ML20133K744
| ML20133K744 | |
| Person / Time | |
|---|---|
| Issue date: | 08/08/1985 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1440, NUDOCS 8508120347 | |
| Download: ML20133K744 (278) | |
Text
- "**"
ORIGINAL r3 UNITED' STATES OF AMERICA
) NUCLEAR REGULATORY COMMISSION In the matter of:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 304th General Meeting (Public Session)
Docket No.
/ \
'R !
Location: Washington, D. C.
- f. ,
[ ,
Jo Not lemove from ACRSOFice ANN RILEY & ASSOCIATES Court Reportors 1625 I St., N.W.
'~
n500120347 000000 I'DH AcHb finn Suite 921 T-1440 Washington, D.C. 20006 (202) 293-3950
1 1 UNITED STATES OF AMERICA 2 NUCLEAR REOULATORY COMMISSION
(
3 ADVISORY COMMITTEE ON REACTOR SAFEQUARDS 4 304th OENERAL MEETING S
6 OPEN SESSION 7
8 Thursday, August 8, 1985 9 Room 1044 10 1717 H 5treet, N.W.
11 Washington, D.C.
12 The Advisory Committee on Reactor Safeguards met in 13 Open Session, pursuant to notice, at 8:3S 4.m., David Ward NJ 14 (Chairman of the Committee 3 presiding.
15 ACRS MEMBERS PRESENT:
16 D. Ward H. Lewis 17 J. Ebersole D. Moeller 18 W. Kerr M. Carbon 19 H. Etherington G. Reed 20 C. Wylie F. Remlok 21 C. Mark C. Stess 22 P. Showmon R, Autmann 23 ACR8 STAFF MEMBERG PRESENT:
24 Raymond Fraley Norman Schwarts O 25 Rhohard Majors A__
2 1 SPEAXERS:
2 5. Bernero D. Scaletti 3 C. Thomas O. Sherwood 4 D. Hankins D. Foreman S 7. Pratt J. Yease11 5 N. s.nderson H. Booher '
7 T.Y. Chang F. Schroeder s G. Bagohl Mr. Hernan 9 J. Thomas G. Cwelina 10 J. Jankovich D. Brinkman 11 D. Persinko 12
([) eae 14 15 14 17 18 19
'20 21 22 23 24 25
3 i PROCEED ! NO S 2 MR. WARD: We will go ahead now with Agenda Item 3 No. 2, OESSAR-!! report, and Mr, Ebersole.
4 MR. EBERSOLE: Thank you.
5 GESSAR-!! Subcommittee met yesterday, in the 6 absence of David Okrent, and I'm sure you all know how 7 impossible it is to replace him; however, I am giving it a 8 try.
9 We had a lengthy discussion right on up through 10 today. I refer you to Tab 2. The various topics we took up, 11 mainly we were following up Supplement 4 of the SER, and I am 12 just going to briefly mention some of the toptos there that wo 13 did.
14 The first topio we will, in fact, take up for the 15 first time today is discussion by the Staff, by Mr. Bernero, 16 about oore melt frequency, et cetera, and the guidelines or 17 the performance criteria for the review of the new standard 18 plant, with some expression of what are the Staff goals.
19 We went into the SSER No. 4 and looked at the 20 outstanding issues, and again, I noted that when issues are 21 resolved, they disappear, and 1 think that is unfortunate. So 22 we pumped up a few and ampitfied them, and I intend today to 23 take the liberty of expanding the fourth topio, whtoh ta 24 security considerations, whleh is going to be very brief 25 because there are not many, and include within that some
4 1 portion of the Staff consideration of 0.E.'s offer for I
( 2 numerous potential design improvements, of whtoh a few have 3 been adopted and are firm now in the design, and the 4 extensions of the concepts that are being offered, such as 5 seismic capahtlity of the UFFS system.
6 We have here -- perhaps one of the more important l
7 toptos is a dotatted discussion of hydrogen and the scrubbing 8 offloiency of the pool and the hot and cold condition. You i
9 can see on page 2 of the agenda yesterday we took up the 10 matter of detonation in some detail, if one wants to pick 11 further discussion on that.
12 Mainly, however, I think that the more interesting l
13 aspects of this stage of the OESSAR-!! design is looking at 14 the potential design improvements, which of these have been i
15 adopted -- there are very few -- have been adopted out of a !
l 16 great list that was prepared by 0.E., and I would like to l 17 suggest some discussion of the cost basis on whleh they were 18 and were not adopted. ,
19 These were considered solely on the basis of 20 reduction of the dose guideline of $1000 per man rem and did i
21 not include consideration of on-site averted cost, whtoh might have made some of these alterations and improvements more !
22 23 attractive had that been done. !
24 My understanding is this is in a Never-Never Land ,
l 25 state of resolution now, and I think we ought to take it up l i
l
5 1 before we really look up OESSAR-!! to see whether we can O
k-)
s 2 justify extensions in characteristics of some of the offered 3 features or perhaps some others whtoh were not considered 4 because of these costs.
5 As you all know, it is very difficult to get a 6 design improvement identified against simply the criterion of 7 the $1000 per man rem, a With that, unless there are other contributions from 9 other members of the subcommittee, which was Carson Mark and 10 Mr. Etherington and Charite Wylie, I am going to turn this 11 over first to the Staff.
12 MR. KERR: Mr, Eberaole, I could get the impression 13 from your comments on cost-benefit analysis that you have the
(
14 impression that this committee has a firm position on that, it 15 seems to me.
le MR, ESERSOLE: No.
17 MR. MERE: It seems to me that might not be the 18 case.
19 MR. EBERSOLE: No, I think it is a floating matter, 20 but I would just point out I am sure sertain improvements were 21 not adopted because of the limitations of this $1000 per man 22 rem basis.
23 I turn it over to the Staff.
24 MR, BERNERO: I would like to take advantage of my 25 infrequent appearances down here to get up at the podium.
i 6
i For the record, I would 1the to say I am Bob 2 Bernero, Director of the Division of Systems Integration in !
t t
3 the Nuclear Reactor Regulation Offtee, and I am happy to say ;
[
4 that in a short time, I will be the Director Notting Water f i
S Reactor bloonsing, so I have a dual interest in the OESSAR.!! >
4 evaluation since ! expect I will be living with it for a good f
7 period of time, as well as having partiotpated in some 4 substantial measure in the conduct of the review, !
9 It was said by the Chairman just before you started
[
10 as part of the housekeeping announaements that the Commission I I
11 la releasing the Severe Accident Polloy Statement, I think it l 1
13 is fortuttous but interesting that it is today it will appear ;
() 13 in the Federal Register.
For your information, you may stumble for a moment, 14 l ,.
15 not recognising it when it appears, It is actually a dual l le nottee. At the last minute it had to be changed editorially 7 17 so that one nottoo withdraws the October 1980 degraded oore l r
14 rulemaking or severe accident rulemaking, and legally that is 19 one things and then the other thing pubitehed book to back i 20 with it is the severe acoldent policy statement, with whleh !
l
! 21 you are fantilar from very long review, 22 Now, that pontoy statement, as you know, emphasises I
33 attention to future plants such as OKSBAR-l! but also covers 24 estating plants, and in essense, the polley statement has the
.. Commise,on s.ying. from .h.t .. ,now tod.y. th. level e, I
. i
D 7
i i technology and safety assootated with light-water reactors is i !
\ 2 safe enough, with conditions, and those conditions are quite ;
3 signittoant, and we are discussing some of them on this one 4 future plant, OE55AR-II, today.
S The conditions for the existing plants are also well 4 known to you because that is the focus, to great measure, of 7 what we are doing with IDCOR, where we are looking at the ;
[
t population of 120 or so estating plants and saying by what {
9 means can we systemattoally and effistently and effeettvely [
10 evaluate severe accident risk in those plants so that we can I
11 discover and suppress outliers.
12 We generally speak of outliers as those sooident ,
() 13 vulnerabilities, severe accident vulnerabilities that stand 14 out from an acceptable background level that would be higher !
15 than desirable, or perhaps even higher than tolerable.
I le That process was done in one case, certainly, at 17 Indian Point, exhaustively over a period of many years, with i
l' 18 hearing, and that was far from perfect and certainly a very 19 expensive and ponderous way to do it. We are looking for a 20 much more offlotent way, 21 For future plants, the severe sooident polley 1
22 statement is much more expliott, and we have been trying to do i
23 the General Electrio OESSARail review in parallel with the i 24 development and promulgation of that policy statement, and it i l
25 bastoally lays out ortleria by whtoh you first and foremost l
t I
i k
+ _ _ - _ _ _ _ _ _ _ - - _ - - _ _ _ _ - _ _ __ !
l
! 8 ,
l l 1 take the existing regulations in their latest form, look 2 carefully at the plant and get that behind you, satisfy l r l 3 yourself that this future plant is going to satisfy the 1 4 regulations in all respects; then you take in serial order the 1
r S unresolved safety issues, generlo safety issues and the like, e and I always regret that the severe accident polloy statement 7 is, if anything, gith in saying to resolve them, to address l
8 them and resolve them.
l I 9 That is a very difftoult thing to do, it is a far j r
I 10 different thing to address and resolve station blackout on the I
l 11 OESSAR-l! than it is to resolve or addresa and resolve A-29 on 12 safeguards. We are intellectually limited on some of them, ;
() 13 14 and it is a rather diffitatt thing to speak dispositively, to say, oh, yes, we have addressed that and resolved it.
l 15 Nut nevertheless, 14 is incumbent on us to look at 16 these plants in order to do that, to take the information base 17 we have, to address the plants to the extent that we can and 18 have that as part of our overall safety judgment in saying
( 19 this future plant is acceptable for future lleensing, for 20 referenoeability in FDA as we speak of it now, or even ii we i
l 21 go in that direotton, if 0,M, chooses, in the not-too-distant I
I 22 future to even certlfy this design by rulemaking through the l
23 Commission, 10 CPN 50, Appendix 0, whtoh has been on the books i 24 for ten years now, ,
j 25 Now, the last condition and the one that raises a r
I i
l ._
9 1 great deal of controversy and difficulty in the future plant 2 review is the use of the FRA, which is, as it should be, a 3 systematio way to display what we understand about the safety 4 of this plant, the vulnerability of this plant, and determine S what is worth doing. What refinements of design, what 6 addenda, what alterations are appropriate, are justifiable in 7 light of the severe accident risk of this plant and the 4 overall cost or cost-benefit of such change.
9 Now, we have here with OKSSAR-!! the first real 10 example of this in the regulatory process, and it differs from 11 the usual review in a very significant way, and I invite your 12 attention on this difficult part of it, which is not does this
() 13 14 plant sattsfy the regulations or not, is this plant sufflotent, does this meet ODC X, Y or 2. No, the issue here 15 is have we gone far enought Are we satisfied that we have 18 turned over all the stonest 17 Now, the Commission has not given us the benefit of it a formally-adopted safety goal or set of ortleria by which to 19 judge this is safe enough. And as I ikke to tell people, even 20 if the Angel Gabriel comes down from heaven with such a safety 21 goal or set of ortleria on a plaque, I don't know how to 22 evatuste against them because we don't have the certainty to 23 say this is exactly the core melt frequency or this is exactly 24 the severe acoident risk off site of this plant.
25 There is just far too much unknown in this. But it
10 1 is a useful backdrop. It is a useful description of acceptable
) 2 risk, just as the PRA work, the source term work, the severe 3 accident consequence work is a useful description of the 4 vulnerability of this plant and the effect on that S vulnerability that one feature or another might have.
6 So as we go in, it is not terribly important that 7 the Commission has not officially adopted a safety goal. There 8 is certainly a range of debate enshrined by many years now --
9 in fact, one of your members devoted a year or so of his life 10 to the safety goal. The range of debate gives a clear idea of 11 the philosophy of acceptable safety and the best available 12 quantitative descriptions of that level of acceptable safety, 13 and we come into this forum, along with the applicant for this 14 FDA, this Itcense, and we come into this forum with the 15 available tools, the state of the art we have on severe 16 accident risk analysis, and we can say how does this match 17 this backdrop, this description of acceptable riskt it In general, if you refer to the Safety Evaluation 19 Supplement No. 4, whtoh is the most useful compendium of the 20 outcome, you . see that, not surprisingly, the General 21 Electrio portrayal of risk and the Staff portrayal of risk are 22 not the same. General Electrio chooses core melt frequency 23 values and acoldent release values that are different from the 24 8taff, The Staff is roughly one order of magnitude higher in 25 the core melt frequency. And in what people call source terms
11 1 in the severe accident releases, we are more like two orders 2 of magnitude higher than the O.E. portrayal 3 That is not surprising and it is not disabling 4 because if you look at it -- and in fact, it is delineated in 5 here in that fashion -- the Staff is saying, for instance, in 6 the source term arena, we choose to use what we would 7 characterise as a higher range of outcome. You know how much 8 uncertainty there is. It is a matter of active debate on the 9 severe core damage phenomenology, core melt phenomena, 10 core-concrete interactions, all of those things.
11 We choose to err on the side of caution, perhaps, 12 using, roughly speaking, two orders of magnitu'de higher 13 releases than O.E. does. Reinforcing that, we choose one 14 order of magnitude higher core melt frequencies. Thus, when 15 we do a cost-benefit analysis, we will more likely justify a 16 change. We will more likely conclude that a certain thing is 17 worthwhile, worth doing. We will draw that line in the sand 18 at what is an acceptable OESSAR-!!, and a stricter position 19 than O.E. would, than they propose.
20 Now, what that says is since we are in great measure 21 erring on the side of caution, that our uncertainty is 22 somewhat skewed then, that it may be that with later 23 information, we will learn not that we should have done more, 24 but we would learn that perhaps some of what we have done is 25 superfluous.
12 1 But with that in mind, and recognizing that we don't 2 have exact answers, the Staff went ahead and made its 3 judgments, and you will see or have seen in the Safety 4 Evaluation Report Supplement No. 4 that we have chosen -- and 5 General Electric, I believe, has acknowledged -- further 6 additions to the design philosophy and the design approach of 7 GESSAR-II, the ultimate plant protection system and certain 8 features associated with hydrogen control and DC power that 9 we think significantly enhance the safety of the plant, and 10 therefore are justified, and they become part of the FDA.
11 Now I would like to add, going back to the range of 12 debate for the safety goals, a very important point. This is 13 not with respect to the arguable levels of safety. We are not
(
14 hanging around a coremelt frequency of 10 to the -4 per year.
15 We are not hanging around a .1 of 1 percent increase in 16 background risk for people who live near a plant that would be 17 built like this.
18 No. In fact, we are haggling in a very difficult 19 arena, in coremelt frequency, by our estimate, where we are 20 already down in the range of 10 to the -5 per year, plus or 21 minus. In GE's estimate, it's another decade lower than 22 that. And in the source terms, we are well below the range of 23 debate.
, r- 24 So with respect to acceptability, criteria for i k i 25 acceptability, this is not a close call. Where it is a close t
13 l
1 call is in cost / benefit analysis. And the cost / benefit
(~) 1 sI 2 analysis in this report, in this action, is indeed based on !
3 the dollars per person-rem.
4 And just earlier, it was mentioned by the Acting 5 Subcommittee Chairman that, why or should the Staff use 4
6 averted economic loss? It is well-known that you can more 7 readily justify a design improvement or a design change using 8 averted economic loss -- the severe core damage and coremelt l
9 cost being immediately of the order of, you know, a few 10 billion dollars or more -- that you could justify it more .
11 likely with averted economic loss than you can with averted 12 offsite public health risk.
() 13 We know that. We have had substantial signals from our own Commission that they are not pursuing the idea of the 14 15 averted economic loss, and as you know, the Staff has 16 recommended it. The ACRS has recommended it in a number of 17 ways.
p 18 No. What we are doing is, we are basing the
\
19 cost / benefit analysis on the averted public health risk, and 20 we are nonetheless satisfied that we are achieving a very 21 high level of safety. We are achieving reasonable j
22 justification of design refinements for this technology, for 23 this kind of reactor, that are good, that are worth building, 24 that are justifiable in the fullest sense of the word.
25 We don't feel cheated because we can't use an i
14 i averted economic loss measure. And if I would say there is a
['T
() 2 lingering concern, that lingering concern is not that the 3 hardware design could be justified to be something different.
4 No, that lingering concern is the one that we would have 5 anyway, no matter which criterion we use, and that is that 6 this comes true when that plant is built. There are no 7 fission products in this blue book. But when we build or 8 someone builds this plant, to make all of this PRA come true 9 is the lingering question and is the one we would want that, 10 in the actual construction and in the operation, we realize 11 the apparent benefits of this design.
12 So with that, I would like to open to your 3 13 questions. The criteria, the lack of a safety goal, are not a
{a 14 problem, because this does not come out to be such a close 15 call 16 MR. LEWIS: Bob, just as a quick obiter dictum -- or 17 you made an obiter dictum -- that the ACRS had recommended 18 including averted economic cost?
19 MR. BERNERO: I shouldn't go too far. In a number 20 of -- if you go back to NUREG-0739, the original -- which 21 really wasn't an ACRS report, there has been a lot of debate.
22 I should have said that more accurately, that some members of 23 ACRS have.
l 24 MR. LEWIS: And some have recommended against it.
lO t 25 MR. BERNERO: Yes, and some have recommended against
15 1 it.
/" N 2 MR. LEWIS: I just wanted to clarify the record on 3 that point.
is 4 MR. BERNERO: And I think in their unique way, the 5 Commission made it abundantly clear which way they are leaning 6 in the final editing of the Severe Accident Policy Statement.
7 MR. LEWIS: But don't lay it on ACRS.
8 MR. BERNERO: No, no. No, the Commission did that.
9 MR. KERR: Bob, you made a comment about the way in 10 which the Staff arrives at numbers for risk and said something 11 about erring on the conservative side or erring on the high 12 side.
() 13 It was my impression that the Gospel according to the Staff was that PRAs are supposed to best estimates. Does 14 15 that mean that when you get to the cost / benefit stage, that 16 you do another PRA that's a conservative one?
17 MR. BERNERO: Well, no. I think -- oh, any number 18 of -- there are some Staff members here who worked with me or 19 for me in the past years -- I have used an expression over and I
20 over again for a long time -- and I think it characterizes 21 what happens in PRA-land correctly and properly -- and that l 22 is, you are trying, to the best of your ability, to make a 23 realistic estimate, but you always approach realism from the
-s 24 pessimistic side of the field.
V 25 Anc' we just had an example a little over a month ago
16 1 -- no, not quite a month ago -- the Davis-Besse incident --
C' 2 and you know the controversy about feed-and-bleed and how many 3 aux feedpumps should you have and what kind of HPI -- and the 4 success / failure state criteria for Davis-Besse in all of the 5 analyses to date were that that plant can't feed-and-bleed.
6 In retrospect, throwing a lot of thermohydraulic 7 bucks into it and people, it looks like that plant can 8 feed-and-bleed, fairly marginally, at least for a long time, 9 but its success state is not what the PRA portrayed it.
10 I'm not uncomfortable with that. We want to be as 11 realistic as we can, but as WASH-1400 did iodine or as we have 12 done with feed-and-bleed capability on Davis-Besse, where you 13 don't have a solid basis to make a realistic estimate, you 14 should be very careful to know where pessimism lies and choose 15 to lean that way.
16 MR. KERR: Well, it seems to me that in almost any 17 situation that you encounter, there are uncertainties. So 18 what you seem to be saying is that if the uncertainty gets big 19 enough, you then become conservative, and I don't know how big 20 is big enough, and I expect nobody else does either.
21 So it would be fair to say that PRAs generally are 22 conservative.
23 MR. BERNERO: Especially if the Staff is sponsoring 24 or doing them.
\,
25 MR. EBERSOLE: Bob, there's a little bit of a hole,
17 1 I think, in that rationale, about feeding-and-bleeding at
'V )
2 Davis-Besse or, for that matter, most other PWRs. Although we 3 have long culled it out, the circuitry and equipment that 4 makes it possible to feed-and-bleed is energized circuitry, 5 comparatively high-voltage, 250 or 125. That circuitry is 6 open and exposed to a hostile environment after a period of 7 feed-and-bleed, and presumably will die in the cold or dead 8 state, at which point the valves close and you can't open 9 them.
10 MR. BERNERO: That's a possibility. But there have 11 been cases -- February 28, 1980, the Davis-Besse -- when they 12 had a loss of the NNI bus -- you know, essentially it was a reproduction of the Rancho Seco lightbulb incident -- and the
( 13 14 operator didn't really know what was happening at Rancho 15 Seco. He turned it on, _and he fed-and-bled 50,000 gallons 16 worth of coolant.
17 MR. EBERSOLE: What was the containment environment?
18 MR. BERNERO: Well, it is primary coolant. It is 19 not that hot. And I consider that a much lesser challenge 20 than failure.
21 Now in Davis-Besse, the implementation of the 22 procedure was somewhat delayed. There is a question, as you 23 may know from NUREG-1154, that the operator's definition of
(~N 24 "immediately do something" was subject to question.
25 MR. EBERSOLE: Well, at least in theory, you can't
18 1 live very long with the PORVs unless you do something to 2 upgrade them.
3 MR. BERNERO: Yes, 4 MR. KERR: May I continue, please?
5 MR. LEWIS: Well, if I could pursue the conservatism 6 for a moment, if you're not coming back to that --
7 MR. KERR: I was going to another subject.
8 MR. LEWIS: Okay. Let me just pursue conservatism.
9 Just for the record, there are those who feel very 10 strongly that an approach to realism from the conservative 11 direction, as you so quaintly put it, is logically flawed, and 12 it can easily lead to non-conservative results. And in your 13 early comments, you made statements like, "Maybe later we will 14 learn that we were too conservative, and we did things that we 15 didn't need to do."
16 Those things that you didn't need to do, there is no 17 way to prove that they didn't make the plant less safe rather 18 than more safe.
- 19 So at least some of us believe that when you have a l
l 20 large uncertainty, you should do your level best not to l
21 approach from the conservative direction, but aim squarely at 22 the middle and acknowledge your uncertainty, and that that's 23 the only way in which you can end up with a result whict. you L
i f
% 24 can characterize in terms af a me-an and an ancer*ainty a s,s u n d I s 25 it.
[ _ _ _ _ _ _ _ _ _ _ _ _ _ _
19 1 That's for the record.
2 MR. BERNERO: May I add to that? l 3 Again, invoking the Davis-Besse as an example of 4 what I think you are driving at, in the Davis-Besse plant, you 5 may know that there is a startup feedpump that sort of has the 6 character of an extra auxiliary feedwater supply. But 7 unfortunately, it was discovered not long ago that it's in the 8 same room as one of the safety-grade auxiliary feedwater pumps 9 and had some piping that was not appropriately designed to 10 withstand high-energy line rupture, et cetera.
11 And in the gavotte that went to "What do you do 12 about this," it ended up that the Licensee proposed, "I will close some valves, and I will padlock the valves, and I will
( 13 14 rack out some breakers, and I will do a total of about five 15 rather bizarre things to assure that there will be no high 16 energy in the line, and therefore no high-energy rupture."
17 And that actually appears to have made the plant distinctly' 18 less safe, because here when that pump was valuable as a water 19 ~ supply in a very real, in a very ordinary sort of incident to 20 begin with, it took extraordinary measures by the plant staff 21 to go down there and get that pump on line, you know, a series 22 of keys and all sorts of things.
23 That is a classic example of how erring on the side s 24 of conservatism may not be so -- that situation of saying, "I 25 will defend against high-energy line rupture.
20 1 But nonetheless, you are faced with the dilemma, 2 whenever you do a PRA -- and I like to go back to that vintage 3 example of, how do you treat the iodine, in WASH-1400 and --
4 what was it? -- Appendix 4 or Appendix 6 -- that said, "It 5 looks like it ought to be the iodine salt, cesium iodide, but 6 there's just not enough data." Well, we have to treat it one 7 way or the other, and we'll just treat it as the more 8 volatile, elemental radioiodine, and as you well know, there 9 was a partition coefficient associated with elemental 10 radiciodine that was quite generous, and there was even 11 controversy at that time, because that didn't gibe with the 12 Reg Guide, but you have to make such a choice, and you have to
) 13 carry that choice through your cost / benefit analysis, and it a 14 will pervade that analysis and give you that pessimistic side 15 of the field skew that I speak of to a realistic estimate.
16 It takes care -- it can be counterproductive; it can 17 make the plant less safe. And that's a major uncertainty.
18 But the central aim -- and then stand back and say, 19 "That's what I have, and there's an uncertainty" -- still 20 leaves you with the need to make a decision, the need to have 21 some fiduciary mark, $1000 per something, to decide whether or 22 not you will turn to these people and say, "I want 23 AC-independent tiki torches for hydrogen ignition. I want the f-s 24 UPPS. I want whatever I want."
U 25 You have to make that decision, and when you do, you
21 1 will consciously or subconsciously choose to reflect your O
ksl m 2 bias, and what I have instructed people for years and continue 3 to instruct them is, be careful, be knowledgeable, be as 4 realistic as you can, but approach realism from the 5 pessimistic side.
6 MR. LEWIS: This is the wrong place for an extended 7 debate on that point, but I think you are really quite wrong 8 on it. But I'm sorry, Bill 9 MR. KERR: You know, after observing how easily you 10 convinced Mr. Bernero of your viewpoint, I think I will forego 11 further comment.
12 MR. LEWIS: I really do apologize.
13 MR. MARK: Cost / benefit, $1000 totally.out of the
(
14 air per man-rem. Nobody knows if it's worth a nickel. It 15 goes into the numerator or the denominator, whether you do 16 benefit / cost or cost / benefit. The estimate of those things is 17 as hairy a step as one can possibly -- probably never seen I 18 before. Until one became more than 21, one never would have l
19 believed such hairy arguments as go into the estimates of a 20 man-rem or some release.
21 In the case of the GE, I think the situation is t
( 22 really a tremendously swampy, soft basis for any kind of I
l l
23 argument. Four-fifths of the total risk comes from seismic, r- 24 The risk from internal is really very modest indeed.
d Some large fraction of the seismic risk comes from 25 l
' 22 1 some assumptions about relay chatter, which aren't based on m
2 any look at any actual relays, but are based on some relays 3 that somebody once said chattered like this.
4 It would be entirely possible to go and say, "Look, 5 make sure that in this plant you use relays which are immune 6 to or largely defended against chatter." Then you would 7 change the whole estimate of seismic risk. You would change 8 the denominator of the cost / benefit equation. You already 9 can't get things down to $1000 for some of the modifications 10 which are thought of. There might be good reasons, but I 11 don't think this is a good reason.
12 You can't get very close -- well, you can get fairly
[
i 13 close to $1000, but on the basis which you simply know could 14 be improved, should be improved, perhaps shouldn't be used at 15 all.
16 So I think the situation here is not an awfully i
17 encouraging one from the point of view of having a numerical 18 basis for decisions, as derived from PRA, as PRAs are done, 19 and can or even can be done, and that one ought to accustom 20 oneself a little more to saying, "We're going to rely on 21 judgment, since that's what we do anyway, and in our opinion, 22 it would be good to have the containment inerted."
23 I don't believe that, but, I mean, you could just 24 say that, It isn't going to come up in a cost / benefit 25 formula. It isn't worth it at all on that basis, even with
e 23 1 inflated risk numbers, and I think we sh3uld perhaps somewhat 2 regroup, rethink, and decide that we're going to assert 3 policy, that we're going to make the best judgment we can, 4 we'll debate it, but we won't debate it on that basis.
5 MR. EBERSOLE: Carson, could you extrapolate that to 6 include security?
7 MR. MARK: Oh, in fact, in security, of course, 8 there is absolutely no underpinning for a cost / benefit 9 analysis of any kind. And if there was yesterday, there is a 10 new one today, with the bombing last night and so on.
11 One shouldn't even be attracted to go that stupid 12 route.
13 MR. BERNERO: I would like to speak to that because 14 I think what you are voicing here is the most significant 15 aspect of this review as compared to other reviews. The point 16 I was trying to make before, that ordinarily when we speak of 17 the acceptable safety of a power plant, prospective design or 18 an actually constructed one, we are speaking of satisfaction 19 of regulations, satisfaction of regulatory requirements that 20 are so carefully crafted that all of this mushiness that you 21 speak of, all of this ill-defined quantitative deliberation is 22 buried back in some regulatory activity that came out in all 23 the splendor that we have in the regulations now. That a 24 single failure criterion, as qualified by the word and the 25 practice of regulation is a sufficient test of reliability.
24 1
And what we have, then, is in the conventional
\~ 2 review, we have masked these difficult problems to a very 3 great extent. We have dealt with them generically. We have, 4 perhaps as in the case of the ATWS rule, gone through some of 5 this very tortured cost-benefit analysis and then backed away 6 from it and said you will have that feature in that plant; a 7 very prescriptive, simple regulation that derives from that 8 information but -- I make this argument and I must make it 9 very strongly -- the decision is not made by the cost-benefit 10 analysis. The cost-benefit analysis can only display in 11 some consistent fashion what you know and what you don't know 12 in order that judgment can be focused on the places where it (p). l'3 has to be exercised and exercised with as much information as 14 you're going to have.
15 MR. MARK: But of course, that presumes that the 16 cost-benefit analysis is.done as well as one knows how and not 17 approached from the left or the right in any conscious way.
18 MR. WARD: Dr. Kerr?
19 MR. KERR: Mr. Bernero, I think I agree with your 20 discussion of the way in which regulations are used until one 21 begins dealing with the severe accidents. As I understand the 22 policy statement, -- and although I have looked at it 23 carefully over the months I'm still not sure I do -- one deals 24 with the severe accident issue at least for new plants on a l O 25 basis which depends rather heavily on PRA and on cost-benefit
t 25 1 analysis. There, they are not regulations to guide one, so Os Y ~ ;:
^'
\ss/ 2 far as I know.
3 And it is to that, I believe, that Mr. Mark was 4 speaking and not to existing plants. And I don't know how one 5 is going to make those decisions, but at least the Severe I
6 Accident Policy Statement would lead one to believe that the 7 PRA and other analyses of that kind would be dependent upon 8 significantly. At least that's my interpretation of the 9 statement.
10 MR. BERNERO: I agree with that. In fact, the 11 reason I invoked the single failure criterion out of the 12 current regulations was that is a severe accident question.
I 13 Core melt frequency is a contest, or an evaluation, that goes b
14 past the single failure criterion. That is a sophomoric 15 evaluation. It happens to give you a basically reliable 16 system to begin with, but then it requires the more careful 17 scrutiny of multiple failures, common cause failures, in order 18 to determine is the occurrence rate or the arrival rate of 19 severe accidents for this design going to be acceptable or 20 not.
21 And admittedly, we are trying to be as realistic as 22 possible but I think the committee has to recognize that you 23 sooner or later will, in some aspect or another, confront 24 issues where you just don't know where realism is. It's a 25 range of debate; a range where you don't have sufficient test
26 1 data or sufficient analysis. And it's only in those regimes 2 that I say you have to be very careful in your judgment on 3 that specific to choose how pessimistic you are or what you 4 use in that particular parameter.
5 But what we're trying to do here with this 6 cost-benefit analysis is to consistently and fairly approach 7 realism as carefully as we can, and then consistently apply 8 that " realistic" cost-benefit analysis to measure 9 possibilities for refining this design and its subsequent 10 implementation.
11 Now, I have to turn to Staff on the relay chatter.
12 Did Staff, somehow, somewhere actually do the cost-benefit of 13 qualified relays? I just don't remember anything on that 14 subject. So that I would have to take that up -- if that were 15 a separate issue --
16 MR. WARD: Did you get an answer?
17 MR. BERNERO: I am getting negative nods.
18 MR. WARD: Well, we can't see those.
19 MR. SCALETTI. No, we did not.
20 MR. WARD: Thank you.
21 MR. BERNERO: The Staff did not. So if you go in 22 this SER Supplement No. 4 -- I forget the table number but
.3 there is a table of the stuff that GE first started to look
- 24 at. Table 15.4. And then the Staff, after they reviewed it, 25 kind of combed the field another way in sharper analysis of
27 1 the principal competitors of modification. But I just don't s
2 think we did look at the chatter-qualified relays, or whatever 3 you might call them.
4 MR. WARD: Dr. Remick has a question.
5 MR. REMICK: Bob, would you help me? Is the GESSAR 6 II a modification of an existing FDA, or is it a review of a' 7 new FDA?
8 MR. BERNERO: Well, I will defer to any correction 9 from Cecil Thomas. The existing FDA's are backward 10 referenceable. And what you have here is a substantially 11 modified design over the original GESSAR, and the issue -- and 12 the phrase used for the issue in the Severe Accident Policy 13 Statement was forward referenceable FDA -- use in future 14 cases. And that's where you get the big threshold of you have 15 to follow the Severe Accident Policy Statement, the CPML rule 16 and all that kind of stuff. So it's really a different FDA; 17 substantially different.
18 MR. REMICK: So you are approaching it from that 19 standpoint.
20 MR. BERNERO: Yes. I am dealing with it as a new 21 FDA. It builds on that previous history.
22 MR. REMICK: I have another question. You have 23 referred to rulemaking for the design certificate in Appendix 24 O. For sometime I have been making the claim that I am not 25 sure anybody knows what that rulemaking would be. Do you know
28 1 if I am wrong? I haven't found anybody that has told me I am s 2 wrong yet, but I keep asking or making that statement, that I 3 don't think anybody has thought out what that means.
4 MR. BERNERO: Well, I would second your comment S because we have never done that. And we have had discussions 6 in Staff that if we ever sit down to do it, we really have to 7 go through the protocols rather carefully, because it hasn't 8 been done. Would the Commission sign SER's, you know, what 9 would be done? It's a rule prerogative in principle but there 1
10 is no precedent, no practice, to which we can look. And 11 internal discussions in the Staff, and I believe to some 12 extent in OGC, have kind of explored this.
13 Perhaps GE would be the first applicant to go for 14 that. And I think part of their decision would be what is it; 15 you know, what sort of resources would it take.
16 MR. REMICK: That's exactly my point. I don't know 17 how anybody can make that decision without knowing what 18 ballgame they're get*.ing into.
19 MR. BERNERO: That's really one of the reasons that 20 the forward referenceable FDA is an alternative in the policy 21 statement that was added, because it was really kind of a 22 Pandora's Box, this rulemaking by Appendix O, including the 23 NRC people. It wasn't too clear what that would take, how 24 long it would take, you know, the hearing mechanisms and all 25 of that to be established.
_ _ _ . , _ _ _ _ _ . , , , _ , , _ . - _ _ _ . _ . , ._y _ , , , - - , , , . , - , _ - , , ,. . - _ - , . . - _
_ _ _ _ _ - ___ ~ - -- _ ____ _. ___.. --. _ - _ _ _ . ~ . _ - _ - - -
I 29 1 MR. WARD: Okay. Jesse?
2 MR. EBERSOLE: No more questions? Let me ask 3 G e n.. r a l Electric, do you have any observations on any of 4 this? On any of the discussion?
i
! 5 MR. SHERWOOD: I am Glenn Sherwood of GE. I could 6 give you a little background on attitude of the GE engineers 7 to cost-benefit analysis. That is maybe appropriate. And I 8 think -- to be perfectly candid with you, I think they really 9 disdain cost-benefit analysis.
10 Our engineers in the Nuclear Division are under the 11 same scrutiny as other GE engineers in our Airplane group and 12 our Steam Turbine, and therefore, they take great pride in 13 their work. But as they have done for decades, they 14 essentially do their engineering by engineering judgment.
15 And we think, for example, that the GESSAR plant has 16 probably the lowest core, melt frequency of all the current 17 generation of plant, if not the lowest, and probably the best 18 off-site risk of all of the current plants.
19 So at least within General Electric, we rely on good 20 engineering judgment, just as we have for decades. We do 21 cost-benefit analysis when we're forced to do it, but we 22 really don't put much stock in it.
23 MR. EBERSOLE: Thank you.
l 24 MR. LEWIS: When you do it, do you do it 25 conservatively?
1
30 1 MR. SHERWOOD: Oh, of course.
2 [ Laughter.3 3 MR. EBERSOLE: Any other questions?
4 [No response.3 5 Thank you, Bob. The second topic on the agenda is a 6 discussion of hydrogen and related matters like scrubbing
? efficiency and the suppression pool and so forth. I believe 8 that is General Electric's topic.
9 MR. SHERWOOD: That will be Dr. Debbie Hankins.
10 [ Slide.]
t.
11 MS. HANKINS: I am Debbie Hankins. I don't know if 12 we had a miscommunication on this agenda item, Jesse, from 13 what you just said, but we had limited it just to the
\
14 discussion of the hydrogen issues. We weren't going to go
! 15 into the source term.
16 MR. EBERSOLE: That's correct. I really didn't mean 17 source term. Just what you have up there.
18 MS. HANKINS: Okay. Thank you.
19 The agenda asked for specific topics, so I have l
20 tried to address those in the order in which they were 21 listed. The first one was on the rate and amount of hydrogen.
22 The generation rates that we see for the full core 23 melt scenarios -- and I will remind you that in a GESSAR PRA 24 we analyzed full core melt scenarios as opposed to the O 25 degraded core scenarios, that for example, the hydrogen
31 1 control owners group is analyzing.
D k ,)
s 2 For those scenarios, the generation rate varied in 3 time during the sequence and for various sequences, from about
, 4 .4 to 1.6 pound mass per second. Those are about the range 5 that HCOG is analyzing in their experimental work.
6 The total in-vessel hydrogen produced, again, varied 7 by sequence from about 1300 to 2300 pound mass. There's only 8 enough oxygen in the 238 MARK III containment to support the 9 combustion of about 2480 pound mass of hydrogen, and that's 10 equivalent to about 67 percent of the active clad as opposed 11 to the hydrogen control rule which states you must analyze --
12 MR. MARK: Could you help me? That 2480, that's the I) 13 total air space -- wetwell or drywell?
14 MS. HANKINS: Wet.
15 MR. MARK: How is it broken down, roughly?
16 MS. HANKINS: I think it is 1.1 times 10 to the 6 17 cubic feet in the wetwell, and what is it, about .4 times 10 j 18 to the sixth for the drywell.
19 MR. MARK: So a quarter of it is in the drywell and 20 three-quarters in the wetwell.
21 MS. HANKINS: Yes. Maybe even more than that; a i
22 higher percentage in the wetwell 23 Ts.is is assuming one burns down to 5 percent p-sg 24 oxygen, and it's enough to support the combustion of 67 25 percent metal-water reaction. The requirements of the CPML
32 i rule are that you must have a system capable of handling 100 (D
k-s 2 percent metal-water reaction in the active clad.
3 Again, we had some discussion yesterday about 4 whether one would go down to 5 percent oxygen or if one would 5 have complete combustion. Theoretically, you could have 6 complete combustion; however, the way we envisione.d these 7 accident scenarios is that the hydrogen would be bubbling up 8 through the pool and you would actually have multiple 9 ignitions. So eventually, once you got the oxygen 10 concentration down to 5 percent, you could not have -- or 11 continue to have ignitions. So that's why we say it is 67 12 percent equivalent.
In terms of hydrogen detonations, when we did the
) 13 14 original PRA we assumed there was about a 26 percent chance of 15 a global hydrogen detonation which would cause simultaneous 16 failure of the drywell and the wetwell. Our understanding of 17 detonations, and in particular global detonations, today is 18 that in fact, they will not occur. And as such, the original 19 PRA results are conservative compared to the analysis, if we 20 were to perform it now in 1985 as opposed to 1982.
l 21 That understanding comes from testimony that was 22 given a the Perry hearings by Dr. Bernard Lewis who is a 23 recognized world expert on combustion phenomena. He says that
(
O 24 in the MARK III geometry you simply cannot have the proper l
25 conditions to have a detonation. We have some disagreement i
i
33 1 with the Staff in that we say that the o.ily way to fail the 2 drywell through hydrogen phenomena is by a global hydrogen 1
3 detonation. The Staff assumed a certain fraction of the time 4 one could have a local detonation that was capable of failing 5 the drywell. We do not agree with those results; we think 6 there is insufficient energy to fail the drywell for a local 7 detonation.
8 Because of their assumption on local detonations 9 failing the drywell, the Staff is showing a high seismic risk, 10 which can then be reduced by about a factor of 2 with the 11 implementation of igniters which would burn the hydrogen in 12 low concentration and preclude the possibility of 13 detonations. Again, we disagree with that analysis so we 14 would not show the same type of risk reduction for ignitors.
15 We do agree with the SER in the sense that there is 16 no risk reduction for hydrogen for internal events.
17 [ Slide.]
18 What we are talking about is this was the original 19 PRA without us. Again, this is the BNL staff assessment of 20 risk. The GE numbers are substantially lower.
21 Today, since GE has committed to the implementation 22 of the UPPS design, we would say the GESSAR design is really 23 here. And if one takes the design and adds then igniters with 24 the back-up power supply, the risk reduction essentially is 25 sero.
_. ._. . _ _ - - _ . .___._ - ._. . _ - - - . ._ _ - - - _ _ _ _ _ . - - = . _ .__ _.__ -
l
. 34 1 The large risk reduction shown here for seismic
%,) 2 events again is related to that assumption of LOCA detonations 3 failing the drywell and we believe that when a realistic 4 analysis, even approached from the pessimistic side, is 5 performed, that these numbers will not show that kind of a 6 risk reduction for seismic events with igniters.
7 [ Slide.)
8 Despite our disagreements, GE has committed to 9 provide a hydrogen control system that is consistent with the i 10 outcome of the HCOG program and the NRC 'eview r of that
. 11 experimental and analytical program.
12 The NRC is requiring an SER for it, that we provide
() 13 a diverse power supply for the igniters, and this is to supply power to the ignitors for the case of station blackout, just 14 15 the dominant core melt event.
16 We find no technical justification for that back-up 17 power supply. However, GE will comply with the SER, we will 18 provide that back-up power supply.
19 In addition, relative to hydrogen, we have also 20 provided the UPPS system. Once again, despite our i
21 commitments, our position is that hydrogen control is 22 unnecessary, that it provides no risk reduction, in that the 23 risk is already low. We find no justification in particular g 24 on a cost-benefit analysis for additional hydrogen control in U 25 the GESSAR plant.
35 1 But, again, we will comply with the requirement.
(
\_) 2 [ Slide.)
3 One of the other issues that is on the agenda 4 related to hydrogen --
5 MR. MOELLER: Excuse me. On your cost-benefit which 6 did not support it, how close was it? Or, you know, was it 7 miles away from --
8 MS. HANKINS: Yes, orders of magnitude.
9 MR. MOELLER: Thank you.
10 MS. HANKINS: Oh, I thi.u if you just look at the 11 internal events, using the Staff BNL analysis, they are 12 showing no risk reduction. If you show no risk reduction and 13 divide it into the cost, it's a pretty high number.
)
14 MR. MOELLER: Thank you.
15 MR. REMICK: Could we have a copy of that slide at 16 the table?
17 MS. HANKINS: It is in SER 4. It was one of the 18 handouts yesterday.
19 One of the questions, because hydrogen phenomena in 20 a BWR, or because of the suppression pool, can be different 21 than what one sees in, for example, large dry containment or 22 even an ice condenser, and that as the steam and hydrogen is 23 delivered to the pool and it starts to bubble up through the s 24 pool, the steam is stripped out, and what is released from the
(
25 surface of the pool is essentially pure hydrogen.
36 1 And that hydrogen then comes upon an ignition source
(
2 such as an igniter, and can burn in what we would call 3 diffusion flames. And the question arose, what would be the 4 impact of those flames on increasing the probability of pool 5 bypass by either failing a sealer or a penetration in the 6 drywell?
7 The drywell equipment hatch in the GESSAR design has 8 a five-foot concrete shield plug between the wetwell air space 9 and the seals itself.
10 So we don't feel that there is any potential for 11 flames impacting those seals.
12 Likewise, with the personnel air look, it also has a 13 cement shield plug.
14 The electrical penetrations which are five foot long 15 are potted with a Portland cement mixture. There was a 16 concern with the MARK-! containment that if you had high 17 drywell temperatures, the compound that is used for potting of 18 the electrical penetrations is an epoxy type substance that, 19 given high temperatures, becomes more fluid. And then with a 20 pressure differential, there could be substantial leakage 21 through those penetrations.
22 The design of the penetrations for GESSAR are 23 substantially different, and the Portland cement material can 24 go to very high temperatures without losing its integrity.
25 So we don't feel that standing flames, either from
37 1 the standpoint of the equipment hatch airlock seals or the ON 2 electrical penetrations, would have any effect on the ,
3 integrity of the drywell 4 I have just one chart, and that will complete my 5 presentation, if it would be all right with you, and that was 6 a question on the ablation of the RPV pedestal. If we could 7 do that, then we wouldn't have to have GE and the Staff --
8 MR. EBERSOLE: Yes, that's the next item on the 9 agenda. You flew right into it, didn't yout to MS. HANKINS: The question arose at the last 11 committee meeting --
12 CSlide.)
13 -- we were discussing whether temperature effects
(
14 could have any impact on the drywell structure and this molten 15 core in the pedestal region. Could the thermal gradient cause 16 stresses in the drywell structure and cause loss of integrity.
17 That question was answered in the negative, in that 18 we did not have any concern from the thermal stress, but 19 another question arose, and that is what about the ablation of 20 the pedestal region, and could it in fact -- the ablation 21 result in the loss of the RPV and loss of drywell structure 22 and potentially containment integrity?
23 We went back to San Jose and looked at this issue.
s 24 One thing we would like to know is that the pedestal 25 design that was analysed and that Jack Rosenthat presented at
38 1 the last committee meeting was a pure concrete pedestal. It 2 was very similar to the design, for example, that they have in 3 Grand Gulf.
4 The GESSAR pedestal is a steel-concrete composite, 5 It is two concentric steel rings. Each are one and a half 6 inch thickness. They are connected with steel shear ties, and 7 the concrete is filled in between. But the primary structural 8 support comes from the steel, not from the concrete.
9 So we did an analysis starting with conditions that 10 were on Rosenthal's chart, and we said okay, let us assume 11 that the first 1.4 meters of concrete have been ablated and, 12 as such, the inner steel ring would also be assumed to be 13 gone.
14 We assumed that the only support is provided by the 15 outer steel shell which would then be at a radial distance of 16 1.8 meters.
17 And further, let's assume that that steel shell is 18 at high temperature.
10 Look at the loads, the weight of the RPV, the shield 20 wall and other equipment and the weight of the pedestal 21 itself. We looked at what that would mean in terms of the 22 load per unit area of that outer steel shell, and found it was 23 about 3.4 kai was the load the outer ring of the pedestal 24 would have to carry.
O 25 The yield strength of stowel at 1100 degrees
__ _ _ _ _ _ - _ = _ . . _. - _ - _ - . ---._- - _. _ _ _ _ _ _ - _ _ . - - . . . _ .
39 1 Fahrenheit is about 21 ksi. There is a tremendous dafference O And because of that, we felt that there was
\ ,/ 2 between 3 and 21 3 a very substantial margin and that the pedestal will in fact 4 carry the loads.
5 As such, we see no loss of the pedestal, the drywell 6 or the containment as a result of this ablation challenge.
7 MR. EBERSOLE: May I ask a question? I gather that 8 all of the structural material between the two shells was 9 standard concrete as well as the foot down below it. The GE 10 design looks strangely almost already as'a sort of a pot which 11 might contain the molten core and some steel from the vessel.
12 I would guess that it might have been of minor cost 13 additions to do something or to put something besides just 14 plain concrete in those areas. Did you-all look at that?
15 MS. HANKINS: You mean in terms of like a core 16 ladle?
- i l
17 MR. EBERSOLE: Whatever, 18 MS. HANKINS: We did look at that very early on in 19 '80 or '81 when we were doing some assessments on 20 core-catchers. And what we found was putting a material down 21 there that would decrease the penetration rate would drive the 22 heat in the upward direction, and we were more concerned about 23 having drywell failure due to the heat in the upward direction 24 than just allowing it to go into the concrete.
- 25 MR. EBERSOLE
- I see. You really have an absorptive
40 1 concept.
\ ,/ 2 MS. HANKINS: Right.
3 MR. EBERSOLE: Then that would either tend to go the 4 other way, that would make it even more absorptive than plain 5 concrete. Did you do that?
6 MS. HANKINS: Well, we didn't want it penetrating 7 the basemat.
8 MR. EBERSOLE: Okay.
9 MS. HANKINS: So there has to be a happy medium 10 there.
11 MR. EBERSOLE: Right. So this is it, plain 12 concrete. Thank you.
T 13 MS. HANKINS: That's all
[G 14 MR. EBERSOLE: Any questions?
15 MR. REED: I thought I saw something on one of those 16 slides that said " committed to UPPS." Is that something new, 17 that commitment?
l l
18 MR. EBERSOLE: We are going to take that up later.
19 MR. REED: Oh. Okay. I just wanted to come to a 20 better understanding.
j 21 MS. HANKINS: We were reminded by the Siaff in 22 yesterday's presentation that in addition to committing to the l
l 23 ignitors, the back-up power supply, that GE is also committed 24 to UPPS as part of the resolution of hydrogen control. And so 25 that is why I added it to the chart.
I l-
41 1 MM. EBERSOLE: Debbie, I think the committee might 2 be interested in your objections to the ignition system, 3 including your grounds that it costs a great deal, in that it 4 carries the burden of the spray system, which is the major 5 part of the cost. As I recall, it was about 1 million for the 6 ignitors and about 9 for the sprays.
7 I don't think the committee -- at least my guess is, 8 they don't understand why the igniter system carries the 9 burden of this, of the sprays. I 10 Could you clarify that?
11 MS. HANKINS: Well, let me clarify that the system 12 that we have committed to, which is consistent with HCOG, is J, 13 just the igniter system with the back-up power supply, which 14 is about $1.2 million.
15 When we did our original assessment of hydrogen 16 control for the cost-benefit analysis, we at that time 17 included a heat removal system. In other words, a way of
. 18 powering the containment sprays independent of the present RHR 19 system. And we did that based on the assumption that that 20 would be needed in order to preclude loss of containment 21 integrity for full core meltdown scenarios.
22 I think BNL and the Staff have done some analyses 23 that say the containment may or may not survive without sprays fN 24 operating during the hydrogen release.
l l
25 Again, that independent containment cooling system L
42 1 upped the cost from a b o u t ' i s ^ w .-1 1 . m . 10-a total of about $10 l s/ 2 million, and that was using the cost-benefit.
3 But again, even if you did the' cost-benefit with the 4 $1.2 million, you are still orders of magnitude away from the 5 cost-benefit ratio of 1.
6 MR. EBERSOLE: And the configuration now does have 7 these sprays or not?
8 MS. HANKINS: No.
9 MR. EBERSOLE: Any questions?
10 [No response.]
11 Well, thank you very much.
12 I am going to ask the Staff to comment on this 13 topic.
14 Do you have anything to say?
15 MR. MOELLER: .In the Staff's comments, it would help 16 me to understand better -- I gather on backfits, you do 17 cost-benefit, but on what you might call a forward fit or a 18 new design, you don't have to justify cost-benefit.
19 MR. BERNERO: No, you do cost-benefit analysis on 20 both. It's just what goes into the equation differs. Clearl 21 if you are backfitting to an existing plant, replacement power 22 cost is going to be a very significant thing for anything, you 23 know, for downtime. Rip-out costs. Backfitting the pedestal, 24 for instance, would be terribly expensive.
O 25 But in a future plant you do the cost-benefit
43 1 analysis, recognizing that you are dealing with lines on paper 2 and engineering costs, looking forward, rather than ripping 3 out and backfitting.
4 But you do the cost-benefit analysis and have to 5 exercise the same precaution in both cases. You can't get an 6 exact equation, you can't get an exact answer that just gives 7 you a numerical decision process. It just can't be done.
8 MR. MOELLER: Well, in the presentation we just 9 heard, though, if there is no benefit of the ignitors, how do 10 you require them?
11 MR. THOMAS: I believe the difference is the numbers 12 that are assumed or the numbers that result from the 13 calculations, I believe the Staff's numbers do show an w
14 incremental benefit, and GE's numbers do not. That is the 15 main difference.
16 One point of clarification: At least I was slightly 17 confused -- I believe the GESSAR design --
18 MR. KERR: Excuse me. I thought Dr. Hankins said 19 that BNL's numbers showed no benefit. Did I misunderstand?
20 Oh, for internal events. Okay.
21 So it's the seismic -- thank you.
22 MR. THOMAS: The other point of clarification, just 23 to make sure, I believe, that GESSAR design does include l
24 containment sprays, but not the separate dedicated sprays that l us/
25 you mentioned; is that right, Debbie?
I l
44 1 MS. HANKINS: Yes.
~g l
\ ,) 2 MR. THOMAS: They do contain containment sprays, but 3 they are not separate from the RHR system, as Debbie 4 described.
5 MR. KERR: Well, what would a spray system be 6 dedicated to, if it weren't dedicated to containment heat 7 removal? I mean you referred to the second one as a dedicated 8 spray system. The other one is undedicated, I take it.
9 MR. THOMAS: Perhaps GE could elaborate on that, 10 since -- since they brought up the subject.
11 MR. KERR: Well, you are the person who used the 12 term. What did you mean by dedicated?
13 MR. THOMAS: Totally independent of the normal RHR 14 system.
15 MR. KERR: Are you talking about two independent 16 systems? .
17 MR. THOMAS: For the purposes of cooling for severe 18 accidents and hydrogen burns and so on.
19 MR. EBERSOLE: There is a sort of a curious 20 relationship, Bill, about the potential for bypass of the 21 drywell that makes these sprays important or unimportant.
22 This design -- it's sort of difficult to get a 23 bypass for this design, but some plants you can -- you have a 24 higher probability of bypass.
O- 25 I understand that those have sprays in part
45 1 dedicated to accomplish a function after bypass of the
\~/ 2 drywell; am I correct? I hear that is more or less the 3 critical consideration, is the potential for bypass or not 4 bypass, or no bypass, rather.
5 This design has just about precluded bypass.
6 MS. HANKINS: The sizing of the containment sprays 7 is based on an assumed bypass from the drywell to the wetwell 8 air space, assumes steam bypass under LOCA conditions. That 9 bypass we realistically don't believe exists, but for 10 licensing purposes, design basis purposes, it is 11 included. Where we would use the sprays would be for, you 12 know, cases where you need additional containment heat 13 removal.
(%d 14 MR. EBERSOLE: If you should get substantial bypass 15 with some of the old designs,-like the loss of a downcomer 16 from PRV or something like that, you will only escape by spray 17 condensation because you are going to lose suppression 18 condensation.
19 MS. HANKINS: Most of those bypass mechanisms are 20 included in the MARK III design.
l 21 MR. EBERSOLE: Right.
22 Any further comments by the Staff on either this or 23 the ablation problem of the vessel support?
24 MR. SCALETTI. We have a presentation by Dr. Pratt r
s h
l l 25 from BNL.
I e
46 MR. PRATT: My name is Trevor Pratt. I am with n
V 2 Brookhaven.
3 [ Slide.]
4 Before I launch into the discussion of hydrogen, 5 perhaps I could clarify the point that Debbie was making with 6 regard to internal events and the change in risk associated 7 with the hydrogen control.
8 This is a copy of Table 15.9, which is in Supplement a
9 4 to the SER. What Debbie was talking about, if you look at 10 the base case with UPPS -- and this, they would argue, is 11 GESSAR now -- you have 33 person rem with UPPS, which is a 12 reduction of quite a significant amount without UPPS. The use fi 13 of igniters would take you down from 33 to 31, which is not d
14 sero but it is not very much.
15 If you assume perfect hydrogen control, you would go 16 down to 23. Now, let me explain these calculations because 17 the headings here may be a little bit confusing.
18 We have to make an assumption. We were l 19 extrapolating a deliberate ignition system into a regime which 20 it was not designed for. That is full core meltdown events.
21 We weren't sure how that machine would work in that 22 environment. So this calculation, which we have labeled 23 " igniters," assumes that you will lose the containment 24 building shortly after vessel failure because of uncertainties
("N l \~
j 25 associated with the core debris coming out of the vessel and 1
47 1 so on, but maintain drywell integrity, m
g, 2 This calculation, with perfect hydrogen control, 3 assumes that you maintain containment integrity and drywell 4 integrity until you fail many hours later into the accident 5 sequence as a result of buildup of partial pressure of gases 6 due to core-concrete interactions.
7 There is an argument, of course, and if you look at 8 some of the hydrogen generation rates during the full core 9 meltdown event, that indeed you could maintain containment 10 integrity during that period of time, and if the UPPS was 11 operating successfully, perhaps that could also help you. So 12 again, to say that this would be the only benefit -- it's a
\_/
) 13 range, if you like, of possible benefits from the system, and 14 certainly this is true that if you inerted the containment 15 building with perfect hydrogen control, that is what you would 16 get, but you might also get there by the installation of the 17 device, as well.
18 I don't know whether t h'a t helps clarify the point, 19 but that is what res11y we were talking about.
20 I have two presentations for you, and we can go 21 through as much detail as you would like. The detailed 22 discussion on hydrogen is not quite as detailed as it was for 23 the subcommittee, and I have taken out quite a few of the 24 Vu-graphs, and again, we have all of the backup slides, so if f-t 25 you do want to go into detail, we can certainly do that.
48 l 1
What I would first like to do, though, is to point I
2 out that there are two differences in the way we have done the 3 assessments, and some of the items that were identified for 4 discussion fall into two categories When we dealt with 5 GESSAR-II PRA review -- and Debbie has already alluded to this 6 -- we were looking at full core meltdown events.
7 The initial submittal did not include any provision 8 for hydrogen control. There was a very high probability of 9 containment building failure in the PRA and signficant 10 probability of early loss of drywell integrity, so we weren't 11 really looking at what were acceptable hydrogen generation 12 rates, how much hydrogen should be oxidised. Debbie pointed 13 out you don't need to oxide a great deal of the cladding
[v 14 before you run into trouble.
15 So there we were looking at full core meltdown 16 events and not dealing with the specific topics that were on 17 the agenda, and again, the containment event trees that went 18 into Supplement 2 of the SER dealt with full core meltdown 19 events and not these degraded core events that we were l
I 20 discussing with HCOG.
21 So the impact of hydrogen control was really only 22 addressed in Supplement 4 to the SER.
I 23 [ Slide.3 24 My second slide goes through some of the 25 interactions that have been going on between the NRC and
49 1 HCOG. Here we are dealing with degraded core events. The aim
( 2 is to maintain containment integrity. Therefore, we are 3 interested very crucially in the amount of hydrogen produced 4 and the rates of hydrogen production so that we could, in 5 fact, design the system and ensure that the system works.
6 And the issues that were brought up by the ACRS 7 related to optimum ignition sources, type of power supply, 8 limitations of the sources and the effect of standing flames, 9 really impacted this particular assessment. And again, I have 10 a Vu-geaph that will touch on the status of our interfaces 11 with HCOG.
12 But the potential for hydrogen detonation, which is 13 one of the items that you brought up, was related specifically 14 to our assessment of GESSAR in the absence of hydrogen 15 control.
16 [ Slide.]
17 What I did with the subcommittee yesterday, and I 18 think it worked well and we will try to do the same thing 19 today, is I put up this Vu-graph, which really outlines the 20 various steps that we in the Staff took to determine the 21 potential for damage to the structures as a result of 22 detonations.
23 We have gone through some of these Vu-graphs already 24 with the ACRS a couple of months ago. It was a rather hurried
' So we can go into more detail if there are any 25 presentation.
i
l 50 1 questions.
/ 2 I don't propose, in view of the time, to go through 3 every one of the Vu-graphs that are in your handout.
4 Basically, we looked at hydrogen generation, and as I say, 5 there were not terribly large differences between our 6 assessment of hydrogen generation and those given by G.E. We 7 were both using the MARCH Code and gave our usual generation 8 rates.
9 Again, looking at distribution, both of us looked at 10 the potential for local distributions to occur in the 11 containment building --
12 MR. KERR: As I remember, the MARCH Code will tell 13 you how much hydrogen you generate if you tell it how much to
\
14 tell you.
15 MR. PRATT: That is certainly true. You know, I 16 have been through this with you many times.
17 MR. KERR: I just wanted to make sure it was still 18 consistent.
19 MR. PRATT: Most of the control that you have, 20 though, is really in the core slumping phase of things.
21 During the initial core degradation --
22 MR. KERR; I am reminded because within the week, I 23 have heard an exposition on what is called tweeking the MARCH 24 Code.
25 MR. PRATT: You are referring to last week's
51 1 presentation, yes. I was there, too. So again, as you point O
() 2 out, you can get wide ranges in hydrogen generation rates and 3 total amounts of hydrogen released. And just a point that was 4 brought up earlier in terms of the way we tried to go with our 5 assessment, when we quant:fied our containment event trees, we 6 did go with a central estimate on what the event tree should 7 look like and that is coming from our assessment in MARCH 8 in its reasonable form.
9 You can go to extremes in MARCH and get more severe 10 consequences and more optimistic assessments and, again, get 11 less impact, so we did hgve a high containment event tree and 12 a medium and a low, and when we looked at our cost-benefit 13 analysis, we were looking at our best estimate containment 14 event tree. We really only looked at the high range when we 15 were looking at our source term uncertainty study.
16 And there again, just so that we are not confused, 17 our high range on the source term was not as extreme, for 18 example, as you have heard from QUEST, where they went very 19 extreme and had an extremely wide range of uncertainty.
20 I have Vu-graphs that I gave to the subcommittee 21 yesterday which will show you very clearly that the upper 22 range of our source term estimates was very close, and in 23 certain cases were not as high as some of the mechanistic 24 calculations that you would get from the suite of codes coming
'" 25 from ASPO.
52 1 So it was an attempt to make a very realistic 2 assessment of what the range might be. We did not go to the 3 extremes that, for example, QUEST went to, and I think that is 4 a very important point 5 MR. MARK: A question. You have headed this slide 6 " Hydrogen Detonations."
7 MR. PRATT: Yes.
8 MR. MARK: Does that imply that you really have no 9 concern over hydrogen deflagration?
10 MR. PRATT: Oh, not at all, n o ~. This is purely 11 because the agenda item pointed out was for detonations, and 12 what I wanted to do is to note on this Vu-graph the various
[~ 13 elements that went into that assessment so that if there were k
14 questions on & particular area, we could then go to the more 15 detailed Vu-graphs and discuss that.
I 16 So no. In fact, at the end I have a couple of 17 Vu-graphs that Jack Rosenthal, I think, gave to you at the 18 last meeting that gives you the general effect of detonations 19 versus deflagrations and how risk changed.around doing that; 20 but most of the time you would expect deflagrations, and it is 21 only very rarely that we get the detonations.
4 22 MR. MOELLER: What does DDT stand for?
23 MR. PRATT: This is the transition of deflagration 24 to detonation, transition, and it's once you start a 25 deflagration, how that would accelerate and go sonic and
- .- . ._ _ - = - -. .._ _- . -- _ ~. - --
03 1 become a ahock wave. So again, as I say, the hydrogen 2 generation rate was based on parametric studies with the 3 MARCH.
4 For distribution, we used an in-house computer code 5 to look at hydrogen distribution and checked that against 6 HECTOR, which was developed at Sandia as part of a research 7 program for NRC.
8 By looking at the distribution as a function of 1
9 time, you could find out over the timeframe of interest when j 10 you might get detonation, when you may get detonable 11 mixtures. And just to address a point that Debbie made 12 earlier, we feel that local detonations will cause damage.
13 There were -- I wouldn't really call them local detonations.
, ~
14 The distribution that we looked at -- and I have a typical one 15 here in your handout --
i 16 [ Slide.]
17 This shows the distribution across the various i
18 volumes that we set up. At the high range, we are talking 19 about a 20 percent mixture of hydrogen, and at the lower 20 range, about a 15 percent. So again, propagation of the shock 21 wave from the enriched area into the lower enrichment, I 22 wouldn't think would affect the shock wave movement.
23 Therefore, I really wouldn't call this a local 24 detonation because it was traveling from the rich region into
\- 25 the weak region and failing at this point. The assumption was 4
..-,---._,.__._._-n,--.,..,..nw .c_.,_.- , - , . . ,
54 1 that the reduction in concentration did not slow down or 2 affect the shock wave.
3 MR. MARK: Well, it will cut the pressure by a 4 considerable factor.
5 MR. PRATT: Moving into this region?
6 MR. MARK: Yes. The detonation, if it propagates at j
7 15 percent, is certainly not the same thing as propagating at 8 30 or 20.
9 MR. PRATT: At 20 percent, the difference is e
10 significant between this? In the calculation that we 4
11 performed --
12 MR. MARK: Something like the factor in the energy 13 source.
14 MR. PRATT: Well, as I mentioned when we met last 15 time, the calculation that was performed by CSQ at Sandia was 16 based on an 18 percent uniform across the containment ,
17 building, so there may be -- what you are saying is for this 18 region there may be a. slight -- well, there could be a 19 significant increase in the pressure pulse.
j 20 MR. MARK: Well, it will -- I'm not sure just how.
l 21 Maybe Lewis can tell us. The energy in the burning front is 22 going to be much affected by the concentration of hydrogen, 23 essentially linearly, and that will reflect in the pressure 24 and the pressure history. That is really quite a bit beside i
O 25 the point. You might get the propagation you speak of.
55 1 MR. PRATT: Yes. Well anyway, just so you know the 2 the calculation that was performed, as we mentioned yesterday, l
3 we did two calculations. One, of course, was the CSQ 4 calculation that was performed at Sandia, which was based on a 4 5 uniform 18 percent concentration, and the other one was to 1 6 look at the shock wave as a simple Chapman /Dugay calculation, i 7 and they were sufficient to fail the structures, we felt. By 8 conversion of the next step --
9 [ Slide.]
10 MR. MARK: On that slide, your conclusion is that 11 the hydrogen distribution, except for a rather transitory 12 period, is effectively uniform. There is a period of -- I 13 don't know what that scale is. How long is it out of s
14 equilibrium?
15 MR. PRATT: Well --
i 16 MR. MARK: For about ten minutes.
17 MR. PRATT: Yes, that's right. But be careful. This 18 we kind of chopped off here. This would be the in-vessel 19 hydrogen production, and then you would expect after vessel 20 failure that this would start to build up again as a result of 21 core-concrete interactions.
22 I have separate slides -- I didn't bring them with 23 me -- that show the effect of the additional hydrogen 24 generation at later times.
25 MR, MARK: Look, it will raise the level, but it
56 1 will still not disturb the uniformity.
\_/ 2 MR. PRATT: Right. And I think when you look at the 3 containment event tree for the later times, you will find that 4 there is much less distribution in the phenomena because of 5 that very fact. What we find is that during the early stages, 6 there is possibility of deflagrations and so on, whereas later 7 on, because of the later times and the fact that it is not 8 distributed as well, you do see a difference in the 9 containment event trees, that's right.
10 [ Slide.]
11 So again, after we move on from the detonation, wo 12 then looked at the potential for detonations to occur, and the 7
b) 13 assumption that we made, as I mentioned, is that if it 14 above 18 percent to 20 percent, we would get detonation.
15 looked at the magnitude of the shock loading. I didn't 16 mention that we used two assessments, one the Chapman /Dugay 17 and the other one the CSQ load, ani then we converted those 18 dynamic loads through an equation that G.E. used, which looks 19 at the materials that the shock wave was seeing, and converted 20 the equivalent shock loading into an equivalent dynamic static 21 loading.
22 Let me say that again. Converted the equivalent 23 dynamic shock loading into an equivalent static load, and g-w 24 compared that static load against the estimates of the U 25 capability of the structures. We did have some discussions
_ . . _ . _ . ._ _ . _ . _ = _ - _ _
57 1 yesterday with the subcommittee on the appropriateness of that
(~~
I 2 in particular.
3 So again, if there any particular quotions that 4 anybody would like to go into on any of these items, otherwise 3 I don't plan on going through all of the Vu-graphs, in the 6 interest of time.
7 MR. WARD: I guess there are none.
8 MR. PRATT: Okay. Let me just move on, then, to the 9 last couple of Vu-graphs, which illustrate the --
10 [ Slide.]
11 This will give you a very brief status of the 12 interface that is going on between the NRC Staff and the 13 Hydrogen Control Owners Group. The Staff position regarding 14 accepability of hydrogen release histories is defined in .
15 letter from Bob Bernero to Hobbs dated June 24, 1985, and I 16 think the issues that the NRC were bringing up related to 17 hydrogen release rates and the amount of hydrogen oxidized 18 were relevant to this particular consideration.
19 And basically we have arrived a t. three cases that 20 the HCOG people should look at in terms of encompassing a 21 range of degraded core sequences that they should model in 22 their quarter-scale test program.
23 The first case assumes that you would have
-~ 24 restoration of basically CRD flow, about 150 gallons per v' 25 minute started at about 3000 seconds after scram, look at
58 1 the hydrogen release rates and generation rates for this O
k, g 2 particular sequence as being typical of a degraded core event.
3 Another limiting --
4 MR. KERR: Excuse me. One does that looking with 5 MARCH 7 6 MR. PRATT: Well, we have looked at it with MARCH, 7, the Staff and the consultants. The Hydrogen Control Owners 8 Group have their own computer code, which is a core heatup 9 code, and a good deal of time and money is being spent on 10 ironing out differences between the version of MARCH that wo 11 have -- bearing in mind the version of MARCH that we used here 12 was different from the version of MARCH that we used in our 13 GESSAR review. We went into this a little bit at the last w
14 meeting. GESSAR was based on --
15 MR. KERR: No, I just asked because this implies 16 that MARCH can tell you what the condition of the core is at 17 50 minutes after scram, sufficiently well for one to know the 18 hydrogen production produced by 150 gallons per minute flow.
19 And I admire -- well --
20 MR. PRATT: I hear what you are saying. As I said 21 before, this is an early stage of core damage, and a good deal 22 of the interface that went on between ourselves and HCOG 23 related to the condition of the core, questions of when you 24 would put off the oxidation as a result of core damage, bypass 25 of damaged reasons and so on.
59 1 So there was a good deal of code comparison in the
- s. 2 calculations. The version of MARCH that we are using here is 3 MARCH II, and it does involve a modeling of the channel boxes, 4 which we didn't have in the earlier version of MARCH. You 5 know, in the old days, we used to have to lump that into the 6 fuel to try to get the right amount of zircalloy and so on.
7 So none of that was in there.
8 We were explicitly modeling these things. There were 9 differences. They had nodes within the fuel an/ we did not and 10 so on.
11 But the predicted differences during the early stage 12 when you're looking at degradation, you are really limited by 13 steam, and it's really not a terribly hard calculation to
)
14 perform; you have X amount of steam and you know what you're 15 going to get out. Really, where the problem arises is when 16 you try to restore the water. In this particular case it's so 17 fast you really don't care. I mean, the reflood process is 18 very fast. You can get a very sharp spike of hydrogen 19 generation and it's all over.
20 Here, it's just about balancing it- And things that 21 MARCH does -- for example, it's a systems code so there's a 22 feedback of pressure. And just the changing of the enthalpy 23 of the steam, or of the water, in the bottom head can give you 24 an additional steam source in MARCH, whereas the HCOO code, t
25 which is just looking at the channel, fixed boundary
60 1 conditions, that specifies the flow rate doesn't see that.
2 So there is a lot of subtlety that went into the 3 reflood process, an additional hydrogen steam sources and so 4 on. So while we have quoted these three cases, there are 5 modifications to the code that will be input to the code to 6 get it closer to what we would feel comfortable with, so t h 'a t 7 we are all talking about the same types of generation rates.
8 But you are right, there is a good deal of 9 uncertainty here. And in Case C it's really an attempt to say 10 okay, we have to reach the 75 percent metal-water reaction 11 which is in the rule, so we have got to just tack on an 12 artificial tail to reach that at the end. But the types of 13 flow rates you get .n here do vary, and of course, you will 14 terminate the accident sequence long before you get to this 15 point by these two cases [ indicating).
16 MR. MARK: What, in pounds per minute or seconds, 17 what does Case A lead to in hydrogen generation rate?
18 MR. PRATT: Oh, I wish I had brought you the 19 graphs. We have literally hundreds. I will have to send them 20 to you. It falls short of the 75 percent metal-water 21 reaction, and it reaches a peak of about --
22 MR. MARK: Well, in some of them you have 20 pounds 23 per minute, or things like that, of hydrogen appearance. Is 24 that consistent with what is up here?
25 MR. PRATT: If I recall, -- and again, I'm thinking
.-. ~ , . - . - ._- - -_ - - - - - _. - -. .- __ _ - _-.
61 1 back to a lot of the sensitivity studies that I saw Dr. Yang 2 do, who is in my group at Brookhaven, but I think the maximum 3 we saw about here was at the peak about 80 pounds per minute, 4 80 pounds per minute peak for a very short time, and then it 5 came down.
l
. 6 MR. MARK: Okay, that's good enough. That means 7 half the water is getting used or converted, t
- 8 MR. PRATT
- In certain cases, that's right.
9 MR. EBERSOLE: Has anyone looked at the probability 10 of this interesting window through which there flows a chance 1
11 of all this hydrogen generation? Because you get get just the 12 right amount of water but not enough to cool the core. It 13 seems like that is a fairly narrow window, if you get just the right amount of water to do all this, but you don't get m
- 14 15 than that and thus accomplish cooling, or none at all. Isn't 16 that a pretty improbable state of affairs?
17 MR. PRATT: Improbable, did you say?
18 MR. EBERSOLE: Improbable.
19 MR. PRATT: Yes. I think that's true. I think the
( 20 way this has been handled, though, is more from a 1
21 deterministic approach.
22 MR. EBERSOLE: That you have exactly the right 23 amount to do the wrong thing, f
~3 24 MR. PRATT: Right.
l' 25 MR. EBERSOLE: And what are the odds that that can t
i
62 1 occur?
I I wouldn't dare guess.
( ,/ 2 MR. PRATT:
3 MR. EBERSOLE: Mr. Bernero is gone. Oh, well, he 4 wouldn't have it either.
5 MR. PRATT: I think there is a difference between 6 this approach and what we are doing in the PRA's in the sense 7 that, you know, this is an attempt to try to scope the types 8 of hydrogen generation rates that one might get, however one 9 does it, and to try to do it to the best of our ability. We 10 are not purposely building into any of this assessment 11 anything that is physically unreasonable. But again, the 12 probability of this occurring, as you say, could be pretty 13 low.
14 MR. EBERSOLE: What you're doing is driving the 15 conditions towards optimization of hydrogen production.
16 MR. PRATT: Yes, this is miserable. You know, this 17 is terrible, and it's terrible to analyse. That's exactly l
18 right, And we spent a tremendous amount of time really i 19 tracking down, you know, really the guts o. f the computer codes l
20 to come up with the difference that we're coming up with here.
l 21 MR. MOELLER: In Case C, how long would it take to 22 reach 75 percent metal-water?
23 MR. PRATT: I'm not sure. I can't answer that, I'm 24 sorry.
25 MR. MOELLER: You couldn't ballpark it for me?
63 l
i 1 MR. MARK: Well, you get 7000 pounds of hydrogen if 2 you take 75 percent of zirconium, so it would take several i
3 thousand seconds --
4 MR. MOELLER: Times 10.
! 5 MR. MARK; Which is to say some number of hours.
6 MR. PRATT: Yes. I just did a little calculation i
4 7 here. The numbers that I seem to remember, I would think it i
8 would take at least 1000 seconds. At least. Additional. !
4 9 Because the numbers that I seem to remember from this 10 calculation were in the region of about 1400 pounds of i
11 hydrogen, and you're trying to get to about 27 or 26. So it's 12 about 1000 seconds, which is maybe a half an hour or so. !
13 MR. MARK: Ten thousand seconds is three hours.
[V]
- 14 MR. PRATT: Okay. The three other points here --
15 and these address some of the concerns of the ACRS -- this 1
- 16 test program is really designed to look at the adequacy of the [
17 deliberate ignition. There was a question asked yesterday 18 about whether or not they were going to look at distribution, 19 and I spoke with the people at Containment Systems Branch at 4
20 NRC. The aim of the program is to try to look at it as built 21 initially.
22 There is some talk that they may not switch on all 23 of the deliberate ignition devices but keep some of them out i
3 24 of the otroult to see what the effect might be. But that 25 would be the only attempt to look at it, as I understand it.
l 1
.,_.m- _
. - _ _ _ - - , _ _ _ , . _ - , _ , _ _ _ _ _ _ _ _ . . . , _ _ _ _,____ , , . _ _ , - , _ _ , _ _ ~
64 1 1 think that was a question you brought up yesterday.
2 And again, they are really not looking at optimum 3 ignition sources; they are really taking the source that will 4 he installed in the plant and testing the as-is situation.
5 MR. EBERSOLE: Well, it's not possible to have too 6 much ignition or too much distribution; it's just an economio 7 or practical matter.
8 [ Slide.]
9 MR. FRATT: The final point which we wanted to touch 10 on was the effect of wetwell hydrogen flames, and I think 11 we're pretty well in agreement with GE in this regard. There 12 was an assessment performed by Appendix A to the Containment 13 Forformance Working Group reports that is NUREG-1037. The
)
14 report is out for distribution and for comment at present.
15 The heat fluxes were provided to the containment 16 performance working group by the containment loads working 17 group, and they are defined in NUREG-1079, and there is 1 16 believe a copy for comment of NUREG-1079 out now.
19 And the calculations reported in Appendix A indicate 20 that the seal temperatures will increase but they were att11 21 below the estimated fatture point. There is an additional 22 point that I have in here whtoh really deals with full core 23 meltdown events and not the standing flames that we see in the 24 degraded core events, which would imply that at later times, 25 because of high drywell temperatures during core-concrete
65 1 interactions, that one could exceed the failure limit of the
\ss/ 2 seals at later times as a result of those high temperatures !
L 3 and then introduce pool bypass at that time.
4 If there are no more questions, the second 5 presentation that I have deals with really the thermal 6 degradation of the support of the vessel. And again, I will 7 abbreviate this presentation relative to the presentation !
6 gave to the subcommittee yesterday.
9 The three items of importance related to ablation of 10 the support, and Debbie has already mentioned that the 11 calculations that the committee were given by Jack Rosenthal 12 were based bastos11y on a concrete support structure. And she j 13 has now shown that the support would be transmitted to the (O}
14 steel. And in fact, if the outer steel supports the vessel, 15 there may not be a problem.
16 What I would Itke to do is skip over the viewgraphs 17 that were given to you by Jack Rosenthat that looked at the 18 ablation front of the temperatures, and show you what the 19 significanoe of the loss of containment integrity would be in 20 terms of the impact on risk, and that might give you some 21 perspective as to what this effect would do. And then there 22 was a final item that you wanted to talk about relating to the l 23 effect of containment venting, and 1 have one viewgraph whleh 24 summarises that effect.
25 So I will skip over the drawings you have of the l
a
i 4 44
, 1 various structures and move to this parttoular slide.
2 (511de.)
3 The table that I showed you earlier at the start of 4 my presentation. Table 15,9 of Supplement 4 to the SER, can be l
5 used because it itemises the contribution to risk of the 4 various fatture modes. You can use that to estimate the 7 impact of early loss of containment integrity, for example, as 4 a result of movement of the vessel. An early loss of 9 oontainment integrity plus loss of drywell integrity. And l
{ 10 simply, if you lose containment integrity, the late l 11 oontainment failure mode calculations, such as the L2 and L3 12 -- the L refers to late containment failure -- would become 12
) 13 and 13) that's an intermediate failure.
14 ff you lose also, in addition to that, drywell 15 integrity then the situations in whtoh you had 3's, whleh it imply suppaession pool scrubbing of all of the fisa 17 products would move from a 3 to a 2. So by simply looking .s le that table, we can go to the bottom line results.
19 !! you lost containment integrity for all o.*the 20 sequenses as a result of this phenomena early on, shortly 21 after vessel failure, the increase in risk -- and this is 22 without Uppt installed -- would be a very modest increase. If 23 you lose containment integrity plus a loss of drywell wall, 24 your person rem estimates would go up by less than a factor of 25 2.
67 1 So again, this is to give you some feeling for if O 2 you go to the very conservative assumptions that you would 3 lose drywell integrity and containment integrity as a result 4 of slippage of the vessel early on, this is the effect you 5 would get. I think the calculation as presented by GE would 6 indicate that this event would not occur earlier on and would 7 ocour much later in the accident sequence. So this gives you 8 an upper bound as to how things might change.
9 CSitde.3 to- There was a question or an iss'ue related to 11 containment venting, and this is a viewgraph bastoally that 12 l've put together and doesn't necessarily represent the staff
() 13 14 position, but I think it does summarise the thinking that's going on.
15 " Clean" venting of the containment butiding would 16 he designed to really address the Class 2 accident sequences 17 plus, too, it tefers to those accidents that lose containment 18 heat removal. And by attempting to vent, you prevent oore 19 damage. And also, the ATWS sequences for.the same reason.
20 You can effectively get an estimate of the impact of 21 this by looking at that table. But Class 2 sequences, by 22 looking at that table, you'll see are signifloantly reduced 23 anyway by the venting capabilities already built into the UPPS
, 24 system. And the only thing, in our looking at the 25 cost-benefit work that we did at Brookhaven, we didn't take a
68 1 great deal of credit for mitigating ATWS sequences as a result f) of the venting. I think this is something that personally,
( ,/ 2 3 you would prefer to try to manage the ATWS sequence rather 4 than going to a venting procedure to help you there, simply 5 because you're dumping such a large amount of energy into --
6 MR. EBERSOLE: May I make an observation at this 7 point for the full committee? The venting concept of the UPPS 8 is just that; it's a concept, and it's an assembly of loose 9 parts, I would call it, not yet integrated and organized into 10 a modular package. And it's General Electric's proposal and 11 the Staff's proposal to leave it this way until the first 12 application or the first buyer appears, and at that time to 13 have a general set of guidelines and criteria against which
[V) 14 they will construct this.
15 I have a little problem with this because it seems 16 to me that it is not a part of a well-developed standardized 17 package; it's a little bit like the early allowance of 18 the industry to let whoever wanted to to design the aux 19 feedwater systems. And I would rather -- and this is just a 20 personal viewpoint -- see this as a pre-identified and well 21 organized picture, both in the context of cperating it --
22 what you're talking about now -- as well as in design 23 configuration including such independent housing as it may 24 have.
\
O 25 At the moment, it's a number of pieces which are i
69 1 scattered about including piping which traverses the 2 containment building possibly into the aux building and I i
3 would call it sprawled. And thus, it has no integral I
4 character. This precludes it being used at this time for 5 consideration, among other things, of secur'ity. And in short, 6 it's just a concept.
i 7 I think the committee should at least study whether 8 they want to leave it as a loose, sprawled concept or have it 9 pulled together at least a little bit tighter. That's all I 10 have to say. Carry on.
11 MR. PRATT: That is the end of my presentation. ,
12 MR. EBERSOLE: Well, in that case, let's have a 13 little recess of 10 minutes until 10:30.
, 14 MR. WARD: And when we come back we will be in 15 closed session.
16 [Whereupon, at 10:20 a.m., a short recess was taken, 17 the subcommittee to reconvene in Closed Session.3 18 19 20 21 22 23
, 24 O 25
70 1 AFTERNOON SESSION
(
\s,) 2 1:50 p.m.
3 MR. WARD: Our next topic is Item 3, the seismic 4 qualification of equipment in operating plants.
5 We appreciate the Staff and representatives from the 6 industry coming in a few minutes early to accommodate us.
7 Mr. Wylie?
8 MR. WYLIE: Okay. The Subcommittee on the 9 Qualification Program for Safety-Related Equipment held a 10 meeting on Tuesday, August 6th. In addition to myself, those 11 in attendance were ACRS members Ebersole, Reed, Siess and 12 Ward, and ACRS consultant Lipinski, and the Staff Engineer, 13 Tony Cappucci
(
14 The objective of the meeting was to review and 15 discuss the NRC Staff Resolution of Unresolved Safety Issue 16 A-46, Seismic Qualifications of Equipment in Operating 17 Plants.
18 The subcommittee had the benefit of presentations by 19 the NRC Staff, Mr. Newt Anderson and T. Y. Chang representing 20 the Staff, and the Seismic Qualification Utility Group, Jim 21 Thomas for Duke Power and Peter Yanev of EQE, and Mr. Dick i
22 Shaftstall of KMC.
l
[
! 23 The NRC Staff presented a status report regarding s 24 the resolution of USI A-46 and its implementation, and SQUG 25 presented their activities regarding their program, the l -
71 1 results to date and their implementation plan.
2 In addition, Mr. Peter Yanev of EQE presented a 3 detailed report of the most recent investigation by SQUG of 4 the March 3rd, 1985 Chilean earthquake, which provided 5 additional evidence that the type of equipment used in power 6 plants survive large earthquakes very well He presented 7 about 50 slides out of 5000 that they collected in their 8 investigation.
9 I won't go into a lot of detail, since Mr. Thomas is 10 here today to cover that portion of the presentation. But in 11 general, it showed that the power plants, industrial plants, 12 oil refineries in general did very well, and the equipment 13 survived very well.
)
14 The subcommittee decided at the end of the meeting 15 to request the NRC Staff and SQUG to present to the full 16 committee a somewhat condensed version of their presentation 17 in the two hours allotted this afternoon. The presentation 18 regarding the Chilean earthquake will be somewhat abbreviated 19 today, but SQUG has offered to present a full report to the 20 ACRS at a later date if the full committee so desires.
21 In considering the proposed resolution of USI A-46, 22 one should keep in mind that the scope of A-46 is somewhat 23 limited, and perhaps a comment regarding its scope will be of
- 24 value.
25 The scope of A-46 requires the verification of the
72 1 seismic adequacy of equipment required at bring the plant to a 2 safe shutdown condition and maintain the plant in a safe l
3 shutdown condition following an SSE. The scope is limited to 4 active mechanical and electrical components, and the scope is 5 further limited by excluding those issues and items which are 6 addressed elsewhere by other USIs, generic letters and IE 7 bulletins.
8 For example, systems interactions are being 9 addressed by USI A-17. Decay heat removal requirements are 10 being addressed by USI A-45. Masonry wall adequacy is covered 11 by IE Bulletin 80-11, And the accident mitigation systems are 12 not included in the scope of A-46.
l '3 As a matter of background, in July of 1983 the ACRS
(
14 was briefed by SQUG on their program and approved a letter 15 from ACRS Executive Director to the EDO, which encouraged 16 continued work along the, lines presented by SQUG.
17 During the 289th meeting, May 10 through 12, 1984, 18 the ACRS was briefed on the status of USI A-46 and l
19 subsequently wrote a letter to the Commission on May 16, 1984 20 endorsing the approach being taken by the Staff based on the 21 SQUG efforts with comments.
I 22 The NRC Staff has prepared its revised resolution 23 package, which consists of NUREG 1030, Seismic Qualification r
i gsr 24 of Equipment in Operating Nuclear Plants, Draft Report for 25 Comment, which is the supporting technical basis for the i
-m ,
73 1 resolution of A-46.
2 It has prepared the regulatory analysis for A-46, 3 which covers the analysis, the plan, the implementation 4 procedure.
5 A draft letter has also been sent -- has been 6 prepared, which will be sent to licensees following resolution 7 of public comments.
8 CRGR has reviewed the package and approved sending 9 NUREG 1030 and the regulatory analysis out for public 10 comment.
11 The package has been sent to EDO to be sent out for 12 public comment.
i 13 No action is required on the part of the ACRS.
[G 14 However, the ACRS may wish to consider writing the letter at 15 the conclusion of today's presentation, in light of the 16 earlier letter which the ACRS wrote.
17 So with that, unless -- first of all, let me ask the 18 subcommittee members if anyone has any additional comments or 19 statements they would like to make.
20 (No response) 21 If not, I will call on --
22 MR. MOELLER: A quick question, Charlie.
23 You are saying this then covers systems which may be 24 covered in other USIs, but this only applies to the seismic 7--
(
25 resistance of those systems, is that correct?
k -
74 1 MR. WYLIE: That's correct, seismic qualifications s,) 2 of the systems.
3 Let me make myself clear. The equipment and 4 considerations which are being addressed in the other USIs are 5 not included in A-46.
6 MR. MOELLER: Well, you were saying like there is 7 another USI to cover safe shutdown heat removal.
8 MR. WYLIE: That's correct.
9 MR. MOELLER: Okay. But this USI covers electrical i 10 and mechanical -- the seismic capabilities of the electrical 11 and mechanical equipment used?
12 MR. WYLIE: That's right. Used to obtain safe i 13 shutdown.
14 MR. MOELLER: Okay. I'm clear.
15 MR. EBERSOLE: Charlie, I would like to make an 16 observation.
17 The robustness of this equipment which enables it to 18 survive these earthquakes, these other shocks, has been 19 acquired and obtained by longstanding developing practice, 20 probably based on the notion that a lot of this equipment is i
21 going to be thrown around on trucks and bounced on the ground 22 and whatever, at which time it will see greater g values and 23 shocks than it ever will in an earthquake.
24 I can't put my finger on any existing control 25 document that says this will continue. In short, a vendor
(
I
75 1 manufacturer might say, "Oh, I am going to pack my device in w
2 cotton, so I can make it flimsy and it will survive the 3 challenge that I thought it was going to have to survive 4 against much more rigorous c i r cums t s.nc e s . "
5 This is a nebulous nature. Whether we can count on ,
6 a persisting character of robustness in this equipment -- l 't 7 seems to me there has got to be some port of controls which 8 have perpetual significance.
9 MR. WYLIE: I think to some extent what you say has 10 some validity to it, in that you do have to ship equipment and 11 it has to hold together when you ship it. But, it also has to 12 perform once it has been delivered. For example, motors have r 13 to be strong enough to deliver their --
(
14 MR. EBERSOLE: I'm thinking about relays.
15 MR. WYLIE: Relays. Again, there are all kinds of 16 relays. I think that is one of the considerations. Relay 17 chatter is one of the considerations.
18 MR. SIESS: Jesse, it seems that your concern is 19 related to future.
20 MR. EBERSOLE: Yes.
21 MR. SIESS: Those are all covered under the existing 22 criteria. This has to do only with the qualification of 23 equipment that is already in operating plants, the 72 of the m.
24 MR. EBERSOLE: You have got to replace these things f-~
25 as they wear out.
76 1 MR. SIESS: Yes. But the replacement presumably
( ,/ 2 will come under the same criteria as new plants.
3 MR. EBERSOLE: I didn't know that.
4 MR. SIESS: I'm not sure.
5 MR. ANDERSON: I don't believe that is correct as 6 you stated it. I think our proposed position with regard to 7 replacement parts is that the utility would have the option of 8 either using the A-46 developed data base, or qualifying it to 9 current criteria.
10 MR. SIESS: Okay. Now, what would apply to, say, 11 Vogtle, the plant Vogtle, which presumably has equipment
~
12 qualified under the existing criteria, right?
[ 13 MR. ANDERSON: Yes.
14 MR. SIESS: Their replacement equipment would not 15 have to be qualified under the existing criteria?
16 MR. ANDERSON: I would have to look to see if V;g -
17 is on our list. I don't believe that it is.
18 MR. SIESS: Let's take a plant that is not on your 19 list. That is what I am talking about.
20 MR. ANDERSON: I think Vogtle is a good 21 example. There isn't any consideration in our proposed 22 resolution that plants that do not now fall within the scope 23 of A-46 would be allowed to use this in the future.
,, 24 However, as we noted to the subcommittee Tuesday,
/
\
25 the IEEE 344 Committee is proposing a provision which would
77 1 include the use of seismic experience data much in the same fashion that we are doing.
(_ 2 3 MR. SIESS: So they intend to go to a relaxation.
4 MR. ANDERSON: Yes, sir.
5 MR. SIESS: Then, I think Jesse has got a legitimate 6 argument.
7 MR. WYLIE: E cept if it is proven by experience 8 that the equipment is adequate and placed in the data base, 9 then you could use it.
10 MR. SIESS: You could use that' model.
11 MR. WYLIE: You could use that model, 12 MR. EBERSOLE: Sure, if you hold to that model.
[~N 13 MR. SIESS: It seemed to me the SQUG type approach 14 resolution of A-46 makes a great deal of sense for equipment 15 that is in operating plants now, equipment for which 16 qualification under the existing criteria would be somewhere 17 between difficult and impossible.
18 I don't see that similar justification for 19 replacement equipment which can be hauled.out to labs 20 somewhere and tested, or for relays which will undergo quite a 21 research program, and maybe will be qualified on some other 22 basis.
23 But, it seems to me the rationale for experience 24 base on existing plants, and the rationale for using 25 experience base on future plants is somewhat different. And,
78 1 I had not thought of it entirely in that light.
( 2 MR. ANDERSON: I would like to add that we did meet 3 with the vendors on much of this equipment.
4 We talked with the NEMA people with regard to 5 cabinet sizes and design practices.
6 Although in some areas the standards are not that 7 well established, the design practice is established. As a 8 matter of fact, does not appear to be a softening or a 9 cheapening of the manufactured materials that is coming out.
10 We can't guarantee this in the future, but I think 11 that we did establish that it has not happened to date.
12 MR. WYLIE: Mr. Thomas had a comment.
' 13 MR. THOMAS: In that area, what the SQUG program is 14 doing in the implementation plan, is trying to come up with 15 some pretty specific criteria for things that need to be 16 checked, looked at in equipment, to verify that it truly is 17 rugged.
18 These walkdowns play a big part in that. What we 19 have really learned is that there are a lot of things that you l
l
- 20 can look at and by using judgment, probably make in most
(
l 21 cases, as good a judgment as to the seismic ruggedness of the 22 equipment as a test is going to give you.
23 As far as the new equipment, I agree, it is out of l
l- 24 the scope of A-46. But some of the criteria and some of the
\-' 25 things we have learned, I would say probably are in many cases l
79 1 as good as a seismic test if you use those as a screening 2 criteria when you look at the equipment and check it out, 3 check the way it is made, its ruggedness, look at cutout 4 sizes, truly verify that you haven't introduced any kind of 5 sensitive area.
6 The relay issue is one thing that is difficult to 7 do.
8 MR. SIESS: Let's take it one at a time.
9 Taking what you have said, for example, wouldn't 344 10 have some of those criteria in it? So, if you simply said, 11 "That is your standard. Meet 344," people would have to look 12 at those things and establish their ruggedness.
13 Or, is it just something between the licensee and
(
14 the manufacturer.
15 MR. THOMAS: The incorporation of the experience 16 data approach to 344 is just beginning. I think what you are 17 saying is a logical way to proceed to where the standard does 18 give you some of that type guidance in regard to what you 19 would be expected to do in regard to applying experience data 20 to a newly manufactured piece of equipment.
21 MR. SIESS: So, specification or committing to 344 22 would provide a fair level of assurance that significant 23 change in weaklings wouldn't get by?
24 MR. THOMAS: Assuming 344 gets developed as we hope
\
25 lit will.
80 1 MR. SIESS: If it doesn't, of course, we could O'
k,j 2 always have a Reg Guide that supplements it That is not an 3 unheard of practice.
4 MR. WYLIE: Any other comments or questions?
5 [No response.]
6 Why don't we call on the Staff to go 7 ahead. Mr. T.Y. Chang is going to present the Staff's 8 presentation.
9 [ Slide.]
10 MR. CHANG: My name is T.Y. Chang. I am the task 11 manager of USI A-46.
12 Today I would like to present to you the status of 13 A-46.
14 The first Vu-graph is an overview of today's talk.
15 I will summarize the A-46 program, then talk about proposed 16 resolutions, scope and basis and implementation requirements.
17 Then Jim Thomas of the SQUG Steering Committee will present a 18 talk on the ongoing SQUG EPRI activities.
19 Among their activities, anchorage guidelines is one 20 thing being currently developed by the EPRI, and also they are 21 in the process of collecting test data from previous 22 qualifications.
23 Relay review procedure is another thing they are 24 looking into, and they are going to propose something on the 25 line of the generic implementation. It is called SQUG generic
81 1 implementation plan.
r~
_) 2 ESlide.3 3 Since the beginning of the commercial nuclear power 4 plant program, design methods and qualification philosophy of 5 equipment have been changing continuously, so it is a logical 6 question to ask whether equipment in those existing plants, 7 w!. : n they are called upon to perform their safety functions, s
8 will they be able to do so or not if an earthquake comes?
9 One possible choice to answer that question is to 10 impose the current licensing criteria on all the equipment in 11 the existing plants, but upon a closer look, it is apparen*
12 us by using the current licensing methods and criteria that is 13 not practical because the current requirement calls for
)
14 testing those equipments on the shake table in the lab.
15 First of all, you may not be able to obtain a piece 16 similar to the ones in the plant, the sort of vintage that we 17 are talking about. Secondly, you have to worry about the 18 down time once you take the equipment out of the plant. So we 19 are looking for an alternative to the current licensing 20 criteria, which is both effective and also cost effective.
21 MR, MOELLER: Excuse me. I don't follow what you 22 are saying. You are saying you are debating whether to apply 23 the current criteria to the equipment at the plants. I thought gg 24 you had to do , t whether you wanted to or not.
l )
%J We are talking about operating plants 25 MR. CHANG:
l 82 now, not licensing plants. l 1
i s
2 MR. MOELLER: You are talking about backfitting?
3 MR. CHANG: Yes. A-46 is concerned only with 4 operating plants. I neglected to mention that.
5 MR. MOELLER: What you are saying, then, is you are J
6 debating or the discussion relates to whether you should take 7 criteria that have been developed recently and back apply that 8 to operating plants?
9 MR. CHANG: Yes. That comes as a logical choice 10 when you look at it because we have that' current licensing 11 criteria and methods in place for the licensing plant.
12 MR. MOELLER: For new plants.
l'3 MR. CHANG: For new plants, right.
(
14 MR. MOELLER: When were these licensee criteria 15 developed?
16 MR. CHANG: The document, the technical requirement 17 actually is described in Standard 344, 1975.
18 MR. MOELLER: So anything that was finished prior to l 19 1975 --
l 20 MR. CHANG: The requirement for this document l
l 21 actually is for those plants. The CP application, if it is 22 docketed after October 1972, then they have to comply with 23 this document.
24 MR. MOELLER: All right.
l i
j 25 MR. CHANG: So the objective of A-46 is to look for 1
i l
83 1 alternative methods in lieu of using the current criteria to 2 assess the seismic adequacy of equipment, 3 MR. MOELLER: Help me again, now. If a plant does 4 not now have an operating license -- you know, say one is 5 under construction -- do you require that all of these 6 critical components be put on a shake table and tested?
7 MR. CHANG: Yes. They have to comply with the 8 current licensing criteria.
9 MR. MOELLER: Which includes --
10 MR. CHANG: Which includes -- well, testing 11 equipment on a shake table is the recommended method in the 12 actual 344 document.
13 MR. MOELLER: If it can be done.
14 MR. CHANG: They have other alternatives, such as 15 analysis, but for equipment, the functionality cannot be 16 assured by structural integrity, then they have to test it.
17 MR. EBERSOLE: Point of clarification. He said 18 all. What really happens is one out of 500 might. Then they 19 invoke the standard construction on electrical apparatus in i 20 particular, right, type tests.
21 MR. CHANG: Type tests, right.
22 MR. EBERSOLE: Highly dependent on consistency in 23 manufacture, assembly, maintenance, et cetera.
24 MR. CHANG: You don't actually have to test each 25 piece of equipment. A-46 was designated as USI in December l
84 1 1980.
2 MR. EBERSOLE: As a matter of fact, couldn't these 3 earthquake experience records constitute type tests of sorts, 4 the actual equipment of a given model, type and so forth has 5 survived a given g value in the laboratory influence?
6 MR. CHANG: This is what they are proposing in the 7 actual document now, trying to use the seismic experience as 8 an acceptable way in lieu of doing the actual qualifications.
9 MR. MOELLER: When you do shake table tests, and I 10 realize we are not necessarily talking about that, what 11 acceleration forces do you use?
12 MR. CHANG: This has to be site specific. It depends 13 on the site.
)
14 MR. MOELLER: Let's say it is site specific. Do you 15 use the SSE or the OBE or what?
16 MR. CHANG: They have to do a number of OBE tests on 17 top of the SSE.
18 MR. MOELLER: Okay.
19 MR. CHANG: During the process, of course, they have 20 to consider the amplification of the seismic input through 21 that building.
22 [ Slide.]
23 MR. ETHERINGTON: Are they putting an actual 24 earthquake spectrum into the test procedure?
25 MR. CHANG: Yes. For each site in the location of
85 1 the elevation, they have to generate the required response 2 spectra.
3 MR. ETHERINGTON: At one time they were using what 4 they called sign beat testing.
5 MR. CHANG: That was the old method 344, 1971 5 method.
7 MR. ETHERINGTON: Has that been abandoned?
8 MR. CHANG: In certain cases if you can justify it 9 that can still be used, but in the current version of 344, the 10 recommended method is to use a random test.
11 MR. ETHERINGTON: That is much better, of course.
12 MR. CHANG: That is much better, yes.
O 13 Upon looking at several potenti.' alternatives, we 14 concluded that the only viable and practical way other than 15 using the current criteria and method is to use the seismic 16 experience data approach. The feasibility of this approach 17 was established by two independent studies. One is done by 18 Lawrence Livermore Labs. The other one was initiated by a 19 utility group, SQUG, Seismic Qualification Utility Group.
20 They started to look at earthquakes in California.
21 The question coming to their mind is, firstly, is there enough 22 information they can gather from those earthquakes? And the 23 second question, obviously, is is there any similarity between 24 equipment in those nonnuclear plants and equipment in the 25 nuclear plants?
86 i To answer those questions, they looked at two 2 California earthquakes. One is 1971, San Fernando 3 earthquake. It is a magnitude 6.5 earthquake. The other one 4 they looked at is 1979 Imperial Valley earthquake, magnitude 5 6.6 6 They looked at a number of nonnuclear plants in that 7 area, and they looked at six nuclear plants and compared the 8 equipment in both categories, and it turned out to be that 9 they are quite similar regardless that it is in a nuclear 10 plant. A piece of pump is a piece of pump, basically. There 11 is no major difference. And for motor control centers, things 12 like that, there is no difference whatsoever.
13 The only obvious difference is that for nuclear 14 plants, you require a piece of paper to prove that.
15 In their pilot program they looked at eight classes 16 of equipment, both in the category of a mechanical equipment 17 and electrical equipment. Those eight classes were chosen in 18 such a way that they will be representative of equipment in 19 the plants, and also they constitute a big percentage of all 20 the equipment in the plant, 21 The eight types of equipment chosen are motor 22 control centers, low voltage switchgear, metal clad 23 switchgear, unit station transformer, motor operator valves, 24 air operated valves, horizontal and vertical pumps.
( What was the criterion used to select 25 MR. MOELLER:
87 1 those particular eight items?
2 MR. CHANG: As I mentioned earlier, they think this 3 constitutes a pretty representative cross-section of both i I
4 mechanical and electrical equipment in the plants. l l
5 MR. MOELLER: Would it represent 90 percent of the 6 risks?
7 MR. CHANG: They did not consider this from the 8 point of risk.
9 MR. MOELLER: Maybe I should have said do they 10 represent 90 percent of those components most likely to be 11 subject to a seismic failure?
12 MR. CHANG: Not from the seismic failure point of I' 13 It constitutes a big chunk of all the equipment U' view.
in the nonnuclear plants and nuclear plants. They probably 14 15 represent 80 percent of all the equipment someone would like 16 to look at. .
17 MR. MOELLER: I would have thought there would have 18 been a basis for the selection. In other words, these were 19 the components most likely to fail or the. components in 20 greatest use or something?
21 MR. CHANG: Keep in mind this is a pilot program.
22 They are starting from scratch. They just want to get some 23 idea as to how the equipment qualification question can be i
24 answered by comparing nuclear to nonnuclear plants from O 25 earthquake experience point of view.
88 1 MR. THOMAS: I would like to add a little to that, fx 4 k 2 What you just said happens to be correct. We tried to pick 3 the equipment that was most utilized in the systems and also 4 the ones that would -- just backing up and taking a big look -
5 might be most susceptible to seismic-type forces.
6 So just as an example, you are going to see hundreds 7 of valves, hundreds of motor operators in plants. You are 8 probably only going to see one inverter. So we picked valves, 9 motor operators for valves. You look at them. So what they 10 said, they are the ones we felt were a group that we could 11 kind of hone in on that was truly representative of very large 12 amount of the equipment.
l 13 MR. CHANG: Again, I think that is from a numerical
[V 14 point of view that that represents a big percentage of the 15 equipment in a plant. When you talk about earthquake loading, 16 every piece of equipment will see earthquake loading in a 17 plant. In their pilot program, the data were collected from 18 the two earthquakes I mentioned, the San Fernando and the 19 Imperial Valley earthquakes, but later, in 1983, Coalinga 20 earthquake happened, so they managed to incorporate the 21 Coalinga data into their pilot program as well 22 In 1984, Morgan Hill and the recent Chilean 23 earthquake. They went there and tried to collect information 24 from those earthquakes in order to expand the data base from
( 25 the pilot program.
_ _ _ - _ _ _ _ . _ _ - . _ . _ . _ _ . _ . . - - - _ . _ _ _ _~___ _ _ _ . . _ . _ _ _
89 1 I think I should mention here the Chilean earthquake C
2 is a big league earthquake, magnitude 7.8. Jim is going to 3 talk about that. He is going to show you some slides. From 4 the building and structural point of view, there is a lot of 5 damage seen there. However, the performance of equipment is 6 very remarkable. You can see from those slides.
7 A panel, SSRAP, Senior Seismic Review Advisory 8 Panel, was formed in June 1983. This panel was jointly 9 selected by SQUG and NRC. The function of the panel is to 10 provide expert opinion and advice as to how to use the seismic 11 experience data, how this should be utilized in our review.
12 There are five members on the panel. They are all b)
( 13 acknowledged experts in this area, and two members are from 14 engineering consulting companies. One is from a university, 15 one is from a test lab and one is from a national lab.
Y 16 Since the formation in June 1983, they started to 17 look into the SQUG pilot program. They actually asked SQUG to 18 go back and reco11 ate *the data in a more systematic fashion 19 that they can readily utilize, and at the end of 1983, they l
20 formulated a tentative guideline as to how the experience data 21 should be utilized to assess the equipment seismic adequacy.
22 That report was revised twice, and the final version t
23 was out January 1985, this year. In that report, the rules as 24 to how those experience data should be utilized were laid out, 25 and in that report, so-called bounding spectra, that spectra
90 1 represents what they think a conservative estimation can be 2 from those earthquakes that should be utilized in our review. ,
i 3 Those are the spectra that the equipment actually l l
I 4 have seen during the earthquake. So in a sense, this is sort 5 of like test spectra in the seismic qualification, but this is 6 a much conservative approach. They reduce the actual size 7 spectra by a certain factor just to make it more conservative.
8 The NRC Staff participated in all the SQUG and SSRAP 9 meetings and closely monitored their progress, and the present 10 proposition from the Staff is based mainly on the seismic 11 experience approach.
13 I might add that when we talk about seismic 13 experience approach, we are not limiting ourselves to only the 14 earthquake experience, but we are covering the test data 15 experience as well. This is the effort that EPRI is currently 16 doing and working on.
17 [ Slide.]
18 Those are either types of equipment I mentioned 19 earlier that are being looked at in the pilot program.
20 Roughly, there are 3000 pieces of equipment included in their 21 pilot program.
22 When we talk about the number of equipment there, 23 the devices, for instance, relay switches and things like that 24 in the electrical cabinets are not counted as one piece of 25 equipment. We only count the electrical cabinet as one piece i
91 1 of equipment. So if you go down to the device level, actually 2 it's much more than 3000 pieces.
3 [ Slide.]
4 These are the conclusions of review by the SSRAP on 5 the SQUG pilot program. First of all, they think the 6 equipment installed in nuclear power plants is generally l
7 similar to and at least as rugged as that installed in 8 conventional power plants. And when this equipment is 9 properly anchored and with some reservations, those 10 reservations are described in the report. They call it 11 caveats or exclusions. I will go into those things a little 12 bit more later on.
13 When those eight classes of equipment are properly 14 anchored, then they concluded that those equipment have an 15 inherent seismic ruggedness and have demonstrated capability 16 to withstand substantial seismic motion without structural 17 damage. This substantial seismic motion they were referring 18 to is actually the earthquake input they have seen at those 19 two earthquake sites in California.
20 MR. KERR: What am I supposed to learn from that 21 second statement?
22 MR. CHANG: From this statement what we're saying is 23 yes, there may be some concerns about equipment. Actually, 24 it's in the area of anchorage. Once you assume the anchorage N
25 is adequate, then the equipment is really pretty rugged as far
92 1 as seismic loading is concerned.
1
( j 2 MR. EBERSOLE: Isn't it just the end result of 3 having to package it for handling prior to installation?
4 MR. CHANG: That could be a result of that.
5 MR. KERR: Does that tell me that anything more 6 needs to be done, or that something needs to be done?
7 MR. CHANG: That tells us that at least we have to 8 look at the anchorage, because this may be an area that will 9 cause some problems. Because during their review of the data 10 base plants -- and we got that input from the ACP program also 11 -- if you don't have anchorage or if the anchorage is not 12 adequate, then the equipment is likely to be shifted during 13 earthquakes, or it can even overturn and cause some damage to U'~N 14 the equipment.
15 MR. KERR: I'm surprised it took a study to reach 16 that conclusion.
17 MR. CHANG: This is actually what you see from their 18 experience.
19 MR. KERR: Thank you.
l I guess, though, one could question ;
20 MR. MOELLER:
21 your first bullet equally. I would that equipment installed 22 in nuclear power plants is as rugged as that installed in 23 conventional power plants. Why wouldn't it even be better if I
24 it has a better pedigree in terms of QA and so forth?
(
i 25 MR. CHANG: Well, I think that's why they say at I
! I i
- - _ . . _ _ _ _ _ _ _ _ . . . _ _ . _ _ _ . _ . . _ _ _ _ _ _ _ , ___ . _ _ _ _ __ _ __ . _ _ _ _ . . _ . _ ._ _ __i
93 1 .least as rugged. Their feeling is probably that equipment in b)
(, 2 nuclear power plants should be more rugged.
3 MR. ANDERSON: That's a necessary conclusion to go 4 'o the point where you're going to use the seismic experience 5 in the plants, to the nuclear plants. We think it's much more 6 rugged. We're trying to establish that similarity.
7 MR. MOELLER; Okay. In that light, I fully 8 understand.
9 MR. AXTMANN: What does the phrase "with some 10 reservations" mean, in the second bullet?
11 MR. CHANG: Those are the caveats and exclusions, 12 because the number of equipment in those plants cannot cover.
[ \ 13 all the equipment in nuclear power plants.
U 14 For instance, when you talk about motor-operated 15 valves, if there are valves only up to -- the motor-operated 16 is only that much apart from the valve body, say six feet.
17 Now, if you have a piece of valve seven feet from the body in 18 a nuclear plant, apparently that piece of equipment cannot be i
19 evaluated by using the seismic experience data base approach.
20 So those are the things they spelled out as the caveats and 21 exclusions to the use of the experience data.
, 22 Other things such as they would think a cutout from i
23 an electrical cabinet cannot be bigger than six inches by 12
-- 24 inches, and things like that. Some of them are based on their 25 engineering judgment; some are based on the extent of l
.g-. -4* A - -,. &-g----M- - - - d ---.-u -- - - - - - A -AA4--.-4-h- M-h-R- h A--'8-=hM - h----1----sL-++4- - = = + - - - - - - M8 .i----a-.*a-wELw--
94
- i 1 equipment they can collect i n t .50se experience data base Q 2 plants.
3 MR. AXTMANN: Thank you.
. 4 MR. CHANG: The number 3 conclusion from this study 5 is that functionality after the strong shaking has ended has 6 also been demonstrated. But the absence of relay chatter
, 7 during strong shaking has not been demonstrated.
8 What they are saying is there is evidence to show 9 that devices like relays after an earthquake when they are a
10 called upon to function, they can always function. But i
11 whether they function or not during an earthquake because of 12 the fact that those earthquakes happened early and that kind 13 of information is lost. So that question cannot be 14 established -- the operability of the equipment during an 15 earthquake.
I 16 MR. MOELLER: Excuse me. How big is a relay? Why i
17 can't i t be tested on a table?
I 18 MR. CHANG: It can, and that is being done for the 19 licensed plants; they are all being tested on tables.
20 MR. MOELLER: But have these tests shown whether it 21 chatters during the shaking? I guess I just don't understand I
i 22 the statement. If it's small enough to test on a table I I
23 would think you would have tested it.
24 MR. CHANG: Yes, indeed. During the test for new 25 licensing plants they observed chattering.
95 1 MR. MOELLER: So they did observe chattering, fi
(% l 2 MR. CHANG: Yes.
3 MR. MOELLER: So we can presume then that there's 4 chattering in the old relays, also.
5 MR. CHANG: It depends on the function. Okay? In 6 certain cases the relay chatter may not be a matter of 7 concern. But in other cases, depending on the function of the 8 relay in the system --
9 MR. ANDERSON: Excuse me. I think one thing you've 10 got to remember is that these conclusions are the conclusions 11 of this panel following their look at the pilot program, and 12 this reflects their conclusion w'ith regard to what they 13 learned from the seismic experience data. It's not
(
14 necessarily the conclusions of our program; just what they 15 said.
16 MR. CHANG: Those conclusions are strictly from an 17 evaluation of the seismic experience data on the eight classes 18 of equipment.
19 MR. MOELLER: Right. And there's nothing, no 20 instrument, in these plants that underwent the earthquake.
21 There's no instrument that would have recorded whether the 22 relay chatter --
23 MR. CHANG: Yes, there's no indication whether they 24 chatter or not during earthquake. So from this conclusion we 25 can see that equipment basically is rugged from the seismic
l 96 l
1 1 point of view. j 2 However, there are at least two areas of concern.
3 One is anchorage; the other is the functionality of equipment 4 and devices such as relays.
5 [ Slide.3 6 Now for those eight classes of equipment, we have 7 the guidelines pretty much laid out how they should be used 8 and the caveats and the exclusion rule spelled out in the 9 SSRAP report.
10 Now how about equipment other than the eight 11 classes? Well, I have to say during their collection of 12 seismic data from those plants, they did not restrict their 13 effort to only the eight classes. They looked at, from the 14 damage point of view, they tried to look for damage to all 15 kind of equipment during the survey. Well, those eight 16 classes are the ones that are being looked into in more detail 1
17 and the data are being presented in a systematic way.
18 However, the ruggedness of equipment other than the 19 eight classes is actually being looked at ,and assured from the l
20 survey as well. So our recommendation for equipment other l
t i .
l 21 than the eight classes is that yes, we believe that they are i
22 rugged from this survey. So we don't require any additional
- .3 collection of seismic experience data for the other classes of i ,,g 24 equipment. However, we will require the utility to document 25 their basis for seismic adequacy for their remaining classes
97 1 of equipment. This can be provided by either verification
/
( 2 that the equipment exists in the data base plant; in other 3 words, they have to prove a similar piece of equipment is 4 really in existence as in those plants surveyed in the 5 California earthquake. So they actually underwent that type 6 of loading.
7 The alternative is by comparing it with the test 8 data. This is currently being collected by EPRI. What EPRI 9 is doing now is they are trying to collect test data 10 information from test labs, from vendors, from AE's, from 11 various different sources. They're trying to come up with 12 some sort of conservative envelope for the test spectra and
/\ 13 they call that GERS, generic equipment ruggedness spectra. So 14 in essence, spectra is similar to the bounding spectra that 15 SSRAP is proposing.
16 However, the GERS come out to be higher than the 17 bounding spectra because normally the way you test, the test 18 input will be higher than the actual earthquake input you saw 19 at the California sites.
20 [ Slide.]
21 In conclusion, after all these studies we have 22 concluded -- SQUG and SSRAP, they all agree, we all agree on 23 these three points. The only three concerns on the seismic 24 adequacy of equipment in the operating plants are the p_
.( )
7# 25 following. The first one is the equipment anchorages. We'd
98 1 better take a look at anchorage to make sure they're adequate.
O)
( 2 The second concern is the relay operability for i 1
3 those relays that are required to function during 30 seconds i
4 or so strong motion earthquake. You had better prove that i
5 it's going to function.
6 The third one is outliers. When we say outliers we 7 mean those things I mentioned earlier, the caveats and these 8 exclusions. We'd better have those rules spelled out clearly.
9 [ slide.]
10 With the background I discussed earlier, the 11 following is the proposed resolutions. Well, let's divide 12 them into two groups; operating plants and new licensing 13 plants. For new licensees there's no requirement from the 14 A-46 point of view. They have to satisfy the current 15 licensing criteria for operating plants. We believe those 16 things should be done by the utility members.
17 First, they have to look at the equipment from the l 18 systematic point of view, develop an equipment list that i
19 should be looked at. Secondly, a walk-through inspection is 20 necessary in order to convince us that two things will be
- 21 answered from this walk-through; one is to verify the r 22 anchorages; the other thing is to identify and address the l
, 23 deficiencies and outliers.
l 24 MR. EBERSOLE: May I ask a question? I think you 25 said you were developing or had developed a set of criteria l
t l
l l
l
99 1 fcr walk-through operations. Where do you say, oh, that
\s, 2 anchorage is adequate, or no, I'm not sure; I must analyze it.
3 MR. CHANG: This is being looked at by the EPRI 4 program as well. The program was started about a year ago.
5 MR. EBERSOLE: Can you eyeball most stuff without 6 analyzing it?
7 MR. CHANG: The idea is to do an analysis 8 beforehand. They actually are looking at each type-by-type 9 equipment, one by one, to look at the configuration -- well, 10 this is from an envelope point of view, the center of gravity 11 and all of that. And then from that point of view and the 12 study of the capacity of anchorages such as expansion anchor n
13 bolts, they can establish how many anchor bolts of what size
)
14 should be needed. Then the end product of that kind of study 15 is a walk-through check 1.ist and sometimes, a screening table.
16 So those will be the end products of the walk-through to use.
17 So in other words, the detail work is done 18 beforehand. Before the walk-through they will have those-19 table guidelines at hand.
20 MR. KERR: The walk-through is just to see if the 21 equipment is in the plant; is that right?
22 MR. CHANG: The walk-through is to make sure that 23 the anchorage is adequate.
g 24 MR. KERR: But you've already done that by analysis.
25 MR. ANDERSON: Let me try that one. The purpose of
100 1 developing the anchorage review guidelines is to get some
/^
(j 2 pre-determined anchorage configurations and anchorage loads 3 for a typical cabinet, and typical anchorage layouts so you 4 can go through and look at each piece of equipment, inspect 5 its anchorage, and you can screen it. Because if you have 6 already determined that that type of anchorage with that ,
7 number of bolts and that configuration is enveloped by one you 8 have already analyzed, then you go to the next one.
9 MR. KERR: The analysis then is a generic analysis; 10 it's not an analysis for each piece of equipment separately.
11 MR. ANDERSON: That's correct, and we think that you 12 can screen out most of the anchorage systems on this basis.
[ 13 The ones that won't pass your screen you have to do a specific 14 analysis on.
15 MR. KERR: You kick it to see if it vibrates or 16 something.
17 MR. ANDERSON: I don't know whether that's going to 18 be one of the guidelines or not.
19 MR. CHANG: For each type of equipment and different 20 types of analysis, you do an analysis first for the EPRI 21 program to come up with what they think is adequate for the 22 anchorage. For each type of equipment and the anchorage they 23 would have that summarized in the form of a checklist and fg 24 screening tables for the walkthrough to use.
- \
25 So, you do the work actually beforehand and you use
i 101 1 the end product to check the adequacy.
2 MR. EBERSOLE: Does this include such things as 3 verifying whether the doors come open on the panel boards and 4 so forth under shock effects?
5 I've seen some switchboards where you could open one e
6 door at a time, but if you open two adjacent ones at the same 7 time you create short circuits.
8 MR. CHANG: I don't think that is part of the 9 anchorage considerations.
10 That sort of thing should be looked at. But, that 11 is not part of the anchorage consideration.
12 MR. EBERSOLE: It might be worth your while to open 13 doors and see -
(
14 MR. CHANG: They are required to open doors to look 15 at the inside devices to make sure they are anchored.
16 MR. EBERSOLE: ,Open two at a time in an opposite 17 direction. Sometimes you see some interesting effects when 18 they bump into each other.
19 MR. ANDERSON: The doors and fas.teners are also part 20 of the experience. They look specifically for these in the 21 data base.
22 MR. CHANG: They look at these to make sure they are 23 not going to swing open.
24 Anchorages are one thing they have to address. The 25 other thing they have to address is to verify the
102 j
1 functionality of equipment. Actually, we are mainly concerned
) 2 about relays here.
3 I did not define the word " deficiencies" here.
4 Deficiencies are obvious inadequacies from the walkthrough.
5 By using A-46 criteria they are obviously inadequate.
i 6 This is different from the outliers. Outliers are 7 things that are not included in the data base. However, it 8 may be okay upon further evaluation. That is why on outliers 4
9 we are relying on additional studies such as testing.
10 MR. KERR: How many outliers is a licensee supposed 11 to find?
12 MR. CHANG: How many?
2
(~'N 13 MR. KERR: Yes.
14 MR. CHANG: We don't expect too many outliers, 15 really. But there will always be some oddball 16 configurations.
17 This outlier is going to catch those outliers.
18 MR. KERR: If you get a report from a licensee that 19 says, "We didn't find any outliers," do you tell him to go 20 back and look again?
21 MR. CHANG: We will talk about how this will be 22 implemented. There are two ways to implement our 23 requirement. One is the so-called generic approach that the 24 utility can treat this as a group.
25 MR. KERR: I was just trying -- see, you are not
103 1 going to be the person that makes a judgment on a report from 2 the licensee. It will be somebody else. I can foresee --
i 3 MR. CHANG: Actually, if they are using the generic 4 approach, the SQUG group will actually sort of certify that 5 good work is done by them.
6 MR. KERR: I can see a letter going out by an 7 applicant -- judging by the letters that I see -- saying, 8 here is some additional questions. One of the additional 9 questions can be, you didn't find enough outliers, go back and 10 look again.
11 MR. ANDERSON: We haven't established any criteria, 12 of course, on how many outliers.
13 MR. KERR: I know you haven't and that is the reason
[~'\
14 I wonder if the reviewer will have to establish criteria.
15 MR. ANDERSON: With the type of review that we are 16 proposing, I think the generic group has taken the 17 responsibility for overseeing this review.
18 What we mean by outliers is outside the data base.
l 19 We don't have sufficient experience data to verify seismic i
l 20 adequacy.
l t
21 MR. CHANG: For implementation, this will be 22 implemented by the issuance of a generic letter. This will be I
i 23 done after the CROR package has been commented upon by the l
, 24 general public and been through CROR again.
I '
25 (Slide) i l
l i
_ . . _ . . _ . . . - . _ . _ _ . . . _ _ _ . . . _ , _ _ _ _ _ _ . _ _ _ _ _ . . . . _ . _ _ _ . . _ . , , , . . _ . . _ _ . . . _ . , . _ . ~ . . _ _ _ _ , , - , _ _ _ , _ _ _ _
104 i Now, about a scope of this A-46 study. One thing 2 that is not mentioned here is the level of earthquake loading 3 that we are considering.
4 For A-46, just like in the new licensing criteria, 5 we are looking at a level only up to one SSE. Not beyond one 6 SSE. This is consistent with the current licensing criteria.
7 Also, this is limited to equipment only, mechanical
< 8 and electrical equipment. Things like piping, ductwork, cable 9 trays, they are not included in the A-46. Again, this is 10 consistent with the current licensing criteria.
11 There are several assumptions. The first assumption 12 is that SSE does not cause LOCA. We think we have a lot of 13 justification on that assumption.
14 From all the earthquake experiences we have seen, 15 and also the latest one from Chile, yes, you do see some 16 piping damages. However, those damages can be summarized into 17 two categories. One is the piping is corroded. The second i
18 one is, if a piece of pipe is connected to the building, let's 19 say, the other one goes through the ground, there is not 20 enough flexibility between the two ends and it may break 21 because of differential anchor movement. We have never seen 22 any pipe break because of inertial load only.
23 Now the reason we do not consider LOCA is more than i
24 that. Because, back in 1979, there were IE bulletins issued, 25 79-02, 07, and 79-14. The utilities are required to go back
105 1 and check the pipe stress analysis and also check the as-built (O) 2 configuration to the design configuration.
3 So we believe that problem is actually being looked 4 at again by this 79 IE bulletin effort. So, we believe there 5 is really no need to consider LOCA in A-46.
6 MR. EBERSOLE: LOCA is a unique phenomenon. It 7 spans from the large LOCA to the small LOCA, and it could 8 include the opening of PORVs on the hypothesis that valves 9 don't close when they open.
10 MR. CHANG: That's why PORV is equipment that has to 11 be looked at for seismic quality.
12 MR. EBERSOLE: Let's talk about the secondary side 13 of PWRs.
("'N 14 Is LOCA in this context also taken to be your LOCA 15 so to speak in the secondary system's feedwater and main 16 steam?
17 MR. CHANG: No. When we talk about LOCA we only 18 talk about the coolant.
19 MR. EBERSOLE: So you will assume then an SSE may 20 bring about a failure in the secondary side?
21 MR. CHANG: No, we don't think so.
22 MR. EBERSOLE: You don't say that.
23 MR. CHANG: We really don't think the piping will 24 break, regardless of what kind of piping you are talking O 25 about, because of the fact that it has been reevaluated again
i l
106 i
\
$ 1 furing the 1979 IE bulletin review.
. s 2 MR. EBERSOLE: Secondary and lower grades of piping 3 are not so intensively analysed as a primary loop.
, 4 MR. CHANG: Up to the main steam feedwater, those 4
5 things were looked at during the 1979 review.
6 MR. ANDERSON: Our assumption with regard to pipe 1
7 break also applies to the secondary and feedwater systems.
8 MR. EBERSOLE: That's what I am getting at. The
\
9 first statement is foreshortened, isn't it?
a 10 MR. ANDERSON: That's true. But it applies to the 11 secondary and feedwater systems. Also the components, the i
12 ADVs, the valving and the other components in the secondary 13 system are part of the scope of the A-46.
14 MR. EBERSOLE: I wish you would expand that first 15 line about a paragraph, about what you consider does not occur l 16 in a LOCA. ,
i 17 MR. CHANG: We just want to point out as far as LOCA l 18 is concerned, we don't postulate that.
19 MR. EBERSOLE: But you also don't --
20 MR. CHANG: Actually, there is much more than --
21 MR. EBERSOLE: Much more than that.
22 MR. CHANG: Yes, i 23 Also, LOCA does not occur simultaneously with or 24 during SSE. We just believe the possibility is too low.
t O 25 We also have conservative assumptions. That is, we
- - ,. . . _ _ .,_ . _ _ , _ _ _ _ _ _ . _ . _ _ _ , _ _ ~ . _ _ _ _ . _ _ _ _ _ _ -
107 1 assume offsite power will be lost during or following 2 earthquake.
3 When we talk about shutdown, we are actually talking 4 about hot shutdown. This is what we are proposing in the 5 package. And we propose that a minimum of 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> --
6 actually the other days I think the subcommittee questioned 7 whether we really need 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> or not. Probably this is on 8 the conservative side. But, we believe 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> might be 9 adequate to do the following things:
i 10 One is, after earthquake you can go down and check 11 the equipment to make sure everything is okay. And, in case 12 you. find some minor damages, then you will have time enough to 13 take care of them.
14 MR. EBERSOLE: On the bottom line, should that not 15 also be expanded to say "and normal heat sinks and i 16 condensers." Offsite power will be lost r. n d normal heat 17 sinks will be lost.
18 MR. ANDERSON: To be rigorous about it, tha t- would 19 be true.
20 MR. CHANG: Equipment scope. As we have heard 21 several times earlier, this is restricted to active electrical 22 and mechanical components, including instrumentation and 23 controls needed to achieve and maintain hot shutdown.
24 As far as tanks and heat exchangers, if they are 25 required to achieve and maintain hot shutdown, then we will
108 1 only look at the anchorages. The tanks and the heat
,,/ 2 exchangers themselves are not included in the scope of A-46.
3 At the beginning, in scoping A-46, we did look at 4 whether we should consider the masonry wall reviews, some aux 5 feed systems, RCS piping, and seismic interaction items or 6 not. Later, upon reevaluation, we found that those concerns 7 are either addressed by IE bulletin or generic letters. So, 8 they have really been taken care of earlier. There is no 9 point to addressing them again here.
10 MR. WYLIE: Are you speaking of the functional 11 review of those systems, or are you talking about the seismic 12 review.
13 MR. CHANG: Masonry walls. Okay, b}
[
14 MR. WYLIE: I understand that one.
15 MR. CHANG: We make sure --
16 MR. WYLIE: What about aux feed system?
17 MR. CHANG: Aux feed again, from a seismic point of
! 18 view, they have been studied by a requirement -- I think it is 19 a generic letter of certain types of aux feed. The seismic 20 adequacy part has been looked at for a number of PWR plants, 21 MR. WYLIE: For anchorages and pumps?
22 MR. CHANG: Yes.
23 Plants affected. This is restricted to operating 24 plants not reviewed to current criteria as documented by SERs.
25 As I said earlier, the current criteria is described l
109 1 in Standard Review Plan 3.10 and IEEE 344 1975 and Reg Guide
- 2 1.100. There are about 49 sites and 72 units in the scope of 3 A-46 review.
4 For SEP plants, only the functional credibility part 5 will be reviewed. The structural integrity part has already 6 been addressed by the SEP program.
7 (Slide) 8 Implementation requirements. Those are steps that 9 we think -- the procedure steps that utilities should follow.
10 First of all, they have to develop the equipment list required 11 for hot shutdown only, of course.
12 And then they have to compare the plant-specific 13 site spectra with the bounding spectra. It has to be 3
14 enveloped by the bounding spectra in order to utilise the 15 seismic experience data base.
16 Thirdly, they have to do a walkthrough inspection to 17 address the concerns of anchorage and deficiency and 18 outliers. And also, they have to address the functional 19 capability question on r e l a y s .,
20 For equipment unique to nuclear power plants, those 21 equipment we believe cannot be addressed by using the seismic 22 experience data base because the data base is from non-nuolear 23 plants. So, for those equipment unique to nuclear plants, we 24 believe you have to check against the test data.
25 And, for replacement parts, if there are equipment
I 110 1 after the review that turns out to be inadequate, or the I i
-w
( ) 2 functional capability part cannot be answered by the review, 3 then we require them to be several ways. Either they have to 4 be tested, or they may be replaced by qualified equipment.
5 By qualified, here, we mean either by using the 6 current licensing criteria, or they can utilise A-46 criteria 7 to evaluate the replacement equipment.
8 MR. EBERSOLE: There is a scattering of walkthroughs 9 beginning at Diablo Canyon that it is required that you look 10 not at the equipment list, the qualified equipment list, but 11 look at the system interfaces to see whether non-required 12 equipment may fall down and ruin qualified equipment. You know, the system interface, seismic interface.
[(- 13 14 Do you intend to do that on your walkthrough? I 15 wouldn't think it is part of your scope looking at older 16 plants?
17 MR. CHANG: I believe that falls in the scope of USI 18 A-17, system interaction. However, we are going to look for 19 interferences during the walkthrough phase. But the 20 systematic review of system interaction will be left until USI 21 A-17.
22 MR. EBERSOLE: You will not do it even in the 23 seismic context? This is sort of a specialised system 24 interaction.
25 MR. CHANG: It is not even in the seismic context.
111 1 MR. WYLIE: For example, the A-17 that would address i
\
7 any electrical or fluid system interactions due to seismic 3 disturbances.
I 4 MR. ANDERSON: That's correct. The USI A-17 will 5 include at least the total scope of A-46. And at that time we 6 will take a comprehensive look if we get a requirement in that 7 area to look closely at seismic type interactions, the type of 8 concern that Mr. Ebersole was referring to.
9 MR. WYLIE: Thank you.
10 (Slide) 11 MR. CHANO: Just a few words on the relay review 12 guidelines.
13 As I said earlier, the actual procedure of the relay 14 review is being studied by the SQUOs group. Those are the 15 guidelines that the NRC Staff thinks should be followed.
16 First of all, t,o identify all relays associated with 17 equipment needed to bring the plant to hot shutdown.
18 Now, for those relays that must function during an 19 earthquake, they have to do one of three things: Either to 20 verify with test data, to replace with qualified relays, or 21 qualify by test.
22 MR. EBERSOLE: Isn't there a flaw with respect to --
23 there is another TAP or USI, whatever it is, that has to do I
24 with control equipment with interacting safety systems, and
(
25 you would not look presumably at the control systems which i
- - - _ -_r---,___...._. - _ _ _ . . , _ _ , . - _ _ . , _ _ _ . _ _________.__,v, , , , , _ . - . _ _ _ _ _ , , _ . . _ - . , _ _ _ _ _ . _ _ . _ _ _ . . _ _ _ _ , ._ , _ _ _ - . - _ . _ . ,
112 1 might interact under special conditions of earthquakes. Yet, lh 2 they might be the ones which would, in fact, work improperly.
(V) 3 MR. ANDERSON: USI A-47, Safety Implications and 4 Control Systems.
5 MR. EBERSOLE: Right.
6 MR. ANDERSON: That's correct. Control systems may 7 be in a little broader context than they are in A-46. As a 8 matter of fact, the seismio concerns on A-47 are pretty well 9 answered by the A-46 data base, because all of the equipment 10 and instruments concerned with an A-47 are represented at 11 least by type, by the same stuff we have in A-46.
12 MR. CHANG: For relays which are not required to
[ 13 function during earthquake. However, they may be called upon D} to function after earthquake. They can do one of the three 14 15 things also. But there is an alternative. The licensee can 16 also show chatter or change of status does not affect plant 17 shutdown. If he can do so, that is the end of it.
18 Relay verification can be deferred until the test 19 data base is complete because we believe the functionality of 20 the relay question can only be answered by a test, and they 21 are in the process of trying to complete the test data base, 22 [811de.3 23 The SQUO group is in the process of developing a
,s 24 review procedure. That includes the identification of relays
- 25 to be evaluated, definition of functionality requirements,
113 1 development of evaluation procedures and review, and those
(_) 2 procedures will be reviewed by NRC Staff and SSRAP when it is 3 completed.
4 MR. EBERSOLE: There are presumably a lot of plants 5 in the field that we don't know anything about yet, about 6 relay chatter, so it seems to me sort of a clear and rather 7 present hazard if we have even r.odest earthquakes. So wouldn't 8 it be prudent to say if we do have relay chatter that we need 9 some assurance -- and we invoke that we won't have a LOCA --
10 that we take advantage of the fact that we now have time, like 11 we do in a power outage, and evolutions that take place which 12 look up the equipment in wrong arrays of various sorts are 13 reversible by the operator.
14 MR. CHANO: During the relay review, they have to do 15 it in a systematic way from a systems point of view, look at 16 systems, at the line drawings, the relays in the line 17 drawings.
18 MR. EBERSOLE: I'm talking about analog overrides as 19 a case in point. If 1 get locked up in some configuration, 1 20 have got to get out of an overlock, overrides would become 21 very important by the operator.
22 MR. CHANG: If this can happen, they better do 23 something.
24 MR. ANDERSON: The evidence that we have seen so far
\
25 gives us a pretty warm feeling that we are not going to get
114 1 into that situation. It turns out the relays are really a lot 2 more rugged than we thought they were. In addition, what we j
3 are finding out is that chatter, even if it occurs, may not be 4 that big a problem because of circuit designs, delays and 5 inertia, that sort of thing.
6 MR. EBERSOLE; You can back out.
7 MR. ANDERSON: We feel comfortable with delaying 8 that verification for some time period until we could get the 9 test data complete.
10 MR. EBERSOLE: I am thinking about the black picture 11 where a bunch of valves open when they should shut, and a 12 bunch shuts when they should open. That can produce some 13 hazards.
V 14 MR. ANDERSON: Due to seismic motion, we think that 15 is kind of unlikely right now.
16 MR. CHANG: That point was evidenced by the Chilean 17 earthquake review. During the review, they were able to talk 18 to operators several days after the earthquake, so they still 19 have those things fresh in their mind. According to what the 20 operators said, the general purpose relays, they functioned 21 without problem. The ones in which they have seen some 22 problem are the protection relays.
23 MR. EBERSOLE: Big relays for power systems.
24 MR. CHANO: That's right, and Jim is going to D 25 elaborate on those further later on.
115 1 Talking about a scope of relay, we asked around and 2 we found out that typically for a relay required for hot 3 shutdown, you are talking about 800 to 1000 relays in a 4 plant. However, we believe the relays needed to function 5 during earthquake will be much, much less than that number, 6 probably 20 or 30. That is just a gut feeling.
7 There are about six to eight types of relays, 8 electromechanical relays, in a typical plant. I think that is 9 all I have for my presentation.
10 MR. MARK: You spoke of the small number that might 11 be required to function of those relays, but there may be a 12 number that could malfunction as well.
I \ 13 MR. CHANG: True; but based on the Chilean 14 earthquake, a 7.8 earthquake, only in the protected relays 15 have they seen some problems. All the general purpose relays, 16 they function well. So that gives us a warm feeling that the 17 scope of the problem is really not that big.
18 MR. MARK: I'm not trying to make a case that there 19 is a horrendous problem, but as well as the things you want to 20 function positively, there are some that if they function 21 negatively, you are in trouble for that reason.
22 MR. ANDERSON: Well, lumping that into one class, !
23 guess, we are concerned about the ones that function when you l
24 don't want them to that can preclude some later function or 25 put the plant in bad condition, so that is part of the set we
116 1 are talking about.
2 MR. MARK: Fine.
3 MR. WYLIE: Any other questions for Mr. Chang?
4 [No response.3 5 If not, I will call on Mr. Thomas to present the 6 SQUG report.
7 Incidentally, there is a handout of his over here 8 someplace.
9 Let's take a break at this time.
10 [ Recess.3 11 MR. WARD: We will come to order.
12 MR. WYLIE: Mr. Jim Thomas of Duke Power Company, 13 representing SQUG, will make the presentation.
14 Jim.
15 MR. THOMAS: Thank you.
16 What I would like to do is start off by really kind 17 of emphasizing what we have done in a little over the past 18 year, start out with maybe where we were in April of last year 19 when we made a presentation to the subcommittee, and kind of 20 bring you up to date from that point.
21 (Slide.)
22 At that particular time, the conotusions of the 23 Utility Group itself, based on the work that we had done with 24 the experience data plus just our own experience in the design 25 and operating of the power plants, is that the seismic
. _ ._. ~ _- _ . - - -.- . _ - - - - _ - - . _ . . - -. -
i 117 4
1 resistance of standard power plant equipment when properly l- 2 anchored was verified during our pilot program, and on that 4
3 basis, explicit seismic qualification of the equipment in the
) 4 operating plants to current standards was not justified.
I .
j 5 Also, based on the broad basis that we were looking i
6 at equipment and the depth of our survey, that the seismic i
7 qualification issue itself was not of a significant concern, 8 and at that time no further action should be required.
9 As you have heard in the discussions by the Staff, i
10 what we had done is we had started this program working with j
11 the Staff and had later decided that we would have a third 12 party to review our work, the Senior Seismic Review and j
13 Advisory Panel, so even though our group had made those 4
14 conclusions, there were still some outstanding concerns by the i 15 Staff and also the ACRS Subcommittee, i
16 CS11de.3
- 17 So based on the concerns that still existed, SQUG 18 continued activities to further our knowledge in the area, t
19 One of the things was to totally define the equipment that 1
20 would be required to perform in a seismic event to assure that 21 you didn't have any safety problems at the plant. So SQUG 22 began work on definition of generic equipment to envelope al!
\
23 plants.
24 We surveyed all the utilities. We had input both by 25 the utilities and the Staff, all utilities and SQUG, to come i
f
- - - - - . - - _ , . _ . , _ _ _ _ - . . _ _ _ . - _ _ _ _ _ . _ . _ . _ . _ . _ . _ - - _ - . - - .r.- . - - _ , .
118 1 up with a list of equipment that would be required. One of 2 their concerns was that we had pretty well fully addressed and 3 resolved most issues for the eight classes, but there was this 4 kind of concern: what about the equipment outside the eight 5 classes; how could you say that is truly represented?
6 So we initiated some work in regard to looking at 7 the data we had. Taking a close look at the surveys that had 8 really included all classes of equipment in that we were 9 trying to find failures, we would always ask the question 10 were there any failures and whether it was in the eight 11 classes or not. We would try to take a look at that.
12 The point was we were not finding failures in any b)
I 13 class of equipment.
Another one was the functionality issue. The 14 13 surveys that were being done were really to look for 16 structural-type damage or damage to the equipment in that some 17 of the earthquakes were a number of years previous but wo 18 haven't had much oppor'tunity and really had not applied the 19 appropriate personnel to address operability during strong 20 motion in regard to experience data.
21 On the second and third items, we kind of enlisted 22 some help from EPRI. The first was to gather some additional 23 test data that we could use to try to demonstrate the 24 equipment beyond the eight classes was very rugged and kind of 23 collate a lot of the testing that various utilities had done
119 1 over the years, various AEs, in regard to the equipment in 7%
%- 2 these plants; and also to take a close look at test data on 3 relays and try to determine really how rugged were the relays; 4 and also we started trying to take a look at exactly what 5 equipment would be required to function, and if it had a relay
- 6 in that system, would relay ohatter give us a problem.
7 So we started trying to exclude some of the relays 8 that may be in these systems by the fact that even if chatter 9 did exist, it would not give you any problems if you were to 10 have a seismic event. Systems that can handle it due to the 11 actual design of the systems.
12 The fourth item that we were working on was to try to develop a kind of generio implementation plan. I think I
( 13 14 will hold that until a little later and get back to it toward i
15 the conclusion of my discussions, but as ! said, we were 16 working in two areas, on items 2 and 3, one with EPRI 17 collecting a lot of data for us, and the other we were trying I 18 to make some systems-type studies.
19 It was in the middle of that work that the Chile r
I 20 earthquake occurred.
21 (811de.3 22 We knew very quickly that the earthquake was a very 23 1arge earthquake. As previously has been said, it was a 7.8 24 magnitude. That had a number of large aftershocks. It 25 affected a very large area. We knew that there were a lot of
120 t facilities in that affected area that could possibly add some
\ 2 information and some background to what we were trying to do 3 with our testing surveys, and we also saw that it was really 4 an opportunity to try to find if there were any holes in our 5 previous conclusions to see if there were failures that had 6 not been seen in other earthquakes that might affect some of 7 our original conclusions.
8 So we decided to immediately send a team as soon as 9 possible after the event, 10 CSlide.1 11 They actually arrived on March 7th to get a very 12 quick look immediately following the earthquake and try to get 13 the data as quick as we could, but we had a number of meetings 14 to really study what we might could learn if we made a real 15 in-depth survey of the earthquake, and what type of team would 16 we need to really contribute significantly beyond what we had 17 learned from previous earthquake surveys.
18 We came to the conclusion that one of the main areas 19 was the operability issue during strong motion, and on that 20 basis decided that we should send some people with 21 systems-type experience to interview operators and to 22 determine really what type of problems operability during 23 strong motion might present, or lack of operability, relay 24 chatter in particular, 25 So we sent a second team in May. We had the EQE to
121 1 essentially continue the work that had been done by the 2 previous team, which was primarily EQE. Also, Dennis Ostrom 3 from Southern California Edison and myself went along to try 4 to look at these things as to what is happening during strong 5 motion. I was looking in particular with regard to the 6 operabilty issue and what happens to relays.
7 (Slide.)
8 We wanted to make sure that we asked the right 9 questions, so we spent quite a bit of time actually 10 interviewing our own operators, various utilities, 11 interviewing design people, trying to determine what type 12 questions should you ask to really find out what kind of 13 problems the earthquake would give with regard to potential
\
14 system misoperation, malfunctions.
15 I won't go through these in detail, but we developed 16 quite a list of questions, to hopefully get us to weak links or 17 weak areas or any kind of contradictions to our earlier 18 conotusions.
19 CSlide.)
20 This is just a continuation of those type questions.
21 MR. MOELLER: Excuse me. Does Chile have any 22 nuclear power plants?
23 MR. THOMAS: They have some research plants. They
, 24 don't have any commerotal power plants.
25 MR. MOELLER: Was this earthquake felt in Argentina?
122 1 MR. THOMAS: Let me get to the next slide.
,/
2 MR. MOELLER: Okay.
3 CSlide.]
4 MR. THOMAS: The earthquake was felt over an area 5 200 miles by 100 miles, a very large area in regard to where 6 the strong motions were felt. I know it is difficult to see 7 here, but what you are seeing on this is actual records that 8 were taken in the horizontal directions. I know that is 9 difficult to read. But the facilities that we were looking at 10 ran from down here in the Rapel area up to north of Las 11 Ventanas and over Santiago, quite a large area, and strong 12 motions being felt. Here is about .64, .57, about .469. So 13 1arge motions over a very large area.
14 To maybe give you a better -- hopefully, you can see 15 this a little better. This is a summary of some of the 16 stronger peak accelerations. The average here of the two 17 horisontal motions. You can see this .58, .64, large 18 magnitude, long area, and also very long duration. The 19 durations were in the neighborhood of 50 seconds of strong 20 motion.
21 C811de.3 22 Just an example of this. You can see, starting from 23 sero, continuing, starting over after 50 seconds, we are 24 getting pretty significant motion. We have quite a bit of 25 data on the motion, the actual strip charts that we can make
123 1 available.
2 (Slide.3 3 This one is one of the largest. You can see that it 4 had a horizontal of .43 in one direction, .67 in another, and 5 a veztical of .86. So we are really looking at a very large 6 earthquake in regard to motion and duration.
7 (S11de.1 8 As I said, there were quite a number of facilities 9 in the area.
10 The second team and this first team collectively 11 visited six power plants that were hydro and fossil. We 12 attempted to get into the research nuclear reactor, but being 13 a military government, we were never able to get access while 14 we were there. The day after we left, we got a message, yes, 15 come on in. That's kind of the way it was.
16 We don't know of any major problems that they had in 17 any type of their control systems.
18 MR. MOELLER: What are the five commercial 19 facilities? What do you meant 20 MR, THOMAS: Well, for example, some of the large 21 apartment buildings may have had some type of equipment, that 22 type of thing.
23 MR. MOELLER: What specifically in the water
[~N 24 treatment plants?
25 MR. THOMAS: They have quite a few pumps, motors, a
124 1 Int of control systems. The other four facilities -- one I'll h
2 get some slides in a moment -- there was a hospital that had 3 an emergency generator. I'll show you what happened there.
4 But, anyway, these facilities contained quite a bit 5 of equipment, a lot of it supplementing the original eight 6 classes, and a lot of it that we took a close look at to try 7 to find problems of anything typical beyond the 8 classes, 8 anything that we would think might have a bearing on our 9 general conclusions.
10 We were searching for problems'.
l l 11 MR. WYLIE: Jim, wasn't one of those refineries a 1
12 copper refinery?
13 MR. THOMAS: Yes, one is a copper refinery. I've
(
14 got some 35 millimeter slides. I'll show you a little of 15 that. One of them was a very large oil refinery, a tremendous 16 number of tanks. It's ki,nd of outside the scope of what we 17 are talking about. But a lot of piping, tremendous amount of 18 tanks.
19 One of the problems we did find was very large tanks 20 did have some problems, around a quarter of a million gallons
.21 type tanks. They did have what we would call design problems 22 for earthquakes.
23 MR. MOELLER: In the water system, do they have 24 elevated tanks on stilts and so forth?
(
25 MR. THOMAS: I don't recall seeing that at either of
125 1 the facilities we visited. I was kind of surprised. It's 2 totally different from what we normally do. I think they 3 treat it and pump it to another area, and in the other area, 4 then it's pumped up to a storage tank.
5 MR. SIESS: They weren't tanks on stilts? The 6 spheres weren't?
i 7 MR. THOMAS: Yes, at the refinery, that is correct.
8 MR. SIESS: But they did no damage to them?
9 MR. EBERSOLE: Did any of the big tanks that sit on 10 the ground burst?
11 MR. THOMAS: Yes.
12 MR. EBERSOLE: They did?
13 MR. THOMAS: Yes. The large tanks --
(
14 MR. EBERSOLE: Like condensate storage tanks?
15 MR. THOMAS: Those types, no, we didn't see any 16 failures of those type tanks.
17 MR. EBERSOLE: Oil storage tanks did, though?
18 MR. THOMAS: Yes, the very large quarter of a 19 million gallon type, the ones that are --
20 MR. EBERSOLE: They're thin as paper, aren't they?
21 MR. THOMAS: They are vertical, they are 60, 70 feet 22 tall, about at least that in diameter.
23 One of the failure modes is they develop what is 24 called an elephant'L foot. The tank at its base just kind of
(~s 25 moves over and poaches out at the base, and you get some
126 1 separation from the wall and the bottom plate, and you lose 2 the contents very quickly. In a number of cases it caused 3 implosion of the top of the tank.
4 MR. EBERSOLE: Were those all contained with dikes 5 or did any of them get away?
6 MR. THOMAS: None of the ones we saw got away. They 7 were contained.
8 MR. EBERSOLE: Does American Petroleum Institute 9 permit that? You know, like around Los Angeles where they 10 have hundreds of them?
11 MR. THOMAS: From what I know -- I'm not an expert 12 in the area -- from what I know, from what we discussed, the 13 way we do things is better than what we saw in Chile.
14 MR. SIESS: I recall Peter saying 10 percent of the 15 tanks failed, ruptured, lost their contents, but that a much 16 larger proportion had distress buckling; is that right?
17 MR. THOMAS: That is correct. They maintained their 18 integrity as far as containing liquid, but they did show some 19 stress and some buckling.
20 Maybe the best thing to do is to show a few of these 21 slides.
22 [ Slide.]
23 The first few I will go through very quickly.
s 24 I intend just to show you that we did see a good bit 25 of damage. We seem to have lost the integrity of the focus.
127 1 We did see a good deal of damage.
2 MR. MOELLER: Is that taken from a t a l'1 building 3 nearby?
4 MR. THOMAS: It is taken from a helicopter. This is 5 the initial team that hired a helicopter to take some aerial 6 shots. This is in San Antonio. You can see some of the 7 collapsed walls. This is in the Port of San Antonio.
8 [ Slide.]
9 You can see the frames that have fallen over. This 10 is not due to inertial forces, but due to the problems with 11 the ground just essentially falling into the water. This was 12 a lot of area that was not really what you would call a good 13 foundation. It was filled earth in many areas.
14 Anyway, you can see the type damage they saw.
15 [ Slide.]
16 One bridge you saw collapse is going into the 17 water. They say a lot of the reason that this probably 18 collapsed was a lot due to some erosion they had seen during 19 the flood. The earthquake kind of finished it off.
20 This one was one that was an older bridge, not in 21 use. It was actually being used for target practice by the 22 military.
23 [ Laughter.3
-w 24 MR. MOELLER: What is the bridge in between? Is
)
25 that a walking bridge?
l
l 128 i MR. THOMAS: That is something that they -- it's a
( 2 quick fix to take peop1<a across the river. This was about 3 three days after the earthquake. This is a temporary 4 structure to provide access for walking.
5 [ Slide.]
6 Some of the taller buildings you can see quite a bit 7 of damage. You can actually see where they had bending in the 8 rebar.
9 [ Slide.]
10 This one is a kind of interesting leaning tower.
11 MR. EBERSOLE: That could be a tourist attraction if 12 they propped it up so it stays that way.
13 MR. THOMAS: It had already been demolished by the 14 time the second team arrived. It was a bit of a danger.
15 MR. EBERSOLE: It could compete with Piza.
16 [ Slide.]
17 MR. THOMAS: This one is one I'd like to talk about 18 a little bit, particularly. This was a hospital in San i
19 Antonio. It's really on two different foundations, this area 20 and this area over to the right. It is a relatively modern 21 structure and --
22 [ Slide.]
23 -- it saw quite a bit of internal damage, as you can 24 see. The collapsed concrete around a lot of those structures 25 really was seeing quite a bit of motion. This is a little bit l
L
129 1 misleading in that some of this is stucco, but you can see s
-/ 2 also concrete actually did suffer some problems.
3 The interesting part was that this hospital had 4 recently been equipped with relatively modern diesel emergency 5 power system, and had a very similar, although smaller, 6 control system as the type of control we use in nuclear power 7 plants.
8 [ Slide.]
9 That is a sideways picture of that.
10 MR. WARD: It wasn't anchored?
11 MR. THOMAS: That was not due to seismic, that was 12 due to -- I'll either get it upside down or -- that's right There is kind of an interesting story about this diesel
( 13 now.
14 control system. The technician at the hospital was home 15 during the time of the earthquake, and he became very 16 concerned that his power. system may have failed during the 17 strong motion. They did lose power almost immediately. He 18 was very concerned about the welfare at the hospital. So he 19 immediately ran from his home, which was, I understand, on the 20 order of a quarter mile or half mile away or so, and when he 1
21 arrived, he found that the diesel had started and was 22 running. And not only that, they continued to run it for a 23 period of weeks to provide not only power to the remaining N 24 portions of the hospital that they could use, but to some I
25 emergency shelters that they set up at a church and a school.
t
n 1 130 1 I would like to point out that this little control l 2 system does have various types of general purpose relays that 3 we have been talking about that needed to be looked at.
4 It went a long way to give us a good feeling that 5 the general purpose type relays are pretty rugged when it 6 comes to seismic, because we know for a fact that this had to 7 be seeing tremendous amounts of forces due to the effects on 8 the hospital.
9 This was in one of the rooms, the back part of the 10 hospital.
11 [ Slide.]
12 M2. EBERSOLE: That was a fast start diesel. I'm
[ 13 wondering whether or not there was an elapsed time of zero b
14 energy where you could have the relay chatter all over the 15 place and it wouldn't matter, anyway.
16 Was there a blackout interval that would have 17 protected you from relay chatter?
18 MR. THOMAS: I'm not totally familiar with the 19 actual circuitry of the diesel control.
l 20 MR. EBERSOLE: If you are in a total blackout, who i
l It's when you're not, that it 21 cares about relay chatter?
22 hurts.
23 MR. THOMAS: We are talking about something with 24 chatter that would get something out of sync and the automatic 25 start signal never properly being accomplished.
131 1 MR. EBERSOLE: If you could have a clean break so 2 there's no power, then relay chatter doesn't really matter; 3 right? If it's all over before you start going again.
4 MR. THOMAS: The relays themselves have power 5 because you've got battery back-up.
6 MR. EBERSOLE: So the batteries persisted?
7 MR. THOMAS: The batteries had to persist for the 8 automatic control system to function, they could help the 9 diesel start. So the relays did see power and they performed 10 directly on the automatic start signals.
11 MR. EBERSOLE: It was a fast start, I guess?
12 MR. THOMAS: I guess you could call it a fast start.
13 MR. EBERSOLE: Okay.
14 MR. KERR: It wasn't tested every two weeks, 15 probably. Or was it?
16 MR. THOMAS: It was tested, I think, about every six 17 weeks. I can get that data. We have it in our data sheets.
18 This was the copper re' finery. I think one -- the copper 19 refineries suffered a lot of damage to its furnaces. The 20 brick fell in on molten copper.
21 [ Slide.1 22 The main thing it added to it was it increased our 23 original data base, particularly in relation to motor-operated 24 valves at high levels.
25 The only problems the plant had, other than the
132 1 collapse of the brick and the furnaces, was a few cases where
) something fell on a mercury monometer and caused sloshing of 2
3 mercury. There were some totally unprotected thermocouple 4 wires, not cable, just pretty small wire that was strung a lot l 5 like a handrail between two different structures, and the 6 relative motion broke the wire.
7 But as far as functionality of the controls and 8 equipment, they had no problems.
9 MR. MOELLER: Excuse me. In the previous slide, 10 that stack stayed up, this is after the event?
11 [ Slide.3 12 MR. THOMAS: Yes.
Was it seismically designed?
( 13 MR. MOELLER:
14 MR. THOMAS: Most of the structures, particularly 15 that type, are designed to a seismic, I think of something 16 like .2 g static design basis.
17 MR. WARD: What was the ground acceleration here, 18 Jim, at the smelter, for example?
19 MR. THOMAS: I think it was in the neighborhood of 20 .35.
I 21 MR. EBERSOLE: Jim, were there any -- in a general 22 context, were there many fires developed as a result of the 23 earthquake?
24 MR. THOMAS: No. That's one of our questions of j
25 anything like that, such as flooding or fires. We saw no l
t
133 4
1 evidence of that in industrial facilities.
l \ss 2 Now in some of the homes and things, they might have 3 had some local fires caused by appliances or something like 4 that 5 [ Slide.)
6 The control facility at the refineries was large 7 enough that they had their own power plant. The interest 8 thing was it was mostly German equipment. So we saw quite a 9 variety of what we had been seeing in our previous surveys of 10 types of equipment, but again they saw no types of 11 instrumentation problems, no falso alarms, no malfunctions, no 12 misoperations.
) 13 [ Slide.]
w.)
14 The Las Ventanas power plant was a very interesting 15 one. I have taken out a lot of the slides that had to do with 16 some piping and some snubbers. I did that in the interest of 17 time, but we did learn quite a bit about what piping can de 18 and what supports are.
19 But again, as far as what we were trying to learn 20 for A-46, they had quite a bit of equipment throughout the 21 plants that added to our data base.
22 [ Slide.]
23 [ Slide.]
24 For example, something I discovered, cable trays and 25 the type of supports they had saw no problems at all.
134 1 [ Slide.]
i
(_s/ 2 The control room itself. Very typical The type 3 switches, the displays, the indicating lights, the alarm 4 panels, all CRT screens, indicators; very similar to the types 5 that you see in the older nuclear operating plants. The 6 structure of the board itself, the way it's designed.
7 MR. WYLIE: Jim, I don't believe you mentioned this 8 is an American-designed plant.
9 MR. THOMAS: Yes. One unit was Charles P. Maine, 10 and the other one was EBASCO, I believe, and on that basis had 11 practically all American equipment.
12 The motor control centers, the switchgear.
13 One of the things that was not in the original 8 14 classes that we were trying to look and see if we had 15 problems.
16 [ Slide.] ,
17 This converter system, very similar, identical, in 18 fact, to some converters that are used in some of the older 19 plants.
20 [ Slide.)
H21 This one is very interesting. It's a burner control 22 system, manufactured by Forney Engineering in Dallas. It 23 contains literally hundreds of relays. Each bay on each site 24 has about 200 relays. There are various types. There are 25 general purpose type relays that we use in various systems.
I 135 1 So we really felt a lot better about the performance of
\ ,) 2 general purpose relays. )
3 MR. REMICK: Question: Did the generating plants 4 ride out the quake? Did they stay operating, or did they go 5 down?
6 MR. THOMAS: A little bit of all cases. Some of 7 them were not operating during the quake, but had some 8 auxiliaries running, that continued to run.
9 Some of them were operating and tripped due to 10 problems on the system. They do see some problems in 11 switchyards. They have some ceramic porcelain type failures 12 that caused failures. One plant tripped. It was kind of 13 interesting.
14 One of the things about Chile is they see quite a 15 few earthquakes. They have seen three major ones, I guess, in i
16 the past 20 years or so, very similar, of this magnitude.
17 So when we talked to operators, we were not getting 18 feedback from just this one instance, but from what they had 19 learned -- one in 1961 and 1970 and one in the '80s.
20 One of the units here, he said, well, it always
, 21 trips off because of vibration. One of the turbines is too 22 damned sensitive.
23 [ Laughter.)
24 So they trip off for various reasons.
25 Most of the cases, when they do trip, it's due to
136 1
1 protecting themselves from what's going on outside. Some of r
2 them did ride through.
3 I think one of the slides that we used in a 4 different presentation is that from a lot of the California 5 experience was if it is below about .3, usually they stay on 6 line. Most of them stayed on line below a .3 ground motion.
7 [ Slide.]
8 Quite a bit of pneumatics, so we got another good, 9 real good feeling about the capability of pneumatics.
10 CSlide.]
11 The hydro plant at Rapel did see quite a bit of 12 motion in regard to the dam itself and the structures. They
() 13 did see millions in damage that they're going to have repair.
14 They didn't cause any leaks or anything, just s o m e '. h i n g 15 they're going to have to go back and fix.
16 The plant is still in operation. It was not in 17 operation at the time. The used it for peaking purposes, but 18 they were able to make' a quick survey and determine that they 19 could continue to use it. But again, back to the electrical 20 equipment.
21 [ Slide.]
22 The mechanical equipment that is required to operate 23 the plant, the same story again. They didn't have any kind of 24 problems with the equipment. They had no misoperations, no s
25 systems starting when they shouldn't or stopped when they
i 137 )
i I
1 should not have, no relay chatter.
I 2 Again, the thing that is quite impressive, with all i
3 of the relays that are used in alarm circuits and the types of 4 relays, we still could not find any evidence of any false 5 alarms.
6 MR. WYLIE: The ceiling didn't fall down.
7 MR. THOMAS: The ceiling didn't fall down.
8 This particular site, though --
I 9 CSlide.]
10 -- back in this direction, there is a hill, and up 11 on top of the hill, overlooking the site of the dam itself, 12 they have a pretty sophisticated switchyard and substation, 13 and inside that structure, which is very strong seismically, i
14 hd a lot of shear walls that ended up showing cracking, very 15 strong shear walls. So we know we've got -- I guess some of 16 the seismic people thought they saw a tremendous amount of 17 amplification due to apology, and that structure is the only 18 place where we found any equipment damage.
19 [ Slide.]
20 So going from the hydro, this is the control room 21 itself from the plant that is down. ,
22 CSlide.]
23 In the turbine building area here, --
i i 24 [ Slide.]
, 25 -- we leave the plant and go up to the substation.
138 1 I believe these batteries are still in the plant also.
y,) 2 [ Slide.]
3 This is up at the substation. They did see some 4 bending of the straps. These are very lightweight straps. No 5 functional problems with the batteries, no failures. But they 6 did have some relay mechanical problems with the internals of 7 five relays. The estimates of the motion are the forces that 8 they say there. We don't have a record there, but just due to 9 the effect of some of the bending of some of the anchorages, 10 somewhere in the neighborhood of .6, .7, .8 caused some 11 internals of some protective relays to either jam or some very 12 small springs, which were soldered at the end of them, to m
13 break lose.
14 It's my personal opinion, I don't really think it 15 was a design problem, but more or less, since it was only 16 three out of something like a hundred relays that broke, I 17 think it might have been more of a manufacturing problem, 18 MR. WYLIE: These are basically line protective 19 relays for transmission lines, and really this is sort of an 20 antique type relay.
21 MR. THOMAS: That's true, but it was the only 22 evidence of equipment failure that was due to or appears to be 23 due to inertial forces, rather than interaction. We've given rg 24 you some evidence previously in previous reports, and we had 25 one other case here where some interaction, like on a motor
139 1 operator interacting by the pipe that it's mounted on swinging 2 quite a bit and crashing the valve into a structure had some 3 breakage. That's not really the type failures that we were 4 concerned about, and even those still remain very rare, 5 [ Slide.3 6 MR. EBERSOLE: This is the only kind of relay 7 failure you saw, right?
8 MR. THOMAS: At the refinery, one of the protective 9 relays caused the emergency power, a generator, the start 10 cycle to trip during the start. The operator immediately 11 reset that particular relay. Then the turbine generator did 12 come up, and no problems after that. That was a protective 13 relay.
14 These relays also were protective relays. We saw no 15 cases of any other type of relay either having any damage or 16 malfunctioning. .
17 MR. WYLIE: These are basically induction disks, 18 protective --
19 MR. THOMAS: Correct. Again, they're older type 20 relays. They're not really modern solid-state.
21 One case of interest, even though I think we did l 22 learn that it was something that probably bears looking at, if 23 you have such an application, is that large oil-filled 24 transformers have in many cases fault pressure relays applied, 25 so that you can get -- if you get any kind of shorts inside,
140 1 you see an increase in pressure due to the oil reacting with O)
(m, 2 that, and you can trip off of that pressure. A seismic event 3 will slosh the oil around, and the relays correctly sense a 4 pressure and trip.
5 That's kind of a systems type thing. I don't expect 6 to see any of that or very limited amounts of that, if any at 7 all, in the type system we're seeing, that needed to function, 8 because as we said earlier, these protective relays are 9 applied like in switchyards and very large transformers and 10 electric lines.
11 This is an example of some of the internals that had 12 a problem. I can read here that this is an MHO measuring 13 relay for phase fault. It actually had a little coil spring
(
14 that came loose, as I was talking about earlier. The end of 15 the spring was soldered, so they had to replace the relay. It 16 didn't cause a problem during the event, but when they were 17 trying to investigate what had happened, they did find a 18 problem with these relays.
19 I don't have .a slide of it, but there was one other 20 equipment problem at that particular substation, and that was 21 some switchgear that was manufactured in Europe. I believe it 22 was French. That is kind of one of the things that we would 23 call an outlier. It was very short. It didn't look anything 24 at all like a switchgear that GE makes. It essentially didn't
\
25 have any structural support, either in the front or the back.
141 1 The only support it had was a very small, little diagonal fs 2 strap that was about three feet long. ,
3 There was a lineup of these switchgears, probably 4 about eight days. All of them bent over in one direction and 5 caused some tearing of the cabinet wall in the back. They 6 crashed into and probably were stopped by some of the GE 7 switchgear that we typically use.
8 Again, they had no functional problems. It worked 9 properly. And after the earthquake, they essentially 10 straightened it up and give it a little bit of a fix in the 11 back where it tore, but they didn't do anything other than 12 straighten it up.
13 So the switchgear was kind of an outlier, as wo 14 called it, and the five relays were really the only equipment 15 damage we saw.
16 That kind of just gives you a taste of the 5000 17 slides that we collected and the type of things that we're 18 talking about.
19 Kind of to summarize where we are at and what we 20 have done in the past year and a half, --
21 CS11de.]
22 -- we have con +1nued to look at earthquakes as they 23 occur, even though we thought the database we had at the time 24 was totally sufficient. We still cannot find any problem. We 25 cannot find any generio area that we think shows any type of
142 1 concern that all of the mechanical and electrical equipment of T
l 2 the type we used in our power plants are not inherently 3 seismically rugged.
4 This is an example of three of the investigations 5 we've made since 1983. We've got two of them fairly large, 6 the Morgan Hill, of course the Chile, and Coalinga in '83.
7 Where we are now, I guess, is probably the best way 8 to summarise.
9 [ Slide.]
to As was stated in the Staff's presentation, we had an 11 effort to identify what types of equipment are needed to 12 assure that you can get to a hot shutdown / safe shutdown during 13 a seismic event. So we did a lot of work to generate such a 14 list, with interaction from the Staff and comments. And while 15 we were doing that, we were saying, "What type of data do we 16 have to support the ruggedness," other than what we had 17 presented in the pilot program, the original eight classes.
18 It was in this area that you can talk about the- EPRI 19 programs that have been initiated to support our work. Also, 20 that involved collecting the test data for equipments that l
21 were not specifically in the eight classes.
22 Also, the Chile earthquake gave us quite a bit here 23 to add in that. So we were trying to develop this list and 24 determine what could we say, based on all of our experience, 25 all of our testing, and all of our engineering judgment. What
143 1 can we put together in the form of documentation of just how N! 2 rugged it is, and then how could you use that documentation if 3 you were to survey a plant to verify that you didn't have 4 anything that was not consistent with our generic conclusions?
5 You've got some outliers or something is a little 6 different. I think the example of the European switchgear, 7 with no support in the front and back, is a good example of 8 things that we would want to look for and verify that we don't 9 have. Most of us feel, just on experience, we're not going to 10 see that anyway.
11 Anyway, we're going to develop the list of 12 equipment, very preliminary. We think the total classes of b)
( 13 generic equipment required is limited to about 21 classes.
14 I'll give you an example. You've seen what the 15 eight classes are. One additional, we kind of lumped 72attery 16 chargers and inverters together, because if you look at them, 17 the cabinets are structurally the same. Essentially, the l
18 internal components are the same. One additional class might 19 be composed of battery chargers and inverters, i
20 We are continuing to work with EPRI. They are also 21 helping us look at the anchorage issue, trying to develop some
! 22 simplistic methods that are technically sound, analytically 23 sound, but would facilitate a walkdown to verify you had
% 24 adequate anchorage.
25 They are also continuing to put together further
_-. . _ -..- . _ - . _ . - . - - - - . . - _ _ _ _ _ _ _ ~ _ - _ _ . .
l 144 i
t 1 test data. This test data, by the way, has, if anything, just 1
% 2 further justified and supported all of the original 3 conclusions that we made based on experience data.
4 [ Slide.]
5 MR. EBERSOLE: May I ask a question? If I'm looking i 6 at the station type batteries, not the 1-E batteries, the ones
- 7 that do the switchyard functions, and I was looking for the 4
1 8 very worst, most fragile installation I could find across all 9 of our plants, what would that be? Are there controls on i
10 that?
i 11 You know, these are inert, stationary, static f
i 12 devices. They sit there like pots. And I could imagine, 13 really, a sloppy designer could just stick it up on wooden i
14 shelves, and say, "I'll I've got to do is hold it up."
15 Yet if I fail that battery early on as a preferred 16 failure, it has drastic consequences in reducing the station's 17 switchgear to total inoperability.
18 MR. THOMAS: I think what you're saying is, correct 19 if such a thing happened. What we are see.ing in totally I
20 non-nuclear applications is any type of application, the
.21 commercially available in general practice are installing I
22 those types of batteries --
i 23 MR. EBERSOLE: Somebody put them on good mounts.
24 MR. THOMAS: Yes. We have not found any battery 4
25 failures.
145 1 MR EBERSOLE: Do you find that to be judgmental 2 requirements or common sense? Just put them under the proper 3 -- in the proper supports. I don't know of any IEEE 4 standards, do you, on mounting the station batteries?
5 Chuck, do you?
6 MR. WYLIE: No. I don't think there are any 7 standards on mounting the batteries.
8 MR. EBERSOLE: So you could just put them on 9 bookshelves.
10 MR. THOMAS: Let me give you one other thing that we 11 get from experience. Some of these emergency diesels at some 12 of the power plants and also at some of the substations, they 13 have battery backups that they can get their diesel or
(
14 whatever started. It's a very small battery system, usually 15 maybe just two or three cells. Most of them were mounted on a 16 rolling frame that had wheels. The method of anchoring that 17 was with a small chain to the diesel frame or just the cable 18 itself.
19 So during the earthquake, obviously these things --
20 we've got a number of slides of where the batteries of the 21 same type, the same type cells, are just sitting there on
' 22 wheels, free to move anywhere, and we could not find any 23 evidence of any malfunctions. We did find evidence of some 4
24 impacts where it had rolled into the frame and chipped some 25 paint off. So we are seeing, really, outliers everywhere we 1
146 1 go as to what you would do in regard to batteries, and we 2 still cannot find any failures, even doing things like that.
3 MR. WYLIE: It's a good thing you didn't anchor that 4 frame.
5 [ Laughter.3 6 MR. THOMAS: One of the final items was attempting 7 to limit the scope of the relays. We have done quite a bit of 8 that through some systems type analysis -- what systems really 9 and truly could be affected by relay chatter, and also at the 10 same time, which ones do you need to worry about the most, 11 such that you make darn sure that you don't have any chatter.
12 MR. EBERSOLE: The ones that are predisposed to 13 produce or not produce this Type V event, coupling the load to
(
14 a high-pressure system would be a most interesting set.
15 MR. THOMAS: That's the one we put on one of our 16 high priorities. Also the ones that may be involved in i
17 assuring you have your emergency power system, DC power, and 18 the resulting AC through the inverters are very high priority i
19 items, to make sure you don't have any problems in relay.
20 Again, the Chile experience and surprisingly the
! 21 test data that we are gathering show that most of the type 22 relays are surprisingly rugged, even in the Chile levels. You 23 can eventually make most relays chatter if they're an 24 electrical / mechanical type, but it's a lot higher than you 25 normally expect for most relays.
147 1 Also, I guess to look at it, just based on my
)
x-) 2 experience of these older plants, a lot of these systems, even 3 in those days, used a lot of solid-state logic in small areas 4 rather than relays. The relays are not everywhere. A lot of 5 cases -- I can't say that that's what happens, but it appears 6 in some cases that solid-state may have been used in some 7 critical applications, because they are worried about relay 8 chatter.
9 MR. ETHERINGTON: Wouldn't the spring weight, the 10 ratio of the spring weight to the mass of the moving part, be 11 some indication of the likelihood of chatter?
12 MR. THOMAS: Yes.
MR. ETHERINGTON: Isn't the spring force much
) 13 14 greater than the weight in the mass in most cases?
15 MR. THOMAS: Yes, that's true. So we're continuing, 16 we're still working on this aspect of limiting -- " limiting" 17 is probably not as good a word as "trying to focus in on" --
18 which areas of relays we really spend our time on.
19 The Chile experience gave a lot better warm feeling 20 in that area. We are continuing to work with EPRI in their I
- 21 work to try to come up with some plant walk-through 22 guidelines, both from looking at outliers and looking at l
23 anchorages.
24 The implementation program itself would involve a 25 test walk-through to verify that what we have come up with on
148 1 these guidelines could truly generically be used. We would
' 2 take a test walk-through some time in the future to feed back, 3 to try to come up with a generic plan for walk-throughs.
I 4 Our plans are for the SQUG plants we would provide 5 seminars to the SQUG members and that the member himself could 6 conduct the actual walk-throughs based on the generio approach 7 developed by SQUG.
8 We feel that SQUG would provide assistance to the 9 individual members in these walk-throughs, and that we are 10 still working on plans, how would we audit these results.
11 We have discussed a SQUG audit of one of its members 12 and also maybe some audits of audits by the senior seismio 13 review and advisory panel and NRC team members.
14 So this is really ongoing work that we plan for 15 somewhere over the next six to nine months to a full year.
16 That essentially concludes that remarks.
17 MR. MARK: I have a question. Perhaps you said.
18 Is there a building code and sort of earthquake level that 19 people are preparing for on these plants that you went 20 through?
21 MR. THOMAS: In Chile?
22 MR. MARK: Yes.
23 MR. THOMAS: In general, the power plants have 24 looked at the structural design aspects from about a .2 g, as 25 far as specs on equipment, they have none. I think that's a
149 1 good point.
O 2 I would like to maybe bring up our last day in 3 Chile. We were able to arrange a meeting with some of the 4 central office people, some of the engineers that both are 5 involved in the operating and somewhat in the design of the 6 plants.
7 We had learned that they do have seismic criteria in 8 design of the structural aspects, and I asked them a question 9 particularly in regard to the control instrumentation, if 10 based on their experience and everything, that in the future 11 specifications or modifications to a plant, or for a new 12 plant, would they include seismic criteria for control
( ) 13 instrumentation or control and mechanical equipment.
14 You have to kind of appreciate we were going through 15 a language barrier, so it took a little time for our questions 16 to get interpreted, and then usually there was a team of three 17 or four people that had to talk about their answer before it 18 would come back to us. So we'd ask a question and we'd sit 19 there for 15 minutes waiting for the answer.
20 And their answer to that particular question !
21 thought was very interesting. Again the question was do you 22 plan to include any specifications for seismic design for any 23 of your modifications in regards to control equipment or 24 mechanical equipment? Their answer was, "Why should we?"
25 They feel that the experience has borne out that they don't
150 1 need to be concerned with that. They have got other things to 2 worry about, they don't have to bother with that aspect.
3 MR. WYLIE: Any other questions?
4 MR. MOELLER: Along those lines one comment one 5 could make, you had mentioned that they had had a severe 6 earthquake in the '70s and one in the '60s, '70s and '80s.
7 Maybe their plants had been pretty well shaken down.
8 [ Laughter.]
9 You know what I mean. What is left there is pretty 10 resistant.
11 MR. THOMAS: That's true. We tried to find out if 12 they had made any modifications over the years to the type 13 equipment we're worried about. Essentially they have not.
[U 14 I think one interesting point, I took out some 15 slides for time's sake, but one of the latter plants they had 16 put in some seismio considerations in piping by installing a 17 snubber. The slides show -- it's quite spectacular -- the 18 snubber ripped out the I-beam to which it was attached or bent 19 it, severely deformed it. The pipe took the snubber and 20 everything with it. No problems for the pipe.
21 [ Laughter.3 22 We had a lot of cases in the slides to where the 23 pipes totally destroyed their support system, ripped it
-~ 24 totally out of the ceiling or out of the floor. All of the V
25 supports were gone, the pipes are still there, still working,
151 1 no repairs necessary.
\s
% 2 [ Laughter.]
3 MR. MARK: One of the distinctions, which I think 4 you people called attention to, was things above 40 feet as 5 compared to things less high than that.
6 Were the things you examined including things on 7 both sides of that?
8 MR. THOMAS: Oh, yes. We added to our data base 9 very much at the Chile earthquake things way above 40 feet. I 10 think the 40 feet -- correct me if I am wrong, Newt -- but th 11 40 feet was kind of an arbitrary number selected by the Senior 12 Seismic Review and Advisory Panel, based on a data base.
13 I think with the Chile earthquake that the arbitrary
)
14 aspect of that is pointed out quite a bit.
15 MR. MARK: Good.
16 MR. WYLIE: Any other questions?
17 [No response.]
18 I would l i k e' to thank the Staff and members of SQUG 19 for their presentation.
20 MR. KERR: Can one assume if there's going to be 21 another earthquake 10 years from there, that you're preparing 22 for that?
23 MR. THOMAS: The people in Chile are assured --
24 assured us that they didn't think that this one was this s
25 decade's big one.
l
\
152 1 MR. WYLIE: Isn't it true they said since 1906, they 2 have had 12 large earthquakes? So they've got their 3 experience.
4 Does the Staff have any other comments?
5 MR. ANDERSON: No.
6 MR. HEllNAN : I would like to follow up a little bit 7 on a concern that Jesse had earlier in the meeting. I think 8 part of it was clarified. As far as new plants like Vogtle, 9 our reg guide -- I'm sorry, Standard Review Plan, Section 10 3.10, covers all plants whose cps were docketed after October 11 27th, 1972.
12 Should those plants -- those plants are required to
- 13 have fully qualified equipment. Should they have to replace l
14 equipment, they would have to go through the same process that 15 they had to go through i.nitially. For the plants, the older 16 plants that come under A-46 resolution, I talked to the EQ 17 branch chief a while ago, they would also have to go through a 18 similar comparison test, should they replace with a different 19 type component.
20 I have got e.tr e c .' ies of the Standard Review Plan 21 if any of you are interested.
22 MR. WYLIE: Thank you.
23 Mr. Chairman, that is our presentation.
24 As I mentioned, the Staff has sent the package to 25 the EDO to be sent out for public comment. We wrote a letter l
133 1 back in May of last year on this program and the question
)
%s' 2 really before the committee now, I guess, of what action we 3 take even now, or later when the public comments are received, 4 whether we write another letter or not.
5 MR. WARD: Is there any sentiment on the committee 6 for writing the letter?
7 MR. REED: I think the data base that this people 8 accumulated is very fine, and they have done a very good job 9 in the categories, the eight categories.
10 I read the past letter and it seems to me that it 11 was calling for more work, and how do you undo a letter, 12 except by producing another letter?
r It would seem to me that ACRS ought to produce a
( 13 14 letter saying what our judgment is now, with this 15 presentation.
16 MR. WARD: Charlie, do you think you have heard 17 enough to be convinced that the program that they have 18 developed since our last letter has covered the concerns that 19 were expressed?
20 MR. WYLIE: I think so. I think they have a sound 21 program and I believe it does answer the questions that would 22 be raised.
23 If you like, I will prepare a letter.
/^ 24 MR. WARD: Why don't you draft a letter and we will
. \m-25 consider it on Saturday.
i i I
154 1 Any other comment?
N
- 2 Bill?
3 MR. KERR: Are you suggesting that the ACRS write a 4 letter saying that enough has been done in some area of the 5 seismic issue?
6 [ Laughter.]
7 I just want to make sure I understood --
8 MR. WARD: This might be a good opportunity --
9 MR. WYLIE: I will turn the meeting back over to 10 you.
11 MR. WARD: Thank you.
12 Let's see. Dr. Moeller, unless your feelings will
) 13 be hurt, we will skip your subcommittee report and put it 14 back to Saturday where it was originally intended, and we will 15 go ahead with agenda item No. 5, which is the NRC maintenance 16 and surveillance program. plan.
17 Mr. Reed.I believe you lead off.
18 MR. REED: Well, that is a surprise. I'm sometimes 19 not ready when agendas get changed that fast.
20 Let me refer the committee to Tab 5. That is the 2; 1 tab that we are working on, and let me point out that I 22 recently became the chairman of the Maintenance Practices and 23 Procedures Subcommittee, although I have been a member for 24 about a year and a half, and I have been very, very surprised 25 by the change in direction and the refocusing of the
155 1 regulatory activities and the improvement that has taken place yo
( l
\_/ 2 in the maintenance program plan over the past one and a half 3 years. I guess we haven't really put out a letter in that 4 period because there has been swift change in transition.
5 However, I did write a letter in May, a proposed 6 letter in May 1984, saying -- which was critical of what was 7 going on in the maintenance program prior. I was concerned in w
8 that draft letter, which never came before the full committee, 9 it never was published, I was concerned what was going to 10 happen as far as we were going to burden the workmen in the 11 work place with lots of unnecessary paper, in my opinion, and 12 all, kinds of activities that would divert them from applying 13 their skill in the work place.
)
14' Now I am pleased. I think what we will see here 15 today -- is Dr. Booher here? I think he will be presenting 16 the status we now have. I think you will see it is 17 considerably refocused and they have selected something like 18 six technical issues that are sometimes called projects which 19 are to be the major regulatory activities in phase one. And l
20 he is in that phase, even though there has been lots of change 21 over the one and a half years that I have been following the 22 maintenance program plan.
23 Now, in addition to the six technical issues of 24 projects, which I am sure he is going to present, there are
\,
25 two issues which we might want to reflect on as he is making
156 1 his presentations, and he may touch on them just a bit.
\~sI 2 Those two issues -- one is a Japanese comparison 3 issue. I think there have been maybe three trips. I know of 4 one by INFO, and there's been one, I guess, by Mr. Denton and 5 others to Japan, to try to figure out the reason that Japanese 6 performance seems to be better.
7 So there has been this USA-Japanese interfacing, and 8 I don't think that will come up today.
9 I have been told by Dr. Booher that that is still 10 under review and perhaps in about a year there will be closed 11 session presentations trying te relate experiences from the 12 Japanese maintenance and the way they do it versus perhaps the b)
( 13 way their people do it with respect to the USA.
14 Now I think we should perhaps think about whether 15 the Japanese comparisons are closely relatable to the United 16 States power plant, nuclear power plant situation.
17 The other thing -- and it shows up in the Japanese, 18 i t ' t. ' an issue -- could be an issue -- is the natural ability 19 selection issue, and that has been taken up in a combined l
l 20 meeting somewhat which was Human Factors and the Maintenance 21 program plan.
22 That was a meeting, I believe, on June 19th, 1985' l
l l 23 where the combined meeting took place.
) 24 I do think we ought to reflect again on whether
~)
j 25 maintenance people should have high mechanical comprehension
157 1 and dexterity. We may not be able to conclude anything if we
\~/' 2 do produce the letter from this presentation today.
3 The last subcommittee meeting was on June 18th, and 4 it was attended by Dr. Kerr, Mr. Wylie, Mr. Ward and myself, 5 and I believe we had presentations from five NRC Staff 6 members. And I guess I am ready to turn the meeting over to 7 Bocher and Company.
8 (Slide) i 9 MR. BOOHER: Good afternoon. I think it has been 10 well over a year since I have been here before the full 11 committee. I have talked to the subcommittee a number of 12 times.
13 As Mr. Reed points out, there have been a number of 14 changes. One change in particular which I am very happy 15 about, is we now have some support in house. We have a new 16 Maintenance Section. Greg Cwalina has been selected as our 17 maintenance leader. He is sitting back there.
18 We have a number of other people with us today,~if l 19 you want to go into any depth on some of the issues. I just j 20 plan to cover pretty much of an overview. We talked to the -
i 21 subcommittee, and I think Mr. Reed covered quite well some of 22 the main issues that might still be handled.
23 Right now, today, I just want to go over these as a l
l 24 major outline of our objectives and scope, current status,
- \/
l 25 programs, summary. We are in Phase I, so we have Phase I l
158 1 projects ongoing. In particular, there is a survey, and we
, f\ 2 will talk a little bit about the progress there.
3 (Slide)
, 4 We almost ran out of space on this slide, so we 5 cannot afford to have too many more meetings on this plan.
6 As you can see, we start out -- last November of 7 1983, Bill Russell had the idea of having a workshop. We had 8 a couple of workshops in town. Brought in the experts from 9 around the area that we knew of, and, also in-house. We also 10 got a pilot study on maintenance indicators.
11 In January of 1984, the Commission came out with 12 policy and planning guidance, which told us to go ahead and g 13 either consider this as part of the Human Factors Program s
14 Plan, or to pursue it alone to come up with some 15 recommendation of what we can do in regulation of maintenance.
16 In May of 1984 was when we had Bill Russell down 17 after his trip to Japan, in closed session to the subcommittee 18 and the full committee n the Japanese study, part 1. At that 19 time, when we talked about the program plan, we had in 20 particular, five phases. A five-phaue program, with emphasis 21 on increased regulation fairly rapidly. This was revised 22 substantially with comments at that time. Dr. Michaelson was 23 head of the Maintenance Subcommittee, and also Mr. Reed's
' 24 comments.
l 25 By June of 1984 we had a draft plan. By this time l
l
159 1 the organization in the industry, NUMARC was formed, and had 2 some comments, fairly severe, on the program plan.
3 In the meantime we had -- through the year we 4 briefed a number of the standards groups, because our emphasis 5 was to try to get the industry moving on their own, if 6 possible. And if that were the case, then we could, as we h* ave 7 in other areas, like training, step back and see how well they 8 are doing, evaluate, and go on from there.
9 By October 1984, we had an updated plan to NUMARC.
10 An ACRS subcommittee was briefed in Seattle on the Japanese.
11 In December of 1984, our first pilot study was 12 completed. Basically, that came up with some indicators. In
- ) 13 the meantime, industry had come up with their own proposed 14 maintenance indicators, some ten indicators.
15 So, by January 1985, we finally had Phase I plan for 16 the plant approved by the EDO. NUMARC had proposed their 17 indicators, and we had a task force to look at their 18 indicators to see if we thought those were good.
19 It turns out that we did think they were certainly 20 as' good as we could come up with at this point. Industry 21 would collect data on them, and we hoped that at that time we 22 would be able to share the data to make an independent 23 evaluation.
j 24 Since then it has become at this stage, we are 1
25 not going to be able to get access to the data, and that has I
l I
160 1 held us up somewhat on that aspect of the program.
[
2 By April, we submitted to the Commission the 85-129,
(
3 the actual program plan, which is pretty much what the 4 committee has before them.
5 And in May of 1985 we started a survey project.
6 And, of course, in June we briefed the subcommittee, 7 and that is where we are today, 8 MR. MARK: Could you say a word more about the 9 remark you made, not being able to get hold of data.
10 MR. BOOHER: Yes. I had that for a little later.
11 MR. MARK: Okay, that is fine.
12 MR. BOOHER: I do have that in the plan.
13 (Slide) 14 Our specific objectives. These have been listed to 15 some degree. We say they are always motherhood, because of how 16 we measure effectiveness, and how we are going to identify 17 these practices, and all these things we would like to do.
18 The main thing I would like to point out, it is not 19 our intent in this program -- at least Phase I -- to address 20 the technical problem of how to maintain specific components.
21 Our objective is to identify factors that contribute to 22 programmatic aspects of maintenance, overall maintenance
! 23 effectiveness.
- 24 Along that line, I believe we have no contest with L
25 industry. We were saying we had something like 35 percent of I - -- - _ . . - _ _ _ _ _ . _ . _ _ , _ _ _ . _ _ __ __ _ _ . . . _ _ _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ _ _ _ _ _ _ _ _
161 1 the problems occurring out there in the field due to p)
\, 2 maintenance. Industry did their own study. They claimed we 3 were conservative, it is something like 40 percent.
4 The only issue they have is they are better at 5 correcting this problem than NRC is by coming out with new 6 regulation. So, that is not an issue. So, we are pretty much 7 in agreement on the objectives. NUMARC went along with this.
8 The only other thing was, they said they thought 9 this was probably going to cover something like 75 percent of 10 all the activities in the plant. Certainly we don't plan to 11 come out with regulations, at least right now, that is going 12 to try to fill that kind of ratchet on the industry.
1G MR. MOELLER: Excuse me. You, of course, have
}
14 ' mentioned NUMARC on the fourth item on that chart, reducing 15 occupational exposure. Is that item closely tied into the 16 group at INPO that is working with industry to reduce doses?
17 MR. BOOHER: That's correct. In fact, since we 18 called this an integrated NRC program, we are really working 19 through the people working on the ALARA pr.ogram. And NRR is 20 not directly in my section.
i 21 (Slide) 22 Now, as far as scope, again I think this is one of 23 the most important things that happened between NUMARC and 24 ourselves back and forth. As you know, in our SALP ratings we
(' 25 have a maintenance category, we have a surveillance category.
162 1 We really kind of defined -- we call our plan Maintenance and 2 Surveillance Program Plan, because we really don't think we l 3 could make major improvements in the industry just by trying 4 to attack maintenance as it was defined in the past, 4
5 particularly when we are thinking of just corrective 6 maintenance, the man out there just trying to do his job once 7 there has been a problem.
8 We also have to address programmatically, what is 9 happening out there with our surveillance and test 10 requirements. We could get into the Japanese a little bit.
11 Really, what we are talking about, is we use the 12 sur.veillance and test program to assure availability, whereas 13 the Japanese use them for preventive maintenance programs.
14 So, we cannot compare one thing against another unless we i 15 broaden our scope to include both.
16 The other thing is, we want to get beyond just 17 safety-related equipment. Again, we are trying to define the 18 status in industry practices in the Phase I Not coming up, 19 necessarily, with rules and regulations to tell industry what 20 has to be done from a regulatory standpoint.
21 So, we are getting beyond safety-related equipment.
22 We are getting into balance of plant.
23 Again, the Japanese do cover considerable -- with 24 their preventive maintenance programs, not only safety-related b
25 equipment as defined, but also others. And also the Salem i
t
{
. . - -, , . _ , , , . , . _ . . . _ - . , - . . . - - - ~ - - _ . . ,
163 1 experience has been the same way. We came out requiring Salem
\ ]
N- 2 to have a preventive maintenance program on certain j 4
3 equipments. They voluntarily have gone well beyond that f 4 because they find it is really cost effective to do that.
5 (Slide) 6 In the program plan we center around -- this is how 7 I got started. We have identified problems. Again, as I 8 said, there is no real contest with the industry in these 9 problems. We do think there is maintenance performance 10 problems. There seems to be a need for improvement in 11 maintenance that is not being performed in many areas. Our 12 inspection reports bear this out. Low SALP scores in some
\ 13 areas do.
j 14 I don't want to get into Davis-Besse, but that is a 15 potential indication there of perhaps some maintenance 16 problems.
17 Of course, resultant failures due to improper 18 maintenance performance.
19 We haven't seen this necessarily improving since we 20 started in 1983. In fact, abnormal reports to Congress are 21 still on the increase. When we did our little calculation on i
22 the percentage of them, versus other causes, maintenance seems 23 to be increasing over the past ten years, and that trend is 24 still there.
25 Maintenance operations interface. This really is
164 1 borne out by the wrong unit / wrong training problem. ,
Generic i
/ 2 issue 102, which we are working with that really is not 3 exactly the same thing. But we are coordinating that with 4 this program plan. What we are doing is looking for 5 short-term solutions first within the program plan.
6 If that means there is some longer term, we will get 7 to that as soon as we come up with our report to the EDO in 8 March.
9 Of course, the challenge to the safety systems, this 10 was a big factor again, comparing the Japanese versus us in 11 terms of trip rates. Maybe ten or twenty times greater in the 12 U.S. versus the Japanese. On the other hand, we look at 13 certain plants within the U.S. and they are performing just as 14 well.
15 So, certainly.it seems to be a possibility that 16 maybe some are doing worse than others. There are some 17 outliers.
18 In fact, when we study the 1983 data on trip rates, 19 40 percent of these were due to equipment malfunctions; 20 another 40 percent -- 82 of them challenged the reactor 21 protector system; 47 percent were due to personnel error 22 during the maintenance and test; and 73 percent of these 23 challenged the system.
24 So, it is certainly a very strong problem, a good i
I 25 indication that we should be looking at this area.
165 1 Of course, occupational exposures. The main thing 2 there is maintenance personnel account for 75 percent of the 3 exposures. If there is some way to reduce that and reduce it 4 systematically, that is something that wants to be considered 5 in the program.
6 (Slide) 7 So, what is the strategy?
8 The strategy of the program plan is not to be 9 reactive, not to just take off and come up with some ideas, 10 because frankly, we can't seem to get too many people to agree 11 on what is the right solution. So, we have taken the broad 12 scope, first of all, in order to cover maintenance and
[ ) 13 surveillance, and also components beyond safety-related v
14 equipment.
15 We are going to focus on six technical issues. We i
16 will enumerate those for you.
17 First, was human error in maintenance.
18 Second, indicators of maintenance efficiency of i
19 performance.
20 Third one was, how useful was that in counteracting l
21 aging and service wear effects.
22 Fourth was management organizational impacts. The 23 amount of resources and the communication back and forth from l . 24 management, does that have an effect positively or negatively N
25 on maintenance performance.
166 1 Of course we don't have adequate criterion and 2 standards in the area at this point to really come up with an 3 answer if someone were to look for it in the industry.
4 The other is the maintenance and operations 5 interface, because that is a key thing, we believe, to such 6 things as wrong unit / wrong train for problems. We are going 7 to use a phased approach. Again, we had the five-phased 8 approach. We are going to have a one-phased approach which is 9 to go out pretty much like we were in training and come up
~
10 with a rule, and work it from that direction.
11 We now have two phases. Phase I is to survey the 12 current maintenance practices and evaluate their (N 13 effectiveness. We have a report due in next April, 1986, on 14 that.
15 The second phase, if we get into it and it is 16 approved by the EDO, will be to then identify distinct 17 maintenance problems, determine their impact on plant safety 18 and worker dose.
19 So, for the next year we are not getting into 20 specific problems. We are trying to look at, are there 21 programm tic problems.
22 Here again the industry says they have done a root 4
23 cause analysis. I have not seen that. I have heard reports 24 on it, in which they agree that there are a lot of problems.
\
25 But, they think they are more plant specific and they cannot
167 1 come up with any programmatic areas that they would recommend b)
(_, ? going forward with some kind of generic solution. j 3 MR. EBERSOLE: Is there a topic in there anyplace 4 that you might call management of maintenance and the 5 motivation to have high levels of maintenance?
6 I get a little disturbed by the notion that the 7 workmen in the workplace --
8 MR. BOOHER: That was one of the technical issues, 9 management and organisation of maintenance.
10 MR. EBERSOLE: When you have a well-maintained 11 plant, it does not always depend on the excellence of the man 12 down at the bottom.
13 MR. BOOHER: Not in the least. I think it has got 14 to go both ways. All the motivation in the world is not 15 going to give them the skills.
16 MR. EBERSOLE: Right. So, how are you going to draw 17 out the limits?
18 MR. BOOHER: At this point we are in an 19 investigating stage, to see whether or not those are what come r 20 up as being key problems. We are not trying to define real 21 limits at this time. It is a tough one. We are trying to be 22 open and objective, and it is very difficult.
23 We just went to a plant on our first survey which I 24 will submit to you was Kewaunee. It was very difficult not to l
O"'
25 use the adjective " good" in front of things that we saw in
168 1 terms of communications and these kinds of things. Because if we are going to call them good, someone else we are going I
( 2 l
3 to have to call bad. l 1
-4 We are really not trying to do that yet. We are 5 trying to define what are the practices out there. Then, 6 hopefully, can start looking at some of the maintenance 7 indicator data coming in, to see if there are some 8 relationships between either management or skills or 9 communications. These kind of things, against the actual 10 performance of the plant.
11 But, at this point we are not trying to preguess the 12 answers.
13 MR. LEWIS: On the same point, Jesse, looking at the
)
14 practices, I am just curious at how deeply you have looked.
15 For example, we have had in the last year or so a number of 16 fairly eggregious maintenance-induced accidents at nuclear 17 power plants. The attitude of management is probably well 18 measured by finding out what happened to the people who 19 committed these sins. Are they still in place? Have they 20 been sent to China for reeducation?
21 Have you made any effort to look at case histories 22 of that kind?
23 MR. BOOHER: In our study we will have six case
/"' 24 histories. We have looked at these things in the past as they 25 have come up. It has been reactive. We always use hindsight
169 1 to say that didn't happen. But it has been very difficult to (x come up with how do you make a good management, a good 2
3 attitude? When you see it there, you know it.
4 MR. LEWIS: I'm not thinking of hindsighting, but 5 just assessing the attitudes of management by what they do to 6 people who make dreadful mistakes.
7 MR. REED: In direct response to that, Hal, I can 8 say one case that I know of where the wrong unit / wrong train 9 was violated in about eight ways. The person was terminated.
10 However, he was caused to be reinstated, at least as an 11 employee of the company, but at a coal plant. It was the 12 union.
13 MR. BOOHER: I can assure you in Phase One we have
)
. 14 a questionnaire in the protocol which very much we are going 15 to cover these areas. We have a whole category for management
- 16 and organization, and those kinds of things are there. I will 17 just say Phase One marching orders are not to prejudge. If it 18 comes out that way, we will see it in that report.
19 MR. LEWIS: I'm not talking about prejudging; I'm 20 talking about collecting specific information. That is not 21 prejudging.
22 MR. BOOHER: No question of specific information.
23 We will collect it in that area. In fact, I think there is 24 kind of a decision tree approach to some of these things. You
\s l
25 find problems there. You are going to find a lot more l
l l
1
170 1 problems on down the line. We are not trying to come up with s 2 the fact that people have made some mistakes along the way as 3 being the root cause of poor maintenance performance.
4 MR. CWALINA: Hal, if I could add to that, as part 5 of the wrong unit / wrong train surveys that we are involved in, 6 those questions are specifically asked at the site visits 7 that were going on: what happens to the people who are 8 involved in those events?
9 MR. LEWIS: That's very good, and I'm glad you said 10 that, but there are some generalities and specifics that very 11 often different people will tell you what our practice 12 is, and you look, and sometimes it is not what is enforced.
13 I have some specifics in mind which this is the
(
14 wrong forum for, but cases where the same mistake seems to 15 have been made over and over again.
16 MR. CWALINA: We do ask the specifics. As a matter 17 of fact, many of the interviews are conducted with the people 18 who actually cause the event, both wrong unit / wrong train 19 ovent.
20 MR. XERR: Has Mr. Reed asked if members are being 21 surveyed to see if they use mechanical aptitude testing? You 22 have asked that question, I . assume?
23 MR. REED: I believe I have pursued that.
24 [S11de.1 25 MR. BOOHER: Phase One projects. This is an
171 1 1 outline. The first two are perhaps the main ones, survey of 2 current practices and maintenance performance indicators. The i
3 others are ongoing kinds of activities, although we have, like 4 8 and 9 on here, we have some activity in that area. In fact, 5 on 9 we have quite a bit of activity. We don't have 6 contractor support for that.
7 CS11de.)
B The site survey of current maintenance practices.
9 We have six tasks. Pacific Northwest Laboratory competed with 10 other labs for this task, and in support'ing this. In site 11 visits we have s o m .. t h i n g like four to six people per visit and 12 about two contractors, one person from our shop, someone from
() 13 the region, from IAE.
14 As I say, there are six sites. I can name those for 15 you if you are interested: Kewaunee, Millstone -- which is 16 going to be done at the end of the month -- Brunswick, Turkey 17 Point, Arkansas 1 and Rancho Seco. Now, we have just 18 completed the Kewaunee visit. We had two purposes there. One 19 was to actually collect the data. We also were using a 20 protocol which was broken into categories, like manager 21 organisation, facilities and equipment, personnel, and they 22 were very helpful in helping us also refine that to see if it 23 was the type of document that we could use to get system data, i 24 MR. REED: Did I notice that was heavily skewed 25 towards PWRs rather than BWRs? Wasn't that about 5 to 17 l
l l
k- .- ._
172 1 MR. BOOHER: I dont' think that was intended.
( 2 Greg, would you comment on that?
3 MR. CWALINA: Yes. The purpose of that, we wanted 4 to get a wide range, which we have done, if you will notice.
5 We had two G.E. plants, two Westinghouse, one B&W and one 6 C.E. so we could hit all the major vendors.
7 MR. REED: I think that is a good idea, to get the 8 types because therein lies the difference in radioactivity and 9 radioactive exposure of maintenance workers. I have been 10 noticing the figures on BWRs versus PWRs recently.
11 MR BOOHER: We also have one from each region. We 12 looked at our own possibility of regional bias, which we 13 wanted to make sure that when we go to look at SALP scores, 14 for example, perhaps one region is not scoring with the same 15 oriteria as another, so our protocol will be used the same on 16 each site.
17 We also have a questionnaire which we are in 18 the process of developing right now which will go to each 19 regional inspector for all plants in order to also gather 20 data, similar data, but again, it is a questionnaire data. We 21 want to use the quickest form we can so by the end of the year 22 we can try to make some kind of statement about the entire 23 industry as far as practices in the maintenance area.
P 24 Of course, these other six sites will be our case
\
25 studies in much more depth.
r 173 1 MR. EBERSOLE: Do you look at maintenance training 2 programs? I just came from Georgia Power. They have extensive 3 maintenance training. Is this a common practice?
4 MR. BOOHER: That is certainly in the category. One 5 complaint we might be getting from administrators is that they 6 are in a radical change in the maintenance training area 7 because of the INPO accreditation process in which maintenance 8 personnel are, of course, some of the people who are to be 9 accredited, and those programs have been behind the operators 10 but it is a fairly radical change for a lot of them, so 11 whatever we see today may not be much representative of what 12 we might see next year, and so on.
[ 13 At any rate, we take that into account. We b} certainly are looking at maintenance training programs.
14 15 MR. MARK: Are you able to talk with the INPO people 16 to find out what approaches they are using?
17 MR. BOOHER: Talk with the INPO people?
l 18 MR. MARK: Yes, i
19 MR. BOOHER: Yes, very much. In fact, that gets j 20 into our whole training qualifications policy statement 21 agreement between INPO, industry and NRC, and we have a 22 certain amount of going along with them to observe their 23 process. We also are doing a certain amount of audits behind l
24 them to see if we agree with them.
25 We are also attending the accreditation meetings, so
174 1 all this has worked into the plan under training. So we are 2 kind of hitting it from another angle than just maintenance by 3 itself. We take that into account.
4 MR. EBERSOLE: Is there adequate incentive in terms 5 of compensation and career for maintenance workers at a 6 nuclear plant? There is a great to-do about being operators, 7 the humongous salaries, which I am worried about a little bit 8 since there may be higher salaries but no better operators.
9 In the case of maintenance like airplane 10 maintenance, I am just as much interested in how the plane is 11 put together as how it is flown.
12 MR. BOOHER: Attending the maintenance 13 superintendent workshops down in INPO, that is one of the big
)
14 complaints. If a person had a choice in the maintenance area 15 that he could he work either fossil plant or nuclear plant for 16 the same salary, he would say: Why should I do it?
17 MR. EBERSOLE: So this is analogous to saying a 18 truck mechanic can work on a 747?
19 MR. REED: Jesse, I think here is another way to 20 give an evaluation and answer to you. My experience is that 21 even though operators are well paid and they get bonuses, many 22 of them, a monthly bonus for maintaining their licenses, what 23 happens in the world that I have been in is that all the time 24 operators and licensed personnel are trying to transfer into s
25 maintenance because of the issue of day work. It is mostly
175 1 day work. There is substantial overtime in refueling events.
2 You will always have a problem. Everyone wants to get into 3 maintenance.
4 4 ! am not sure that is true of all plants, but the j 5 plants I am familiar with.
6 MR. EBERSOLE: He said the maintenance people get 7 the same salary ratings and so forth as commercial coal plants 8 and so forth. There is no distinguishing degrees of l 9 excellence or anything.
10 MR. REED: There is grease, coal and dirt in coal 4
11 plants, and radioactivity -- I think really maintenance 12 workers have not yet -- if you talk to workers -- have not
/') 13 been clubbed to death with respect to paper and diversions, 14 and I think they pretty much are still plying their skills, 15 and they ply it in a much cleaner atmosphere.
16 MR. LEWIS: Could I ask a question out of ignorance 17 just for people who actually work in it? When we say
- 18 maintenance, I suddenly realise I don't know what it means.
i 19 Do maintenance workers spend most of their time maintaining in 20 the sense of adjusting, tuning things? Do they spend most of 21 their time conducting tests required by the tech spoos or do 22 they occasionally sometimes fix things?
i 23 MR. REED: In my experience, the testing, the 24 periodio and routine ongoing testing is generally done by 25 Operations and computerized callup system on backshifts.
{
176 1 MR. LEWIS: That is not done by Maintenance.
O s- 2 MR. REED: That's true, except for instrumentation 3 and control if you want to call that also maintenance. I am 4 thinking of electrical and mechanical maintenance. During the 5 nonrefueling period when you are in operation, the maintenance 6 personnel are doing valve adjustments here and there and 7 overhauling and getting prepared for the next refueling. That 8 kind of thing.
9 Of course, during refueling, they are very much 10 involved in the big aspects, such as head lifts and pump 11 overhaul and those kinds of things.
12 MR. LEWIB: So it is the sort of thing I would
) 13 normally think of as maintenance.
14 MR. KERR: I expect they spend a significant 15 amount of their time filling out forms and signing things too, 16 don't they? .
17 MR. REED: It is a growing thing, the paperwork 18 aspect.
19 MR. KERR: Mr. Booher, in your trip to Kewaunee did 20 you encounter any significant surprises?
21 MR. BOOHER: I have not been totally -- I didn't 22 personally go. I wasn't on the site visit. I don't know if 23 we have anyone here who has actually done -- Brinkman from I&E 24 was on that.
s 25 MR. BRINKMAN: I'm Don Brinkman. I'm with the
177 1 Office of Inspection and Enforcement.
2 I accompanied Dr. Booher's people on the visit. I 3 don't believe we really discovered any surprises there other 4 than we were pleased to see the attitude of the workers at the 5 Kewaunee plant. They seemed to be very diligent in their 6 work.
7 MR. KERR: You were surprised at that?
8 MR. BRINKMAN: I was surprised at the level of their 9 interest in their work.
10 MR. REED: Weren't you also surprised by the number 11 of personnel or lack of number of personnel?
12 MR. BRINKMAN: I was quite surprised at the small 13 staff they had.
)
14 MR. MARK: Were you well received?
15 MR. BRINKMAN: Yes, we were.
16 MR. MARK: That should have surprised you.
17 [ Laughter.3 18 MR. WARD: Does anyone else have any suggestions for 19 things Mr. Brinkman might have been surprised at?
20 [ Laughter.)
21 MR. KERR: I was curious because if he wasn't 22 surprised at something, it seems to me there wasn't there a 23 whole lot of point in making that trip. You spend a lot of 24 effort and personnel on something of this sort. I would 25 anticipate that you would encounter something unusual.
178 1 Apparently not.
2 MR. BOOHER: You must realize also this plant was 3 picked first because it does have a good record. One of the 4 main purposes of this was not necessarily to find surprises or 5 problems; it was to see how well we could use this protocol 6 because we are going to be using it again on other plants 7 ultimately. If it works, we hope to use this in Phase Two as 8 actual probable assessment criteria tool for regions and for 9 audits and this sort of thing.
10 So we want to make sure we are catching everything 11 in terms of being able to describe it, not necessarily good or 12 bad at this point.
13 MR. REED: I can't pass up this opportunity. I'm w
14 sorry. I might point out Kewaunee is one of those plants that' 15 uses natural ability selection testing and has used it for 16 years.
17 MR. BOOHER: We had to add that one in.
18 MR. LEWIS: So does my university, but it doesn't 19 help. .
20 MR. BOOHER: Maintenance performance indicators.
21 Again I say we started something Research had been working on 22 indicators, developing this for some time. Everyone agreed it i
{ 23 would be nice if we had some quantitative measurements to help 24 our subjective opinions used by the experts.
b 25 CS11de.3 4
179 1 When NUMARC said they were going to come up with and 1
2 develop the indicators, that sounded great, and we have held 3 off considerably waiting to reach some kind of agreement, 4 perhaps we could get this data so that by the end of next year 5 we could have independent evaluation of their statistical 6 analysis of the data, and see if these things really do what 7 we hope they will, 8 In our case we would like to see whether or not 9 there is some kind of distribution. Some we might want to put 10 more concentration on than others. Unfortunately, at the 11 present time, we received a letter from NUMARC -- I believe it 12 was just last month, a few weeks ago -- which stated that it 13 looked as though we would not be able to receive the data from 14 NUMARC, because INPO has really collected the data for them, 15 and they are having a difficult time, as you know, getting any 16 kind of access to plant-specific data which would be given to 17 INPO.
18 Also there is a lawsuit going on right now which is 19 another problem. If we get the data, would we be required to 20 give it in the Freedom of Information Act? This is all up in 21 the air. Right now it looks like we are going to have to go 22 out and collect data on as many of these indicators ourselves, 23 as long as it is publicly available data.
24 MR. REED: Are you in agreement -- I believe there 25 are 10 INPO indicators which looks very solid to me.
180 1 MR. BOOHER: We saw nothing wrong with those. We 2 added two other ones which INPO itself said they wanted to 3 collect on those two others, but they didn't know how to do 4 it. That was going to be a follow-on. As far as sitting 5 around this room with experts saying, hey, this is okay, 6 everyone is in agreement. As far as having any real data to 7 follow up and validate it and see if it really relates to 8 something that we can call safety, we don't have that. But at 9 least it is certainly the best starting point we could think 10 of.
11 As I say, we are not going to get that data directly 12 from INPO.
13 MR. REED: As I remember the indicators -- and I see 14 three of them here on your next page right away -- they are 15 available to the NRC every month?
16 MR. BOOHER: That is right. Those we can get.
17 Others we are going to try and get.
18 MR. REED: I noticed those others that look on 19 radiation, there is a minor separation there. I think you are 20 trying to oull out for maintenance workers. That might not be 21 available, but of course under 10 CFR 20, all the other 22 radiation information --
23 MR. BOOHER: I'm not discouraged. I think the fg 24 problem is we could have been doing this six months ago, but
- U 25 we waited to get -- agreed to do this together, and we didn't l
l I
i l
I 1
l
l 181 )
l 1 want to go off doing something -- collecting data at the same l 2 time when we are going to get it from another source. And 3 that was just something that was unfortunate, but we can't do 4 much about it.
5 I do think we can get that data. Not only that-- I 6 don't know, maybe we might even be able to get some 7 voluntarily, because the industry does have it in the proper 8 form now.
9 A lot of the big problem with NUMARC was having 10 cooperation from utilities, what they were asking for, not 11 everyone had it in the same form. There was a great deal of 12 work to get it. How do you define preventive maintenance or T 13 corrective maintenance in terms of reporting these kinds of 14 figures?
15 MR. EBERSOLE: I wonder if you could tell me a 16 little bit about the make,up of NUMARC a r.d its qualifications 17 and authority and what makes it so good.
18 MR. BOOHER: What makes it so good, it is staffed 19 totally by --
l 20 MR. EBERSOLE: How big is it?
21 MR. BOOHER: How big an organization?
22 MR. EBERSOLE: Yes.
23 MR. BOOHER: As I understand, it doesn't really have
-s 24 a budget at this point. It's all been volunteer, and so it is v But they claim to represent the 25 a pretty small organization.
L_
182 1 entire utility industry.
2 MR. REED: Jesse, I think there are about 50 k
3 executives, utility executives, who are on the NUMARC 4 membership list. Is that correct?
5 MR. BOOHER: That's essentially it. It was formed 6 at the time when we were thinking about coming up with 7 operating experience rule, something like 50 people appeared 8 to the Commission one day and all said we're vice presidents 9 or above, and here is our initiative, we will take care of the 10 problem.
11 And I think it was a grand step forward. Because if 12 they made a decision, it was going to happen, they had the 13 money to back it up. And we found it has come through in the
)
14 operating experience beoause of new plants being licensed.
13 That's where it came down. Either they have shift advisers 16 for the requirements we set, or six months, so that was a real 17 hard problem to correct'.
18 MR. REED: Arb you going to keep the pressure on to 19 try to pry that data loose?
20 MR. WYLIE: Yes, definitely.
21 MR. REED: The plant people, in the final analysis, l 22 it comes down to the poor plant organization to produce the i
i 23 data. So anything that can unburden repetitious data, fine, 24 you can keep the pressure on. I'm sure you should be 25 successful
I
- l 183 1 MR. BOOHER: The pressure is still coming from 2 helping Vic Stello's group. This may not be -- particularly 3 if this lawsuit gets settled, something could be worked out.
4 Also perhaps we can get the data without having it 5 specifically identified which plant it is. If we have 80 6 points, we can do some statistical analysis.
7 On the other hand, they said we could figure out who 8 the plants were if they did that.
9 [ Slide.]
10 At any rate, I think things are moving along on 11 schedule. The main thing, industry is still going to have 12 this data when they present it a year from now. We would like 13 to be able to say yes, that's good data, it's validated right, 14 it really does show the trends. I think we should have some 15 kind of independent look at it. And that's what we will be 16 working on.
17 MR. AXTMANN: Did you select Salem with a dart 18 board?
19 MR. BOOHER: Salem was selected strictly because 20 they're the only plant that we have required to have a 21 preventive maintenance program, and they have now the 22 experience of having come from having none to having one, both 23 in terms of also a lot of cost data on what that has taken --
. 24 MR. EBERSOLE: They inherited that from that event, N/ 25 didn't they?
184 1 MR. BOOHER: Certainly did.
)
N ,/ 2 MR. EBERSOLE: I guess Davis-Besse will inherit the 3 same thing?
4 MR. BOOHER: I can't speak to that at the present 5 time.
6 Oh, Project 4 was a participation in standards 7 groups. A number of the standards groups have been doing work 8 and since 1983 they really did get off the dime in a lot of 9 areas, particularly IEEE has been working -- every couple of 10 months have been meeting and trying to come up with some good 11 practices themselves. ANS 3.9 was a subcommittee formed. The 12 chairman there is actually Tom Fitzgerald from American
/M He's been trying for five years to try to 13 Nuclear Insurers.
(N-14 get something going. That's kind of in a hold status right i
15 now because 1NPO says they are also providing good practices 16 along the same line.
, 17 And we do have members in all of these. Tommy Lee l
18 has been designated as a member, who comes from the I&E groups 19 and works in our section and is off at one of those meetings 20 right now.
21 CSlide.]
l 22 Program integration. This is nice work, but we are
(
l 23 really trying very, very hard to get our act together within 24 NRC. I think just as a good example, research reliability 25 program, they have been coming up with a little questionnaire l
l l
185 1 or survey themselves, rather than going off to different
{D
\s,/ 2 plants. They are going to be coming with us on the Millstone 3 plant. They are looking at it for a different purpose. They 4 are trying to look at what types of reliability programs are 5 out there. They are not all the same, and get data on that.
6 But they will be working together on that.
7 We looked at some of the programs that are shown in 8 the program plans.
9 [ Slide.]
10 We actually have a pretty good write-up, I'm 11 impressed with the write-up. I didn't write it, some of the 12 people on my staff did. It really tells you what some of the 13 other areas are doing. When we specify actual interface areas
)
14 for each of these programs, the phase one or phase two, and 15 what it would be we would expect 16 Also when they have a report, we don't just look at 17 it and quickly comment. We are actually analyzing it. If wo 18 can't do it, we have our contractor do it and write interface 19 descriptions of how one program relates to the other, and how 20 one could help the other, and that will be in our final report 21 as well 22 So that is a fairly big effort on the Staff's part, 23 but it's another thing that we want single phase solutions in 24 these areas.
(
25 [ Slide.)
186 1 The sixth program is the analysis of the Japanese k 2 vs. US maintenance programs experience. I cannot say a whole 3 lot here.
4 John Jankovic, who is the project manager of the 5 program, is here today. We have a NUREG CR-3883 which is on 6 the street, which tells the results of the comparison we did.
7 Again we have stopped work in that area because 8 indsutry has sent their own team over to look at the Japanese 9 practice. We found, I guess it's VEPCO has also been doing 10 things on their own, trying to adopt some of these practices, 11 and they seem to be very happy with it.
12 I think what is going to happen next year when we 13 get into phase two is we hope these results will be very f^N
\
14 useful in helping determine the appropriate balance between 15 surveillance testing and preventive maintenance. What is the 16 trade-off there. That's going to beccme the real crux of the 17 issue. Can you back off on sone of the surveillance and test 13 requirements in return for an increase in preventive 19 maintenance. Is there a tradeoff there. Because that is 20 really what we are comparing, when we are comparing the 21 Japanese vs. us.
22 MR. WARD: Are the Japanese looking to do just the 23 opposite?
24 MR. BOOHER: The Japanese are not looking to do more O 25 surveillance and tests. They are looking to do less
187 1 preventive maintenance. They think they are overdoing it
(,, 2 because of their long outages.
3 Of course, that again is required by the 4 government. So they are looking for a little data to say that 5 they can shorten those.
6 MR. REED: I know at the subcommittee meeting you 7 mentioned how you had paired plants for evaluation, and we had 8 some comment about that, and I wondered if perhaps they should 9 have been paired on performance statistics more than vintage.
10 How do you still feel about that? For instance, I 11 have the feeling that if a Kewaunee plant was compared against 12 say the better Japanese, you would not find all that much 13 difference.
14 MR. BOOHER: That might be a comparison in the 15 future. What has been done has been done.
16 John, would you like to give us the rationale for 17 why we did it that way? I know it seemad at the time it was 18 the only way to do it.
19 MR. JANKOVIC: I am John Jankovic, from the 20 Maintenance and Surveillance Section of Dr. Booher's staff.
21 I would like to point out that one of the objectives 22 of the study was to identify a number of philosophies, 23 management approaches to maintenance. And we did look at 24 differences, not just similarities.
\
25 In that respect the report was successful, because
I, 188 1 we identified the Japanese approach, emphasizing preventive s ,/ 2 maintenance as the primary element of their maintenance 3 philosophy, while we didn't find it at those plants which were 4 selected from the U.S.
5 MR. REED: Well, I keep thinking that maintenance is 6 very much tied to human beings, their aptitudes, skills, 7 vocation, and these kinds of things. And I'm not so sure that 8 philosophy -- that is how many times -- whether you do lots of 9 surveillance testing or whether you do balance of plant 10 maintenance, yearly, in depth -- I'm not so sure that that is 11 going to lead us to maintenance improvement in the long run.
12 So I keep digging away at look at the human aspects 13 in the maintenance activity. And I was happy to hear that you 14 surveyed Kewaunee and came away impressed. But I also know 15 that Kewaunee is that kind of dedicated, motivated, coupled, 16 small organiaction, and I am sure they minimize paper and they 17 have done a good job. The record speaks for it.
18 MR. BOOHER: There may be similarities between the 19 Kewaunee attitude and Japanese attitude. It may be difficult 20 to get the rest of the industry --
21 MR. REED: There's a thread there. That's why I'd 22 like to see if it can be put together.
23 MR. BOOHER: As I say, VEPCO, for example, is trying 24 to evolve on this, really trying it.
25 John, can we say something about that?
189 1 MR. JANKOVIC: About two weeks ago VEPCO management sx
%s 2 gave a presentation to our Division of Licensing, and the 3 subject was their upgraded approach to maintenance. And they 1
4 told us that they tried to implement the Japanese quality 1
I 5 circle approach, and they had their maintenance superintendent l 6 here and even members from the first selected team. There 7 were stx craft people in a team plus the foreman, and they get 8 indoctrination about the quality, importance of quality of 9 their work, and they get a little training, not technical 10 training, but team approach to work.
11 Within that team they are free to do assignments to 12 each other. But maybe the most important aspect of this team 13 approach is that before they start a job, they have a pre-job V)
I 14 discussion, what are the critical aspects of the work, what 15 they have to watch out for in particular, and once they finish 16 the job, they have also a little discussion. Was the 17 procedure correct, was their approach correct, what 18 improvement can be done.
19 This was the first time from Surry, and they --
20 VEPCO will have another team in September in place at North 21 Anna. They are scheduled, of course, for transferring the 22 entire maintenance department to this approach, to this team 23 approach, by next September.
24 MR. REED: I believe, and I have read some on that, d 25 I believe the quality circle approach -- I believe it probably
190 1 is a better thing to work where you have large organizations
(%
k, s
2 and lots more people, and that is the way to couple things 3 together. There is a difference, though, I think, between 4 that and the Kewaunee thing, where you have very small 5 organizations who, on an individual basis, do perform like 6 teams.
7 So the human side-is very important --
8 MR. BOOHER: Certainly we as human factors 9 specialists don't want to ignore that at all.
10 MR. JANKOVIC: I would like to add a closing remark 11 to our visit to Kewaunee. One of the observations we made was 12 that the personnel in the maintenance department is unique 13 there, because they have been there since the beginning of the
)
14 plant. They participated in the construction. They are very 15 dedicated to their organization, and their turnover rate is 16 less than 2 percent a year, 17 MR. REED: Do you think that is unique? If you were 18 to go out and check all the plants in the United States, y'ou 19 think it's that unique?
20 MR. JANKOVIC: At the moment we think it is unique.
21 We will be able to tell you more half a year from now when we 22 finish our survey.
23 MR. BOOHER: I guess we need to move along.
I 24 [ Slide.3 25 The seventh topic is one that we do not have
191 1 actually formulated yet. Mr. Reed has discussed that with 2 other members of the Staff, and we have his comments and 3 recommendations. This is an area that we inherited from the 4 Research organization because their funds have been cut in the 5 Human Factors and Research. So we do not have an actual 6 ongoing task right now. This is sort of a list of things we 7 feel we probably need, in that area, and we will be working, 8 of course, closely with INPO and the accreditation program in 9 anything we do there.
10 [ Slide.]
11 The eighth area is, we've tagged along on a bigger 12 research study, which is a feasibility study investigation,
) 13 investigating ultrasonic testing. Apparently Pacific 14 Northwest Lab is doing what the call round-robin experiment or 15 test of inspectors. They have six sets of two sets of 16 inspectors. They have been trained by EPRI on pipe crack 17 detection, and they want to see how well this holds up in a 18 different situation. All we're doing is participating in that 19 exercise at a very low level, but looking for any human 20 factors aspect that might come out of that.
21 I think one of the findings that they're coming up 22 with is, inspectors who are on there are complaining because 23 they don't get any feedback and knowing whether or not they're
, 24 doing the thing rightly or wrongly. There seems to be a wide l xY 25 range among the inspectors, where supposedly they are all i
192 1 qualified at the same level They're getting something like a 2 39 percent difference in their performance on these people who 3 have been trained.
4 I don't know how the data is going to turn out 5 ultimately, but at this point, it just was a chance for us to 6 look at some human error potential in something like crack 7 detection.
8 [ Slide.]
9 Our final project in Phase 1 is to participate in 10 the human error in wrong unit / wrong train units. At his 11 point, we are about halfway through looking at the plants.
12 The, goal here was to look at short-term solutions, low cost, 13 and what we're doing, of course, is starting with the LERs and 14 actually going down to the plant and trying to analyse the 15 root causes, in many cases talking to the person who committed 16 this error to find out why they did it.
17 Again, preliminary information is not real 18 surprising. One area is labeling. Another area is 19 procedures. But a couple of other areas seem to be getting 20 into the area of maybe like fatigue. Some people thought 21 these courses were very exhausting at the time and tend to 22 have sort of a mindset that was just sort of going, not really 23 thinking, down old paths.
24 It was still too early to come up with any real 25 answers here.
1
193 1 MR. REED: I hope if you are interviewing these f-'s
(,) people and trying to find out why they did it, as you said, 2
3 that you realize that the individual will tend to rationalize 4 his action, and that you ought to be checking with an 5 evaluation. I'm sure the companies make evaluations.
6 MR. BOOHER: I think this is probably just one data 7 point. It's got to come all the way down to look at it.
8 Drew, you participated in some of these. Would you 9 care to give Mr. Reed a flavor for how this is done?
10 MR. PERSINKO: My name is Drew Persinko. I'm in the 11 Maintenance Surveillance Section.
12 I've been going on the wrong unit / wrong train
/-
13 interviews thus far. As Dr. Booher said, interviewing the 14 particular individual who made the mistake is only one of the 15 points that we talk to. We also speak to the maintenance 16 supervisor, production supervisor, somebody very knowledgeable 17 with the particular events that have occurred, so we get both ,
18 sides of the picture.
l 19 We just felt it would be good to speak to the 20 individual. We thought we could get some of his viewpoints, 21 what he was seeing as he was walking out there. I think that 22 has been useful That's where we get some of the idea that.
23 his mind -- he had been practicing on something for a week, g 24 and his mind was set on it. It seemed like almost nothing 25 would have changed his mind at that point.
4
194 1 MR. REED: I hope you people are very cautious in
(,) 2 doing that kind of thing. You can get crosscurrents to union 3 grievance proceedings and all kinds of activities. It's like 4 going out and asking a criminal, perhaps in a jail, why he is 5 there, and he certainly would say, "I'm not guilty, and I 6 shouldn't be here."
7 So you've got to be very careful when you get into 8 the human aspects of a utility activity. I'm a little 9 surprised that the utilities aren't a bit turned off by you 10 getting into that depth.
11 MR. PERSINKO: Most, in fact just about all utilities have been very interested in what we've been doing.
~
12 13 We realize the problems that may arise with unionization. In 14 one case -- only one case out of all the interviews that we 15 have conducted -- the individual wanted somebody there frou 16 the union. But this wasn't forced upon him. He didn't have 17 to speak to us, if he didn't want to. But they all agreed to 18 it, and it had been checked out with the union ahead of time.
19 For the most part, all the utilities and the 20 individuals involved have been very receptive to what we have 21 been doing.
22 [ Slide.3 23 MR. BOOHER: As a wrap-up, just to give a feel for 24 the progress we have made since the past year, both NRC and C) 25 industry have made progress, I think it probably would be
195 1 worthwhile if some of the industry members could come and O 2 brief the ACRS -- NUMARC, INPO -- I don't know how long it's 3 been since you've had a discussion from them on what they are 4 doing, particularly in maintenance.
5 I won't go through all these, because we've already 6 covered most of them. WY dc have the new section. We are 7 staffed up. We're into Phase 1. We're on schedule. Draft 8 reports should be coming out like maybe March of next year.
9 Industry, in the meantime, we have kept the pressure on.
10 Standards groups are formed; they're actually coming along 11 with some things.
12 I think this next year is going to be very, very .
critical on whether there is true progress, and we will have a
) 13 14 baseline from them on to measure progress, 15 MR. REED: I'm a little surprised by your 16 recommending that maybe the ACRS should have the industry come ,
17 and brief them. I had not considered that. I am finding that 18 the Staff, I think, is doing a very good job in having 19 industry come in. They might see things differently. I could 20 ask them the question.
21 Is the NRC Staff interfering and going down and 4 22 interviewing people and raising a ruckus?
23 MR. BOOHER: We're open to that kind of criticism.
24 We work very hard to avoid it. I don't think we're going to d
25 have that at this point.
4
196 i
1 MR. REED: We will ask the question of the Full 2 Committee, if they want to be briefed by industry. Quite 3 frankly, I have been so pleased with the progress in the last 4 year and a half that I don't feel it's necessary.
5 MR. BOOHER: I'm glad to hear that. Myself, I am 6 just a little impatient. I feel some of these things, like 7 the indicators in particular, we hoped to be a little further 8 along on.
9 Also, I'm not quite sure at this point, without a 10 little more communication, of how well industry is doing i
11 themselves on some of the tasks they said they are doing that 12 we have kind of backed off on.
13 MR. REMICK: What you want is indicators of
[
J 14 individual plants, is that it? You don't want the industrial 15 averages.
16 MR. BOOHER: That's right.
We don't want an 17 amassment of averages. The Regions have pointed out very 18 definitely that we certainly don't want anything to replace 19 something like our subjective experts with. numbers that we 20 cannot really tie to what does that mean.
21 Also, we would like to see what is their plan for 22 collecting the data, and what do these numbers mean.
23 MR. WARD: Do you have any indication from sampling, 24 for example, that you can expect there to be large significant 25 differences, a significant range in the values of these
-. _ _ _ ~ , . _ _ _ _ _ __ __
197 1 indicators?
2 MR. BOOHER: We don't have anything at the present 3 time. I really don't know how it's going to turn out. We 4 just have a feeling -- in order to pick those indicators, I 5 have a lot of confidence that those do reflect differences in 6 performance.
7 MR. WARD: You said a lot of people have agreed on 8 those being valid indicators subjectively?
9 MR. BOOHER: The reason they've done that is, within 10 the NRC, it was a task force which was headed by at least 11 Branch Chiefs or above, which had representation from the 12 AEOD, who have worked on problems of this nature, from IE, 13 from two Regions, and all agree that these indicators, by (O) 14 name, are fine. But again, the problem is, numbers mean 15 nothing if we don't know what lies behind them. And those 16 indicators, if valid and reliable, are great.
17 Industry, themselves, spent quite a bit of time 18 through the NUMARC organization and INPO -- they went through 19 something like a hundred possibilities and narrowed down to 20 these ten. One was: Are these something that would look like 21 it would be meaningful Secondly, are they something they can 22 get the data on?
23 Even so, it was difficult getting data on those 24 ten. I know C. O. Woody said that was one of their most
"}
a 25 difficult problems, because at the beginning, some of the
198 1 utilities said, "We're going to have to back off some of the
(\
\s ,/ 2 work we're doing in maintenance in order to get you this kind 3 of data." Now that they have it, the skids are greased to 4 provide that kind of data.
5 MR. MARK: You s. a i d , I believe, close to when you 6 started, that from the LERs, I think, the number of 7 maintenance problems is increasing rather than falling off.
8 MR. BOOHER: That was the abnormal events reports to 9 Congress. It wasn't necessarily the LERs. The number is 10 significant, events that have been reported to Congress, Last 11 year, we presented a curve showing something like four or five 12 of those a year, maybe. The percentage of those which have 13 maintenance involvement or maintenance causes have been
(
14 increasing.
15 MR. MARK: Is there a measure of that increase? Has 16 it doubled in five years, or has it gone up ten percent in 17 five years, or what?
18 MR. BOOHER. I don't have it off the top of my head.
19 John, do you recall.
20 MR. MARK: I don't want a number.
l 21 MR. JANKOVICH: We plotted the regression equation 22 for the data. When you compare the absolute number of 23 abnormal occurrences to the number of maintenance-related, we 24 find that the slope is twice as much as the number of evens, i \
i 25 So the maintenance-related evens go up twice as fast as the
199 1 number of other all together.
I k 2 MR. MARK: Is there a term in there that has to do 3 with the details of the reportings that are now available as 4 compared to some time ago?
5 MR. JANKOVICH: The reporting system has been 6 constant during the period, the past ten years. There were 'n o 7 changes in this reporting system, not like in the LER system.
8 MR. MARK: It's the only thing in the agency that 9 has that property.
10 (laughter.]
4 11 MR. BOOHER: Again, we say we think it just was what 12 got us into this area. We used whatever indications we could 13 to see if there is a problem. Industry is not disagreeing
)
14 with us, at least in the meetings that we have both been to 15 together.
16 MR. REED: Is that the end? That's the total s
17 presentation, the one-hour presentation? ,
18 MR. BOOHER: Yes.
19 MR. REED: It looks like it's very timely, and I 20 would like to thank you for it. I think that's a good way to l
l 21 make the presentation. One person, if he can handle it all, i
22 rather than the others who did in subcommittee, who put it all 23 together.
l 24 I would like to thank you and the people from your
\
l 25 group that are here to fill in on some of the answering.
I 200 1 Now I would like to ask you a question: Do you want 0 2 a letter from the ACRS?
3 , MR. BOOHER: I would like a letter. I have talked 4 with Bill Russell. I think, yes, we would like a letter. I 5 think you have seen where we started. It's been over a year.
6 We are into Phase 1. We would like a letter, if it has -- you 7 have the plan; you can see where we are. Have you got 8 criticisms or further directions, anything that can be done?
9 We would welcome it.
10 MR. REED: Suppose the letter had a comment in it 11 that we think the plan is moving along fine? I personally 12 feel that this is a very complex issue, the whole maintenance f 13 issue. It's human-related. There are ways of doing 14 maintenance one way, and you can come out with a very good 15 job. There are ways of doing it another way in a different 16 situation, a different nature, different time, and you come 17 out with another recommendation.
18 So I would like to think that at least the 19 subcommittee is kept informed before you make final 20 conclusions and put them on the street. Are we going to be 21 informed?
22 MR. BOOHER: I don't see any problem. For example, 23 we intend, as soon as we get the protocol finalized, even 24 though we're using that, and also the questionnaire, we would 25 like the subcommittee to look at these to see if they are
201 1 reasonable, these indicators. We would like to work along 5
2 with the subcommittee.
3 Again, Phase 1, the whole idea is not to come up 4 with answers, but to try to get the best grasp of the problem 5 that we can, so that in Phase 2 we can actually work on the 6 problems, and everyone is in agreement that these are the 7 problems that we should be working on.
8 MR. REED: So you would try to show the subcommittee 9 draft proposed releases and documents, perhaps before you're 10 too far down the pike with them on the street.
11 MR. BOOHER: I think that would probably be 12 reasonable. We would have to clear that higher up, but I have 13 no problem.
14 MR. REED: We would like to be informed, I think.
15 Thank you very much.
16 MR WARD: Okay,. Let's take a break, a ten-minute 17 break, and come back for the future activities discussion.
1B [Whereupon, at 4;30 o' clock, p.m., the transcribed 19 portion of the meeting was concluded.]
20 21 22 23 24 25
O 1 CERTIFICATE OF OFFICIAL REPORTER 2
3 4
5 This is to certify that the attached proceedings 6 before the United States Nuclear Regulatory Commission in the 7 matter of: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS e
9 Name of Proceeding: 304th General Meeting (Public Session) 10 11 Docket No.
() 12 Place: Washington, D. C.
13 Date: Thursday, August 8, 1985 14 15 were held as hereen appears and that this is the original t
16 transcript thereof for the file of the United States Nuclear .
l
! 17 Regulatory Commission.
18 (Signature) ,
,g l (Typed Name 'of Reg 6rter) /Su'zagfe B. 55dng 20 21 22
[ 23 Ann Riley & Associates, Ltd.
24 25 I --
s f%
U 1 CERTIF1CATE OF OFFICIAL REPORTER 2
3 4
5 This is to certify that the attached proceedings 6 before the United States NucIear ReguIatory Commission in the 7 matter of ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 8
9 Name of Proceeding: 304th General Meeting (Public Session) 10 11 Docket No.
12 Place: Washington, D. C.
Is cate: Thursday, August 8, 1985 14 15 were held as herein appears and that this is the original 16 transcript thereof for the file of the United States Nuclear 17 Regulatory Commission.
19 .
3 'g 39 (Signature) 3, ,
/Q [
(Typed Name of Reporter) Mimie Meldzer 20 21 22 23 Ann Riley & Associates, Ltd.
24 25
i '
l 1 I l
- i lO l l
GESSAR II SEVERE ACCIDENT ISSUES l
4 i
i 4
i i
i A PRESENTATION TO THE ADVISORY l
- COMMITTEE ON REACTOR SAFEGUARDS ,
i I
O WASHINGTON, D.C.
i I
l l GENERAL ELECTRIC COMPANY
' AUGUST 8, 1985 l
\
s\ :
I O
- - . - - . . , - _ . . _ _ . . _ . _ . . _ _ . _ . . - - . . - _ - - _ . _ - - . _ _ _ ., ...---,,--_.__-,.y---,_.,-,..-_,,_,-
HYDROGEN ISSUES g
o RATE AND AMOUNT GENERATION RATES VARY FROM 0.4 TO 1.6 LB M 1300-2300 LB M TOTAL IN-VESSEL HYDROGEN ONLY EN0 UGH 0XYGEN TO SUPPORT COMBUSTION OF 2480 LBM HYDROGEN (+ 67 PERCENT OF ACTIVE CLAD MWR) o HYDR 0 GEN DETONATIONS INSIGNIFICANT RISK REDUCTION FOR ADDITIONAL HYDROGEN CONTROL (BASED ON PRA RESULTS WITH DETONATIONS)
CURRENT UNDERSTANDING--LOW LIKELIHOOD OF DETONATIONS IN MARK III RISK EVEN LOWER THAN ORIGINAL PRA RESULTS SER SHOWS NO RISK REDUCTION FOR HYDROGEN CONTROL FOR INTERNAL EVENTS, FACTOR OF 2 FOR SEISMIC RISK (BASED ON DRYWELL FAILURE BY LOCAL DETONATIONS, GE ANALYSES g
l DISAGREE) o GE COMMITMENT: PROVIDE A HYDROGEN CONTROL SYSTEM CONSISTENT WITH OUTCOME OF HC0G PROGRAM AND NRC REVIEW NRC REQUIRING DIVERSE POWER SUPPLY FOR IGNITERS (BEYOND HC0G POSITION OF POWER FROM EDG)
GE FINDS NO TECHNICAL JUSTIFICATION FOR DIVERSE POWER SOURCE o GE POSITION:
HYDROGEN CONTROL UNNECESSARY--ABSOLUTE RISK ALREADY LOW NO JUSTIFICATION FOR IGNITER SYSTEM ON COST-BENEFIT BASIS O
O EFFECT OF STANDING FLAMES ON SEALS e ISSUE: CAN STANDING FLAMES FROM HYDROGEN DEGRADE DRYWELL SEALS LEADING TO P0OL BYPASS?
e ASSESSMENT:
DRYWELL EQUIPMENT HATCH HAS A 5 FOOT CONCRETE SHIELD PLUG PERSONNEL AIRLOCKS ARE DOUBLE SUBMARINE DOORS WITH CEMENT SHIELD PLUG ON WETWELL SIDE ELECTRICAL PENETRATIONS ARE 5 FOOT LONG AND POTTED WITH A PORTLAND CEMENT MIXTURE e CONCLUSION:
NO EFFECT OF STANDING FLAMES ON DRYWELL SEALS O DAH r
O ABLATION OF RPV PEDESTAL o PEDESTAL IS A STEEI.-CONCRETE COMPOSITE CONSTRUCTION TWO CONCENTRIC STEEL SHELLS CONNECTED WITH STEEL SHEAR TIES CONCRETE FILLED BETWEEN THE SHELLS o EVALUATED SUPPORT CAPABILITY AFTER ABLATION ASSUME LOSS OF 1.4M OF CONCRETE ASSUME ONLY SUPPORT IS OUTER STEEL SHELL ASSUME OUTER SHELL TEMPERATURE IS 1100*F o RESULTS LOADS ON OUTER SHELL WEIGHT OF RPV 2300 KIPS WEIGHT OF SHIELD WALL + E0PT WEIGHT OF PEDESTAL 2700 KIPS 1100 KIPS TOTAL 6100 KIPS i
COMPRESSION IN STEEL SHELL = 3.4 KSI YIELD STRENGTH OF STEEL AT 1100*F = 21 KSI o CONCLUSIONS PEDESTAL WILL CARRY LOADS - SUBSTANTIAL MARGIN NO LOSS OF PEDESTAL, DRYWELL OR CONTAINMENT STRUCTURAL INTEGRITY O
O .
GESSAR-II PRA REVIEW DETAILED DISCUSSION OF HYDROGEN PRESENTED BY TREVOR.PRATT -
BROOKHAVEN NATIONAL LABORATORY
.O ueTON. New YORK 11973 l PRESENTED TO THE ACRS AUGUST 5, 1985 BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVER51 TIES, INC.(llll
O HYDROGEN ASSESSMENTS GESSAR II PRA REVIEW:
- BASED ON FULL CORE MELTDOWN ACCIDENTS
- INITIAL SUBMITTAL INCLUDES NO PROVISION FOR H 2
.. CONTROL DU8(NG SEVERE ACCIDENTS
' [' .
- CONSEQUENTLY, VERY HIGH PROBABILITY OF EARLY CONTAINMENT FAILURE AND SIGNIFICANT PROBABILITY 0F EARLY LOSS OF DRYWELL INTEGRITY (VIA DETONATION)
- CONTAINMENT EVENT TREES IN SUPPLEMENT N0 2 SER (NUREG-0979) DO NOT CONSIDER Hz CONTROL
- IMPACT OF H2 CONTROL ADDRESSED IN SUPPLEMENT NO. 4 TO SER BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(ILll
HYDROGEN ASSESSMENTS (Cont.) .
HC06/NRC INTERACTIONS:
- DEALS WITH DEGRADED CORE ACCIDENTS (CORE REMAINS IN-VESSEL)
- AIM IS TO MAINTAIN CONTAINMENT AND DRYWELL
-' INTEGRITY.BY H2 CONIRSL. .
.O - a^1es ^wo anouwt or n2 cewea^1 tow aae ineoatant FOR DESIGN OF H2 CONTROL DEVICE
- ISSUES RELATED TO DELIBERATE IGNITION:
OPTIMt!M IGNITION SOURCES TYPE OF POWER SOURCE LIMITATIONS OF IGNITION SOURCES EFFECT OF STANDING FLAMES BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(llli
O HYDROGEN DETONATIONS H 2 GENERATION H2 DISTRIBUTION
.~ ..
- - POTENTIAL FOR DETONATIONS
,.O -
IGNITION SOURCE
- DDT
- MAGNITUDE OF SH0CK LOAD
- RESPONSE OF STRUCTURES TO LOADS BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED Ui,WERSITIES, INC.(llll
I H, GENERATION .
l l
~
- RATE OF RELEASE FROM PRIMARY SYSTEM . ,
l
- TOTAL AMOUNT GENERATED
- GE AND BNL APPROACHES SIMILAR (BASED ON MARCH CODE)
- GE AND BNL PERFORMED SENSITIVITY STUDIES TO ASSESS IMPACT OF UNCERTAINTIES
..O
- FOR
REFERENCE:
- MASS ZR IN CLADDING 72,000 LB
- MASS ZR IN BOXES 64,000 ts POTENTIAL H2 MASS 6,200 ts
- 1,700 ts H2 PRODUCES 20 VOLUME PERCENT IN CONTAINMENT O BROOKHAVEN NAll0NAL LABORATORY l} lj l A5500ATED UNIVERSITIES, INC.(Illl
l H, DISTRIBUTION l
GE CALCULATED DISTRIBUTION USING IN-HOUSE CODE
- BNL DISTRIBUTION BASED ON IN-HOUSE CODE AND HECTRE CALCULATIONS AT SNL
-O EXAMPLE: 50 ts/ MIN FOR 27 MINUTES (1350 LB TOTAL Hz RELEASE)
BROOKHAVEN NAll0NAL LABORATORY l} gyl A5500ATED UNIVERSITIES, INC.(llll
O
~~
H2 - 50. LB/ MIN,FOR 26.667 MIN .
C d
S.
g . . . . .
A.
g . . . . . . . .
_2 .
,9 R. . .
5 O 8 a:
- 29. .
zo .
b z
S ggR. . . .
EE E
E e- .
c
- g. .
8 d'
8 i, g
0.0 t$.0 25.0 36.0 45.0 56.0 50.0 TIME - MINUTES, 9 ZONES w
QUANTIFICATION OF CET H2 IGNITION: .
- DELIBERATE IGNITION DEVICE NOT INCLUDED
- HIGH PROBABILITY OF IGNITION AT TIME OF POWER RESTORATION H2 FLAMMABILIT.Y LIMITS: __ .
- >4 VOLUME PERCENT UPWARD PROPAGATION
-.O -
>9 VOLUME PERCENT DOWNWARD PROPAGATION H2 DETONABILITY LIMIT
- >18 VOLUME PERCENT PROBABILITY OF H2 EVENTS BASED ON PROBABILITY OF POWER RESTORATION AND FRACTION OF TIME LIMITS ARE EXCEEDED BROOKHAVEN Nail 0NAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(llll
COMPARISON OF CONDITIONAL PROBABILITIES -
FOR H, PHENOMENA (CLASS' I TRANSIENT WITH LOOP - POWER RESTORED PRIOR TO VESSEL FAILuilE) .
ItE EL GLOBAL DETONATION 01 00 GLOBAL COMBUSTION 03 0 66 O
LOCAL DETONATION 03 0 08 LOCAL COMBUSTION 03 0 26 BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Ill1
O DYNAMIC LOADS RESULTING FROM H, PHENOMENA GE AND BNL CALCULATE SIMILAR LOADS FOR H2 DEFLAGRATIONS -
GE ASSESSMENT OF H2 DETONATIONS P
PEAK
-' - *= 17 (STRUCTURE PARALLEL TO WAVE)
P INITIAL
-O P PEAK
- = 41 7 (STRUCTURE PERPENDICULAR TO WAVE)
P INITIAL
- WAVE TREATED AS EQUIVALENT TRIANGULAR PULSE NRC ASSESSMENT OF H2 DETONATIONS BASED ON CALCULATIONS AT SNL USING CSQ CODE O BROOKHAVEN Nail 0NAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl
RESPONSE OF STRUCTURES
- DYNAMIC SH0C.K, LOADS CONVERTED INTO EQUIVALENT STATIC ;
LOADS BY DYNAMIC LOAD FACTORS
- EQUIVALENT STATIC LOADS COMPARED AGAINST CAPACITIES OF STRUCTURES
.~ .. __ .
- GE CONCLUDED:
DETONATIONS FAIL CONTAINMENT GLOBAL DETONATIONS FAIL DRYWELL ROOF UNDERWATER NRC CONCLUDED:
DETONATIONS FAIL CONTAINMENT POTENTIAL FOR FAILURE OF DRYWELL WALL AS WELL AS ROOF UNDERWATER BROOKHAVEN NATIONAL LABORATORY l)l)l AS$00ATED UNIVERSITIES, INC.(1 til
HYDR 06EN BllRN EVENT TREE FOR RE! .ATION OF POWER REFORE CORE StilMP i
GLOBAL EVENT DETONATION DRYWELL BREACH FAllllRE IN REMARKS (SPATIAL DIST.) (CONCENTRATION) (STRESS ANALYSIS) DRYWELL HEAD DRYWELL WALL Fall l
0.0 l 66 10 '
NO FAIL Y
t i
Fall allENCHED
+ 5 N
8 5 FAIL NO QUENCH l 34 24 2 NO FAIL
. 76 j -
NO FAIL i
BROOKHAVEN NAll0NAL LABORATORY l} l} l A5500ATED UNIVERSITIES, INC.(IIll
Table 15.1 O Conditional consequences predicted by the staff for internally initiated events and probability of -
occurrence with and without UPPS, per reactor year Release Early Early Latent - Probability category" Person- .
fatality injury fatality rems I
w/o UPPS w/UPPS !
1-T- L3 0 0 40 7 x E5"" 3 x E-6 9 x E-7 1-T-E3 0 0.0005 200 3 x E6 8 x E-6 1 x E-6 1-T-12Q 0 3 200 '
, 3 x E6- 1 x E-5 1 x E-6 2-T-83 0 .
0 300 5 x E6 ~:
4 x E36 - 4 x E' -7.- .
' ATWS 0 '
1 400 '
6 x E6 3 x E-6 3 x E-6 -
14T-I2 0 6 500 8 x E6 3 x E-6 3 x E '
1-SB-f1 0.006 it- 600 - 9 x E6 1 x E-9 1.x E-9 .
"See definitions in Table 15.15.
- 7 x E5 = 7 x 105..
- Notes: ,
.- (1) All conditional'mean consequences were calculated using the upper range BNL source term values described in SSER 2.
If2) The calculations assumed the Shippingport site, with public evacuati~on within 10 miles and relocation 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after plume passage.
(3) Mean consequences were computed over 91 different weather conditions.
9 0 -
e
O DEFLAGRATION I -$LohA ~- NO IGNITORS -
'- FAILURE OF WETWELL SEAL ASSUMED UNIT PROBABILITY
- SMALL PROBABILITY OF RPV PIPE BREACH OR DRYWELL SEAL FAILURE LEADING TO El OR E2 RELEASE, - - -
g OTHERWISE 8 RELEASE
- ABOUT FACTOR OF 3 IN
.' PERSON-REM CONSEQUENCES O LOCAL - WETWELL SEAL MAY FAIL
- SMALL PROBABILITY OF RPV PIPE
~
BREACH OR DRYWELL SEAL FAILURE LEADING TO E2 OR E3 RELEASE OTHERWISE L3 RELEASE
- ABOUT FACTOR OF 4 IN "'
l PERSON-REM CONSEQUENCES
l l
LOCAL DETONATIONS
. O .
1-SB-El PORTRAYS DRYWELL AND WETWELL EARLY FAILURE CREDIT FOR PRIMARY SYSTEM RETENTION AND POOL g
SCRUBBING OF VOLATILES -
PERSON-REM CONSEQUENCES ABOUT AN ORDER OF MAGNITUDE GREATER THAN 1-I-L3 :.
1 - T - 12, 1 - T - 120 PORTRAYS'DRYWELL HEAD FAILURE
.._ _ g DUE TO DETONATION SHOCK LOAD
- ABOUT FACTOR OF 3 IN PERSON-REM CONSEQUENCES DEPENDING ON FAILURE LOCATION-O g l
l O .
O STATUS OF HCOG CONSIDERATIONS
- NRC STAFF POSITION ON ACCEPTABLE HYDROGEN RELEASE HISTORIES DEFINED IN LETTER FROM BERNER0 TO HOBBS, DATED JUNE 24, 1985
- CASE A: 150 sm STARTED 3100S AFTER SCRAM
- CASE B: 5000 em FLOW
- CASE C: CASE A FOLLOWED BY 0 1 ts/s H2 UNTIL 751 MWR
- AB0VE WILL BE USED FOR 1/4 SCALE TEST PROGRAM
- HCOG TEST PROGRAM TO CONFIRM ADEQUACY OF DELIBERATE IGNITION
- HCOG TEST PROGRAM WILL NOT TEST FOR OPTIMUM IGNITION SOURCES BROOKHAVEN NATIONAL LABORATORY l} g)l
' ASSOCIATED UNIVERSITIES, INC.(I(11 l
l
O ~
EFFECT OF STANDING WETWELL HYDROGEN FLAMES DETAILS PROVIDED BY DR PARCZEWSKI (NRC) IN APPENDIX A TO NUREG-1037 (CPWG REPORT)
HEAT FLUXES PROVIDED BY CLWG.(NUREG-1079)
.O - SEAL TEMPERATURES ARE SIGNIFICANTLY ELEVATED BUT REMAIN BELOW FAILURE
- LATER IN ACCIDENT HIGH DRYWELL TEMPERATURES DURING CORE / CONCRETE INTERACTIONS MAY CAUSE SEALS TO EXCEED FAILURE LIMIT BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(IIll
f- 7 4
O .
l
~
- l '; ... ,
EFFECT'0F A CORE MELT ON VESSEL SUPPORT INTEGRITY l iN.I l l
l
. PRESENTED BY l TREVOR #RATT BROOKHAVEN NATIONAL LABORATORY v- O _
UPTON, NEW YORK 11973 A
1 PRESENTED TO THE ACRS
_. AUGUST S, 1985 BROOKHAVEN NATIONAL LABORATORY l} g)l
! A5500ATED UNIVERSITIES, INC.(llll
l O
TOPICS Y. i
, ' _. i , is :
ABLATION OF SUPPORT u . :
- SIGNIFICANCE OF LOSS OF CONTAINMENT INTEGRITY FOLLOWING SUPPORT FAILURE
- EFFECT OF CONTAIN!(EW VENTING O BROOKHAVEN NATIONAL LAB,0RATORYl} g)l ASS 00ATED UNIVERSITIES, INC.(llll
(
j
_ i
.r ,,,, -SHIE' D .
.~.;*,
sut:.Dwo
/. .
. "?
- .? . . jf -CCCA1NMsNT f
'. 4 bO '
, 5.'
HsAo r.
', #~'
, UPPER POOL 3w.,J;",gd.b wm .r- -- .** 7 p P :
nrw
.., e e g . I. -
w wr
. .x
. c.,
- , .m'
- . m . . . ..-. . =v '
.,..,..~..,...' pSRV DtSCHARGE RsAcTon .
w
.g-1, -
D '.,:
~
$ '.*. . . . ; .'s P R
- ,,,, REAl" TOR -\.
- n. .. F.. 7
/ -
1.. sHisto WALL R..
- gf ,r .:-
j/
- .< :. .; ,. ,,, 0RywstL s.
- a. . , . ..
- . $1 -
- :-
..; W"a- ,
< '.. J.
,.i .. .. ,; .
Y- ': :- -
- :: -won wAu.
'(
/ -
p 'p .. ..,. 5 ')l.
Jr ,W G
gli g Q:
~'
W' ';
- l':'D $W f; .: -
o===
f .5
~.
?"** M..;l.' i $
.T:kh....h.;5. ..::: .a. .nl'.::\.'.'.i??.*
~~
R_..i.4 'Wii'"*"
- .: ::.:c.:i.:c.::::i..;:w:,:::: :i:.i.:. :. :r... g ::;
'I rg ... ... ... .. ..... ........ . . . .......
I l.i /
lIk o2 Wy ""P'f'tN@.?
I l
/f $ L d il I "
l .
I Figure 15.1 Principal feat,ures of MARK III c6ntainment -
1
_ _ _ _ _ ~ _ . _ - - - _ - . . , _ . _ . e ,0
g O.
'453 5 53.e5.g3. M g a sug, pass . -- - qm , guemaggg ,a gu g,3 -
CESSAIN2 TQUXLP CORCON-MARCH 2-15.1 5-21-85 y
a.
a -
m . ..
U gg .0 .
4 INTER - g
+ ...
D. -
E me. ~CORCON.
a .
~
X
% h< ..
A,,
O ,
U 304- - _ _ _ - _ _ _ _ _ . . _ . , _ _ , _ _ _ _ _ _ _ , , _ _ _ _ ,
. J .
M' E
g ...
l 08 2000 #9e 6008 e008 30000 82004 34000 36000 38000 2000 0 1
TIME - (MINUTE) .
l VOLUME No.1 8 r.
g
- 4
J
.d \
- o k S Q ,
o i
e B
o o - - Oe g 00
- ~
e .
o -
D i o
VESSEL. b
. T:5.1700 X d' y .6n c. . . . . _ . . _ .
og g 80 06 -
o s
~'
8,*e o I
l pg 3_.4 O
1
- 4 o 0 3 g
~
')
O -
4 1,gg p - - - -
g Q 0 O
-T9 , .~ --
O o .
o.
.O a . g
~
- # o g # p g i p ao 8 oo -
6.lM 00 '
l 0
- o '
3.6m o O O 7.7K : o 0 e ;
o 0 0 i '
O D TO WETWELL' CORIUM g 1.6M 0 0 0 - 0 4
000 8 o - f f o l, ,
6 3 0 b 3 0 9
g 'D O o -
gg 'Q 3 i -e 3M r g e, o 0, o O ,O o O o o o 4.6M TO PENETRATE 3 o 0 0 0 i BASEMAT d "o 5
- O
- o O g0 C g o o , o b g
' *o o "
, O ..o o s
- - - ---.-,---,_.__---,----.--.,?_------
LATION RATES
^
= j }@0NCRETE)(I ABLATION ~- I INITI AL) + k ABLATION h~
l j = 10 - 20 W/cM2 (LOWER CAVITY)
. y=12-3 W/cM2 (SURROUNDING) p = 2.5 y/cM ,3C = 1. J/p/K, 2,5 240 J/Gn g = ABLATION RATE = 10-20 cM/HR (LOWER CAVITY)
INITIAL ._
= 12-3 cM/HR (SURROUNDING)
,, fC 140 cM RQIAL, PEDESTAL INTEGRITY DOUBTFUL IHERMAL GRADIENT f T=*<BTf g - .
1 HR I INITIAL = 300K __
+ 25 cM g 20 HR T INITIAL _ _ _. _ _ _._ __
-< 60 cM e
O :
SIGNIFICANCE OF LOSS OF VESSEL SUPPORT MEASURE EFFECT RELATIVE TO RISK ESTIMATES IN TABLE 15 9 0F SUPPLEMENT 11 TO SER (NUREG-0979)
- EARLY LOSS OF CONTAINMENT INTEGRITY
. ' .~ . -- .
- LATE CONTAINMENT FAILURES (L2, L3) BECOME EARLY FAILURE (12, 13)
Q
- EARLY LOSS OF CONTAINMENT INTEGRITY PLUS LOSS OF DRYWELL INTEGRITY COMPLETE POOL SCRUBBING SEQUENCES (E3, 13, L3, B3) BECOME PARTIAL POOL SCRUBBING SEQUENCES (E2, 12)
'I O BROOKHAVEN NATIONAL LAB,0RATORYl} l)l ASSOCIATED UNIVERSITIES, INC.(llll
I Table 15.1 Conditional consequences predicted by the staff for internally initiated events and probability of occurrence with and without UPPS, per reactor year Release Early Probability
' Early Latent Person-category
w/o UPPS w/UPPS 1-T-L3 0 0 40 7 x E5** 3 x E-6 9 x E-7 1-T-E3 0 0.0005 200 3 x E6 8 x E-6 1 x E-6 '
1-T-12Q 0 3 200 3 x E6 1 x E-5 1 x E-6 2-T-B3 0 0 300 5 x E6 4 x E-6 4 x E-7 ATWS 0 1 400 6 x E6
^
3 x E-6 3 x E-6 1-T-I2 0 6 500 8 x E6 3 x E-6 3 x E-7 1-SB-El 0.006 10 600 , 9 x E6 1 x E-9 1 x E-9
,tSee definitions in Table 15.15. , , ,
"7 x ES = 7 x 105 Notes:
"_ (1) All conditional mean consequences were calculated using the upper range BNL source term values. described in SSER 2.
(2) The calculations assumed the Shippingport site, with public evacuation within 10 miles and relocation 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after plume passage.
(3) Mean consequences were computed over 91 different weather conditions.
es e O .
GESSAR II SSER 4 15-31
i O O O h Table 15.9 Public risk from internal events (person-rees per unit per year) for GESSAR II base case y and with design modifications l m <
} O '
Unlimited J m generator !
i XI Base UPPS and cand UPPS Unlimited f l
" GESSAR case perfect 10-hour DC Unlimited [andperfect generator i i
- w/o Perfect with hydrogen battery charger UPPS and generator ihydrogen and UPPS :
j Release
- UPSS Ha control UPPS control capacity generator ignitors and UPPS ~ control igniters
\ .
l 1-T-E2 3 -
0.5 -
1 , 1 -
0.3 ','- -
) 1 -
0.2 -
0.4 0.3 - O.1 ~-
1-T-E2Q -
l j -
- ..s. P l' l 1-T-E3 23 . 4 -
10 8 M* 9',j 2* -
4
- r. ,i .
1-T-I2 22 -
3 -
9 7 -~*'. , 1 -
T, g 1-T-12Q 31 -
4 -
12 10 - ' . 1 - -
b 1-T-I3 12 -
~2 -
5 4 h 2- 0.5 -
0.5 1-T-L2 * '
l' "
1-T-L3 2 22 0.6 3 1 1 0. 6 - 0. 5 2 0.5 l 1-58-El 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 j II-T-83 20 20 2 2 20 20 2 2 2 2 I
ATWS 18 18 18 18 18 18 'S 18 18 10 i
i Total 131 59 33 23 76 68 31 25 22 25 I
- See Table 15.15 for a description of the release categories.
j i .
! i l
O SIGNIFICANCE OF LOSS OF VESSEL SUPPORT (CONT.)
EARLY LOSS OF CONTAINMENT INTEGRITY:
GESSAR W/0 UPPS: 131 PERSON-REM PER YEAR WITH EARLY' LOSS: 139PkRSON-REMPERYEAR N---
- EARLY LOSS OF CONTAINMENT INTEGRITY PLUS LOSS OF DRYWELL INTEGRITY:
GESSAR W/0 UPPS: 131 PERSON-REM PF YEAR WITH EARLY LOSS: 227 PERSON-REM PER YEAR 1
l l
BROOKHAVEN NAll0NAL LAB,0RATORYl} g)l A5500ATED UNIVERSITIES, INC.(llll
O errect os cowtaiwaewT vewrive
- CLEAN" VENTING:
- ATTEMPTS TO MITIGATE CLASS 2 AND ATWS SEQUENCES
- MEASURE EFFECT RELATIVE TO RISK ESTIMATES IN SER
- CLASS 2 S500ENCES TrliiNIFICASTLY REDUCED BY UPPS
~' - ABILITY TO MITIGATE ATWS BY VENTING UNCERTAIN
- VENTING AFTER CORE DAMAGE: I
- MINIMAL IMPACT ON EARLY H2 PHENOMENA HENCE H2 CONTROL NEEDED EVEN WITH VENTING l
BROOKHAVEN NATIONAL LABORATORY l} g)l ASSOCIATED LINIVERSITIES, INC.(lll1
O SOUG Q)NCLUS10tG
- SEISMIC RESISTANCE OF STN0ARD POWER PLANT EQUIRB1T, WHEll PROPERLY l ANCHORED, WAS VERIFIED DLRING TE PILOT PROGRNi,
- EXPLICIT, SE!911C QUALIFICATION OF THIS EQUIRBIT IS 70T JUSTlFIED,
- SEISMIC QUAllFICATION IS KIT A SimlFIC#iT SAFETY CONIRN, TEREFORE, RlRTHER ACTION IS P0T EQJIRED.
O l
1 O
O V
SUBJECT:
QUESTIONS FOR OPERATOR INTERVIEWS DU IN CHILE 1.
What was plant status prior to earthquake?
- What are 2.
Does the plant have any special earthquake procedures?
they? Are copies avail &ble? -
l t During strong motion in the event did automatic If auto-action of the p a 3.
systems take place? What were these automatic actions?In
? What the abs action did not take place, should it have? action (base '
What alarms were i>nitiated? or manual)?
action did he take?Did the plant respond properly (auto o
infomation? Were 4.
Af ter the strong motion was over, what was Were the plant status?
systems any auto-actions which were Was the needed operator taking required place?
to take manual resetting to nomal?If so, what action did he take and did the plan action?
it was supposed to?
Was off-site power lost?
q 0 Was auxiliary power lost?
Was diesel power available?
Q 8 8
8 Was d.c. power available and was 1oad or to relay chatter?
were there any 5.
In determining failures or damage to equipment:i l nature? ) Were misoperations or malfunctions of equipment to relay chatter?
there misoperations or malfunctions of equipment due Type of relay?
8 What type of equipment?
8 8
Any damage due to inducedMaintained improper 0
Any problems with momentary contacts on switches?
contacts? Problems?
Any mercury switches?
8 Any system change of state not attributable l ?
to r 8 Were any printed circuit cards broken or other such fai 0
Were there problems with cables or cable teminations?
i ?
6.
Were there structural failures Pipe supports?which affected systems func Large pipes 2 1/2" small?
air? ? Other degraded 7.
Were there reduced or increased flows in cooling systems O
v functions?
CRT's?
- 8. Any damage to Control Boards?
- 9. What worked that wasn't expected to work?
What failed that was expected to work?
- 10. What people related problems were experienced? Access to tools, procedures, damage control equipment, comunications, etc.
- 11. What secondary events occurred? Fires? Spills? .
- 12. Any problems with equipment in operation? Cranes? Portable equipment? Maintenance in progress?
- 13. Is there' seismic monitoring equipment at the site?
- 14. For problems encountered at power plants, were they considered a
" systems" problem or an " operation" problem?
9 Steam cycle?
8 Condensate?
8 Feedwater?
4 Power?
O l
l 0
.~
O Questions related to switchyard functions:
- 1. Have there been any design changes to prevent damage that would require action?
2.
Are there any special switching arrangements ' developed for earth response? Any line isolation provisions?
3.
Are there any special provisions for starting if off-site power is lost due to the earthquake? (blackstart) 4.
Were any problems encountered in synchronizing the plant with the transmission system after the earthquake?
5.
Were there any degradations in the Anyrelaying comununication syste special procedures What type system - microwave or carrier?
related to degraded comunications?
Did it create any problems?
- 6. Does dispatch system exist?
O e
O
, SQUG PROGRAM OUTLINE O
- 1. SCREEN ESSENTIAL EQUIPMENT LIST 0 COVERED IN SQUG PROGRAM .
0 OTHER DATA AVAILABLE (EXPERIENCE, TEST) '
0 ENGINEERING JUDGENT
- 2. DOCUMENT SEISMIC RDGGEDESS QF EQUIPMENT
~
0 ASSIGN RUGGEDNESS LEVELS WHICH CAN BE' JUSTIFIED 0 IDENTIFY EXCEPTIONS /YE.NERABILITIES FOR EACH EQUIPENT CLASS O DEFINE DATA NEEDS, IF ANY
- 3. C0FLETE/ REVIEW EPRI PROGRAMS, DEVELOP ANCHORAGE INSPECTION GWIBELIES
! e ANCHORAGE o TEST DATA ASSIMILATION O
', 4. DEVELOP SIMPLIFIED APPROACH FOR DETERMINING REQUIRED !
SEISMIC RUGGEDNESS IN NUCLEAR PLANTS O O ELEVATIONS LESS TRAN 40 FEET 0 HIGHER ELEVATIONS
- 5. ATTEMPT TO LIMIT SCOPE OF RELAY FUNCTIONALITY REQUIREMENTS ON GENERIC BASIS
- 6. DEVELOP PLANT WALK-THROUGH 6UIDELINES AND TEAM
- 7. PERFORM ' TEST" WALK-THROUGH
- 8. DEVELOP PLANS FOR SQUG' MEMBER IWLEENTATION O O SEMINARS 0 GENERIC SQUG TEAM APPROACH 0 SSRAP/NRC AUDIT
\
O -
i l SQUG ACTIVITIES PROMPTED BY ACRS/ STAFF CONCERNS 4
- l. DEFINITION OF GENERIC EQUIPMENT REQUIRED TO ACHIEVE
- 2. DEVELOP RATIONALE TO ASSURE SEISMIC RUGGEDNESS OF EQUIPMENT BEYOND THE 8 CLASSES DEFINED IN THE PILOT PROGRAM.
i
- 3. FUNCTIONALITY DURING STRONG-MOTION (PRIMARILY RELAYS) ll . SOUG'S GENERIC IMPLEMENTATION PLAN
. EQUIPMENT SCREENING
. PLANT WALKDOWN PROCEDURES 4
. AUDIT FUNCTION O
- O O -
O USI A-46 ACRS PRESENTATION AUGUST 8, 1985
SUMMARY
OF USI A-46 PROGRAM T. Y. CHANG N. R. ANDERSON PROPOSED RESOLUTION, SCOPE AND BASIS i
IMPLEMENTATION REQUIREMENTS STATUS OF ONG0ING SQUG/EPRI ACTIVITIES J. THOMAS (SOUG) l -
ANCHORAGE GUIDELINES TEST DATA BASE DEVELOPMENT RELAY REVIEW PROCEDURE SOUG GENERIC IMPLEMENTATION PLAN i
O O -
O BACKGROUND
- SEIMSIC SAFETY MARGIN IN OPERATING PLANT EQUIPMENT MAY VARY CONSIDERABLY
- SEISMIC QUALIFICATION OF EQUIPMENT IN OPERATING PLANTS NEEDS TO BE REASSESSED l
- PROBABLY NOT PRACTICAL TO SEISMICALLY QUALIFY OPERATING PLANT EQUIPMENT USING CURRENT CRITERIA ,
- NEED TO DEFINE ALTERNATIVE METHODS
- r. y. ,
l O O -
O SEISMIC QUALIFICATION OF EQUIPMENT USING SEISMIC EXPERIENCE DATA l
STAFF ESTABLISHED FEASIBILITY OF USING EXPERIENCE DATA (LLNL STUDY)
SOUG CONDUCTED PIL0T PROGRAM 10 COLLECT AND EVALUATE SEISMIC EXPER
- ADDITIONAL EXPERIENCE DATA COLLECTED FOR C0ALINGA, MORGAN HILL, CHILE EARTHOUAKES I
- SSRAP FORMED JUNE 1983, JOINTLY SELECTED BY SOUG AND NRC SSRAP ISSUED REPORT IN JANUARY 1985
! SSRAP DEVELOPED RULES FOR USE OF DATA NRC STAFF PARTICIPATED IN DATA EVALUATION AND CLOSELY MONITORED SQUG/S PROPOSED STAFF POSITION BASED ON USE OF SEISMIC EXPERIENCE
O O -
O SCOPE OF PILOT PROGRAM
- GATHERED AND DOCUMENTED EQUIPMENT AND EARTHOUAKE PERFORMANCE DATA FOR EIGHT CLASSES OF EQUIPMENT MOTOR CONTROL CENTERS LOW-VOLTAGE (480 v.) SWITCHGEAR i METAL-CLAD (2.4 To 4KV) SWITCHGEAR UNIT SUBSTATION TRANSFORMERS MOTOR-OPERATED VALVES AIR-0PERATED VALVES Il0RIZONTAL PUMPS AND MOTORS VERTICAL PUMPS AND MOTORS
" INVESTIGATED DATA ONcm 3000 ITEMS OF EQUIPMENT IN CONVENTIONAL (NON-NUCLEAR) PLANTS
_ _ _ - - _. . = _ . _ _ .
O O -
O SSRAP CONCLUSIONS FOR 8 EQUIPMENT CLASSES j EQUIPMENT INSTALLED IN NUCLEAR POWER PLANTS IS GENERALLY SIMILAR AND AT LEAST AS RUGGED AS THAT INSTALLED IN CONVENTIONAL POWER PLANTS I THIS EQUIPMENT, PHEN PROPERLY ANCHORED AND WITH SOME RESERVATIONS, HAS AN INHERENT SEISMIC RUGGEDNESS AND HAS A DEMONSTRATED CAPABILITY TO WITHSTAND SUBSTANTIAL SEISMIC MOTION WITHOUT STRUCTURAL DAMAGE FUNCTIONALITY AFTtR THE STRONG SHAKING HAS ENDED HAS ALSO BEEN DEMONSTRATED, BUT Tile ABSENCE OF RELAY CHATTER DURING STRONG SHAKING HAS NOT BEEN DEMONSTRATED 1
e
O O -
O EQUIPMENT BEYOND 8 CLASSES
\'
4 NO REQUIREMENT FOR COLLECTING ADDITIONAL SEISMIC EXPERIENCE DATA
!
- THIS CAN BE PROVIDED BASIS FOR SEISMIC ADEQUACY MUST BE DOCUMENTED FOR EACH EQUIPMENT TYPE.
BY:
VERIFICATION EQUIPMENT EXISTS IN DATA BASE PLANTS TEST DATA CURRENTLY BEING COLLECTED BY EPRI/SQUG
't
- .- - -- ----- .a
O O O l
THREE CONCERNS
- 1. EQUIPMENT ANCHORAGES
- 2. RELAY OPERABILITY
- 3. OUTLIERS
O O -
O PROPOSED RESOLUTION OPERATING PLANTS DEVELOP EQUIPMENT LIST PERFORM WALK THROUGH INSPECTION VERIFY ANCHORAGES VERIFY FUNCTIONALITY OF EQUIPMENT (RELAYS)
IDENTIFY a ADDRESS DEFICIENCIES AND OUTLIERS NEW LICENSEES l
NO REQUIREMENTS IMPLEMENT BY-GENERIC LETTER
i ,
O '
O -
O SCOPE OF SEISMIC ADE0VACY REVIEW ASSUMPTIONS SSE DOES NOT CAUSE LOCA LOCA DOES NOT OCCUR SIMULTANE0USLY WITH OR DURING SSE 0FFSITE POWER WILL BE LOST DURING OR FOLLOWING SSE
- MAINTAIN HOT SHUTDOWN FOR A MINIMUM 0F 72 HOURS, EQUIPMENT SCOPE ACTIVE ELECTRICAL AND MECHANICAL COMPONENTS INCLUDING INSTRUMENTATION NEEDED TO ACHIEVE AND MAINTAIN HOT SHUTDOWN
- ANCHORAGES ON TANKS, HEAT EXCHANGERS REQUIRED TO ACHIEVE AND MAINTAIN HOT SHUTDOWN
- NO REQUIREMENT T0 (1) REVIEW MASONRY WALLS, (2) REVIEW SOME AUX FEED SYSTEMS (3)
INSPECT RCS PIPING (4) REVIEW SEISMIC INTERACTION ITEMS PLANTS AFFECTED ABOUT OPERATING PLANTS NOT REVIEWED TO CURRENT CRITERIA AS DOCUMENTED BY SE 49 SITES, 72 UNITS. SEP PLANTS WILL BE REVIEWED FOR FUNCTIONAL CAPABILITY ONLY
1- .
O O .
O IMPLEMENTATION REQUIREMENTS DEVELOP EQUIPMENT LIST VERIFY ENVELOPE OF SITE FREE FIELD SPECTRA BY APPROPRIATE B0UNDING SPECTRA .
WALK-THROUGH INSPECTION ANCHORAGE REVIEW IDENTIFICATION AND REVIEW 0F." DEFICIENCIES" AND " OUTLIERS" IDENTIFY ALL EQUIPMENT THAT MUST FUNCTION DURING STRONG SHAKING RELAYS ARE MAJOR CONCERN REVIEW 0F EQUIPMENT UNIQUE TO NUCLEAR PLANTS i
l REPLACEMENT PARTS
i '
O O -
O RELAY REVIEW GUIDELINES NRC GENERAL REVIEW GUIDELINES IDENTIFY ALL RELAYS ASSOCIATED WITH EQUIPMENT NEEDED TO BRING PLANT TO HOT SHUTDOWN RELAYS WHICH MUST FUNCTION DURING STRONG SHAKING:
VERIFY WITH TEST DATA REPLACE WITH QUALIFIED RELAYS QUALIFY BY TEST RELAYS WHICH MUST FUNCTION AFTER STRONG SHAKING:
VERIFY, REPLACE OR QUALIFY AS ABOVE; OR LICENSEE SHOW CHATTER OR CHANGE OF STATE DOES NOT AFFECT PLANT SHUTDOWN RELAY VERIFICATION CAN BE DEFERRED UNTIL TEST DATA BASE COMPLETE
o RELAY REVIEW (CONTINUED)
.o -
o l SQUG DEVELOPING REVIEW PROCEDURE IDENTIFICATION OF RELAYS TO BE EVALUATED DEFINITION OF FUNCTIONALITY REQUIREMENTS DEVELOPMENT OF EVALUATION PROCEDURES 4
REVIEW BY NRC STAFF AND SSRAP i
CHILEAN EARTHQUAKE CONFIRMS NEED TO REVIEW RELAYS SCOPE OF RELAY REVIEW TYPICAL BWR (DRESDEN/LASALLE) i 1000/1200 RELAYS 6/8 RELAY TYPES l
TYPICAL CE PWR (CALVERT CLIFFS)
- 1100 RELAYS
- 6 RELAY TYPES, 25-30 MANUFACTURES l
i -
TYPICAL B8W PWR (0CONEE)
!
- 750/900 GENERAL PURPOSE / INCLUDING PROTECTIVE RELAYS
!
- 25 MANUFACTURES 1
l i
O O -
O l RELAY REVIEW (CONTINUEU)
TYPICAL M PWR (ZION)
- 1100 RELAYS
- 7 RELAY TYPES l
l
_ _ _ a
I t
O J
NUCLEAR REGULATORY COMMISSION MAINTENANCE AND SURVEILLANCE PROGRAM i
O .
PROGRAM MANAGER DR, HAROLD R. B00HER, CHIEF LICENSEE QUALIFICATIONS BRANCH DIVISION OF HUMAN FACTORS SAFETY /NRR l
GREGORY C. CWALINA, SECTION LEADER MAINTENANCE / SURVEILLANCE SECTION O
r f l 4
i i !
i l, ,
f f
4 1
4 4
4 i
4 OUTLINE )
i I
I OBJECTIVES & SCOPE i
l i CURRENT STATUS .
f
~ '
PROGRAM
SUMMARY
i@
PHAS5 I PROJECTS t
SURVEY PROGRESS i
t I
i i
i l
1 c
I i~
r 3 MAINTENANCE PROGRAM DEVELOPMENT U
NOV 83 NRC MAINTENANCE WORKSHOP MAINTENANCE INDICATOR PIL0T STUDY INITIATED JAN 84 COMMISSION POLICY AND PLANNING GUIDANCE MAY 84 US/ JAPANESE STUDY PART I COMPLETED ACRS BRIEFED (MAINTENANCE SUBCOMMITTEE AND FULL COMMITTEE JUN 84 DRAFT PLAN TO NUMARC ASME BRIEFED PUC BRIEFED JUL 84 IEEE BRIEFED SEP 84 ANSI BRIEFED
[)
OCT 84 UPDATED PLAN TO NUMARC ,
ACRS - MAINTENANCE SUBCOMMITTEE - JAPANESE STUDY BRIEFING DEC 84 MAINTENANCE INDICATOR PILOT STUDY COMPLETED JAN 85 PLAN PHASE I APPROVED NUMARC PROPOSED INDICATORS NRC INDICATOR TASK FORCE FORMED MAR 85 REVISED MSPP (NUMARC COMMENTS)
APR 85 SUBMITTED TO COMMISSION (SECY-85-129)
SURVEY PROJECT INITIATED MAY 85 JUN 85 COMPLETED REGIONAL BRIEFING / COORDINATION ACRS - MAINTENANCE SUBCOMMITTEE 1
( .
NRC SPECIFIC OBJECTIVES DETERMINE EFFECTIVENESS OF CURRENT MAINTENANCE PROGRAMS IDENTIFY PRACTICES WHICH REDUCE HUMAN ERROR RATE IN PERFORMANCE OF MAINTENANCE IMPROVE EFFECTIVENESS OF MAINTENANCE PROGRAMS IN ASSURING
())) OPERABILITY OF SAFETY SYSTEMS REDUCE UNNECESSARY AND UNANTICIPATED RADIOLOGICAL EXPOSURE TO MAINTENANCE PERSONNEL -
. DETERMINE REGULATORY APPRGACH TO ASSURE EFFECTIVE MAINTENANCE PERFORMANCE
___ __ _._~ _ ,._- , ,-.- .,__ ,--- __ _ . __ .. , - _ , _ - . _ _ _ . - _ _ . .
MSPP SCOPE ALL ASPECTS REQUIRED TO CARRY OUT A SYSTEMATIC MAINTENANCE PROGRAM SURVEILLANCE AND TEST ACTIVITIES EQUIPMENT REMOVAL FROM/ RETURN TO SERVICE POST-MAINTENANCE TESTING
~
- MAINTENANCE MANAGEMENT / ADMIN, CONTROL
- PERSONNEL SELECTION, QUALIFICATIONS, TRAINING
.O -
PROCEDURES DOCUMENTATION THOSE COMPONENTS WHICH AFFECT PERFORMANCE OF SAFETY SYSTEMS O
l
1 IDENTIFIED PROBLEMS ,
2 3
4 a
l
. 1. MAINTENANCE PERFORMANCE :
- 2. FAILURES DUE TO IMPROPER PERFORMANCE i
- 3. MAINTENANCE /0PERATIONS INTERFACE O -
- 4. CHALLENGES TO SAFETY SYSTEMS d
- 5. OCCUPATIONAL EXPOSURES
! 4 l
i O
I
t i
i f
IO i
1 i
STRATEGY I
i i
i BROAD SCOPE l FOCUS ON TECHNICAL ISSUES 1
l USE PHASED APPROACH i INTEGRATE STAFF ACTIVITIES l
! COORDINATE INDUSTRY INITIATIVES i
l I
.I i
i l
l O -
l l
.y FIGURE 3.1 Pe0ERm PHASES PROGam 3 .
I INITIATIONJ % TEAR 1 YEAR 2 . TEAR 3 i
Iqu4RTER 1 2 3 4 1 2 3 4 1 2 3 4 Phase ! Survey and Evaluation A A ,
1 EDO Review ^^
Phase II Identification of Probles and Ispacts lA .A A, no nevis. l An .AA
- I i
! i I 8
- a try m ru l l 1-r . gic , A A ,A A svama ms acurs n I ,
i ,
1i gayo A A A A mi A i i i O
l
PHASE I PROJECTS i
- 1. SURVEY OF CURRENT PRACTICES
- 2. MAINTENANCE PERFORMANCE INDICATORS
- 3. MONITOR INDUSTRY ACTIVITIES
- 4. PARTICIPATE IN STANDARDS GROUP
- 5. PROGRAM INTEGRATION (NRC AND INDUSTRY)
- 6. ANALYSIS OF JAPANESE /U.S. MAINTENANCE PROGRAMS
- 7. MAINTE' NANCE PERSONNEL QUALIFICATIONS
- 8. H.F. IN IN-SERVICE INSPECTION
- 9. HUMAN ERROR IN EVENTS INVOLVING WRONG UNIT OR WRONG TRAIN (GENERIC ISSUE 102) i O
% __. 4 3 - a J. - -
- * - _ _ _ma _ _ - _
SURVEY OF CURRENT MAINTENANCE PRACTICES
[]) ,
TASK 1 -
DATA COLLECTION AND ASSESSMENT l l
TASK 2 -
SALEM PREVENTIVE MAINTENANCE PROGRAM TASK 3 -
QUESTIONNAIRE TASK 4 -
MAINTENANCE REVIEW PROTOCOL i
TASK 5 -
SITE SURVEYS O .
TASK 6 -
SUMMARY
REPORT i
i t
O O O
SUMMARY
APPROACH AND SCHEDULE TASK 3 ,
Develop i&E MAY 15.1905 ouestionnaire Administer Anahre Data Prepare MARCH 15,1906
- goe ,-.y, R ,p.re TASK 1 Estab Data Base Perform Cemparative UPdate Data Base f tre )1 L
Structure Anahsis t TASK 2 TASK 6
[ stab Maint identiN Data Celket Data Arrany Visit Sakm pr, par, Taxenemy -
feeeds on Salem PN and Assess" j p,,,, ,
Salem $tte - -
Report gn,g/CR ~
Visit PM TASK 4 Develop Maint Conduct Pilot Complete Review - Site Visit -
Protoeel Re w Preteoe1 Protoce:
START TASK 5 Determire Arrange Site Conduet Site Correlete Site Plants le Visit - Visits Visits Visit Reperis 1
() '
- 2. MAINTENANCE PERFORMANCE INDICATORS SCOPE: -
- MONITOR INDUSTRY INDICATORS
- DEVELOP NRC INDICATORS IF WARRANTED MILESTONES:
PILOT STUDY, PNL, JAN - APR 1984 ENDORSEMENT OF NUMARC INDICATORS, MAY 1985
({]) .
VERIFICATION OF INDUSTRY DATA DECISION ON NEED FOR NRC INDICATORS i
l
- i l
NUMARC MAINTENANCE PERFORMANCE INDICATORS I. DATA AVAILABLE THROUGH EXISTING SOURCES
- 1. UNIT FORCED OUTAGE RATE
- 2. UNIT EQUIVALENT AVAILABILITY
- 3. N0. OF UNPLANNED AUTOMATIC SCRAMS DUE TO MAINTENANCE II. DATA TO BE OBTAINED FROM UTILITIES
- 1. TOTAL RADIATION RATE PER UNIT DUE TO MAINTENANCE O 2.. OVERTIME WORKED BY MAINTENANCE PERSONNEL
- 3. LOST TIME ACCIDENT RATE NO. FOR MAINTENANCE PERSONNEL
- 4. AMOUNT OF OUTSTANDING NON-0UTAGE CM WORK
- 5. RATIO 0F HIGHEST PRIORITY NON-0UTAGE CM WORK
- REQUESTS
- 6. PM ITEMS OVERDUE
- 1. PERCENTAGE OF MAINTENANCE REWORK
- 2. SAFETY SYSTEM AVAILABILITY O
- 3. MONITOR INDUSTRY ACTIVITIES SCOPE:
SUMMARIZE FINDINGS ASSESS APPLICABILITY DOCUMENT ACHIEVEMENTS MILESTONES:
- COORDINATION WITH NUMARC JUNE, OCTOBER 19814 JAN, FEBRUARY 1985 ATTEND EPRI SEMINARS, MARCH, APRIL 1985 REVIEWEPRI/M}TREPORT, JUNE 1985 O
- I 1
0 -
- 4. PARTICIPATE IN STANDARD GROUPS SCOPE:
ENC 0URAGE INDUSTRY INITIATIVES PROVIDE.NRC CONTRIBUTIONS MILESTONES:
ANS 3.9 SEPTEMBER 1984 - SUBCOMMITTEE FORMED Q ,
MARCH 1985 -
OUTLINE CIRCULATED FALL 1985 -
NUMARC DRAFT DUE ASME OPERATIONS AND MAINTENANCE COMMITTEE 4 A'PRIL 1985 -
NO PM STANDARD SPRING 1985 -
SEMINARS COMPONENT STANDARDS DEVELOPMENT CONTINUES IEEE W.G. 3.3 -
" PRACTICES" OCTOBER 1984 -
W.G. FORMED Q SPRING 1985 -
SCOPE DEFINITION
.O f
< 5. PROGRAM INTEGRATION SCOPE:
IDENTIFY RELATED PROGRAF 1S INTERPRET CONTENTS AND SCHEDULES PREVENT OVERLAP <
' PRESENT UNIFIED POSTURE MILESTONES:
O PROGRESS REPORTS, FINAL REPORTS ARE BEING
. REVIEWED i
l I,
- O
()
NRC PROGRAMS RELATED TO MAINTENANCE RESPONSIBLE ORGANIZATION QUALITY ASSURANCE PROGRAM, R.G. 1.33 IE SYSTEMS IMPORTANT TO SAFETY IE SAFETY IMPLICATIONS OF CONTROL SYSTEMS (USI A-47) NRR COMPREHENSIVE REEVALUATION OF STANDARD TECHNICAL SPECIFICATIONS RES rs SURVEILLANCE AND TEST REQUIREMENTS
\' > (ECCS OUTAGE CRITERIA) RES ,
NUCLEAR PLANT AGING RESEARCH RES EFFECTIVENESS OF INDUSTRY ALARA PROGRAMS NRR EQUIPMENT QUALIFICATION - R.G. 1.89 NRR RELIABILITY RESEARCH RES IMPROVING 00ALITY IE TRAINING RULE - SECTION 306 WASTE ACT NRR O
i .
- 1 l
- 6. ANALYSIS OF JAPANESE /U.S. MAINTENANCE PROGRAMS a
SCOPE:
1 I COMPARE OPERATING EXPERIENCE COMPARE MAINTENANCE REQUIREMENTS ANALYZE ORGANIZATION AND MANAGEMENT
^
- MILESTONES
- APRIL 198f4' -
PROGRAM INITIATION MAY 1985 -
PROJECT COMPLETED O
- NUREG/CR-3883 AND 3883P JULY 1985 -
PUBLISHED 4
1 l0 f
y _ c. ..-
_---.-_,.-_,.____.___.___..__.._,__..,__.-,,-_,,_..._.__,-._,__.__,_..,-..m,__,,,_._ _ . . , . , - _ . - . - , _
O .
t 7, MAINTENANCE PERSONNEL QUALIFICATIONS NEED TO DETERMINE THE KSAs REQUIRED FOR MAINTENANCE JOB TASKS Q
T0 IDENTIFY THE RELEVANT SOURCES OF THE KSAs IN TERMS OF EDUCATION, TRAINING, AND APPRENTICESHIP PROGRAMS TO CONDUCT AN ANALYSIS OF APPLICABLE INDUSTRY GUIDELINES AND STANDARDS AGAINST JOB RELEVANT KSAs AND SOURCES l
i O
i i
i
- 0 -
]
- 8. HUMAN FACTORS OF IN-SERVICE INSPECTIONS SCOPE:
1 i
IDENTIFY HUMAN ERROR POTENTIAL ;
A MILESTONES:
,f MAY 1985 -
PROJECT INITIATION l
1
! JULY 1985 -
DATA COLLECTION l
4 DECEMBER 1985 -
FINAL REPORT l -
!O 1
i l
i 1
1
!O
() '
- 9. HUMAN ERROR IN WRONG UNIT / WRONG TRAIN EVENTS SCOPE:
I IDENTIFY PROBLEMS ANALYZE. ROOT CAUSES DEFINE ACTIONS FOR RESOLUTIONS l
MILESTONES:
- JANUARY 1984 -
AE0D REPORT r
MAY-JULY 1984 -
AEOD/NRR C0 ORDINATION
[ ]) ,
l JUNE-DECEMBER 1985 -
SITE VISITS
! JANUARY 1986 -
FINAL REPORT l
i l
l l
lO l
1 PROGRESS
() NRC MAINTENANCE SECTION FORMED TASK FORCE REVIEW NUMARC DRAFT INDICATORS PARTICIPATION NEW STANDARDS EFFORTS REVIEW JAPANESE MAINTENANCE /0PERATIONS EXPERIENCE PARTICIPATION IE INSPECTION FEEDBACK NRC REGIONAL / RESIDENT INSPECTORS G ~ INDUSTRY STANDARDS EFFORT INITIATED (IEEE, ANS) ANALYZE MAINTENANCE-RELATED CONTRIBUTION TO EVENTS REVIEW / REVISE INP0 PLANT EVALUATION OBJECTIVES / CRITERIA IN MAINTENANCE AREA DEVELOP PRELIMINARY PERFORMANCE INDICATORS REVIEW INP0 DRAFT MAINTENANCE GUIDELINES EVALUATE STATE OF MAINTENANCE IN INDUSTRY O
- - - - + - ,, . - - -,. . -, , , ..--..--m -., . , -_ ------ -. .---- _-_.-.- . -,-- - ---+ * --}}