with in the PRA.
16 DR. SIESS: It points out the efficiencies in PRAs.
17 That is certain things they don't look down in enough detail.
18 Also points out deficiencies in the qualification 19 requirements. We worry about a lot of things that aren't
- 20 important, presumably get confident from all of these things 21 we worry about that is not important, which is wrong.
22 MR. MICHELSON: One other thing that I also just 23 recalled in reading this, it is, it is very difficult to tell 24 for sure when you expose a large amount of equipment to a O 25 common environment like when you have an accident inside of HERITAGE REPORTING CORPORATION -- (202)628-4888
11 1 containment or outside of containment, it is hard for me to 2 tell for sure whether the PRA studies have caught this common
( ]3 3 mode effect and applied it to common environments to all the 4 components concurrently, which is the way an accident goes, 5 exposes everything concurrently, and what is the effect of 6 that concurrent exposure inside of containment? Because you 7 have only singled out a few pieces of equipment. You singled 8 out some HPCI and RCIC and fans, for instance, but you didn't 9 single out all the equipment insides of containment whose 10 malfuncton now is a result of exposing it to that kind of 11 environment, introduce many more questionc and problems that, 12 in mitigating the accident.
13 DR. SIESS: I thought what they looked at was the 14 report.
15 MR. MICHELSON: No. That's what I, I--was important 16 one, viewed one at a time and some of them turn out not to be l 17 important when viewed one at a time, but I didn't get the 18 feeling that you viewed them all concurrently and then said 19 this one wasn't important. That's the part I didn't get.
20 Maybe you can clear it up during the discussion because it is 21 very important if you claim that you are exposing all 22 concurrently, which is the case, then you have to show that, 23 that your study of individual components in finding them 24 non-important was from the viewpoint of their being exposed O' 25 and failing concurrent with all the rest of these things.
HERITAGE REPORTING CORPORATION -- (202)628-4888
-- ~ __
12 1 That didn't really come through, f'
.L] 2 CHAIRMAN WYLIE: Suppose we move ahead.then, and I 3 believe Moni, you are next.
4 MR. DEY: Okay. Good morning. My name is Moni Dey.
5 I am the project manager for the equipment qualifications 6 scoping study in the Division of-Regulatory Applications in 7 the Office of Nuclear Regulatory Research.
8 We have had some opening remarks from the 9 Subcommittee and let me try and answer some of them as I go 10 through my brief presentation, and we can clarify some of the 11 points as we go through the technical presentations ano 12 summarize at the end if you wish.
.,f s 13 A brief historical background--as you are aware, the 14 NRC funded EQ research for qualifying safety-related 15 electrical equipment to function during design basis 16 accidents. For the last ten years, that was primarily to 17 support the EQ rule regulation, 50.49 and the reg guide and 18 accompanying standards to confirm whether the rule adequately, 19 and the standards adequately address equipment operability 20 during design basis conditions, and I would like to stress 21 that, that that was primarily directed to design basis 22 accidents, and supporting the EQ rule, which is for electrical 23 equipment and for, it is based on deterministic considerations 24 or design basis accidents.
O 25 MR. MICHELSON: Moni, when expressing the word HERITAGE REPORTING CORPORATION -- (202)628-4888
13 1 design basis accident, let me ask you did you consider a pipe
(} 2 break outside of containment as a design basis accident?
3 MR. BUSTARD: Pipe break, main steam line break, or 4 high energy line break is a design basis accident.
5 MR. MICHELSON: All high energy; how about moderate 6 energy line breaks?
7 MR. BUSTARD: No.
8 MR. MICHELSON: Why did you say one yes and one no?
9 What is--the regulatory history is a letter was written a long 10 time ago to look at both high and moderate energy line breaks; 11 1972 and '73 those two letters came out, but did that make it 12 a design basis accident then when the utilities were told they 13 had to look at the consequences of such breaks?
14 CHAIRMAN WYLIE: Are you talking about outside?
15 MR. MICHELSON: Inside, too; that letter covered all 16 pipe breaks. It is focusing primarily on outside of 17 containment, those two.
I 18 MR. DEY: Bill Farmer, would you like to address 19 that?
20 MR. FARMER: You are referring to something beyond 21 the main steamline break. Some plants do use the main steam 22 line break as the--
23 MR. MICHELSON: But it is generally postulated to be 24 inside of containment for purposes of the questioner and all
[~)
that sort of thing, but for purposes of, for outside of 25 t
HERITAGE REPORTING CORPORATION -- (202)628-4888
14 1 containment purposes, are those breaks design basis accident?
/~h 2 MR. FARMER: I don't think they are called that as
(_/
3 such, although the licensees do address the exposure of 4 equipment in say the auxiliary building to. environments 5 greater than normal operating.
6 MR. MICHELSON: What classes do we put these things 7 in it we don't call them design basis? Because it is part of 8 the design basis of the plant to high and moderate.
9 CHAIRMAN WYLIE: Let's take, for example, a 10 feedwater line.
11 MR. MICHELSON: Take a smaller one. The big ones I 12 think are a little easier to justify as design basis, although
- 13 it isn't clear outside of containment that they are, either.
14 CHAIRMAN WYLIE: I tried to recall--don't they 15 really look for pipe breaks and impingements? But as far as 16 the other environmental conditions, I'm not sure they really 17 look at that.
18 MR. MICHELSON: Well, they are supposed to. the 19 Muso and O' Leary letter said you look at the consequences and 20 that ever since I have wondered are these now design basis 21 accidents? And that's why I put it up this, on the slide. I 22 wondered is that what you are talking about? Or what design
! 23 basis accidents-are we talking about purely Chapter 15? Is i 24 that what you mean?
25 HR. FARMER: Under the EQ rule, it is as far as I HERITAGE REPORTING CORPORATION -- (202)628-4888
15 1 know, purely Chapter 15.
()
U 2 MR. MICHELSON: Okay. Then there is another whole 3 class that we will have to talk about from time to time how 4 you handle those because you are supposed to be designing for 5 them.
6 -MR. FARMER: I think they are addressed but not in 7 this formal listing fashion.
8 MR. MICHELSON: The report is a little mixed up on 9 what are external events and what are internal events. Your 10 definition in the report is different than any I have seen 11 before.
12 MR. BOHN: I don't understand that comment.
13 MR. MICHELSON: PRA people have a definition of 14 external events. It is not the same as you are using in here.
15 Some of the things you call external they consider internal 16 and vice-versa. I will find it later on. We will talk about 17 it, chapter and verse. Go ahead.
18 MR. DEY: This research program was terminated in 19 1986, and the results of research are documented in NUREG/CR 20 301, and Bill Farmer from tb- Office of Research was project 21 manager for that, for that study, and summarized the status on 22 equipment qualification, issues, research, and the status of i 23 the resola. ion as a result of ten-year program.
24 The main objective of this, well, let me come to O 25 that later. Following that, the ACRS in 1986, recommended HERITAGE REPORTING CORPORATION -- (202)628-4888
16 1 funding of EQ research to assess the survivability of 2 electrical equipment for design basis, broad spectrum of N'JT 3 accidents, including severe accidents, stated that the 4 research is viable to preventing and mitigating accidents 5 should they occur.
6 So the EQ risk scoping study was initiated to 7 determine the risk impact, to primarily determine the risk 8 impact of not doing the research that was being conducted to 9 support the EQ rule, the reg guide, and the accompanying 10 standards.
11 Now as a result of the ACRS letter, we have 12 attempted to broaden that scope to a certain extent, to 13 address beyond design basis accidents, severe, severe 14 conditions, but that isn't the main focus and objective of the 15 study, although we have tried to cover it, and we have, 16 frankly we have not been able to cover all conditions, all the 17 full spectrum of accidents and severe conditions, as you will 18 note later from the presentations, that there are difficulties 19 in assessing the risk impact of equipment during severe 20 accidents due to lack of PRA methodology, but the study does 21 highlight areas, provides a starting' point and highlights 22 areas that may be further studied.
23 HR. MICHELSON: Your statement bothers me a little 24 bit. In looking at where the shortcomings are, I generally O 25 don't seem to find them as shortcomings and methodology at PRA HERITAGE REPORTING CORPORATION -- (202)628-4888
17 1 but rather in the availability of data and of adequate
-2 modeling where once if you can draw them, model of the, what
(')T 3 you are trying to figure out, and if you have got the data to 4 plug in, the PRA methodology comes through very nicely for you 5 I think. Is that correct?
6 HR. DEY: That is what I meant.
7 MR. MICHELSON: Shortcoming is really not in the 8 methodology, but in the availability of data, and the adequacy 9 of modeling because we aren't smart enough to model all the 10 things that could happen and so forth, but if we were, the 11 methodology certainly would work. That's the way I would do 12 it.
13 MR. DEY: Now as you mentioned, there is two areas 14 which are lacking in, in PRAs currently. One is the data, and 15 which affect the limits the reliability data base for varioua 16 components like valves, motor-operated valves, and SOVs, and 17 also the HRA human reliability analysis models for plant 18 instrumentation, which is, which is vital during severe 19 accidents.
20 HR. MICHELSON: Also the models of physically what 21 is--and you point out in your report that you don't have the 22 models, don't go down and chase the water through a conduit 23 leak and that sort of thing. It could if you wanted to, but 24 it doesn't simply because you don't have that detailed an O 25 understanding of the arrangement for the particular plant.
HERITAGE REPORTING CORPORATION -- (202)628-4888
18 1 MR. BOHN: This is Mike Bohn. There is another area
(~N 2 that is, that is lacking, and that has to do with recovering uI 3 from accident, or trying to. Let's say your PRA hypothesizes 4 an accident sequence that is brought down to say shutdown but 5 you have to hold it there for a couple of weeks before you can 6 take some sort of final action.
7 PRAs don't address what you have to have to hold it 8 in that position and what information, instrumentation is 9 required, and it is those areas where some of the EQ issues 10 will have an impact, but we can't quantify them right now 11 because the PRAs just do not have, they do not consider the 12 sequence after it goes to hot shutdown. That's a limitation 13 that we ran into looking at a number of different issues.
14 MR. MICHELSON: This was brought out in the report.
15 It would have been awfully nice, though, if you kind of pulled 16 this together in one place, and focused on this kind of an 17 issue. It doesn't come through because you have got to read 18 the whole report to keep picking it up hero and there, and it 19 really is an extremely important point, but it didn't come 20 through to the reader as well as it perhaps should. It is 21 there no doubt, and it is a better, better attack on that 22 problem than I have seen in a long time.
23 MR. BOHN: We will try to focus on that a little 24 more.
O 25 MR. MICHELSON: Summary would be one place. It is HERITAGE REPORTING CORPORATION -- (202)628-4888
f 19 1 in there some, but you just didn't bring it all together and 2 really nail it down. It is a serious problem with FRA and it
_( }
3 is something that people who look at the bottom line had 4 better realize, and hear all the kinds of shortcomings, what 5 has been done.
6 MR. BOHN: My understanding is that as part of the 7 follow-up of the 1150, there is going to be some limited work 8 focused on these type of long term or beyond, beyond EOP 9 actions, and as that word defines some of the options further 10 in PRA context, I think there is the opportunity to re-examine 11- some of the EQ issues in light of what they come up with.
12 HR. MICHELSON: When you do it, you need to--you
- 13 talk about this stuff piecemeal in this report. You talk 14 about the blackout and the effects on the RCIC room 15 temperature. You ignore the rest of the building and what it 16 is warming up and what the solid state electronic back in some 17 control cubicle is doing to perhaps even interfere ultimately 18 with RCIC when you try to bring the power back on and so 19 forth. It just is piecemeal.
20 MR. BOHN: That's a valid observation. We had 21 trouble getting our hands around this program because it was, 22 involved so many different issues, and yet the PRAs didn't 23 have the level of resolution to address them, and so we 24 focused on trying to identify at least a number of issues l
O 25 where we could clearly say hey, there is a problem here. And l
l HERITAGE REPORTING CORPORATION -- (202)628-4888
20 1 it was, it did come out looking piecemeal by the time we were 2 done.
(~]Y
\- .
3 MR. MICHELSON: What I would like to--I don't think, 4 1 didn't think you would do any more than that because there 5 was a finite limit to any of these studies, but somewhere 6 there ought to have been new paragraphs pointing point out the 7 other things that weren't done that could be potentially 8 contributing significantly to the end results and that the 9 reader at least has got to keep this in mind.
10 MR. BOHN: We appreciate your comment.
11 MR. MICHELSON: recognition of the shortcomings of 12 the whole piece of work.
7- 13 (Slide)
V) 14 HR. DEY: Other than the follow-up 1150, the staff 15 is currently going through the implementation of the severe 16 accident policy statement, and depending on the outcome of 17 that implementation and development of beyond emergency, 18 current emergency operating procedures the study has indicated 19 there may be need for investigating equipment qualification l 20 along with determining those strategies. Okay.
21 Briefly go over the objective of the study, put it 22 into perspective, it is to determine to the, che electrical 23 equipment that are essential to preventing core melt accidents 24 as well as mitigating consequences, and a list of potential O 25 equipment of equipment that are important will be presented to HERITAGE REPORTING CORPORATION -- (202)628-4888 >
v ., ..,-,- - --,,_--.- -- -- ,- -- -..
- - - , - . - ,. - - - - - , ,-,w-- , ,-.,e- -
21 1 you.
2 We also will present the EQ assumptions and issues
["}
3 pertaining to the performance of these important equipment and 4 what may be areas that warrant further, that warrant further 5 investigation.
6 Now the question came up what the results of the 7 study will be used for. As I mentioned, it is primarily to 8 determine the risk impact of the issues that are, that were 9 documented in Reg CR 4301, which were the EQ issues as a 10 result of the past research. And if they are, if any of those 11 issues are important enough to be identified as generic 12 issues, and I think you will see that with the information we 13 have from the study, tnere are not, but certain equipment have 14 been identified whcre the reliability data will be necessary, 15 and depending on the outcome of that data, one would be able 16 to get a better estimate of the risk impact, but at this 17 point, we don't see any generic issues.
18 However, the research study has recommended some 19 areas to concentrate on in the inspection and application of 20 the EQ rule, and some areas to stress and some areas to not 21 stress as much based on risk significant PRA considerations.
22 And I think the study EQ, this is the first time EQ 23 has been looked at from a PRA viewpoint, first as an example 24 have been modeled in the past in PRAs. EQ as we have O 25 discussed this me..ning is treated rather crudely in PRAs.
HERITAGE REPORTING CORPORATION -- (202)628-4888
22 1 This is the fi.rst time where EQ which is based on conservative 2 design basis considerations is being looked at from a PRA 3 probabalistic risk significant viewpoint, and it provides 4 insights as to one, for current reactors which areas should 5 be, should be stressed, which areas for which we may need more 6 data, and for future reactors, maybe we want to treat EQ 7 differently depending on the classification of accidents.
8 MR. MICHELSON: I think you said that fire had been 9 looked at from the EQ s'.ewpoint? Did I understand you 10 correctly?
11 MR. DEY: No. Fire has been analyzed in PRAs.
12 MR. MICHELSON: Oh, in PRAs, because we haven't 13 looked at it, except for a few cable thermal barriers, we 14 really haven't looked at equipment qualification at all from 15 the fire viewpoint, from the PRA viewpoint. Therefore, we 16 have really not looked at fire because we didn't have any 17 data. We don't know the susceptibility of equipment to damage 18 from fire results such as heat or smoke or whatever. Without 19 a data base I don't think--you can do all the PRAs you wish, 20 but you can't get any results because you don't have any 21 numbers to plug in, so people are purported to have done fire 22 PRAs but they really haven't because they don't have any data 23 base to work with on the qualification of the equipment for 24 those environments.
O 25 CHAIRMAN WYLIE: Your first bullet, that's not HERITAGE REPORTING CORPORATION -- (202)628-4888 I
23 1 exactly correct, is it? The risk of just electrical? It is 2 all equipment, is.. t it?
. (~'}
(-
3 MR. DEY: Okay.
4 CHAIRMAN WYLIE: It is all, important to safety.
5 MR. DEY: Yes. The, developed for--the study 6 applies to equipment, all equipment important to safety, but 7 a. we go into analyzing the EQ, the details of the equipment, 8 we have analyzed electrical equipment.
9 CHAIRMAN WYLIE: Principally electrical equipment, 10 but there is other considerations in that report--pumps, 11 heating up of pools.
12 MR. BUSTARD: Sure, but pump has an electrical
- 13 component to it, for example, a check valve will not have 14 electrical component to it, and the check valve will there;' ore 15 be outside the scope of this study.
16 MR. MICHELSON: Let's take the containment fans, for 17- instance. Let's assume for the moment that the motor is well 18 qualified. The spray doesn't bother it a bit, but how about 19 the mechanical loading of the fans by that 8 tenths of an inch 20 per hour rain storm that is falling into the fan blades? You 21 don't think that has a whole lot of load to the motor? When 22 you are talking to heating and ventilating people, that seems l
23 like it is hard to believe that those fans can handle that i
I 24 kind of a mechanical load under such conditione. I didn't n
25 find it too well discussed in the report in the mechanical HERITAGE REPORTING CORPORATION -- (202)628-4888
__=.7... . _ . _
24 1 loading aspects.
(} 2 MR. BOHN:
This'is Mike Bohn.- This is question has
~
3 come up in a number of views, and let me try to clarify it.
1 11 From the PRA perspective of this problem, we 5 attempted to include all equipment that was required to do 6 something~following an accident, to mitigate an accident, so 7 we tried to look at all equipment regardless of how it failed.
8 We tried to identify those pieces of equipment that saw a
. 9 harsh environment as a result of the accident. Then from the 10 EQ point of'. view, our insights and our knowledge of what 11 failure levels are likely to happen even though they are very.
12 rudimentary are based on test data on electrical equipment, so 11 3 our insigh' 'f where things start to go back come from the
.O 14 limited ; cst data and the NRC's EQ research program that has 15 been going on.
16 MR. MICHELSON: I don't think you need test data to 17 know what happened to get that much water into the fan blades.
18 This is well protected. There is no way that-19 containment spray can get to the fan blade, but if the water 20' gets to the fan blades, mechanically you will burn the motor 21 up. It is not a pump. It is a blower.
L 22 MR. BOHN: I understand, and that's a valid comment.
23 I am just saying thit when we tried to draw insights from, 24 from actual, where we tried to estimate where failure levels O 25 occurred, there we were restricted to EQ issues associated HERITAGE REPORTING CORPORATION -- (202)628-4888
. _ . . ~ . . _ . _ _ . - _ , _ _ . . . _ - - - ~ . . _ . _ , - . . .
^
25 L1 with the electrical equipment, but in terms of determining
(} .2 what was important and what would see harsh environments, we 3 .tried to look at all equipment so these insights should be 4 valid regardless of the failure modes.
5 MR. MICHELSON: Weren't_a number of these pieces of 6 equipment qualified by analysis and not by testing?
7 MR. BUSTARD: It was true in general when it comes 8 to--
9 MR. MICHELSON: I wonder whether we are that smart 10 yet to, you know, know enough to qualify a piece of equipment 11 on analysis alene.
12 MR. BOHN: When, when a piece of equipment is 11 3 qualified by testing, we make assumption as to accelerated 14 aging and temperature relations, and some of these issues, of 15 course, we don't know that well and there is an ongoing.
16 program that Bill Farmer is in charge of to look at aging '
17 effects and-artificial aging and ways of monitoring aging so 18 yes, these are issues that, analytical issues that enter enter 19 into even testing equipment.
20 MR. MICHELSON: Because qualification by testing has 21 been a rather limited sort of qualification. I think more 22 stuff was analyzed qualified than was tested. Is that ,
23 correct? Or do you have a feel for it? Most of the big
'24 pieces of equipment tested like the fan motors, that sort of O 25 thing?
HERITAGE REPORTING CORPORATION -- (202)628-4888
i.
26 1 MR.. BUSTARD: Generally when the NRC'goes to a 2' ~ utility and-does an inspection, they want to see some test
- -}
'3 data. That's the preferred thing to look at.
4 MR. MICHELSON: It is not required.
5 MR.. BUSTARD: Not in every case.
6 HR. MICHELSON: Regulations allow you to qualify by 7- analysis?
8 MR. BUSTARD: That is correct.
9 MR. BOHN: I don't believe fans are tested.
10 MR. FARMER: I once looked for this very item 11- because everybody said well gee, the fans all operated at THI, 12 and were amazed they lasted a year, and I tried to find 13 qualification data and the only thing I found was reference to
-O 14 Westinghouse and one, one of the fan-manufacturers having 15 -conducted an EQ test, but I didn't find that, the test data, 16 so.I don't know how far, whether they just irradiated a piece f
, 17 of. equipment. I doubt very seriously if it is subjected to 18 water spray. I imagine it was a chamber type of 19 qualification.
20 MR. MICHELSON: I think it probably was a static 21 test in a test chamber and thermal dynamically took it out of 22 the chamber.
'23 MR. BUSTARD: Information let's say in e utility
. 24 FSAR that I know of does refer to test data in containment 25 fans. Now again, it is proprietary test data and I didn't HERITAGE REPORTING CORPORATION -- (202)628-4888
j ;- - 27 1 have access to cny. details, so I don't know how far they took 2' the test.
3 MR. MICHELSON: Could you have gotten access if you 4' 'needed~it, if you wanted it? I don't think the contractors 5 are limited to non-proprietary information in doing their 6 work. It is just that they don't explore it that far.
7 MR. BUSTARD: That much detail.
8 HR. DEY: Okay. Objective No. 3 was to determine if 9 a need exists for further research to characterize areas of 10 large uncertainty.
11 As I mentioned, the objective of the study was 12 two-fold, was, is to determine the risk significance, and two, 13 if one could reduce large areas of large uncertainty, and the 14 study presents some areas where further data may be necessary 15 to get, to determine a better esuimate of the risk 16 significance of-the components.
17 (Slide) 18 MR. DEY: Let me give you, give you the status of 19 the project, and schedule. We have the draft final report in 20 front of you, and we request comments from the Committee in j
21 July. Now what we would like to do is, we are presenting at 22 this meeting the results of the study, and as a result of 23 considering those recommendations, if you could, if you could 24 give us an indication of where you think further research is O
\~# 25 warranted, we would like to use that as input to'tard our HERITAGE REPORTING CORPORATION -- (202)628-4888 )
28 1 planning.
() .2 Okay. We have had a peer review team go over this 3' report. We have had two meetings. The team consisted ef.
4
. George' Slater from EPRI. -Ken Kennedy from Duke Power was
'S representing-a number of utilities'and industry organizations;-
6 Sal Perfagnio from Franklin Research Institute who has been 7 involved in the EQ area for a number of years; and Andy 8 Woelford from INELL. That was primarily to try and tie up to 9- some of the other EQ work that is going on, which you asked-10 about.this morning. Unfortunately, Jerry Weidenheimer wasn't 11 able to be here this morning. He is the project manager.for 12 EQ work in the mechanical equipment area. That's primarily 13 being conducted at INEL. I think you have been briefed by 14 that recently. The peer review teams were very useful and I 15 think contributed significantly to the report. We don't have 16 the final comments in as yet from the peer reviews, but they 17 provide, provided preliminary comments, global comments at the 18 meetings and preliminary comments in written form to us which 19 recommends, which is reflected in the report.
20 We hope to have the final NUREG report out by the 21 end of July, and a research information letter from the 22 Division of Regulatory Applications summarizing what the staff 23 intends to do by the end of this fiscal year, and in
- 24 developing that, we would like your comments on what areas you O 25 think as a result of the study that we might want to consider i
HERITAGE REPORTING CORPORATION -- (202)628-4888
,,-.# _ . . - . _ . . = . . _ _ _ _ . _ - _ . _ . . _ _ . _ _ , , _ , _ , - . _ . _ . _ _ . . . , _ . , _ _ _ _
29 1 as part of that plant.
(~) 2 DR. SIESS: Tony, relating to the risk that might V
3 result from let's say inadequacies in our present requirements 4 for environmental qua..ication, to what extent will this be 5 factored into the individual plant examinations of severe 6 accident policy implementation?
7 MR. DEY: I'm not sure I can answer to what extent.
8 I know it will be, it will be addressed.
9 DR. SIESS: I mean it is a part of it. The idea of 10 the IPE is defined risks that we didn't know about from 11 whatever cause, but now the IPE, if it is done by a PRA, 12 probably won't do any better job in finding these thingr than
- 13 PRAs in the past have. IDCOR methodology have some mechanism
~,s 14 for taking into account the common mode failures due to harsh 15 environment, which is what presumably you are talking about 16 here, the harsh environment being a severe accident 17 environment?
18 MR. DEY: Yes. They have addressed the question of 19 equipment survivability during severe accidents, and they have 20 a report out on that a couple of years ago, and we have taken 21 a look at that report and we will make comparisons between 22 what this study has come up with and what the IDCOR came up 23 with, so in answer to your question, I think they have taken a 24 look at it.
)
25 DR. SIESS: If they have taken a look at it and the HERITAGE REPORTING CORPORATION -- (202)628-4888
30 1 IDCOR methodology would look at that sort of thing, it seems
(} 2 to me the IDCOR methodology would be better than the PRA which 3 I think is significant. We have been urging a PRA rather the 4 IDCOR methodology.
5 MR. DEY: One area which we have comments on in the 6- peer review teams and the IDCOR makes these statements in the 7 reports is the assumption on corrective action on what the 8 operator is able to do to terminate the accident. Depending 9 on those assumptions, that determines the need for equipment.
10 He can't take corrective action and mitigate--
11 MR. MICHELSON: The problem is, though, when they 12 start going through these kind of rituals they fail to 13 recognize that the event that they are facing may be causing a 14 large amount of misleading information to be fed to the 15 operator and that sort of thing which would, now makes the 16 operator action quite different and sometimes hurtful even 17 instead of helpful. That's in this study. They just say oh, 18 by the way, remember people ca. do things, but people can do 19 the wrong things, too, very easily if they are misled by 20 incorrect information; TMI, for example.
21 DR. SIESS: The study did address information--
22 MR. MICHELSON: The effect of elevated temperature 23 on the solid state cabinets that send information on to the 24 control room they don't address, and misinformation to the O 25 control room, for instance.
HERITAGE REPORTING CORPORATION -- (202)628-4888
31 1 DR. SIESS: I didn't'go through that much detail,
(} 2 but I thought I saw conclusions about th- instrumentation 3 needed to follow the course of an accident and suggested that 4 qualification of some of the instrumentation called for by Reg 5 Guide 1.97, more work needed to be done there.
6 MR. MICHELSON: Qualifying inside of containment, 7 Chet; it is outside of containment where you can always get 8 into this. That is how many severe accident scenarios can 9 generate, by scenarios outside of containment ultimately 10 leading to core melt. We just don't do it very well because 11 we don't know how to do it.
12 Severe fire, for instance, we don't know how to do 13 the consequences of a severe fire because we don't have any 14 data on how equipment responds to elevated temperature and 15 smoke and so forth, and in fact the fire protection systems i 16 are going off because they think there is a fire and there 17 isn't one in another area. It is hot and then just heat and 18 smoke, and we don't include that sort of thing in the studies.
19 DR. SIESS: What is the answer then? Qualify 20 everything for everything?
l 21 MR. MICHELSON: There isn't any easy answer, but l
! 22 certainly we have to recognize that this is a shortcoming, and 23 we have to start dealing with it.
24 DR. SIESS: Safety-related equipment, as they point l
O 25 out in the report, we have tended to rely on redundancy, and i
i HERITAGE REPORTING CORPORATION -- (202)628-4888
32 1 of course, the harsh environment can cause concrete common ;
(} -2 mode failures either inside or outside of containment.
3 MR. MICHELSON: That's right.
4 DR. SIESS: Outside of containment, if it is not 5 safe, not safety-related, we don't necessarily rely on 6 redundancy. Probably don't even bother to qualify it if it is 7 not safety-related.
8 MR. MICHELSON: If it is safety-related, we don't 9 necessarily qualify it, either.
10 DR. SIESS: But the point is I think that what we 11 call safety-related for our design basis deterministic 12 accidents and what we say is not safety-related for those 13 accidents may not be the same for severe accident.
f-C/
14 MR. MICHELSON: That's right.
15 DR. SIESS: Particularly when you get into accident 16 management, you are going to be using equipment that is not 17 normally considered safety-related. It is there. Again, it 18 just simply points up to me that we step over the line from 19 design basis accidents to look at the severe accidents, and 20 PRA, we move into a different world.
21 MR. MICHELSON: Well--
22 DR. SIESS: This is an attempt to sort of bridge 23 that--
24 MR. MICHELSON: There is another world I think we
aren't looking at, and that is, you know, I asked the question 25 n
HERITAGE REPORTING CORPORATION -- (202)628-4888
33 1 a little earlier what is a design basis accident? I think T 2 there is a world that we aren't looking at which is beyond the
('J s_
3 design basis and not included in that basic design basis set, 4 but not severe accidents yet, but which can-lead to severe
. accident such as I notice in the, this report, the Station 6 Blackout is considered a non-design basis event. Clearly it 7 is an important event. I hear your report says it is 8 non-design basis, so I have got to have a new definition. I 9 thought power blackout was a design basis event. I thought we 10 had to design for it and therefore it was design--but if, 11 because you design for it doesn't mean it is a design basis 12 accident anymore.
13 DR. SIESS: A design for loss of off-site power.
(9
~~/
14 MR. MICHELSON: But Chet, the problem is 15 nomenclature. We are talking about only two things--design 16 basis accidents and severe accidents. I am saying it is a 17 third set, non-design basis accidents, which t,re not in that 18 set you normally call severe accidents.
19 DR. SIESS: These are initiators of PRAs.
20 MR. MICHELSON: Not initiators; sometimes they are 21 follow-on. Sometimes they are initiators.
22 DR. SIESS: They are the thing, I mean blackout is 23 initiator, so starting points for getting into a severe 24 accident.
25 MR. MICHELSON: But it is a non-design basis event.
HERITAGE REPORTING CORPORATION -- (202)628-4888
34 1 We don't--which means to me we don't design for it, but it is
(~T 2 a poor' nomenclature. I think we do design for it.
LJ 3 DR. SIESS: I prefer Chapter 15 for design basis.
4 It has been :o long since we looked at SAR most people have 5 forgotten what Chapter 15 is.
6 MR. MICHELSON: We have--the Commission better start 7 thinkir.g of new definitions because we are getting confused 8 now between severe accident and design basis accident, and not 9 recognizing there is another whole set of blackout is one of 10 them. Pipe breaks outside of containment are another one.
11 Fire is another one. Is fire a design basis accident?
12 MR. DEY: Those of you who may' remember our proposal 13 for the advanced reactors, we got into this problem. We just 14 called them event category one through four, and just assigned 15 a number.
16 HR. MICHELSON: That's fine, too. At least we 17 understand what we are talking about when we write a report.
18 DR. SIESS: Design basis accident clearly is an 19 accident you design for, design the plant for. Anything else
! 20 are accidents you examine the plant for, and I think the gap l 21 between the examining and severe accident, designing it for i 22 the design basis, some of the initiators are just arbitrary, l
23 assumed without thinking how you get there.
24 MR. MICHELSON: Also I think the requirements may be I
( 25 different for severe accident and for the these non-design i
~
l HERITAGE REPORTING CORPORATION -- (202)628-4888
35 1 basis accidents such as pipe breaks outside of containment.
(')
V
- 2. They have, may have more severe requirements than the severe 3 accident because they are more likely presumably.
4 MR. KOLACZKOWSKI: Dr. Siess, I would like to make a 5 clarification for the record, and that is that relative to the 6 IDCOR PRA if you will, or IDCOR approach for the IPEs, yes, 7 they have at least looked into the concern about survivability 8 of equipment, but they have no better way or approach than the 9 PRAs do in terms of trying to really quantify the handling of 10 tha model. There is no glorious way to handle it out there 11 some place.
12 DR. SIESS: Assume--then you could have a panel of 13 experts to decide what the probability is; back to that, 14 right?
15 MR. KOLACZKOWSKI: Yes.
16 MP. . MICHELSON: Assuming it is failed may not be 17 conservative. There may be a more serious problem of getting 18 unwanted action from it instead of simple loss of function.
19 DR. SIESS: The design you produce is certainly, may 20 not be the right design.
l 21 MR. MICHELSON: For instance.
22 CHAIRMAN WYLIE: Okay, j 23 DR. SIESS: We design plants with a ( artain set of 24 design basis accidents, which lead to certain types of 25 designs, certain levels of redundancy, certain levels of HERITAGE REPORTING CORPORATION -- (202)628-4888
36 L 1- equipment qualification, and what we are-in the process of (j- 2 doing now-is using risk based analyses to see how-safe we make 3 them.
11 This Committee keeps pointing out, we compare those 5 results with the safety goal. We should use that to decide 6 whether the regulations are providing an adequate review of 7 safety. We don't expect the. design plants to get severe core -
8 damage down to ten to the minus 8. The object is to design 9 plants so that the residual risk we used to call class
'10 accident and environmental studies, that was the residual 11 risks at an acceptably low-level. What is acceptable 12 presumably is defined by the safety goals except the safety 13 goals are not all that clear, either, you know. This is 14 objectives, this is quantitative goals, not in our lifetime 15 core melt and ten to the minus~6 release of something that is 16 called large, and again you can't design a plant without 17 having a design basis.
18: MR. DEY: We will clarify the report. The intent 19- was in the report for design basis to mean the design basis 20 considered in the EQ rule, and since that, that is a starting 21 point for the study, and anything beyond that is non-design 22 basis, but we will clarify that.
23 MR. MICHELSON: Your clarification really is, design 24 basis is Chapter 15 type event only? Is that what you are 25 going to clarify?
HERITAGE REPORTING CORPORATION -- (202)628-4888
-. .__ _ _ _ _ . . _ - , , . . _ . _ _ _ ~ . . _ . _ - _ . _ . - _ _ . _ . , . _ - . . , . . .
37 1 MR. DEY: Yes. We will, I will send you the peer
(~)
1X /
'2 review comments as when they become available.
3 MR. MICHELSON: The report does clearly go beyond 4 design basis in some areas some of the discussion?
5 MR. DEY: Yes.
6 MR. BOHN: Well, in fact virtually all the PRA 7 sequences are beyond design basis.
8 MR. MICHELSON: Before they get done.
9 MR. BOHN: So every one of them leads beyond. You 10 can't really find a sequence that is truly one, only one with 11 the design basis accident condition as defined for the EQ 12 requirements, and some of the conclusions that we come to in 7s 13 this study is that certain areas especially related to Gl 14 radiation are perhaps over, over-harsh in terms of 15 requirements, and some temperature / pressure conditions are 16 under-harsh, and there is possibilities of changing the 17 requirements to be more realistic.
l 18 MR. DEY: I was wondering if Bill Farmer, if you 19 would like to summarize the EQ worl: going on now for yourself 20 and Jerry and your division?
21 MR. FARMER: I can give you a quick overview 22 whenever you want, and it will have to be a little sketchy and 23 Jerry's work.
24 I am Bill Farmer of the Office of Research, Division 25 of Engineering.
HERITAGE REPORTING CORPORATION -- (202)628-4888
38
- 1 As far as the EQ research that is going on, as you 2 know, originally we had electrical, it was divided into
' {}
3 electrical and mechanical work. The electrical was done
-4 predominantly at Sandia with a.few additional-tasks being 5 conducted at Franklin and elsewhere.
6 The majority of the mechanical. work was being dona 7 by INEL. All of the electrical work was essentially unfunded 8 after Fiscal '86. The'only work that has been going on since 9 then has been mainly the work of trying to complete certain 10 unfinished tasks, and write up and edit publish reports on the 11 work that had been conducted up to them then.
12 The mechanical work has been ongoing, and in fact 13 this past year, there were a number of tests conducted in the 14 HDR German facility in which the mechanical EQ group had 15 installed some motor-operated valves in the piping at the HDR, 16 and these valves were then subjected to, to various seismic 17 loads as a part of the overall HDR program.
18 The main motivation for the HDR program was the 19 dynamic response of piping to seismic loads, and they had a 20 special shaker installed to provide the input forces function.
21 That work I believe was finished last last month. There was r 22 initially some problems encountered with the motor-operated 23 valves. It was never fully established what, why they, one of 24 them behaved slightly erratically. We tend to believe it
~O 25 could be very well related to the power supply inadequacy.
HERITAGE REPORTING CORPORATION -- (202)628-4888
39 1 They had a DC generator which apparently was not adequately
(~}- 2 sized. It was replaced with an AC generator in the final test
%s 3 last month, and the valves did behave perfectly. That is, 4 they closed under seismic loads, with the hydraulic forces 5 present.
6 The other EQ work that I am aware of in the 7 mechanical area is a test that has been going on for the past 8 two months at Wyle Labs in Huntsville, and there they are 9 testing some, I believe there are around six or eight inch 10 motor-operated valves under blow-down forces simulating what 11 would happen if you had a breakdown stream of the valve and 12 intenled to close it.
~s 13 And those tests as far as I know, well, the last (J \
14 test is being conducted, as a matter of fact, this sek. Now 15 in the future, it is my understanding that the mechanical 16 equipment qualification research is budgeted somewhere in the 17 vicinity of about a million dollars for Fiscal '89. That work 18 continued at INEL and they are devoted primarily to looking at 19 containment isolation valves of one kind er another, and their 20 response to generally more severs accident loads than a DBA.
21 As far as the EQ research is concerned, there has 22 been no specific plans made to fund any EQ research in '89.
23 The option has always remained, and been understood that if as 6 24 a result of the risk study conducted by Sandia for Moni, if l
(~)
'~'
25 there was a strong need to reintroduce or re-initiate some 1
HERITAGE REPORTING CORPORATION -- (202)628-4888 L
~
1 e
40 i
1 form of electrical research, that funding would have to be
- 2 established by the management through. discussion and-
{
3- reallocation.
4 That pretty much is where the work stands as far as 5- Fiscal '89. Now what you might look for, too, in the way of
'6 finishing up, the one task that is not finished but is almost 7 there, is the one at Sandia in which we were looking at the 8 mechanistic NUREG 0956 and have determined the radiation 9 levels for a series of accidents from'DBAs to severe accidents 10 for various' locations in containment for a typical PWR and a 11 typical BWR.
12 That work was done by BCL and is finished. Sandia 13 took the output from the BCL source term analysis and has been 14 doing dose calculations for both gamma and beta radiation, and 15 those calculations are essentially done, and the report is in 16 a draft stage, and hopefully after it gets through peer 17 review, will be published later this year.
18 I don't think we are prepared to comment other than 19 there does seem for severe accidents to be a slight increase 4
20 in dose levels over those generally assumed in the Reg Guide 21 189, Rev. 1, which was published in support of the rule 50.49, 22 but that seems to be related to the fact that the new source 23 term gives far more credence to the particulate release in 24 containment than was given by the TID 14844 terms which had a O 25 deterministic 1 percent, but it is not a major increase, and I FERITAGE REPORTING CORPORATION -- (202)628-4888
. _ . . . _ . . . , _.__.-_-.___-____..._______..,_.-_.a _
41 1- don't at'the moment anticipate really'any change in the EQ 2 rule as-a result of'this study.
}-
3 .The other items were in complete reports which have 4 been finished up by Sandia. One was on'the post-accident 5 monitors where we conducted a high_ dose' rate test. This was
~
6 done because one of the health physics 1 scientists at PNL had
~
7 made studies indicating that internal heating within the >
8 gaseous chamber when superimposed on the steam temperutures of 9 a LOCA would lead to temperatures high enough to destroy the 10 post-accident monitor, and mostrof the EQ qualification work 11 :has been done, as you know, at levels like half a megarad per 12 hour, as we exposed a chamber to a simulated LOCA at about 4 13 megarads per hour, and the chamber behaved fine, so as far as 0
14' we could determine there, the concern was not real.
15 The other piece of work which has been finished and 16 published is, of course, a severe accident--not a severe 17 accident, but the hydrogen burn equipment survivability which 18 is sort of another issue, and that has been finished. We have 19 got one more report produced, and that was the tests of 20 various electrical penetration under severe accident
-21 c ondi tio.'s , which was done at Sandia. All the QLRs and 22 testing was done, and we have been having trouble getting the l-23 report written because the principal scientists have gone off 24 on SDI programs. We think we will get that report, though, 25 Sandia has promised to get that report out by the end of this
' HERITAGE REPORTING CORPORATION -- (202)628-4888 o
42 1 year by reassignment of somebody else.
(} 2 That basically is it. Now there are some areas like 3 fire and seismic where there is a small amount of cork going 4 on; nothing on fire research as such other than a very small 5 program looking at the effect of aging, and on seismic, there 6 is a very small effort under the aging program being done at 7 Brookhaven Labs and that seismic was not environment, include 8 environment. I think that is pretty much it as far as the 9 areas that you have addressed in terms of equipment survival.
10 MR. MICHELSON: I wonder if I might ask Sandia a 11 question then. I would like to ask you a question.
12 When you did your EQ study, did you look at any of 13 these solid, the possible problems with solid state t7 ,i sJ 14 electronics in the building under harsh environmental 15 conditions?
16 MR. BOHN: No, we did not.
17 MR. MICHELSON: So if you don't even look at the 18 particular problem area, and yet Research plans its future 19 actions only on EQ study, I think you are going to miss some 20 big boats because the Sandia study was still rather limited, 1
21 certain components, certain kinds of problems. That was not 22 overall or all-encompassing study, as near as I could tell 23 from reading it.
24 MR. BOHN: That is certainly correct, and it started 25 off focused on so-called the traditional equipment HERITAGE REPORTING CORPORATION -- (202)628-4888
L 1
!^ , 43 1 qualification protocol issue, simultaneously irradiate versus (p 2L temperature, can you do them sequentially? What-order? And 3 soon we found that those were second order of effects to a 4- large extent. We got into the'other issue--
5 MR. MICHELSON: Starting facing down the avenues
'6 that led you to--
7 MR. BOHN: That is correct.
8 MR. MICHELSON: But you didn't go and look for other 9 avenues that had not been worked on before, and still out 3 10 -there in the unknown, and then now Research uses that as the 11 besis for any future program, I think they are going to miss 12 potentially some boats.
13- How are you planning to cover these other areas to 14 be sure that they aren't outliers that could give you serious 15 difficulty?
16 MR. DEY: Let me clarify that this study was set up 17 primarily to determine the risk impact need for the EQ 18 research electrical equipment.
19 MR. MICHELSON: Look at solid state electronic l
20 equipment used in breakers, used in adverter, used in control 21 cabinets, used extensively throughout some of the plants.
22 Didn't even look at that area as near as I could tell.
23 MR. BOHN: Let me say the whole area of 24 instrumentation and control is one that we did look at under 25 the general heading of instrumentation and control without HERITAGE REPORTING CORPORATION -- (202)628-4888
-. ,. .. .. - - .. - --, _ . _ . . - . , . - . . . . - - - - - - . -. -, - . , - . - . ~ . . - . . . . - _ - . _ . . -
44 1 trying to go down to subcomponent level because we didn't have 2 any reliability information on it. 3 One of the conclusions of the study is that common 4 mode failures of instrumentation can play a real significant 5 role, but we couldn't go any further with it with the 6 information and the level in which they are included in PRAs, 7 but that is one of the conclusions of-the study, that that's a 8 whole area that probably deserves a study of its own, and 9 that's, the solid state aspects is Just one aspect of the 10 overall policy. 11 MR. MICHELSON: Research is going to pick up on that 12 observation and look into it or have somebody else look into 7- 13 it or do something to make sure that that is not an important (v) 14 avenue to be pursued? 15 MR. DEY: As Mike mentioned, we have looked at some 16 areas. We have some conclusions, but let me point out that 17 this study was not aimed at and was not planned to address all 18 the areas, so this study would be used as one area. 19 MR. MICHELSON: I don't see any future plan here so 20 far in the work along those lines? Through '89 at least I 21 didn't hear that you are doing anything on it. I think you 22 said no EQ research in '89 on electrical. 23 CHAIRMAN WYLIE: That is funded; they haven't been 24 allocated any funds to do any-- 25 MR. MICHELSON: That to me equates to no work. HERITAGE REPORTING CORPORATION -- (202)628-4888
45 1 CHAIRMAN WYLIE: Well, that's right. I think what 2 Moni said a while ago, if I understand him correct, he was U'~T 3 looking for us to give some recommendations in this area based 4 on the studies, and-- 5 MR. MICHELSON: I just want to make sure I 6 understood what I think Sandia, how far they had gone and they 7 didn't go very far and for good reason, but I think it is an 8 area that DOD I think will go on this, and I think it is 9 something that perhaps the Committee ought to consider whether 10 or not we think it is important, and if we do, then it ought 11 to go in our letter. 12 DR. SIESS: We have got to recommend--when we make 13 recommendations, we don't supply the monies. We don't assign 14 the priority. I put this ahead of most of the so-called 15 severe accident research that is being done. 16 MR. MICHELSON: This is what is going to get us into 17 a severe accident if we don't head it off at the pass 18 potentially. That's why you ought to look into it. It has a 19 potential to get you into some very bad situations, and we 20 have seen small evidence of that already just in normal plant 21 operation. 22 MR. DEY: This study did address that particular 23 question. 24 MR. MICHELSON: But it cut off very quickly into the 25 development of it, and I didn't hear you picking it up as an HERITAGE REPORTING CORPORATION -- (202)628-4888
46 m 1 area that you would want to' pick up on,-ao I couldn't, you {} 2 know, I just had to use the opportunity to make the speech.
. 3 MR.. FARMER: There is some activity, there are 4 activities planned under the aging program which don't address 5 the accident part of it, but do address the reliability.
6 aspects,- and instrumentation and control is one of the areas 7 that Oak Ridge supposed to look at in the Fiscal '89 aging 8 activity, the question of, and it would address the aging 9 -aspects of' electronics and solid state devices, but most of 10 the aging programs look at reliability from the point of view 11 of normal operations. They do look at normal operating 12 temperatures, and high cabinet temperatures, but they wouldn't
- 13 necessarily look at accident conditions.
14 MR. MICHELSON: From the viewpoint of 15 instrumentation, this Sandia report did touch a little bit on 16 another very serious potential problem with instrumentation, 17 and that is this--you have got all the hot and cold legs of 18 the sensing line, levels and so forth, under rather elevated 19 conditions during normal operation, and when you go down the 20 reactor, you are going to go down to the lit.e. You get done, l 21 I am not sure what you have in terms of sensible 1 22 instrumentation, even on steam generators and so forth in the 23 steam line break, and you touched on this a little bit, but 24 probably don't have good information with which to decide, but O 25 it is a very fine example of equipment that seems to be l HERITAGE REPORTING CORPORATION -- (202)628-4888
-. ~ . _ , _ _ _ _ . _ _ _ . . . _ . _ . . . . _ ._ . . . . _ . . . _ . _ _ . . _ _ . - . . , _ . _ . . ._
47' 1 running fine, tested fine, and once you have the accident, it
} 2 becomes potentially disturbed, and what state I don't know.
3 MR. BOHN: That's especially true for delta P 4 instrumentation where you are dealing with small differences 5 in pressure, ringing of the lines, temperature differences, 6 those types. 7 MR. MICHELSON: All steam lines blow down from one 8 -break for a while. All the valves get closed. I don't know. 9 You get flashing on the good generators, the operator doesn't 10 know'which line is broke or what to isolate or how to maneuver 11 his feedwater because all the instruments have gone crazy, all 12 common mode from the blow-down. 13 We haven't had this happen yet because we don't get 14 many big breaks. Once it happens it is too late to find out 15 that none of it works. 16 MR. BOHN: That is an example that we will go 17 through later on in the presentation. R18 CHAIRMAN WYLIE: You said the work being done on the 19 aging program was, electronic equipment was under normal 20 conditions, not DBA? 21 MR. FARMER: Not DBA; basically all the aging 22 program for the most part address normal operating conditions. 23 In fact, the only program that goes into really accident 24 conditions is the cable aging program, which is being done at (n.) 25 Sandia, where cables are being aged to simulate 20, 40, HERITAGE REPORTING CORPORATION -- (202)628-4888
48 1 60-year life, and then in order to determine whether they are 2 Jtill qualified, will be run through a DBA or a LOCA type (~')i L. 3 accident simulation. 4 CHAIRMAN WYLIE: But no equipment has been looked at 5 in that standpoint? 6 MR. FARMER: No, not except in terms of whatever 7 exists in the literature, but basically the focus in aging is 8 on operational reliability and operational aging. 9 CHAIRMAN WYLIE: It seems like to me that, I guess I 10 missed that along the way. I thought that they were looking 11 at aged equipment and their capabilities to cope with DBA. 12 MR. FARMER: They are, but only from the point of
'3
. view of what is existing. What I meant was there is no 14 specific testing work being ontemplated with aged equipment 15 to determine whether it would survive under DBA. They look at 16 it from the point of view of whatever exists in the 17 literature.
18 MR. MICHELSON: An example of that are air-operated 19 solenoids or solenoid valve for air systems where you have got 20 maybe a kind of oily air system or some plants have, and they 21 run along that way for ten years, accumulate the crap inside 22 the valve, and then they suddenly expose that valve to 23 elevated temperature from the inside of containment from an 24 accident, and we don't know whether that valve will work O O 25 anymore or not. That may become varnish. That oil becomes HERITAGE REPORTING CORPORATION -- (202)628-4888
t 49 l 'l varnish as soon as you elevate the temperature a ways. k' ("> 2 MR. FARMER: I should mention one piece of the EQ
%.)
3 research which is also finishing up, in fact I have the report 4 on my desk, is the work that Franklin did on the solenoid 5 valves. 6 MR. MICHELSON: Did they try using aged equipment? 7 MR. FARMER: We got valves from, if I am correct, 8 from Arkansas Nuclear and from Duke Power, and the Duke Power 9 were Valcors and Arkansas were ASCOs. We bought new valves of 10 the same model, artifically aged them, trying to simulate the 11 same life or which had been order of four to ten years, 12 depending on which set of valves, and we ran those through 13 DBA. We had e horrible time. It was very, very difficult. 14 We first placed the ASCOs. We spent almost six 15 months trying to simulate aging, and every--we kept dropping 16 the--and every time we did the lasimers would deteriorate, and 17 hang the whole valve up, and we would go back and replace it, 18 and lower the temperature and do other things, and we finally 19 gave up. We simply just didn't have time or money to 20 continue. 21 MR. MICHELSON: Did the old valves survive fine? 22 MR. FARMER: We ran the old valve through the test 23 along with the new valve, and they behaved for a short time 24 -all right, first few hours. They leaked badly, though, after 25 about like 16 hours. The ASCIs we also-- HERITAGE REPORTING CORPORATION -- (202)628-4888
' ~ '
,x 50
)
l' MR. MICHELSON: Leaking badly.is bad news if the 2 valve has to be on on accumulator.
'[
~
3' MR. FARMER: The ASCIs we had a lot of trouble with, 4 too, but it was primarily-due to sticking. The ones that we 5 got,1we got two sets =from ASCO. One set had, we.found the 6 first time we tried them, that they 9 tick, and we disassembled
-7 them and there was some sort of organic material.
.8 MR. MICHELSON: These are, have been routinely 9 surveyed, though, in the plant, haven't tney?
l I 10 MR. FARMER: . hey had removed them and they had 11 cxperienced sticking problems in the plant. We cleaned _them 12 all up, and ran them ;hrough. From then on they-- 13 MR. MICHELSON: You don't want to clean them up when 14 you run your test. 15 MR. FARMER: We had never run a test with them the 16 way they we. because they were, they just stuck. You 17 couldn't run a test. 18 MR. BOHN: In the seismic testing area, the issue of 19 cleaning up relay contacts, run a test, and they get relay 20 chatter, they go in and clean or file the contacts, et cetera. 21 Then they work the second time around, and that again is not 22 very reassuring because that's not going to happen when they 23 are installed in the plant. I have seen this at-- J 24 MR. MICHELSON: Earthquake is going to clean away 25 contacts. HERITAGE REPORTING CORPORATION -- (202)628-4888
_1 - 51 1 'HR.,-BOHN: I have seen this in reviewing fragility, ,
'2' arrive at various seismic-PRAs.
F(~ 3l MR. MICHELSON: In the aging program, it doesn't 4' ~ seem that we' coordinate the aging effects with, you know, 5- accident response under severe accident, EQ kind of-issues.
-6 They don't.seem to equate.that way too well.
7- MR. FARMER: On the--I will just quickly finish
~
8 up--on the SOVs,'on the ASCOs, it was the, you find it of 9 interest. The new valves, after we had aged them, also stuck, 10 and we had long. discussions with ASCO people as to what the 11 term was, and we were, we thought maybe they are putting _ 12 grease on some part of the solenoid, and they contended no, 13- but we are never able, so again we have to clean those up in 14 . order to get them to go through the test, and it finally 15 turned out only two valves out of, oh, I think of the order of 16 12 or 14 went through successfully, and one amczingly enough I l7 was a Buna end, had Buna end seats in it, which everybody 18 feels-- 19 HR. MICHELSON: Gness we better put this on the list 20 of agenda for the next valve meeting, the solenoid.
- 21. MR. FARMER: The other one was EPDM material which 22 we ran with a nitrogen blanket or nitrogen process gas, 23 because one of the things we had, were trying to find out is 24 some manufacturers use nitrogen to, in their qualification O 25 process. Others use oxygen, and we wanted to know if you do 5
HERITAGE REPORTING CORPORATION -- (202)628-1888
, . . , . . . ~ . . , . . , - , - . . - , _ , . , . - .- . , . - - . - - . - - - - - . - . _ - _ . . - - . . - . - . . - . ---
52 1 this, does, is the qualification really a valid qualification? 2 And we found that there was a distinct difference. The ({ 3 oxygen, ones which were tested with oxygen deteriorated badly. 4 The EPDM with nitrogen didn't, indicating that oxygen played a 5 significant role. 6 Now if you have nitrogen supplying your process gas, 7 which it does in some instances, in the qualifications is 8 okay, but if you qualify it with nitrogen and then say you are 9 going to put it on the air line where you have oxygen, why it 10 leads to some questions, but overall, we had a great deel of 11 difficulty is all I can say, and we-- 12 MR. MICHELSON: None of this is in the accident PRA. 13 That data is still not useful for PRA purposes. It is only 14 very small sample, normal operational reliability. 15 MR. BUSTARD: We do touch upon this data, though, in 16 this program. 17 MR. MICHELSON: It was brought in a little bit and 18 mentioned, some of the problems, but once you elevate the 19 temperature inside of containment, y~s don't have a data base 20 other than the one in the EQ program and-- 21 MR. FARMER: Goir.g back to Charlie's question about 22 the aging program in DBA, if you look at most of those Phase 1 23 reports, where they have surveyed the behavior of different 24 components, you will . ir.d that the statistics they have used 25 to compile operational reliability have been from LERs and HERITAGE REPCRTING CORPORATION -- (202)628-4888
53 1 NPRDS data file. You will find very little discussion of 2 accident environments on the behavior of aged components
-/(_)N~
3 primarily because the data in many cases just doesn't exist 4 other-than EQ data. 5 CHAIRMAN WYLIE: Thank you, Bill. Maybe we should 6 move on and proceed. 7 MR. BUSTARD: My name is Larry Bustard, and I am 8 from Sandia Labs in Albuquerque, New Mexico, and what I am 9 going to be doing right now is giving you an introductory, 10 some introductory presentation on the EQ risk scoping study, 11 and I would like to start with a summary of project objectives 12 and what some of the EQ issues were that we considered in this 13 study. 14 So to begin, and some of these we have been , 15 discussir.g already this morning, our scope is electrical 16 aafety-related equipment that must function durirg harsh 17 environments. Examples are things like cables, valve 18 actuators, pump motors. Not included in our scope is 19 mechanical and safety-related equipment that must function 20 during harsh environments. Examples would be check valves,
- 21 globe valves, et cetera.
22 (Slida) 23 HR. BUSTARD: Our objectives are to use existing 24 PRAs as the basis for study to assess the impact of electrical 2S equipment, enviionmental qualification or lack thereof on l HERITAGE REPORTING CORPORATION -- (202)628-4888 i
54 1 reactor risk and its uncertainties, and to identify any 2 analyses or testing that may be necessary to reduce the risk ("')s
\_
3 or its uncertainties stemming from the lack of qualification 4 of equipment important to safety. 5 (Slide) 6 HR. BUSTJ"D: As Moni has been mentioning this 7 morning, we havo been looking at different types nf equipment 8 qualification issues. One type of issue is the, associatad 9 with NUREG/CR-4301 issues. NUREG CR-4301 was a status report 10 on equipment qualification issues research and resolution. 11 That was written by Sandia Labs to summarize a number of 12 equipment qualification issues that have been investigated 13 over a several year period with NRC funding. 14 In essence, many of these historical issues that are IS diccussed in this report focus on design basis event equipment 16 qualification practices, and ask how appropriate those 17 practicos are, and what I mean by that is that an EQ 18 practitioner has to make decisions on how to structure an 19 equipment qualification test from a practical standpoint. 20 Does he use a sequential environmental exposure to simulate 21 the accident conditions that would occur in an actual 22 accident? And so one question would be is that sequential 23 application of environments good enough to simulate the 24 simultaneous reality that might occur in a plant? 25 So as I said, this report focused on the historical HERITAGE REPORTING CORPORATION -- (202)628-4888
55 1 issues associated with design basis event EQ practices. 2 In addition to those issues, there are some issues fi s_/ 3 associated with the fact that severe accident PRAs are 4 different in some respects than this design basis event EQ 5 practice that is ongoing in the industry. 6 In particular, PRAs highlight important non-design 7 basis event accident environments sequences. Some severe 8 accident sequences, for example, have environments that exceed 9 the typical levels of temperature, pressure, or steam 10 conditions that would be associated with the design basis 11 event, environment, and if we are going to use PRA to try to 12 analyze equipment qualification issues, we need to deal with 13 thic difference between the PRA perspective and the EQ O 14 perspective. And hence, that's part of our scoping study. 15 A second major thing that has already been touched 16 on several times this morning is that if we are going to use 17 PRAs to analyze equipment qualification issues, somehow we 18 have to come to grips with the fact that EQ does not generate 19 equipment accident reliability information, and yet PRA being 20 a probabalistic risk assessment technique, uses equipment 21 reliability information in its analysis, so in essence then we 22 have a set of issues tha'c are the historical issues mainly 23 associated with a design basis event, EQ practices, and we 24 have the non-design basis event issues associated with severe 25 accident environments, important sequences that are beyond the
!!ERITAGE REPORTI'!G CORPORATION -- (202)628-4888
56
- 1 design basis, and the lack of equipment reliability
/~T 2 information.
(_) 3 MR. MICHELSON: Sounds like from what you have said 4 that if it is a non-design basis accident, then it is in the 5 severe accident category. 6 MR. BUSTARD: Not necessarily. 7 MR. MICHELSON: Not necessarily; it is the correct 8 way, and so again, you are fuzzying it up, but if it is not 9 design basis, it is something else, and right now we call that 10 I don't know what. 11 DR. SIESS: Non-design basis. 12 MR. MICHELSON: Non-design basis might be a good way s 13 or thinking about it because some of them aren't--in other 14 words, the severe accident rule I think was not intended to 15 cover such things as a lot of these pipe breaks outside of 16 containmant and so forth. It wasn't--those are supposed to be 17 mitigated within Chapter 15 kind of results. They aren't 18 supposed to go beyond. We have a water line break outside of 19 containment, but yet it is a non-design basis I guess accident 20 because it is not antlyzed in Chapter 15. 71 UR. SIESS: We have to distinguish between 22 non-design basis and beyond design basis. 23 MR. MICHELSON: Yes. That's sort of the idea. 24 MR. BUSTARD: That is one reason why we chose the O 25 word non-design basis, because we recognize that not HERITAGE REPORTING CORPORATION -- (202)628-4888
~ - _- - .- -
, 57
[., 1 everything was'beyond' design basis. 2' HR. MICHELSON: But you threw them all together,in 3 your discussion just a moment ago, and in the report.it is 4 ' thrown together. .It is not clear. This never comes out. : 5 MR. BUSTARD: Okay. Well', as I mentioned,-if we go 6 .to NUREG/CR-4301 -there are a number of-issues that were 7 highlighted by that report in your handout. I give all 30 P 8 issues that are highlighted byfthat report. I don't want to
- 9. go through them all right now. I will just throw one vugraph 10 ;up here on the machine to give you a feel for what types of 11 issues that report raised.
-12 For example, here are several of the issues that 13 were raised with respect to methods.for simulating accident O 14 conditions. It is the sequential exposure of equipment to 15 radiation followed by steam which is the approach used by the 16 IEEE standard, and adequate simulation of the accident 17 conditions in a plant, or are synergistic effects important?
18 That will be an example of one issue that gets extensive 19 discussion in NUREG/CR-4301, 20 Another one which Bill Farmer mentioned has some 21 work being finished up, getting finished upright now, would be 22 can gamma radiation source be used to simulate the mixed 23 accident beta / gamma radiation conditions in a plant? 24 If you look at the calculations that are done to 25 characterize a design basis accident radiation condition, the , HERITAGE REPORTING CORPORATION -- (202)628-4888 :
58 1 equipment, much of the radiation is beta radiation, yet in an
~T 2 equipment qualification testi that beta radiation is simulated (G
3 using gamma radiation, so another one of the historical EQ 4 issues that is applicable both for a design basis event type 5 environment as well as a severe accident or a non-design basis 6 event accident would be this issue can a gamma radiation 7 source be used to simulate the mixed beta / gamma radiation 8 conditions during an accident? 9 (Slide) 10 MR. BUSTARD: There are a number of other issues 11 discussed by NUREG/CR-4301. They include things like how can 12 we appropriately simulate the aging environment? Do we need 13 to consider dose rate effects during aging? Can we employ 14 things like the Euremas methodology to do thermal accelerated 15 aging, or must we use something different than that, et 16 cetera. 17 There is also some special topics that were pointed 18 out in NUREG 4301. One of them which has direct bearing to 19 the issue of reliability in the use of PRA is this first one-- 20 are there test approaches such as multiple sample, fragility, 21 and overstress testing, which can complement qualification 22 type testing to achieve a broader safety perspective? 23 As I have said earlier, one of the issues we have to 24 deal with in this study is the fact that PRA by its nature O 25 using reliability information to generate its conclusions and HERITAGE REPORTING CORPORATION -- (202)628-4888
w L 59 t. I 1 EQ does not provide reliability information when it does its rg 2 type testing. L-) 3 (Slide) 4 MR. BUSTARD: Some examples of the non-NUREG/CR-4301 5 type issues that are more associated with the fact that we 6 move beyond the design basis accident when we talk about PRA 7 assessments of risk, would be the following--we have some 8 important non, some important accident environments, 9 sequences, that are not considered in the current 10 qualification process. 11 For example, right now there is a lot of emphasis 12 and discussion within the PRA community and within the NRC 13 regarding steam venting from BWR dry wells and suppression O V 14 pools to the reactor building, and when one vents from the dry 15 well to the reactor building, one creates a steam and high 16 temperature environment in that reactor building that 17 includes--this is more severe than what, let's say your solid 18 state electronics have been qualified for or potentially more 19 severe than what they have been quelified for much. 20 There is also severe accident stean pressure and 21 radiation conditiont ,at are not enveloped by the historical 22 EQ conditions that may occur inside containment, and finally, 23 we have in PRA models the use of non-safety related equipment 24 that is not qualified. For example, some of the older plants O k> 25 use components to achieve feed and bleed that aie not part of HERITAGE REPORTING CORPORATION -- (202)628-4888
60 r 1 EQ master lists. (} 2 The Surry analysis as part of NUREG 1150 took 3 partial credit for containment fans that were considered by 4 Surry not to be safety-related, so we have issues associated 5 with basically PRA versus EQ perspectives. I have mentioned 6 the reliability issues. The question that's appropriate for 7 this study is what PRA risk uncertainties occur because EQ 8 does not assure that normal operation reliabilities will be 9 maintained during harsh environment conditions? 10 And PRA typically employs normal operation 11 reliabilities to describe harsh environment equipment 12 operation. 13 CHAIRMAN WYLIE: Let me interpret you. Would this 14 be a good place to break before you get into the rest? Why 15 don't we do that? Let's take-- 16 MR. BUSTARD: Sure. 17 CHAIRMAN WYLIE: Fifteen minute break. 18 (A brief recess wa; taken.) 19 CHAIRMAN WYLIE: Let's resume our meeting. 20 MR. BUSTARD: Right before the break, we were 21 talking about project objectives and EQ issues, and simply in 22 summary, PRA insights and analysis technique are to be used to 23 assess the risk significance and uncertainty associated with 24 equipment operability issues, and the equipment operability O 25 issues we examined include the traditional EQ issues such as HERITAGE REPORTING CORPORATION -- (202)628-4888 f
~ 3 61 1 those discussed by NUREG/CR-4301, and issues associated with (~T 2 EQ versus PRA perspectives and approaches regarding equipment
' 'u.)
3 opcrability during harsh environments. 4 (Slide) 5 MR. BUSTARD: In just a few minutes I would like ~;o 6 go through the project approach, but before I do that, I want 7 to spend just one or two vugraphs summarizing some of the PRA 8 and EQ perspectives that were important to developing the 9 project approach for this study, and in some cases, were I 10 rather important to the conclusions of the project. 11 From the PRA perspective standpoint, there were four 12 major factors that impacted our project approach and 13 conclusions. One is a PRA perspective of what is risk 7s
\
14 significant accident timeframes. From an EQ standpoint, 15 equipment tends to be qualified for much, much longer 16 timeframes than what PRA tends to consider to be important 17 from their perspective. For example, if you look at 18 containment fans, containment fans maybe he qualified for one l 19 year operability after initiation of an accident. 20 If you go look at cable qualification reports, some i 21 cable manufacturers have actually tested their cables for a 22 hundred days or 365 days after initiation of an accident, so 23 from an EQ standpoint, we have very long accident timeframes 24 that are considered in the qualification process for selected O 25 pieces of equipment. HERITAGE REPORTING CORPORATION -- (202)628-4888 I
r 62 JL The PRA perspective in contrast is that equipment r3 2 operability, most of its risk significance comes from
'NJ 3 operability'early after initiation of an accident sequence, 4 and that can be anywhere from a few hours to a few days, and 5 that's an important difference in perspective.
6 A second thing that I want to mention is that from 7 the standpoint of PRAs, PRAc model things like containment 8 fans, and containment sprays that are associated with 9 post-core melt accident management, but right now, there is 10 increased interest within the NRC and other places regarding 11 post-core melt accident management strategies, and as those 12 strategies are further developed, they will be incorporated 13 into PRA, and at that time, we can assess the risk 14 significance of what equipment is involved and the EQ issues 15 that are associated with it. 16 For the purposes of this study, we limited our scope 17 to current post-core melt accident ' management strategies that 18 are modeled by PRAs. We did not-- 19 MR. MICHELSON: Your item 1, I think when you 20 developed it in your report, tended to indicate there is a 21 short-term timeframe and a long-term timeframe. 22 I think there is also a very short-term timeframe 23 wherein equipment is exposed to yet another set of, kind of 24 operating parameters and hasn't yet--in other words, the 25 system hasn't settled down yet to that intermediate operation. HERITAGE REPORTING CORPORATION -- (202)628-4888
63 1 For instance, core sprays on a boiler when it comes (~ 2 on right after a big depressurization. That core spray line V) 3 is filled with steam and water together, and now you are 4 throwing in a cold water insurge into it. Its operability 5 isn't going to be the same as it will be after that line is 6 purged again and it settles down to an intermediate operation'. 7 Some of the short-term effects can be quite severe because 8 they are transient in nature, and you didn't seem to dwell on 9 operability of equipment and reliability during this 10 short-term transient, prior to reaching some steady state 11 accident operation. 12 MR. BUSTARD: We did identify the issue of flashing
- 13 in reference to like piping as one of the short-term phenomena 14 that might have an impact on equipment operability, in 15 particular the instrumentation that tells us what the pressure 16 and levels are inside after BWR PV. You are right that when 17 we focused our investigation further after the identification 18 process and said what five or six items are we going to look 19 at in more detail, we did not focus on that BWR reference like 20 detector issue.
21 MR. MICHELSON: That is one example of a number of 22 very short-term effects that disappear shortly thereafter but 23 are quite significant in the short-term. There is some 24 electrical transient problems also in the very short-term ' (') 25 having to do with the sequence loading of, on diesels and { HERITAGE REPORTING CORPORATION -- (202)628-4888 i
64 1 things of this sort, which get in;o some fluid hydraulic {} 2 problems which are quite interesting in the short term. 3 For instance, if you make a transfer, you shut the 4 pumps off, you also killed the power to the valves so that a 5 pump restarts under valves wide open condition, things of that l 6 sort, and perhups even partially drained pipes, draining was 7 going on while it was waiting for its restart in the sequence,
- l. 8 these are very short-term effects which introduce the new i
I 9 problems that are not seen if you look at the intermediate 10 time. 11 MR. BUSTARD: Okay. Okay. As I was saying, our 12 study is limited to the current post-core melt accident 13 management strategies as modeled by PRA. 14 Another important perspective from a PRA that 15 impacted our approach was that PRA typically assumed that 16 normal operation reliability applies during the accident 17 conditions. This is particularly true for the front end 18 portion of the PRA analysis, and we, we took that and had to 19 deal with that assumption when we developed our project 20 approach. 21 And finally, something that has been discussed today 22 already is the fact that PRA does not have in its models a lot l 23 of detail regarding plant status instrumentation and ho/ the 24 operator is going to respond to possibly erroneous plant A
\- 25 status instrumentation.
HERITAGE REPORTING CORPORATION -- (202)628-4888
65-1 (Slide) ("g 2 MR. . BUSTARD: From an EQ perspective, there were U 3 several issues or realities that impacted our project 4 approach. 5 One is the level of detail with which equipment 6 qualification describes equipment compared to the level of 7 detail with which PRA describes equipment. 8 Equipment qualification is typically concerned with 9 things like transmitters, cables, connectors, pressure 10 switches, motor-operated valves, solenoid operator, et cetera. 11 In contrast, PRA tends to model systems and does not give as 12 much attention to how the cable is going to respond in that 13 system. The system is more of a black box and it does not get O 14 down to the level of detail that EQ does. 15 Again, getting back to the differences between 16 equipment operability perspectives for EQ and PRA, I have 17 already mentioned that PRA's tend to assume normal operation 18 reliability during harsh environment conditions. EQ 19 unfortunately does not provide us with harsh environment 20 reliability information. EQ's approach is to assume that 21 systems are redundant, equipment is redundant, and therefore, 22 if I can do a type test on a single component, I can show that 23 I have achieved my safety margin. 24 Bill Farmer mentioned this morning a Research 25 program regarding equipment qualification source terms for the HERITAGE REPORTING CORPORATION -- (202)628-4888
66 1 - radiation environment, and what I wanted to mention here was /~% 2 that in the course of your study, we examined the different V 3 environments that one would achieve during the sequences and 4 asked ourselves are they more severe than what would occur 5 during the design basis event, the equipment qualification-6 type environments? And one runs into situations where the 7 steam and the pressure and the temperature conditions are more 8 severe, but on the whole, the radiation conditions are not 9 going to be substantially more severe even in the severe 10 accident situation than they are in the equipment 11 qualification situation and so that is just an input to this 12 study. 13 And finally, as has been mentioned this morning, 7s 14 PRAs deal with the more diverse set of accident sequences than 15 does equipment qualification, Station Blackout, ATWS events, 16 transients with losses of pressure and cooling. There is a 17 whole variety of sequences that are important to PRAs that 18 have not been used to develop the environmental conditions for 19 equipment qualification. 20 The way we dealt with that was to say that we have 21 this set of issues simultaneous versus sequential or normal l 22 reliability versus accident reliability, and that those issues 23 mostly transcend whether we are talking about the traditional 24 LOCA event or whether we are talking about other types of b
'd 25 sequence environments and so we can examine those issues for HERITAGE REPORTING CORPORATION -- (202)628-4888
67 1 all the sequences that are important to PRA. 2 HR. MICHELSON: Let me ask on that point, the first NJ) 3 approach I think was to qualify this equipment for the, for 4 the design basis events covered by Chapter 15 of the FSAR, but 5 have we gone beyond Chapter 15 in some cases? Have we indeed 6 gone outside of containment for high energy line breaks and 7 qualified certain equipment for that harsh environment? 8 MR. BUSTARD: Yes. The utilities do. 9 MR. MICHELSON: We go beyond the Chapter 15 but not 10 beyond the design basis? 11 MR. BUSTARD: Let me put it this way. NUREG 0588, 12 which was one of the documents that preceded 10 CFR 50.49, 13 very clearly requires utilities to examine events such as main 14 steam line breaks and high energy line breaks, and when you go 15 to a utility, there is a number of equipment outside 16 containment that is part of their qualification documentation. 17 MR. MICHELSON: Main steam and feedwater are part of l 18 Chapter 15 events, but high energy line breaks such as say the 19 auxiliary feedwater steam line would not be in' Chapter 15 20 because the thought always was that is bounded by the other, 21 and it is not. It is a breakdown, compartments far removed 22 from the main steam chase, so then we realize that one day 23 along with a number of other things--and we went back and did 24 the pipe break outaide of containment kind of analysis, but O' 25 when we did equipment qualification, have we indeed qualified HERITAGE REPORTING CORPORATION -- (202)628-4888
68 1 equipment for some of these harsh environments other than from ('T 2 main steam ~and feedwater? Gi 3 MR. BUSTARD: I believe-- 4 MR. MICHELSON: I think we have, but I am not sure. 5 MR. BUSTARD: I believe for a BWR they qualify for 6 your high energy line breaks outside containment, outside the 7 dry well. Yes. I have seen qualification reports. 8 MR. MICHELSON: Presumably the environment is a 9 little less severe maybe than inside of containment? 10 MR. BUSTARD: It generally is substantially less 11 severe than inside containment. 12 MR. MICHELSON: Now the other thing that we did for 13 pipe breaks outside of containment that I have seen, and it 14 says that sort of equipment is qualified for mild environment. 15 Then they define mild environment, but I, I wonder do they 16 actually do some environmental testing in that particular 17 environment, that mild environment? 18 MR. BUSTARD: Now what do you define as being a mild 19 environment? 20 MR. MICHELSON: I don't. The licensee defined the 21 conditions as a mild environment, and then he says the 22 equipment is qualified I thought for that. 23 MR. BUSTARD: Let me give you an example and here is 24 where we run into definition problems. If you go to let's say O 25 like Grand Golfs' equipment qualification submittal and you HERITAGE REPORTING CORPORATION -- (202)628-4888
I. 69 1 start going through that submittal and you look at what-type ('N 2 of equipment they are qualifying and for what environments
. \-)
3 they do qualify their HPCS pump for a, like 150 degree 4 Farenheit. peak temperatures t'nat occur very soon after the 5 start of a like type environment, that's not considered mild 6 environment in the sense that mild environment as I understand 7 it, is anything beyond normal environment. But that is-- 8 HR. MICHELSON: Mild is beyond normal? 9 MR. BUSTARD: Yes. 10 MR. MICHELSON: But up to what then? Beyond normal 11 and up to what is mild? 12 MR. BUSTARD: No, no, no. Wait a minute. 13 MR. MICHELSON: Mild is normal. 14 MR. BUSTARD: Maybe I have made a mistake here. 15 Mild is normal environments. 16 MR. MICHELSON: Okay. 17 MR. BUSTARD: Anything beyond normal environments is 18 qualified or harsh. 19 MR. MICHELSON: Harsh is.to some level. 20 MR. BUSTARD: So you get a small transient thermal 21 exposure up to to 150 degrees Farenheit being a basis for 22 performing a qualification effort. 23 MR. MICHELSON: During normal operation, is not 24- experiencing a normal environment by that definition because 25 normal might be up to 110 degrees for an instrument room but HERITAGE REPORTING CORPORATION -- (202)628-4888
'70 1 the room never gets that hot normally. Maybe it gets 80 or 85
{} 2 normally, but it could go up to 110 so the argument was it is 3 designed for 110 but mild doesn't, normal doesn't mean what it 4 normally'is at. It maybe means what it is designed for. I 5- don't know. 6 MR. BUSTARD: Right. Now you asked-- 7 MR. MICHELSON: The nomenclature gets very 8 confusing. Somebody tells you oh, that's qualified for mild 9 and/or that's qualified for harsh. I don't know what he is 10 talking about. 11 MR. BUSTARD: You asked the question earlier do they 12 have test data for some of these conditions like I just 13 referred to the 150 degree Farenheit, and in that particular 14 case, what they used was the test report developed for an in 15 containment applicaticn for similar components, so they had 16 tested something to 340 degrees Farenheit and then showed by 17 analysis that that created qualification for this 150 degree 18 Farenheit type environment. 19 MR. MICHELSON: It would be helpful some day if you, 20 somebody in the regulatory area got some of these definitions 21 straightened out as to normal operation, mild environment, 22 harsh environment, so forth, exactly what is meant, because 23 they use it in reports all the time and I'm never sure what 24 they mean because we have seen solid state electrical cabinet 25 not too long ago that was exposed within its design basis and HERITAGE REPORTING CORPORATION -- (202)628-4888
71 1 yet it went and did very strange things and it didn't even
/~ 2 reach its design basis temperature, but it was certainly an \~})
3 environment beyond its normal operating condition, up to--what 4 was that, Charlie? McGuire--190 degrees it got up to? 5 CHAIRMAN WYLIE: Well, the temperature inside the 6 control room-got up to 90 degrees, but inside the cabinet went 7 to 120. 8 MR. MICHELSON: But it was designed for 140 degrees 9 ambient room temperature I thought. 10 CHAIRMAN WYLIE: Yes. 11 MR. MICHELSON: And it only got up to 90 and it 12 already started its behavior, but it was beyond its normal 13 operating temperature which was I guess 80 degrees or 14 something like that.
- 15 CHAIRMAN WYLIE: Well, yes. Inside the cabinet it l
l 16 was up to 120, but on the components they really didn't know 17 what it got to. 18 MR. MICHELSON: These nomenclatures are, ought to be 19 squared away. 20 MR. BUSTARD: Okay. Well, I have mentioned some PRA 21 perspectives that were important to our project approach and 22 also some EQ perspectives that were important to our project 23 approach. Your vugraphs provide more detail on these, but 24 what I thought I would do so that we can catch up in time here O 25 is to move on to our project approach, having just given you a HERITAGE REPORTING CORPORATION -- (202)628-4888
72 1 short overview on each one of these issues. (s]- 2 (Slide) 3 MR. BUSTARD: And that's, the project approach is at 4 the back of your handout. 5 MR. BOHN: Second to the last slide in the first 6 handout. 7 MR. BUSTARD: Right. 8 MR. MICHELSON: Does that mean you are skipping all 9 the rest of the slides? 10 MR. BUSTARD: Skipping the analysis framework type 11 slides. What I have there is like three or four slides that 12 talk about the PRA risk significant accident timeframes, but I s 13 just verbally gave you some of the information that is on 14 those slides, and to help catch up here, I thought we could 15 skip the detail that is in all those slides. 16 (Slide) 17 MR. BUSTARD: The project approach is as follows-- 18 several candidate equipment operations that must be 19 accomplished in a harsh environment were used to scope the 20 risk significance of EQ issues. 21 And what I mean by this is first, we chose 22 appropriate equipment operations that had to be performed in 23 harsh environments. 24 Second, we attempted to characterize the risk O 25 significance of those equipment operations, and also to HERITAGE REPORTING CORPORATION -- (202)628-4888
73 1 characterize which of the EQ issues were most applicable to 2 those QA operations, and from those two sets of ( }. 3 characterizations, we developed an. assessment of the risk 4 significance of the various EQ issues. 5 To go into a little bit more detail regarding that, 6 we developed a list of candidate risk significant equipment 7 operations whose operability is desired during harsh 8 environments. Allen will be talking in more detail about that 9 in a few minutes. Basically it involved looking at historical 10 risk achievement analysis as practiced by current PRAs and 11 historical PRAs, consultation with PRA analysts, and through 12 an examination of the timing of accident progression, and 13 mapping the likely risk significant equipment with harsh 14 environment locations with respect to the timing of that 15 accident progression. 16 (Slide) 17 HR. BUSTARD: Once we had a list of candidate risk 18 significant equipment operations, we reviewed qualitative EQ 19 data sources to identify those EQ components for which 20 accident reliabilities may differ from normal operation 21 reliabilities. 22 MR. MICHELSON: In this. regard, did you look at 23 inside of containment where you have subcompartments that 24 might be exposed to, to rather short-term but more harsh 25 conditions than, you know, than in the longer term, as pipe HERITAGE REPORTING CORPORATION -- (202)628-4888
74 1 breaks and depressurizes into a suLcompartment and it finally (~T 2 depressurizes on into total containment? Generally I thought U 3 the numbers that were used for EQ were the mixed mean total 4 containment temperatures and pressures and so forth, but do 5 you actually qualify on subcompartment conditions, transient 6 conditions, or on steady state containment conditions? 7 MR. BUSTARD: Qualification, for example, with 8 respect to the radiation source term that is used for 9 qualification, is based on an instantaneous release to the 10 containment, and the qualification number that is typically 11 used is a center of containment value. 12 MR. MICHELSON: That's a mixed mean value? 13 MR. BUSTARD: That is a mixed mean value. 14 MR. MICHELSON: How about in the case of 15 temperatures and pressures? 16 MR. BUSTARD: Generally what the utilities do is 17 they try to come up with bounding calculations so they may 18 examine several different sequences and different compartments ! 19 and then pick a bounding calculation which covers the 20 different sequences and in the different compartments. 21 MR. MICHELSON: They find the worst case? 22 MR. BUSTARD: Yes. 23 MR. MICHELSON: Worst in the transient sense or 24 worse in the steady state mix means sense? O\# 25 MR. BUSTARD: I would have to go back and check HERITAGE REPORTING CORPORATION -- (202)628-4888
75 1 that, but an example would be a large break LOCA is going to
/~'j 2 dump a lot of steam into containment rapidly, and then may
-V 3 cooldown more rapidly than what a small break LOCA would have, 4 and so the qualification profile that is used as a generic 5 qualification profile would have the rapid temperature rise 6 associated witn the large break LOCA, and the slower tail 7 decay in heat, decay associated with small break.
8 MR. MICHELSON: You missed the point. You have just 9 made it in a way, and that is talking mixed mean. 10 MR. BUSTARD: That is mixed mean. 11 HR. HICHELSO4: How about the surge line, the 12 pressurizer if it were to rupture, for instance, and elevates, 13 or the nozzle on the vessel or whatever that elevates a 14 subcompartment, very high pressure and fairly high temperature 15 momentarily? By momentarily I mean for maybe a fraction of a 16 minute. This could be severe enough to disrupt the component 17 for future use thereafter, but I don't find it included in the 18 equipment qualification. 19 MR. BUSTARD: I would have to go check a 20 plant-specific analysis. 21 HR. MICHELSON: That is really a question. Does 22 that include that transient effect, or only the steady state? 23 MR. BUSTA?0: I would have to check that. As I 24 mentioned, the first step in our project approach was to D' 25 develop a list of candidate risk significant equipment. t HERITAGE REPORTING CORPORATION -- (202)628-4888
76 1 The second step was to review qualitative EQ data t (} 2 sources to identify those components for which accident 3 reliabilities may differ from normal operation reliabilities. 4 This provided a data base to justify the harsh environment 5 parametric risk achievement analysis that was done in the 6 later stages of the project, and it also helped focus the 7 selection of equipment operations for further study. 8 Third step was to cbcoca several of the equipment 9 operations that were identified in step 1 for further study. 10 (Slide) 11 MR. BUSTARD: And then within the context of a 12 scoping evaluation, we evaluated in more detail the selected 13 equipment operations. For exar.ple, we determined how PRAs 14 model the equipment operation. We assessed for which 15 sequences and environments PRAs suggested equipment operation 16 was important. We with evaluated the risk impact of equipment 17 operations, and we assessed the applicabilities of the 18 historical equipment on practicality issues such as the EQ 19 issues and NUREG/CR-4301. ! 20 After doing this fo!! several equipment operations, 21 we were able to decelop some more general conclusions and 22 recommendations for the study. - 23 (Slide) 24 MR. BUSTARD: The project approach circumvents O 25 numerous constraints imposed by current EQ and PRA practices. HERITAGE REPORTING CORPORATION -- (202)628-4888
h 4 77 , 1 It provides a reasonable method for eliminating certain EQ 2 issues from further consideration, and it is useful-for
}
3 identifying the direction of further study. 4 (Slide) 5' HR. BUSTARD: In a minute, Allen will be describing 6 in more detail the first step, which was to develop a list of 7 candidate risk significant equipment, and after he does that, 8 I will be describing in more detail our evaluation in more 9 detail of selected equipment operations, and I will be 10 providing at the end of the day c summary of our general 11 conclusions and recommendations, but just to give you a feel 12 for what the general conclusions and recommendations are so 13 that you can think about them during the remainder of the 14 presentation, let me just quickly go over them in a short, 15 very synopsis way. 16 One general conclusion is that PRA accident timing 17 perspectives provide a basis for modifying EQ practices. 18 A second conclusion is that PRA perspectives suggest 19 that the importance of the accident radiation dose is i 20 overemphasized in current EQ testing. 21 A third conclusion is that when PRA perspectives are ! 22 combined with the existing EQ perspectives, research data 23 bases and perspectives, several historical EQ issues can be i 24 con-idered not risk significant. ; 25 A fourth general conclusion is that when PRA HERITAGE REPORTING CORPORATION -- (202)S28-4888 j
78 1 perspectives are combined with the existing EQ perspectives,
- /~l 2 several of the historical EQ issues have possible risk \_/
3 significance. 4 MR. MICHELSON: What you don't show is a >
- c. *sion 5 about things that haven't been thought significant.
6 MR. BUSTARD: There is a few more. 7 MR. MICHELSON: Go ahead. 8 MR. BUSTARD: Okay. Lack of information regarding 9 the reliability of certain components in accident environments 10 creates non-conservative bias in the important cident 11 sequences and base case core damage frequencies. There are 12 risk important system operations that rely on either
, 13 non-saf ety related equipments or must be performed during
_A 14 environments that are not characterized by the design basis 15 vent EQ conditions. 16 The risk importance of plant status equipment is 17 difficult to quantify using current technology and data 18 sources and the risk importance of certain EQ issues coulc not 19 be charecterized by the scoping atudy using existing PRA 20 models and EQ data bases, and I said we will get into more 21 details about these. 22 MR. MICHELSON: One of the things that you didn't 23 appear .o do and thae, is to look for potentially risk l 24 significant issues that have not been highlighted by the 25 previous PRA which led you down this particular path. There HERITAGE REPORTING CORPORATION ~- (202)628-4888 y rm- m w F
i 79 L 1 wasn't, in the old PRA you didn't pick it up in this study; (} 2 tor instance, the electronic equipment exposed-to non-design 3 basis environment. Is it significant? Maybe it isn't, but 4 you didn't really look. 5 MR. DEY: I think the methodology which Allen will ! 6 present is intended to cover some of that. 7 MR. MICHELSON: It could, but you think it did? 8 MR. DEY: I don't think .t covers all of it. 9 MR. MICHELSON: It certainly didn't cover the effect l 10 of elevated temperature on solid state control components. 11 MR. BUSTARD: Let me respond to that in this way. 12 If you go to let's say appendix A of the report, or if you 13 look at the way the analysis of individual equipment 14 operations was structured, your question is covered by the ; 15 analysis technique. Now you are right. We did not focus on 16 electronic equipment outsido containment and ask those 17 questions that are given in Appendix A for that equipment, and 18 therefore, we don't have conclusions for outside containment 19 equipment associated to a medium energy pipe break, but they 20 could be developed using this methodology. 21 It was a matter of choosing which equipment 22 operation we vere going to look at in more detail. When Allen 23 gives his presentation, one thing he will point out is that 24 there is a low pressure injection equipment located in the O 25 reactor building of 3 BAR that is potentially, will HERITAGE REPORTING CORPORATION -- (202)628-4888
80 1 potentially see environments after a BWR vents its dry well or 2 ruptures'its dry well in a TW sequence. That was one of the
-V(')
3 equipment operations that we considered analyzing further in 4 this study. We didn't have the means to do that from a time 5 and dollar sense, and that waald have addressed a number of 6 your issues because once we get into the reactor building of a 7 BWR, and we are talking about high temperature steam 8 conditions, we are basically enveloping some of the conditions 9 that you have been referring to this morning. 10 MR. MICHEISON: Not 1ecessarily--perhaps. The other 11 problem is you didn't look at non-safety related equipment and 12 its interactive effect. 13 In the case of high energy steam, I know at least 14 one plant that has rather extensive building heating 15 arrangement using steam from the auxiliary boiler and it is 16 not in any PRAs or anything. Wanders around and heats areas, 17 and so forth. That sort of thing is not there, but if it 18 ruptures, it certainly creates adverse environment in an area 19 where there is some solid state control equipment. Now 20 whether that is risk significant, I don't know, and I don't 21 know from this study whether that sort of thing is risk 22 significant either. 23 MR. KOLACZKOWSKI: I will address your comments. 24 MR- MICHELSON: Okay. O 25 MR. BUSTARD: Do you want to come up? HERITAGE REPORTING CORPORATION -- (202)628-4888
81 1 MR. KOLACZKOWSKI: My name is Allen Kolaczkowski. I (} 2 work for Science Applications International Corporation. We 3 are basically a contractor to Sandia National Labs on this 4 project. 5 I am one of those guys that you are attacking most 6 of the morning. I.am actually a PRA analyst. I am a PRA 7 practitioner, and so in a more practical sense yes, I am l l 8 having to deal with a lot of the very concerns and problems f 9 that have been brought up here this morning. 10 Part way into my present-tion, I will try to, Mr. 11 Michelson, address your comments about what did we do and what 12 did we not do, okay? And you are right. There are some very 13 definite limitations on the analysis, and I will try to say as f3 O to why, why we made those limitations. Perhaps they are not 14 15 spelled out as explicitly as they should be in the report. I 16 think that is something that we need to consider. 17 (Slide) 18 MR. KOLACZKOWSKI: So as Larry mentioned, what I am 19 going to cover is basically how we came up with what are the 20 potentially important equipment operations in both BWRs and 21 PWRs that could be risk significant, and could be impacted by 22 an EQ issue? And as Larry pointed out, because there are 23 certain limitations in the PRA process, et cetera, you will 24 see that in doing this, we had to go beyond what current PRAs D d 25 do or what current reports of PRA study cover and we had to HERITAGE REPORTING CORPORATION -- (202)628-4888
82 1 come up with another process, and I will try to describe that 2 process a litt.'.e bit. 3 (Slide) 4 MR. KOLACZKOWSKI: This so-called task 1 basically 5 was made up of two, two-prong approach, although within the 6 second approach there was actually also an elicitation of 7 expert opinion to see if in fact we were comp 2ete. The 8 purpose of the task again was primarily to identify from a PRA 9 perspective that important equipment or important operations 10 which could be impacted by a harsh environment, and therefore 11 could be risk significant. 12 HR. MICHELSON: Your view now is both inside and 13 outside of containment? 14 MR. KOLACZKOWSKI: That is correct, and you will see 15 that in just a little bit. 16 As I said, primarily involved a two-prong approach, 17 although within this second approach there was also the 18 elicitation of expert opinion to see if they were complete. 19 The first involved the literature review. We went l 20 out and said okay, what do PRAs say is the important piece of ! 21 (quipment? And we also looked at a veriety of studies of PRA 22 results where people have taken a variety of different PRA 23 results, and then tried to meld that together in one report 24 saying what are all these PRAs telling us? And I will have a, O 25 present a list in a short while as to the study we actually HERITAGE REPORTING CORPORATION -- (202)628-4888
1 83. l' looked at.
. j"') 2 We went chrough and did a literature survey to find
- U 3' out what did the PRA, what is the PRA world-basically telling l'
4 us right now is important? But as we pointed out, and I think f f
-5 we have.already mentioned a number of cases this' morning,'PRA ,
) 6 has a lot of limitation relative to the-being able to handle 7 EQ issues, anu o we had to come up.with-another wry beyond 8 what current PRA does to look at what equipment or operation 9 might be risk significant, might be affected by EQ issues, and 10 make sure'that those, that list of candidate equipment or 11 operation was put on our list. 12 So we came up with another process in which we 13 -looked at equipment importance from an accident mitigation or 14 prevention scandpoint, where that equipment was located, and 15 what potential environments might that equipment see and then 16 through a mapping process which I will go through in a short 17 while, you will see how we basically go through a formalized
. 18 process.such as this.
19 We cams up with yet a larger list and what just came 20 out of the literature review process, I will describe both of 21 those in this presentation. 22 (Slide) 23 MR. KOLACZKOWSKI: The literature review part of the 24 process primarily involved these four major steps to come up O 25 with a representative set of PRA studies to look at, and I HERITAGE REPORTING CORPORATION -- (202)628-4888 f_ .
84 1 will show you what that list is in Just a moment. 2 We looked at those studies for what were those 3 studies saying as to what is the risk significant equipment or 4 operations in both BWR and PhRs? Now unfortunately, though, 5 most of these studies, in fact virtually every one, although 6 some of them did look at EQ a little bit, nearly all the 7 studies did not look at what equipment is important and also 8 at the same time, say could EQ affect that equipment? And sc 9 we had to take this list of important equipment and we had to 10 ask ourselves could that equipment see a harsh environment and 11 therefore, EQ would have affect on the operability of that 12 equipme ,s 13 So the point is not everything that is important e ! s m/ 14 would always see a harsh environment, and so some subsets of 15 this list of important equipment is equipment that would also 16 see a harsh environment and so therefore EQ could perhaps 17 impact reliability of that equipment. We had to go through a 18 reocess of screening what came out of this literature review 19 for where could hsrsh environment exist, and as a result, come 20 up with some subset of that list of equipment. 21 MR. MICHELSON: Isn't there another block, though, 22 that says here is equipment which does see a harsh environment 23 which is not thought to be risk significant but because of 24 exposure would be? l
' - MR. KOLACZKOWSKI: That came out of this part.
25 HERITAGE REPORTING CORPORATION -- (202)628-4888
c 85 1 MR. MICHELSON: The used-- {} 2 ,MR. KOLACZKOWSKI: First part is what does the
, 13 literature say is important? And we have got to come up with 4 subset and because-of.your.very' concern, Mr. Michelson, we 5 said we have got to'come up with another process that says 6 what might be risk significant-isn't thought to be now because-7 of the EQ issues.
8 MR. MICHELSON: Okay. 9 MR. KOLACZKOWSKI: That's why we did the other. 10 Okay.
-11 (Slide)
? 12 MR. KOLACZKOWSKI: This is a list of studies that we 13 looked'at.- The current draft versions of the NUREG 1150 14 reports are represented. 15 Representative of some of the if you will state of 16 -the art in PRA, and the NRC sponsored studies--we had easy 17 . access to them, et. cetera, so we locked at those particular 18 PRAs in their current draft form. As the Committee may be 19 aware, 1150 is currently undergoing a revision based on both 20; NRC comments, public comments, whatever. The new 1150 reports 21 I guess will be out somewhere around the end of the calendar 22 year, but we did have access to the current draft versions, 23 and looked at those, and to see what those reports were saying 24 were in fact important eouipment or operations. O 2S And I won't go through the list, but you can see HERITAGE REPORTING CORPORATION -- (202)628-4888
86 2 there are a variety of both NRC sponsored work, also IDCOR work, and I will talk about that IDCOR report in just a little ['} 2 3 bit, and also some EPRI sponsored work down here, so we tried 4 to get'a wide variety of both NRC and industry sponsored work 5 to look at. 6 HR. MICHELSON: What is important here is which ones 7 of those reports correctly have even considered external 8 events? How many of those? 1150 did not. How about some of 9 the others and which ones did consider external events and did 10 they do a reasonably adequate job of it? 11 MR. KOLACZKOWSKI: I can't, I can't run down the 12 list and tell you yes, no, yes, no, although I do know some of
, 13 them. For instance, the TAP-45 work did.
14 MR. MICHELSON: A little bit. 15 MR. KOLACZKOWSKI: Certainly most of these I would 16 say the emphasis is on internal events. 17 MR. MICHELSON: So your focus then became driven to 18 internal events? 19 MR. KOLACZKOWSKI: On the literature review part, 20 yes. 21 MR. MICHELSON: Okay. 22 MR. KOLACZKOWSKI: One thing you will find in a lot 23 of these studies, not necessarily all of them, but in a lot of 24 these studies, is that they looked at things using two risk b'
'- 25 importance measures--the risk reduction and risk increase.
HERITAGE REPORTING CORPORATION -- (202)628-4888
87 1 This is PRA's attempt to look at how important might a piece f"% 2 of equipment be if its reliability were to be suddenly (_) 3 increaced significantly or decreased significantly? Okay. 4 But it usually does it on a component-by-component 5 basis, and again this is one of the limitations of the current 6 PRA process and the current PRA calculations that are 7 typically performed, and that is it does do these calculations 8 on a component-by-component basis, whereas when you are 9 dealing in a world of EQ, you are talking about an environment 10 that is affecting many, many different pieces of equipment all 11 at once, and so it is a much more of a broader common cause 12 effects. 13 HR. MICHELSON: You mean, well, one is common cause O 14 and one is single component failure. I think that is what you 15 meant by the first part of your statement, didn't you? 16 MR. KOLACZKOWSKI: Yes. Really-- 17 MR. MICHELSON: One failure at a time? 18 MR. KOLACZKOWSKI: These calculations are usually 19 performed only one failure at a time. 20 MR. MICHELSON: That gives you a quite different 21 answer. t l 22 HR. KOLACZKOWSKI: Certain amount of insights, but 23 not everything that we need. 24 MR. MICHELSON: Yes. 25 MR. KOLACZKOWSKI: Okay. So again, just keep in l HERITAGE REPORTING CORPORATION -- (202)628-4888
88 1 mind that a lot of the results that came out of the literature (} 2 review process or a lot of what the, these studies have done 3 is looked at things using this risk reduction and risk 4 increase importance measures where they basically, in the case 5 of risk reduction they say what about what would happen if you 6 assumed that the equipment was totally' reliable?- How much 7 would you reduce core damage frequency or risk and those 8 things that have a high value out if that importance measure 9 says that maybe if I can spend some money and resources to 10 make that equipment more reliable I get a big risk reduction 11 so it is worth spending some money in that direction. 12 On the other hand, the risk increase, you look at 13 what if the reliability of the equipment went to zero? That 14 is the unreliability of the equipment went to 1.0; how much 15 would risk or core damage frequency increase? Okay? And that. 16 tells you something about if that equipment's reliability is, 17 suddenly gets much worse than what the PRA nominally assumes, 18 that in fact it makes a big impact on the core damage 19 frequency, and using these two types of measures, quite often 20 these studies of PRA results have come up with a variety of 21 conclusions as to what might be important pieces of equipment 22 or operations. 23 HR. MICHELSON: From the environmental qualification 24 viewpoint, that information is of very limited value? O 25 MR. KOLACZKOWSKI: Understood. HERITAGE REPORTING CORPORATION -- (202)628-4888
89 1 MR. MICHELSON: Environment by definition is common 2 to a number of components and not iust to one component. 3 MR. KOLACZKOWSKI: This only gives you a portion of 4 the insights that we really need. 5 MR. MICHELSON: In fact it might mislead you? 6 HR. KOLACZKOWSKI: Yes. You have to be careful. I 7 am not going to go through all these slides. There is a, I 8 think about three, four, or five slides that follow this one 9 that goes through each one of our, of the literature review 10 results, and makes some comments relative to what insights 11 came out of the study itself and our comments on those, but if 12 you were to look at each one of these, what you will basically 13 find in most cases is that there is typically some shortcoming O 14 in the study relative to our needs and our use on this EQ PRA 15 project. 16 Very often in the comment celumn you will see our 17 comments say something like did not consider equipment's 18 potential for seeing a harsh environment. Okay. Or an 19 example, little was done on specific instrumentation and 20 subcomponent modeling. 21 So the point I want to make here is that although 22 there were some insights as to what are important pieces of 23 equipment that might be risk significant, and with us lacking 24 at it through EQ eyes saying yes, we could see certain 25 accident conditions where that equipment would see a harsh HERITAGE REPORTING CORPORATION -- (202)628-4888
90 1- environment, typically considering the harsh environment in a (" 2 direct sense in that study was not done, and so these people
?u.)T 3 were looking at what equipment is important, but they weren't 4 thir. king about the fact that could that equipment see a harsh 5 environment or not? And that was true of most of the studies 6 here, so that's why the literature review, although it gave us -
7 some insights, was really not appropriate for our needs and we 8 had to go on to this other formalized process that I will 9 describe. 10 Let me just jump ahead about three or four slides. 11 (Slide) 12 MR. KOLACZKOWSKI: Oh, I do want to make a comment 13 about the IDCOR work because that came up earlier at the 14 meeting as to what the IDCOR reviewed. The IDCOR task 17 15 raport did look at equipment survivability and degraded core 16 environment, and they basically'came up with a series of 17 equipment that potentially could see an adverse environment in 18 excess of qualification levels in certain accident sequences, 19 but all they really did was in fact first of all as you can i ! 20 see in our comments column ' hey only looked at a limited , 21 number of certain acts sequences. 1 22 Secondly, there was no perspective on what the 23 increased failure rates might be due to the environmental 24 stress imposed on the equipment. Okay. And hence there was 25 no means of coming up with a PRA perspective in a quantitative l HERITAGE REPORTING CORPORATION -- (202)628-4888
91 l senseDas to what the effects on core damage frequency might 2_ be. 3 In1this study what you find they basically did was 4 they went through a limited number of accidents and said wnat 5' kind of environment might this equipment see then.if that. 6 environment is below what the equipment is qualified for? 7 They assumed it would work at its nominal failure rate. 8 That's the bottom line. And so even though they did go and 9 look at the harsh environment a little bit,'again it fell far 10 short of really what our needs are on this program. And 11 again, there really is no mechanism right now in the PRA 12 process-for really handling these increased unreliabilities on 13 .the equipment and harsh environment because there is no data.
-14 .MR . MICHELSON: Did IDCOR consider concurrent' 15 failures?
16 MR. KOLACZKOWSKI: I don't think so. 17 MR. BUSTARD: IDCOR looked at one at a t'ime, and in 18 -fact they didn't look at'it from a system standpoint. They 19 decided there were like four components that they could not 20 prove that they would have equipment survivability for, sud 21 one of them was the cable, and they didn't say what was thst 22 cable connected up to? What system did that impact? It w3s 23 basically here are a sets of components, can I show that I can 24 somehow envelope the equipment severe accident environments by O 25 qualification environments, either through direct envelopes, HERITAGE REPORTING CORPORATION -- (202)628-4888
. . . . . . . . - - . . . . _ - - . , . . . . - . - . . - , . . . . . _ - - . - , - , - . - , = . - - , _ . .
92 1 or through heat transfer type calculations, et cetera? 2 So, for example, in that program they say it.may see
).
3 a couple of hydrogen burns, but we can show based on the mass 4 of the item that its temperature rise will never exceed 5 qualification parameters in the timeframe that it has to 6 operate, and therefore, it will survive this severe accident 7 environment. That gives no information on reliability and it 8 gives no information on system impact. It is just extending 9 qualification of severe accident. 10 MR. MICHELSON: One component failing at a time? 11 Hydrogen burn, of course, a number of components are 12 concurrently exposed to the same burn potentially. 13 MR. BUSTARD: Right. 7s b 14 MR. MICHELSON: They didn't try to integrate the 15 collective effects? 16 MR. BUSTARD: No. In task 17 report they basically 17 asked can we show equipment survitability? 18 MR. MICHELSON: Wasn't even directed at single l l 19 failure but rather just at the just show that the transient l 20 will not destroy the equipment? 21 MR. oUSTARD: Could we qualify this equipment to 22 more severe environments? 23 MR. KOLACZKOWSKI: So despite the fact that a 2^t of I '24 studies-- ( 25 MR. MICHELSON: Did IDCOR deal with outside of 1 l HERITAGE REPORTING CORPORATION -- (202)628-4888 l -
93 1 containment events, or just inside of containment events? (3 2 Just looking at what happens when you get the core too hot and () 3 it starts to do things? 4 MR. KOLACZKOWSKI: As I recall, it primarily 5 stressed degraded core environment kind of stuff. Didn't 6 think about--that's right. 7 MR. MICHELSON: Didn't focus on what was happening 8 out of containment at the same time? 9 MR. KOLACZKOWSKI: No. 10 MR. MICHELSON: It might be what caused degraded 11 core started outside of containment to begin with. 12 MR. KOLACZKOWSKI: Yes. So the point I have tricd 13 to make here is that there are some study of PRA results. 14 They talk about what could be important risk significant 15 either pieces of equipment or operations. Because they 16 weren't looking through the eyes of an EQ pen, and thinking 17 about EQ environments, they fall short of what our total needs 18 are in this project. However, there are still some valuable 19 insights that do come out of those studies. 20 They are listed here. That despite those 21 limitations, we could, the literature review did come up with 22 a, with us coming up with a list of equipment that in fact we 23 recognize could see a harsh environment in certain accident 24 scenarios and in fact these studies were saying yes, these are O 25 potentially risk significant pieces of equipment in the plant. HERITAGE REPORTING CORPORATION -- (202)628-4888
94 1 And you can see the list. I won't read it here for you. l ("T 2 But as we pointed out, because of these limitations V 3 of these studies, we had to go beyond the literature review 4 process and we had to try to go beyond or transcend the 5 limitations of the PRA process, and that's why we wanted to 6 get this other formalized process which I am now going to 7 describe. So with that, in summary, although we did identify 8 certain pieces of equipment that could be risk significant, 9 could see a harsh environment when they need to operate, we 10 needed to go beyond this literature review process, and that's 11 what I am going to describe now. 12 First when you look at this thing, it is a rather 13 complicated looking diagram, but let me try to take you 14 through if briefly, and explain to you what we did. 15 Okay. As we said, these studies have a number of 16 limitations so we needed to go beyond current PRA techniques 17 and try to think of things a little bit more in a global 18 sense, and be looking at things through these EQ eyes if you 19 will, and so we came up with this process to do that, and let 20 me try to take you through some of the major boxes, describe 21 what we are doing. 22 First thing we did was we defined a general class of 23 accident categories, and I will be presenting those for you in 24 just a moment, but they basically embody in a very general way O 25 all the possible accidents that could occur, okay, in the HERITAGE REPORTING CORPORATION -- (202)628-4888
95 1 power plant. 2 Enough said on that; I will come back to each one of-L'~T ) 3 these in just a moment. 4 Once you have this general class of accidents 5 defined, you then list what equipment might I need to operate 6 in order to either prevent that class of accident or at least 7 mitigate that class of accident so it doesn't become worse? 8 And that list of equipment includes both safety-related 9 equipment and non-safety-related equipment. Anything that the 10 operator might use in the way of indication, control, the 11 actual pumps and valves themselves, whatevar, that list of 12 equipment was made for each one of those classes of accidents. s 13 We then defined environmental boundaries in the (J \ 14 plant, and again this was done in a very gross way considering 15 this being a scoping study, and you will see how we took the 16 plant and divided it into basically four major environmental l l 17 areas. 18 We then said can we get our hazardous environment in l-19 each one of those boundaries? Depends on which kind of 20 accident we have. And this was not defined in such a way that l 21 we worried about are we above qual limit or below qual limits. 22 The point is if we could determine at all that the 23 environment in this environmental location could get worse 24 than the normal environment, and here by normal, Mr. O 25 Michelson, I mean what they normally see, okay? If there was HERITAGE REPORTING CORPORATIOli -- (202)628-4888
96 1 .any increase at all in temcerature, pressure, humidity, j }. 2 whatever, we said that's a harsh environment for purposes of 3 our definition in this study. 4 Once you have the accident category defined, you 5 know what equipment needs respond to it. You know where the 6 environment is going to get harsh for that kind of accident. 7 You need one other piece of input, and that is again because 8 in the PRA world, PRA defined systems in a more gross sense 9 than the EQ people do. That is, to a PRA person, a valve 10 really includes the valve and the valve motor and the 11 connector where the cable comes in and the cable and so on and 12 so forth. We had to take this list of equipment and break it 13 down to subassemblies to put it into EQ terms and worry about 14 where is the connector, where is the cable, that kind of 15 thing. Where are these things typically located relative to 16 these environmental beundaries that we defined earlier? 17 Once you know the accident you are dealing with, the 18 equipment that you are dealing with, where that equipment is 19 located, and where the environments could be harsh, you can 20 then say somethi.c about is this equipment going to be needed 21 to operate, to respond to this accident, and yet it is going 22 to see a harsh environment at the same time? If that's the 23 case, then I at least have the potential for that equipment to 24 be risk significant, No. 1, and second, I know it is going to O 25 need to operate in a harsh environment. HERITAGE REPORTING CORPORATION -- (202)628-4888
y 97 1 MR. MICHELSON: Did you look at all for the
<~T 2 equipment that although it is not needed for mitigation of the b
'3 event, it is necessary to protect it so it doesn't produce 4 unwanted actions?'
5 MR. KOLACZKOWSKI: I was going to, I am going to 6 come to your concerns one by one shortly, but to answer your 7 question right now, to answer 3;our question right, now the 8 answer la no, we did not cor. sider spurious actions of 9 equipment in this study, and I will try to address this as'to 10 why we didn't do that. 11 Okay. Onco you overlap all of this information, I 12 know now the accident, the environment, the equipment that 13 needs to respond and where that equipment is located. If I U 14 map all that information on top of one another, I basically 15 can then come up with a is this a potential EQ issue or not? 16 If this equipment is going to need to operate and 17 therefore could be potentially risk significant, but it is 18 going to see a harsh environment during that operation, then I 19 am concerned about it. When we came up with this list, 20 ultimately we did have a number of PRA and EQ people look at 21 the list to say can you think of anything that we messed? And 22 as a means of sort of a sanity check if you will, and so we 23 did ask a number of people to look at our list to see if they 24 felt we had missed anything. 25 MR. MICHELSON: Defining environmental boundaries HERITAGE REPORTING CORPORATION -- (202)628-4888
98 1 per se outside of containment events, that meant you chased the ventilation ducts and the sewer lines and whatever? {} 2 3 MR. KOLACZKOWSKI: Somewhat; I am going to define 4 that. 5 MR. MICHELSON: That is what you meant by trying to 6 define--what is the zone of influence of event? 7 MR. KOLACZKOWSKI: That is correct. That is 8 correct. 9 MR. MICHELSON: Okay. 10 MR. KOLACZKOWSKI: Let me go into each one of those 11 just a little bit, and I am going to start addressing some of 12 your comments, Mr. Michelson. 13 Okay. Again rather than specific accident 14 sequences, as normally defined in a PRA process, four generic 15 accident categories were considered in order to come up with a 16 means of scoping all the possible accidents that could occur, 17 and those four accident categories were these as they are 18 defined here. 19 This first catetory says that I have had some 20 initiating event in the power plant, but I am, at least 21 initially in the accident, I am cooling the containment and I 22 am cooling the reactor vessel. That would be, for instance, a 23 design basis LOCA accident, large LOCA inside containment. I 24 am putting a severe environment inside the containment, but O 25 suppouedly the equipment is responding to that. You can take HERITAGE REPORTING CORPORATION -- (202)628-4888
99 1 that accident and you can look at what is all the equipment rg 2 that needs to respond to that accident? Okay. V 3 Once you know what that list of equipment, you can 4 say now on the basis of that accident, where could the 5 -environments be harsh? And then you can start saying is that 6 equipment still likely to keep working given I have that harsh 7 environment wherever it might be? 8 Similarly I have these other categories where the 9 accident keeps progressively getting worse and worse. I could 10 have an accident in which the vessel is being cooled, for some 11 reason the containment is not being cooled. All right. 12 Ultimately that means that if there is certain equipment 13 inside the containment that I need to continue to operate, O 14 that environment is getting harsher all the time because I am 15 not removing the neat of containment even though I am cooling 16 the vessel. 17 DR. SIESS: You said containment is not being 18 cooled, but your slide says containment failed. 19 MR. KOLACZKOWSKI: I'm sorry. I'm sorry. Yes. 20 Okay. It fails because it is not being cooled. 21 HR. MICHELSON: There is another category then. 22 Case where the containment was in a failed state at part of 23 initiating event such as isolation valve doesn't close. 24 MR. KOLACZKOWSKI: That could be also. D' 25 MR. MICHELSON: Containment was already failed and HERITAGE REPORTING CORPORATION -- (202)628-4888
100 1 the vessel is being cooled fine, but the containment contents
.(] 2 were now being released side, v
3 MR. KOLACZKOWSKI: This is meant to encompass that 4 type of accident also. 5 DR. SIESS: That really means no containment? 6 MR. KOLACZKOWSKI: Ultimately it means r3 7 containment. 8 DR. SIESS: Release to the atmosphere. 9 MR. KOLACZKOWSKI: And you will see an example later 10 in our pres 1ntation of thrt kind of an accident, the so-called 11 TW accident in a boiler. 12 DR. SIESS: More ways of getting there than just
- f. 13 losing cooling or even failure to isolate, and have a burn, 14 explosion.
15 MR. KOLACZKOWSKI: Yes. That encompasses all of 16 those items. 17 MR. MICHELSON: The release may be to the 18 surrounding building or it may be to the environment outside 19 the building. { 20 MR. KOLACZKOWSKI: That is correct. 21 MR. MICHELSON: That gives you two different l 22 problems. 23 DR. SIESS: What is the significance of vessel l 24 cooled for that accident? No--primary system still in tact? l ("h
For instance, as lcng as you stay 25 MR. KOLACZKOWSKI:
HERITAGE REPORTING CORPORATION -- (202)628-4888
101 1 within this class accident, it does not get worse, progress {} 2' down to this kind of accident. You don't have the radiation 3 yet. 4 DR. SIESS: Vessel cooled means vessel is in tact? 5 MR. KOLACZKOWSKI: Yes. Actually means more the 6 core is in tact as well. 7- DR. SIESS: Primary system in tact you said? 8 MR. KOLACZKOWSKI: Yes. I am cooling the core. I 9 am removing the heat out of the core, but I am not, but the 10 containment has either already failed or I am not removing the 11 heat from the containment. 12 The point is the containment boundary ultimately 13 , gets lost either because of the initiating event or ultimately 14 gets lost. 15 DR. SIESS: If you are cooling the core, what is 16 heating the containment? 17 MR. KOLACZKOWSKI: It could be either I have a break 18 and I continue to take heat out through the break into the 19 containment or in a TW scenario, it is where I am cooling the 20 core but all the heat is going out the SRV to the suppression 21 pool. I am not cooling the pool, for example. Ultimately 22 that heat is just being constantly deposited in the pool and 23 containment pressure and temperatures are rising. l 24 CHAIRMAN WYLIE: Would that include bleed and feed? C:) 25 MR. KOLACZKOWSKI: Yes, but I am not removing the l HERITAGE REPORTING CORPORATION -- (202)628-4888
102 1 heat out of the sump water ultimately. 2 DR. SIESS: The vessel cooled is not the proper--
)
3 MR. KOLACZKOWSKI: No. The core is cooling. 4 Similarly I could be, I could be--the containment is okay, and 5 it is being cooled, et cetera. The containment is'in tact. 6 Its' integrity is there, but I'm not cooling the core, and 7 ultimately it is going to lead to vessel breach and vessel 8 failure. Okay. Now I am going to create the potential of 9 hydrogen burns insides the containment, and all that kind of 10 stuff. 11 DR. SIESS: You see, I think I know what you are 12 saying, but it says vessel fails. Change that--if you start 13 with LOCA, you already had a breach in the primary system. 14 MR. KOLACZKOWSKI: It should say core. It should 15 say core. 16 DR. SIESS: You mean the core is still in the vessel l 17 or not in the vessel? 18 MR. KOLACZKOWSKI: More than that, and more than 19 that, the core is okay. It is not melting. 20 DR. SIESS: Okay. l 21 MR. KOLACZKOWSKI: In this case, or it is not i 22 melting in this case. Here I am going to lose the core. Here l l 23 I am going to lose the core because I have already had some l-l 24 random failures of some other equipment, but now that is
/"N O 25 putting a harsh environment on pieces of equipment that I am L
l I HERITAGE REPORTING CORPORATION -- (202)628-4888
103 1 still trying to keep the containment cooled with. Okay? (} 2 MR. MICHELSON: One more class of events that is 3 possible, - and that is where you have a pipe bl .mk outside of 4 containment. 5- MR. KOLACZKOWSKI: I am going to coine to that in 6 just a moment. 7 MR. MICHELSON: Containment is fine. 8 DR. SIESS: Where would TMI be? Between the second 9 and third? 10 MR. KOLACZKOWSKI: I guess in a way it would be 11 between these two I guess. Right. I mean because if we did-- 12 DR. SIESS: You. 13 MR. KOLACZKOWSKI: We weren't cooling the core for a 14 while, okay, but we were still doing okay on the, as far as 15 the containment was concerned. 16 DR. SIESS: Eventually. 17 MR. KOLACZKOWSKI: We get back to this stage.
- 18 DR. SIESS
- Didn't get the vessel cooled.
19 MR. KOLACZKOWSK1: That's right. We could lose the 20 containment entirely. Lastly, we are not removing any heat or l 21 whatever, its integrity is going to be lost, or is already 22 lost, and we ere also not cooling the core. There is still 23 certain pieces of equipment that you would like to see l 1 24 operable just to mitigate the consequences of the event such 25 as containment spray, spraying down the fission products, and l HERITAGE REPORTING CORPORATION -- (202)628-4888
104 1 so at least the consequences of the accident aren't as severe (~N 2 as they would be otherwise, so there are certain pieces of G 3 equipment that you would like to see operate under these 4 conditions, that are going to see some sort of harsn 5 environment because of the radiation being released, and maybe 6 perhaps pressures and temperatures rising in the containment 7 to the point of final fails. 8 DR. SIESS: Lose the sprays, you probably wouldn't 9 fail the containment. 10 MR. KOLACZKOWSKI: You may not. Depends on-- 11 DR. SIESS: Recovery doesn't come into the picture, 12 does it? 13 HR. KOLACZKOWSKI: All recovery does is it moves 14 from one category back into another perhaps. 15 DR. SIESS: Yes. If you lost your sprays and the 16 containment was heating up, you still woula be in the third 17 category. If you got your sprays back on before it failed,- 18 would you stay in that category? . 19 MR. KOLACZKOWSKI: That is correct, but it says, it 20 says that you are relying on those sprays to be working under 21 condition where containment could be heating up. You are 22 trying to stay in this category so it doesn't become this 23 category. 24 Now there is a fifth category, to answer first of r C 25 all one of your comments, Mr. Michelson, there is a fifth i HERITAGE REPORTING CORPORATION -- (202)628-4888
105 1 categcry, and that is--I guess you could say it would fit 2 ultimately into one of these, but we considered at one time a (O~T 3 fifth category, and that was the very point that you are 4 making, treating accidents that are starting outside. 5 See, all of these accidents starts with basically 6 the harsh environment occurring somewhere inside the 7 containment and working its way out. Okay. You could start, 8 though, at harsh environment outside such ac a steam line 9 break outside the containment or whatever. One time we had a 10 fifth item here that we were going to cover. When we started 11 down that road, what happened was we found out that it was 12 very dependent on plant-specific design. 13 In other words, where the pipes really run through 14 what pump, rooms or through--I have seen some control rooms 15 where or pipes going through the control rooms and stuff, you 16 know, which rooms are the pipes going through? If that pipe 17 was to break, where is the water going to go? Where is the 18 steam going to go? And we found out very quickly that it was
'19 so plant-specific that we didn't see a way of handling it in a 20 scoping study with the resources and the schedule that we had.
21 And I think we chould be honest with ourselves and recognize, 22 perhaps explicitly say in the report that there is a class of 23 accidents that starts with a harsh environment is outside, the 24 very ones that Mr. Michelson has been talking about, that we O 25 just felt we could not handle in the scoping study. Perhaps HERITAGE REPORTING CORPORATION -- (202)628-4888
106 1 that is going to have to be handled on a plant-specific basis 2 and whether that is handled through the IPE process or ggg 3 whatever, I don't know, but yes , there is another class of \ 4 accidents we had that were listed on here at one time, and as 5 we started down that road, we said we can't handle this in any
% 6 gentric way. There is just no way to include this because the 7 pipes could be almost anywhere. Stairwells can be anywhere 8 for the water is going to go who knows where, et cetera, and 9 that's a very plant-specif..c kind of problem.
10 MR. MIChELSON: Left out of the study which that 11 was, left ort of yrs.r documentation then that, you know, 12 unless you warn the reader somehow, he is dull and happy. 13 MR. KOLACZKOWSKI: I agree. 14 DR. FIESS: De PRAs handle ; hose? 1 15 MR. KOLACZKOWSKI: They are starting to. In terms 16 of internal flooding analyses and things of.that nature, I 17 think the FRA procer-s is working on the problem, and you know, 18 as the evolutionary process of PRA centinues, hopefully it 19 will get better. 20 I guers what I can say, it is starting to do it, but i 1 21 what PFA typically coes is it will assure--let's say yot 'a'e 22 aipe break in a certain room. PRA will usually, at this int, PRA anae st will assume the equipment in that very room
- is failed, all equipment, and he starts looking at where does ~ g >j
g/$ir 2L the water and/or the steam go? Okay. I, b g I
107 1 And if the water is going down stairwells to lower 2 elevations in the plant, he will follow where that water might { 3 Eg o , where it will ultimately end up. Wherever that water ends 4 up, he now starts worrying about submergence of equipment, 5 things like that. The water is going to collect, considering-6 drains, all that stuff. If the water looks like it is going 7 to submerge certain equipment, then he ascumes that equipment 3 gets failed also. 9 Steam, you know it is a little harder to say where 10 does all the steam go? Aad I think, to be quite honest, PRA 11 analysts don't treat where the steam goes very well, so they a i 12 will make a conservative assessment in terms of within the 13 room, also make a conservativa assessment as to where the 14 watar or condensation ultimately ends up. They probably don't 15 treat steam very wel?., and then they assume overything else 16 works that is nominal failure rate. That is what the PRA 17 analyst does today. 18 MR. MICHELSON: I think you are exactly right except 19 when you said the he assumes something failed. You really 20 meant you lose the function? 21 MR. KOLACZKOWSKI: Yes; spuriour,. 22 MR. MICHELSON: Didn't look for whether failure mode 23 might be unacceptable? 4 24 MR. KOLAC2nOWSKI: Sput.ous signals are typically O 25 not treated. i HERITAGF. REPORTING CORPORATION -- (202)628-4888
108 1 MR. MICHELSON: Don't analyze whether failure is, 7"T 2 okay, the kind of failure you get from the steam or from the U l 3 submergence or whatever? You just ignore it? 4 MR. KOLACZKOWSKI: Let me take this-- 5 MR. MICHELSON: I know.from LERs that you can't 6 ignore it. We see in little ways how when water does get in 7 the equipment or steam or whatever that you do get unwanted 8 actions and this doesn't fail greatly and go away. It just 9 doesn't cease to function. It functions in an adverse way. 10 MR. KOLACZKOWSKI: Becausa we are defining the types 11 of accidents we look at, I think I should address your other 12 two points, and that is I think we need to make something, put 13 something in the report about the limitations. 14 Second of all, another class of accident that we did 15 not deal with is the loss of HVAC such as what occurred in 16 Station Blackout. When all of these accidents began we looked 17 at harsh environment starting if you will in the primary 18 system and ultimately working its way out, and we did not 19 consider the fact that for instance, in Station Blackout 20 accident, maybe all the equipment in the control room is e 21 heating up because I have no HVAC in there. 22 Again, because of certain plants have certain rooms 23 on safety-related HVAC and certain ones do not, et cetera, it 24 just looked to us 3ike to really examine what equipment would O 25 be affected under loss of total HVAC and where that equipment HERITAGE REPORTING CORPORATION -- (202)628-4888
E109 1 might be, just looked like a bigger problem than we could ( 2 tackle in the scoping study, and we felt that maybe that would 3 be better handled in a more plant-specific kind of way, and I 4 think we need to spell that out in the report. , 5 Last, the third kind of thing that we did not 6 consider in this study which perhaps we should list as a 7 limitation is your spurious actions. We did not consider 8 spurious actions, and the reason for that 2.s that if you want 9 to start postulating spurious actions, you basically make the 10 problem endless because I can define an infinite variety of 11 nossible spurious actions. 12 MR. MICHELSON: I don't think there is any doubt of 13 that, but you have to be very careful to recognize that they 14 may be quite risk significant, but you don't know it, and you 15 better warn the reader that take only a limited comfort in 16 your bottom line because it is just that. It is a partial 17 bottom line. You don't, you have left out a lot of things 18 that might, might be cuite--you just don't know because this 19 was an endless job. 20 DR. SIESS: If the problem is endless, I presume , 21 unsolvable, is you said, and if it might be-- 22- MR. i!OLACZKOWSKI: I am nc? ready to say that yet. 23 DR. STESO: And if-- 24 MR. KOLACZKOWSKI: Within our resource, we couldn't O 25 do it. HERITAGE REPORTING CORPORATION -- (2021628-4888
m 110 1 DR. SIESS: Hight be risk significant, as Carl said, [} 2 then where does that leave not-the reader, but the person that 3 is worried about risk? 4 HR. KOLACZKOWSKI: Except for the spurious action 5 one, I think as far as the HVAC or the pipe breaks, I think 6 what it says is that, I think what it says is you rely on a 7 lot of redundancy when you try to, to for instance, if you 8 have a safe shutdown room that is located away from the rest 9 of the plant or whatever, and it is on its own HVAC or diesel 10 power source or whatever. You try to combat some of common 11 cause environments by making a common cause environment. 12 DR. SIESS: It is if you can think of it, you can 13 fix it? 14 MR. KOLACZKOWSKI: Spurious signals, I think that is-15 a real problem. No matter how much radundant equipment you 16 have, if something is operating in a manner that you had not 17 predicted or in fact this is actually extcerbating the 18 accident, you can't get control the thing. That's a real 19 difficult problem. 20 DR. SIESS: That's true, and if it is a significant 21 contributor to risk, at what point do our general design 22 criteria break down? 23 MR. KOLACZKOWSKI: I don't have an answer for you. 24 O*. SIESS: What--do the regulations permit this? O 25 MR. KOLACZK0WSKI: Let me say this one sa,ing grace. HEPITAGE REPORTING CORPORATION -- (202)628-4888 L _. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
111 ' 1 'We might have relative to spurious signals, usually--not r~ 2 always, but usually, some sort of active component has to be , b) 3 performing some action, whether it is a pump or a valve has to 4 move or whatever, in a way that we did not want it to. 5 Usually, usually--not always, but usually that takes power to , 6 do that. 7 If you have L narsh environment, it is very, I don't 8 think it is probabalistically very high that you will get a 9 spurious signal and yet not lose the power source. If it is 10 harsh enough to create some sort of shorting, partial short 11 inshortingg in cable or whatever, it is probably harsh enough 12 you will lose the power source to begin with so you lose the 13 signal to begin with. O 14 DR. SIESS: You do a FMEA. Is that binary? 15 MR. KOLACZKOWSKI: In terms of any one failure mode, 16 yes. 17 DR. SIESS: It usually is. 18 MR. KOLACZKOWSKI: You see what happens is relay 19 stays shut if the relay stays up. 20 DR. SIESS: But the valves are binary. Electric 21 power is not binary. It would be low or high or-- 22 MR. MICHELSON: Air pressure. 23 DR. SIESS: Air pressure is not binary, 24 MR. KOLACZKOWSKI: Well, but you look at it in a 25 sense that you ssy whrt if the power is more than what I want HERITAGE REPORTING CORPORATION -- (202)628-4888
112 1 it to be or the voltage is higher than what I want it to be 2 versus what would happen if the voltage is less.
-{ }
3 DR. SIESS: Look at single things FMEAs, is there 4 any method suitable for doing multiple FMEAs in various 5 combinations of time sequences? Because this is the sort of
~
6 thing-that-- 7 MR. KOLACZKOWSKI: That is a difficult problem. 8 DR. SIESS: Is it impossible, or is it just never 9 been done? 10 MR. KOLACZKOWSKI: I don't way to say it is 11 impossible. It probaoly hasn't been done adequately. 12 DR. SIESS: How do we, how did we think about this 13 type of problem? We can't say it is experience. f. 14 HR. KOLACZKOWSKI: We don't have the experience of 15 these EQ harsh environments and what their effects would be, 16 that's right. 17 ER. SIESS: Experience we have under normal l l 18 conditions. We have had problems with these things. Go on j 19 and over with somewhere in between as Carl pointed out. l 20 MR. KOLACZKOWSKI: I don't have an answer. If I l l i 21 did, I am sure so would a lot of other people and we would be 22 doing something about it. I don't have an answer. 23 MR. MICHELSON: The harsh environment will not be 24 harsh, but in the case of elevated temperature like McGuire 25 when the solid state cabinet heated up, the power supply hung i HERITAGE REPORTING CORPORATION -- (202)628-4888 l
113 1 in there, but the controls are what went screwy first, so you 2 don't always lose, as you elevate the conditions, you don't ( 3 always lose what you would like to lose first. Maybe we need 4 a thermal monitor that just says as soon as it gets too warm. > 5 MR. KOLACZKOWSKI: Shut the power off. 6 MR. MICHELSON: There are ways of thinking about it. 7 DR. SIESS: That is where most of the heat is coming 8 from. 9 MR. MICHELSON: Most of the heat is coming from the 10 power supply, but the room when it got warm, then starts 11 elevating the cabinet temperatures. 12 MR. KOLACZKOWSKI: The point I want to make, I think 13 we went a long way to defining a lot of the possibility 7s d 14 accidents and so on, but yes, you a:e right, we did not 15 consider some common modes outside the containment where that 16 is the initiating event, and we did not consider the 17 possibility of spurious actions just because we could come up 18 with an infinite variety of those and said well, there is no 19 way of treating that within this scoping study. 20 DR. SIESS: Did you look at individual spurious 21 actions? 22 MR. KOLACZKOWSKI: No. You could. That would be 23 something that you could kind of-- 24 DR. SIESS: That would be easier. 1 25 HR. KOLACZKOWSKI: You could take one if you will HERITAGE REPORTING CORPORATION -- (202)628-4386
114 1 system or train of a system at a_ time and say what about if- {} 2 this thing was operating in such a way or failed in such a way 3 that it is in fact exacerbating the accident? 4 CHAIRMAN WYLIE: I think he said earlier-- 5 DR. SIESS: You could take three. It is when it 6 gets to be 40-- 7 CHAIRMAN WYLIE: I think what he said earlier was 8 that many of these things are plant-specific, and on a generic 9 basis, you can't really do-- 10 DR. SIESS: So are PRAs and they don't do it or FMEA 11 . don't do it. 12 CHAIRMAN WYLIE: It is not practical from that
. 13 standpoint to try.
14 HR. KOLACZKOWSKI: Not for this scoping study; we 15 need to acknowledge, therefore, as Mr. Michelson has been 16 saying, that-- 17 MR. MICHELSON: The key question is are they risk 18 significant anyway? Maybe we ought to at least take one plant 19 somewhere and dig in deep enough, one plant at least to find 20 o r. t . 21 CHAIRMAN WYLIE: I think in some plants they are and 22 some plants they are not. 23 MR. MICHELSON: I think that is probably the right 24 answer, but we haven't even tried on any one plant as far as I O
~
25 know, really dig into these kinds of effects as far as I know , HERITAGE REPORTING CORPORATION -- (202)628-4888
115 1 partly because it takes a large research effort to determine (~'} 2 -the effect of elevated temperature and-- V 3 DR. SIESS: When you had curves of postulated 4 ceterioration with elevated temperature it is a terrific job 5 to find out if these--it is like sneak circuits that got 6 formalized, but these are sneak of a different category. 7 MR. MICHELSON: In the case of at least one plant I 8 am aware of that they hung all the heating, all the cooling 9 for the entire plant on two chillers. You lose those 10 chillers--they are in that room, the came room about what, 50, 11 60 feet apart, something like that. They are in the same 12 room, 50, 60 feet apart. If you lost both those chillers for 13 whatever reason, you really do affect a lot of 14 environments--the control room, the instrument room, the power l 15 switchgear are all on the same chiller, two chillers for the l l 16 whole plant. 17 CHAIRMAN WYLIE: For that particular plant you could , 18 make a, an analysis of that one, l 19 DR. SIESS: You still have a problem what got 20 affected in all those rooms? 21 MR. MICHELSON: You have to do a lot of calculating, 22 and then some testing. 23 CHAIP. MAN WYLIE: Another plant that doesn't do that 24 is entirely different. O 25 MR. MICHELSON: It is, it has a different set of HERITAGE REPORTING CORPORATION -- (202)628-4888
116 1 problems. 2 DR. SIESS: A fails one way and B fails another.. It 3 might be different than if you switched them. 4 MR. MICHELSON: I think you are making progress that 5 we even admit such problems exist. 6 MR. KOLACZKOWSKI: There you go. 7 MR. MICHELSON: This report for the first time 8 begins to make some admissions about where we are short. 9 DR. SIESS: I can't help but wonder how many of 10 those problems exist, something being required in the design. 11 MR. BOHN: One is that the effect of spurious 12 actuation, if it is similar to the sorts of things that we 13 observed, relay chatter is dependent on the state of the 14 circuit. When you get into it, I mean the effect may be 15 different depending on whether you are trying to switch into 16 an emergency mode or you are bringing the plant up or bringing 17 it down. We have seen this relay chatter, so it is not even 18 the fact that you just have one circuit you have to look at. 19 You have to look at sort of state of the circuit. I guess the 20 other place where it might really be significant is if it 21 affected your interlocks. That to me, it might be more-- 22 DR. SIESS: That was the concern. 23 MR. BOHN: Just spurious actuation of valve here or 24 there, you affect interlocks, you are going to damage A
\- 25 something; you generally try to use it.
l i l HERITAGE REPORTING CORPORATION -- (202)628-4888 l
117 MR. MICHEL3ON:
~
1 There are also relay races. 2 Interject different part of the race and you get different 3 results, and the thing is very' confused. 4 In the case of solid state circuitry, at least the 5 experts tell me that sulid state devices will not necessarily 6 go to the same direction each time you elevate the + 7 temperature. It will not behave the same way. It isn't 8 consistent in its behavior, and that really complicates the 9 issue then. At least a relay has to close contacts, may be 10 more consistent, but solid state device 1.s not consistent in 11 behavior. At least that is what one expert told me. 12 DR. SIESS: You are, your point about interlocks, 13 those are happening today in normal operation. I read the O 14 other day somebody had done something, fixa.d an interlock or 15 something when he couldn't do something else. Took quite a 16 while to find out what it was. I forget what it was. I think 17 Haddam Neck, but I can't remember what it was. Normal 18 operation; those are all there. 19 MR. KOLACZKOWSKI: I could go through each one of 20 these other, these items in detail, or if you would like, I 21 can go right to the end. I mean basically I have shown you 22 the process overall with-- 23 MR. MICHELSON: Go through the environmental 24 boundaries at least. I think that would be the next most
' 25 important consideration.
i HERITAGE REPORTING CORPORATION -- (202)628-4888
118 1 MR. KOLACZKOWSKI: Okay. Well, again, in the gross (} 2 way in order to include this in a generic scoping study, we 3 primarily broke the plant up into these four primary or these 4 four areas. First of all, what is happening inside the 5 primary boundary itself? .I am talking about within the 6 reactor coolant system because there is some equipment in 7 there or attached to it that in fact we might want to be 8 interested in such as the IDCOR thermocouples or whatever. We 9 also said the next major environmental area was inside the 10 primary containment, but treated it as if you will a single 11 environment. Okay. 12 DR. SIESS: Are you going to tell me what you mean
- 13 by primary containment?
14 MR. KOLACZKOWSKI: On the boiler, I am talking about 15 the dry well, wet well, and on the PWR, I am talking about the 16 actual containment structure. 17 DR. SIESS: Okay, i 18 MR. KOLACZKOWSKI: Then we said further out from 19 that is inside the reactor building oa if there is some 20 annulus building around the containment if you will, okay, 21 where if the environment got harsh within the containment and 22 the containment were to either be vented or rupture, that 23 might be the next area where the environments could then r 24 proceed to. This classified that in a loose, in loose terms t 25 as the reactor building. HERITAGE REPORTING CORPORATION -- (202)628-4888 ,
119 1 DR. SIESS: Some PWRs have it and some don't? 2 MR. KOLACZKOWSKI: Yes. And then inside the other 3 auxiliary buildings was basically everything else. That would 4 be the control room, the turbine bay, diesel generator rooms 5 for, in separate buildings, what have you, so everything else 6 was just lumped together in this other fourth category. 7 Okay. 8 DR. SIESS: Must be some dividing line between that 9 and the environment. 10 MR. KOLACZKOWSKI: Yes. Let me just jump ahead a 11 few slides. Once we defined what equipment might need to be-- 12 DR. SIESS: The doors have to be closed. 13 MR. KOLACZKOWSKI: Yes. Once we define what O 14 equipment might need to operate to respond to those four major 15 accident categories I talked about, we also determined where 16 that equipment was located and where the en/aronments might 17 get harsh, and for an example, let me just .aow you this as an 18 example. This is about five or six slides into the ! 19 presentation. 20 If you were to have a loss of coolant accident type 21 of scenario, let's say inside the containment. Now I am 22 talking about where this is just an example of one inside the 23 containment. As long as I am, as I stay within the category 24 of the containment integrity is okay, and the primary vessel 25 is being cooled, then my harsh environment is occurring only HERITAGE REPORTING CORPORATION -- (202)628-4888
120 1 in environmental srea No. 1, that is, within the primary {} system itself, and also in environmental area No. that is, ' 2 2, 3 within the containment boundary, and so we define for purposes 4 of the scoping sthiy, that indeed we are getting an increased 5 or harsh environment la the form of temperatures or pressures 6 or mechanical shock or what have fou in these two areas,.but 7 as long as I stay within that class of accident, that the 8 containment integrity is okay and the vessel is being cooled, 9 I don't expect any harsh environment occurring out in the 10 reactor building, or in any of the other auxiliary buildings. 11 Similarly, if a containment ultimately gets 12 breached, in this form of an accident, but I'm still cooling 13 the reactor, I'm now going to extend that harsh environment ; 14 beyond the containment boundary depending on where it 15 ruptures, and I am now going to get it into, for instance, the 16 surrounding reactor building or annulus. Okay. And so we did 17 this sort of mapping for all our various accident types and 16 for different kinds of whether we had a LOCA or transient 19 condition to determine where the environment could possibly be 20 getting harsh, and again I mean, harsh means anything beyond 21 the real, the normal environment that exists in that 22 environmental boundary. 23 When you know where the equipment is located--and we i 24 did that. Here is an example of that where for instance, we O
\- 25 took the emergency service water system on this--could be a l
l HERITAGE REPORTING CORPORATION -- (202)628-4888
, 121 1 PWR or BWR, and we basically by looking at a variety of FSARs
(~y 2 and plant schematics that we had available to us, we broke
%)
3 that equipment down or that system down into major pieces of 4 equipment and tried to say where is that equipment normally 1 5 located in a plant relative to our definitions of those 6 environmental boundaries? And once you know where the, which 7 environmental boundaries are going to see a harsh environment 8 and what equipment inside those environmental bouncaries could 9 be affected, you can then start saying something about is 10 there a chance that this equipment could fail when it needs to 11 operate? 12 (Slide) 13 MR. KOLACZKOWSKI: I am just going to jump to the O 14 end for the sake of time. In going through this mapping 15 process, and then taking what we learned from that and 16 combining it with what came out of the literature review 17 process, and also then asking experts in both PRA and in EQ to 18 look at this list and say do you think it is reasonably , 19 complete, have we missed anything, we came up with this list 20 of I think it is 20 I believe equipment or operations that I 21 appeared to be action that could be risk significant, and we f 22 could define accident scenarios in which the equipment would 23 need to operate at a time wher. the equipment would be secing a 24 harsh environment. 25 And as you can see, they range from like on the HERITAGE REPORTING CORPORATION -- (202)628-4888 >
122 1- boiler, there are instrumentation types of issues here. There q 2 is also, we are dealing with pieces of equipment such as
)
3 safety relief valves, motor valves, containment isolation 4 valves, pumps, various MOVs, high radiation monitors, 5 whatever. On the PWR, you see in a kind of similar nature, 6 again things such as valves and pumps, valves and pumps. We 7 also have containment fans that typically does not exist in a 8 boiler. There are one or two boilers that do in fact have 9 containment fans, but typically they do not. Containment 10 fans, there is a lot of more instrumentation typically inside 11 a PWR containment that could potentially see a harsh 12 environment than what you normally see in a boiler. Usually 13 the boiler has the instrumentation outside and brings the 14 reference legs outside. The PWRs tend to have a little more 15 equipment inside such as pressurizer level and 16 instrumentation, and these t re things that you will, though 17 may not directly affect the operability of the equipment, they 18 affect the feedback that is coming back to the operator in 19 terms of his understanding of what is going on in the 20 accident, and therefore, there is still a potential risk 21 impact if that equipment, if that instrumentation fails. 22 HR. MICHELSON: Lot of interlocks on that. 23 MR. KOLACZKOWSKI: That's right. Hight be 24 interlocks as well. 25 MR. MICHELSON: Lots of interlocks surface. HERITAGE REPORTING CORPORATION -- (202)628-4888
123 1 MR. KOLACZKOWSKI: So this formed the basis if you
/~ 2 ~w ill of the equipment that we felt could be risk significant, b}
3 could see a harsh environment when it needs to operate, and 4 then as Larry will talk about this afternoon, wa then on the 5 basis of this list, picked out some example ones to look at in 6 a little more detail so that we could try to draw some general 7 conclusions of the study ultimately. 8 DR. SIESS: Is this list longer or shorter than the 9 ones that I would get from 50.49? 10 MR. KOLACZKOWSKI: I'm not sure how to answer that. 11 MR. FARMER: It is basically shorter, Al; 50.49 is 12 supposed to cover all safety-related. 13 MR. KOLACZKOWSKI: That's right. That's right. 14 Yes. This list would be shorter because-- 15 MR. DEY: Shorter in the number of equipment, but it 16 is larger in the kinds of harsh environments of equipment. 17 MR. KOLACZKOWSKI: I think we are comparing apples 18 and oranges here. 19 MR. BOHN: Talking about the EQ master list. 20 MR. KOLACZKOWSKI: When we say low pressure 21 injection pump motors on this list, we mean the motor, the 22 cabling, the connector, whatever, but we say it in a FSAR 23 utility. In a PRA space we say it more in a system space or 24 overall component way or as the master list would actually 25 list the corresponding cable, connector, whetc.rer. We still HERITAGE REPORTING CORPORATION -- (202)628-4888
124 t-1 mean that equipment is encompassed in this item that we call
] 2 low pressure injection pump motor so we are comparing apples
)
3 and oranges a little bit, but yes, typically there are items 4 on those lists that would not be on this list and what we are 5 saying is from a PRA perspective, if you assume that that 6 equipment were to fail with probability of one, forgetting the 7 spurious actuation for a moment, it is not risk significant. 8 We are basically saying from a PRA perspective, it is not 9 clear as to why it should be on the list. 10 DR. SIESS: I guess I am still not clear. If I made 11 up a list following criteria on 50.49(b), would it include 12 items that you don't have on that list? 13 MR. KOLACZKOWSKI: Probably yes, 14 DR. SIESS: Now I'm considering--you say HVI pumps. 15 That's not the pump that you are worried about, but that's the 16 things that it takes to make that pump operate? 17 MR. KOLACZKOWSKI: That is correct--the cable to go 18 to the motor so that runs it, the connectors, the penetration 19 that the wiring comes through. 20 DR. SIESS: On that basis you still think 50.49(b) 21 that you have things that aren't on-- 22 MR. BUSTARD: 50.49 has a longer timeframe 23 perspective than what Allen does as a PRA analyst. For ' 24 example, 50.49 would have equipment like residual heat removal 25 system that is not always well modeled in PRAs. HERITAGE REPORTING CORPORATION -- (202)628-4888
125 1 MR. MICHELSON: I would expect to see the cervice 2 water pumps on this list simply because they are required for 3 the, some of the other pumps.to work, providing cooling water 4 for oil on the pumps. 5 MR. KOLACZKOWSKI: From an EQ perspective, and 6 recognizing that we weren't worrien about breaks let's say in 7 the service water rocre area, we said it is hard for us to see 8 where the service water pimps would ever see. harsh environment 9 and so from the purposes of is it risk significant and will it 10 see a harsh environment when it needs to operate, service 11 water pumps don't make the list because'we couldn't define 12 harsh environment other than if you did this initiating event 13 in the service water pump to initialize the accident, as 7 ' ()~ 14 pointed out, we really didn't cover those because those are 15 very plant-specific. 16 MR. MICHELSON: You didn't look at the intake 17 structure accidents? Is that what you are saying? 18 MR. KOLACZKOWSKI: That is correct. 19 DR. SIESS: We create a harsh environment, but it - 2v wouldn't fail because of harsh environment. 21 , MR. MICHELSON: Oh, yes, they will fail in harsh 22 environments. It has already happened on a couple of plants. 23 Didn't get both trains, fortunately. 24 HR. KOLACZKOWSKI: As I pointed out, we didn't cover 25 start breaking pipes in any room anywhere out in the plant nERITAGE REPORTING CORPORATION -- (202)628-4888
126 1 because we said that's just too big a project for us to handle {} 2 in a generic scoping study, because that could be very 3 plant-specific. Okay. 4 Did I answer your question? Yes, you would see 5 items on the list most likely that would not be on this list. 6 MR. MICHELSON: From your exposure of looking at a 7 number of PRAs to do this work and so forth, dit fou tind that 8 PRAs handled the question of pipe breaks outside of 9 containment? It is purported to--we ask the PRA person of 10 which you are one, you say yes, I handled pipe breaks outside 11 of containment. 12 MR. KOLACZKOWSKI: I will try to be as honest with 13 you as I can, 14 I would say if you are doing, if you are not doing 15 an internal flooding type of an analysis or something like 16 that, or fire analysis, that no, Idould, they are not looked 17 at. 18 If you are-- 19 MR. MICHELSON: Do you think that people are doing 20 flooding analysis as a part of their PRA? 21 MR. KOLACZKOWSKI: They are starting to get more and 22 more. 23 MR. MICHELSON: Do you think when they do it they 24 appropriately follow the flood and take out the equipment for 25 the environment that they are exposed to and do, at least do HERITAGE REPORTING CORPORATION -- (202)628-4888
127-1 the kind of a level of study you have done here?
-v/"} 2 MR. KOLACZKOWSKI: As I pointed out, such as in 3 flooding, they will typically assume that the equipment in the 4 room where the initial break occurs is failed. They will 5 follow the water paths and then wherever-it collects, if it 6 can be, get high enough to submerge equipment, they will 7 assume that equipment fails. We really don't handle the steam 8 very well.
9 MR. MICHELSON: To submerge is fairly easy to do. 10 How about the case where the water is? 11 HR. KOLACZKOWSKI: Usually they will assume-- 12 MR. MICHELSON: That is where I find they don't know 13 how to do it. 14 MR. KOLACZKOWSKI: Usually they will assume that the 15 spray--they will try to comeu with some probabalistic argument 16 that well, the equipment would survive 90 percent of the time 17 if it is sprayed on and 10 percent of the time it won't, and 18 it is really based on very limited test data and mostly 19 engineering judgment. 20 HR. MICHELSON: When it gets to electrical cabinets, 21 they don't seem to know what to do at all. 22 MR. KOLACZKOWSKI: Very true. 23 MR. MICHELSON: They ganerally just say well, I lost , 24 the function of the cabinet. 25 MR. KOLACZKOWSKI: Very true; you have got to HERITAGE REPORTING CORPORATION -- (202)628-4888
128 - 1 remember PRA is where maybe we were doing LOCA analyses back [} 2 in, you know, 1960 and '70s, it is still evolving, still 3 growing, and you know, PRAs is going to have to evolve, learn 4 how to-- 5 MR. MICHELSON: The problem is people come up with 6 these answers. The core melt probability of this plant is ten 7 to the minus 6, non-problem, but they don't tell you all the 8 things they didn't do in the process of ending up with low 9 probability, and they didn't tell you what importance risk 10 significance. All those left out things might have-- 11 DR. SIESS: They are not supposed to tell you the 12 bottom line and you are not supposed to pay any attention to 13 the bottom line! O 14 MR. MICHELSON: You tend to do it. 15 DR. SIESS: Do you think we have come that far on , 16 core melt of LOCA analysis since the 1960s? 17 I was hoping we would make a lot more progress than 18 that. 29 iThere was a brief pause in the proceedings.) 20 DR. SIESS: Must have been something I said! 21 MR. MICHELSON: You set it off, whatever it was. 22 MR. KOLACZKOWSKI: That's all I had as far as the 23 formal part of the presentation.
.24 The main pointa I want to make is that we did do a O 25 literature review. There are a lot of limitations relative to HERITAGE REPORTING CORPORATION -- (202)628-4888
129 1 looking at things through EQ perspective, so we had to go 2 beyond that and we tried, but this process of coming up with 3 general categories of accidents, general environment 4 locations, where the equipment typically is located, et 5 cetera, mapping those one on top of another, come up with a 6 list that says this equipment might need to operate in an 7 accident, it could see a harsh environment, therefore it ought 8 to be on the list. 9 MR. MICHELSON: I notice a-- 10 MR. KOLACZKOWSKI: Spurious accidents waren't 11 limited to HVAC across the board and also we didn't handle the 12 breaks initiating outside. 13 DR. SIESS: Didn't find anything in that process O 14 that you wouldn't have found with 50.497 15 MR. BUSTARD: That is not necessarily true. Well, 16 by his process he would find that you need PORVs for feed and 17 bleed for PWR. 18 DR. SIESS: They are not safety related. 19 MR. KOLACZKOWSKI: They are not on 50.49, 20 DR. SIESS: Only non-safety related; of course, 21 50.49 is only electrical. 22 MR. BUSTARD: There are electrical components that 23 support-- 24 DR. SIESS: Electrical equipment failure could r O- 25 prevent satisfactory accomplishment of safety functions-- HERITAGE REPORTING CORPORATION -- (202)628-4888
130 1 wouldn't the PORV?' 2 MR. BUSTARD: -If you note it is at the end of
)
3 50L19--there is basically a grandfather clause so the newer 4 plants-- 5 L SIESS: I am talking about the criteria of 6 50.49, not the a,,11 ation of it. 7 MR. BUSTARD: Oh, okay. The newer plants do include 8 f eed and bleed components or. their saf ety list. 9 CHAIRMAN WYLIE: 50.49 speaks to non-safety related 10 as well as safety-related; item 2, 11 DR. SIESS: That is what I just read. He says that 12 was back to some of the older ones, but your list is 13 essentially a suoset of what I get by very careful process of 14 50.49, and you found the things that I do not have or things 15 like the time related and radiation related, general insights, l l 16 some things that aren't needed for a year. I 17 MR. BUSTARD: Wait a minute. We are also talking 18 about different accidents for which this analysis has to be j 19 done compared to the analysis. 20 DR. SIESS: Different environments, though. l ( 21 MR. BOHN: The significance of this list here is 22 that it is equipment which when it occurs in a PRA, is not, t 23 does not have the correct reliability information, that the l 24 reliability for these pieces of equipment should be based on 25 condition reliability, whereas instead what they actually use HERITAGE REPORTING CORPORATION -- (202)628-4888
-131 1 is the historical normal operation reliability. That's the /~N 2 meaning of this list.
D 3 DR. SIESS: Difference is that I couldn't care less 4 what they do in a PRA. I am interested in what happens in a " 5 plant if we have an accident, which may or may not be what the i 6 PRA is talking about. 7 HR. BOHN: You can also extrapolate this to say that 8 particular list according to the accident scenario involved 9 will probably see conditions beyond which it was qualified 10 for. 11 DR. SIESS: I was not addressing the environment 12 which was qualified. I r..m si.mply looking at the equipment 13 list. I was trying to see if thia approach which you have 14 through very logical approach identified any equipment to be 15 important to the health and safety of the public, not a PRA 16 analyst, that wouldn't have been identified by following the 17 definitions of 50.49. 18 HR. DEY: It would be those equipment identified 19 which are important in environments that are not covered by 20 the design basis. 21 DR. SIESS: You name a piece of equipment that was 22 on that list that I would not find in 50 in a master list or t l 23 whatever you call it that somebody developed, i 24 HR. ROLACZKOWSKI: I'm not sure, but let me take a O 25 stab at that maybe, rake the HSIVs here. They are probably l HERITAGE REPORTING CORPORATION -- (202)628-4888 i - - , , , - _ - - - _ _ _ _ , _ _ _ _ , _ . _ . _ . . , . _ , _ _ _ , _ , , _ . . _ _ ,
132 e 1 on the 50.49 list because I need to be able to isolate the 2 containment in certain conditions and so the inboard MSIVs and 3 outboard need to be able to close the risk significance, 4 however, of this item, although that can be risk significant. 5 From this perspective, we are worried about not only that 6 function, but what about if I needed to reopen the MSIV so I 7 could get back to the condenser, remove the heat out of the 8 reactor, okay? Even though this item might be beyond 50.49 9 list, it probably is only concerned with the closure and not 10 with reopening. 11 DR. SIESS: You are answering Parts B and C of the 12 question I haven't asked. My original question was simply the 13 list, the items might be on there for different reasons. O 14 Might be on there for the wrong reasons. They might be 15 associated with wrong criteria, but I was just trying to 16 figure out if the process has expanded the list. 17 MR. BOHN: The only item that we are aware of is 18 PORVs. 19 DR. SIESb. Char.ge the basis for the list. 20 MR. KOLACZKOWSKI: Only item might be the PORVs for 21 some plants. Now the PORVs are also getting on lists for 22 power plants. 23 DR. SIESS: You see what process you have gone 24 through, where I sit I now understand better why they are on 25 the list and what to do about it, then simply going through HERITAGE REPORTING CORPORATION -- (202)628-4888 . . - . . ~ . - . . . - . - - . - _ _ , , - , . - , . . . - _ - . , - - - . - - . - _ . - - . - . . _ .
133 1 the 50.49 process. O
%)
2 MR. KOLACZKOWSKI: 50.49, really a lot of 3 regulations are not based on risk concepts. 4 DR. SIESS: A lot you said? Fine. You answered my 5 question. 6 CHAIRMAN WYLIE: Any other questions? If not, let's 7 take a break for lunch and be back at 1:20. 8 (A brief recess was taken.) 9 I 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 O 25 HERITAGE REPORTING CORPORATION -- (202)628-4d88
134 1 AFTERNQON SESS I Q ll 2:20 p.m.
/') 2 CHAIRMAN WYLIE: Let's resume the meeting here.
U 3 MR. BUSTARD: What I would like to talk about now is 4 some of the additional examination of selected equipment 5 operations that we performed. 6 Having established a set of equipment operations, 7 the next step was to look at some of those in more detail to 8 assess their risk significance within the context of important 9 PRA sequences and environments and to assess possible 10 equipment qualification issues that would be applicable to 11 that equipment operation, again within the context of the-12 environments and the sequences, and the timing need of the 13 equipment when based on PRA analysis. 14 The particular issues that we--wait a minute. The 15 particular issues that we looked at in more detail during the 16 EQ risk scoping study included the PORV components that are 17 useful for feed and bleed, and--in a PWR, and that would 18 include solenoid operated valves, and in some cases, the 19 motor-operated valve block valves, and of course, for each one 20 of these components, we did assess applicable impact of 21 penetrations, splices, cabling, et cetera, on the operability 22 of the components. 23 We also looked at steam generator, detection 24 transmitters, containment fans and high range radiation 25 monitors for PWR components, and for BWR components we looked l t HERITAGE REPORTING CORPORATION -- (202)628-4888
135 1 at the SRV and MSIV solenoid operators, the MSIV bypass valve, 2 MOVs, and high pressure injection pumps, the high range 3 radiation monitors. 4 So with that statement, why don't we start with the 5 steam generator level detectors, and PORVs, and I am going to 6 move right to the next vugraph here. 7 And the significant thing to keep in mind when we 8 talk about feed and bleed, which relies on the PORVs which in 9 turn rely on the solenoid operators and possibly the block 10 valves, and auxiliary feedwater which relies on steam 11 generator level detection, is that what we have is a small 12 break LOCA or a transient-induced email break LOCA type 13 sequence occurring in the PWR, and we need to depressurize the O 14 primary system sufficiently so that we can achieve core 15 cooling through the high pressure injection systems, or the 16 low pressure injection systems, whichever are applicable. And 17 there is two ways that that can be accomplished. 18 One way is we can depressurize the primary system by 19 opening up the PORV, basically performing the feed and bleed 20 function. 21 The second way we can achieve core cooling is to 22 feed water through the auxiliary feed to the steam generators 23 and dump heat out of the primary system through the steam 24 generators, so in the small break LOCA environment, we have O
\~/ 25 these two possibilities from a probabalistic risk assessment HERITAGE REPORTING CORPORATION -- (202)628-4888
136 1 perspective for maintaining cote cooling, and what this table 2 _shows is some analysis out of the draft NUREG 1150 offort for '{ } 3 Surry, and what I have for Surry is a base case core damage 4 frequency of about 2.5E to.the minus 5. That represents the 5 contribution to core damage frequency for Surry for all 6 important accident sequences. 7 The entries in this particular table represent the 8 contribution for four particular sequences, and they happen to 9 be small break sequences, and transient-induced sequences that 10 are similar to small break sequences. And in the draft NUREG 11 1150 analysis, these four sequences give a contribution to the 12 base case core damage frequency in the range of 4E to the 13 minus 8, or approximately a tenth of a percent of the total 14 core damage frequency, so they are not super significant 15 contributors to core damage frequency based on current PRA 16 analysis. 17 Now how was this number 4.5E to the minus 8 18 calculated? It was calculated using probabalistic techniques, 19 but what were the inputs? The input on the auxiliary feed 20 failure rate was roughly 3.5E to the minus 5. The auxiliary 21 feed was considered to be rather reliable for this 22 calculation, and that reflects the fact that Surry has three 23 different modes for performing auxiliary feed as well as can 24 crossconnect to the neighboring Surry unit so auxiliary feed O 25 is considered rather reliable. HERITAGE REPORT'NG CORPORATION -- (202)628-4888
137 ; 1 In addition, the alternative technique to auxiliary [} 2 feed, namely feed and bleed, had a reliability number 3 associated with that. operation of 3.6E to the minus 2, and 4 that's mainly controlled by operator response to an indication 5 to actually punch the button and start doing feed and bleed. i 6 The equipment reliability to perform feed and bleed is 7 considered better than this, but--by the PRA analysts so I 8 have numbers in the 10 to the minus 2 range and 10 to the 9 minus 5 range from a ?RA perspective for responding to this 10 small break LOCA, and that gives me contributions to 10 to the 11 minus 8. That is small compared to my total base case core 12 damage frequency. 13 Now from an EQ perspective, the feed and bleed 14 operation relies on appropriate operation of the pilot i 15 operated relief valve, and that relies on appropriate 16 operation of the solenoid operator, and possibly the 17 motor-operated block valve which for many plants may be closed 18 part of the time and have to be opened if I am going to do 19 feed and bleed. There is very little redundancy in this 20 system. 21 PRAs typically assume that one needs two of two i 22 PORVs to open to successfully accomplish feed and bleed, and 23 each PORV has one of one solenoid operators, and one of one 24 block valves, for example, so that there is very little . O 25 redundancy up here, and we have components, MOVs and SOVs,
- HERITAGE REPORTING CORPORATION -- (202)628-4888
138 1 which historically have shown up in numerous information {} 2 notices that the NRC has put out. As Bill Farmer mentioned 3 this mortgage, they have d6ne testing on solenoid operators as 4 part of the equipment qualification research at Franklin Labs, 5 and the question is can we characterize that SOV operability 6 by a reliability down in the few percentage points for 7 unavailability, or should we move in this direction based on 8 the fact that EQ does not give us normal operation reliability 9 for harsh environments such as the steam environment that 10 would be generated during a small beeak LOCA? 11 On this side, cur alternative cooling mode is 12 auxiliary feed, and as I mentioned, there is numerous 13 injection paths for that, and many of the components 14 associated with those injection paths are located outside the 15 primary containment, and therefore, are not going to be 16 impacted by this small break LOCA environment. 17 However, there is the steam generator level 18 detectors, and associated cabling, electrical penetrations, 19 and connections, that help provide the operator indication on 20 how to achieve appropriate auxiliary feed. In some plants, 21 they, the steam gerierator level detection is a, is basically 22 used by the operator manually. He controls auxiliary feed 23 manually after looking at the indication in his control room, 24 and in other plants it is done automatically. Auxiliary feed 25 rates are controlled automatically based on the signal coming HERITAGE REPORTING CORPORATION -- (202)628-4888 r ___
139 1 off the steam generator level detectors. (} 2 The steam generator level detectors have been 3 tested. For example, Barton transmitters were tested as part 4 of NRC sponsored research program at Sandia, and the main 5 feature about that type of instrument is that it is going to 6 be susceptible to moisture condensation or moisture intrusion 7 effects in the circuitry. If you are going to have terminal 8 block in the circuit, you would have moisture condensation-9 problems. That could drive the steam generator level 10 detection. 11 Likewise, if you had electrical penetrations that 12 allowed for moisture condensation with age, and leakage across 13 the steam generator level signal inputs, you could get false 14 signals coming out of this side. 15 So what does that do? That suggests that from an EQ 16 standpoint, we may not have as high a reliability as 3.5 to 17 the minus 5. We should be moving in this direction. Likewise 18 we may not have as high reliability for the SOVs and the HOVs 19 in which we are moving in that direction. As has been pointed i 20 out many times this morning, we lack accident reliability data 21 to say exactly how far we should go in either direction, but 22 what one sees when one looks at the table is that this 23 requence which was, is and was considered insignificant at a 24 one tenth of 1 percent contribution to the base case core O 25 damage frequency, does get larger and becomes a significant P HERITAGE REPORTING CORPORATION -- (202)628-4888
140 1 contributor to core damage frequency and our auxiliary 2 feedwater failure rate gets to 1 percent, for example, and our 3 PORV failure rate gets to 10 percent. We then have a very 4 significant contribution to base case core damage frequency 5 and in fact it is a large as what is the total contribution 6 to base case core damage frequency currently considered by 7 PRAs. 8 To just summarize some of the things I have said 9 with respect--in vugraph form, the PORV feed and bleed 10 function relies on solenoid operators, and motor operators 11 historically were not included on utility EQ master lists for 12 the older plants, have little redundancy, have been the focus 13 of numerous NRC information notices, inspection findings, and O v 14 research programs, and lack data to quantify reliabilities 15 during accident conditions. ) 16 CHAIRMAN WYLIE: I don't know. On the PORVs, are t 17 they required to report all failures of those? I don't know. 18 Are they? Since it is not on the EQ list? 19 MR. KOLACZKOWSKI: I think on individual plant 20 basis. 21 CHAIRMAN WYLIE: You don't really have the data base 22 because a lot of stuff hasn't been reported? 23 MR. MICHELSON: Depends on what happens from the 24 time they found the field. If they just found the block valve 25 closed-- HERITAGE REPORTING CORPORATION -- (202)628-4888
(. - 141 1 CHAIRHAN WYLIE: I am just talking about the PORV. r~s 2 MR. MICHELSON: The block valves are part of the b 3 problem. The block valves are not opening. 4 CHAIRMAN WYLIE: In discussions with Glenn Reed, he 5 always had a lot of concern with the internally operated PORV, 6 and talked about crud build-up and what have you inside there 7 presenting problems. From his experience they were all bad 8 from that standpoint. They could be worse than what the 9 numbers show. 10 MR. BUSTARD: The steam generator level detectors, 11 as I mentioned, do have substantial redundancy. For example, 12 Westinghouse plants typically would have three transmitters 13 associated with every steam generator. However, these
# circuits would be sensitive to common cause failure resulting 14 15 from moisture condensation or moisture intrusion, and so this 16 is the type of EQ issues that you would really want to be 17 careful about with respect to this.
18 (Slide) 19 MR. BUSTARD: This part of this, of this table; 20 MR. BELL: I am Russell Bell. I am with NUMARC. 21 That issue, the moisture intrusion, is that a new issue? It 22 is well known and it is understood that you have to qualify 23 your interface for these instrumants. I see it called into 24 question repeatedly in the report. I'm not sure it is, I'm 25 not sure it is proper. HERITAGE REPORTING CORPORATION -- (202)628-4088
142 1 MR. BUSTARD: Well, let me respond to that by saying 2 that many of the issues that are being evaluated in this study 3 are not issues that we have just come up with for this study. 4 Simultaneous versus sequential, accident simulation 5 techniques, have been called out in literature references like 6 NUREG/CR-4301 for quite a while, and the question that we were 7 trying to ask in this study was could we take PRA perspectives 8 and provide some perspective on many of these historical 9 issues? It was not necessarily a goal of the study to only 10 focus on new issues that have never been identified before. 11 MR. MICHELSON: The moisture intrusion is the major 12 problem because although there are qualified ways of sealing 13 conduits and so forth, they all have to, sealant has to be O 14 installed in situ, and after installation, so although factory 15 qualifications work pretty well, some of these field installed 16 qualifications don't work so well because there is no way to 17 test that necessar:ly, any way to test it after it has been 18 installed. Sometimes it was installed with little care and 19 some cases wasn't even installed they found out later, gone 20 back and sealed the conduit entrances and so forth. 21 HR. BELL: Programmatically you do have procedures 22 in place to handle that? 23 MR. MICHELSON: Oh, yes. There is procedures all 24 right. 25 MR. BELL: What is the significance of PRA study HERITAGE REPORT 2.M CORPORATION -- (202)628-4888
143 1 citing the importance of feed and bleed where as that has not 2 been an EQ concern in the past? f-)
%)
3 MR. BUSTARD: The steam generator level detection 4 circuits have been, I mean transmitters, Barton transmitters 5 that are used in steam generator level, steam generator level 6 circuits, have been qualified. 7 MR. BELL: Absolutely, and that's as you say, 8 relating to the aux feed system which is redundant and 9 qualified and you know, supposedly up, down and all around. I 10 suppose that's the reason feea and bleed has been 11 de-emphasized for EQ, that they have not most times gone to 12 qualify as you mentioned the PORV and the-- 13 MR. BUSTARD: What you are saying is that because I O 14 have such high redundancy on this option, I don't require as 15 high accuracy or redundancy or whatever over here? 16 MR. BELL: I am suggesting that, yas. I come from 17 an EQ perspective, and that's, that may be a typical comment 18 you might expect. 19 HR. BUSTARD: Sure. I could believe that that's the 20 perspective of some utilities. I think what we are doing in l 21 this study is saying how can we look at that perspective from 22 the standpoint of risk analysis, and PRA analysis, and that 23 has not been done before as far as I know. 24 One thing to point out is that PRAs typically do 25 things like risk achievement analysis, as Allen mentioned HERITAGE REPORTING CORPORATION -- (202)628-4888 _-_ ,__ _ _ _ _ . . _ _ . . . . _ . . . - _ _ _ _ . , - . _ _ - ~ , _ _ - - _ . _ , - , .
144 1 carlier today, to look for.what are important pieces of
~
2 equipment that if they were-to fail would have a large impact 3 on risk. 4 If we were to do that in this situation, the steam 5 generator level detection components are' eliminated from the 6 reduced fault trees early in the PRA process because of 7 truncation and other things that are going on, and therefore 8 when they do get around to doing risk achievement analysis, 9 you don't pick up the importance of the steam genarator level 10 detection having an impact on your auxiliary feed, feed rate, 11 which is considered highly reliable by PRA. Okay. 12 And what we are saying is there is a possible link 13 here with the steam generators level detection circuitry. In O 14 certain conditions, for example, if you had terminal blocks in 15 that circuit, you could-- 16 MR. BELL: Again, there shouldn't be. At this point 17 in time, you shouldn't-- 18 MR. BUSTARD: Couple months ago they found terminal i 19 blocks in PWR plants, and I agree with you. They shouldn't 20 have. 21 MR. MICHELSON: There is no regulation against it, 22 is there? 23 MR. BUSTARD: Terminal blcaks were highlighted by an 24 NRC information notice several years ago. l j
) 25 MR. MICHELSON: That requires no action.
l l HERITAGE REPORTING CORPORATION -- (202)628-4888 l' f
c-145
- 1. Information notice doesn't require anything, just informing
(} 2 .you of terminal' blocks are permitted, but certainly inside a 3 containment they have to be qualified, or show that you can
.4 accept the failure.
5 MR. BUSTARD: That's right. 6 MR. BELL: Okay. Thank ycou. 7 MR. BUSTARD: Okay. So what type of' activities 8 could be performed to reduce the uncertainty associated with 9 actual plant operation with respect to the feed and bleed and 10 the auxiliary *.eed operation? 11 For feed and bleed, one could assess the solenoid i 12 operated valve and motor-operated valve reliability during 13 small break LOCA, steam, temperature environs. 14 One thing I want to stress here is that I don't have 15' radiation in this statement. This operation from a PRA 16 perspective is performed prior to core melt, and therefore 17 prior to the development of a radiation condition within the 18 plant. 19 The Franklin study that has looked at SOVs has 20 recommended that one monitor and assess the long-term aging 21 behavior of the elastomeric components associated with the 22 SOVs. And there is various operations aad maintenance issues, 23 for example, conduit, sealing on SOVs, to make sure that you 24 arc not going to get moisture condensation effects in insida. O 25 MR. MICHELSON: Did you look at overvoltage effects P HERITAGE REPORTING CORPORATION -- (202)628-4888 i
r-146 1 on the'SOV? r~% 2- MR. BUSTARD: No.
\~)
3 MR. MICHELSON: Just simply failing the power supply 4 in an overvoltage condition,.most of power supplies can swing 5 up as well as swing down and not necessarily just quit all 6 together. In other words, they degrade. If they get 7 overvoltage, which has happened in at least one occasion on a 8 SdRAM system in the past, that the solenoids no longer 9 function. You get baked out in the hot coil. Just simply 10 'will not be able to open the relief valves. 11 MR. BUSTARD: Let me make one comment with respect 12 to what you said about the operability of the steam generator
. 13 level decectors, and the SOVs and the MOVs, that that should
( 11 be covered by current operations. 15 If you look at the information notices, MOVs have 16 had something like 17 inforniation notices over a ten-year 17 period. The evidence is not there saying that problem is 18 decreasing with time. Once one problem gets solved, another 19 one seems to assert itself. For Limitorque motor operators, 20 they discovered degraded internal wiring due to space heaters. 21 A couple of years ago they were calling out another problem. 22 For SOVs, the information notices are starting to , 23 show the impact of not performing maintenance activities or 24 repair activities according to manufacturer's suggestions and (- k 25 things like that. HERITAGE REPORTING CORPORATION -- (20?)628-4888
147 , 1 So what we are saying is that given the historical [} 2 types of problems, the research that has been done by the NRC 3 on SOVs, the possibility from moisture condensation effects 4 that we could move somewhere in that direction, and in this 5 direction, but right now we don't know how far. We don't know 6 how big our uncertainty is for the contribution of these four 7 sequences to base case core damage frequency because we are 8 not sure how far down this path we are, nor are we sure how 9 far we are over on that path. 10 And to deal with that, one recommendation is to 11 develop estimates regarding the accident reliability of steam 12 generator level detection circuits. Likewise, this particular
. 13 table has on this axis the auxiliary feed failure rate, not 14 the steam generator level failure rate. We have to develop 11, realistic human reliabili ty models that relate the f ailure 16 rate for level detection to the reliability of the auxiliary 17 feed system if we want to make a definitive quantitative PRA 18 analysis for this particular set of sequences.
19 Some other activities that could se done to reduce 20 the uncertainty regarding that portion, tb4 steam generator 21 level portion of this issue, as I said, would be to ensure 22 that terminal blocks are not employed in the circuits, 23 determining the impact of humidity and electric and 24 penetration assemblies and address various maintenance and O 25 operation issues. HERITAGE REPORTING CORPORATION -- (202)628-4888
148 1 -(Slide) f3- 2 MR. BUSTARD: Several historical EQ issues are not
\/ ,
3 risk significant for this particular set of equipment 4 operations. As I mentioned earlier, from a PRA perspective, S we are interested ir. avoiding core melt with these equipment 6 operations so we don't have a radiation environment inside 7 containment. And therefore, many of the historical EQ issues 8 that have arisen because of the radiation environment, the 9 instantaneous release requirements of the NRC don't apply, and 10 that would include the differences between simultaneous versus 11 sequential accident simu'ation techniques, beta / gamna 12 equivalence. 7, 13 In addition, from a PRA perspective, we are 14 interested in short-term equipment operability over the course l ', of a few hours to a few days. We are not interested in 16 long-tern equipment operability, a hundred days, and therefore 17 we don't have to worry about post-accident acceleration 18 techniques, nor do we have to be as concerned with the issue 6 I 19 of oxygen in the test chamber environments during the accident 20 simulations because our current research data suggests that 21 these, and this--and these are all long-term equipment 22 operability issues rather than short-term equipment 23 operability issues that would arise beginning with accidents. 24 (Slide) O 25 HR. BUSTARD: One thing that is appropriate to
- HERITAGE REPORTING CORPORATION -- (202)628-4888
149 l 1 mention is that additional study is warranted to determine 2' whether the current radiation EQ requirements impact risk (
- 3 adversely.
4 What we have for tce steam generator level i 5 transmitters is that the current accuracy specifications that 6 the utilities place on those transmitters is based on the 7 radiation portion of the equipment qualification test. For 8 example, if you go to Barton's specifications for their 9 transmitter, they specify that during the radiation 10 environment, the accuracy would be about plus or minus 10 11 percent, bat during the steam and pressure environment, it is 12 approximatnly plus or minus 3 percent, so these devices are 13 much more accurate in the steam temperature conditions that 14 they would be expected to work in to achieve this function, 15 but from a plant basis where they have to commit to j 16 instantaneous release of radiation at the start of the 17 accident, they have to tie their accuracy to a less accurate 18 specification, and both Westinghouse and a utility have 3 19 indicated to me that that means an increase, a decreased 20 operating window and therefore, increased SCRAM rates for the 21 plant. That has not been quantified in any way, and the true 22 risk impact has not been determined yet, and that is a 23 worthwhile thing to perform. 24 (Slide) O 25 HR. BUSTARD: Let me move on to another example of HERITAGE REPORTING CORPORATION -- (202)628-4888
150 1- equipment operations that we identified and analyzed in a (} 2 'little bit more detail. And for this one, I am also going to 3 move directly to tho table. 4 What I have here is a BWR TW sequence, and the , 5 analysis is, the draft is performed in conjunction with the 6 draft analysis for NUREG 1150 at Peach Bottom. The TW i 7 sequence is a transient that has the subsequent loss and 4 8 suppression pool cooling, and so this is an example where 9 initially I am maintaining vessel cooling and therefore I am 10 maintaining core integrity, but I've lost my ability to lose 11 the heat that is coming out of the core and going into the 12 suppression pool. I have lost my ability to take that heat 13 out of the primary containment, and so my primary containment 14 is gradually heating up. First the suppression pool gets hot, 15 and then it gradually pressurizes and increases the 16 temperature in the dry well. , 17 Now from a PRA perspective, how do we deal with that i 18 typical sequence? The first thing that a PRA event tree would 19 do would be to rely on the high pressere injection systems to 20 maintain vessel cooling as the suppression pool heats up. 21 However, at a certain time in that accident , t 22 sequence, several hours into the accident sequence, the \ l 23 avspression pool has reached temperatures of about 210 to 240 24 degrees Farenheit, and at that point, the PRA analysts assume . O 25 that the high pressure injection systems fail. This is one of - t HERITAGE REPORTING CORPORATION -- (202)628-48d8 ,
151 1 the few cases in PRA analysis where'they do not rely on normal s 2 operation reliability to describe the front end sequences, _(d 3 They have actually employed a step failure rate where early in 1 4 the accident sequence, I have normal operation reliability, B and once my suppression pool reaches 210 to 240, I am up te 6 failure rate of one for the high pressure injection system. , 7 Given that the high pressure injection system is no 8 longer available to maintain core cooling, there are 9 alternative techniques available to maintain core cooling. 10 One technique available to the operator is to depressurize the 11 primary system using manual actuation of the SRVs, and upon 12 depressurization of the primary system to rely on the low 13 pressure injection system to maintain core cooling. O 14 The current assumption in PRAs for the draft 15 analysis of Peach Bottom with respect to this sequence was 16 that the failure rate for that action would be 4E to the minus 17 3. 18 MR. MICHELSON: But you are still pumping from the 19 - suppression pool which is what, overheated? 20 MR. BUSTARD: That's right. 21 MR. MICHELSON: And the RHR, although it is designed , l 22 to operate at high temperatures under closed loop conditions, 23 isn't necessarily designed to operate at high temperature I 24 pumping from the suppression pooling. Its high temperature 25 rating is also accompanied by a very high pressure which you
- HERITAGE REPORTING CORPORATION -- (202)628-4888 i
152- , l' get on closed cycle operation of RHR from the vessel back to (} 2 the vessel. You don't get that same operation pumping from 3 the suppression pool. 4 MR. KOLACZKOWSKI: Remember in this sequence the S containment is pressurized so even though the temperature is 6 rising in the pool and in the containment environment, 7 pressure is rising also. 8 MR. MICHELSON: There is a whole lot more to that i 9 system, unfortunately. It returns its water back to the 10 suppression pool through some relief, through some, through a 11 globe valve which has to now flash the water back down to , 12 whatever containment pressure is. That system won't work at 13 high temperature even though RHR does, and when you are in RHR 14 closed cycle mode, you are pumping there the vessel, back to 15 the vessel. You are not breaking down the pressure. The 16 whole system is at about the same pressure. Even though it is 17 all 300 degrees, it is all in about the same period. In this 18 case, you are dropping the pressure from 250 pounds, 300 19 pounds discharge to the pump back down to the suppression pool 20 which is maybe at 30 pounds, whatever. 21 MR. KOLACZKOWSKI: In this accident the suppression 22 pool cooling mode is failed, that's why the TW sequence is-- 23 MR. MICHELSON: Now you are saying you go to RHR7 , 24 HR. KOLACZKOWSKI: No, no--only for the vessel O 25 injection mode. ! HERITAGE REPORTING CORPORATION -- (202)628-4888 ,
o 153 v, 1 MR. MICHELSON: Where are you getting the water 2 from? 3 MR. 2.0LACZKOWSKI: From the suppression pool. 4 MR. MICHELSON: You have got to pump the water out 5 of the suppression pool and you are going to return it just to. S the reactor vessel. 7 MR. KOLACZKOWSKI: Which in turn is going to boil it 8 off to the SRVs back to the pool. 9 MR. MICHELSON: Three hundred, 400 pounds of the 10 vessel, okay. As long as you don't go back to the pool, you 11 are going to be all right; sidestep some of it to the pool. 12 MR. KOLACZKOWSKI: By virtue of the sequence the 13 pool cooling is failed due to I can't get the valves open to 14 the suppression pool or whatever it might be. 15 MR. MICHELSON: That's right. That's how the 16 suppression pool-- 17 HR. KOLACZK0WSKI: That's right. 18 MR. MICHELSON: You are saying RHR works. It just 19 doesn't work-- 20 HR. KOLACZKOWSKI: We can eject to the vessel but l 21 can't get the heat out of the containment. 22 MR. MICHELSON: How about the room cooling for RHR t 23 since those rooms are sitting right next to the suppression ; I , 24 chamber on the older model BWRs? What is keeping the rooms O 25 cool so the RHR pump can work? You didn't include heating and f r r HERITAGE REPORTING CORPORATION -- (202)628-4888
154 1 ventilating. 2 MR. KOLACZKOWSKI: It is likely the HVAC system in 3 RHR rooms are still working. 4 MR. MICHELSON: Cooling the air in ,he room which is 5 being heated by the suppression pool; suppression pool is 6 uninsulated. You talk about 250 degrees or something, you are 7 talking about a terrific heat load on the room cooler. 8 MR. KOLACZKOWSKI: Granted. 9 MR. MICHELSON: I doubt the room cooler can keep up 10 with that heat load. If the room is isolated, you are all 11 right, but--on older BWRs it is not. On the newer ones it is. 12 MR. KOLACZKOWSKI: Go back to our list of equipment. 13 You will notice that-- 0 14 HR. MICHELSON: Peach Bottom is not. 15 MR. KOLACZKOWSKI: That the low pressure injection 16 pumps are on the list and it is because of reasons like you 17 are talking about, but this is dealing with the SRV issue so 18 let's assume that the pumps are working for the sake of the 19 argument. 20 MR. MICHELSON: I don't see how you can assume that. 21 You are saying I will use my high pressure. You will use 22 really low pressure injection, but low pressure injection 23 can't be cooled because of the suppression pool is too hot. 24 Then I don't see if it is one and the same--it gets 25 unavailable. HERITAGE REPORTING CORPORATION -- (202)628-4888
~
155 1 MR. KOLACZKOWSKI: Ultimate thing you con do (} 2 on--boilers will typically have means to crosstic one of the 3 service water pumps so that you can actually put service water 4 into the reactor vessel. 5 MR. MICHELSON: Some boilers have that. 6 MR. KOLACZKOWSKI: So we can say that, that's what 7 is working is the pressurizing so that you can put the service 8 water into the reactor vessel and those pumps sitting out by 9 the river somewhere. 10 MR. BUSTARD: This sequence is a slowly developing 11 sequence in the sense that it occurs over many hours. It is 12 not something where we have to respond to it in a very short 13 time period. The high pressure injection system is assumed to l 14 fail several hours into the accident sequence, and it is at 15 that point that one wants to have some methods available to 16 take it to place, and one method involves depressurization of 17 the primary system using solenoid operated valves that are 18 located inside the dry well to operate the SRVs, so the , 19 solenoid operators in this case are seeing these environments 20 that are increasingly getting more harsh as the suppression 21 pool heats up. 22 Now as I-- 23 MR. MICHELSON: My only point was outside of . 24 containment the environments get more harsh, too, as the 25 suppression pool heats up because it is uninsulated. Over a i HERITAGE REPORTING CORPORATION -- (202)628-4888
156 1 period of time it is going to get almost as hot outside as it [} 2 is in, the water is in the pool. You are talking about hours 3 now. 4 HR. BUSTARD: Okay. Now the second method available 5 to the operator to respond to the loss of the high pressure 6 injection system is to reconnect the power conversion system 7 so that the heat from the core can be dumped to the power 8 conversion system and stop raising the temperatures of the 9 suppression pool, and upon initiation of this accident 10 sequence, the plant has closed its main steam isolation 11 valves. And PWR, PRAs have a, based on again normal operation 12 information, have a recovery function and it describes with 13 what probability the operator can recover his power conversion 14 system as a function of time after initiation of the accident 15 sequence. That recovery probability currently does not take 16 into account the fact that the HSIVs, inboard MSIVs, are 17 seeing this harsh environment and may not be able to open due 18 to the harsh environment. 19 Typical number for the recovery probability is in 20 the 20 to the minus 3 range, 7, 10 to the minus 3, from a PRA 21 perspective, and what we have for this sequence at Peuch 22 Sottom is given the normal operation reliability for the SRVs 23 system, and the normal operation reliability for the HSIV 24 system, we get a core damage frequency contribution in the 10 0 25 to the minus 8th range, which is down around a half to 1 HERITAGE REPORTING CORPORATION -- (202)628-4888
157 , 1 percent of the total Peach Bottom core damage frequency, so () 2 again the TW sequence does not look like a major contributor 3 to core damage frequency at Peach Bottom, just like the 4 sequences in the PWR case that we gave previously-- 5 MR. MICHELSON: Well, I don't understand why. It is 6 entirely likely I might be in the lower right-hand corner of 7 your chart. 8 MR. BUSTARD: are saying why aren't we down 9 here? 10 HR. MICHELSON: Yes. Inside of containment now; 11 your SRV solenoid sees it, and so does your VV. If one ! 12 goes, they likely both go, very similar if nos anticol, and 12 why aren't you down in the corner? Then you would be quite 14 significant. 15 HR. BUSTARD: Our conclusion of this study is that, 16 like I pointed out before, we believe that there is enough ( 17 information like the research tests that have been done on 18 solenoid operators, and the substantial number of information 19 noticee and things like that, to suggest that we possibly , 20 should be moving in this direction, and this direction, but 21 there is not sufficient reliability information for those 22 accident conditions to characterize exactly where we are on , 23 this table. t 24 MR. MICHELSON: If the suppression chamber is O 25 already over 240, which I thought was 220 or thirty, somewhere HERITAGE REPORTING CORPORATION -- (202)628-4888
158 1 in that range, which was the number you had in your report, if
-('i 2 it already is that hot inside of containment, you think those
\./
3 solenoids will still survive? 4 MR. BUSTARD: We are talking 200 to 240 degrees 5 Farenheit, and yes, in some of the NRC tests, they took them 6 up to 340 degrees Farenheit, and they didn't start to get 7 failures until maybe four hours into the test sequence. 8 MR. MICHELSON: We are talking 10 12 hourc into the 9 event I thought we start to need these things. I thought by 10 then the probability of failure would be one for both of them. 11 MR. KOLACZKOWSKI: Let me just add a comment to 12 that. When you first need the SRV, it would be like six, s 13 eight, ten hours into the scenario, and you are about 210 to 14 240 degrees. Then you get a, be able to make sure they stay 15 open, and in fact if the containment should subcequently 16 either be vented or failed, you then ask for the SRVs to be 17 reopened later on in order to make sure that the core cooling i 18 remains. I 19 There is a possibility that the SRVs and MSIVs will 20 undergo a transient by which you will see temperatures in this l 21 sequence as high as about 350, 360 degrees, and pressures in 22 the containment approaching 200 pounds gauge before the 23 containment finally fails. 24 MR. MICHELSON: MSIV won't even work because the l
) 25 solenoid won't lift.
l 1 l l l HERITAGE REPORTING CORPORATION -- (202)628-4888
159 1 MA. BUSTARD: That's right, but the issue is after (} 2 they lift, after the containment fails, can we recover the low 3 pressure injection system and depressurize the primary systedf 4 MR. MICHELSON: This is- you are assuming 5 containment failure already? 6 MR. KOLACZKOWSKI: Either failure or vented; no. 7 Remember that the environment is getting harsh before the 8 containment has failed. At some point the containment 9 pressuri-en enough so the SRVs cannot remain open. 10 MR. MICHELSON: That's right. 11 MR. KOLACZKOWSKI: Now you have got a race because 12 you are not injecting low pressure coolant. The SRVs are 13 closed and the containment is approaching its failure point. O-14 Let's assume that we vent the containment or that it fails at 15 that point. Okay. Now the containment depressurizes my 16 instrument air pressure, not nitrogen. Pressure is now 17 sufficient to be able to reopen the SRVs again. Back pressure 18 is not high. Now the question is, now the question is will 19 the SRVs reopen, or can I get the MSIVs open? 20 MR. MICHELSON: How about vent all of this into the 21 reactor building now and-- 22 MR. KOLACZKOWSKI: That's another issue. I 23 MR. MICHELSON. This is a flaky business at best. I L 24 believe much of this is, really stretches the imagination. O 25 MR. KOLACZKOWSKI: All that Larry is saying is that HERITAGE REPORTING CORPORATION -- (202)628-4888 i
160 1 currently PRAs will estimate that a typical probability 2 failure of'this equipment under the very scenario we described 3 puts you on a sequence probability that somewhere over in the 4 upper left-hand corner of the table in the case of Peach 5 Bottom, and I think the point Lary is trying to make or is 6 about to make is that there are a number of iaformation 7 notices, et cetera, relative to SOV type of operation and 8 recognizing that this equipment will be seeing for many hours 9 temperatures of higher than 300 degrees, pressures 10 approaching a hundred pounds, that whether the equipment 11 continues to work and perhaps even work after the containment 12 has failed, with the notices saying that maybe there is some 13 suggestions that qualitatively that may not be true, you start O 11 moving to the right on the SRV. You start moving dew; on 15 MSIVs, and we don't knov where we are at. 16 MR. MICHELSON: If you believe that, then it is very 17 risk significant. 18 MR. KOLACZKOWSKI: Yes. The more and mers of those, 19 those failure probabilities go up, the more you move toward 20 the lower right-hand corner of the table. 21 MR. MICHELSON: You have an update already on the 22 equipment to know that it is getting very questionable that it 23 will survive and operate under those conditions. Therefore, 24 you are in the lower right-hand corner.
/ 25 MR. BUSTARD: I'm not, I don't want to go all the HERITAGE REPORTING CORPORATION -- (202)628-4888
161 1 way to the lower right-hand corner without getting additional
- (~'\ 2 'ini rmation.
's /
3 MR. MICHELSON: I wouldn't want to race there,
-4 'either, but I think that is where you are headed pretty fast.
5 If you believe any of this, you have got to kind of end up 6 there. 7 MR. BELL: That is not the purpose of the study. 8 The purpose of the study is to point out there may be a trend 9 in that direction. I hope you are not going to conclude that 10 we are in the lower right-hand corner, and I know you are not. 11 MR. BUSTARD: No. What we are saying right now, 12 there is not sufficient accident reliability information for 13 SOVs in this type of environment to tell exactly where we are 14 on this table. 15 Once again, let me point out that we have not yet 16 lost core cooling in this situation. What we are concerned 17 with is maintaining core cooling. Therefore, we don't have a 18 radiation environment inside the dry well, and most of the 19 test experiences that have been reported on, for example, that 20 give us some negative information regarding SOVs after four or 21 five hours worth of operability, were performed on SOVs that 22 had seen 200 megarads worth of irradiation, so there was some 23 substantial conservatism with respect to radiation in the 24 existing test data, and that-- O 25 MR. MICHELSON: As I recall, the SOVs were tested at HERITAGE REPORTING CORPORATION -- (202)628-4888
162 l' fairly high. temperature but not high pressure. 2 How far did they go? k) m 3 MR. BUSTARD: Well, my understanding of what 4 Franklin did was that in their first series of tests, they 5 went up to like-340 degrees Farenheit, but only went up to 6 about 80 psi. 7 MR. MICHELSON: Did they go that high? 8 MR. BUSTARD: Yes. And in their second test, they 9 actually did 340 degrees Farenheit with saturate steam, which 10 means they were up at about 110 psi. So they had some il failures let's say of their Valcors four or five hours into 12 the test. In one sense that test was conservative. It had _ 13 200 megarads worth of irradiation. It was several hours at 14 these high temperatures. In the other sonst, it was not 15 conservative because it didn't go out for the 15 hours 16 durations or so that we are talking about in this particular 17 scene. 18 MR. MICHELSON: The valves, they tested pneumatic 19 seals for the wiring, the factory integral type seals, or were 20 they normally field installed? 21 MR. FARMER: Very special seals that were made up of 22 potting, RTV. They are probably more-- 23 MR. MICHELSON: Are those factory installed? In 24 other words, the next guy that has one of these is going to O 25 have the same seal that you tested, or is he going to have his HERITAGE REPORTING CORPORATION -- (202)628-4888
163 P 1 own in situ? a/') 2 MR. FARMER: He will have his own. Q
- 3. MR. MICHELSON: Some of those have.already been 4 shown not to be so good or maybe not even there.
5 MR. FARMER: The effort in the test program was to 6 remove the seal. 7 MR. MICHELSON: They will, will go in a hurry 8 without a sealant, the coil compartment. 9 (Slide) 10 MR. BUSTARD: Let me just make several consents 11 about this set of equipment operations. Both the SRVs and 12 MSIVs have substantial redundancy, and we don't need to use 13 all of them to achieve our operation, so that's an additional 14 conservatism that we have. 15 As I have mentioned, research testing and NRC 16 inspection experiences have highlighted concerns regarding the 17 important components, namely, the SOVs and the MOVs. The 18 reason we raise MOVs is because to open the MSIV may require 19 opening the bypass valve first to equalize pressure. 20 MR. MICHELSON: You are depending upon non-safety 21 grade air system in many cases to fill accumulator tanks to 22 keep the SOVs going, isn't that correct? For BWRs now? 23 MR. KOLACZKOWSKI: You have the instrument nitrogen 24 system that is normal. O 25 MR. MICHELSON: Not every plant has the same, r HERITAGE REPORTING CORPORATION -- (202)628-4888
164 1 .MR. KOLACZKOWSKI: That is usually backed by 2 instrument air and certainly the ADS portion of the valves on 3 th'e BWR will have their accumulators as well. 4 MR. MICHELSON: How much time, how many cycles are 5 on the accumulator? 6 MR. KOLACZKOWSKI: Very typically you hear. numbers 7 of like five cycles or steady continuous for I forget how 8 long. I can't remember. 9 MR. MICHELSON: We are talking steady continuous. 10 How long do you think the accumulator will last? 11 HR. KOLACZKOWSKI: Move from one valve to another 12 valve to another valve assuming they keep operating. 13 MR. MICHELSON: Eventually you may need only one at O 14 a time. 1 15 MR. KOLACZKOWSKI: You are low on decay heat and 16 getting lower all the type. You only need one valve. 17 MR. MICHELSON: I just wondered has anybody gone 18 through and timed out what the cycles would be for 10, 12 19 hours? You see how many they have got to air to even run them 20 for 10, 12 hours, and also what is the effects of the elevated 21 temperature on the loss of air from these valves? 22 See, tolenoids and so forth may leak severely at 23 elevated temperature and drain down the accumulator while it 24 is trying to hold the main valve open, in which case you might 25 find that you've only got one cycle of operation and no more, HERITAGE REPORTING CORPORATION -- (202)628-4888
. - . - - - - - - - , . . . - _ . . - . . - _ . - . . . - . , . - . . . - . - . , - . - . . . . . . _ ~
r-165 1 and I doubt the test would have been that sophisticated. 2 MR. KOLACZKOWSKI: I'm sure they-haven't.
. g/~T 3 MR. MICHELSON: There is a lot of uncertainties in 4 this business, and you don't want to lose site of them.
5 MR. KOLACZKOWSKI: All right. 6 MR. MICHELSON: That's a big one. 7 MR. BUSTARD: Right, and to help overcome that 8 uncertainty, reliability testing during these types of 9 conditions'is warranted. 10 (Slide) 11 MR. BUSTARD: As I mentioned previously, once again, 12 for this particular set of equipment operations, extended s 13 radiation exposures are not applicable from a PRA standpoint, 14 and PRAs are only interested again in a few days of accident 15 operability. Hence for these particular equipment operations, 16 we can again conclude that certain EQ issues are not risk 17 significant. 18 MR. MICHELSON: Did you look at the accumulator 19 pressures now in terms of the fact that the containment is 20 raising its pressure and you did point out in your report a 21 couple of cases you worried about the back pressure on the 22 solenoids? 23 I am wondering if that has been factored in in terms 24 of how many cycles of operation you get off the accumulator O 25 because when you get down to where yeur equilibrium, your HERITAGE REPORTING CORPORATION -- (202)628-4888
166 1 pressure accumulates, is down maybe 40 pounds and that's the 2 outside area. You obviously don't work at all.
'3 MR. BUSTARD: That's beyond the level of detail we 4 could do in this scoping study.
5 MR. MICHELSON: We may not have five stages of 6 operation even for elevated pressure conditions beyond our 7 normal design basis. 8 MR. BUSTARD: As I mentioned, several of the 9 historical EQ issues are not applicable to these components in 10 this particular set of sequences either, because PRA is not 11 interested in an extended radiation exposure, or because PRAs 12 are only interested in a few days of operability. 13 (Slide)- 14 MR. BUSTARD: Let me go through one more example and 15 that's for your high pressure injection systems on a, on BWRs. 16 They are important during a number of accident sequences such 17 as power loss transients, anticipated transients without SCRAM 18 events, small and intermediate break LOCAs, et cetera. They l 19 are located outside containments. 20 However, our sequences such as TW sequences or l 21 Station Blackout sequences can create environments either l 22 internal or external at these pumps that are more harsh than l 23 their design basis EQ requirements. For example, Station 24 BlacP- d transient events can lead to high suppression I () 25 pool . cures since the suppression pool becomes a suction HERITAGE REPORTING CORPORATION -- (202)628-4888
167 1 source for the pumps early in the accident sequence, and
} 2 Station Blackout'can produce high temperatures for the HPCI 3 and RCIC rooms.
4 As I have already mentioned when we were talking 5 about the TW sequence, PRAs do currently assume failure of the-6 pumps when the suppression pool temperatures exceed 210 to 240 7 degrees Farenheit. What we note is the designing and 8 qualifying of the HPCS pump to operate during higher 9 suppression pool temperatures could have a substantial impact 10 on overall core damage frequency. 11 For example, based on the draft NUREG 1150 analysis 12 of Grand Gulf, one would reduce the core damage frequency by 7 ,s 13 76 percent if you could show the HPCI pump could operate at O 14 higher temperatures than 210 to 240 degrees Farenheit. The 15 risk impact of the HPCI/RCIC pump environmental failure is 16 uncertain. Unlike the HPCS pumps, HPCI and RCIC don't have 17 their own dedicated diesel generator. They are relying on 18 station batteries for operator indications, et cetera, and 19 right now, PRAs consider that three factors would contribute 20 to failure of HPCI or RCIC several hours into the accident 21 sequence. They are battery depletion, possible high room 22 temperatures due to loss of cooling in the room that contains 23 HPCI and RCIC, and finally, the high suppression pool 24 temperatures that are the suction source for HPCI and RCIC. O 25 What we have is that solving from an environmental HERITAGE REPORTING CORPORATION -- (202)628-4888
d 168 1 qualification standpoint the environmental issues, namely, the 2 high room temperatures, and the high suction temperature { 3 issues, may not have a very large impact on risk because that 4 doesn't resolve the battery depletion issue, and so unless the 5 battery depletion issue is resolved, the environmetatal issues 6 are not going to solve the problem of the operability of the 7 HPCI and RCIC system during a Station Blackout. 8 For a TW sequence, we don't have to deal with the 9 consequences of battery depletion or high room temperatures 10 because we have the HIVC system and we have the station power. 11 What we have lost is the cooling to the suppression pool and 12 we have to deal with the impact of high suppression pool 13 temperatures on HPCI and RCIC. 14 Right now, because of this alternative cooling mode 15 that I mentioned earlier, namely, my ability to depressurize 16 the system through the SRVs or the reconnection of the power 17 conversion system, my HPCI and RCIC does not appear to be risk 18 significant because my alternative systems give me core damage 19 frequencies down in the 5 to the 10 to the minus 8 regime. 20 If, however, if the TW sequence, I felt that this was an 21 appropriate assessment with respect to my SRV operability or 22 MSIV operability, and I was further down here, then designing 23 and qualifying HPCI and RCIC to higher suppression pool 24 temperatures could have a large impact on risk. O 25 DR. SIESS: That 76 percent reduction, did that, 'I HERITAGE REPORTING CORPORATION -- (202)628-4888 .- .. ... . - . - , . - .- - . - _- . - . . -_- ,_ - . - . - _ _ . , - . - - - - . . -.=
169 1 didn't understand how that was divided between the battery 2 depletion and the high temperature. Was that all due to the {; 3 high temperature assuming the batteries were there? 4 MR. BUSTARD: No. For a plant such as Grand Gulf, 5 the HPCS pump, they have a dedicated diesel. 6 MR. KOLACZKOWSKI: And battery and service water 7 system. 8 MR. BUSTARD: What happens, the other issues don't 9 get you. The only issue that does get you is the high 10 suppression pool temperature. 11 DR. SIESS: Somebody improved the batteries; they 12 become, this becomes level? 13 MR. KOLACZKOWSKI: More important with the HPCI and O AS RCIC systems, yes. 15 MR. BUSTARD: Yes. Again, from an EQ standpoint, 16 current equipment qualifications specifications require time 17 periods on the order of a hundred day operability and include 18 a radiation environment for HPCI, RCIC and HPCS, and that 19 would not be risk significant from a PRA perspective. 20 DR. SIESS: Suppose I had a pump that I bought to 21 use in a non-harsh environment. What changes would I have to 22 make to that pump to meet the current specs for the hundred 23 days? 24 MR. BUSTARD: Would you! If you bought a pump that O 25 was net designed for harsh environments-- HERITAGE REPORTING CORPORATION -- (202)628-4888
170
.1 DR. SIESS: What would I have to do to it besides
{} 2 shipping it off to somewhere to be qualified?
- 3. MR. BUSTARD: It would depend on the pump. That's 4 the first thing that a manufacturer might do would be ship it 5 off and see if they could qualify it. If they could not 6 qualify it, they would have to look at why it failed.
7 For example, were the motor windings in the pump 8 made of insulation systems that were not designed for high 9 temperature or long-term high temperature use? If that was 10 the case, you would want to change the pump design to get 11 better motor windings. There would be a number of other 12 failure modes that would be brought into, into play there. 13 That would, you know, if you have--how do you get electrical 0- 14 input to the pump? What type of connection device do you use? 15 The French tested some of their motor operated valves, for 16 example. 17 CHAIRMAN WYLIE: Basically rely on a motor also like 18 a pump, and they found numerous failure modes, greases, that 19 could not tolerate the radiation environment, terminal blocks 20 that had too large of leakage current, stater windings that 21 were not appropriate, and then they design motor-opetated 22 valve to survive the types of environments that were specified 23 for its application. 24 DR. SIESS: These are all electrical issues. O 25 MR. MICHELSON: There is mechanical issues, too, HERITAGE REPORTING CORPORATION -- (202)628-4888
i e' 171 1 such as seal cooling.
?() 2 DR. SIESS: I didn't hear any. %/
3- CHAIRMAN WYLIE: He mentioned lubrication. 4 MR. BUSTARD: They had bearings problems in the 5 French test also. 6 MR. MICHELSON: Your primary, your RHR pumps, you 7 have to go back and do something about seal cooling because 8 that seal and so "orth is sized only for relatively low 9 *emperature. Depends on the selection. RHR probably will get 10 by. 11 DR. SIESS: All the peripherals come under the 12 heading of pumps? You say pumps? 13 MR. KOLACZKOWSKI: Yes. We say pumps. 14 DR. SIESS: Goes back to the bus. 15 MR. KOLACZKOWSKI: Whatever it takes to really make 16 that pump operate. 17 DR. SIESS: Sealed cooling, how far do you go down 18 the line? Service water? 19 MR. KOLACZKOWSKI: Yes. I mean ultimately, though, 20 you are going to run out through the bearing cool lines. Once 21 those lines leave this environmental area, that is going to be 22 harsh, now you are just dealing with the service water itself 23 which is located some other separate building. 24 DR. SIESS: Service water doesn't supply enough O 25 water to cool it at that temperature. HERITAGE REPORTING CORPORATION -- (202)628-4888
172 1 MR. MICHELSON: Put in a bigger cooler.
/'i 2 DR. SIESS: You will disable it.
.( / 3 MR. MICHELSON: No. You just put in a bigger 4 cooler. 5 MR. KOLACZKOWSKI: These pumps, Dr. Siess, these 6' pumps that he is talking about, that HPCI/RCIC pumps and so 7 on, they are cooled by the water they pump. That's what the 8 problem is. 9 MR. MICHELSON: RHR is not. 10 MR. KOLACZKOWSKI: I am talking about HPCI and RCIC. 11 So the suppression pool water that they are now, they are 12 using as intake is rising in temperature 240, 300 degrees,
- 13 whatever, and yet that's a self-cooled pump, doesn't rely on 14 external service water supply, so that is what happens.
15 Ultimately you start getting seal problems. You could 16 eventually get to bearing problems. 17 MR. MICHELSON: Turbine problems, too; you have to 18 have another water to cool the turbines. 19 HR. KOLACZKOWSKI: So both the NRC contractors have 20 gone back to people like Terry Turban and said give us your 21 best shot at what temperatures before the pumps will quit, and 22 that's what IDCOR and the NRC contractors have gotten back to 23 people like Terry Turban and so on, that they are willing to 24 guarantee a number in the low hundred range. After that, they O 25 kind of go, oh, I don't know. HERITAGE REPORTING CORPORATION -- (202)628-4888 __ - __ - . . _ _ ._. _ ~ . _ __ _ _ _ _ _ _ . _ ~
173 1 DR. SIESS: ' Conservative high, conservative low?
'T 2 MR. KOLACZKOWSKI: I like to think they are (V
3 conservative low.
.4 MR. MICHELSON: You have got to figure out how to 5 cool'the electronic governor on it, for instance. You can't 6 run that thing up to those temperatures, figure out some kind 7 of exotic room cooler for the electronics on it.
8 MR. KOLACZKOWSKI: Once the seals start to go, you 9 start putting steam into the room. 10 MR. MICHELSON: Hect in the room will get the 11 governor. 12 MR. BUSTARD: In the remaining section of the
- 13 handouts that you have, we talk about PWR containment fans,
\s 14 and high range radiation monitors. they are also covered in 15 the draft report. Given thet it is 2:30 and we have to be 16 finished by 3:30, I thought I would move on to the general 17 conclusions and recommendations of the study, and that's the 18 handout. 19 MR. MICHELSON: I thought our schedule was five 20 o'cloch. 21 HR. DEY: You have another agenda item. 22 MR. MICHELSON: Your part is done. 23 MR. KOLACZKOWSKI: We have to make room for soinebody 24 else. O 25 HR. BUSTARD: The first two vugraphs in your handout HERITAGE REPORTING CORPORATION -- (202)628-4888
174 1 are the ones I showed early this morning to bring you up to 2- speed on what we would be talking about later this afternoon. (J') 3 And let me review them once again. 4 From a general conclusion standpoint, without 5 getting deeply involved in the specifics of any one given 6 equipment operation that we examined, we have some general 7 conclusions, and the first one is that PRA accident timing 8 perspectives provide a basis for modifying EQ practices, and I 9 will be providing an additional vugraph on each one of these 10 in a minute. 11 Second one is the PRA perspective suggested the 12 importance of the accident radiation doses overemphasizing the 13 EQ testing. 14 The third general conclusion is that when PRA 15 perspectives are combined with the existing EQ perspectives, 16 several historical EQ issues can be considered not risk 17 significant. 18 Fourth one is when PRA perspectives are combined 19 with existing EQ perspectives, several historical EQ issues 20 have possible risk significance. 21 The fifth one is that the lack of information 22 regarding the reliability of certain omponents in accident 23 environments creates non-conservative biases in important 24 accident sequences and base case core damage frequencies. O 25 No. 6 is that there are risk important system HERITAGE REPORTING CORPORATION -- (202)628-4888
175 1 operations that-rely on non-safety-related equipment or on 2 equipment that must function during non-or beyond design-basis {'s
.J 3 event environments.
4 Seventh is that the risk importance of plant status 5 equipment is difficult to quantify using PRA techniques, as 6 currently practiced, and the eighth is the risk importance of 7 certain EQ issues could not be characterized by the scoping 8 study using existing PRA models. 9 MR. MICHELSON: Before you leave that, I thought 10 item 5, I sense that in reading the report that this lack of 11 information regarding reliability of certain components in 12 accident situations was leading to an underestimate of the 13 potential risk involved and not necessarily the O 14 non-conservative biases. 15 I got the sense of, from the report that there are 16 many areac in which not knowing how it performs and knowing 17 that it is going to be less reliable than we anticipated 18 simply because we are throwing in all these other factors, 19 none of which worked for improved reliability, all work for 20 reducing reliability, that indeed the situation is worse than 21 we predict. 22 MR. BOHN: This is what we are trying to say. 23 Non-conservative shift point voltages are lower than we think 24 they ought to be. 25 MR. MICHELSON: That means that it is not being HERITAGE REPORTING CORPORATION -- (202)628-4888
-176
'l-~ biased in a conservative direction?
(~T- 2 HR. BUSTARD: -No. That's, that's not the case. %/ 3 DR.~SIESS: It is biased in non-conservative 4 directinn? 5 MR. MICHELSON: Maybe I don't know how to' read. 6 MR. BOHN: What you said is exactly right. The risk 7 is too low. 8 MR. MICHELSON: The risk is too low. 9 MR. BOHN: We don't know exactly how much. 10 HR. MICHELSON: It is higher than we think, but we 11 don't know how much higher necessarily? 12 DR. SIESS: You don't really know that the risk is-13 higher. You just know that the calculated risk on this basis 14 is calculated to be higher than what you get on another basis; 15 not your idiom that close to the truth. 16 MR. MICHELSON: The risk is lower on a calculated 17 basis. 18 MR. DEY: There is uncertainty. 19 DR. SIESS: The calculated risk would go up, could 20 go up if you had this information. There are other 21 conservative biases in PRAs. 22 MR. MICHELSON: Yes. 23 DR. SIESS: Let's don't confuse PRAs with real life! 24 MR. MICHELSON: They are far from it. O 25 MR. BUSTARD: Okay. General conclusion No. 1 was HERITAGE REPORTING CORPORATION -- (202)628-4888
7, 177 1 that PRA_ accident timing perspectives provide a basis for 2 possibly modifying EQ practices. The conclusion from PRAs is 3 that the core damage frequency and risk are substantially 4 dependent (n1 equipment operability during the first few hours 5 to days of an accident sequence. 6 This contrasts with EQ practice that may qualify 7 some components for up to a year. And therefore, the 8 implication of the PRA conclusion is that less emphasis be 9 placed on demonstrating long duration accident equipment 10 operability. 11 Once we draw that conclusion, and we combine that 12 conclusion, request certain perspectives from our EQ data , 13 base, we can come to the conclusion that certain EQ issues are O 14 therefore not risk significant in the context of PRA, and 15 examples would be certain things like the correctness of 16 post-accident acceleration techniques, the use of simultaneous 17 versus sequential accident simulations, and the importance of 18 oxygen within the LOCA test chamber. 19 (Slide) 20 MR. BUSTARD: The second general conclusion is that 21 the importance of the accident radiation dose is 22 over-emphasized. The PRA perspective is that PRAs do not 23 consider substantial in containment radiation conditions to 24 exist until core melt has occurred. PRAs calculate the core O 25 melt won't start until about half an hour to several hours HERITAGE REPORTING CORPORATION -- (202)628-4888
178 1 after accident sequence initiation. In the case of like a TW 2 sequence, it can be a day after accident initiation. {} 3 Moreover, if core melt should occur, PRAs determine 4 that the risk significant containment failures will occur 5 -within a few hours to a few days of that core melt, so the 6 overall PRA perspective is that for, for a portion of the 7 accident sequence, I don't have to worry about radiation 8 conditions, and that portion of the accident sequence for 9 which I do is of limited duration--a few days, and risk goes-10 down beyond that. 11 DR. SIESS: Suppose the PRA is over conservative, l 12 and the containment failure doesn't occur in a few days? Does 13 this change the picture? Or do you get sufficient decay in 14 fission products that the radiation still isn't a concern? l 15 MR. KOLACZKOWSKI: The point here is that-- 16 DR. SIESS: There was 90,000 curies in TMI 17 containment for several months. I assume that didn't do any 18 damage, or wouldn't. 19 MR. KOLACZKOWSKI: The point is that if you just 20 forget about what is going on in terms of the accident 21 scenario and just say what does my source term look like as 22 time proceeds, and then I release that, whatever point I want 23 to, as time goes on, the source term is getting, as we all 24 know, lower and lower and lower so that even if that source O 25 term were to be released, PRAs suggest, you know, the risk l HERITAGE REPORTING CORPORATION -- (202)628-4888
- 1 179
- 1. from that is fairly low.
2 -DR. SIESS: I-was talking about damage survivability. 3 of the equipment. 4' MR. KOLACZKOWSKI: It may. 5 DR. SIESS: The statement is made-that PRAs 6 calculate the failures will occur within a few hours to a few 7 days. Therefore, the equipment isn't going to be subject to 8 radiation for more than a few hours or a few days. 9 MR. KOLACZKOWSKI: No, that is not the implication. 10 The implication is that because PRAs say that containment 11 failures'are likely to occur within a few hours to few days, 12 that-- 13 DR. SIESS: Later you don't care. fs - U 14 MR. KOLACZKOWSKI: Later than this even though the 15 equipment may see radiation environments longer and therefore 16 subsequently fail, the source term that you will release will 17 be less and so the consequences of the accident will be less. 18 DR. SIESS: So if applied in there somewhere that 19 you realize that it is or is not a large release. 20- MR. KOLACZKOWSKI: Yes, I understand. 21 DR. SIESS: You wouldn't want to quantify it, would 22 you? I am taking a poll! 23 HR. KOLACZKOWSKI: I understand, but what the PRA is 24 saying is that the consequences, the risk of the accident is a O 25 lot worse early in the accident than it is a lot later. HERITAGE REPORTING CORPORATION -- (202)628-4888
180 1 MR. BUSTARD: They are also saying that the 2 probability of containment failure is higher early in the (~} 3 accident if I have a vessel breach, and I get direct 4 containment. 5 DR. SIESS: I thought PRA said if I lose containment 6 cooling, probability of containment failure is one? 7 MR. BUSTARD: That's if I put that in conjunction 8 with things like direct containment heating or hydrogen burn. 9 DR. SIESS: No. Lose containment heating and have 10 nothing but decay heat. Where is it going? Increase 11 suppression in containment, eventually lower pressurizer? 12 MR. KOLACZKOWSKI: Ultimately it would if you didn't 13 do something, but the boiler answer to that is we vent, and we 14 do c controlled, a controlled venting of the containment. 15 DR. SIESS: If I lose containment heating anyway, 16 period, large dry, lot of--if I lose containment heating, the 17 probability of containment failure is one; only a matter of 18 time. If I get cooling back, if I vent it, lots of things I 19 can do, but if I don't do anything, it seems to me a basic 20 postulation of PRAs it is going to fail. 21 MR. KOLACZKOWSKI: Yes. 22 DR. SIESS: But the consequences will be negligible? 23 MR. KOLACZKOWSKI: Will be less. , 24 MR. MICHELSON: You are looking at--your venting O 25 capability of the, at least the ones that you looked at in HERITAGE REPORTING CORPORATION -- (202)628-4888
181 1 your study, did you find all of-them had such a venting
.g- '2 capability, reasonable or reliability of operation?
.)
3 DR. SIESS: I think it is built in. The test we 4 made cui steel containment in concrete conditions says it is 5 vent. 6 MR. BOHN: That is what I was getting at. 7 MR. MICHELSON: The statement that was made I 8 thought the inference was they would be vented by operator 9 action, not by failure. 10 MR. KOLACZKOWSKI: I wasn't implying any reliability 11 in my numbers. 12 MR. MICHELSON: Some cases the provisions aren't 13 there that are operable in the ranges that we are really O 14 talking about, these big containment-- 15 MR. KOLACZKOWSKI: I understand. 16 !!R . BOHN: That is exactly right. 17 DR. SIESS: They don't have torque, you could open 18 any valves at all. 19 MR. MICHELSON: You have got to find a valve you can 20 open the vent with. 21' DR. SIESS: As I recall, if you can get an opening 22 equivalent to about six square inch hole, containment won't 23 overpressure. 24 MR. MICHELSON: You have to find such a vent. l 25 DR. SIESS: One of those 500 openings ought to be l l HERITAGE REPORTI!!G CORPORATION -- (202)628-4888
'r i 182 1- about right.
(3-
%/
2- MR. BELL: Combination. 3 DR. SIESS: Probability is already well over zero. 4- -(A, discussion was held off the record.) 5 MR. BUSTARD: Why don't I start over again, continue 6 along the line of discussing the radiation overconservatism 7 in, possible overconservatism in current EQ testing? 8 Steam generator level detectors are an example where 9 the current radiation overconservatism may impact risk 10 adversely, and that's something that could be looked into and 11 should be looked into a little bit more. 12 General conclusion No. 3, when PRA perspectives are
- f. 13 combined with the existing EQ research perspectives, several 14 historical EQ issues can be considered not risk significant.
15 Here are some examples. 16 What I mean by combining PRA perspectives with EQ 17 perspectives, Bill Farmer mentioned this morning the example 18 where there was a concern high range radiation monitors may 19 fail when the high temperatures associated with the steam 20 conditions of an accident were superimposed with the high 21 radiation dose rates, and the experimental evidence suggested 22 that that was not the case. 23 We have not found in our testing any in containment 24 component--so far, any in containment component so far where O 25 there is an immediate or early impact based on whether we do HERITAGE REPORTING CORPORATION -- (202)628-4888 :
183 1 simultaneous versus sequential testing. -What we have is 2 longer duration impasse like a cable may degrade more rapidly O' 3 in a simultaneous test than it does in a sequential test, but 4 that won't show up until a week or two into the accident, and 5 then if you combine that perspective with the perspective from 6 PRA, that risk is strongly impacted by early accident 7 performance. We can come to the conclusion that this issue 8 does not have a strong risk impact. 9 In contrast, when PRA perspectives are combined with 10 the existing EQ perspectives, several historical EQ issues do 11 have possible risk significance. 12 For example, if we look at the steam generator level 13 detectors that have a role to play in auxiliary feedwater, O 14 clearly we want to maintain adequate sealing and protection of 15 those circuits from moisture intrusion, and given the 16 potential risk impa:t of not doing that through auxiliary 17 feed, we can say that when the PRA perspective is combined 18 with the existing EQ perspective on what would happen if we 19 had adverse condensation errects, that there is potential risk 20 significance associated with this issue. The same is true for 21 a number of other issues, and these are some of the examples 22 out of the report. 23 (Slide) 24 MR. BUSTARD: General conclusion No. 5, and this O 25 gets back to the wording discussion we had earlier, is that HERITAGE REPORTING CORPORATION -- (202)628-4888
4 184 x 1 the lack of information regarding the reliability of certain (} 2 components in accident environments creates a non-conservative 3 bias in the important accident sequences and base case core 4 4 damage frequencies that are determined by PRA *echniques. 5 Just to summarize that particular conclusion, the EQ 6 process rarely supplies accident equipment reliability 7 information. PRA analysts' techniques possibly underestimate 8 the importance of harsh environment equipment operations 9 because of various features of how they perform risk 10 achievement analysis in current methodology. 11 Hence when we did our study, we looked to see 12 whether there were sequences where equipment operation during 13 harsh environments could have an impact on sequence O 14 probabilities, and two sets of sequences that were identified 15 were degraded harsh environment equipments may impact current 16 PRA perceptions, and they were the transient-induced small 17 break LOCA PWR sequences that involved operability of the 18 PORVs, and the steam generator level detection circuits in 19 harsh environment, and the BWR transient sequences with losses 20 of suppression pool cooling, the required operability of 21 solenoid operators and motor operators in a high temperature, 22 high pressure environment. 23 DR. SIESS: Both of these are severe accident 24 sequences? O 25 MR. BUSTARD: The small break LOCA is not a severe HERITAGE REPORTING CORPORATION -- (202)628-4888
185 1 accident sequence. 2 DR. SIESS: It doesn't say small break LOCA. It
~ )-
3 says--oh, that's small break LOCA sequence. That is a design 4 basis accident? 5 MR. KOLACZKOWSKI: At least within the envelope of 6 design basis accident. 7 DR. SIESS: The transient induced is a severe 8 accident, something beyond the design basis and start ing of f 9 with the transient? 10 MR. BUSTARD: Right. 11 DR. SIESS: Okay. Thank you. 12 MR. MICHELSON: Small break LOCA can be the 13 initiator of severe accident. 14 DR. SIESS: Yes. Anything can be the initiator. 15 HR. MICHELSON: Potentially. 16 DR. SIESS: I gather for small break LOCA itself, 17 well, you know, unless this initiated a severe accident, there 18 wouldn't be any risk. 19 MR. BUSTARD: Right, but we are talking about 20 equipment operability. In this particular example, we are 21 talking about equipment operability during environments that 22 are enveloped by traditional EQ profiles. 23 DR. SIESS: Small break LOCA created an environment 24 which could affect the equipment and thus leads down the line 25 to a more severe accident with some risk to the public? HERITAGE REPORTING CORPORATION -- (202)628-4888
r; < 186 1 MR. BUSTARD: Right; r'ight.
- 2. (Slide) 3 MR. BUSTARD: General conclusion No. 6, there are.
4 risk important system operations that rely on 5 ncn-safety-related equipment or on equipment that must 6 function during environments that are-not characterized by.the . 7 design basis event, EQ parameters. 8 And examples of those are the fact that for some 9 plants, the PORV, solenoids and block valve motor operators 10 are not included on EQ master lists, and hence the system 11 operation relies on non-safety-related equipment. 12 Other examples that are tied more closely to the 13 fact that the equipment must function during non-DBE 14 environments is that the HPCS pump operability when subjected 15 to high tempe.ature working fluids, from high, the high 16 suppression pool, high temperature suppression pool, the SRV 17 solenoids, MSIV solenoids, MSIV bypass valve motor operators 18 during TW sequence, and the low pressure injection equipment 19 after BWR containment rupturing or vent. 20 DR. SIESS: I understand the last three are 21 environments due to accidents beyond the design basis. 22 MR. BUSTARD: Yes. 23 DR. SIESS: But the first one is not? 24 MR. BUSTARD: That's true--to environments that are (~) 25 within the design basis, but the equipment may not be HERITAGE REPORTING CORPORATION -- (202)628-4888
187 I 1 considered on an EQ master list. (~) 2 DR. SIESS: Now why does the regulations permit U 3 equipment to be classified as non-safety related if its 4 failure during the design basis accident could lead to a core 5 damage? 6 MR. DEY: Assumption for other failures in that 7 sequence. 8 MR. BOHN: Feed and bleed just wasn't considered-- 9 feed and bleed just isn't--used to be considered a safety, a 10 safety option. In fact, until a couple of years ago most 11 plants didn't have procedures for feed and bleed. Still today 12 that PORVs are not tested in a feed and bleed mode. They were
,.s 13 just not, ?.his was just not considered an optio.t. ! )
U 14 DR. SIESS: I know, but you are stating a fact, but 15 the question was why do the regulations permit these 16 components to be clensified as not safety-related when clearly 17 they are safety related and their failure during a design 18 basis event could lead to severe core damage, which is what I 19 am hearing? 20 MR. KOLACZKOWSKI: You have to remember that 21 licensing power plants is not based on probabalistic grounds. 22 It is based on definite set of regulations, criteria. Among 23 those is the single failure criteria, and all the licensee had 24 to assume was that he had a small break LOCA, and then let's f-i'"3/ 25 say he failed one of his three auxiliary feedwater pumps. HERITAGE REPORTING CORPORATION -- (202)628-4888
188-L1 DR. FIESS: These were not safety related because 2 they were redundant. 1( ) 3 MR. KOLACZKOWSK1: Essentially, as Mike pointed out, 4 the feed and bleed mode wasn't even considered or had to be 5 considered because you didn't lose all of the aux feed in that 6 scenario. 7 DR. SIESS: These become of concern here only 8 because the environmental conditions may lead to a common 9 failure of redundant trains? 10 MR. KOLACZKOWSKI: And/or if you introduce the 11 concept of probabilities that there is a probability of 12 failure beyond a single failure. 13 . MR. BELL: Of the aux feed system. 14 DR. SIESS: There is always probability of failure. 15 MR. KOLACZKOWSKI: Yes, we understand. We know 16 that. 17 DR. SIESS: Even if they are safety related. 18 MR. MICHELSON: It is almost one. 19 DR. SIESS: No--for safety related. 20 MR. BOHN: The question is--if indeed a plant does 1 21 institute feed and bleed into their emergency procedures, they . 22 have to, of course, submit a plan for showing they have done 23 appropriate thermal hydraulic calculations. 24 The question is if they do that, does that also 25 imply that then the equipment required has to be looked at i HERITAGE REPORTING CORPORATION -- (202)628-4888
I 189
-1 from'a safety, viewpoint? I don't know.
2 DR. SIESS: The redundant systems--you didn't answer 3 me I guess because safety-related systems are made redundant 4 with single failure rate, and now you are saying that they 5 were not, safety-related systems were made redundant to meet 6 the single failure criteria, but the single failure criteria 7 didn't apply to non-safety-related systems, so how did these 8 get to be non-safety-related systems if they have got to work? 3 MR. KOLACZKOWSKI: As I pointed out--lets assume 10 they didn't have to work at one time. All you had to do was 11 take the small LOCA, and assume one failure in the auxiliary 12 feedwater system. You still have plenty of aux feed. You are 13 out of the accident in licensing space. You are done. O v 14 As soon as you start saying there is a probability I 15 can lose all of aux feed either because of the steam generator 16 level detectors, and/or just random failures of aux feede now 17 the next thing you have to go to is feed and bleed and now you 18 are relying on, in some plants, a non-safety-related PORV to 19 work, but it didn't have to be co.isidered in licensing space. 20 MR. MICHELSON: Thermal hydraulics may not be ; 21 working, for instance, but rather the natural circulation-- 22 DR. SIESS: I hear you, and if you say it fast, I 23 understand it, but I have to think about it. 24 MR. KOLACZKOWSKI: Like Station Blackout isn't a O' 25 design basis accident because in that you have got some loss HERITAGE REPORTING CORPORATION -- (202)628-4888
190 1 of off-site power and failure of all your diesels. es 2 DR. SIESS: The diesel is the safety-related system.
)
]\m/ 3 MR. KOLACZKOWSKI: I understand. 4 DR. SIESS: And we are talking about 5 non-safety-related systems that must function during a design 6 basis accident. 7 MR. KOLACZKOWSKI: If a sufficient amount of 8 safety-related equipment is, has already failed; in this case, 9 aux feed. 10 MR. DEY: The definition of design basis accident 11 aould include that failure of aux feed system functioning so 12 it wouldn't-- 13 DR. SIESS: Design basis accidents that assumed a O 14 large LOCA and a loss of off-site power simultaneously; others 15 that didn't assume, only assumed one failure. 16 MR. BELL: It is just this sort of exchange that 17 makes me wonder whether, I was concerned whether the PORV 18 should even have been brought into the realm because it again 19 is tne aux feed system, single failure, et cetera. 20 MR. MICHELSON: But the aux feed system does, you 21 know, good if the thermal hydraulics isn't working for you. 22 Start bleeding down the primary side and if your thermal 23 hydraulics doesn't work for you, then our answer is we will go 24 to feed and bleed. You can't say you are going to go feed and r^ 25 bleed unless it is qualified to work under the conditions HERITAGE REPORTING CORPORATION -- (202)628-4888
191 1 existing at the time. Thermal hydraulics looks pretty good on
-fg 2 Westinghouse and does not look quite so good on B&W, Just
; b.
3 don't know that. 4 MR. BELL: How many plants do you think have it in 5 their emergency procedures to use feed and bleed? Does 6 anybody have a feel for that? 7 HR. BOHN: Right now, it is most of them I 8 understand. 9- MR. MICHELSON: The staff told me all of them the 10 other day, but I didn't want to vouch for that. 11 HR. DEY: Option-based procedures. 12 DR. SIESS: How many plants have emergency 13 procedures? O 14 MR. MICHELSON: They say all of them. 15 MR. BUSTARD: Why don't we move on to-- 16- MR. BOHN: Let me make a comment here. Although it 17 is beyond the scope of this work, if one were to examine this 18 from the context of external events where you are guaranteed a 19 loss of off-site power essentially, so the reliability of the l P l 20 aux feedwater system, no longer do you have the two electrical i l 21 pumps, you are down to turbine pumps, so the reliability l 22 number for aux feedwater is no longer the five times 10 to the 1 23 minus 5 number up there but is now down to 10 to the minus 2 24 range, now the importance of feed and bleed becomes even more. 25 DR. SIESS: Where did you get that guarantee, the HERITAGE REPORTING CORPORATION -- (202)628-4888 L e ,e- - . . - - - - - , - ,
--.,--..-,-,-.,-.-..n, -e _,-_.,e,..w._-,,-m-n, ,,----,----,,----.,-,--,-_..-,,.v,.n.- nn--.,-,-- ..- -
192 1 loss of off-site power? Guaranteed to m- means probability of 2 one.
'{ }
3 MR. BOHN: I won't say that. 4 DR. SIESS: I can't think of any external event that 5 is going to lose off-site power with the probability of one. 6 MR. MICHELSON: Sure--fire. 7 DR. SIESS: That is not probability of one for 8 reactor year. You are not going to get a fire every year. 9 MR. MICHELSON: Given the fire. 10 DR. SIESS: I know, but cut the wires. 11 HR. BOHN: Let's say for all intents and purposes 12 the probability of losing off-site power is greater than a
- 13 tenth, but based on the historical data it is somewhere around 6 14 probability of a half or something; loss of ceramic insulators 15 in the yard.
15 DR. SIESS: Two tenths G earthquake. l 17 MR. BOHN: No, sir. It is quite a bit lower than 18 that. It is between 118 and .25 or something. l 19 DR. SIESS: That is pretty close to .2. 20 MR. BOHN: I'm sorry--two thirds of G. 21 DR. SIESS: .2G. 22 HR. BOHN: I'm sorry. I thought you said two thirds 23 of G. 24 DR. SIECS: It is for lots of plants in the eastern O 25 U.S. about 10 to the minus--probability per year. Let's don't t HERITAGE REPORTING CORPORATION -- (202)628-4888
193 1 gets the loss of off-site power, you know, guarantee. 2 MR. BOHN: It is a probability of .5 which PRA space f]v 3 might as well be one. That's why I-- 4 DR. SIESS: 6.5. 5 MR. BOHN: Off-site' power is .2. 'You have .2. 6 DR. SIESS: .2G is one. 7 MR. BOHN: Pardon? 8 DR. SIESS: I would say at .2G earthquake, the 9 probability of loss of off-site is one, but the probability of 10 loss of off-site-power is not .2 or anything else. It is 11 whatever the probability of earthquake. That's a conditional 12 probability. 13 MR. BOHN: That is correct. Given an earthquake, 14 you are very likely to lose off-site power. That was my 15 initial statement, so given that situation, then all I am 16 pointing out is that the probability of the reliability of the 17 aux feedwater is no longer the very small numbers that you 18 used in internal event analysis where you have two electric 19 and one turbine pump. In fact if you have lost off-site 20 power, you don't have two motor pumps, you are down to one l 21 turbine pump, if you sort of look at that table in the light 22 of external event, you see that the relative importance of 23 this issue can be significantly greater. 24 DR. SIESS: Relative to something else. 25 MR. BOHN: Relative to those numbers. I HERITAGE REPORTING CORPORATION -- (202)628-4888 I
194 1 DR. SIESS: Dynamiting the transmission towers (y 2 around the plants is about the same as that probability you V 3 are looking at. 4 MR. MICHELSON: But in some parts of the country the 5 probability of a tornado taking out the transmission line is 6 higher than seismic. 7 DR. SIESS: I don't think so. 8 MR. MICHELSON: One in 5,000 a year is Sequoyah. 9 DR. SIESS: That is two times 10 to the minus 4, and 10 I say 10 to the minus 3, 10 to the minus 4. 11 MR. MICHELSON: During that range; I thought the 12 SSE-- 13 CHAIRMAN WYLIE: Given the tornado hit the plant is O 14 one 15 MR. BOHN: The 10--it is not dominant sequences in 16 most of the seismic PRAs today. That is one of the reasons I 17 brought the issue up. 18 MR. BUSTARD: Let me move on to general conclusion 19 No. 7, which is that the risk importance of plant status 20 equipment is difficult to quantify using correct methods. 21 The plant status instrumentation is not well modeled 22 by PRA fault trees, and PRAs basically assume that the plant 23 status information is reliable, and based actuation failure 24 response on human error in response to reliable information, 25 so we were not able to devwlop a lot of conclusions on plant HERITAGE REPORTING CORPORATION -- (202)628-4888
195 1- status instrumentation due to this lack of models. l (T 2' What'are our recommendations? We have several () 3 general recommendations. 4 (Slide) 5 MR. BUSTARD: And they include the following-five 6 . items. One is the development of accident reliability data 7 for selected components and accident sequences. 8 MR. MICHELSON: liow do you do that? 9 MR. BUSTARD: That would be something that would i 10 have been worked out. We have, I mean in the case of an SOV, E 2 11 their volume is small enough that you.could put many of them 12 into a test chamber and test the SOVs. 4 13 MR. MICHELSON: Now you have got to have accident. O 14 and flow conditions and in some cases, I mean not the SOV 15 necessarily, but for components, you really are talking about 16 the real world of the component is in at the time of accident 17 and addressing the accident. 18 MR. BUSTARD: I am not, I am not trying to 19 generalize this to every EQ component in every accident. What 20 I e.n trying to do is say there are selected components that 21 can be highlighted from PRA analysis and selected accident 22 conditions where there is uncertainty in the PRA analysis 23 because one lacks reliability information luring accident 24 conditions. 25 DR. SIESS: Wait a minute. Which number are you HERITAGE REPORTING CORPORATION -- (202)628-4888
196-1 talking about? ('] '2 MR. BUSTARD: Number 1, No. 1. V 3 DR. SIESS: That says development of accident 4 reliability data. 5 HR. BUSTARD: For selected components and selected-6 . accident sequences. 7- MR. MICHELSON: Big enough sample to.be meaningful. 8 DR. SIESS: When you have that data, you are going 9 to make PRAs? ' 10 MR. BUSTARD: When I have that data, I am going to 11 make PRAs. 12 DR. SIESS: Do PRAs using it? 13 MR. BUSTARD: I could use PRAs to get a better 14 assessment of the risk, yes. 15 DR. SIESS: Or better assessment of the importance 16 of the qualifications of risk. 17 HR. BUSTARD: Yes. , 18 DR. SIESS: If you can't get the data--suppose you 19 get that data fmr three components and the seven components 20 that you can't get it for, is the gain and is the reduction in 21 uncertainty noticable? I hate to bring up the word 22 uncertainty, but clearly that's-- 23 MR. KOLACZKOWSKI: If we could resolve the steam
^
24 generator level of the reliability and the PORVs, we couldn't O 25 do the MSIVs for some reason, we could at least resolve better ,. r i HERITAGE REPORTING CORPORATION -- (202)628-4888
197 1 in a quantifi' cation sense whereby for along those two axises (~y- 2 if we move at all, we went on those two aux feed and the PORV V 3 axis, we could resolve some PWR sequences. We wouldn't be 4 able to resolve the BWR sequence. 5 DR. SIESS: What does resolve mean? What-do we do 6' with it? 7 MR. KOLACZKOWSKI: Have better estimate; now that we 8 are factoring in the EQ induced failure rate of the equipment, 9 we have a better estimate as to what core damage frequency or 10 the risk is from those sequences. 11 MR. MICHELSON: Reduce uncertainty of that risk. 12 DR. SIESS: The thing is you are introducing the 13 uncertainty. It wasn't there before. O 14 MR. MICHELSON: That's true. 15 DR. SIESS: You have discovered it. Now you have 16 got to do something about reducing it. 17 MR. BOHN: I think that first bullet, what we say-- 18 DR. SIESS: What do we do about it? If the risk is 19 low enough, we don't do anything. 20 MR. BUSTARD: Let me-- 21 DR. SIESS: Right? 22 MR. KOLACZKOWSKI: Right. 23 MR. BUSTARD: Let me go to some of the examples here 24 for this particular bullet. 25 MR. MICHELSON: You are coming back to the other HERITAGE REPORTING CORPORATION -- (202)628-4888
198 1 slide later? (}
~
2 HR. BUSTARD: I am coming back. No doubt about it. 3 DR. SIESS: You have got a slide on each of the 4 items?
~
5 HR. BUSTARD: Okay. Let's go down here to an issue 6- that we have not spent a lot of time talking about today, 7 which is the low pressure injection equipment reliability 8 during steam temperature conditions after BWR containment. 9 venting and rupturing. 10 One reason we haven't discussed that in detail today 11- is because as part of NUREG 1150 analysis, they did an expert 12 opinion solicitation on exactly this issue, so they said wo 13 need to know the accident reliability information for the low 14 pressure injection system in a BWR venting type scenario, and > 15 they gathered some experts together and tried to do an initial 16 assessment of that. 17 DR. SIESS: That is one way to get the data. 18 MR. BUSTARD: That's right. 19 DR. SIESS: A group of experts will always come out 20 with a larger spread than a smaller group of experts. I don't > 21 know anybody has done a study on that. 22 MR. BUSTARD: What I am driving at is that in some 23 cases, the PRA community has already realized that there may 24 be a need to move beyond normal operation reliability date, O 25 and try to develop accident reliability data. HERITAGE REPORTING CORPORATION -- (202)628-4888
199 1 What we have here is after going through this EQ 2 risk scoping study and bringing in some EQ perspectives, we 3 highlight areas where developing data would be useful such as 4 solid operated valve reliability during PWR small break LOCAs, 5 and BWR TW steam pressure conditions, motor operator valve 6 reliability again during the small break LOCA or TW steam 7 pressure conditions, steam generator level detection 8 reliability during small break LOCAs, steam temperature 9 conditions and the HPCS reliability under high suppression 10 pool. 11 DR. SIESS: Now you are asking us, you know, what 12 kind of research needs to be done? And you are saying here 13 that one thing that would be is to develop better reliability O 14 data on these components under the harsh conditions. Now 15 those data would then be fed into PRAs and telling you how 16 important it was that these particular components be reliable, 17 or be more reliable than they are. 18 The results might be that it is a no never mind. 19 Making it more reliable isn't going to help. > 20 MR. MICHELSON: I misunderstood then. I thought 21 that was a part of this study showing what-- 22 MR. KOLACZKOWSKI: Show him another example of one 23 of those. 24 DR. SIESS: You think that making it more reliable 25 than you think they are--you don't know how reliable they are HERITAGE REPORTING CORPORATION -- (202)628-4888
200
'l now so how do you know where you are on that matrix?
2 MR. BOHN: I think these issues, that calling it (V") 3 reliability data, I think is an overstatement of what we are 4 suggesting here. 5 What I would suggest is, for example, they are 6 . qualified at a certain leve1~, and supposing one were to test 7 ten of these valves or something. Maybe it is the relief 8 valves, and one found out in fact that seven or eight out of 9 ten didn't function at that level. 10 Now this would say that No. 1, we are not sure what 11 reliability is, but it is certainly not the reliability down 12 at the historical normal environment level. 13 And secondly, you would conclude that in fact if you 14 did have, if the reliability was ten or something like that, 15 you would say that in an accident situation, I really can't 16 depend on whatever options use those, and so instead of trying 17 to get data at that point, the logical thing to do would be to 18 upgrade the valves in some sense. 19 DR. SIESS: Look at this diagram. Clearly as you 20 move down into the right, you degrade the system. Were are , 21 you now? 22 MR. KOLACZKOWSKI: That's the point. PRA analysts 23 think they are in the upper left. All we are saying is--if 24 you start thinking about EQ induced failures of equipment, and t-25 we have identified sequences in which this equipment would HERITAGE REPORTING CORPORATION -- (202)628-4888
201 1 have to operate in a harsh environment, we might be worst 2 case, way down in the lower right. 3 DR. SIESS: Okay. 4 MR. KOLACZKOWSKI: All we are saying in order to 5 determine where we really are, we either need data base-- 6 DR. SIESS: Probability you are up to the left? 7 MR. MICHELSON: Of course. 8 DR. SIESS: You are nodding. You don't think that 9 is likely? 10 MR. KOLACZKOWSKI: Oh, all I am saying-- 11 DR. SIESS: I start with the PRA analysts. 12 MR. KOLACZKOWSKI: He hasn't considered environment. 13 DR. SIESS: Therefore, you think from the basis 14 that, of your studies, that there is no way you can go except o 15 down to the right? 16 MR. KOLACZKOWSKI: That's right. That is correct. 17 DR. SIESS: No probability at all things are better 18 than they are assuming they are? 19 MR. KOLACZKOWSKI: Assuming that our nominal data 20 base is reasonably accurate, yes. 21 DR. SIESS: I am not assuming anything. I really 22 don't know what it is. ; 23 MR. KOLACZKOWSKI: Environment effects can only make 24 things worse. That is true. They can't improve it any. 25 DR. SIESS: That data base, they have taken absolute f HERITAGE REPORTING CORPORATION -- (202)628-4888
202 1- good best estimate reliability data base for their point? 2 MR. KOLACZKOWSKI: Right. (~)) 3 DR. SIESS: There is not something that is going to 4 make this any better? 5 HR. KOLACZKOWSKI: That is correct. 6 MR. MICHELSON: I guess there is a small finite 7 chance they were wrong to begin with and it even is better 8 than-- 9 DR. SIESS: A lot of PRAs, it is more than small and 10 finite. If they have taken a good best estimate, if they are 11 in a data base that used extreme values, to be conservative, 12 they could well have moved over far enough to cover some part 13 of what is in here. 14 HR. KOLACZKOWSKI: All we are saying, if we can 15 collect data or get some information from the PRA perspective 16 as to what is the reliability of the equipment, given it would 17 have to operate in this environment, if it turned out that 18 they are just as reliable in that environment as the numbers 19 that the PRA analysts was using before, then we wouldn't move 20 off the left part of the curve. 21 DR. SIESS: With that data things move down into the 22 right, then somebody can make the decisian two times 10 to the 23 minus 7 is good enough. 24 Now suppose it isn't good enough. We now have O 25 information somewhere in the research we are talking about to HERITAGE REPORTING CORPORATION -- (202)628-4888
203 1 tell us what we can do to make it better? {J 2 MR. KOLACZKOWSKI: Hopefully the tests would 3 indicate that, the test data would indicate what are the 4 failure modes, and how could we harden the equipment.to 4 5 operate in that or find a redundant way of performing these 6 functions? 7 DR. SIESS: Okay. 8 HR. BOHN: The unstated concern here is that when 9 qualification on equipment is performed, certain items fail, 10 you keep testing until you find one that does work. That's 11 the way equipment qualification works. They don't have, you a 12 don't report the ones that don't work. You just report that I i 13 you have several that do work. And from--you know, we have 14 all had experiences where they have tested several and they 15 finally get one to work and that's the one that they report, 2 16 which implies to us that this sort of type testing approach 17 does not provide much reliability even at the qualification 18 level. f 19 DR. SIESS: They are all alike. They are not l 20 proving anything. I'm surprised anybody accepts that. l 21 MR. BOHN: That's the whole basis of the, our EQ 22 prograrc l 23 MR. MICHELSON: You are getting closer, Chet. 24 DR. SIESS: You keep testing until one of them O 25 passes? HERITAGE REPORTING CORPORATION -- (202)628-4888
- 4. 204 1 MR. BOHN: That is correct, sir.
/'N 2 .MR. BELL: Don't suggest that is--they might try a U
3 little difference in the seal mechanism and get the apparatus 4 to work, and-- 5 MR. MICHELSON: One of the big problems has been 6 seals. 7 DR. SIESS: That is not testing the same component. 8 MR. MICHELSON: They don't use that seal out in the 9 field. That is installed by a different group of people, 10 perhaps different specs. Hopefully yot have got a qualified 11 seal spec. 12 DR. SIESS: If that is the kind of EQ program we 13 have got, forget about it. 14 MR. MICHELSON: You are getting close. 15 DR. SIESS: This isn't going to change that. 16 MR. FARMER: I think you said it correctly. They 17 don't, they do try to do things, but the process isn't one of 18 radical change. It is one of refinement until you do pass. ! 19 DR. SIESS: There is nothing in in progress is going i ! 20 to change it. You tell them to qualify it to twice the l l 21 temperature and twice the pressure, they can still do it the 22 same way you are talking about. 23 HR. BOHN: What I was suggesting the qualification i l 24 of, I would like to see a test of--say it is clear that the ( 25 relief valve on BWR are crucial items. You don't have to do a l HERITAGE REPORTING CORPORATION -- (202)628-4888
._ - - - _ . .- . - . . . , . . ., ., . , ,.. - . - . ~ . . - . - . . - - . . , , . - . .
205 1 PRA to recognize that. Everyone knows~that that's why they (} 2 put 18 of them on it, to provide redundancy, and yet the 3 concerns of common mode failure that are raised here would say 4 we would like to have you test ten of these things under these 5 adverse conditions, and let's see how many really work. 6 MR. MICHELSON: I am not sure all'18 qualified. 7 DR. SIESS: The statistician would shoot you. A 1 8 statistician would shoot you. 9 MR. BOHN: That's why I don't want to get hung up in 10 sample sizes. 11 DR. SIESS: From ten tests? It is ridiculous. 12 MR. BOHN: Suppose out of 17, seven fail? 13 DR. SIESS: Suppose one out of those ten tests fail? 14 MR. BOHN: How would you feel about depending on 15 these things under adverse environments? 16 DR. SIESS: On diesel they had to prove that there 17 was a 95 percent chance that they would start? 18 MR. MICHELSON: That was reliability test. 19 DR. SIESS: They had 298 tests without failure. 20 MR. BOHN: I understand. I understand that, but 21 that is not the issue here. To me as a rational engineer, if 22 you tested ten of these things, and two of them failed, you 23 might say this is not an appropriate qualification test. We ; 24 are depending too heavily on those safety valves. 25 DR. SIESS: You are still wrong. The rational t HERITAGE REPORTING CORPORATION -- (202)628-4888
206 1 engineer would want to know why those two failed and the other eight passed. Without approaching probabalistically the [}- 2 3 sample ten, you look~for the reason for those two failures, 4- you would say it is not a random thing or if it is, that is 5 not enough margin. That is what you would do rationally. 6 MR. BOHN: I agree. 7 MR. BUSTARD: I only have a few more 8 recommendations, 9 DR. SIESS: You have given entirely too much concern 10 to getting us statistical data base used only in PRAs. PRAs 11 don't make anything safe or unsafe. 12 MR. BUSTARD: I have a few more recommendations to 13 go over in the next five or ten minutes, and the second 14 general recommendation we have coming out of our study is that 15 additional definition of risk importance should be determined , 16 for equipment important to core melt arrest and post-core melt F l 17 accident management strategies, and for plant status 18 instrumentation. 19 The third general recommendation is that some 20 efforts on redefinition of the EQ regulatory basis could be 21 pursued and the examples of areas that could be investigated 22 would include the instantaneous radiation release requirement, 23 and the current practice of qualifying equipment for very long 24 durations. O 25 (Slide) HERITAGE REPORTING CORPORATION -- (202)628-4888 j 1
207 1' HR. BUSTARD: The fourth recommendation is that PRA r i
] )
2 techniques coupled with EQ data bases and insights can provide 3 improved guidance regarding operations, maintenance, and ; 4 inspection activities. For example, it can highlight the 5 priority for making sure that the steam generator level 6 detection circuite are immune from moisture intrusion and 7 condensation effects. 8 And finally, PRA techniques again when combined with ! 9 EQ perspectives can be used to provide improved guidance , 10 regarding equipment aging issues. And for example, one thing 11 that could be further investigated would be humidity, aging : 12 effects on the electrical penetrations with polymid 13 insulation. 14 MR. MICHELSON: The word electrical penetration, do ; 15 you mean the ones penetrating through the containment wall? 16 MR. BUSTARD: Yes. i 17 MR. MICHELSON: Not the electrical penetration where 18 you come off a cable train, go into a conduit, seal up the 19 conduit? t 20 MR. BUSTARD: No. That is seal. [ t 21 MR. MICHELSON: That is tested, too, I hope. I 22 think it is. There have been prc; rams that determine how to i 23 make up such a seal or--that is qualified, is that correct? f 24 MR. BUSTARD: That's right. Those are qualified [ 25 also, but so are-- l HERITAGE REPORTING CORPORATION -- (202)628-4888
208
.l' MR. MICHELSON: Not necessarily a part of the test
(} .2 of the solenoid valve, for instance?
-3 MR. BUSTARD: No. In fact, there has been cases
'4 historically where a manufacturer will take his component and 5 simply qualify it and net qualify the sealing technique 6 because he doesn't know what sealing technique the utility is
- 7. going to use.
8 HR. MICHELSON: Uses his own in the laboratory, but
'9 that isn't necessarily the one used in the field.
10 HR. BUSTARD: Tnat's right. 11 CHAIRMAN WYLIE: Any further questions? 12 MR. MICHELSON: How about general recommendation No. T F 14 CHAIRMAN WYLIE: That was the one he just read. 15 MR. BUSTARD: I just did that one. ! 16 MR. BELL: One quick question--the figures in the 17 upper left-hand corner of those tables, where the PRA 18 analysts, they come up with that reliability, is that based on 19 safety grade equipment or commercial grade or are they-- 20 DR. SIESS: They say there is not any difference. , 21 MR. KOLACZKOWSKI: You have to understand where the 22 data base comes from in the PRA, and that is basically just 23 going through and saying this valve has failed, and therefore 24 I count its failure somewhere in the overall data base that I O 25 am building. Okay. r I HERITAGE REPORTING CORPORATION -- (202)628-4888
l 209 l' MR. MICHELSON: It is not exposed to adverse 2 environments. It may nave failed because of adverse { 3 environment. That's just lost in the numbers. Most of them 4 are-- . 5 DR. SIESS: They said they can't--safety grade, I 6 non-safety grade. 7 MR. BELL: Our biggest premise is that-- 8 DR. SIESS: I like to attack the biggest premise. I 9 That is safety grade quality equipment counteracts the effect l i 10 of environment such that perhaps there is no difference in 11 reliability under accident coaditions. 12 MR. KOLACZKOWSKI: We don't know that, i 13 MR. BUSTARD: Let me answer that this way. There ( j 14 are-- i l 15 DR. SIESS: Wrap it with the paper, they will be 1 l 16 able to withstand-- l 17 MR. MICHELSON: It doesn't have conduit seals and so 18 forth if it is not safety related. You don't 19 environmentally-- 20 MR. BELL: I understand that. Take steam generator 21 level, safety grace. 22 MR. MICHELSON: They may not have the same gas, get 23 on the cover plate, things of this sort. 24 HR. BUSTARD: Moreover, I think, I don't think you O 25 can assume that just--let me give you an example out of HERITAGE PEPORTING CORPORATION -- (202)628-4888
210-1 every-day life experience. We have an aut> mobile that is 2 designed with tires to go 20,000 miles. Say that we try to
}
3 design an automobile to survive the Indianapolis 500. They 4 have tire blow-out 500 miles, so the assumption that 5 safety-related equipment is going to maintain normal operation 6 reliability in harsh environment has not been demonstrated. 7 MR. BELL: I understand. But has it been not 8 demonstrated? 9 MR. BUSTARD: There 1s-- 10 MR. BELL: I understand there is a question, but I 11 am not so sure we throw up all the red flags. 12 MR. BUSTARD: It is counterdemonstrated in the sense 13 that Franklin Labs has done testing on ASCO operators and 14 shown that after four hours worth of testing exposure, Valcor 15 operators start to degrade substantially, and they are 16 safety-related valve operators. 17 DR. SIESS: Have they tested non-safety-related l 18 valve operator and found it also failed the same way or it 19 fails in two hours? 20 MR. BELL: Maybe it does better. 21 MR. FARMER: They tested ones that were class 1E and 22 ones that weren't. They were essentially the same valve 23 except it hadn't been qualified, and the behaved essentially i l 24 the same. Qualification hasn't really changed the (~)
\' 25 reliability.
HERITAGE RE?ORTING CORPORATION -- (202)628-4888
211 1 DR. SIESS: That is, you know, in a way that's very 2' good because that broadens the data base. The data base can ( }_ 3 now include all valves of a'given type without regard to 4 whether they were category one or not, but we have asked PRA 5 analysts what distinction they make between reliability and 1E 6 and non-category 1E, and they say we just lump everything into 7 the data base. 8 MR. KOLACZKOWSKI: That is correct. 9 MR. MICHELSOM: For non-accident conditions. 10 DR. SIESS: Data base they are using. 11 MR. MICHELSON: It is non-accident data base. That 12 is fine as long as-- 13 MR. BOHN: You have to understand the data baso 14 comes from things like LERs which are reportable occurrences 15 related to safety systems, so to a large extent, they are 16 biased toward safety-related systems. 17 MR. MICHELSON: We find the LERs also show that, 18 that equipment should be protected against moisture intrusion 19 is not protected against moisture intrusion when we get 20 inadvertent actuation of spray, and yet you have to keep 21 asking yourself gee, I thought this equipment was qualified i 22 for that. 1 23 MR. BOHN: We are involved in the ongoing old INE 24 ongoing inspection program of plants, and we keep finding O 25 things exactly like this, that most of the inspections, and on HERITAGE REPORTING CORPORATION -- (202)628-4888
212 1 an isolated basis of course, but these seem to be out there 2 quite a bit.
}-
3 MR. MICHELSON: We are depending upon the non,'the 4 non-accident reliability to be indicative-of accident 5 reliability. When we do PRAs we are depending on that fact. '- We are not necessarily depending on it 6 DR. SIESS: 7 but analysts do it. 8 MR. MICHELSON: It is the only thing we have got. 9 We are using it. 10 DR. SIESS: Speak for yourself. I am not using 11 bottom line PRAs for anything, and that's just one of the 12 reasons why I don't. 13 CHAIRMAN WYLIE: Okay. 14 DR. SIESS: I think you guys have done a tremendous 15 job, and I am not arguing against Carl here. It seems obvious l 16 to me that there is no mechanical engineer or electrical 17 engineer--if I take a piece of equipment involving electric 18 motors and wires and so forth, and subject it to steam and 19 water anc high temperatures and higher pressures, it is a, for 20 example, a lot more likely to fail than if I don't. That's j P 21 just common sense. 3 22 What you are saying is can we quantify it, and can 4 23 we quantify that and stick it into PRA and quantify its , t 4 24 effects on health and safety of the publie? And if it isn't I ( 25 enough, what can we do about it? HERITAGE REPORTING CORPORATION -- (202)628-4888 ,
213 1 Well, like most other problems, if.you know what is 2 wrong, you can fix it. Now here you suspect something' wrong, 3 you don't know how important it is, and there is a proposal 4 how to get at it. If you find out it is important, we will r 5 fix it. There are all sorts of ways to fix it. The plants - 6 may end up looking like these things we have got out there . 7 now, but that's how they got that way. We fixed everything we 8 could think of, t 9 MR. MICHELSON: One of the suggestions earlier that 10 got passed over lightly because--I don't know why, that is 11 gee, if we just had a separate room out here with this certain 12 minimum amount of equipment needed to safely remove the decay 13 heat, we would be all right, and this has been proposed from ' O 14 time to time. Decay heat removal, but none of the facilities, 15 the agencies-- 16 DR. SIESS: What happens in that room if everything 17 is in the scme room? 18 MR. MICHELSON: There is another set that is more 19 isolated from it. 20 DR. SIESS: Going to be redundant bunker, dedicated? 21 MR. MICHELSON: No. Doesn't have to be redundant. 22 It is reparated from the rest of the plant. The rest of the 23 plant is separated from it. 24 CHAIRMAN WYLIE: Okay. Does staff have any othe.r-- 25 DR. SIESS: Exactly how they come to that. HERITAGE REPORTING CORPORATION -- (202)628-4888
214 1 MR. MICHELSON: That is why they came to it in tV(') 2 Germany. 3 DR. SIESS: I am not sure yet. he buy them, 4 MR. DEY: Yes, sir. 5 CHAIRMAN WYLIE: Okay. Moni, have you got something 6 else to say? 7 MR. DEY: No, other than if you would like a 8 briefing in the July meeting-- 9 MR. MICHELSON: Where does this, where do you go 10 from here with this thing? I mean just drop it in the black 11 hole? 12 EMR . DEY: No. As i mentioned this morning, we have 13 a plan to prepare a research letter by the end of September, ' 14 and we would like your input toward that fact. 15 DR. SIEPS: Report the results of this? 16 ER. JEY: T: that is an appropriate approach. 17 DR. SIESS: A RIL reports the reaults of research to 18 other offices, doesn't it? 19 MR. DEY: Right. 20 DR. SIESS: Seems to me-- 21 MR. DEY: It gives the recommendations. 22 DR. SIESS: What you are getting back from here is 23 recommendation for needed research. You don't need to write a 24 RIL to tell yourself you have research. O 25 MR. D',Y: Yes. HERITAGE REPORTING CORPORATION -- (202)628-4888
, . . , . . . ._- _ _ _ _ - . _ . ._. . _ _ . . _ _ _. .- .~ ~. - . _ . - . . . .
215 1 DR. SIESS: Any proposals for research are going to r~T. 2 ~come out of Research. What-I thought you were going to do V 3 with this stuff is now, is what research programs would help 4 to implement this? 5 MR. DEY: That's what this letter would do. 6 DR. SIESS: Why--oh, you have to use a need request. 7 You can't initiate a research request in Research for this 8 until--you have to sell it to NRR? 9 MR. DEY: Right. That letter would go to NRR with 10 concurrenta of an appropriate action plan. 11 DR. SIESS: RIL doesn't usually you have an action 12 plan. Does a research letter usually have an action letter? _ 13 MR. DEY: It has recommendations. , 14 DR. SIESS; You can make recommendations, ask for 15 the NRR concurrence. 16 MR. OEY: Yes. d 17 DR. SIESS: Is that in here? 18' MR. DEY: I have not indicated that, The results of 19 the study indicate there are no generic issues that would come 20 out. 21 MR. MICHELSON: There aren't any generic issues in 22 -here? 23 DR. SIESS: Not any new ones.
- 24 MR. DEY
- Not any new ones.
25 MR. MICHELSON: I got all excited on some of the HERITAGE REPORTING CORPORATION -- (202)628-4888
,0 216 1 things you-found and then you say gee, .this is not a new issue f
2 anyway. I think it is an issue of the qualification of
}
3 equipment to meet accident conditions. That's a generic 4 issue. 5 MR. DEY: I think they have three things. One is 6 any new generic issues. I don't think we do. We have closed 7 out some issues, and that was one of the objectives of the 8 study that were documented in the summary report of research 9 which which done in the past ten years. A lot of those issues 10 will be closed out. 11 Now third-- 12 DR. SIESS: You mean some of these issues on hou you 13 do environmental qualification? 14 MR. DEY: Right. 15 DR. SIESS: To me what this program has said is it, 16 what yoa qualify ia a heck of a lot more important than how l ! 17 you qualify it. A'.1 these questions about simultaneous versus , 18 sequential and this versus that and--are not nearly as 19 important as the fact they are not qualitied in certain things 20 like say the steam generator level indicators and so forth. 21 Now that is sort of a broad statement. It seems to 22 me also you said a lot of things we have been worrying about 23 aren't important and forget about them. Let's concentrate on 24 doing right things. Can you under the regulations require for C:) 25 the qualification for an environment that is harsher than that HERITAGE REPORTING CORPORATION -- (202)628-4888
217-1 coming from the design basis accident? F'i 2 MR. DEY: Not under current rule.
"V '
3- MIG KOLACZKOWSKI: You can on the basis of beyond 4 design basis accident go back and make the licensee now have-5 to move that envelope further up and tell them they have got ! 6 to qualify for severe accident. 7 MR. BOHN: There is one avenue as Reg Guide 197. 8 That's the only avenue that, where severe accident 9 qualifications is required. i 10 DR. SIESS: Had-to be qualified for; so you really i y 11 haven't got a legal basis. Now to find out that something ! 12 isn't qualified and it contributes to risk, can be done l i 13 through the implementation of severe accident policy, >:b e O
- You can't require it if they
! 14 individual plant examination. l' i 15 find out that steam generator level iadication is a, you know, l 16 real problem. They do the IPE, they still have to decide i 17 whether to fix it, cost / benefit and all of the.t stuff, 18 MR. KOLACZKOWSKI: Industry can, could ask a similar 19 question and saying that well, if you are telling me that some 20 of these issues are not risk significant, then doggone it, why 21 can't the regulations be relexed? 22 DR. SIESS: That makes sense. There is a good 23 answer. Nobody has relaxed a regulation yet. The regulations 24 only go one way. 25- MR. MICHELSON: PRA is right, and you find it is a HERITAGE REPORTING CORPORATION -- (202)628-4888
218 1 non-problem, then fine. You have got to do the PRA's {J 2 correctly first, and we are not doing them correctly
'3 presently.
4 DR. SIESS: Let's take a piece of equipment that is 5 now qualified for harsh environment inside containment. Do 6 your studies suggest it should be a harsher environment by 7 whatever terms you want? I got the impression that those 8 things that were qualified, the level of the environment 9 wasn't all that much off. Is that right? 10 MR. BUSTARD: That was true for like the steam 11 generator level detectors. That is not true for the SRVs and 12 the SOVs that are on, the MSIVs on the TR.
. 13 DR. SIESS: In terms of what?
14 MR. BUSTARD: Pr3ssure and temperature can be much 15 more extreme during a BWR TW sequence or more extreme in the 16 sense of much longer (.uration. 17 MR. KOLACZKOWSKI: Also higher. 18 HR. BUSTARD: And higher. 19 HR. BOHN: For example, three to five times the 20 temperatures are some of the numbers out of the performance 21 electrical aquipment report, and time spans are on the order 22 of hours instead of-- 23 DR. SIESS: Three to five; give me two degrees, one 24 at this and one at this. O 25 HR. BOHN: We are involved in that. HERITAGE REPORTING CORPORATION -- (202)628-4888 _. _~- ._ _- . _ _ . - _ -_ _ _ . - , . - _ _ _ _ - _ _ _ _ ___ _
219 1 DR. SIESS: Inside containment was steam line break t'] 2 accident which was 300 degrees or something. (/ 3 MR. BUSTARD: Typical qualification parameters for 4 inside containment, typical would be on the order of 340 5 degrees Farenheit. 6 DR. SIESS: That is steam line break. 7 MR. BUSTARD: And what we-- 8 DR. SIESS: Pressures of 60. 9 MR. BUSTARD: Pressures of 60 to 80 psi; however, if 10 they have been qualified using saturated steam conditions, 11 they will actually be qualified to pressure more like 110 psi. 12' DR. SIESS: Hundred and ten; you are not going to
~
13 oet much more than about a three times the pressure. They 14 fail in what kind of temperatures? 15 MA. DUSTARD: I have that in my report somewhere. 16 MR. KOLACZKOWSKI: It has approacned 400 degrees. 17 DR. SIESS; Instead of 340. 18 MR. KOLACZKOWSKI: Up to like Peach. Bottom's 19 estimate failure pressure now is about a, about 175 pounds, so 20 the SOV might have to survive 170 slow rise, 175 pounds and 21 depressurization after that. 22 DR. SIESS: Now your proposal to try to improve the 23 data base would be to try to get some of the things that are 24 now qualified for lower pressures looked at, for higher O 25 pressures some of the things that aren't qualified for HERITAGE REPORTING CORPORATION -- (202)628-4888
220 1 anything to be looked at.for harsh environment?
] 2 MR. MICHELSON: Also_you have got to know the 3 reliability of those things that are already qualified 4 currently for the environment, but we know nothing, limited 5 amount about the reliability of them. That's the thing that 6 is missing.
7 DR. SIESS: Well-- 8 MR. MICHELSON: I think that's a generic issue 9 myself. 10 DR. SIESS: It seems to me that things that are now 11 qualified for 340, if they are going to see 400, and are now 12 qualified for 400, I don't see where you are getting. 13 MR. MICHELSON: Once. having qualified them for 400, 14 now you also have to go back and estaolish chair reliabilit,y 15 of operation at 400. One qaalification, one piece of 16 equipment, doesn't establish reliability. 17 DR. SIESS: Well, that, I can't go along with that 18 kind of a recommendation because if I am going to establish 19 reliability on a statistically valid basis, I think I am out 20 of this world. 21 MR. MICHELSON: That's right. I think you are. 22 DR. SIESS: So I think at some point engineering 23 judgment is going to have to be used in place of PRA. 24 MR. MICHELSON: Probably more than one would have to O~' 25 be tested. HERITAGE REPORTING CORPORATION -- (202)628-4888
E 221 3 1 DR. SIESS: I understand why it failed. 'If I testad 2 a hundred critical components and 99 of them stayed within a 3 band like'this, and one of them ended up down here, that 4 doesn't give me high confidence. I want to know what trent 5 wrong :ba that one so it can't go wrong in the next one. I 6 don't know what it is. Until I know what it is, I can't be 7 sure that it isn't going to be in the ones I buy and put in my 8 plant. It is not something that obscure. 9 MR. MICHELSON: That's the point. 10 DR. SIESS: Purely statistical basis, this sort of 11 thing; quality control people can explain it. That's fine, 12 but production stuff, we turn out thousands of them, you watch 13 the variation, but-- 0 14 !!R . KOLACZKOWSKI: What w.a are trying to say is the 15 agency is going to begin to make decisions on the basis of not 16 only the deterministic regulatory criteria, also going to 17 involve PRA, either through the IPE process on current plants, 18 or through_PRAs for future plants. Should we ever build any 19 through the design, design certification approach, where PRAs 20 would have to be submitted as licensing documents, then the 21 PRAs need to be getting better and as accurate as possible and 22 all we are saying is that in order for PRA to begin to treat 23 EQ in a quantitative sense, we need data. 24 DR. SIESS: But now find somebody that is interested 25 in something other than EQ and they are going to need data, HERITAGE REPORTING CORPORATION -- (202)628-1888 _ _ . . ~ . - . . _ . _ . ~ . . . _ - . _ _ _ _ -..._ -.--.. _ ,_ ~._-._ _ - .
222 1 too. /'s 2 MR. MICHELSON: You have got to--EQ, EQ doesn't V. 3 always by that definition include the operating 4 characteristics.under accident conditions like isolating 5 blow-down and so forth. Some people think EQ means the 6 environment around the component, not the process fluid 7 passing through, but it has got to include the process fluid 8 passing through. 9 DR. SIESS: You have got to look at the whole
.10 picture. There is another kind of qualification that goes on 11 for a lot of components, and that is seismic, and I don't know 12 what the reliability is of the seismically qualified piece of 13 equipment. I put it on a shake table and I pushed it up 14 there. I haven't done enough of that. I am a structural 15 engineer. We don't build a hundred buildings to see how many 16 fall down. We build up our probabalistic approach by 17 elements, r.ot on the finished product.
18 MR. MICHELSON: Yours are not active components. 19 DR. SIESS: They fall down and we find out why they 20 fall down and try to fix that on the next one. We don't get 21 it statistically. 22 MR. BUSTARD: You still are trying to assess in some 23 way reliability. You are just not doing the test mode. 24 DR. SIESS: That's right. You do it by looking at O 25 what failed and what, why it failed. If we see a failure, we HERITAGE REPORTING CORPORATION -- (202)628-4888
223 1 try to find out what failed. Then we.go back and look at our {} 2 building codes to see how come, and if we can make a change in 3 the building codes that will prevent that failure the next 4 time, we do it. You can see the parallel between that and the 5 regulations.
'6 HR. BUSTARD: I believe my vugraphs were general 7 enough to say assess the reliability of the components rather 8 than necessary to test.
9_ DR. SIESS: I think, you know--but what bothers me 10 is that so many times we say we want this because we need it 11 for PRA. It is not our business, doing PRAs. Our business is 12 protect the health and safety of the public and PRA is one 13 tool that gives us some insights, and the least insightful 7.) O thing we get out of the PRA is the bottom line. 14 15 HR. MICHSLSON: Using it for safety goal bottom 16 line. ! 17 DR. SIE$S: Using it for safety goal because t t 18 somebody stated safety goal in probabalistic terms. 19 MR. MICHELSON: Unfortunately. I 20 DR. SIESS: If you state it in probabalistic terms, 21 you can only show compliance by probabalistic means. They 22 were careful enough not to define any of the terms exactly l 23 enough so nobody knows what it means. 24 MR. BOHN: PRA-- 25 DR. SIESS: IPE is going, is supposed to look for HERITAGE REPORTING CORPORATION -- (202)628-4888
~. - . . . . . ..
224 _ 1. outliers. And you know, maybe that steam generator level is
~
an outlier, but you don't have to have a very precise PRA to
; 2 3 find outliers. When you look at the ones that you have found 4 in the past, they were pretty obvious type things.
5 MR. BUSTARD: If we go to the PWR'TW sequence right 6 now as I pointed out, that's in the draft NUREG 1150 analysis. 7 That is considered a very small contribution to the overall 8 core damage frequency at Peach Bottom, and yet if you look at 9 that, that box we had up there, it could becomo very 10 important. 11 DR. SIESS: It is an outlier. It is going to be an 12 outlier. If the PRA has done what, the more realistic
, 13 probability failure in harsh environment, is a bio jump in 14 risk is what you are saying. That's an outlier.
15 MR. BUSTARD: That is an example of outlier. 16 DR. SIESS: Like the ceiling falling down at Indian 17 Point. Until somebody looked at it, it wasn't a probability. 18 MR. BOHN: How would it be an outlier since all of 19 them are qualified to the same, to the same standards, 20 presumably would affect the same component variety of plants? 21 DR. SIESS: I didn't mean plant by plant. It is 22 outlier in the risk profile of the plant; risk profile. 23 HR. BOHN: Well, I don't understand. You keep 24 emphasizing-- O 25 DR. SIESS: You are going to look, all you are HERITAGE REPORTING CORPORATION -- (202)628-4888
_ , . m .. _ _ _ _ . . . ._ _ 4 225 l' supposed to look for in the IPE is outliers. The plants 2 aren't as safe as we thought they were because of something we 3 have overlooked. That was sort of an ad hoc definition of an di outlier. That's the way--the severe accident statement says 5 we think most plants are safe if they fix the USIs and GSIs 6 and meet all the regulations, and since we made some PRAs,
-7 found things we hadn't expected like the building banging.
8 together, the ceiling falling down, and if we think, if people 9 look for them and find these unexpected modes of failure, that 10 they will be plenty safe. We are going to find some of them 11 or fix them. Outlier is something we didn't expect that makes 12 the risk bigger than we thought it was. That's a good 13 qualitative safety goal I guess. O
.14 MR. BOHN: All we are pointing out in this study 15 here is that a PRA is used to rank components relative to 16 concern, and there is a couple of areas that we pointed out 17 where equipment is not qualified for harsh environments in 18 dominant accident aequences, and to make judgments, many 19 things that we pointed out essentially get dropped out of PRAs 20 because they are so unlikely because they have redundancy, and 21 yet what we are pointing out is harsh environment can give you l
22 common mode failures for which they no longer drop out and can 23 be important, and like the seismic case, many of these things i 24 are easier to fix than to quantify.
'O 25 I certainly don't want anyone going away thinking we l
l HERITAGE REPORTING CORPORATION -- (202)628-4888
226 1 want to. develop reliability data bases on these equipment just 2 for doing PRAs. I think we have said that several times. It (')'T s
.3- is easier to fix than--
4 DR. SIESS: Watch your language then because-that's 5 the impression you give somebody. 6 MR. BOHN: If you know you have a problem, for 7 example, let's say you do a qual test on ten components, again 8 getting.away from the sample size, but you find out that in 9 fact a significant number, whatever that number is, one or 10 nine-- 11 MR. MICHELSON: Say you are testing one. That is 12 generally what you are doing. 13 MR. BOHN: You have to test more than one. Let's 14 say you test ten of them. You find out in fact half of them 15 really don't work which you think they are--at this point I 16 don't want to develop a reliabili+.y data base. I want to-- 17 DR. SIESS: Ain't qualified, period. 18 MR. BOHN: You want to fix the problem and put in a 19 piece of equipment that you can rely on, and that's the 20 approach. 21 MR. MICHELSON: What does that tell you about 22 accident operation reliability? 23 MR. BOHN: You are presuming it is under accident 24 condition appropriate to the-- O 25 MR. MICHELSON: But it isn't, of course. It is rare HERITAGE REPORTING CORPORATION -- (202)628-4888
227 1- that we could truly test under accident conditions. We only 2 simulate the environment now, not the process thac is going {} 3 on. This equipment is not under dynamic operation. When you 4 do this test, it is not trying to isolate a pipe break or 5 something, for instance. 6 HR. BOHN: I understand that, especially the issue 7 with the PORVs. It might have to cycle ten or twelve times in 8 two-phase blow or something like going through while they have 9 to do it. I understand. But one has to start somewhere, and 10 I think-- 11 DR. SIESS: Aren't mechanical engineers smart enough 12 to design a PORV that will cycle 12 times under a given set of 1] temperature and pressure? 14 MR. BOHN: I am sure they are. Until re test it, we 15 don't know, and they he.ven't been tested. 16 DR. SIESS: Bu'c I an saying Liven the conditions, 17 you know, aren't they as likely to come up with one that will 18 work as the people that design containments are to come up 19 with containment that will work? 20 HR. BOHN: Do you k"ow why we don't test them now? 21 DR. SIESS: Go back. What you are saying-- 22 MR. BOHN: They tend to be fragile and fail during I
- 23 the test.
24 DR. SIESS: You are saying what you have discovered 25 here is a PRA is being made with reliability data based on ! HERITAGE REPORTING CORPORATION -- (202)628-4888 i
228
.i normal operation and things are not likely to be a whole lot 2 better than that. .And in a harsh environment, common sense
(')) 3- tells you that they are liable to be whole lot less than that, 4 and you have got some quantitative studies on various degrees 5 shown in some cases, certain pieces of equipment, they fail in 6 the harsh environment, can really affect risk, significantly. 7 That, to me that's an outlier type of thing. 8 Now uhat is the fix? To qualify them for that 9 environment, some way to require it. If it was IPE, and it 10 af f ects risk enot gh, they have got to make a cost / benefit. 11 Maybe it is cost beneficial on that basis. The intermediate 12 step of getting some reliability data was simply to say 13 previous statement was qualitative, we think it is worse, and 14 tf it is enough worse, it is going to be an outlier. Maybe we 15 have got to find out how much worse it is, and that I haven't 16 got so much confidence in. 17 I asked you how you are going to do it. You said I 18 haven't thought about that yet. I think if you start this. king 19 about it, ar.d start thinking about it in terms of 20 statistically oriented people and find out what it would take 21 for them to get a number they would be willing to use, it may 22 turn out to be impossible. Thoir standards of what they will 23 take aren't so darned high--they will take expert in, lets 24 face it, but the thing is it may be difficult to devise a O 25 reasonable test program that is any better than expert opinion HERITAGE REPORTING CORPORATION -- (202)628-4888
229 1 for some things. 2 MR. BUSTARD: If we are talking about containment 3 fans-- 4 DR. SIESS: If it turns out a reasonable assumption 5 -regarding degradation, there is a significant degradation in 6 safety, increase in risk, then tx) me as an outlier, as you 7 said, a lot of them can easiy be fixed. Maybe this can. 8 MR. BOHN: The ACRS problem as I see it is not 9 getting reliability data. We have, we perform sensitivity 10 studies basically. What you don't know, and what--you don't 11 know that, so you don't know how to proceed basically, is you 12 don't know what that reliability is even at the qualification 13 level.
/.
(~ 14 I mean regardless of the limitations with modeling 15 and during the dynarsic accident conditions, if you went to a i 4 16 qual test, actually tested ten of them, and got some number of 17 reliability of ten worked and then compared that against the 18 normal reliability and find wide discrepancies, then you would 19 have the basis for saying hey, we do have a generic problem 20 and we think something needs to be done, but right now, all 21 you have is a sensitivity study. 22 DR. SIESS: Suppose you test one and find out that 23 it won't work? They are assuming it will. You don't have to 24 test ten. O 25 MR. BOHN: I will agree with that. HERITAGE REPORTING CORPORATION -- (202)628-4888
230 1~ DR. SIESS: If you test one, find-out it does work 2 just as well, you are going to be suspicious because there is
}
4 3 no reason it should work. 4 MR..HICHELSON: We have been discussing these in the 5 cold--LERs'in the reports, we have seen they don't work. 6 Those are kind of tests, too. 7 MR.'BOHN: I think it would be reasonable to do some 8 limited testing even'at, even though the tests might be 9 imperfect on a very few selected components, to just see what 10 kind of results you would get to determine where you would go 11 at this point especially since you are going to have to accept 12 the IPE severe accident results, and again they will be using 13 the old normal reliability data. 14 DR. SIESS: There are two kinds of tests you can 15 make. Let's take you are going to do ten or a hundred. One
- 1G is you can set up a set of conditions thac they must meet, and 17 you can test X number of units on a pass / fail basis.
18 The other is to vary the conditions and find out at 19 which level of temperature, pressure, or combination it fails. 20 Then you have some probability distribution. You say the i 21 probability of surviving this temperature, you know, is 95 22 percent. 23 Now these are entirely different things both in the 24 number of tests required, in the significance of the results. O 25 Just setting one set of conditions on a pass / fail basis PRA, I l HERITAGE REPORTING CORPORATION -- (202)628-4888
4 231-1 1 people won't know what to do with that, either. They will 2 stick that in, maybe they just stick that, come to the branch 3 point, say okay, eight of them passed, two of them failed, .I 4 put 80/20 there. 5 MR. KOLACZKOWSKI: That is probably what we would do 6 because we have nothing better. 7 DR. SIESS: And yet you wouldn't feel real happy 8 about it? 9 MR. KOLACZKOWSKI: No. 10 DR. SIESS: Would you rather have ten sets of ten?
.11 MR. KOLACZKOWSKI: I would like to get a little more 12 information than to say this could be a problem and therefore, 13 Mr. Utility, you have got to harden all your PORVs right now.
O 14 DR. SIESS: You can't make them do that. 15 MR. KOLACZKOW5KI: I shouldn't because I don't know , 16 it is a problem now, and what we are trying to propose is 17 let's get a little bit more information. Granted we are going 18 to crawl before we are, we can walk before we can run. 19 DR. SIESS: Some of the places you want more 20 information is on the equipment that has never been qualified 21 at any level; PORVs, non-safety related. 22 MR. KOLACZKOWSKI: I only see two choices. We 23 -either assume the PRA guy is right and it is not important, or 24 we go to the other extreme and say it is a problem, and 25 everybody has got to go out and do something. HERITAGE REPORTING CORPORATION -- (202)628-4888
..- -. - . - , . . . - . . . -. - .- . . - . . , - , . - - . ~ .
232 1 MR. MICHELSON: The study decided it was a {} 2 non-problem. There were serious flaws in it also by the 3 assumption that they make concerning one component at a time, 4 other. kinds of problems. So you have got to--if the PRA was 5 done right, then I.think what you say is correct, but the PRAs 6 so far haven't necessarily been done right. They have been 7 done with various degrees of rightness. 8 DR. SIESS: Don't laugh. Expert opinion, the.whole 9 seismic analysis is based on it. Everybody says that 10 dominates. And that curve of seismic has expert opinion. 11 MR. MICHELSON: We get to fire people, they tell us 12 the fire dominates. Depends on which hat they are talking
- 13 under on a given day.
14 CHAIRMAN WYLIE: Okay. Let's move on and wrap this 15 up. And we have got two hours at the main meeting. I don't 16 know exactly what day yet. You will have to talk to Rich to 17 get that nailed down, and I guess what we are looking for is 18 from what I can tell is synopsis, two-hour synopsis vf what 19 you have done here, we have talked all day about, so why don't 20 we shoot for that then? I don't think we can at this time say 21 any more than that. l. 22 MR. DEY: Okay. I think two hours is adequate. l i 23 MR. MICHELSON: Only thing they can leave out, is 24 there anything they could leave out of today's meeting all I
) 25 together?
HERITAGE REPOhTING CORPORATION -- (202)628-4888
=
233 1 CHAIRMAN WYLIE: I will leave it up to their 2 judgment to get the message. I think, I think--
-)
,)
3 MR. MICHELSON: I am not sure there is. 4 MR. BOHN: We can certainly compact it. 5 CHAIRMAN WYLIE: I think they ought.to take a look 6 at it with the idea of getting the message across that you 7 wanted to get across, and see what you have got to do to do 8 it. 9 MR. BOdN: That is no problem. 10 CHAIRMAN hYLIE: I think it is a good piece of work. 11 Obvously there-is a lot of interest from what was discussed 12 here today, and thank you. 13 We will come back ten after four.
/~'s G' 14 (A brief recess was taken )
15 CHAIRMAN WYLIE: We have an update on the 16 implementation of USI A-46, and T.Y. Chang is going to present 17 that. 18 MR. CHANG: My name is T.Y. Chang. I am the task 19 manager of USI A-46. 20 In October last year, we, I reported the status of 21 USI A-46 to the to the Full Committee, and today's i 22 presentation, the purpose of this is two-fold--first of all, 23 is to bring you up to date on where do we stand on the A-46 24 program? And secondly, I would like to report to you the O 25 lessons learned from the Nine Mile walk-down. i HERITAGE REPORTING CORPORATION -- (202)628-4888
234 s 1 (Slide) {} 2 MR. CHANG: A-general overview of status of A-46 3 program, the seismic experience data gathering is essentially-4 complete. A trial training program developed by the SQUG and 5 conducted by SQUG at the Nine Mile site took place December 6 last year, and after that, in February 1988, in February '88, 7 a trial walk-down at Nine Mile site took place. This is the 8 pilot plant for a BWR plant. As you recall, a walk-down on 9 the PWR plant took place in May last year. That's at Zion 10 Unit 2. 11 Also the preliminary draft of the relay screening 12 procedures was issued in July last year, and two trial years
, 13 of the procedures took place also at Zion and Nine Mile before 14 tne actual seismic walk-down.
1S MR. MICHELSON: Was that procedure very long? 16 HR. CHANG: It is in a book form. 17 MR. MICHELSON: It is a big thick book? 18 MR. CHANG: It is in a binder. It is prepared by, 19 by MPR for EPRI. 20 MR. ANDERSON: That's basically a systems review 21 procedure followed by the seismic capacity review of the 22 relays. 23 MR. MICHELSON: Detail on each type of relay you 24 mean? O 25 MR. SCHMIDT: You have two copies that have, that HERITAGE REPORTING CORPORATION -- (202)628-4888
-~ .- . . . . . _ , . _ _ - - _ __ _ _ _ - . _ . . - _ _ , , . _ _ _ _ _ . . _ . - _ _ _ -
= 235 1 were provided to your staff.
, 2 HR. MICHELSON: I haven't seen it. I don't want to 3 see the whole thing. Maybe the procedures are the little 4 short part of the beginning of it or something.
5 MR. ANDERSON: No. It is pretty long and 6_ complicated procedure. 7 MR. MICHELSON: Would you identify yourself for the 8 report? 9 MR. SCHMIDT: Bill Schmidt; I am the technical 10 coordinator of the SQUG group. 11 MR. CHANG: We have two gentlemen representing-SQUG 12 here today. Mr. Bill Schmidt is associated with MPR, and he 13 is as, he said, the technica'_ coordinator for A-46 program, 14 and Mr. Dick Schaffstall. He is with SMC. He is the 15 administrative coordinator for the SQUG activity. 16 And also we, from NCC we have Mr. Newt Anderson as 17 well. 18 MR. MICHELSON: Was the relay screening procedure 19 just mainly addressing that relay chatter question? 20 MR. CHANG: Yes, but first of all, there are mainly 21 two aspects of that procedure. First of all, you have to 22 identify the so-called essential relay. You have to identify 23 the systems going through the system review, and look at the 24 circuits and identify which relay is required to function () 25 during and right after earthquake. Those are the ones we are HERITAGE REPORTING CORPORATION -- (202)628-4888
236 1 concerned about because the chatter will affect its function.
~
2 MR. MICHELSON: Data sheets in that book told the
-{)
3 ' reviewer about the relay chatter aspects of that particular-- 4 MR. CHANG: How to go about doing the system review, 5 and at the second aspect of the procedure is to compare that, 6 the relay ruggedness, with the test data to make sure that it 7 is going to function. One important document is the, provided 8 ed by SQUG, generic implementation procedure. This document 9 ties together all the elements in the SQUG program, and this 10 one is provision zero. Before that we are, they have issued 11 three versions of the GIP, but all those versions are draft 12 versions.
-13 Okay. The A-46 implementation schedule, most 14 specifically I think this is geared to the SQUG program for 15 the' members of the SQUG group. There are 37 members in the 16 SQUG now, 34 domestic utilities, and the three foreign 17 utilities. That represents I think 65 plants out of all the 18 69 plants required to perform the A-46 walk-down.
19 As I say, said earlier, the revision zero version of 20 the GIP has been sent to NRC, and I believe two copies, as 21 Bill Schmidt mentioned, were sent to the ACRS also. 22 And within 30 days, we are supposed to review the 23 document and pass our judgment on the document, and it will be 24 issued in the form of a generic SER. This will be the first O 25 formal staff endorsement of the SQUG approach for the A-46 HERITAGE REPORTING CORPORATION -- (202)628-4888
237 1 review. 2 MR. MICHELSON: That will apply only to old plants? { 3 You are not endorsing it in any way to be used in future 4 plants in lieu of any kind of seismic qualification work? 5 MR. ANDERSON: Not at this time. 6 MR. MICHELSON: If it will be decided to do that, 7 would that have to be a rule? 8 MR. ANDERSON: I don't know. 9 MR. MICHELSON: You would officially announce it and 10 you would discuss it then? 11 MR. ANDERSON: Yes. 12 MR. MICHELSON: Right now this is only the old 13 plants that have inadequate documentation? 14 MR. ANDERSON: Yes, who haven't met current 15 licensing requirements. 16 MR. CHANG: We define very clearly the A-46 program 17 only applies to the operating plants, not for the future 18 plants. 19 MR. MICHELSON: Thank you. 20 MR. CHANG: And once the utilities receive the 21 generic SER, they will try to come up with a plant-specific 22 implementation schedule. That is, within 60 days they have to 23 get back to us, let us know about their specific, 24 plant-specific implementation schedule. O 25 Now the reason that we would like to have this HERITAGE REPORTING CORPORATION -- (202)628-4888
238 1 information is that originally SQUG would like to have (} 2 everything finished, then come back to us and tell us about 3 the implementation schedule, but we staff think that this is 4 going to delay the implementation procedure because the 5 walk-down usually will take place during the reactor outage, 6 and there are only so many outages every year and there are so 7 many other activities going on. If you don't get, try to 8 block out a time for the A-40 walk-down at early time, then it 9 is going to be delayed, so we talk to SQUG and we both agree 10 that at this point, there is enough information in the GIP for 11 them to--and endorsement of those will be enough for them to 12 come up with implementation schedule even though there are 13 still some activities going on and those issues will be 14 addressed later on by revisions of the GERS and staff in turn 15 will try to issue the supplement to the SER to address our. 16 review on those issues. So those will come out lator. 17 Revision 1 will come to us in September, and within 30 days we 18 are going to write an SER on that, and the finnl GIF with all 19 the issues closed out will be sent to us February next year, 20 and within 40 days, we are going to issue our, the NRC final 21 generic SER. 22 Utility then will conduct training seminars for the 23 member, members of the SQUG. The present schedule calls for 24 conducting those seminars in the second and third quarters of 25 1989, so the plant-specific implementation of A-46 for the HERITAGE REPORTING CORPORATION -- (202)628-4888
239 1 SQUG members will start in.the fourth quarter of '89.
/^Y 2 (Slide) ?%)
3 MR. CHANG: Here is a tabulation of the information 4 we received from SQUG as to how many plants will' conduct their 5 walk-down in what timeframe. As you can see, the majority, 6 the members in.the parentheses signifies the total number of 7 plants going to be walked down during that year. I~think mere 8 than half of the SQUG plants will be walked down during the 9 year 1990. 10 (Slide) 11 MR. CHANG: Here is a slide that summarizes the 12 ongoing activities, the SQUG ongoing activities. The NRC 13 staff and senior seismic review advisor panel have been 14 following those activities very closely. We have been , 15 participating in the meetings on a regular basis with the SQUG , 16 and their centractors, as far as the EPRI and EPRI 17 contractors. 18 On the issue or anchorage guidelines, basically l ' i l 19 there is agreement by all the parties on the criteria to be l 20 used. EPRI and its contractor is presently working on the 21 criteria for the review of tanks and heat exchangers 22 anchorage. That will include also for thin walled tanks the 23 structure shell wall integrity issue also. 24 MR. MICHELSON: Why are you worried about the shell
' 25 wall unless it causes the piece of equipment to fall? !
HERITAGd REPORTING CORPORATION -- (202)628-4888
240 1 MR. ANDERSON: -Hell, the issue of tank integrity {} 2 came up-in USI A-40, and we got SQUG to agree to take on that 3 part of A-40 issue within A-46, which.means that we are 4 concerned about not only the~ tank anchorage, but_we are 5 concerned about some of the older tanks that may have been 6 designed with the old rigid wall criteria and may not have 7 enough wall strength. 8 MR. MICHELSON: If it doeF fail, when they do the 9 walk-down, if they determine that the tank wall will fail, 10 what do they do with the contente of the tank? Do they 11 analyze what happens to the contents? 12 MR. ANDERSON: You take steps to assure that it 13 doesn't fail. 14 MR. MICH5LSCN: Is that what they do? 15 MR. ANDERSON: Yes. 16 MR. MICHELSON: Not because of water that might be 17 in the tank, but just because of arbitrary rule that says 18 tanks can't fail? 19 MR. ANDERSON: I suppose you could take the approach 20 if it failed-- 21 MR. MICHELSON: If I understand, understood the 22 criteria correctly, and unless it is, has been changed, right 21 Low if you find a tank can topple because of weak anchorages 24 or whatever, and if it doesn't impact directly a piece of O 25 safety-related equipment that is needed, then you do not have 4 HERITAGE REPORTING CORPORATION -- (202)628-4888
n-D s 241 1 to chase the contents, the water, to see if the water impacts
/^ 2 the safety-related equipment. Has tilat changed any?
b) 3 MR. ANDERSON: That I think is in the process of 4 changing right now. I recognize that point. We have debated 5 this. 6 MR. MICHELSON: Went over it many times.
-7 MR. ANDERSON: Been over it many, many times. I 8 know the question. I think you know my answer, but--
9 MR. MICHSLSON: There is no regulatory requirement 10 to chase it as I understand it? 11 MR. ANDERSON: That's right. The flooding review 12 under A-17 is tne partial answer to it, and also the SQUG has 13 a subcommittee that is looking at the incorporation of_other 14 issues, and one of the thingn-- 15 MR. MICHELSON: A-17 is not including er.ternal 16 avents? 17 MR. ANDERSON: I know, be.t they are looking at water 18 sou'ccJ and flooding. 19 MR. MICHELSON: Hut not as a seismic consequence? 20 MR. ANDERSON: No. Well, from any consequence, 21 whether it be seismic or whether somebody drives a truck into 22 it. 23 MR. SCHMIDT: Evaluation of tanks in the SQUG 24 progress is more tanks which are in the scope of the A-46, 25 that is, they are in safe shutdown systems. It doesn't HERITAGE REPORTING CORPORATION -- (202)628-4888
-. - _ . _ - _ _ _ ~ --
242
~
1 address the interaction concern that you mentioned. 2 MR. MICHELSON: You are going to look at the non, or 3 other tanks that could fall, interact, and this, I thought 4 that was included. 5 MR. ANDERSON: From seismic-systems interaction 6 basis, that is correct. ! 7 MR. MICHELSON: I thought that was included. 8 MR. SCHMIDT: Oply as it affects physical 9 interaction, but not flooding. 10 MR. MICHELSON: Right. If the tank were to fall on 11 the pieces of equipment you needed, then-- 12 MR. CHANG: You want to make sure that doesn't 13 happen. O 14 MR. MICHELSON: Wouldn't have to worry about the 15 water that was in the tank? 16 MR. SCHMIDT: Right. These criteria that Dr. Chang 17 is referring to are actually criteria to evaluate the shell 18 for buckling of those tanks which are needed as part of a safe 19 shutdown train. 20 MR. MICHELSON: These are part of the essential 21 equipment you want to make sure-- 22 MR. CH4NG: That's right. 23 MR. MICHELSON: Okay. 24 HR. CHANG: Only for those tanks that are needed for 25 the hot shutdown. i l HERITAGE REPORTING CORPORATION -- (202)628-4888
h 243 1 MR. MICHELSON: Okay. 2 MR. CHANG: On the GERS--generic equipment [ 3 ruggedness spectrum--EPRI is going to issue a supplemental 4 report pretty soon. There was a report issued by EPRI last 5 year. I think it is around May time, timeframe. There are
'6 still a number of GERS issues going to be resolved.
7 MR. MICHELSON: I don't recall--Charlie, did we ever 8 discuss GERS, G-E-R-S? I 9 CHAIRMAN WYLIE: I don't'think so. 10 MR. CHANG: Those are the spectra developed from the 11 test data. 12 MR. MICHELSON: This is just developing a seismic 13 spectra to b2 used? 14 MR. CHANG: This is sort of like a seismic capacity 15 of equipment. The spectra signifies the equipment capacity, 16 earthquake capacity based on the test data. 17 MR. MICHELSON: Used during tha walk-down? I 18 MR. ANDERSON: That's right. 19 MR. SCHMIDT: Generic qualification. 20 MR. MICHELSON: I didn't recall. j 21 MR. ANDERSON: I thought we had this discussion when l l 22 rou were here. 23 DR. SIESS: Spell it out. 24 MR. CHANG: Spell it out--generic equipment O 25 ruggedness spectra. HERITAGE REPORTING CORPORATION -- (202)628-4888
- - , - . _ _ - . _ _ _ _ _ _ _ _ _ . . , _ . . , _ _ . . , . . _ . . _ . _ . - - _ _ , . _ . _ . . ~ . _ . _ _ _ _ . . _ . . - - . _ . _
244 1 DR. SIESS: Okay. Now I remember.
~
2 MR. CHANG: We used two, we used data base from two 3 sources. One is the earthquake data base, and the other one 4 is the test data base. This is related to the test data base. 5 DR. SIESS: That's the EPRI? 6 MR. CHANG: That's the EPRI study. 7 DR. SIESS: ,Taking all the data base and trying to say'-- *
~ '
,,,8 . . . ,
9 ,, MR. CHA,NG: Sort of like envelope of-the test data, 10 DR. SIESS: I dor.'t know whether the Committee'has 11 ever looked at that. I haven't. 12 MR. MICHELSON: I don't recall, but that doesn't 13 prove much. O 14 MR. CHANG: The relay evaluation procedures, there 15 is a concern that those seismic testings mostly are done in , 16 middle of '70s or so, and their relays are probably older than 17 that timeframe in the older plants, so in order to address the 18 relays in those older plants, that people don't have the test 19 data, EPRI started a test, testing program on those older 20 relays. 21 MR. MICHELSON: You are only using tests that l 22 actually monitor each and every contact with chatter of the f i 23 appropriate frequency? 24 HR. CHANG: Basically. 25 MR. MICHELSON: Some people qualified the stuff l HERITAGE REPORTING CORPORATION -- (202)6'.8-4888
245 1 before and after the test. They tested before and after and 2 they shook.it to make sure it didn't mechanically come apart, 3 both open, closed and solenoid both energized, de-energized? 4 MR. CHANG: Yes--all configurations. 5 MR. SCHMIDT: They have different ruggedness levels. 6 MR. MICHELSON: It was a pretty good test. 7 CHAIRMAN WYLIE: How many relays of the test was 8 involved?.. Do you know? . 9 MR. CHANG: ,There are two series of, well, the first 10 series contains about 24 relays, and there are 24 more to be 11 tested, so all together there will be 48 relays to be tested. 12 CHAIRMAN WYLIE: These are all types of relays? la MR. CHANG: All types. 14 CHAIRMAN WYLIE: Protected relays? 15 MR. SCHMIDT: They are all types, and we 16 concentrated on trying to get these older relays from the 17 population of relays that are most widely used, and we were 18 able to do that to the extent that utilities could find these 19 old relays in their storerooms and were willing to donate them 20 for testing. 21 CHAIRMAN WYLIE: I see. j 22 MR. SCHMIDT: These are 1960s vintage relays. 23 CHAIRMAN WYLIE: These represent protective and 24 industrial type? The whole spectrum? O 25 MR. CHANG: Whole spectrum. HERITAGE REPORTING CORPORATION -- (202)628-4888
2E 246 1- CHAIRMAN WYLIE: Just a matter of information, what 2 is the lowest seismic fragility level that you found in the 3 relay chatter? 4 MR. SCHMIDT: The bulk of the relays are in the 5' range of about four to five G spectral acceleration and above.
.6- There are a few protective and more fragile types that are 7 down in the one, as I recall, one to one and a half G spectral 8 acceleration range, and those relgys are, . pretty sure if they 9 are in central circuits, will be a problem in most of these 10 plants. The peak of the--
11 DR. SIESS: What frequency does that come in? 12 MR. CHANG: I think that is within 4 to 16 hertz. 13 That's the tost, IEEE test for the plateau, between 4 and 16 O 14 hertz peak, 4 to 16 hertz. l l 15 DR. SIESS: Relays start in the range of 4 to 16? l 16 MR. ANDERSON: Somewhere in that range. l 17 MR. CHANG: Plateau stems from 4 to 16 hertz. 18 DR. SIESS: I thought when we got high frequency 19 reports at Surry people said that might be important with 20 respect to relays, but that was up in the 22, 23 range. You 21 are saying the relays don't chatter after that? 22 MR. CHANG: We don't say that, but what I am saying 23 is that the requirement by IEEE is to test it with the plateau 24 from 4 to 16 hertz. 25 MR. MICHELSON: Inference is they don't chatter in HERITAGE REPORTING CORPORATION -- (202)628-4888
y __ 247
'l that range,.they don't chatter outside that range, either. Is 2 that the' inference or--
V 3 MR. SCHMIDT: The input levels are. lower as you get 4 out at higher frequency, but what Dr. Chang is saying is that 5 the test table spectra shape has most of its inputs'in the up 6 to 20 hertz, and when they chatter, that defines a level-of 7 that curve which is a qualification level, and then that's 8 , what is published as a qualification' curve. 9 DR. SIESS: Those members of the staff who are not 10 concerned about high frequency earthquake input because of 11 relay chatter were wrong? 12 MR. CHANG: We have a concern for relays that 13 . chatter in the high frequency range. As a matter of fact, we O V 14 have collected a number of LERs that has the bumping of 15 cabinets that causes chatter of relay and those bumping 16 normally were, create high frequency input to the relay. 17 DR. SIESS: I don't understand why the IEEE spectrum 18 is limited to 4 to 16 hertz. There has been a couple of 19 earthquakes that have shaken a couple of existing power plants 20 above the OPE level, up in the high frequency range, which 21 doesn't bother me a bit, but somebody says gee, but that's up 22 in the range where the relays will chatter. 23 MR. ANDERSON: Well, the first place, a lot of types 24 we are looking at axisting test information on these relays, p (J 25 and this is the range they were tested in. HERITAGE REPORTING CORPORATION -- (202)628-4888
- 248 1- Now we 'do have a research program at Brookhaven on-2- 'their fragility program where they are doing single frequency
)
3 testing of relays and discrete frequencies. They raise the 4 amplitude until it chatters. 'Then they chart. 5 DR. SIESS: This is historical? 6 MR. ANDERSON: This is for the most part historic, 7 but the results of the Brookhaven study show that that then 8 generally have a peak on the frequency tests. Some are maybe 9 around 6 to 8, cnd some other will have peaks as high as maybe 10 16 hertz, but it seems to me in that range where the chatter 11 really occurs. 12 CHAIRMAN WYLIE: These are new modern relays? New, 13 modern? 14 MR. ANDERSON: Yes. 15 DR. SIESS: Really modern? 16 MR. SCHMIDT: We also do have open action item to go 17 back and address specific relays that are-known to be 18 sensitive to high frequency, and a list of those is being 19 generated based on records of relays that have chattered 20 because of these impacts of cabinets and so forth. And we do 21 plan to go back and look at those and make sure that, that we 22 have somehow got them covered. We haven't quite figured out 23 how to do this, but we don't think there are very many. We j 24 -think they are mainly protective. n 25 MR. MICHELSON: What is the shortest duration of l HERITAGE REPORTING CORPORATION -- (202)628-4888 l
, l' 249 1 chatter?
2 MR, SCHMIDT: That's an interesting subject. The
'{
3 test data is all based on 2 milliseconds chatter because 4 that's the industry standard. 5 MR. CHANG: This is something I am going to go into 6 'also. 7 HR. MICHELSON: Okay. , 8 MR. CHANG: In the relay testing-- 9 MR. MICHELSON: Before you move that, that slide 10 confused me. I thought these were tests yot* were doing and 11 you needed some more relays. One of your bullets. the other 12 one is you completed a series of tests. I thought this was 13 all going on now somehow. And what you are saying is you just 14 are digging up the old data? 15 MR. CHANG: The testing of those old relays are 16 going on right now. 17 HR. MICHELSON: you are doing the testing now. This 18 is not old test data. This is new test data? 19 MR. CHANG: Because there was no old test data on I u 20 those old relays. 21 MR. SCHMIDT: Historical test data base, primarily 22 newer model relays; in order to try to answer a question of do I 23 the old model relays perform the same way, there are series of 24 what I would call confirmatory tests tha; will duplicate the O 25 historical tests and are intended just to demonstrate whether I i HERITAGE REPORTING CORPORATION -- (202)628-4888 ; l r - - - , . - - - - , _ - , _ _ . _ _ , _ , , , , _ , ,,m. . _ _ . . _ _ , _ , _ _ _ _ _ _ _ _ _ , _ , __ ._ ,,
250 1 the1old-relays perform'as well. That's what those-- 2' MR. CHANG: That's the main purpose. Another issue f~}
-3 is related to.the in cabinet amplification because the relays 4 -are going to'see, amplify input compared to false input and t 5 for some. cabinets and the location of relays, it could see 6 quite high seismic input, so as one way to reso1Ve that, SQUG 7 proposes to maybe we can look into the relay chatter criteria.
8 In che IEEE requirement, it says if it chatters 9 longer than two milliseconds, that constitutes a chatter. In 10' other words, that the device that is controlled by this relay 11 downstream is going to malfuncton. This is the assumption. 12 Now SQUG proposes that maybe we can relax the 13 criteria somewhat. They propose to attach the relay and see 14 what happens, if the relay chatters in the range of 10 to 15 15 milliseconds, whether this is really going to affect function 16 of the devices controlled by the relay or not. 17 MR. MICHELSON: Is the assumption always that there 18 is another relay downstream, or you might have-- 19 MR. CHANG: Could be a number of things downstream. 20 MR. ANDERSON: Depends on the circuit, but what you 21 have to do is reproduce the circuits. If there are other 22 relays they have to object-- 23 MR. MICHELSON: If it is an electronic equipment 24 downstream, that has been being actuated, you are talking 500 0 25 microseconds. HERITAGE REPORTING CORPORATION -- (202)628-4888
- m. ;
251 l' ~MR. SCHMIDT: What-this-is testing-- I - 2 MR. MICHELSON: This is'only for relays. 3 MR. SCHMIDT: 'It is only for-electromechanical 4 devices like solenoids and other relays. 5 MR. CHANG: Contactors. 6 HR. SCHMIDT: Switchgear and contactors'that have a 7 substantial inertia,~and the preliminary tests we have run 8 indicate that relay-chatter in the 10 or 15 second timeframe 9 does not cause those to actuate, which is not surprising. 10 MR. MICHELSON: It will cause another relay. 11 MR. SCHMIDT: No. We have run preliminary tests 12 with HFA, HGA typv relays on the table, and we can't make
. g, 13 those drop out in 10 to 15 milliseconds, either.
b 14 MR. MICHELSON: Used to say it was two milliseconds 15 dropouts, that really weren't right. 16 MR. SCHMIDT: I think it was a very conservative 17 number. 18 MR. MICHELSON: I remember in those days we worried 19 about relays sequence or whatever, and 2 milliseconds was the 20 number we were using. That is what GE said it would be. 21 MR. ANDERSON: I think based on the qualification 22 tests, that-- 23 MR. SCHMIbT: The tests we are running will 24 obviously be limited to those kind of devices which we test, 25 but so far, they are showing that we can tolerate 10 or 15 HERITAGE REPORTING CORPORATION -- (202)628-4888
q 252
'l milliseconds chatter.
. (') ; '2 MR. MICHELSON: Before they actua3--
wr 3- MR. CHANG: Also preliminary results indicate that 4 once you allowed the chatter to go to 10, .0 and 15 5_ milliseconds, the ruggedness of the relay can be increased a 15 lot up to maybe two times the seismic level that, that it can 7 take if you name a 2 millisecond criteria. 8 'On cable tray and conduit criteria-- 9 MR. MICHELSON: Let me.ask on'the relays, do you 10 include the contact as the relay? 11 MR. SCHMIDT: All of the, all of the switching 12 devices that show up in the circuits get evaluated. Those 13 which are electromagnetic or solid state devices get evaluated 14 as part of relay evaluation procedure. The other contacts 15 that show up that are basically things like manual switches, 16 limit switches on motor-operated valves which are mechanically 17 driven, we have already concluded that those, those are 18 inherertly rugged. That is, they are not prone to change of 19 state, and they are reviewed, but they are, concluded that 20 they are rugged, so they are accepted as is. 21 MR. MICHELSON: Solid state stuff-- 22 MR. SCHMIDT: Solid state stuff is being evaluated 23 separately, but the switchgear, contactors, relays, if they 24 are in circuits that are essential, they will get evaluated O 25 either with individual relay capacity data, GERS, or as part HERITAGE REPORTING CORPORATION -- (202)628-4888
y 253 1 of-the' equipment GERS for the things like motor controlled or A- 2 switchgear. O-L3 MR. MICHELSON: You say you don't get'an actuation 4 10 to 15 milliseconds? You do not close the contacts? 5 HR. SCHMIDT: That is correct. Or open the 6 switchgear, right. 7 HR. MICHELSON: That is true when you energize and 8 de-energize? , 9 MR. SCHMIDT: We will find out. Preliminary test's 10 have shown it works. The tests are scheduled for the second 11 half of-the year. 12 MR. CHANG: And SQUG is also working on the cable 13 tray conduit review criteria. At present, criteria for the 14 cantilever bracket support types of relay, I mean cable trays 15 have been completed, and reviewed and sort of agreed upon by 16 the staff and the SSRAP members and they are continuing 17 working on the other types of cable trays such as trapese 18 frame type and raw end type. 1 19 DR. SIESS: Are the criteria expressed in terms of 20 strength or stiffness? 21 MR. CHANG: Thee are two aspects in the review; one 22 is quantitutive. The other one ic qualitative. So there are l 23 a number of caveats they have to review, and try to satisfy. 24 That's qualitative based on the data base collected. O -25 For instance, the span between the two is supports HERITAGE REPORTING COREJRATION -- (202)628-4888 i
. . - - - - , - , , . . , , , - - , - , , - - , , - - - , - - - - , - , -.-n,-,n-., -=~,--,--,,-,n - - , , - --,--,n, n w- . n,.,, e - , , , - .n-,,--
'l 254 l 1
1- cannot be longer than ten feet and so forth.- Qualitatively, , I l ( - 2 they have to do a number of quasi-static check in the vertical 3 direction and-in the horizontal direction to make sure that I 4 the supports are adequate. . 5 DR. SIESS: In terms of what, stress or deformation? I 6 MR. CHANG: In terms of stress. 7 DR. SIESS: Cantilever uni-strut and something i 8 coming out here, two bolts back on this side? q r 9 MR. CHANG: That's right. i 10 DR. SIESS: How do they fail? The whole thing 11 coming out of the wall or just one cantilever? 12 MR. CHANG: Usually the deformation is local, one i [ 13 possibly at the anchorage, if you have expansion angle bolts { 14 and organize the support to the ceiling, and that's a weak [ 15 link there, f 16 DR. SIESS: This thing is supported on the floor? 17 MR. CHANG: No, but it is hanging down from the 18 ceiling, i i 19 MR. ANDERSON: Hanging down from the ceiling with a l i L 20 cross-arm. I t 21 MR. SCHMIDT: Hight be atrut based or not. [ f 22 DR. SIESS: I am still--did you give me one or two i 23 examples of how they fail? f 24 MR. ANDERSON: We don't have very many examples of i 25 failing. I HERITAGE REPORTING CORPORATION -- (202)628-4988 l
255
~
l' DR. SIESS:- How do you develop criteria against 2 failure? 3 MR. ANDERSON: The bolts pull out or they fail; 4 local deformation and buckling. 5 MR. SCHMIDT: We-are looking for the weak links. 6 DR. SIESS: They can pull out the bolts this way? 7 MR. ANDERSON: That is one. 8- MR. CHANG: There is no real cases of failure in a 9 data base. It is hard to-- i 10 DR. SIESS: Somebody must have tested them. 11 CHAIRMAN WYLIE: You put a force on them or 12 something and then if it passes that force without failing, 13 they are good? O 14 MR. SCHMIDT: It is a T. The actual experience data 15 provides very little information on actual failures so what is 16 being done is some straightforward quantitative checks are 17 being put together to check for weak links in the system for 18 both vertical and dead weight and seismic loads, and they are 19 protecting you against such things as for a cantilever bracket . 20 with strut support, a lead which would, for example, put a 21 load on the side load on the tray and pull the anchor bolts 22 out of the ceiling. , 23 Another mode would be for net section collapse of 24 the vertical member right at the strut support. Those are the ( 25 kind of analytical checks that are being made to assure that HERITAGE REPORTING CORPORATION -- (202)628-4888
!~ 256 1 those kinds of simple failures won't occur. 2 DR. SIESS: Designed essentially for 1G vertical?. f'I J 3 MR. SCHMIDT: No, they are not. We don't know what 4 they are designed for. 5 DR. SIESS: I'm sorry, No earthquake means 1G. 6 MR. SCHMIDT: 1G is dead weight. 7 7 DR. SIESS: That's what I mean. That is presumably l 8 what they were designed for. 9 MR. SCHMIDT: Presumably designed for a lot more 10 than that. 11 DR. SIESS: Oh. l 12 MR. CHANG: I think it varies from plant to plant. 13 There are a lot of different vintage plants. 14 .MR. SCHMIDT: The simple analytical collection, give 15 you an example, typically checking these devices for loads l 16 like twice dead weight 2G and like 3G. i 17 MR. CHANG: Vertically. 18 MR. SCHMIDT: As conservative static checks, that 19 will give you good assurance that they won't come down. 20 CHAIRMAN WYLIE: The criteria, Chet, used for design 21 in cable trays was the design for the weights of cable plus a 22 whole crew up there pulling cable in the trays, walking the 23 trays. 24 DR. SIESS: And with some argin? 25 CHAIRMAN WYLIE: With margin, yes. HERITAGE REPORTING CORPORATION -- (202)628-4888
257 1 DR. SIESS: So now what--are we looking for the same 2 margin? 3 MR. CHANG: The purpose of the check is to make sure 4 that the cable-tray configuration is covered by the data base. 5 This is the purpose. -They looked at the' worst cases in the 6 data base and they found by using this check-- 7 DR. SIESS: I thought you were' giving me some , 8 crit eria, some quantitative strength criteria. You are just 9 going back and comparing with the data base I understand.
~
10 MR. CHANG: The purpose of that is to make sure it 11 is covered by the data base. 12 MR. SCHMIDT: There is a physical comparison with 13 the data base. Are they physi: ally like the ones in the data 14 base? Do they have parameters tant fall within the bounds of 15 the data base? We have said that's not enough. In addition 16 to that, we will go, on a selected number of them, we will 17 also numerically quantitatively check them to make sure that 18 there is ample margin on dead weight and horizontal load. 19 DR. SIESS: Your data base, as I recall, was 20 predominantly made up of equipment that wasn't designed for ; 21 any seismic. 22 MR. ANDERSON: That ir correct. l 23 11R . SCHMIDT: Correct. 24 MR. ANDERSON: And using this, these analytical ( 35 checks on a number of cases in the experience data base, we i HERITAGE REPORTING CORPORATION -- (202)628-4888 .
258 1 have predicted failure with these checks, and that the trays 2 didn't fail, so we felt that using this check--
)
3 DR. SIESS: Okay. And you really are expecting to 4 ' find things that couldn't take twice the' dead load? 5 MR. ANDERSON: No. We don't expect to find any, but 6 that's a check to ensure that we do have that type of, kind of 7 capacity in the tray system. 8 DR. SIESS: .You check everything in the plant? < 9 MR. CHANG: No. They look at the. worst case, the 10 most heavily loaded. 11 DR. SIESS: Forget about it. I just can't + 12 understand all this checking of loads and stresses on cable ; 13 trays when there are a lot of other structural components in ( ' 14 the plant that were designed the same time, by the same people 15 and the same earthquake, concrete wa'lls and steel columns. 16 MR. ANDERSON: And it is just another check because , 17 there are a lot of people, a lot of our reviewers were a 18 little bit uneasy because we are, we have the application of 19 experienced date. base to the nuclear plant, and this is just 20 additional check. 21 DR. SISSS: I guess so. If you have never seen one 22 fail and don't know how it fails, I guess maybe you get-- l 23 MR. ANDERSON: You have to look pretty hard to find 24 anything that resembles the failure in data base. O 25 MR. MICHELSON: The water coming out of non-safety HERITAGE REPORTING CORPORATION -- (202)(28-4888
259 1 tank.that fallsnon the equipment or on the. floor rather.
- . 2 Don't worry about that water?
3 DR. SIESS: Cable trays in plants, try to figure out what happens if one of them fails and it is, I think it is 5 going to hang up there on the cables. 6 MR. ANDERSON: That has happened. 7 DR. SIESS: You know - 8 CHAIRMAN WYLIE: You are probably right. 9- DR. SIESS: We are, I keep worrying that we are 10 worrying about the wrong things. 11 MR. MICHELSON: Why don't we put it on something 12 perhaps a little more interesting? 13 MR. SCHMIDT: There are a large number of people who 14 share your cpinion. 15 DR. SIESS: There are lots of other things I worry 16 about in plants besides cable trays. 17 MR. ANDERSON: There are a cadre of people who think 18 you have to do a three-dimensional finite analysis on cable 19 trays. 20 DR. SIESS: Those are the ones I worry about! The 21 guys that are doing three-dimensional pipe and element t 22 analyses on cable trays, yes, or pipe restraints or pipe 23 supports for that matter. That's what worries me, instead of 24 somebody out there making a quick check on statics and I U 25 calculations. l l HERITAGE REPORTIHG CORPORATION -- (202)628-4888
4-260 1 MR.~ ANDERSON: Economically it is very lucrative for, J , 2~ analysts.
.3 MR. CHANG: The last ongoing item is the seismic 4 study. EPRI tried to collect actual earthquake data from 5 nuclear plants basically in Japan and also commercial. building 6 seismic data in this country, and trying to-see whether they 7 can generate means to estimate floor spectra for some plants 8 .because for some older plants that information is simply not 9 there. The other plants, the information may be too 10 conservative.
11- CHAIRMAN WYLIE: Back on your relay evaluation, are 12 you going to publish this data that you find after you do all- > 13 this testing and everything for the record somewhere? ( MR. SCHMIDT: It is, much of it has been published l 14 - 15 and it is being published under EPRI reports, so it will be 16 subject to the same distribution rules as any other EPRI
- 17 document.
18 CHAIRMAN WYLIE: That's good. Brookhaven work come 19 out as NUREG? 4 , 20 MR. ANDERSON: As NUREG, yes. 21 CHAIRMAN WYLIE: How wide scope is Brookhaven? L 22 MR. ANDERSON: It has been going on for a couple of i 23 years, and couple of times a year they publish a NUREG on it. ! 24 The program is continuing and we have been successful in ! O 25 getting input into that. They are helping us work the HERITAGE REPORTING CORPORATION -- (202)628-4888
261 1 problem. (~)) 2 CHAIRMAN WYLIE: That is interesting because when we 3 reviewed GSAR Brookhaven did the PRA work, and they found that 4 relay chatter was the most contributor to core damage in that 5 PRA, and they were using relays fragility of .06G. I don't 6 know where they ever came up with that. - 7 MR. ANDERSON: I don't, either, but the GSAR PRA was 8 done some time ago. 9 CHAIRMAN WYLIE: But the .06G never existed I don't 10 think. 11 MR. ANDERSON: The lowest one we have ever seen I 12 think had a ZPA of .2.
,_s 13 MR. CHANG: That number is the spectra acceleration ks 14 or the ZPA value.
15 CHAIRMAN WYL12- I ara not sure. I asked them where 16 they got the number. They say they picked it up out of U.S. 17 Army Corps of Engineers' study someplace. It had no 18 relationship to power plants, but that's the reason I was 19 curious about the publishing of this data, because I think it 20 really would be valuable in the PRAs. 1 21 MR. ANDERSON: I think most of the recent PRAs have 22 taken a much more realistic view of the relay chatter. 23 MR. MICHELSON: Relay chatter is actually in there 24 as a variable or just as a function versus no function? 25 MR. ANDERSON: I understand that on the more recent HERITAGE REPORTING CORPORATION -- (202)628-4888
262 -
~1 PRAs, trat they look at it first from a standpoint of whether i 2 or not it will chatter, it has a fragility level for-chatter,
{ 3 and they do look at the circuits to determine the' impact of 4 the chatter, but I don't think that has been done on very many 5 of the PRAs. Mike Vaughn, who was here before, has done a lot 6 of that work. i 7 MR.-MICHELSON: I will chat with him about that. 8 Thank you. 9 MR. CHANG: Okay. The second topic is under Nine 10 Mile plant walk-down. I want-to go through this quickly. ' 11 This is the pilot plant walk-down for a BWR plant, and it took 12 place at the end of January this year to early February. 13 The main objective of this walk-down, which is not () 14 listed here, is to try to test if' people try to use utility t 15 engineers to do the walk-down, whet her it is feasible or not. 16 The Zion walk-down was performed by consulting engineers. l l 17 MR. MICl!ELSON: When you do your walk-down, there is [ ] 18 a procedure you talked about a little earlier, the guide book. 19 What does that guide book say about looking at fire 20 protection when you are walking down to make sure that it is ; 21 of such a nature that it is not going to inadvertently actuate 22 during a seismic event, spray the equipment that is not 23 qualified, things of that sort? 24 MR. ANDERSON: No, it doesn't look at that, unless there are piping in the vicinity. I 25 u HERITAGE REPORTING CORPORATION -- (202)628-4888 2 _, , - . _ . _ , , . . _ . . . . _ _ _ _ _ . _ . _ _ . _ ~ _ _ _ _ _ , _ , _ _ _ _ _ , . . - -
263 - 1 MR. .MICHELSON: Look at pipe to make sure it will 2 not break open and spill its contents; that is a GDC ( 3 requirement. Of course, fire protection ought not to fail and ,. 4 affect any safety-related function. 5 MR. ANDERSON: Well, it is not part of this program. 6 MR. MICHELSON: It isn't. What program is that one? 7 MR. ANDERSON: This has fire protection systems. 8 MR. MICHELSON: It is not in fire protection, no. 9 DR. SIESS: Didn't they have to do seismic 10 interaction? Weren't there seismic interaction walk-downs? 11 MR. ANDERSON: That's right. Seismic interaction is l 12 included as part of the A-46 review. 13 MR. MICHELSON: If fire protection pipe falls down 14 and hits a piece of equipment, then it is-- 15 MR. ANDERSON: Or in the vicinity; can't interfere. 16 DR. SIESS: If it falls down and hits a piece of i l 17 equipment, you will get high frequencies, too. l 18 MR. MICHELSON: Hitting the floor even will 19 generate--you don't have to hit the equipment to get the high ( 20 frequencies, either. You hit the floor. 21 DR. SIESS: The Germans were interested in the 22 frequencies developed when an airplane hit the containment. 23 They were interested in relay chatter from that source. 24 MR. MICHELSON: Ever drop a load from one of those 25 big cranos just a wee bit and hit the floor with it? Do you HERITAGE REPORTING CORPORATION -- (202)628-4888
264' 1 hear'the whole-building? ; 2 MR. SCHMIDT: I could say one thing, that the l
.(
3 seismic interaction concerns that are being addressed under 4 the SQUG program are those which are being judged to be real, 5 real seismic interaction concerns, not theoretical. 6 MR. MICHELSON: If the fire protection is not 1 7 seismically qualified, it isn't a real problem unless.an 8' earthquake actuates the fire protection. That's what we have l 9 to find out. Then you have to tackle it somewhere; apparently 10 not in this walk-down. ; 11 MR. ANDERSON: A-46 does not do that. 12 HR. MICHELSON: We have got plenty of inadvertent 13 actuations that have taken out safety-related equipment. It ' 14 is a very real phenomena. LERs are about a, by the stack show 15 that. 16 MR. CHANG: The other purpose of this pilot-plant 17 walk-down is to test the generic implementation procedures to 18 determine what improvements could be made to the training 19 course and to the GIP, and as a side product, is to verify the 20 seismic adequacy of the equipment. 21 MR. MICHELSON: Let me pursue this relay chatter j 22 then from the viewpoint of non-qualified equipment that could 23 chatter and cause such t' *.ngs as fire protection actuation. ; i i i 24 You are not looking at that either then? See, they , () 25 found mercury 2 relays in one of the plants recently, i I HERITAGE REPORTING CORPORATION -- (202)628-4888
..-.. . . . , - . _ . - . - - . - . , . , - , . . - , - . - . . - - - . _ - - . ~ , , - . . - . - -
. . ~ . .
265 1 MR. ANDERSON:' I know they did. f(} 2 MR. MICHELSON: I was hoping this was going to 3 ferret out the last of the mercury 2 relays. 4 MR. ANDERSON: That was mercury relay that 5 interfered with the diesel operation, and when we saw that 6 LER, we called up-one of Bill Schmidt's people at MPR doing
'7 the reviews, asked him pointedly if the relay review that we
-8 are doing on A-46 would have~ caught that, and they considered 9 it for a while and said indeed it would.
10- MR. MICHELSON: Why would it?
- 11 MR. ANDERSON: Because they trace back, when they 12 review the diesel, they look at all connections to it, all the 13 connections to the services and trace back the contacts. They 14 would have found that contact and mercury switch out.
15 MR. MICHELSON: If they happen to be using mercury 16 switches in the CARDOC fire protection system, they wouldn't 17 catch it because it is not direct wire. 18 MR. ANDERSON: In that connection, if that had not 19 been connected, that contact directly to the diesel support 20 system, they would not. 21 MR. MICHELSON: Picked up as interface to the diesel 22 system. Don't pick up spatial interfaces. That's all? 23 MR. ANDERSON: That's right, i 24 MR. MICHELSON: That's what it amounts to. 25 MR. CHANG: There is a lot of preparation work to be 4 HERITAGE REPORTING CORPORATION -- (202)628-4888 !
266 1 done before the walk-down. -I don't want.to go into details. 2 There are, three seismic review-teams were involved in the 3 walk-down. Each team has four members, three seismic , 4 engineers and one systems engineer, and on each team there is 5 a person with the professional. license, and as I mentioned 6 ' earlier, all the .aembers were formed by utility engineers. 7 Now this, those facility engineers are not 8- necessarily supplied by only Niagara Mohawk people. Some of 9 the SQUG plants, they provided their own-utility engineers to 10 do the walk-down. : 11 DR. SIESS: What constitutes a seismic capability 12 engineer? 13 MR. CHANG: There are a number of requirements 14 listed in the GIP. They have to have at least five years of 15 experience. The person has to be a degreed engineer. He has 16 to have experience in seismic, either seismic qualification or ; 17 seismic design and so forth. 18 MR. SCHMIDT: Has to have completed the SQUG 19 training course. 20 MR. MICHELSON: Again, this is, is it safe to say 21 that really what A-46 looked at in its resolution was 22 equipment that was supposed to be seismically qualified but t t t 23 for one reason or another either the documente. tion lost or 24 wasn't--but i t doer.. ' t look at non-safet; equipment that maybe
) 25 should have been seismically qualified?
J i ! HERITAGE RESORTING CORPORATION -- (202)628-4888 ,
267 1 MR. ANDERSON: I don't t hint. that is strictly
(~') 2 correct. A-46 was developed as kind ot' alternative for the v
3 utilities to meet what we consider to be acceptable seismic 4 qualification for their equipment. Now we said that they did 5 not, they don't, did not need to have all safety equipment 6 within the scope, but they need to shut down the plant. Okay.
7 We did not say that it had to apply to all seismic class one 8 equipment or everything defined in Reg Guide 129.
9 MR. MICHELSON: You did say it applied to non-safety 10 equipment if it could physically interact?
11 MR. ANDERSON: For the systems interaction part, but 12 for the basic equipment, that we want to protect seismic.
13 MR. MICHELSON: I think that's the part where I have
( )
20 MR. CHANG: No. The requirement that as a team, l 21 the colle:tive experience should have, should be such that it ;
22 has a coverage of all aspects. It may not be on a particular l 23 engineer, but as a group, they should have a collective ,
i 24 expertise on all aspects. {
l
]
( 25 MR. ANDERSON: One of the things the NRC was quite i 1
i HEF.ITAGE REPORTING CORPORATION -- (202)628-4888 .
_-...,.--__.~..,_m .,_,,... _ ,_ _____,,,_-.._ _ _m____-,____,__m, __.-,,-,,__,_m..,_-.. ._.-
269 1 concerned about is the qualifications of people to do this ;
2 review because a lot of this necessarily has to be judgment.
/}
3 when you look at something. Is it going to inte ract? Although they have the guidelines and caveat to follow, there 4 5 is a lot of judgment that needs to be exercised, and so we are 6 really concerned that we get people in there who really have 7 the capability of making those judgments. 8 As a matter of fact, SSRAP also has that same - 9 perspective that they, I think the observation was made by the i 10 SSRAP that at the Nine Mile point walk-down, they come up--the i 11 less the level of the experience, the more-outliers they come ; 12 up with. The more experienced the people are, the more, 13 better they are at resolving problems in the plant. If they 14 are not experienced, they want to come out and analyze them. l 15 HR. CHANG: The walk-down took all together about-16 two weeks. It took a long time to get the radiation training 17 and all that, but the actual walk-down was about six days, and i 18 I stayed with all those teams for the most part the duration, i 19 and the senior seismic review advisory panel, and other NRC 20 reviewers, they came into our end of the walk-down to do some 21 sort of audit of their work. i 22 (Slide) 23 HR. CHANG: Results of the walk-down, there are a ' l 4
- 24 175 pieces of equipment that was identified as the, as O 25 required for safe shutdown, and all those 175 pieces of
, i l HERITAGE REPORTING CORPORATION -- (202)628-4888
270 1 equipment were looked at during the walk-down.
.2 This is a list of breakdown of the category in terms 0('A -
3 of categories of equipment. Out of the 175 equipment, 145-4 passed walk-down review and 30 equipment were considered to be outliers, and I am going to talk about what is the definition of outliers on the next slide. 7 DR. SIESS: Especially when you get five out of 8 five. 9 MR. CHANG: Pardon? 10 MR. ANDERSON: Looked at five and all five were-- 11 MR. CHANG: All five; I think probably that's 12 related to the anchorage. 13 DR. SIESS: If you looked at five motor control 14 centers, they are were all deficient in anchorage, does that 15 suggest you looked at another five? 16 MR. ANDERSON: Outlier isn't really saying they are 17 deficient. That says that they were unable using the 18 screening tables and the guidelines that they had to make 19 judgments of their seismic adequacy in the field. 20 MR. CHANG: May simply mean-- 21 MR. MICHELSON: Have they done that yet, or is that 22 coming? 23 MR. ANDERSON: They haven't resolved the outliers 24 yet. I ' ( 25 DR. SIESS: I would be interested in looking at this HERITAGE REPORTING CORPORATION -- (202)628-4888
271
~1 slide as to whether.those. tive motor control centers are:all 2 in one category or whether they-are spread among these six }}
- 3. categories.
4 HR. CHANG: I don't have that specific information 5 here. 6 DR. SIESS: They were probably anchorage concerns. 7 MR. CHANG: Could simply be they need more 8 information to determine whether it is adequate or not. 9 Mr.. MICHELSON: Qualifications required on the motor 10 control center? 11 MR. ANDERSON: I don't think that there were. As I 12 remember, I think that the anchorage configuration on those 13 motor control centers were on fairly old plant. I don't think 14 it was really covered in the screening guidelines that they 15 had. 16 HR. SCHMIDT: There are two kinds of problems with 17 electrical cabinet. One of them was questionable or marginal 18 anchorage, and the other one were potential interaction 19 problems because there was a limited, because there were 20 adjacent cabinets which were not bolted together but which had 21 a gap, and the teams were concerned that they might bang 22 together and cause a jarring ar something that would disrupt
.23 relays hnd so forth, so they got put on the list as a concern 24 side by side.
O 25 DR. SIESS: Those things are fairly-- HERITAGE REPORTING CORPORATION -- (202)628-4888
272 1 MR. SCHMIDT: They are about this wide and maybe
'T 2 three feet' deep, and they sit side by side and some of them k /
3 .weren't bolted together, had a gap between them. 4 MR. MICHELSON: Actually did anything to them? 5 MR. SCHMIDT: They are in the process right now of 6 resolving those outliers. 7 MR. MICHELSON: You don't know yet. 8 MR. SCHMIDT: In the case of the anchorage, they are 9 in the process of screening. 10 DR. SIESS: Rubber strips, an inch and a half 11 machine bolts. 12 MR. MICHELSON: There are more than five motor 13 . control centers of interest, though. Is that the only ones 14 you needed for safe shutdown following earthquake? 15 MR. SCHMIDT: I could dig up how many there are. 16 There are only a few. There were a total of nine I think that 17 had anchorage concerns. 18 MR. MICHELSON: Only five that were in the 19 walk-down. 20 MR. CHANG: There are eight that has anchorage 21 concerns. 22 MR. MICHELSON: I was looking at tne previous chart. 23 It says motor control centers. 24 MR. CHANG: Five out of five there. O 25 DR. SIESS: Only five were looked at, only five HERITAGE REPORTING CORPORATION -- (202)628-4888
l 273 1 needed-- 2 MR. CHANG: I think I recall since Bill mentioned 3 that I think two of them are because of the gap effect between 4 the two. 5 DR. SIESS: That is not the question. The question 6 is was there, were only five looked at? Are there only five 7 needed for safe shutdown? 8 MR. CHANG: Yes. I think based on the information 9 here, only for five M C are required. 10 DR. SIESS: No sampling? 11 MR. CHANG: Every piece of equipment is looked at. 12 MR. MICHELSON: I find it hard to believe--they 13 might all be valves, but I suspect they aren't, there are only O 14 five valves you need. 15 DR. SIESS: It is an old plant. 16 MR. MICHELSON: It is possible, but I find it hard 17 to believe. 18 MR. ANDERSON: I thought, T.Y. just mentioned to me 19 that they are expanding that equipment list there and being a 20 trial walk-down, they selected the equipment list, but I think
, 21 they are going ba:L and doing some more now, so you wouldn't--
22 MR. SCHMIDT: I don't want to, want you to be 23 misled. The A-46 doesn't require that you evaluate all safe 24 shutdown equipment in the plant. It requires that you select 25 a safc shutdown train with redundancy and that you evaluate HERITAGE REPORTING CORPORATION -- (2021628-4888
7-274' . l' that'one. f3 )2 MR. MICHELSON: That is what you did here?
\_J
.3 MR. SCHMIDT: That however many was on here is what 4 was'needed for one train plus. redundancy.
- 5. MR. MICHELSON: That train might be one, in one 6 division or in one train there might be two or three systems-7 in one train that you need for safe shutdown. I suspect there 8 is more than one system. Decay heat removal is the principal 9 one.
10 MR. SCHMIDT: There are substantially more.than five 11 valves and more than five pumps, too, so I suspect that there 12 is more than'one control in those MCCs. 13 MR. MICHELSON: Not generally one valve, only one 14 MCC. 15 DR. SIESS: Doesn't call them, doesn't say cabinets. 16 It says conters. Is center more than a cabinet? 17 CHAIRMAN WYLIE: That might be. 18 MR. MICHELSON: Normally a motor control center is 19 one module in a big panel. Each one is a motor control l 20 center. l l 21 MR. ANDERSON: This is supposed to represent the 22 whole equipment list. l 23 MR. JCHMIDT: I have got nine pages of equipment 24 here that is in that equipment list, and it has got at least
/~T
(/ 25 five different cabinets that are called, that are designated i HERITAGE REPORTING CORPORATION -- (202)628-4888
275
-1 as' motor control centers, and it has got a number of other L /"% ; 2 called power boards and power supplies and it may just be the G
3 designation. 4 MR. MICHELSON: There are dozens of motor-operated o 5 valves. 6 MR. SCHMIDT: The expansion of the list that is 7 being done right now is being done not because of A-46 but 8 because Niagara Mohawk has decided that it makes sense to, 9 while they are at it, to qualify all of the like pieces of 10 equipment that might be in any of the safe shutdown trains, so 11 they don't have to face up to which train is, has been 12 reviewed, which train hasn't been reviewed, they don't have to 13 address questions of LCo. They don't have to worry about 14 replacement parts being different and so they are in the 15 process of expanding the list for those reasons. 16 DR. SIESS: They won't take two manyears to decide 17 on that? 18 MR. SCHMIDT: No. It is, surprisingly it is not 19 that much of an increase in the scope. 20 MR. CHANG: Well, the finishing of outliers, are the 21 equipment may not be, that once the equipment is--it doesn't 22 mean that the equipment is not adequate seismically. It 23 simply means that one of the following things could be, that 24 it is being accessible. In this case, nine pieces of O 25 equipment, most of them are in the highly radioactive area and HERITAGE REPORTING CORPORATION -- (202)628-4888
276 L 1 it requires' extensive paperwork to have them ready to be 2 looked.at, so itRis simply there wasn't time enough to perform l _' s 3 that during the walk-down. 4 ' Anchorage concerns, interaction concerns, equipment
- 5. not in the experience data base, it could be that.the piece of 6 equipment is not in the data base at all, or if it falls 7 outside of the requirement there, if it doesn't satisfy the 8 caveat of the description of the equipment, or it could be 9 that the seismic demand exceeds capacity.
10 Anothar possibility is in the case of valve body and 11 yoke, there is a requirement that the people shouldn't use 12 valve with a body and yoke of the material of cast iron so if 13 you don't have that information, you are outlier until you O 14 have access to further informatien for confirmation. r 15 DR. SIESS: On next to the last one, seismic demand 16 exceeds capacity, how do you determine that in the walk-down i L 17 without a calculatien? l 18 MR. CHANG: That information was prepared before the l j 19 walk-down. They would have the floor spectra at each floor i 20 level ready for walk-down people, 21 DR. SIESS: That would be the demand. How would f 22 they get the capacity? 23 MR. CHANG: Capacity is listed in the EPRI book. 24 MR. ANDERSON: Generic equipment. 25 MR. CHANG: And also seismic experience; there is HERITAGE REPORTING CORPORATION -- (202)628-4888
277 1 _the bounding spectra available also.
/~~ 2 DR.'SIESS: Thank you.
()1 3 MR. SCHMIDT: The walk-downs are not a good, term. 4 There is both. office review, evaluation, as well as in-plant 5 as part of it. 6 (Slide) 7 MR. CHANG: The total efforts for the Nine Mile 8 walk-down is about two manyears, and the breakdown in manhour 9 in terms of percentage is as follows--can see a lot of time is 10 spent on the walk-down preparation, 28 percent, and I, the 11 identification of safe shutdown equipment, 18 percent, and the 12 training course took about 23 percent of the total manhours 13 there. 14 This is different from the Zion walk-down. Zion, 15 there was only short course conducted, only one day, one day 16 course. For the Nine Mile walk-down, a five-day course, eight 17 hours a day was conducted. That's why it took a big chunk out 18 of the total manhours. 19 MR. SCHMIDT: Approximately 50 hours per person. 20 MR. CHANG: The walk-down itself took about 26 21 percent of the time. 22 (Slide) 23 MR. MICHELSON: On the walk-down, in that case it 24 included, the analysis associated what they saw when they O 25 walked that 26 percent? HERITAGE REPORTING CORPORATION -- (202)628-4888
[ . 278 1- MR. SCHMILT: On the anchorage evaluation. 1^% 2 MR. MICHELSON: Preparation analysis you did before b '1 3- you walked down? 4 MR. SCHMIDT: Primarily to get the drawings and the 5 ' data you needed and to identify the equipment, fill out the 6 appropriate data sheets so that people were ready to conduct 7 the walk-down and had all the information that they needed. 8 MR. MICHELSON: How much of the efforts went into 9 the post-walk-down analysis that you did after you-- 10 DR. SIESS: It is on the screen. 11 MR. CHANG: What is the question? 12 MR. MICHELSON: I gather it is going to be answered 13 in a moment. 14 DR. SIESS: Something called walk-down follow-up. 15- Is that the answer? 16 MR. CHANG: Follow-up is the resolution of the 17 outliers. 18 MR. SCHMIDT: That was the last couple of days and 19 two-week period the time was spent gatharing, gathering 20 information that was missing during the actual physical 21 walk-down. The time required to resolve the outliers is not 22 included in that. 23 MR. MICHELSON: Is that more time on the qchedule? 24 MR. SCHMIDT: That is utility specific and he is, O 25 you know, undertaking to do that. I
- HERITAGE REPORTING CORPORATION -- (202)628-4888
279 1 MR. MICHELSON: Once walk-down was dene, then the
/~1- 2. utility had to worry about figuring it out?
T) s 3 MR. ANDERSON: That's right. They have to debate s 4 with us on what schedule and how they resolve them after it is 5' 'over. 6 DR. SIESS: Since they were exploratory, or
.7 experimental, was any attempt made for duplication?
8 MR. ANDERSON: At what? 9 DR. SIESS: Duplication, two teams walk the same 10 arra and see if they get the same answer. 11 MR. CHANG: Actually, to a certain extent that was 12 done. Two teams, they may have a duplication of certain . 13 amount of equipment list on same, on the walk-down, and it is 14 interesting that-- 15 DR. SIESS: Had four teams or three? 16 MR. ANDERSON: The three SRTs and observer team. 17 DR. SIESS: Those were the people. How many teams? 18 MR. CHANG: Three teams, three walk-down teams, but 19 there is one observer team that consists of NRC people. 20 DR. SIESS: How were the teams divided? By a oa in 21 the plant or components that they looked at? 22 MR. SCHMIDT: The way that was done is the 23 documenthtion of 175 pieces of equipment was prepared in 24 advance of the walk-down, and was distributed to the three O 25 primary walk-down teams on a daily basis, and was distributed HERITAGE REPORTING CORPORATION -- (202)628-4888
280
, -1 to them in a way that each team was required to evaluate all n
2 of.the different types of equipment.
.f S U
3 However, there'was not overlap. In other words, ) 4 there was no duplication of' effort, but there was duplication-5 in the sence that every team reviewed'every type of equipment 6 at.the plant. .The monitoring function war performed by the 7 fourth team which was referred to as observer team, who was 8~ made up of people who developed all this methodology, and who 9 broke up and accompanied the SRTs during their rounds in the 10 plant, and basically just monitored. 11 DR. SIESS: The motor control centers were not all 12 looked at by the same team? 13 MR. SCHMIDT: That is correct. O 14 DR. SIESS: Looked at by three teams? 15 MR. SCHMIDT: That is correct, for the entire scope 16 in each case of the study, but each, but of the five, only one. 17 team would have looked at each, you know, at any one cabinet, 18 but all three teams would have looked at NCC, for example. 19 Plus we had the observer team double-check their conclusions, 20 plus the SSRAP and NRC representatives also went out and 21 looked at any outliers. 22 MR. CHANG: In the case of Mine Mile. 23 DR. SIESS: That was my concern. 24 MR. SCHMIDT: Those were checked as well.
- p, V 25 DR. SIESS: Every component was looked at by more HERITAGE REPORTING CORPORATION -- (202)628-4888
281 1 than one ceam it you count the observer team? {} .MR. SCHMIDT: Oh, absolutely. - 2 3 DR. SIESS: Hight have been they were looked at by
.4 more than one of the--
5 MR. SCHMIDT: Likely looked at by three different 6 teams. 7 DR. SIESS: How much record do you have on whether 8 there was an outlier found by one team and not by the others 9 for the same thing, two out of three or something? How 10 consistent were the teams in finding outliers? 11 MR. SCHMIDT: The answer to your question is, I 12 think is generally consistent, but in some cases, some team 13 ' members who were let's say more conservative may have raised 14 more outliers than other teams. l 15 I think the botton line was that the very 1 16 experienced people, namely the SSRAP, concluded that, and one 17 of their conclusions of the walk-down was that where the SRTs' 18 ok inions or judgment differed from SSRAP's, they were 19 generally on the more conservative side than was SSRAP's. 20 DR. SIESS: Then why was the, you know, he said 21 something about concerned about that the teams be adequately 22 experienced. From what you are saying, it seems to me you get 23 a very good result by taking a team of less experienced people 24 and let them do all of it, and then have a more experienced 25 team just check the outliers. 9 HERITAGE REPORTING CORPORATION -- (202)628-4888
282
.1 MR. ANDERSON: I suppose.
2 DR. SIESS: Less experienced people found more
)
3 outliers. 4 MR. ANDERSON: You need a level of experience where 5 people understand what, when they are seeing an outlier. You 6 don't want people-- 7 DR. SIESS: I was just told that, that-- 8 MR. SCHMIDT: Less experienced people tended to find 9 more-- 10 DR. SIESS: This is a conservative approach and then 11 you check them with the experience. 12 MR. SCHMIDT: There is also a conclusion in the
.s 13 report of the walk-down that, for example, in the seicmic 14 review teams, the procedures required two qualified seismic 15 engineers to do the walk-down and double-check, consult each 16 other.
17 In the walk-down at Nine Mile Point we had a third 18 more junior seismic capability engineer on the team just to 19 see how he would do. The experience levels on the more senior 20 people ranged in the range, I think the numbers were like on 21 the average 17 years of experience in scismic and 22 qualification type work. The more junior engineer typically 23 was in the six to ten or six to twelve years of experience 24 range. O 25 The conclusion that was reached by the team members HERITAGE REPORTING CORPORATION -- (202)628-4888
283 1 was that, that of those groups they generally wnsws up They had pretty consistent . 2 agreeing on the conclusions.
' (J'l 3 conclusions. .I think from my observations when I was there, 4 the less experienced people tended to do more analysis, more 5 double-checking perhaps than the very senior people.
6 DR. SIESS: What I am getting at, as you go on with 7 this program, what you did there required a minimum of nine 8 people with so-called seismic capability plus whatever was on 9 the-- 10 MR. SCHMIDT: Only required one team. We did three 11 in order to get a cross-check and train more people. 12 DR. SIESS: In the future do you think you will go 13 to one team? 14 MR. SCHMIDT: What we concluded in the report, that 15 one team is enough, two teams provided a good opportunity for 16 crosstalk and consistency checks. 17 MR. CHANG: It is interesting to compare the Zion 18 walk-down results and the Nine Mile walk-down results. The 19 numbers are pretty comparable--items of equipment that needs 20 to be reviewed, one is 159, other one is 175, and so forth. 21 But the interesting thing here to see is that Nine Mile spent 22 more time on walk-down preparation. It is almost twice as 23 Zion. I think it has some effect on the walk-down time spans. 24 If you spend more time on the preparation, I think it tends to O d 25 end up with a shorter walk-down time. HERITAGE REPORTING CORPORATION -- (202)628-4888
e-284
- 1. DR. SIESS: Combined; I find it interesting-that an
- ~j 2 old plant like Nine Mile, one finds about one out of six 3 outliers, and somewhat more modern plant like Zion came out
'4 with one out of four. Is that--
5 MR.-ANDERSON: 'I think there is observations. I 6 know Bill probably-- 7 DR. SIESS. Zion has done some fixing. 8 MR. ANDERSON: Nine Mile is a'very spacious plant, 9 and there is a lot of room in it and they don't jam everything 10 together and you can see things, and so they tend to fix 11 things. They are easier to put in. 12 My impression at Zion is it is a, everything is 13 jammed together. We saw some systems interaction problems O 14 there that I thought were my own impression pretty bad. It is 15 because I think it was our fault, we forced people to jam 16 rooms full of stuff and they didn't have any place to put it. 17 DR. SIESS: Pipe supports, pipe restraints. 18 MR. SCHMIDT: Other observation--Zion and Nine Mile 19 aren't that much different in age. Zion is earlier '70 plant. 20 DR. SIESS: But has Zion had seismic PRA? Did they 21 do anything significant as a result of that? 22 MR. SCHMIDT: I don't know the answer. 23 MR. ANDERSON: I know a number of people have walked 24 through Zion, but I was very impressed w:' th the depth of the O 25 review that was done in Zion and they found, they found HERITAGE REPORTING CORPORATION -- (202)628-4888
285 l' outliers that I don't even know how they saw them. They would 2 have to'be~a contortionist with double mirrors to get to them. 3 They found some outliers there, very consciously went through 4 for the' equipment scope. They did what I consider to be an 5 excellent review. 6 DR. SIESS: Maybe we should have'a 3D computerized 7 display for these plants and nothing would be inaccessible. 8 MR. CHANG: On the lessons-- l l 9 DR. SIBSS: It would take two manyears to develop 10 it. 11 MR. CHANG: On the conclusions of the Nine Mile 12 walk-down, first of all, this is the ESRAP observations and 1 13 conclusions. The SSRAP members, they felt that, that the O 14 walk-down team members had received adequate training and 15 adequate and qualified job of evaluating equipment. 16 DR. SIESS: Let me back up a minute. Team members 17 are not all from Niagara? 18 MR. CHANG: No. 19 DR. SIESS: This are people that are available for 20 future walk-downs? 21 MR. CHANG: That's one of our comments. This may 22 not be typical for the future walk-down. 23 DR. SIESS: I said are they available for future 24 walk-downs? Are they professionals now? 25 MR. SCHMIDT: No. HERITAGE REPORTING CORPORATION -- (202)628-4888
286 1- MR. CHANG: From other. utilities. 2 MR. SCHMIDT: They were all utility engineers except 3 for-the observers. 4 DR. SIESS: Not Niagara Mohawk? 5 MR. SCFMIDT: No. They came from a number of 6 volunteers, including the people who were going to be working 7 on Hatch later this year. 8 DR. SIESS: Utilities volunteered their people, 9 those people to do walk-downs themselves? 10 MR. SCHMIDT: For their own plants.
'll DR. SIESS: You are not developing a cadre of 12 walkers? The utilities are trying to build up experience by 13- having their people participate in other programs?.
14 MR. SCHMIDT: Right. 15 DR. SIESS: Is there a systematic approach to this 16 among the utilities? 17 MR. SCHMIDT: Dr. Chang said they are scheduled for, 18 when all of this methodology is complete, and training 19 programs are all finalized, there will be widespread training 20 given to all utility members next year. It will be required 21 of anybody who does the walk-down. 22 DR. SIESS: That training means participating in 23 other walk-downs like you have teams here? 24 MR. SCHMIDT: No. It means having gone through, O 25 meeting the qualification requirements and'naving gone through HERITAGE REPORTING CORPORATION -- (202)628-4888
287 1 the trainng course, which probably will be in the order of a 2 week give or take. 3 MR. CHANG: As Bill Schmidt mentioned earlier, SRT 4 and SSRAP felt SRTs' judgments on outliers probably were more 5 conservative compared to SSRAP's conclusion. And in general 6 SSRAP felt that for the vintage considered, Nine Mile is, 7 probably is one of the better plants from a seismic point of 8 view. 9 NRC observations and conclusions--generally the NRC 10 staff agreed with the SSRAP assessment of the Nine Mile 11 walk-down. 12- On the qualification of the walk-d i engineers, it 13 appeared that the utility engineers are doing a good job, but L. 14 we want to emphasize that people doing the walk-down still 15 have to satisfy the minimum requirement described in the 16 generic implementation procedure. The feeling is that as soon 17 as that basic requirement and the experience is satisfied, 18 then after people go through the training course, usually we 19 think they can do a pretty decent job. 20 However, as pointed out by Dr. Siess earlier, we 21 think that the caliber of the Nine Mile walk-down members may 22 not be representative of future walk-down teams. 23 DR. SIESS: Which direction? Better or worse? 24 HR. ANDERSON: We are concerned that they, that some 25 utilities may put some ringers in on us or bring people in who HERITAGE REPORTING CORPORATION -- (202)628-4888
288 1 really don't have the degree of experience we think is 2- necessary. If they want to use their own people, they' don't 3 have'anybody of that level of experience, you know, we are 4 afraid they may just put them into the training course and 5 . send them through and they may not even recognize an anchor, 6 let alone know how to judte its adequacy. 7 I don't know that's the case, but that's something 8 we are concerned about, and Nine Mile Point was utility 9 engineers, but'they were carefully selected, and for the most 10 part they were very competent and experienced structural 11 people. I an not sure all the utilities have that level of r 12 experience.
-13 DR. SIESS: What disciplines?
14 MR. ANDERSON: Structural and civil and seismic 15 engineering; they were-- 16 MR. CHANG: Some people's background maybe i 17 mechanical, but it was a collective team experience. . r 18 DR. SIESS: Visually what would happen when the l , 19 earthquake hit? ! 20 MR. ANDERSON: People had that kind of experience, 21 worried about that kind of problem. 22 MR. MICHELSON: Had at least a little bit of 23 electrical experience? 24 MR. ANDERSON: Yes. They are familiar and have O 25 experience with the seismic capacity and capability of the HERITAGE REPORTING CORPORATION -- (202)628-4888
l 289 1 electrical equipment, and-- l 2 MR. MICHELSON: Part of the effect of relay chatter, 3 that sort of thing? 4 MR. ANDERSON: That is done by system engineering. 3 There is systems engineer from the plant who accompanies the 6 team. 7 MR. MICHELSON: He is part of the team? 8 MR. CHANG: Relay review is done separately. 9 MR. SCHMIDT: Completely separate at a different 10 time than the walk-down; it was done specifically by l 11 electrical people. l l 12 MR. CHANG: By the way, the manhours shown earlier 13 does not include a relay review effort. That's extra effort. O 14 The layout of the Nine Mile, it is quite open, is a 15 considerable open space at Nine Mile. That contributad to the l 16 lack of seismic itteraction hazards. In other words, as a 17 result of this effect, we have seen very few seismic 18 interaction concerns. 19 MR. MICHELSON: Wasn't that inherent in the, that's 20 a BWR 4, isn't it? 21 MR. SCHMIDT: BWR 2. 22 MR. MICHEL3ON: But it is the, it is the four 23 corners of the basement kind of arrangement, pumps and 24 everything as that is inherent to be well spread out as i
+ 25 opposed to Zion where you pack them into little cubicles and HERITAGE REPORTING CORPORATION -- (202)628-4888
290 1 stack them side by side? 2 MR. SCHMIDT: They-were their own
}.
3 architect / engineer. They have'a' tendency to spreadLthings.out' 4 even more. 5' MR. MICHELSON: Might have been a bigger box. 6 MR.'SCHMIDT: Turbine-building and so forth has more 7 . space. 8 MR. MICHELSON: BWRs are inherently well spread out 9 anyhow. 10 DR. SIESS: Until you get inside the-- 11 MR. MICHELSON: Because of four corner arrangement 12 the electrical generally feeds the four train physical 13 arrangement. 14 MR. ANDERSON: On the other hand, Zion in some 15 areas, particularly in auxiliary feedwater room, it is hard to 16 .even see through it, let alone walk through it. It is a mess. 17 MR. MICHELSON: Sort of like Sequoyah. 18 MR. CHANG: Well, in general, we felt the walk-down 19 was conducted quite smoothly and lessons learned from Zion 20 walk-down and the pre-planning was, contributed to this ,
- 21 successful walk-down.
22 DR. SIESS: Mr. Chairman. I have a question but it 23 is not addressed to to Mr. Chang. Can I ask it? 24 CHAIRMAN WYLIE: Sure. 25 DR. SIESS: Very short answer--if I wanted somebody HERITAGE REPORTING CORPORATION -- (2021628-4888
. .~ ,_. -~ . - - . _ , . .- - . _ . _ . - _ _ _ _ _ - _ . _ , _ . _ _ _ , . - ._.-._... ._. _ __.. _ __,- _ _.
291 i 1 to explain to me the differences between the A-46 walk-downs I and the seismic margin program, who would I ask? {) 2 3 MR. ANDERSON: I guess you could ask me. 4 DR. SIESS: I won't ask you now, but I will put it 5 on the list for me. 6 MR. ANDERSON: As a matter of fact, on this--I will 7 just make a comment at the Hatch walk-down in October, there 8 will be an A-46 walk-down and seismic margins walk-down, get 9 an insight how they can-- 10 DR. SIESS: There is difference in why we are doing i 11 them, too. 12 MR. ANDERSON: Yes. 13 DR. SIESS: Would be interesting to have that O 14 somewhere in the picture. 15 CHAIRMAN WYLIE: Carl, do you have anything to say? 16 MR. MICHELSON: No. What kind of action are we 17 anticipating on this? 18 CHAIRMAN WYLIE: Nothing. This was for our 19 information. We will report it. 20 DR. SIESS: You are going to be at this for the next i 21 six years, huh? 22 MR. ANDERSON: Most of us are going to retire on p 23 this program I think!, 24 CHAIRMAN WYLIE: This was their career! O 25 DR. SIESS: This is all the plants before when? HERITAGE REPORTING CORPORATION -- (202)628-4888
292 l-1 MR. ANDERSON: It is cps before about 1972,-but the L I
- 2. real criteria was those that had not been reviewed for seismic
{'} 3 qualification of the current criteria. 4 DR. SIESS: Okay. 5 CHAIRMAN WYLIE: Thank you, Dr. Chang. Adjourned. 6 (Whereupon, at 6:40 p.m., the meetir.g was 7 adjourned.) 8 l 9 10
- f. 11 12 14 15 16 17 18 19 20 21 22 23 24 O 25 HERITAGE REPORTING CORPORATION -- (202)628-4888
r 1 CERTIFICATE O 2 3 This is to certify that the attached proceedings before the 4 United States Nuclear Regulatory Commission in the matter of: 5 Name: ACRS--Subcommittee on Reliability Assurance 6 7 Docket Number: 8 Places Washington, D.C. 9 Date: June 14, 1988 10 were held as herein appears, and that this is the eriginal 11 transcript thereof for the file of the United States Nuclear 12 Regulatory Commission taken stenographically by me and, 13 thereafter reduced to typewriting by me or under the direction 14 of the court reporting company, and that the transcript is a 0 1s true end eccurete record of the fore 9eine groceedinee. 16 /S/ [ Y wr d f - 17 (Signature typed): Catherine S. Boyd 18 Official Reporter 19 Heritage Reporting Corporation 20 21 22 23 24 25 O Heritage Reporting Corporation (202) 628-4888 4
,3- .o i - l7 0 ' *
.'"hl j_. ., . b l- I, p- . .. l {. , l. b i-1 Meeting with the ACRS Subconnittee l on'Reltability Assurance on the 1: , " Equipment Qualification Scoping Study
-(FIN: A1823) l June 14, 1988 l
Presentation By-
'Moni Dey, Project Manager Advaaced aeactors 5 ceaeric issues Breach
- O.-
Division of-Regulatory Applications Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Connission
,p-f O
y
4 PROJECT INITIATION
~
()I - SINCE'1975, NRC HAS FUNDED EQ RESEARCH TO STUDY THE~ METHODS FOR QUALIFYING' SAFETY-RELATED ELECTRICAL EQUIPMENT TO DEMONSTRATE ABILITY TO FUNCTION DURING AND 'F0(LOWING DESIGN BASIS ACCIDENTS WHICH PRODUCE HARSH ENVIRONMENTS. THE PROGRAM HAS TERMINATED IN 1986. RESULTS OF THE RESEARCH ARE DOCUMENTED IN NUREG/CR-4301, "STATUS REPORT ON EQUIPMENT QUALIFICATION ISSUES RESEARCH AND RESOLUTION <" ACRS LETTER (6/11/86) RECOMMENDED FUNDING OF EO RESEARCH TO ASSESS SURVIVABILITY OF ELECTRICAL EQUIPMENT WHEN SUBJECTED TO HOSTILE. CONDITIONS INCLUDING SEVERE ACCIDENTS. STATED () "..... RESEARCH ARE VITAL TO PREVENTING ACCIDENTS AS WELL AS L MITIGATING THE CONSEQUENCES OF ACCIDENTS, SHOULD THEY OCCUR." EO SCOPING STUDY WAS INITIATED TO DETERMINE THE RISK IMPORTANCE/ PRIORITY FOR NRC FUNDED EO RESEARCH. O __,. . - ~ - - OBJECTIVE OF' EQ SCOPING STUDY
.(]):
TO DETERMINE THE RISK ' SIGNIFICANCE OF: -
- 1. ELECTRICAL'EQUIPMENTTHATAREESSENTIALTOPRbVENTING CORE-MELT ACCIDENTS AS WELL AS MITIGATING THE CONSEQUENCES OF SEVERE ACCIDENTS:
- 2. E0 ASSUMPTIONS AND ISSUES PERTAINING TO THE PERFORMANCE OF THE ABOVE EQUIPMENT. -
THE RESULTS OF THE STUDY WILL BE USED TO: () 1. DETERMINE THE NEED FOR IDENTIFYING NEW GENERIC ISSUES; A. DEVELOP / REVISE REGULATIONS'(50.49) AND/0R B. SUPPOR,T. REVISION OF NRC GUIDES (E.G. R.G. 1.89)
- 2. SUPPORT IMPLEMENTATION OF THE COMMISSION'S SEVERE ACCIDENT POLICY FOR EXISTING AND FUTURE PLANTS. (WHAT EQ CONCERNS SHOULD BE ADDRESSED?)
- 3. DETERMINE IF A NEED EXISTS FOR FURTHER EQ RESEARCH TO CHARACTERIZE AREAS OF LARGE UNCERTAlf4TY.
- 4. PRIORITIZE ANY ADDITIOilAL E0 RESEARCH.
J. ( PROJECT STATUS / SCHEDULE ., DRAFT FINAL REPORT DISTRIBUTED FOR NRC AND ACRS COMMENTS ACRS COMMENTS REQUESTED IN JULY PRELIMINARY COMMENTS FROM PEER REVIEW TEAM RECEIVED AND INCOR0 RATED INTO THE DRAFT REPORT. FINAL COMMENTS DUE SHORTLY. l FINAL NUREG REPORT TO BE PUBLISHED BY JULY 29, 1988 RESEARCH INFORMATION LETTER TO BE ISSUED BY SEP".tBE8 30,1989-l l 6 l O l O 4
'h .
O EQ-Risk Scoping Study: Project Objectives / EQ lssues O l l i l O l
EQ-RISK SCOPING STUDY - OBJECTIVES
- 1. Use existing PRAs as a basis for study.
1 1
- 2. Assess the impact of electrical equipment environmental qualification or lack thereof on reactor risk and its uncertainties. '
- 3. Identify any analyses or testing that may be necessary to reduce the risk or its uncertainties g stemming from lack of qualification of equipment important to safety.
O
O EQ-RISK SCOPING STUDY - SCOPE Scope: Electrical safety-related equipment that must function during harsh environments. Examples: Cables, valve actuators, pump motors. , Not included; Mechanical safety-related equipment that must function during harsh environments. s Examples: Check valves, globe valves. O
EQ-RISK SCOPING STUDY - EQUIPMENT ISSUES O
- 1. NUREG/CR-4301 issues.
NUREG/CR-4301: Status Report on Equipment Qualification issues Research and Resolution. Many of these issues question the appropriateness of DBE EQ practices.
- 2. Other issues important non-DBE accident environments / g sequences.
Lack of equipment accident reliability information. O
l O NUREG/CR-4301 EQ ISSUES: Generic issues Related to Methods for Simulating Accident Conditions. Is the sequential exposure of equipment to radiation followed by steam an adequate simulation of the accident or are synergistic effects important? Are the effects of superheated steam important to duplicato, and what are the appropriate methods? Is steam impingement and/or thermal shock O important to duplicate, and what are appropriate methods? Are radiation dose-rate effects a factor in equipment degradation / failure and how must they be simulated? Can a gamma radiation source be used to simulate the mixed (accident) beta / gamma radiation? How does the presence of oxygen affect accident-simulation test results? O
O NUREG/CR-4301 EQ ISSUES: Generic issues Related to l I Methods for Simulating Accident Conditions (Continued). Is the variability of chemical spray composition significant in the accident simulation? Can the effects of postaccident environments be satisfactorily accelerated? Does the method, or choice, of accelerated-aging affect the behavior of equipment and materials O during accident simulation? Does the hydrogen burn issue affect the qualification method chosen? la submergence adequately simulated in accident qualification tests? O
O NUREG/CR-4301 EQ ISSUES: Generic issues Related to. Methods for Simulating Aging Conditions What are the realistic plant ambient environments, and how do they impact the choice of accelerated-aging methods in the aging simulation? When is the Arrhenius methodology invalid? Are dose-rate effects a factor in aging simulation, and how must they be simulated? O can sequential exposures be use'd to simulate simultaneous radiation / thermal environments and does the order of the sequence matter? I is mechanical stress an important factor (i.e., in seal / gaskets) in the methods used for accelerated aging? O 4
. . , , - - - - - - - . . - . . n -- , , , . - - - - _ - . - - - -- - _ , - - - - . - - , . . . . - . - - - . - - - - - . ~ - - - - . - - - -
O NUREG/CR-4301 EQ ISSUES: Generic issues Related to Methods for Simulating Aging Conditions (Continued) What is the effect of oxygen in ambient and accelerated aging? Does humidity affect aging, and how can it be accelerated? Are there analytical and experimental techniques which can be used to correlate real-time aging and accelerated aging (methods)? O What aging and degradation experiences are available from actual plant use, and can they be correlated with accelerated aging methods? O
O NUREG/CR-4301 EQ ISSUES: Special Topics Related to Equipment Qualification Are there test approaches (such as multiple sample, fragility and overstress testing) which can complement qualification type-testing to achieve a broader safety perspective? What can we learn from the TMI-2 accident and research regarding simulation of accident conditions? O Are the procedures given in the IEEE Standard = a>r qualifying specific types of electric equipment adequate? 3 is post-accident monitoring (PAM) equipment adequately quallfled to meet the requirements of Regulatory Guide 1.977 What are the applicable qualification methods for i advanced systems? 4 1 O
O NUREG/CR-4301 EQ ISSUES: Special Topics Related to Equipment Qualification (Continued) What are realistic accident environments, and how . do they impact the choice of simulation methods? Can criteria be established for selecting the appropriate simulation method for equipment and materials? Do the standards and Guides inclusively and adequately encompass the equipment qualification issues? O What are the appropriate accelerated-aging methodologies for nuclear station batteries? Can radiation damage thresholds be established for ' organic materials, electronics, and equipment generally? O
l O QINER ISSUES: Important non-DBE accident environments / sequences. Examples: Steam venting from BWR drywell to reactor building. Severe accident steam, pressure, and radiation conditions. O PRA modelling of non-safety-related equipment.
"feed and bleed" components non-safety-related fan coolers O
OTHER ISSUES: Lack of equipment reliability information. What PRA risk uncertainties occur because: EQ does not assure that normal operation reliabilities will be maintained during harsh environment conditions. EQ does not measure equipment harsh environment reliabilities. g PRA typically employs normal operation reliabilities to describe harsh environment equipment operation. O
1 O ga_ Risk Scoping Study: Project Objectives / EQ lssues Summary: PRA insights and analysis techniques are to be used to assess the risk significance and uncertainty associated with equipment operability issues. Equipment operability issues include
- 1. Traditional EQ issues such as those dis ussed by NUREG/CR-4301.
O
- 2. Issues associated with EQ versus PRA perspectives and approaches.
O
ANALYSIS FRAMEWORK: PRA and EQ Perspectives O Important to EQ-Risk Scoping Study Project Approach. PRA Perspectives:
- 1. PRA risk-significant accident time frames.
- 2. Post-core melt accident management strategies. ,
- 3. PRA equipment operability assumptions.
- 4. PRA plant status instrumentation analyses and modelling.
O
O ANALYSIS FRAMEWORK: PRA and EQ Perspectives important to EQ-Risk Scoping Study Project Approach. EQ Perspectives:
- 1. EQ equipment description.
- 2. EQ equipment operability perspectives.
O a. EQ radiation source terms.
- 4. EQ-important accident sequences.
O
PRA RISK SIGNIFICANT ACCIDENT TIME FRAMES O PRAs only model the first few days of an 3 ] accident sequence. Core damage frequency and risk are substantially dependent on the first few hours to days of an accident sequence. EQ, in contrast, quallfles some equipment for very long accident durations. g Containment fan motors - one year operability Cables - 100 day to one year operability PRA perspective:could have a large impact on EQ practices. 9
O PRA RISK SIGNIFICANT ACCIDENT TIME FRAMES Three plant damage states are applicable to this issue.
- 1. Core integrity is being maintained.
- 2. Core melt has occurred but the vessel has not been breached.
O s. Core melt and vessel breach have occurred. O .
O PRA RISK SIGNIFICANT ACCIDENT TIME FRAMES: Core integrity is maintained. Decay heat loads reduce substantially over a few day period. l After one day - a factor of 10 reduction After a week - a factor of 35 reduction Reduction of decay heat loads provides operators substantially more flexibility to respond to equipment failures. O More time to respond to equipment failures. Additional techniques become available to perform core cooling. Example: BWR CRDM pumps l
Conclusion:
Most core damage frequency contribution is from first few hours to days g after accident initiation.
O PRA RISK SIGNIFICANT ACCIDENT TIME FRAMES: Core melt and vessel breach have occurred. Containment failure due to direct containment heating will fall containment rapidly. Containment failure due to core-concrete interactions will take hours and is unlikely if containment heat removal is available. Engineered safety features (i.e., containment sprays) and natural plateout will reduce O airborne radiation concentrations over a several hour period. Zion small-break LOCA sequence: NUREG-0956 lilustrates that a delay in containment failure from 2 hours to 15 hours reduces risk by an order of magnitude.
Conclusion:
Most risk contribution occurs early after core melt and vessel breach. O 1
O PRA RISK SIGNIFICANT ACCIDENT TIME FRAMES: Arrested core melt sequences. This scenerlo is not well modelled by current l PRAs. Within a week of accident initiation: Plant personnel have had adequate time to repair those systems outside containment that contributed to initial core melt. O Plant personnel have had time to develop alternative and possibly innovative core cooling methods.
Conclusion:
Most risk contribution occurs ' early after accident intitiation. O
PRA RISK SIGNIFICANT ACCIDENT TIME FRAMES: Summary Core damage frequency and risk are substantially dependent on equipment operability during the first few hours to days of an accident sequence. This conclusion is important to the EG-Risk Scoping Study. O O
POST-CORE-MELT ACCIDENT MANAGEMENT STRATEGIE 3 PRAs currently provide limited modelling of post-core melt accident management strategies. Increased interest in post-core-melt accident management strategies may necessitate changes in PHA models and/or equipment operability programs. EQ-Risk Scoping Study analyses is limited to g currently modelled strategies (i.e., containment sprays, fans, etc.) l O
O PRA EQUIPMENT OPERABILITY ASSUMPTIONS PRA typically models harsh environment accident conditions using normal operation equipment reliabilities. EQ process rarely supplies equipment reliability information for harsh environment accident conditions. Thus, insufficient information exists to quantify O risk impact when-Harsh environment equipment reliability differs from normal operation reliability. Hence, EQ-Risk Scoping Study could not rigorously requantify PRA models to assess the importance of various EQ issues. A combination of parametric and qualitative analyses was performed to scope the importance of various EQ issues. O
O PRA EQUIPMENT OPERABILITY ASSUMPTIONS . 3 PRA does employ risk achievement measures to identify equipment that could substantially impact risk if failure rates increase above nominal values. Only limited harsh environment safety-related equipment is highlighted by these calculations. Analysis tends to be one component at a time; EQ equipment is redundant. O EQ equipment not always modelled in fault trees. Risk achievement is reported more for "front-end" analyses than "b.ack-end" analyses. Hence, EQ-Risk Scoping Study employed additional techniques to identify potentially risk significant harsh environment equipment. O 1
O PLANT STATUS INSTRUMENTATION: PRA Analyses and Modelling A substantial amount of in-containment EQ equipment monitors plant status. Risk importance difficult to assess. Diverse instrumentation available to operator. O Lack of human reliability PRA models for plant status instrumentation. Hence, during the EQ-Risk Scoping Study we were unable to assess the risk importance of selected plant status instrumentation. l 9 .
O EQ EQUIPMENT DESCRIPTION l l 3 EQ testing and research has historically focussed on components such as cables, electrical penetrations, transmitters, etc. Many PRA models do not describe systems with the detail necessary to assess importance of "EQ component" failures. Hence, the EQ-Hisk Scoping Study O Parametrically varied a "PRA system" reliability to assess risk importance. Qualitatively assessed which EQ issues were most likely to impact system performance via their impact on system subcomponents. 9
l O EQ RADIATION SOURCE TERMS
- EQ assumes an instantaneous release of part of the core inventory to containment.
During the EQ-Risk Scoping Study we assumed that the EQ equipment source term is adequate . to describe severe accident in-containment radiation conditions. Additional research regarding this issue O is currently being performed by the NRC. l important equipment operation: i containment sprays. i O 4
IMPORTANT ACCIDENT SEQUENCES O EQ is for LOCAs, MSLBs, and HELBs. PRAs suggest a more diverse set of important accident sequences. Examples: Station blackout, transients, ATWS sequences.
O . Draft NUREG-1150: Contributors to Core Damage Frequency (PWRs) Event Tvog Surry ZlGD Seouovah LOCA 7.0E 6 = 28% 2.7E 5 = 18% 5.1E-5 = 59% Transient <1 E-7 1.0E-7 Less of Bus 5.0E-G = 20% 2.2E 6 = 3% Loss of Offsite Power 1.1 E-6 = 4% 1.0E 6 = 1% ATWS 1.6E 6 = 6% <12 8 1.1 E 6 = 1% i i Interfacing I LOCA 9.0E-7 = 4% 1.1 E-7 3.3 E-7 i Loss of Servic6 water or component cooling water <1 E-7 1.2E-4 = 79% 2.7E-5 = 31% Station l Blackout 9.5E-6 = 38% 3.0E-6 = 2% 4.1 E-6 = 5% l Other 5.0E 8 <1E 6 = 1% 2.6 E-5 1.5E 4 1.0E 4 O
1 O DRAFT NUREG-1150: C ntributors to Core Damage Frequency (BWRs) Event Type Peach Bottom Grand Gulf Station Blackout 7.0E-6 = 86% 2.8E-5 = 99% ATWS 1.0E-6 = 12% 1.8E-7 = 1% O LOCAs 1.0E-7 = 1% <2 E-8 Transients with loss of long term heat removal <1 E-8 <1 E-8 1 ( Transients with l loss of all coolant injection 6.8E-8 <1 E-8 0 8.2E-6 2.8E-5
O IMPORTANT ACCIDENT SEQUENCES The technical importance of an EQ issue may apply to non-DBE environments. The draft NUREG-1150 assessment is based mostly on normal operation reliability statistics. Would harsh environment equipment operability information alter these O conclusions? 6 O
ANALYSIS FRAMEWORK: Summary Several EQ snd PRA practices / perspectives were important towards defining the project approach and conclusions.
- 1. PRA risk-significant accident time frames.
- 2. Post-core melt accident management strategies.
- 3. PRA equipment operability assumptions. g
- 4. PRA plant status instrumentation analyses and modelling.
- 5. Eu equipment description.
- 6. EQ equipment operability perspectives.
- 7. EQ radiation source terms.
- 8. EQ-important accident sequences.
O
O PROJECT APPROACH Several "candidate" equipment operations that must be accomplished in a harsh environment were used to scope the risk significance of EQ issues. O
O PROJECT APPROACH
- 1. Developed a list of candidate risk significant equipment whose operability is desired during harsh anvironments.
Historical risk achievement analyses Consultation with PRA analysts Examining the timing of accident progression and mapping likely risk significant equipment O with harsh environment locations. i O
. I O PROJECT APPROACH
- 2. Reviewed qualitative EQ data sources to identify those EQ components for which accident reilabliltles may differ from normal operation ;
reliabliltles. Provided qualitative data base to , ustify harsh environment parametric risk achievement analyses for the candidate equipment operations. Helped focus the selectlon of equipment O operations for further study. l 3. Chose selected equipment operations for further study. l 0
O PROJECT APPROACH 1 l l
- 4. Evaluated in more detail the selected equipment ;
operations. Determined how PRAs model the equipment operation. ! Assessed for which sequences / environments PRAs suggested equipment operation was important. Evaluated risk impact of equipment operation. O Assessed applicability of historical equipment t operability issues. l I l
- 5. Developed generalized conclusions after evaluation of several equipment operations for both PWR and BWR plants.
O L--_--__-___________-.___-_._. - _ . __ _ . . --
O PROJECT APPROACH Circumvents numerous constraints imposed by current EQ and PRA practices. Provides a reasonable method for eliminating certain EQ issues from further consideration. Useful for identifying the direction of further study. O l l r O
1 1 O O O ( 3 l . l l l EQ - RISK SCOPING STUDY RISK SIGNIFICANT EQUIPMENT / OPERATIONS FOR BWRs/PWRs l IN HARSH ENVIRONMENTS i ! ) i i u N ) 4
)
s E N I L T U O N
- ~ 1 O
I S . K S U S C A S T I D W E Y I G V O R L E O Y V D R O O A K H M S T M A E U _ T M S
- ~
( l 4 l i,
TASK 1 OVERVIEW PURPOSE
- lDENTIFY FROM A PRA PERSPECTIVE IMPORTANT EQUIPMENT OPERATIONS WHICH COULD BE IMPACTED BY A HARSH ENVIRONMENT PROCESS
- LITERATURE REVIEW
- REVIEW REPRESENTATIVE PRA iMPORTANCE STUDIES TO IDENTIFY RISK SIGNIFICANT EQUIPMENT
- FORMALIZED PROCESS
- RANKING OF EQUIPMENT BASED ON EQUIPMENT IMPORTANCE, EQUIPMENT LOCATION, AND i POTENTIAL ENVIRONMENTS FOR EACH ACCIDENT i TYPE REVIEWED N ]
LITERATURE REVIEW PROCESS DETERMINE A REPRESENTATIVE SET OF PRA STUDIES TO REVIEW V REVIEW STUDIES FOR INSIGHTS ON RISK SIGNIFICANT t EQUIPMENT l V SCREEN THE INSIGHTS FOR POSSIBLE E.Q. SIGNIFICANCE TABULATE RESULTS
i ! O e G l r 3 i l STUDIES REVIEWED
. ASEP-NUREG-1150 :
! i IMPORTANCE RANKING BASED ON AGING CONSIDERATIONS - NUREG-4144
~
l . ECUIPMENT SURVIVABILITY IN DEGRADED CORE ENVIRONMENTS - l IDCOR TASK 17 l l . RISK IMPORTANCE TO PRIORITIZE EQUIPMENT QUALIFICATION - NSAC/36 i
. SIGNIFICANT CONTRIBUTORS TO CORE DAMAGE - NUREG-3762 STATUS MONITORING DURING ACCIDENT CONDITIONS - NUREG-1440 ELECTRIC EQUIPMENT ACCIDENT PERFORMANCE - NUREG-4537
! SIGNAL VALIDATION EFFECTS ON COMMON CAUSE FAILURES - EPRI-NP-5081 l . CRmCAL COEPONENTS AND MONITORING TECHNIQUES FOR NUCLEAR POWER j PLANT LIFE EXTENSION - SAND 87-0184 i
. DECAY HEAY REMOVAL ANALYSIS -TAP-45 I
i i i !
g _ _ 4
~D
( STUDY REVIEW SCREENING TECHNIQUES
- QUALITATIVELY CONSIDER EQUIPMENT LOCATION IN TERMS OF POTENTIAL FOR SEEING A HARSH ENVIRONMENT l USE RISK INCREASE AND RISK REDUCTION CALCULATIONS TO IDENTIFY IMPORTANT EQUIPMENT
- DEFINITIONS 1
RISK REDUCTION: l A MEASURE OF HOW MUCH THE RESULTS ARE REDUCED GIVEN A SPECIFIC EVENT IS ASSUMED TO BE TOTALLY RELIABLE. A LARGE VALUE INDICATES THAT A SIGNIFICANT REDUCTION IN THE CORE DAMAGE FREQUENCY / RISK IS POSSIBLE BY IMPROVING THE RELIABILITY ASSOCIATED WITH THAT EVENT. RISK INCREASE: OPPOSITE OF RISK REDUCTION. A LARGE EFFECT INDICATES THE IMPORTANCE OF MAINTAINING THE RELIABILITY OF THE SPECIFIC EVENT AND NO T LETTING IT GET WORSE. .
O o e ( ) LITERATURE REVIEW RESULTS COMMENTS ON OTHER STUDY REVIEWED INSIGHTS GAINED PERSPECTIVES NEEDED NUREG-1150 Reactor pressure sensors are - Little done on specific important to safety from a PRA instrumentation and sub- ( (ASEP) component modeling perspective and are in a location that could see harsh environments during many - Considered only core damage accident sequences frequency and no risk High Pressure Core Spray Pump - Did not consider equipments seals are important to reactor potential for seeing a safety and have the potential harsh environment for seeing a harsh environ-ment during several accident sequences NUREG-4144 Both PWR and BWR power operated - No consideration of equip-(Aging Considerations) relief valves are important to ment location in terms of reactor safety and could see a environment potential harsh environment during many accident sequences - No consideration of total risk BWR relief valves are - Only one plant NSAC/36 (Importance ranging for important to reactor safety E.Q. Prioritization) and could see a harsh enviro- - No direct consideration of ment during many accident equipment location in terms sequences of environment I
i O O O I ( h LITERATURE REVIEW RESULTS COMMENTS ON OTHER STUDY REVIEWED INSIGHTS GAINED PERSPECTIVES NEEDED NUREG-3762 Power operated relief valves - No consideration of equip-(Identification of and associated motor operated ment location in terms of Significant block valves are important to environment potential
, Contributors to safety and could see a harsh } Core Damage) environment in many accident - No direct risk increase /
- scenarios risk reduction considera-
- tion
{ Solenoid operators for PORVs i and containment isolation - No concept of total risk valves are important to safety core damage only . 4 and could see a harsh environ-ment in many accident scenarios i Limit switches, connectors, and , cabling are important to safety and could see a harsh environ-l ment in many accident' scenarios NUREG-1440 Many instrumentation groups, - No explicit ranking of (Light water reactor some of which are not covered importance l instrumentation by regulatory guidance are 4 importance ranking) important to reactor safety - No risk increase / risk
- and have the potential for reduction calculations I
harsh environment exposure ! - No direct consideration of potential environmental impact based on location
- No consideration of total risk (core melt only)
O e e r h LITERATURE REVIEW RESULTS COMMENTS ON OTHER STUDY REVIEWED INSIGHTS GAINED PERSPECTIVES NEEDED NUREG-4537 Solenoid operators for primary - Only one plant considered (Electrical Equipment pressure control relief valves Performance in and main steam isolation - No total PRA perspective Accidents) valves are important to safety in terms of risk increase and could see a harsh environ- or risk reduction for core ment in many accident scenarios damage I - A limited number of accident sequences considered EPRI NP-5081 Historically there have been a - Examined only one event I (Signal validation number of cases of common with one PRA i techniques) cause failure of vessel level
- instrumentation. This - Examined function unlikely instrumentation is similar to see environment stress to other important reactor during performance time-instrumentation groups frame Pointed out that much of the - Examined a function that instrumentation uses D/P as is frequently tested to a means of sensing this insure capability reducing diversity l Pointed out several sensors share a common sensing line reducing the level of redundancy l
e - ^ (
~ ~
h LITERATURE REVIEW RESULTS COMMENTS ON OTHER
, STUDY REVIEWED INSIGHTS GAINED PERSPECTIVES NEEDED SAND 87-0184 Reactor pressure system, low - No direct consideration of (Component monitoring pressure recirc sump, and potential environment for plant life accumulator outlet MOVs are impact based on location extension) all important to reactor safety and inside containment - No consideration of equipment operability time i
frame in terms of equip-
, ment risk due to environ-ment
- Examined only one PRA TAP-A45 Instrumentation for safety - Limited scope PRA (Decay Heat Removal injection system actuation perspective (no large LOCA
, Unresolved Safety Issue) is important to reactor safety or ATWS) i' and could see a harsh environment in many accident - Simple containment model scenarios makes risk impact compari-i sons difficult i High and low pressure recirc.
system suction MOVs are - No direct consideration ! important to safety and could of equipment location 1 see a harsh environment in with respect to enviro-l many accident scenarios ment
i G e e ( 3 l LITERATURE REVIEW RESULTS i i 1 j COMMENTS ON OTHER GTUDY REVIEWED INSIGHTS GAINED PERSPECTIVES NEEDED 1 ) IDCOR 17 Containment fan motors, 600-V - No perspective on increased j (Equipment Survivability power cables, H sonitors, failure rates due to j in a Degraded Core pressure transaktters, and environmental stress at or i Environment) splice / penetration assemblies below qualification levels ' are important to safety and could see an adverse - No PRA perspectiva in terms environment in excess of of effect on core damage or qualification levels in several total risk l accident sequences 1
- Only a limited number of i
dominant accident sequences considered l 4 i
l O a
)
g ! MAJOR ISSUES FOUND FROM i' ! LITERATURE REVIEW l l - i i . PWR PORVs AND BLOCK VALVES , I
. BWR SRVs 6
! . BWR VESSEL PRESSURE AND LEVEL SENSORS , ; j ! MSIV SOLENOID OPERATORS
. PWR AND BWR CABLES AND CONNECTORS FOR CONTAINMENT AND VESSEL INSTRUMENTATION '
i
. PWR CONTAINMENT HEAT REMOVAL SYSTEMS ,
l ( J I
0 O O ( m. I LITERATURE REVIEW
SUMMARY
IDENTIFIED PbRVs, SRVs, MSIVs, PRESSURE AND LEVEL SENSORS, AND CONTAINMENT HEAT l REMOVAL SYSTEMS AS POTENTIAL E.Q. SIGNIFICANT l I EQUIPMENT l
. POINTED OUT NEED TO DEVELOP A PROCESS THAT
! FOCUSES ON IDENTIFYING RISK SIGNIFICANT EQUIPMENT WITH E.Q. POTENTIAL AS PAST WORK HAS NOT BEEN DESIGNED TO ACCOMPLISH THIS l
~
r h TASK 1 l
- FORMAL PROCESS FLOW -
l DEFINE ACCIDENT DEFNE ENVIRONMENTAL CATEGORIES BOUNDARIES
'r ir DEFINE HAZARDOUS I ENVIRONMENTS FOR THE ENVIRONMENTAL BOUNDARIES FOR EACH ACCIDENT CATEGORY ,7
,7 LIST EQUIPMENT UST EQUIPMENT IWORTANT TO SAFETY s . SUBASSEMBLY LOCATION PRA MODEUNG FOR EACH /
BASED ON ENVIRONMENTAL KNOWLEDGE ACCIDENT CATEGORY BOUNDARIES 5r 1r INPUT FROM FOR EACH ACCIDENT CATEGORY EXPERTOPNION DETERMINE EOUIPMENT SUBASSEMBLY ON POTENTIAL ENVIRONMENTAL EXPOSURE E.O. ISSUES BASED ON LOCATION 2 5r DOES PRA CONSIDER ENVIRONMENTAL O #8 SAVE AS A NO ' YES 4 POTENTIAL EO. O STRESS FOR THIS > A POTENTIAL ISSUE EQUIPMENT ISSUE FAILURE RATE
O
)
( TASK 1 OVERVIEW PROCESS FLOW-r- DEFINE ACCIDENT L DEFDJE EMW:#diTAL J CATEGORIES : BOUNDAPTEE M-
'r '
r DEFINE HAZARDOUS ENVIRONMENTS FOR THE ENVIRONMENTAL BOUNDARIES FOR EACH f ACCIDENT CATEGORY ,7 l ,, UST EQUIPMENT LIST EQUIPMENT IMPORTANT TO SAFETY g , SUBASSEMBLY LOCATION PRA MODELING FOR EACH / BASED ON ENVIRONMENTAL KNOWLEDGE J ACCIDENT CATEGORY BOUNDARIES l i INPUT FFIOM FOR EACH ACCIDENTCATEGORY EXPERT OPINON DETERMINE EQUIPMENT SUBASSEMBLY ON POTENTIAL ENVIRONMENTAL EXPOSURE E.O. ISSUES BASEC ON LOCATION
^
1 l 1f l DOES PRA CONSIDER l DGCARD AS SAVE AS A YES
---> O NO / ENVIRONMENTAL STRESS FOR THIS & A POTENTIAL POTENTIAL E.O.
EQUIPMENT (SSUE ISSUE FAILURE RATE
O (
-)
TASK 1 METHODOLOGY
- DEFINE ACCIDENT CATEGORIES -
~
RATHER THAN SPECIFIC ACCIDENT SEQUENCES, FOUR GENERIC ACCIDENT CATEGORIES WERE CONSIDERED FOR THE ENVIRONMENTS ASSOCIATED WITH THEM
- CONTAINMENT COOLED, VESSEL COC_dD l - CONTAINMENT FAILS, VESSEL COOLED l - CONTAINMENT COOLED, VESSEL FAILS l
i
- CONTAINMENT FAILS, VESSEL FAILS l
1 ! ( ) l
-w .- -- .. . - _ - - __ - - - - _ _ - _ - _ _ _ _ _ _ _ _
( 3 0 em 58 db i'
"s!
sLLI ' 7 . lil nll. is . sl!! 80 m 5S l ,E oc3 Lwl
';@ a m 5:
s 3 LU u 5 se 3
> co !@El $
O [3
,_ <a l[llh r
g! g l4l! l*
<e e
!"lgl!gmu 8
g gh g --+
$ o!h b y) c. 00 !!=
(' U $ @
> lfs nu v
- s8
!18 a
<s Elgg h ll!
1 ahhn ' l llis . e ( ) l l'
O e e f h TASK 1 METHODOLOGY '
- DEFINE ENVIRONMENTAL BOUNDARIES -
FOUR LOGICAL PLANT AREAS WERE DEFINED BASED ON SEVERITY OF POTENTIAL ACCIDENT ENVIRONMENTS INSIDE THE PRIMARY BOUNDARY INSIDE PRIMARY CONTAINMENT INSIDE THE REACTOR BUILDING INSIDE AUX BUILDINGS (OTHER) l . Y ]
h
- ~ GE N
L G ED D LF O MW ' O AN RK P L L ^ AI NA TE OT NU
^ TES I N TAT E Ot S
NC M S P EO N E I MLOI A P YIRR I A L ULV D T A 7 QBN EMEU N - N T ENO - W E-E M ES NI OR RA S SOB I S LAD BE US S D SA r, Y E INN B I W V U ,' L B Y V O EO YM EB REE RL N I F E E H H CY OSR GSU EASN EF D S U T LA R O R AEO D OTR G TBOO AU PT CSXA I LS AIH T ARN T T VSS R ASE Z FN OT TM SC FA E g/ TTEC NNLO EEAL !e REER EN Rs PDAO MUT? E OE ANN r i OMTN N SSNF PLA HEOI ET ENOSIUA R , I CP EO OORS m ' E MnRN CIU AQ MD DCI VE QF 1 C N O F NiAE V DDI I R fN N C HEOS NE NT ES RE O EI DV N OA UC C AER A KR EI NIV B T E B RMN N ORE SP ES DE I I CR E FE T E A- CO AG D m T - EET Y- Y. < N I A T R
- FC T FE O e..
E N A' G D E S'H E 7, M I TOCT P AA U T EG O QNRT AE E AO N SLE TTF SR OE AAI U UO .C ETS P C VNS M A AEI I SO T P N L - l M O AS S - OMiE DPNU I l FOES TS - T TO UR I. P EPO ~ - lNPf4E XO - E - O ( - ( \
O a . e l ( h l TASK 1 METHODOLOGY ! - LIST EQUIPMENT IMPORTANT TO SAFETY - i i i
- BASED ON SEVERAL FACTORS i
l 1 - EPGs PLANT KNOWLEDGE l PRA PROCESS l i I ( _ J i 1 .
\ O r 3 TASK 1 METHODOLOGY
- LIST EQUIPMENT IMPORTANT TO SAFETY -
s EXAMPLES
- 1. Containment Holds, Vessel 1. Electric and Mechanical Scram Cooled 2. Rod Position Indication
- 3. SLC (for ATWS events) l
- 4. PPI's l
- 5. SRV l
- 6. ADS
- 7. High Pressure Cooling Systems
- 8. Low Pressure Cooling Systems (Incl. RHR)
- 9. CRD
- 10. Cond/FW
- 11. RHRSW
- 12. RWCU (isolation during ATWS)
- 13. Fire H2O
- 14. MSIV position - control and indications
- 15. Service H2O
- 16. CST
- 17. Turbine / Steam Dumps
- 18. Drywell Systems (Coolers, fans, etc; isolation valves) and Atmospheric Indications (temperature, pressure)
- 19. SP temperature, level
- 20. DG Systems, electrical systems and
. associated support systems
- 21. MVAC l
- 22. ESW
O e -
~
( 3 TASK 1 OVERVIEW PROCESS FLOW - DEFINE ACCIDENT DEFINE ENVIRONMENTAL CATEGORIES BOUNDARIES 1r ir
' DEFINEHAZARDOUS ENVIRONMENTS FOR THE ENVIRONMENTAL BOUNDARIES FOR EACH
,7 ACCIDENT CATEGORY ,
7 M UST EQUIPMENT UST EOttlPMENT IMPORTANT TO SAFETY % . SUBASSEMBLY LOCATION PRA MODEUNG FOR EACH / BASED ON ENVIRONMENTAL KNOWLEDGE ACCIDENT CATEGORY BOUNCARIES ! I l i NPUT FRCSA FM EACH ACCsDENTCATEGORY i I EXPERTOPNION DETERMINE EQUIPMENT SUBASSEMBLY l ON POTENTIAL ENVIRONMENTAL EXPOSURE l E.O. ISSUES BASED ON LOCATION
'V DOES PRA CONSIDER SAVE AS A DISCARD AS
--+ POTENTIAL E.O. I NO / ENVIRONMENTAL STRESS FOR THIS YES
- A POTENTIAL ISSUE EQUIPMENT ISSUE FAILURE RATE
O
~
( 3 TASK 1 METHODOLOGY
- HAZARDOUS ENVIRONMENT DEFINITION BASED ON ENVIRONMENTAL BOUNDARY AND ACCIDENT CATEGORY-BASED ON KNOWLEDGE OF THERMOHYDRAULIC CODE PREDICTIONS, PREVIOUS STUDIES, OPERATIONAL l EXPERIENCE, AND EXPERT KNOWLEDGE EXAMPLE AREA 'I !! III IV LOCA ACCIOUli SCENARIO l I
cent i t a l n t.r.n.H.I none non. Prs =ry v==i a i 1 l I Centei -mat er.ech N t.r.M.H.I t.P.M.H.I Wan. Pri ry v.=.i a l t l l ENVIBONMENT KEY t - Elevated Temperature H - High Humksty T - ExtremeElevationinTemperature R - High Radletion P - ElevatedPressure I - Impingement bySteam orWater M - MoctionicalShocks H2 - Hydrogen Generation
w as Ile b 8E B e g8 o< ig v 8 e h " Ug8!
!&!f l5l 's! ;
>o CC .a 85 g8 n
I v i LLI 1
>m s - g a
s Em5 til /9 g O@ l 's m , lgd 2 8 isl!lgg_ E$ = !ll llll I~' ea s. gie l <' Ig E a i h ie g- ,, s8 s! 2Is a ga <s sii l2l! 1 ss W lis O
- ~
( . _ j
,-,---.y.-- , - -. . , - - - , m v., --w-- - , - , - - - , - - - ,,, - -n.a.----, ,, - ---- -
-~ ~
O
~
F h l TASK 1 METHODOLOGY
- LIST EQUIPMENT SUBASSEMBLY LOCATION BASED ON ENVIRONMENTAL BOUNDARIES - _
BASED ON PLANT SCHEMATICS AND FSAR
. EXAMPLE
=
I II !!! IV EQUIPMEllT/5YSTEM (PRSAARY) (CONT) (Rx BLDG) (AUX BLDG) Pl .itu.CV.Vs.lWF. W1.Pu.Mo.in. PS. E5V I m . m .vi.ny. , v..rs.cv.a..P. l l l l EQUIPMENT / COMPONENTS l Pu - Pump Po - Power Supply Mo - EAotor B - Electrical Harhare i Va - Valve Hy - Hydraulic Actuation (alr, oil,H2) WI - Wire / Wiring CV - Check Valve Ta - Ta.WAccumulators MOV~ Motor Operated Velve l De - Detector /Sermor Hx - Heat Exchanger l In - InstrumentettorMndication SOV- Solenoid Operated Velve Am - Amplifiersand LoDic CircuMry ADV- Air Operated Valve / Damper i Pl - Piping l . 1
O e _
^
F h TASK 1 OVERVIEW
- PROCESS FLOW -
i __ DEFINE ACCCENT DEFINE ENVIPONMENTAL CATEGORIES BOUNDARIES I 1 ir DEFINE HAZARDOUS 3 ENVIRONMENTS FOR THE ENV:RONMENTAL t BOUNDARIES FOR EACH y ACCIDENT CATEGORY ,y j UST FOUIPMENT LIST EOutoMENT l ' IMPORTANT TO SAFETY l 3 ', SU8 ASSEMBLY LOCATL% PRA MODELING FOR EACH a
/ 3ASED ON ENVIRCNMENTAL KNOWLEDGE ACCJDENT CATEGORV BOUNGAKES l
iP ' E i GNPUTFROM . FOR EACH ACC8 DENT CATEGORY - EXPERT OP;NION CETERM;NE EQUIPMENT SU8 ASSEMBLY d Ord POTENTIAL ENVTRONMENTAL EXPOSURE I E.O. ISGUE3 BASED ON LGCA'OON . N M m,
=
1P
, /DOES PRA SAVE AS A
/ CCNSIDER DtSCARD AS
-+ < m A - ENVIRONMENTAL ygg POTENTIAL E.G. STRESS FOR THIS & A POTENTIAL l
ISSUE EQUIPMENT . ISSUE l
. FAILURE J
RATE 7 2
- O _
=
2 TASK 1 METHODOLOGY ! - EQUIPMENT QUALIFICATION ISSUE SCREENING - i l SCREEN ON PREVIOUS RESULTS ! OVERLAY PRA INSIGHTS - KEEP MOST SIGNIFICANT RESULTS SOLICIT EXPERT OPINION l'
O a - 3 I( FORMALIZED SCREENING PROCESS EXAMPLE SYSTEM COMMINENT APEA/[Irfit SAFETY FUNCll0ll/ INTERFACES EQ POTENTIAL /PRA INSIGHTS A. [lectric None I-II/ N/A Place reactor in a subcritical. No EQ losues are evident, a scram should occur and safe condition. prior to the occurrence of a significant EQ type ( leechanical environment. l I Scram and l Rod Pos. Indication C. PPI's 1. Vessel Level and II/t.H.M.I Provides vessel level input Current Pila models do not typically account for Pressure Detector signals to all high/ low flashing in detector piping. Core theruel conditions, j - piping pr m ure Injection an automatic condensation of stems on piping lepingenent or depressertration systems, and mechanical shock could cause thermal gradientsin operator indication of vessel detector piping and me/ result in periods of tuo level. phase flow isading to erroneous detector Indicaties and possible prevention of access to les pressure injection systems. As core t; 2 ,- M c conditions approach saturation temperatures and pressures, tuo phase flow may result. The concern here is not detector piping failures but erroneous detector Indicstlen or detector failure hosed en the environments the piping is emperiencing. 9 L _ _ _ _ _ _ _ _ _ _ -
S &
^
f
~ ~
3 FORMALIZED PROCESS RESULTS OPEP.ATIONAL ANALYSIS ISSUES
- l BWR issues PWR issues i
- LOWPRESSURESENSORS CONTAlffENT FANS
- ItttnNCE LEG DETECTOR PIPIflG CONT'TSPRAYPlfPS/?MORS j (PRESSURES, LEVELS)
(INSIDECONTAlfFOR) i i SAFETYRELIEFVALVES PRESSURIZER IB/El> PRESSURE INSTRU. . INB0MD MSIVs & BYPASS VALVES I R)RVs & fMOR-0PERATED BLOCK VALVES CONTAlif0fT ISOLATION VALVES CONT'T PRESSLE INSTRTENTATION ), ) HIGi PRESSLE INJECTION PltFS STFliM GEN. LEVEL Itit.CIURS LOW PRESSURE INJ. PLPP fEORS PRIl%RY SYSTEM RIDS i LOW PRESSlE INJ. tels CONTAltfBITISOLATICHVALVES HIGH PANGE PADIATION IDilTORS HYDROGENititCTURS llIGH RANGE RADIATIGiIUlITORS Operational Analysis to include Consideration of Applicable Penetrations, Splices, and Cabling.
. 4 O
EQ-RISK SCOPING STUDY: GENERAL CONCLUSIONS / RECOMMENDATIONS o O
4 GENERAL CONCLUSIONS g
- 1. PRA accident timing perspectives provide a basis j for modifying EQ practices.
- 2. PRA perspectives suggest that the importance of the accident radiation dose is overemphasized in EQ testing.
- 3. When PRA perspectives are combined with the existing EQ perspectives, several historicai EQ ,,
issues can be considered not risk significant. l O l 4. When PRA perspectives are combined with existing EQ perspectives, several historical EQ issues have possible risk significance. I O
l GENERAL CONCLUSIONS (Continued) O
- 5. The lack of information regarding the reliability of certain components in accident environments creates non-conservative biases in important accident sequences and base case core damage frequencies.
- 6. There are risk important system operations that rely on non-safety-related equipment or on equipment that must function during non-DBE environments.
- 7. The risk importance of plant status equipment is O difficuit to quantity.
- 8. The risk importance of certain EQ issues could not be characterized by the Scoping Study using existing PRA models and EQ data bases. .
O '
l GENERAL CONCLUSION #1: PRA accident timing perspectives provide a basis for modifying EQ O practices. ) i Conclusion from PRA Core damage frequency and risk are substantially dependent on equipment operability during the firm few hours to days of an accident sequence. l EQ imolications
- Less emphasis on demonstrating long duration accident equipment operability.
O f Certain EQ issues are not risk significant, for example: Correctness of post-accident acceleration techniques. Use of simultaneous versus sequential accident simulations. Importance of oxygen within the LOCA test chamber. O i
.._- --.-__-_.__ . _- _L
GENERAL CONCLUSION #2: The importance of the O accident radiation dose is overemphasized. l P_BA Perspective PRAs do not consider substantial in-containment radiation conditions to exist until core melt has occurred. Core melt doesn't start until half an hour to several hours after accident sequence , Initiation. O PRAs calculate that risk significant containment failures will occur within a few hours to a few days. EQ Perspective EQ requires all equipment to survive an instantaneous release to the containment of part of the core inventory. EQ radiation specifications are typically based n I n9 term a ident exposures. O
GENERAL CONCLUSION #2 (Continued) g EQ Implications The following EQ issues are impacted. The importance of simultaneous versus sequential accident simulation techniques. The importance of post-accident radiation dose , rate effects. The significance of beta versus gamma G effectiveness in producing equipment damage. Use of realistic accident environments during equipment qualification. l l
- l
l . . GENERR CONCWSION #2 (Continued) O Steam generator level detectors are an example where the current radiation overconservatism may impact risk adversely. O O
GENERAL CONCLUSION #3: When PRA perspectives are g combined with the existing EQ research perspectives, several historical EQ lssues can be considered not risk significant. l Examples:
- 1. Simultaneous versus sequential application of accident radiation, steam, chemical spray, and pressure conditions.
- 2. Accident dose rate effects.
- 3. Beta / gamma equivalence. $
- 4. Oxygen presence within the test chamber during LOCA simulations.
- 5. Correctness of post-accident simulation techniques.
I O
l O GENERAL CONCLUSION #4: When PRA perspectives are combined with existing EQ perspectives, several l historical EQ issues have possible risk i significance. l l Examples:
- 1. Adequate sealing / protection of safety-related circuits from moisture intrusion / condensation effects.
- 2. Use of air versus nitrogen process gas for solenoid operators.
O
- 3. Choice of ambient environments as a basis for equipment qualification.
- 4. Humidity aging effects.
- 5. Choice of accident environments for equipment qualification.
- 6. Use of alternative test approaches (multiple sample reliability assessment and/or fragility l testing) to complement current qualification l type testing.
l 0
GENERAL CONCLUSION #5: The lack of information g l regarding the reliability of certain components in accident environments creates non-conservative biases in important accident sequences and base case core damage frequencies. EQ process rarely supplies accident equipment reliability Information. PRA analysis techniques possibly underestimate the importance of harsh environment equipment operations. O Risk achievement analysis is typically performed only one component at a time; accident environments can affect multiple components. Using normal operation reliabilities, PRAs l may screen out quallfled equipment prior to performing risk achievement analysis. ! Risk achievement analysis is not consistently performed and reported for g "backend" PRA analyses.
GENERAL CONCLUSION #5 (Continued) O Two sets of sequences were identified where degraded harsh environment equipment may impact PRA perceptions. Transient induced or small break LOCA PWR sequences. BWR transient sequences with loss of suppression pool cooling. O O
GENERAL CONCLUSION #6: There are risk important g system operations that rely on non-safety-related equipment or on equipment that must function during non-DBE environments. Examples: PORV solenoids and block valve motor operators during transient induced and small break LOCA environments. HPCS pumps when subjected to high temperature working fluids. O SRV solenoids, MSIV solenoids, and MSIV bypass valve motor operators during TW sequences. Low pressure injection equipment after BWR containment rupturing or venting. O
GENERAL CONCLUSION #7: The risk importance of O plant status equipment is difficult to quantify. Plant status instrumentation not well modelled by PRA fault trees. PRAs assume plant status information is reliable and base actuation / failure response on human error. O O
O EO-RISK SCOPING STUDY: GENERAL RECOMMENDATIONS O O
GENERAL RECOMMENDAMNS: , O
- 1. Development of accident reliability data for selected components and accident sequences.
- 2. Additional definition of equipment risk -
importance.
- 3. Redefinition of EQ regulatory basis.
- 4. Improved guidance regarding operations, maintenance, and inspection activities.
O 5. Improved guidance regarding equipment aging issues.
1 GENERAL RECOMMENDATION #1: Development of harsh g environment reliability information for selected equipment. Examples: 1 SOV reliability during PWR small break LOCA and BWR TW steam / pressure conditions. MOV reliability during PWR small break LOCA and BWR TW steam / pressure conditions. Steam generator level detection reliablllty during small break LOCA steam / pressure @ conditions. Low pressure in,ection equipment reliability during steam / temperature conditions after BWR containment venting and rupturing. HPCS pump reliability during high suppression pool temperature conditions. O
GENERAL RECOMMENDATION #2: Addlilonal definition of O equipment risk importance. Equipment important to core-melt-arrest and post-core-melt accident management strategies. Plant status instrumentation. O O mi i um
1 GENERAL RECOMMENDATION #3: Redefinition of EQ O regulatory basis. Examples: Instantaneous radiation release requirement. Equipment operability time requirements. l O 6
l O GENERAL RECOMMENDATION #4: Improved guidance regarding operations, maintenance, and inspection activities. i Examples: Priority for steam generator level detection and high range radiation monitor circuits. l
- Priority for SOVs and MOVs used for BWR SRVs and MSIVs and PWR PORV applications.
O l l O
GENERAL RECOMMENDATION #5: Improved guidance g regarding equipment aging issues. l l Examples: Aging assessment and/or condition monitoring of batteries. Humidity aging effects on electrical penetrations with polyimide insulation. O i . O
l O O O 4 ? I i ) ADDITIONAL EXAMINATION OF SELECTED EQUIPMENT OPERATIONS i i l .
ADDITIONAL EXAMINATION OF SELECTED EQUIPMENT OPERATIONS O PWR Components Selected for Further Analyses:
- 1. PORV components (SOVs and block valve MOVs)
- 2. Steam generator level detection transmitters
- 3. Containment fans
- 4. High range radiation monitors BWR Components Selected for Further Analyses:
- 1. SRV and MSIV solenoid operators O
- 2. MSIV bypass valve MOVs
- 3. High pressure injection pumps l l
- 4. High range radiation monitors
, Analysis of each equipment operation to include consideration of applicable penetrations, splices, and cabling.
O STEAM GENERATOR LEVEL DETECTORS AND PORVS (WITH ASSOCIATED BLOCK VALVES) Small Break LOCA: Multiple system, common-cause failure rate for steam generator level detection and PORV operability must be less than 1 E-3. therwise, current perceptions regarding O important core damage sequences and base case core damage frequencies may need to be modified. O . d
,w -,-- --. - - - - , , , ,%-. c.--,w--,,.-,,-,-,,wy,-.r.,v,,
IMPACT OF FEED AND BLEED AND AUX FEEDWATER FAILURE RATES ON SURRY CORE DAMAGE FREQUENCY Feed and Bleed Failure Rate g' 3 6E 2 Qd M es E b!
;--l 3.5E 5 4.5E 8 1.3E 7 3.8E 7 14 0.001 .1.3E 6 3.6E-6 1.1E 5 if G D 0.010 1.3E 5 U;E 5 1.1 E 4 -
'E s 0.100 1.3E 4 3.6 E-4 1.1E 3 (i
Surry Base Case Core Damage Frequency = 2.5E 5 Table entries represent additive contribution of S3 LP3 , S2LP3 ,T 20LP ,and 3 T 3MOLP 3sequences t to core damage frequency. O L .. . . . . ..
O SteanLGaugrator Level D31eclors and PORVs (With Associated Block Valvesl The PORV "feed and bleed" function relles on solenoid operatoi's and motor operators that: l a. Historically were not included on utility EQ master li.sts.
- b. Have little redundancy.
O c. Have been the focus of NRC Information notices, inspection findings, and research programs.
- d. Lack data to quantify reliabilities during accident conditions.
O
i O Steam Generator Level Detectors and PORVs (With , Associated Block Valves) The steam generator level detection circuits have substantial redundancy. These circuits would be sensitive to commcn-cause failure resulting from moisture condensation or moisture intrusion. O l I O
l l O Activities to reduce uncertainty regarding PORV feed and bleed operation: Assess SOV and MOV reliability during small break LOCA steam / temperature environments. Monitor / assess the long term aging behavior of l SOV elastomeric components. Address various operations / maintenance issues. O l l l l lO l
Activities to reduce uncertainty regarding steam $ generator level detection: 1 Develop estimates regarding the accident reliability of steam generator level detection circuits. Assess impact of level detection failure on reliability of auxiliary feedwater. Insure that terminal blocks are not employed in these circuits wherever steam conditions might occur. O Determine the impact of humidity aging on "KAPTON" electrical penetration feedthroughs (if they are employed in SG level detection circuits). Address various operations / maintenance issues. I O
O Steam Generator Level Detectors and PORVs Several historical EQ issues are not risk significant. Differences between simultaneous versus sequential accident simulation techniques. Presence of Oxygen (test chamber environment) during accident simulations. l l Beta / gamma equivalence. O j Post-accident acceleration techniques. t O
I p Steam Generator Level Detectors and PORVs Additional study is warranted to determine whether the current radiation EQ requirements impacts risk adversely. Example: steam generator level transmitters ; O O
O s_ws_ safety-Reiief valves and uain Steam isolation Valves TW Sequence: Multiple system, common-cause failure rate for MSIVs and SRVs must be less than 1E-3. Intermediate Break LOCA Sequence: SRV common-cause failure rate must be lower than O 1 E-1. Otherwise, current perceptions regarding important sequences and base case core damage frequencies may need to be modified. O
IMPACT OF SRV AND MSIV COMON-CAUSE FAILURE
- RATES ON PEACH BOTTOM TW SEQUENCE FREQUENCY i
i ! PEACH BOTTOM CDF = 8.2E-6 i SRV FAILURE RATES I (COMON-CAUSE) 4E-3 1E-2 1E-1 1.0
- h 7E-3 5.2E-8 1.3E-7 1.3E-6 1. 3'Z-5
\
# a, u
"g 1E-2 7.4E-8 1.9E-7 1.9E-6 1.9E-5 l 3 i da 5 g 1E-1 7.4E-7 1.9E-6 1.9E-5 1.9E-4 Eo E~ 1.0 7.4E-6 1.9E-5 1.9E-4 1.9E-3 O O O .
. 1 O BWR Safety-Relief Valves and Main Steam isolation Valves Both the SRVs and the MSIVs have substantial redundancy. However, research testling and NRC inspection experiences have highlighted concerns regarding the important components, SOVs and MOVs. Reliability testing dur:ng steam / pressure conditions O is warranted. O
BWR Safety-Relief Valves and Main Steam ) l Isolation Valves Extended radiation exposures not applicable to PRA ' risk significant SRV and MSIV operability. PRAs only interested in a few days of accident , operability. ! Hence, several EQ issues are not risk significant. 1 Importance of simultaneous versus sequential accident simulations. g importance of including oxygen within the LOCA test chamber. Correctness of post-accident acceleration techniques. Equivalence of beta / gamma radiation. l l
O HPCI. RCIC. and HPCS Pumps important during a number of accident sequences. Power loss transients Anticipated transients without scram (ATWS) events Small and intermediate break LOCAs Located outside containment. O ) However, non DBE sequences can produce degrading environments (internal and external) at the pumps. Station blackout and transients events can lead to high suppression pool temperatures. (The suppression pool is a suction source for the pumps.) l Station blackout can produce high HPCI and RCIC room temperatures. 1 O
1 HPCI. RCIC. and HPCS Pumps O PRAs currently assume failure of the pumps when suppression pool temperatures exceed 210-2400F. Designing and qualifying a BWR/5/6 HPCS pump to operate during higher suppression pool temperatures Would reduce Grand Gulf core damage frequency by 76%. .. O O
O HPCI. RCIC. and HPCS Pumps Risk impact of BWR/4 HPCl/RCIC pump environmental failure is uncertain. Station blackout battery depletion causes failure coincident with any environmentally induced j failures. For TW sequences, risk impact will depend on the j reliability of: O MSIVs SRVS l Low pressure injection equipment (after containment venting or rupturing). l l l ! O
- IMPACT OF SRV AND MSIV COMON-CAUSE FAILURE '
j RATES ON PEACH BOTTOM TW SEQUENCE FREQUENCY i . PEACH BOTTON CDF = 8.2E-6 SRV FAILURE RATES ! (COMON-CAUSE)
- 4E-3 1E-2 1E-1 1.0 s
m i W 7E-3 5.2E-8 1.3E-7 1.3E-6 1.3E-5 \ &C EN 1E-2 7.4E-8 1.9E-7 1.9E-6 1.9E-5 ! 5 't l 25
- u. IE-1 7.4E-7 1.9E-6 1.9E-5 1.9E-4 g
= m E~ 1.0 7.4E-6 1.9E-5 1.9E-4 1.9E-3 O O -O .
i O HPCI. RCIC. and HPCS Pumps Current EQ specifications that require 100 day operability and include a radiation environment are not risk significant. Small risk impact for HELBs with failure to isolate steam line (to HPCI, RCIC, or RWCU systems). Initiation pipe break frequencies are less than 1E-6. O l
- O
Hiah Rana _e Radiation Monitors O l I One of the few indications available to the operator indicating that core melt is occuring. Indication of core melt may (?) be important towards emergency response decisionmaking. High range radiation monitors lilustrate the difficulty in assessing the risk importance of plant status instrumentation. Surry draft NUREG-1150 analyses p'rovides risk impact g associated with early versus late evacuation. O
g AOBABUTY CF EXCEEDNG 200 REM WH BODY O .o , EAR'v Ccta ruera sit'as
,o _M
- 1. COtRNJE NORMAL ACTMTY 70 ~
- 2. BASEMENT 9-ELTER 1 START EVACUATION 1 FH AFTER SELEASE 80 -
- 4. START EVACuAT)CN .5 FRS BEFCRE REL 80 -
N e cases esans av se.ocArno mm CCNT4.euTED NwA5 6 >R$ MOM PE.EXE 4o - so - so - l 2 l 1o --
+
g 3
, , , . m N . > . , , . I 1 Vi_E 3 MLES 5 MLES 10 AtES !
(' D! STANCE FROM REACTOR g PROSASUTY 0: EXCEEDNG 50 REV WH BODY O 90 4 2 2 80 - 7o - 2 1 6o - So - l 40 - l 3o _ l 20 - 1 to 9 2 3 $ ! , , , , , , i , , 0
- 1 MLE 3 MILES 5 MILES 10 MLES DISTANCE FROM REACTOR Figure 10.3 Effectiveness of emergency response actions: early O containment failure (Surry)
i { 190 l AVERACE PH BODY DOSE TO INDIVIDi'AL IN MID PLllME 100 l b l 90 - EARLY CONTANAENT FAllBE i 2 80 - i
- 1. CONTINUE NORMAL ACTMTY 70 - ;
- 2. BASEMENT SHELTER
- 3. START EVACUATION 1 FR AFTER FIELEASE
! 60 - ~ ^ N AU. CASES PmSONS ARE RELOCATED MOM t 2 50 - NAMNATED AREAS 6 HRS FROM RE1 EASE l' l g- 40 - . , 2 l 8 - 30 - l I . 3 1 20 - 10 - 4 2 1 4 g3 o 3 MILES 5 , , 5 MILES
, agm , ,
10 MLES 4 W 2 3 20 MILES 4, DISTANCE FROM REACTOR Effectiveness of emergen esponse actions: early containment 9 failure (Surry) (continu
O O o . 25 AVERAGE WH BODY DOSE TO INDIVDUAL ("in etunc) 1 LATE CONTANAENT FAR1FE 20
- 1. CONTINUE NORMAL ACTMTY
- 2. BASEMENT SHELTER
- 3. START EVACUATION 1 FO AFTER RELEASE
- 4. START EVAQJATION .5 Ff1S GEFORE REL
_ 15 N ALL CAES PERSCNS AFE FELOCATID FRCN b COtJTAh*4ATTD AREAS 6 FRS FROM FEEASE O w E 2 10 1 5 3 2 I* 4 2 3 1 4 3 4 2 3 4 0 5 MLES 10 MLES 20 MILES 3 MILES DISTANCE FROM REACTOR Effectiveness of emergency response actions: late containment f ailure (Surry) (cont.inued)
, ROBABL TY OF E'/CEEDNG 200 REM WH BODY LATE CCMAne T FAL'.RE
- 1. CCtfTNJE NORMA ACTNTTY l
- 2. BASEMENT SHELTER !
so - 1 START EVACUATION 1 m A-M CE. EASE E 4. START EVACLATKN .5 PS BEFORE REL h 1 N au. curs sescrs Am: m_ccAvio mm ccenmTio Am.As e e mm m.i.Ase a - so - 2 3 4 ' ' ' ' ' ' ' ' 0 10 MLES 1 MLE 3 l4_ES 5 MLES DISTANCE FROM REACTOR
,g PROBA9UTY OF EXCEED:NG 50 REM WH BODY 90 -
1 80 70 CC 5 SO E 3 r va - so 2o 1 i to - 4 2 3 4
- , , , m_ , , , , ,
o u i , 1 MLE 3 MLES 5 MLES 10 MLES DISTANCE FROM REACTOR Effectiveness of emergency response actions: late containment failure (Surry)
l O sigh Range nadiation uonitors At time of core melting, containment pressures, temperatures, and radiation conditions are weil bounded by typicai qualification parameters. l l Several DBE EQ issues that could impact radiation monitor performance. Aging dose rate and synergistic effects. adequate spe iri ation of qualification acceptance O criteria based on system requirements. Proper sealing against moisture intrusion. 1 l 1 1 0
\ High Range Radiation Monitors EQ issues that are not risk significant: Importance of simultaneous versus sequential accident simulation techniques. importance of oxygen during LOCA simulations. Correctness of post-accident acceleration techniques. O Adequacy of the radiation source term and radiation simulation techniques There does not appear to be a risk basis for requiring a factor of two accuracy over the range of 1 rd/h to 1E+7 rd/h. O
O P_WR Containment Fans Containment fan risk significance is not large.
- Addition of fans at Surry would decrease latent cancer risks by approximately 10%
Fans impact risk at Zion by approximately a factor of two. O . l 1 l l l o
1 PWR Containnient Fans O Uncertainties regarding fan reliability could be ! reduced by the following efforts. l l Assess fan reliability during and after core melts ! that produce substantial amounts of aerosols. Assess impact of humidity on fan cooler operation. Insure adequate fan cooler maintenance. ! l Assess intensity and impact of multiple hydrogen g burns in the vicinity of fan cooler components. O
a e O pwR Containment Fans Several EQ issues lack PRA risk significance. Importance of simultaneous versus sequential accident simulations. Importance of including oxygen within the LOCA test chamber. Correctness of post-accident acceleration techniques. O l O ~ - _ - - - - _ ._- - - - _ _ _}}