ML20054M089
| ML20054M089 | |
| Person / Time | |
|---|---|
| Site: | Shoreham File:Long Island Lighting Company icon.png |
| Issue date: | 07/07/1982 |
| From: | Atomic Safety and Licensing Board Panel |
| To: | |
| References | |
| ISSUANCES-OL, NUDOCS 8207090236 | |
| Download: ML20054M089 (214) | |
Text
- ._ ____ _ _ _ __ _ _ _ _ _ _ _ __
NUCLEAR REGULATORY COMMISSICN O mWW A R E IKL, Ym
,O 8trOaz raz arOMIC SarsrY.aND tICzNSING eOnaD s
In the Mattar cf: :
LONG ISLAND LIGHTING COMPANY :
- DOCKET NO. 50-322-OL (Shoreham- Nuclear Power Station) :
O CAar: July 7, 1982 PAGIs: 5833 - 6033 AT: Riverhead, New York
,e c l r A tWA be;-- L fot<YS D' W""
yn Epv
.fkal
.unensox /' armnxa O -
F- Q 40 0 vi_ginia Ave. , S .W. Wa shd .g .nn , D . C. 20024 C Talaphene: (202) 554-2345 m :f & v.scnv
5833
() 1 UNITED STATES OF AMERICA 2 NUCLEAR REGULATORY COMMISSION 3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4 - - - - - - - - - - - - - - - - -x 5 In the Matter of 6 LONG ISLAND LIGHTING COMPANY s Docket No., 50-32 2- OL 7 (Shoreham Nuclear Power Station) a 8 -----------------x 9
to Riverhead Town Hall 11 200 Howell Avenue 12 Riverhead, New York 11901 13 . Wednesday, July 7, 1982 l 14 The hearing in the above-entitled matter l
15 convened, pursuant to notice, at 9:00 a.m.
16 BEFORE:
17 LAWRENCE BRENNER, Chairman 18 Administrative Judge 19 JAMES CARPENTER, Nember 20 Administrative Judge 21 PETER A. MORRIS, Member 22 Administrative Judge 23 WALTER H. JORDAN, Assistant to the Board
() 24 Administrative Judge 25 ALDERSON REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
i 5834 1 APPEARANCES 2 On behalf of Applicants 3 ANTHONY F. EARLEY, Esq.
O 4 W. TAYLOR REVELEY III, Esq.
5 T. S. ELLIS III, Esq.
6 Hunton & Wil11aas ,
7 707 East Nain Street 8 Richmond, Ya . 23212 9 On behalf of the Regulatory Staffs to EDVIN REIS, Esq.
11 RICHARD RAWSON, Esq.
12 Washington, D.C.
13 On behalf of Suff olk Countys 14 LAWRENCE COE LANPHER, Esq.
15 KARLA J. LETSCHE, Esq.
16 Kirkpatrick, Lockhart, Hill, 17 Christopher & Phillips 18 1900 N Street, N*.W.
19 Washington, D.C. 20036 20 21 22 23 O 24 25 O
ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554-2345
5835 O coxIsxIs 2 WITNESSES: DIRECT CROSS REDIRECT RECROSS BOARD 3
Vogin Joksimovich, 4 Edward T. Burns and Robert M. Kascsak (Resumed) 5 By Ms. Letsche 5843 By Mr. Rawson 5905 6
. (Af ternoon Session. . . . . . . . . . . .page 5932)
Vogin Joksimovich, 8
Edward T. Burns and 9 Robert M. Kascsak (Resumed)
By Mr. Ellis 5933 to By Judge Morris 6012 By Judge Jordan 6016 11 By Judge Morris 6020
~
E E E I E I I. E 13 l
p v IDENTIFIED RECEIVED BOUND IN TRANSCRIPT 14 NUMBER 15 Suffolk County 22 5903 -
5903 16 Suffolk County 23 5903 5903
' Suffolk County 24 5903 5903 18 i Suffolk County 25 5903 5903
. 19 l Suffolk County 26 5903 5903 r 20 LILCO 14 5988 5989 5989 21 22 i
RECESSES:
23 Morning - 5894 O 24 Noon - 5931 Afternoon - 5966 25 i
ALDERSON REPORTING COMPANY,INC, l
400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5836 O ' za s s I n 1 " a s.
2 ( 9:00 a.m.)
3 JUDGE BRENNER: Good morning.
4 I don't know if Mr. Lanpher is here. He said 5 something off the record that I wanted to put on the 6 record this morning. I can do it even without him.
7 MS. LETSCHEa Judge Brenner, he is in a 8 meeting. He should be here shortly. He is not here 9 right now.
10 JUDGE BRENNER: Well, I can just go ahead and 11 do it. In open session but off the record yesterday, 12 Mr. Lanpher informed us that indeed with respect to 13 emergency planning we can consider the filing of 14 contentions received yesterday from the County as 15 including and superseding that separate filing that we 16 did not wish to deal with, not because it wasn't worthy, 17 but because we didn't want to deal.with separate filings l 18 from Mr. Shapiro and on behalf of his clients.
19 Related to that subject, we discussed two l
20 filings from the staff at the beginning of next week. '
l 21 One will ba the sta f f 's position on the recently filed 22 emergency planning contentions, and the other is going I
23 to be the staff's further report with respect to the 24 involvement of SAI in work for the staff and LILCO with 25 respect to systems interaction. The reason I mention it O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5837
() 1 again is to clarify the filing dates. We requested that 2 each of those be filed Honday morning, but only require .
- 3 that the emergency planning filing be filed Monday 4 morning. The requirement for the SAI filing is Tuesday 5 at the hearing, although we pref er it Monday morning.
6 The reason I emphas,1ze that is because of Judge 7 Carpenter's schedule, it is imperative that we receive 8 the emergency planning filing first thing Monday 9 morning.
10 Also, in considering my remarks yesterday as 11 to the staff's filing on the situation with regard to 12 SAI, I think I was remiss in not emphasizing that we 13 certainly ild appreciate the staff 's filing and also 14 appreciated the fact that the staff made a preliminary 15 filing to apprise us as soon as possible of what th e 16 staff did know, and we certainly prefer that to a 17 blanket request for an extension of time while the staff 18 was gathering up the other facts, and we certainly had 19 no problem with the staff waiting a week to collect the 20 other information so as to give us a full report. Our 21 concern was from the written filing it appeared the time 22 period was open-ended, and that concern was quickly 23 alleviated when the staff gave us the filing date.
() 24 The f'arther comments were to take advantage of 25 the fact that the staff was going to make another filing O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, O.C. 20024 (202) 554-2345
5838
() 1 so that we could have that filing better focused on what 2 ve were interested in. And I had some concern that -
3 perhaps the staff would be spending the bulk of the time 4 gathering up information that would be peripheral to S what our main interest is, that is, a long, lengthy list 6 of contrac,ts as opposed to direct answers to the 7 questions that we reiterated yesterday.
8 So it was in that spirit, and we did 9 appreciate certainly receiving the preliminary filing to which was helpful as a preliminary filing.
11 I guess I should add that it is a fact of life 12 in hearings that you usually don't hear when things are 13 fine; you only get comments as to the further O 14 information we wan t, and requests for further 15 information should not be taken as criticism of the 16 information that has been provided unless we expressly 17 so state.
18 (General laughter.)
19 JUDGE BRENNER4 We have received yesterday 20 Suffolk County's response to LILCO's actions to 21 supplement witness panels, plus several related 22 matters.
23 I guess I don't frankly recall what the
() 24 related matters are. As I read the filing yesterday, i:
25 is focused on the notions to supplement witness panels, O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
i 5839
() 1 is that not correct, Ms. Letsche?
2 MS. LETSCHE: I think that is right, Judge 3 Brenner. I think the title there is a reference to the 4 title of the LILCO pleading. Frankly, I am not greatly 5 familiar with our filing.
6 ,
JUDGE BRENNER: You are correct. I missed the 7 quotation marks in reading it, and indeed, the LILCO 8 pleading did have related matters, including the 9 notification of additional testimony as to one matter.
10 Well, we are going to overrule the objections 11 of the County and grant LILCO's action to supplement 12 witness panels. The written objections of the County 13 appear to us to be just a rephrasing of the general 14 objections previously heard. We think the 15 supplementation, for reasons we have already discussed 16 several times, will be very helpful to the record, to 17 the Board, and should be to the County, too, if they are 18 interested in getting answers to questions as opposed to 19 just asking questions that can't be answered by the 20 witnesses present. Although the written answer alleges 21 prejudice, we don't see any basis for prejudice to the 22 County.
23 In addition, as we emphasized several times,
() 24 the County has always been free to supplement its 25 witness panels. It continues to be free to do so, O
ALDERSON REPoHTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5840
() I although by now we had hoped to catch up so that we 2 could get reasonable advance notice of a week as a 3 ressonable yardstick when a panel would be 4 supplemented.
5 We won't enforce that for the rest of this
, 6 month, so that the County could add witnesses on 7 rela tia vely short notice, although certainly some 8 notice, subject to the particular adjustments, if 9 requested and if we agree they are necessary, similar to 10 the adjustment we made for the County with respect to 11 Dr. Lucks. However, after this month, after July, 12 absent the particular circumstances as to a particular 13 request to add a witness, we will require one week's O 14 notice by all parties to add a witness.
I 15 The adjustment, as you will recall, with 16 respect to Dr. Lucks was that the County had an 17 opportunity to indicate whether they needed further time
! 18 to ask further questions beyond the initial question.
19 That is not the only possible adjustment, but it is j 20 one. .
21 MS. LETSCHE: Judge Brenner, as I indicated 22 earlier, Mr. Lanpher is in a meeting this morning, and 23 he is more familiar with tr.at opposition than I an. I l
() 24 was frankly hoping that if I had known you were l
25 immediately going to rule on this, I would have l
ALDERSoN REPORTING COMPANY,INC, l 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5841
() 1 indicated to you tha t we would prefer to have it dealt 2 with when Mr. Lanpher was.present so that he could 3 respond to any concerns that you have.
4 I think there might be some cla rifications 5 that he would want with respect to the Board's ruling, 6 and I would request that when he is available, -which 7 should be later today, that he have an opportunity to 8 address that if he feels it is necessary to do so.
9 JUDGE BRENNER: I as certainly generally aware 10 that he is the counsel for the county that dealt' with 11 this matter, and if we were going to have argument on 12 i t, I would have given you advance notice. But I didn't 13 envision any argument, and that is why I just went ahead 14 and announced our ruling. Certainly he can come back 15 and seek any clarification he wants, and it doesn't have 16 to be today. It aight be more efficient for him to take 17 a look at the transcript and come back later this week.
18 But as you wish.
19 (Board conferring) 20 JUDGE BRENNER: We have nothing else this 21 morning.
22 I should also let you know that as f ar as we 23 are concerned we have no pending matters before us other
() 24 than the emergency planning contentions which we are 25 going to work on and for which we will receive some O
ALDERSON REPORTING COMPANY,INC, ,
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
i 5042
() 1 further filings as indicated yesterday. I make that 2 announcement in case we lost something in the flurry of 3 papers.
4 Does anybody have something pending before 5 us?
6 HR. REVELEY: We hd some actions to strike, 7 Judge, but those are the only things I'm familiar with 8 othat than the energency planning.
9 JUDGE BRENNER: Yes, thank you, and I should 10 have added pending before us for which a ruling has now 11 arrived, and we scheduled responses for those notions to 12 strike, but I did forget about them momentarily just 13 now.
14 Are there any other matters?
15 (No response.)
16 JUDGE BRENNERa Good.
17 We are ready to continue with the cross 18 elanination. We have your estimate in mind at the end l
19 of the day. I hope you do.
20 Whereupon, -
21 V0 GIN JOKSIMOVICH, 22 EDWARD T. BURNS and 23 ROBERT H. TASCSAK,
() 24 called as witnesses by counsel for Long Island Lighting 25 Company, having been sworn previously, resumed the stand O
ALDERSoN REPORTING COMPANY,INC, ,
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5843 and were examined and testified further as follows:
(]) 1 2 .
CROSS EXAMINATION -- Resumed I 3 BY MS. LETSCHE:
4 0 Mr. Kas:ssk, yesterday Dr. Burns stated in 5 response to a question from the Board that the Shoreham 6 PRA contained information identifying systems 7 interactions. Aside from your review of the accident 8 frequency data contained in the Shoreham PRA, has LILCO 9 performed any separate analysis of the systems 10 interaction identified in the Shoreham PRA?
11 A (WITNESS KASCSAK) We have performed a review 12 of the products of the analysis, of the PRA analysis 13 which shows the sequences that lead to the fault trees 14 and event trees that produce the sequences that we i 15 talked about yesterday. Embodied within those f ault 16 trees and event tries are the dependencies and the 17 modeling of the system dependencies that Dr. Burns had 18 talked about yesterday. We do review those results and I
19 see if that information is accurate and whether any 20 particular action is required as a result of those 21 results.
22 Ihere are a number of activities pending as a 23 result of the studies, and our review process is just
() 24 beginning because the draf t has been in our hands some 25 time now, but we are still in the process of reviewing
()
ALDERSoN REPORTING COMPANY,IN,C, 400 VIRGINIA AVE., S.W WASHINGTON, D.C. 20024 (202) 554-2345
5844 1
i
(} 1 it and discussing it with the peer review group.
2 0 Can you describe for us what the analysis of 3 the systems interactions identified in the PRA is? I O 4 mean, you said you do a review of something. Can you be I
5 a little more specific as to what that analysis is with '
6 respect to systems interactions?
7 A (WITNESS KASCSAK) The process itself through 8 the logic trees, the fault trees and event trees, depict 9 dependencies, the intra- and intersystem dependencies, 10 and therefore the intra- and intersystem interactions 11 are depected on these trees and are shown, presented in 12 a logic format on the trees. The process also produces 13 through the computer processing of the data that is on 14 the trees what sequences affect both system reliability 15 and what sequences affect certain event sequences that 16 lead to failure or core vulnerable situations.
17 It is those sequences that contribute in the 18 highest extent to the public risk, and it is those 19 higher order sequences that are looked at, and we would 20 look at the sequences that produce the largest 21 con trib ution to system unreliability and the event 22 sequences that lead to the events that prod uce the core 23 vulnerable situations and try to evaluate whether or not l
() 24 those interactions that are produced through those 25 sequences are representative of what we would expect and O
ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
.I 5845 O i nether or not ener re any unique sequences that are
- 2 produced particular to the Shoreham plant, and then we j i
3 would consider evaluating what might need to be done to I '4
',. reduce those probsbilities of those events occurring.
5 i 6 7
8 9
10 .
11 12 13 O 14 15 16 17 18 19 20 ,
21 22 23 24 25 O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
. _ . ~ _ - _ . _ .
5846 0 Am I correct, then, that your analysis of
(]) 1 2 system interactions identified in the Shoreham PRA is 3 limited to a review of the accident sequences that have l 4 a high probability of risk?
5 A (WITNESS KASCSAK) It is a review of the fault 6 trees and the event trees. The fault trees depict the 7 system reliabilities and how you lead to -- how certain 8 sequences affect the reliability of the system. And you 9 can look at what dominant sequences or dominant 10 dependencies or interactions within that system- or 11 support systems to that systen affect the reliability of 12 that system.
13 And we talked about functional dependencies, 14 which show how redundant components or support systems 15 contribute to the functioning of that system and how 16 those f ailures would affect the reliability of that 17 system.
18 We also have the contributions of spatial 19 interactions, like ventilation systems or other systems 20 Lik e th a t that may interact with that system and affect 21 its reliability. We have the human aspects, the human 22 errors we talked about that are modeled in the fault 23 trees; the maintenance callibration considerations that
() 24 are built into the fault trees.
25 And we look to see how those dependencies
()
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5847
() 1 affect the system reliability and which ones dominate 2 that situation in terms of producing a system that ,
3 fails. There are literally hundreds of sequences in 4 each fault tree and probably thousands of sequences in 5 each event tree. All of these are interactions, and you 6 just don't look st every one because many of these 7 interactions are perfectly acceptable.
8 You look at the interactions that produce 9 unacceptable results, and you try to concentrate on 10 those that dominate your effect on system reliability or 11 affect situations that would lead to a core vulnerable 12 situation.
13 Q Let me try to understand. Do you kind of look .
14 at the results which identify dominant contributors or 15 dominant sequences, and then look at the fault trees 16 related to those for interactions?
17 A (WITNESS KASCSAK) Well, you keep on using the 18 word "d o m in a n t . " It isn't only necessarily dominant 19 sequences we look at. We look at all of them. There 20 are for each class, there are many sequences. Even 21 though we might call that a dominant class, there are 22 many, many sequences that contribute to that overall 23 core vulnerable sta te f rom that class. And you look at
() 24 all of the sequences that have the larger contributions 25 to that particular class.
O
, ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5848
() 1 Q Mr. Kascsak, I'm using the word " dominant" 2 because th a t is the word that you keep using. You said 3 s minute ago that you don't look at all of the 4 interactions because there are lots of them depicted on 5 these event trees and fault trees, and I'm trying to 6 understand what it is you actually do look at, which of 7 the interactions you do look at, how do you decide which 8 ones you 're going to look st.
9 Can you identify for me which ones it is you to look at in this review? -
11 A (WITNESS KASCSAK) Well, as I say, all of 12 these sequences are produced, and there is a term called 13 " cut sets" which looks at in an individual sense each .
14 one of the sequences. And those cut sets are available 15 to be looked at to evaluate how particular sequences 16 contribute to the failure of the system or to the 17 failure of an event that leads to a core vulnerable 18 situation.
19 0 Have you looked at all of those cut sets?
. 20 A (WITNESS KASCSAK) No, we haven't looked at 21 all of then yet. We have looked at some of them. We 22 have looked at the f ault trees and the event trees and 23 the dats that went into the fault trees and even trees.
() 24 We have not completed our review of the information that 25 has been produced by this analysis.
O -
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5849 What you have completed, what you have done so
(]) 1 Q 2 far, has that review been documented, the LILCO review?
3 A (WITNESS KASCSAK) Yes, there has been some 4 documentation. Any comments or information that we do 5 pass on to SAI or discuss with them is documented.
6 0 Has your -- what conclusions has your review 7 thus far reached with respect to your systems 8 interactions identified in the Shoreham PRA?
9 A (WITNESS KASCSAK) Well, there are a number of-10 things that came up-during the review process, two items 11 in particular that we decided to make a design change as 12 a result of our review of the analysis. There are a 13 number of other things pending and maybe more. things 14 pending as we continue our review, that have initiators, 15 to go back and look at things in some more detail, to 16 possibly consider design changes that could affect 17 system reliability or the reduction in core vulnerable 18 frequency.
1 19 0 What are the two design changes?
20 A (WITNESS KASCSAK) The two design changes --
21 one design change was a review of the containment 22 failure situation in looking at the ultimate capability 23 of the containment, and in that review we determined
() 24 that there was a situation where, on the equipment hatch 25 and the personnel hatch, there were viewing windows on ,
1 1 :
, ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5850
(} 1 those hatches that were susceptible to failures at lower 2 pressures than most of the other structures in the 3 containment.
O 4 So in order to bring the ultimate strength of 5 the containment up to an equal level and use that 6 information in the PBA analysis, we made a modification 7 to those viewing ports within the equipment hatch and 8 personnel hatch.
9 These changes are well beyond the normal 10 design basis of the plant. They are looking at the 11 ultimate capability of the containment and also looking 12 at the system fault tree and the event sequences dealing 13 with the use of the high pressure co,olant systems, and 14 particularly how RCIC could come into play in certain 15 sequences.
16 We made a modification to the trip setpoint on 17 the RCIC turbine to allow the turbine to operate during 18 a LOCA. The initial trip setpoint was at about 25 psig 19 and we made a modification to allow it to operate at 50 20 psig. And .we had to review the system perf ormance and 21 ensure that that could operate properly at those 22 conditions. So that modification has been made.
23 We have also two current studies under way,
() 24 one dealing with the HSIY closure. Presently the HSIV 25 closure signal opera tes at level two and we have a study O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5851
(} 1 under way looking at changing that to level one. This 2 would allow the plant to -- it helps you during certain 3 overpressurization transients. It allows the normal O 4 condenser heat sink to be available for more of the 5 sequences and a longer period of time. We are 6 performing an analysis now to see if, from an accident 7 analysis point of view, that present: no problems.
8 The other evaluation that is under way is the 9 consideration of adding a black start capability to the 10 7as turbines that are presently located at the Shoreham 11 site.
12 JUDGE JORDANS That was what, black start?
13 WITNESS KASCSAKs. Black start, meaning the gas 14 turbines could ba started directly at the site rather 15 than requiring auxiliary power offsite to start the gas 16 turbine.
17 This is a 50-megawatt gas turbine and its 18 capacity is such that it can supply sufficient power to 19 operate all of the ECCS equipment during an accident 20 situation. The analysis showed us that the loss of 21 offsite power event and the interactions with the l
22 offsite power, although generally within the bounds of 23 other contemporary analysis, there might be some room
() 24 for improvement if we added to the gas turbine the 25 ability to start the gas turbine from onsite.
l I
l ALDERSoN REPORTING COMPANY,INC, 400 VIRGINfA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5852 l 1 So we are looking at that capability and
[}
2 seeing how, it we add that capability, how it would 3 affect the analysis, the ability to improve the O 4 reliability of the plant, the ability to improve or 5 lower the frequency of core vulnerable situations as a 6 result of that modification.
7 And here is an interaction where we are 8 talking about offsite power, supplying power to run the 9 plant in a positive way, to reduce the frequency of 10 having to rely on onsite power or restoring offsite 11 power quicker in the event -- than might normally be the 12 case.
13 And as I say, these are examples of what our 14 analysis has produced thus far, and we will continue to 15 look at these things and we will continue to use the PRA 16 as a tool to investigate plant modifications or events 17 when they occur in the plant, to see if they are unusual 18 events and may lead to maybe a precursor to a more 19 serious situation, and take action accordingly.
20 BY NS. LETSCHEs (Resuming) 21 0 These design changes you've mentioned are 22 designed to improve the reliability of the plant 23 overall, is that right?
() 24 A (WITNESS KASCSAK) Right, to improve the 25 overall system reliability or reduce the probability of O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5853
(} 1 an accident situation leading to a cora melt.
2 Q I'm not sure you answered my question, which 3 was, where have you documented in a review of systems O 4 interactions these design changes you told me about are 5 relevant changes. But can you tell me where you've 6 documented any review of systems interactions that may 7 have been identified in the Shoreham PRA?
8 A (WITNESS KASCSAK) There are literally 9 thousands and thousands of interactions that are 10 depicted through this analysis, and every time you pick 11 up an event tree or a fault tree you look at these 12 hundreds or thousands of interactions. Through the 13 review process you evaluate the dependencies and how 1
14 these interactions are coupled in such a way to produce 15 the overall effect of affecting the reliability of the 16 systen or the plant itself.
17 And by looking at how these systems interact 18 through these models and which of the sequences are 19 significant to the overall success or failure of the 20 system or the plant, you are in f act looking at system 21 interactions. And if we found what we consider to be 22 unacceptable interactions that were produced by these
- 23 models, we would take some action accordingly.
() 24 Now, it turns out that most of these things we 25 talked about are in some uay interactions and to a O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5854 l
1 limited extent are improvements to or will help
(
2 eliminate unacceptable interactions. To that extent, we s 3 haven't found any other what we consider to be
-] 4 unacceptable interactions, although there are many, many i 5 acceptable interactions that we . have looked at and have 6 concluded are acceptable.
7 0 Has LILCO's documentation of its review of 8 systems interactions identified in the Shoreham PRA --
9 is that documentation limited to the things you have 10 just testified about, these two design changes and the 11 two ongoing analyses.
12 A (WITNESS KASCSAK) I think the best way I can 13 . answer that is that these reviews are performed by
(
14 people who are familiar with the entire design of the 15 plant and are familiar with all of the other studies 16 that we have discussed before -- the design processes, 17 the deterministic studies. And they use all of that 18 other information that has already been produced in 19 concert with the PRA analysis to make their judgments.
20 So it is not as if they use the PRA in a 21 vacuum. They use it coupled with all of these other 22 pieces of information, which also depict how the plants 23 should respond, how systems should function as a unit.
() 24 And all that information is used together to make 25 judgments as to whether or not there is a consistency O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5855 1 and whether or not the plant is responding in the way it 2 is expected to be responding.
3 Q My question, Mr. Kascsak, is a simple one Is O 4 there any documentation of the systems interactions 5 review other than the things you have mentioned?
6 A (EITNESS KASCSAK) Other than what I have 7 mentioned, there is no other documentation.
8 JUDGE BRENNER I'm a little co6 fused by the 9 last exchange, and maybe I'm the only one. I don't knov 10 if that means that there is documentation of the items 11 you discussed or there is no documentation of any of it 12 other than your testimony here.
. 13 UITNESS KASCSAK4 Well, I think I indicated 14 previously that there is documentation of our review of 15 the products of the analysis, the event trees, the f a ult 16 trees, all of the data bases that went into the 17 analysis. And c bviously there is documentation 18 associated with any design change or study that's under 19 way.
20 Other than that, there is no other 21 documentation.
22 JUDGE JORDANa But you did say that that 23 included the systems interactions, which were in fact
() 24 literally thousands, you said. And that is documented.
25 So therefore I 'an also a little puzzled when you say O l 1
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5856
() 1 there was no documentation.
2 WITNESS KASCSAK4 No. She's asking me if 3 there's any additional documentation than what I 4 explained, and I'm saying that the review of the event 5 trees and the fault trees are in fact reviews of systems 6 interactions, because they do depict these 7 dependencias. Those reviews are documented.
8 Absent tha t documenta t' ion, there is no other 9 explicit piece of paper that talks about system 10 interactions studies. I think that is what the question 11 dealt with.
12 JUDGE BRENNER: I tell you, I've been sitting 13 here for many, many weeks and reading a lot of testimony O 14 in some frustration -- and I'm making this as a general 15 comment; it's not any criticism of any particular answer 16 --trying to grapple with how the PRA might be applied, 17 aside from the general overall use of the PRA in the 18 particular con text of this contention, how it might be 19 applied as a check or an initial inquiry or however you 20 vant to characterize it, to identify systems 21 interactions worthy of some further consideration or 22 study.
23 And I have heard a lot of generalities in
() 24 questions and answers, and enough blame for everyone to 25 share in terms of the inefficiencies, I think, these O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
I 5857 ;
() 1 last few weeks. And now I start hearing some specifics 2 that I hear for the first time, anyway, and maybe the 3 technical members and Judge Jordan and other Board 4 members heard things I didn 't, and when we go over the 5 record we will discuss it and go over that together 6 among ourselves.
7 That was a long prelude to my next inquiry, 8 Jhich would be, if I wa'nted to physically get my hands 9 on some sort of summary documentation ~of the 10 interactions, or other modifications for that matter, 11 that LILCO is considering making as a result of the 12 preliminary inquiry through the PR A combined with the 13 other techniques mentioned in the testimony, is there a 14 way I could do that? Does that exist?
15 WITNESS KASCSAK I think I attempted to 16 enumerate basically the ones that I have identified to 17 date, and those are documented. At this point there is 18 no other. There are review efforts under way and 19 possibly those reviews will produce other 20 considerations.
l 21 JUDGE BRENNERa I understand you're still 22 working on it, and certainly that was emphasized from 23 the very beginning of the testimony on this contention.
() 24 And I also understand that there is documentation of the 25 matters,you discussed. But there is documentation and i O l l
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5858 1 then there is documentation, and I'm trying to get a 2 handle on whether it would be necessary to pore through 3 many, many pages in diverging places or whether there 4 isn't some reasonably concise summary of potential 5 interactions being looked at as a result of the increase 6 so far.
7 (Pause.) I I
8 MR. ELLIS: Judge Brenner.
l 9 JUDGE BRENNER: Wait a minuta. Let me get an l 10 answer to the question, unless you know the answer.
11 MR. ELLIS I think I do.
12 JUDGE BRENNER: Well, let's let the witness 13 answer. Then I will swear you in.
14 ( La ug hter. )
15 WITNESS KASCSAKs One of the suggestions that 16 has been made to us previously within the peer review 17 group context would be to develop in a more concise way 18 how system interactions are addressed in the PRA and to 19 add a section to the report that deals with a discussion 20 of the dependencies and how they are modeled and how 21 they are in fact reviewed within the program.
22 And we have discussed this with SAI and they 23 are in fact in the process of producing such a piece of O 24 the report.
25 JUDGE BRENNER Let me interject, and I guess O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W, WASHINGTON. D.C. 20024 (202) 554-2345
5859 1 this is a comment for your counsel and not for you. I
)
2 would have expected that very thing to have been in the 3 written testimon'y and it is not.
O 4 But putting that aside -- and that information 5 was helpful, Mr. Kascsak, - but my particular question is 6 not going to a discussion of'the methodology at this 7 point, but rather whether there is some reasonably 8 concise compilation, whether it be one document or 9 several documents, that would* indicate these 10 modifications or changes that are being considered as a 11 result of the work so far on the PRA. That is, the ones 12 you 've testified about.
13 And what troubles se is you inserted the words 14 "for example" from time to time. So that is my 15 question. It is the documentation as to that; does that 16 exist 17 UITNESS KASCSAK4 Well, I would have to look
, 18 back, I think, and see what kind of documentation l
l 19 exists. The process is not all that formalized in the 20 sense of producing a given sequence enat is then 21 reviewed and checked. Although that is th0 process that 22 is performed, it is not performed in a manner that is 23 retrievable to the extent of documenting each individual sequence that is looked at.
(]) 24 25 There are pieces of paper that depict these t
O ALDERSoN BEPORTING COMPANY,INC, 400 V;RGINIA AVE, S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5860
() 1 sequences and we can show probably our documentation of 2 our review of those, but not the actual thought process 3 that went into determining whether or not we felt 4 anything further should be done as far as evaluating in 5 any more detail that particular sequence. That is a 6 judgment call and it is just not documented.
7 JUDGE BRENNER: Well, again, that is broader 8 than what I asked. That is all of the thought processes 9 tha t went 'into it. That might or might not be helpful, 10 but what I'm asking about now is not an application of 11 the thought processes, but you identified here this s2 morning for the first time to my recollection a number 13 of modifications that LILCO is considering, not that you
( 14 are making all of them.
15 WITNESS KASCSAKs Well, two that we have made 16 and two that we are considering.
17 JUDGE BRENNER: All right. And I would like 18 to know if I can pull together one, two, three or four 19 or that sort of number pieces of paper, whether it be an 20 exchange between LILCO and SAI or just pieces of paper 21 within LILCO or that type of thing, that would sta te, 22 these changes are going to be made, or these changes are 23 being looked at, and some tie to the PRA study as part l () 24 of the reason for thinking about those changes.
25 WITNESS KASCSAK4 I can certainly produce the O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
i 5861 ,
documentation on these four items that I've mentioned
(]) 1 2 and couple that to the review of the analysis that we 3 have. And I think I might even in fact have that with 4 se.
5 JUDGE JORDANS These items you mentioned, 6 these four items, came as a result of these analyses of 7 the various fault trees. Now, it isn't necessarily true 8 tha t it was a systems interaction that made this become 9 not a domin* ant but a significant contributo r. There are 10 other possible things that could have done it --
11 component failures and whatever, isn't that true?
12 You're not saying that it was entirely systems -
13 interactions that made this happen?
14 WITNESS KASCSAKs Well, in the broad sanse of 15 dependencies, there are interactions that come into play 16 here, and in many cases they are within the same system, 17 that's true. But there is a certain amount of 18 dependencies that exist and, for example, the RCIC 19 modification which allows that system to operat'e during 20 the LOCA allows that system to be added to the sequence 21 of systems that can respond to the accident sequences, 22 and in that respect it .s a positive system interaction 23 rather than maybe a negative system interaction.
() 24 JUDGE BRENNEB M r. Ellis, I didn't mean to 25 cut you off completely. I just didn't want to divert my O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5862
(} 1 own thought. And if you want to make any comment, you 2 are free to do so now.
3 MR. ELLIS: No, sir. I hope the Board got the O 4 information it was seeking. I think it did. I thought 5 I heard the question answered. I think it is helpful.
6 I accepted the Board 's comment concerning the comment 7 for counsel and I welcome that, because it is helpful 8 for us to know what it is that is on the Board's mind so 9 that we can address those issues. And I know we are 10 sometimes wide of the mark and we apologize for that, 11 and we will try ts do better. We appreciate the 12 guidance. -
13 JUDGE BRENNER: I am.just very surprised -- I
, 14 shouldn't say "we" -- that I think I heard this for the 15 first time this morning. And the expertise we have on 16 the Board any have discerned some of this earlier and 17 maybe I missed it. -
18 MR. ELLISa Well, I was sort of a newcomer to 19 the field myself. But I am bound to say I saw the 20 relationship earlier, and it just may be different l
21 perspectives on it, and we will do our best to clarify 22 in the remainder of the testimony.
23 JUDGE JORDANS Judge Brenner realizes I do
() 24 concur in his remarks.
25 JUDGE MORRIS Mr. Kascsak , to help my i
(
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 1
5863 understanding of what's going on, could you tell me when
(]) 1 2 the review of the PR A draf t began by LILCO employees?
3 WITNESS KASCSAK Well, the process is 4 ongoing. It commanced with the initial products, which 5 were the initial fault trees just depicting the 6 dependencies and showing how the plant was modeled.
7 JUDGE MORRIS: When was that?
8 WITNESS KASCSAK4 Around September or October 9 of last year.
10 JUDGE MORRIS 4 And about how many people are 11 involved?
12 WITNESS KASCSAK I would say there are 13 somewhere between four and five people involved on a 14 continuous, routinely continuous basis, and then there 15 are many other people, maybe twice that many, involved 16 on a part-time basis.
17 JUDGE 50RRIS: This was four or five that are 18 essen tially full-time reviewing?
19 WITNESS KASCSAK Essentially. They have 20 other responsibilities also, but they are in the review 21 process on a continuing basis.
22 JUDGE MORRIS: Which organization are ther 23 in?
() 24 WITNESS KASCSAK4 Host of the ones that are on 25 a full-time basis are within the nuclear engineering O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345 l
5864
(} 1 d ep a r tm en t . There are other reviews taking place by our 2 operations staff, not on quite as regular a basis, 3 although we are attempting to get more of that input.
O 4 It is somewhat difficult at times because of other 5 activities going on.
6 JUDGE MORRIS: What is the scope of this 7 review? Do you intend to review all of the fault trees 8 and event trees?
9 WITNESS KASCSAK: Yes. And we have --
10 JUDGE MORRISa And are there specific 11 directions as to what the people should be looking for 12 when they review it? -
13 WITNESS KASCSAKs Yes, particularly to ensure 14 the models are accurate and complete, initially; for 15 them to look at the data that is inputted into the 16 models as the second step. First the models are 17 produced without the data included, and then the next 18 step would be for the trees to be produced with data, 19 and then eventually the data is processed and you see
! 20 these individual sequences, the dominant sequences, the 1
21 cut sets that we talked about, to produce the higher 22 order of failure modes within the system and within the 23 trees, the event trees.
() 24 JUDGE MORRIS: Is.there activity by LILCO 25 employees to review the way in which minimum cut sets l
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5865
() 1 are determined?
2 WITNESS KASCSAK Well, early on we had 3 discussions with SAI on how that was handled within the 4 computer programs and what information would be given to 5 us.
6 JUDGE MORRIS: I'm looking for sort of a QA 7 function on what SAI is doing in lopping off the 8 branches of these big trees.
9 WITNESS KASCSAK4 Well, again, as I mentioned 10 before, obviously there are many, many sequences and .
11 many of th e sequences are insignificant relative to 12 their contribution of success or f ailure, particularly 13 failure, which is what we are concerned about. And many 14 of those saquences, th e f act th a t they are of trivial 15 contribution would be of such a nature that they 16 wouldn't warrant an extensive review.
17 JUDGE NORRIS That takes some judgment on the 18 part of the reviewer, I guess, as well as the original 19 analyst.
20 WITNESS KASCSAK4 Well, it's true, it takes 21 judgment on the part of the analyst who's involved in 22 producing that information and judgment on the part of 23 the LILCO reviewers as to whether or not we've looked
() 24 far enough. And tha t's true, those are judgment calls, 25 and hopefully those judgmen ts are correct. But they are O
ALDERSON REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
I 5866 !
() 1 based upon a fairly good knowledge of the system and of 2 the data that went into the system.
3 JUDGE MORRISt How long do you think this 4 review process will go on?
5 WITNESS KASCSAKs We were hoping to complete 6 the process by the end of this year. That is still our 7 goal. I think we have to reassess this.
8 JUDGE MORRIS: So you are somewhat more than 9 halfway there?
10 WITNESS KASCSAKs I would say that's true. .
11 JUDGE MORRIS: So the documentation has not 12 been any formal attempt to write a progress report that 13 summarizes particular aspects of the review, but it is 14 more documentation of individuals' or a small subgroup 15 progress reports, or the documentation would be the 16 recommendation or correspondence with SAI that 17 corresponds with the documentation ?
18 WITNESS KASCSAK Well, we do have weekly, t
19 sort of biweekly and monthly progress reports of all of 20 the activities within the divisions and the departments, 21 and certainly any significant outcomes of the review 22 would be documented in those reports.
23 Each individual product, event tree, fault
() 24 tree, or section of th e report, that is reviewed and 25 comments are generated, those comments are available and O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5867
(} 1 they are produced by the individual reviewer, and then 2 they are reviewed by a section supervisor, and then 3 formally transmitted.
O 4 JUDGE NORRISs Are there any current plans to 5 sort of collect and finally document portions of the 6 review or the total review itself ?
~
7 WITNESS KASCSAK We have plans to formally 8 document to management this entire process, to ensure 9 that the initial goals of our effort were achieved, to 10 document what transpired as a result of this particular 11 analysis and how we would continue to further use the 12 analysis in the future.
13 This has been something we have clearly 14 defined as something we will be doing.
15 JUDGE MOESISs Has work started on that at -
16 all? .
17 WITNESS KASCSAK Not in the formal sense, 18 no. All of these pieces will be put together eventually 19 to compile that final report.
20 JUDGE MORRISs Would you expect the draft of l 21 that final report to be available before the end of the 22 year?
23 WITNESS KASCSAKa I would say it should be
() 24 available before the end of the year. I don't know how 25 much before the end of the year.
()
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W WASHINGTON, D.C. 20024 (202) 554-2345
5868 1 JUDGE MORRISs Well, I was thinking of 2 December.
3 WITNESS KASCSAKs I would ssy there's a good O 4 possibility that that could be available.
5 JUDGE MORRISs Thank you very much.
6 JUDGE BRENNER4 Give us a couple of moments, 7 please.
8 (Board conferring.)
9 10 .
11 12*
13 lO ,4 i
15 16 -
17 -
18 19 20 21 22 23 24 25 O
ALDERSON REPORDNG COMPANY,1NC, i
j 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5869 JUDGE BRENNER: Well, we are going to think
(]) 1 2 about this, and we hope counsel preparing further 3 questioning, including red i rec t , will think about it 4 also. And we will proceed with the cross examination at 5 this point.
6 BY MS. LETSCHE (Resuming):
7 0 Mr. Kascsak, the people who are currently 8 reviewing the results of the Shoreham PRA, what criteria 9 do they use for determining what systems interactions 10 are acceptable and what are unacceptable?
11 A (WITNESS KASCSAK) In reviewing the sequences 12 that are produced at the fault tree level and event tree 13 level, they will look at what the sequences -- and 14 again, these cut sets depict the sequences that occur 15 within the trees, and they will look through those 16 sequences to see which ones contribute significantly to 17- the unreliability of the system or to the event, that 18 even taking place.
19 It is a matter of seeing if anything within 20 those seyuances is unusual, or whether or not there are 21 certain possible situations where one particular 22 sequence may dominan t unavailability. And if that is 23 the case, you would say well, maybe I should really look
() 24 at that one particular one.
25 Now, if there are hundreds and they are all O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
a 5870
(} 1 equally distributed and they all look generally like 2 they should look, then you probably don't spend a lot of 3 time on them. But you try to flesh out those ones that 4 are important, and the ones that may be unique or 5 particular to contributing significantly to that 6 particular system unavailability or event taking place.
7 0 Are they looking at the probability numbers?
8 A (WITNESS KASCSAK) Yes, there are probability 9 numbers associated with each of the sequences.
10 0 What probability would be one that was unusual?
11 A (WITNESS KASCSAK) There isn't any one number 12 you can say. I mean, the cut sets do cut off at a 13 certain number. I believe the number is something 1.ike
]s s 14 10
-8
, where you say anything below that is just noise 15 really and really doesn't contribute significantly to
- 16 the overall unavailability, or even sequence probability.
17 From then on, it is just a matter of judgment 18 as to how you look at these. And as I say, if there is 19 one particular sequence that stands out as dominating, 20 or a couple, then you try to look at those in particular.
21 Q What number is it th a t would mean a number of 22 these sequences dominated? What does +.he differential 23 have to be?
() 24 A (WITNESS KASCSAK) I don't think I can give you 25 an explicit answer to that. It is just, again, a matter l
l (2) l
( ALDERSON REPORTING COMPANY,INC, 1
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5871 of the analyst and the reviewer looking at that to see
(]} 1 2 if there is something unusual about that particular 3 sequence or sequences, and whether or not some O 4 particular -- and maybe one particular sequence 5 dominates many trees. And you may consider that 6 sequence, therefore, contributing to.
7 If you were able to reduce the f ailure 8 probability of that sequence, you would not only improve 9 system availability for that one system, but you would 10 improve it for other systems, also. So it is kind of 11 all of that that comes into play together.
12 Q If there aren't any that this particular 13 reviewer decides are unusual, I take it the review 14 doesn't proceed any further with respect to interactions 15 in that sequence? Is that right?
16 A (WITNESS KASCSAK) That is true for this 17 initial phase of the study, but as we have indicated 18 before, this study is going to be used as an ongoing 19 tool; we are going to feed back into it operating 20 experiences to evaluate whether or not those operating 21 experiences were reflected properly in the analysis; 22 whether or not probably or possibly the events that 23 occurred occurred at a frequency greater than we would
() 24 have initially anticipated, and we would evalua te why 25 that is happening.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5872 1 He would also feed into this model future 2 modifications to evaluate how those modifications would 3 affect system reliability or plant response, to see if, 4 in fact, that modification is an improvement in 5 reliability, or refine it in such a way that it will 6 produce the optimum reliability to the plant.
7 Q In terms of what exists right now, though, in 8 the PRA, setting aside what might be sided to it later 9 as you get some operating experience, the review of 10 . systems interactions will not be pursued if the reviewer 11 determines, in looking at these cut sets, that there 12 isn 't anything unusual? Is that right?
- 13 A (VITNESS KASCSAK) Yes. ,If the SAI analyst and 1
i 14 the LILCO reviewer and possibly another aspect would be i
l 15 some peer review group or review of these sequences, but 16 if all of those. reviews conclude that there are no 17 unacceptable interactions, then they will not proceed 18 any further. That is true.
l l, 19 Q Is there anything in particular -- you said 20 there isn't a particular number you can tell me for the 21 c ri te ria . Is there any standard that these people are 22 looking at to decide if something is unacceptable or 23 unusual?
24 A (VITNESS KASCSAK) The standard would be how 25 the particular systems are used within the plant, how O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5873 they would respond to different initiating events and
(]} 1 2 transients, and looking at that, looking at whether the 3 availability and the failure rates that are depicted on
.O 4 these models are reasonable for these kinds of concepts, 5 or typical of what we would expect, and whether or not 6 they would be anything unique.
7 And that is based upon knowledge of other 8 plants, it is based upon knowledge of information, all 9 of these other pieces of information that are supplied 10 to us relative to how the plant should operate, what its 11 design basis is. And if these trees should the plant is 12 responding in a way that is different than it was 13 designed or that we expected, then we would further
\-) 14 investigate that.
15 0 Is there any documented standard that tells 16 you wha t it is you are expecting?
17 A (WITNESS KASCSAK) Well, there are design bases 18 of the plant that tell you how systems should operate, 19 how they should -- for example, the accident analysis 20 that we have in Chapter 15 tells us how the plant should 21 respond to certain event's.
22 0 That doesn't have probability numbers.
23 A (WITNESS KASCSAK) No, but it has system
() 24 dependencies. How systems are used in concert to l 25 respond to a particular accicent or transient, and how O
ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5874 i
l
~
1 they would come into play to mitigate that event.
(}
2 0 So if there isn't any difference in -- strike 3 that. l 4 (Counsel for Suffolk County conferring.)
5 Mr. Kas:sak, is there any document that the 6 reviewer has at his disposal for his use to tell him 7 what is an acceptable or an unacceptable or expected or 8 an unusual systems interaction?
9 A (WITNESS KASCSAK) The only documents he would to have would be the design documents of the plant which 11 would show how the plant was designed, and how certain 12 systems should performs what their service conditions 13 are, how their performance should be, what their .
14 performance criteria is. And in that respect, he would 15 be aware of how the systems should respond, and if for 16 some reason there was an in teraction, -- let's say a 17 non-safety system, or for some reason, an electrical, 18 there was an electrical interconnection in some way that 19 caused that system to be unavailable and not perform its 20 intended function -- then.he would be aware of that 21 through his kno'wledge of what the design basis of the 22 plant was. And it is somewhat of a qualitative type of 23 review.
() 24 0, Mr. Kascsak, is it your testimony that the 25 Chapter 15 analysis is the basis for anticipated systems O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASH!NGTON, D.C. 20024 (202) 554 2345
l 5875 l
performance?
(]} 1 2 A (WITNESS KASCSAK) No, it is not. All I stated 3 was that is just another piece of information that O 4 analyzes how the plant responds to certain accidents and 5 transients, and identifies how certain safety systems 6 come into play to mitigate those events.
7 (Counsel for Suffolk County conferring.)
8 0 Mr. Kascsak, how do you assure that there is 9 any consistency in the way that these different 10 reviewers you have make their determinations on what is 11 unusual and what is unacceptable?
12 A (WITNESS KASCSAK) Well, the formal review 13 process is supervised through my division through the 14 nuclear systems supervisor. And his orchestrating of 15 this review and his review of the review, if you will, 16 is the assurance that there is a consistency in what 17 people are doing and how that -- what the products of 18 those reviews are.
19 Q Mr. Kasesak, what is the systems interaction 20 that is involved or that was involved in your i
21 modification to the viewing window in the containment?
22 A (WITNESS KASCSAK) That is an interaction of l
23 the over-pressurization transients that take place
() 24, within the containment. So upon an accident you would 25 pressurize the containment, fill the containment up with
(
l l
[ ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTC 4, D.C. 20024 (202) 554-2345
5876
() 1 steam, and depending upon how much pressure is produced 2 within the containment, you would have potentially a 3 containment failure. So the pressure within the O 4 containment, the steam would interact with the 5 containment structure and produce a failure.
6 By modifying those windows, we were able to 7 increase the overall structural integrity of the 8 containment such that we could now accept a higher 9 pressure within the containment before we would be 10 concerned about a breach of that containment.
11 Q You are concerned about a failure of this 12 window resulting from pressure, right?
13 A (WITNESS KASCSAK) Tha t ,is true . .
14 0 There isn't any other system or any other 15 component that is reacting with the window, is there?
16 It is just the pressure.
17 A (WITNESS KASCSAK) Well, the interaction is the 18 fact that you have had an accident and you produced 19 steam, so whatever -- the core itself is producing 20 steam, so that the reactor system is producing steam.
'21 That steam then interacts with the containment vall and 22 produces -- potentially affects the integrity of that 23 vall.
(), 24 Q Is it your testimony that any time you have a 25 transient or an accident that might have an effect on I
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5877 any particular component, that you have a systems
(]) 1 2 interaction problem?
3 A (WITNESS KASCSAK) I think a lot of people have 4 attempted to define systems interaction. We have said 5 before it is a dependency; it is a dependency of one 6 component system or systems on another. And in the 7 broadest sense, you can call any of these dependencies 8 interactions.
9 And yes, there are literally, every single 10 time you do anything in the plant, interactions take 11 place. And many of them are acceptable. And in fact, 12 most of them are perfectly acceptable and, in fact, are 13 planned. .
14 Q Mr. Kascsak, yesterday, Dr. Burns stated that 15 systems interactions that were identified in the 16 Limerick PRA had been incorporated, if appropriate, into 17 the Shoreham PBA. Has LILCO performed any comparison of 18 the systems interactions identified in the Shoreham PRA 19 to see if any of those were not present at the Limerick 20 plant?
21 JUDGE BRENNEHa Could I get that question 22 again, because I think I thought you were going to say 23 something else and therefore, I let my mind drift. If
() 24 rou could repeat it.
25 MS. LETSCHEa I can, Judge Brenner. ;
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5878
() 1 BY MS. LETSCHE (Resuming):
2 O Yesterday, Dr. Burns mentioned that systems 3 interactions that had been identified in the Limerick 4 PRA had bean incorporated, where appropriate, into the 5 Shoreham PRA. Has LILCO performed any comparison of the 6 systems interactions identified in the Shoreham and in 7 the Limerick PRAs to determine if any of those present 8 at Shoreham were not present at Limerick?
9 A (WITNESS K ASCSAK) We have performed no such 10 review. I don't know whether SAI has performed such a 11 review.
12 0 Has SAI performed such a review, Dr. Burns?
13 A (WITNESS BURNS) There has been no formal 14 review. I as personally aware of all of the systems 15 interactions that were identified at Limerick; howe 7er, 16 the Limerick analysis was more limited in scope in that 17 it did not --
the purpose of it was not to directly 18 calculate or provide information on systems interaction.
19 Its purpose was solely for the assessment of 20 public risk and solely to sa tisf y a request by the NRC 21 for that purpose. And, therefore, while the engineers 22 at Philadelphia Electric, the analysis engineers at SAI, 23 the reviewers at General Electric were all aware of the
() 24 interactions that could lead to system f ailure or 25 sequence failures, that inf ormation is not documented in O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
l l
5879
(} 1 --
formally documented in that PRA.
2 And there has been no formal comparison 3 between Limerick -- no published formal comparison O 4 between Limerick and Shoreham, although General Electric 5 who was also a consultant to Limerick -- excuse me -- to 6 LILCO on the Shoreham PRA, has performed a comparison of 7 Limerick and.Shoreham and identified areas where there 8 are differences. And those differences are now being 9 documented but have not yet, to this point, been 10 documented.
11 (Counsel for Suffolk County conferring.)
12 0 Mr. Kascsak, do you know when this GE review 13 that Dr. Burns.has mentioned is going to be completed?
14 A (WITNESS KASCSAK) Well, I think for the most 15 part, the differences have been identified and now we 16 are in the process of evaluating why those differences 17 exist and whether they are significant or not. And that 18 is an ongoing process.
19 0 Wen you say differences, are you talking about 20 hardware differences?
21 A (WITNESS KASCSAK) No. Some of it is hardware, 22 but some of it is differences in how the plant responds 23 to a given sequence. And certain assumptions that maybe
() 24 vent into the analysis. It is many different things.
25 0 Is this report something that LILCO plans on l (
i ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5880 1 reviewing as part of its overall PBA review?
(}
2 A (WITNESS KASCSAK) Do we plan on pursuing these 3 differences?
O 4 Q Well, I guess my question is: what -- does 5 LILCO intend to do anything with the GE report?
6 A (WITNESS KASCSAK)'Well, GE made us aware of 7 these differences and we are pursuing them to see if 8 there is anything significant in the fact that they do 9 exist, and if there is something significant we will ",
10 evaluate it.
11 Q Is that -- your awareness of these differences 12 from GE, is that documented anywhere?
,13 A (WITNESS KASCSAK) Yes. The differences are 14 documented in a preliminary report that GE did transmit 15 to us.
16 (Counsel for Suffolk County conferring.)
17 0 What are the diff erences that GE has 18 identified for you?
19 (Panel of witnesses conferring.)
20 MR. ELLIS: Could I have the question again, 21 please?
22 JUDGE BRENNER: Can you repeat it?
23 BY MS. LETSCHE (Resuming):
() 24 0 What are the differences that GE has 25 identified for LILCO?
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5881 1 A (WITNESS KASCSAK) I an afraid I can't recall 2 the details. It has been a number of months since that 3 piece of information was produced. Many of them deal O 4 with assumptions that went into the analysis.
'1 I 5 6
7 8
9 [
10 11 12 1
i 13 14 15 16 17 18 l
! 19 20 21 22 i
23 24 25 O
ALDERSON REPORTING COMPANY,INC, l 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5882 0 They deal with the me thodology of the PRA?
(]) 1 2 A (WITNESS KASCSAK) No, not with the 3 methodology, with the input. It may be, in fact, that 4 the input should be different, that there are 5 differences in the plant that require the input to be 6 different, but it wasn 't obvious to GE, so they wanted 7 to point that out to us, and we are pursuing whether or 8 not there was an error in the input or whether or not 9 the input was proper. '.
10 JUDGE BRENNER: Let's go off the record.
11 (Discussion off the record.)
12 JUDGE BRENNER: Let's go back on the record.
13 BY MS. LETSCHEa (Resuming) 14 0 nr. Kasesak, the input and assumptions that go 15 into a PRA can affect the results of that PRA, isn't 16 that right?
17 A (WITNESS KASCSAK) Of course.
18 MS. LETSCHE: Judge Brenner, I would like to 19 request, or maybe I can just ask Mr. Kascsak or Mr.
20 Ellis, a copy of this GE analysis. It sounds to me like 21 it is a very relevant document if it discusses the 22 assumptions and input that went into the Shoreham PRA 23 and differences between that and the Limerick PRA. In
() 24 light of GE's statements concerning both the Limerick 25 and the Shoreham PRA and LILCO's statements concerning O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5883 the comfort they have drawn from the Limerick PRA, I
(]) 1 2 think it is a most relevant document.
3 JUDGE BRENNER: Okay, you've made your O 4 request. Let's lay it to one side for now. After the 5 completion of the examination of these witnesses, of 6 LILCO's witnesses by all parties, we will come back to 7 your request, if counsel are ready to address it. I 8 will give them a chance to consider it and I will le t 9 you make the request with somd greater support, if you 10 wish, or you can just stand on the comments you just 11 made. ,
12 WITNESS KASCSAK If I could just make a quick 13 comment, as we have stated many times, the analysis is i O 14 under review, and it is being reviewed by many people, 15 and this piece of information is just one of the pieces .
16 of information that have come back through the review 17 Cycle that indicate some questions as to whether or not 18 some of the input was consistent with other input in 19 another analysis. But it is no different than other 20 comments we are getting from peer review group members 21 and people within LILCO, our operating staff. I mean, 22 there are all kinds of reviews taking place right now, 23 and this is just one of them, and like the culmination
() 24 of this entire review process, will produce a report 25 that we feel is proper and is state of the art. And I
()
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554 2345
l 5884 l l
l
() 1 don't know why you should single out any one particular 2 part of the review as being significant.
3 JUDGE BRENNER: So you didn't ask for enough.
4 (General laughter.) .
I l
5 JUDGE BRENNER: One reason I indicated the 6 procedure we would take is to give Mr. Kascsak and any 7 other technical experts a chance to confer with counsel 8 so he can get an orderly response which presumably will 9 include some of the gomments you just made, which I am 10 sure were helpful to your counsel as well as to 11 everybody else.
12 (Counsel for Suffolk County conferring.) ,
13 BY MS. LETSCHEs (Resuming) 14 0 Mr. Kascsak, yesterday you mentioned certain 15 high grade dif ferences between Shoreham and Limerick, 16 particularly the automatic standby liquid control system 17 and the containment vents that are in the Limerick 18 plaut.
19 Has LILCO analyzed the significance of either 20 of those features in the Limerick PR A results?
21 A (WITNESS KASCSAK) We are aware that those 22 were design inputs that went into the Limerick analysis, 23 meaning they were assumed to be part of the design basis
() 24 when the analysis was performed. Even though that may 25 not be the case for the second decision, the containment O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
.. - _ - - . ,-. . - ~ _ -
i 5885
(} 1 vent is something that-I think they have made a later 2 decision not to install, but in any case, we were aware 3 that they were design assumptions that went into the O 4 analysis, and we were certainly aware that those same --
5 tha t similar design assumptions were not included in the 6 Shoreham analysis. And in that respect, we are aware 7 that there will be differences in the plant response as 8 a result of the fact that one plant has a slightly 9 different design than the other.
10 And in looking at the results, it will be 11 obvious to us why the results are different in some 12 cases, because in fact those designs are within that ,
13 plant and not within Shoreham.
14 Now, it is a matter of us assessing whether or 15 not the Shoreham design is adequate, as it is basically, 16 as it is presently designed, and whether or not there 17 are any reasons for us to consider to make some more i 18 changes.
19 I think preliminarily, based upon the resultc -
20 we have seen and the fact that our reliability numbers 21 and core melt frequencies are somewhat similar to 22 Limerick, even though they have these design changes, it 23 would indicate that they are probably not necessary for O 24 saerea -
25 0 Have you looked at whether or not the systems O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D C. 20024 (202) 554-2345
l 5886
() 1 were contributors to any of the differences in dominant 2 sequences or in probabilities of particular sequences in 3 the Limerick PRA?
4 A (WITNESS KASCSAK) Yes, and they are.
)
5 0 What criteria is LILCO going to use to 6 determine whether or not the LILCO numbers are adequate 7 to justify not adding those systems to Shoreham ?
8 A (WIINESS K ASCS AK ) Well, I think as we 9 in[dicated --
10 MR. ELLISs I would like to register an 11 o bj ec tion . I don't know about adding systems to 12 Shoreham, how that relates to systems classification, 13 but the witness had already started to answer. But I 14 don't see how this line of questioning rela tes to 15 classifications or interaction.. I think it is an 16 additional line of questioning relating to ATWS.
17 JUDGE BRENNER It is starting to drift out 18 again, but not too far yet, but I think we have seen 19 this morning, we will allow that question in a moment,-
20 but I think we have seen this morning how much more 21 efficient it is when we draw the focus back in within 22 the context of the contention. So I may jump in with a 23 few questions myself, but let's see where it goes.
() 24 MS. LETSCHE: Thank you, Judge Brenner.
25 BY MS. LETSCHE: (Resuming)
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5887 0 Do you recall the question, Mr. Kascsak?
(]) 1 2 A (WITNESS KASCSAK) Yes, I do.
l 3 As v a have stated in our tentimony, our 4 initial gotl is to ensure that there is nothing unique 5 or nothing disproportionate relative to the Shoreham 6 design, relatjve to its contribution to public risk.
7 The generally accepted and to date probably the best 8 benchmark to ovaluate a plant performance against vould
. 9 be the W ASH-1u00 results, at least, the core vulnerable, ,
10 the core melt frequencies and the consequence curves 11 tha't were produced in WASH-1400 as some standard by 12 which at least some judgments were made tha t tha t-13 prosented a reasonable. risk. 'So'we wanted to ensure
\- 14 that the Shoreham plant was at least equal 'o t or better 15 than, or within certain uncertainty bounds of the 16 analysis, that the Shoreham plant f(11 within those 17 bounds.
18 And that is our initial goal, is a required -
19 part of th e product of this analysis, af,ter we assure 20 ourselves that the.Shoreham plant is certainly equal to 21 or better than any of its contemporary plants. Nov it 22 is a matter of asking some personal judgments as to 23 whether or not we want to go a step further and improve
() 24 tb2 plant in certain voys, and we will.use the analysis 25 to tell us what types of' sequences or systems within the ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5888 plant are important to public risk and where we might
(]) 1 2 best put our attention both in plant maintenance and LER
~ 3 review and future modifications of the plant, and 4 certainly as part of the second order review of this 5 analysis we will in a sense prioritize what systems and 6 sequences are important to the plant and see how we may 7 vant to improve them in the future.
8 0 Do you know whether or not the Shoreham PR A
. 9 identified any systems interactions that were not .
10 present at Limerick because of the additional systems 11 that you know were in the Limerick plant?
12 JUDGE'BRENNER: Do you want to modify that 13 question to be because of the additional systems that l
14 were assumed in the Limerick PRA?
15 MS. LETSCHE: Yes.
16 WITNESS KASCSAKa I would say no, and I would 17 say no particularly because I think it was the event 18 sequences that these modifications that we talked about 19 vent to an attempt to let's say prevent the sequences 20 from leading to the failure, the core failure. Those 21 sequences were well known, and in fact, they were 22 certainly well known in the industry such that Limerick, 23 considering their site conditions and other things, gave
() 24 them more incentives to make those modifications. Those 25 same incentives or motivations don't exist for Shoreham O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 l l
5889 I
(]) because of different site characteristics and to some 2 extent diff erent design features within the plant, but 3 certainly something like ATWS and those sequences were 4 well .\nown in advance, and it was just a ma tter of a 5 management decisin as to how far you go with adding 6 systems within a plan to mitigate that event. We have 7 systems within the plant to mitigate ATHS and we have 8 added certain systems within the plant already on 9 Shoreham, RPT, ARI, alternate rod insertion, and scram to discharge volume modifications, all to enhance the 11 ability of the Shoreham plant to mitigate an ATWS.
12 It is certainly cleu2 that if we were to add 13 more things to the plant, we wcaid ao better in response 14 to some of thse events, but the question is do you need 15 to, and that is the judgment you make by looking at 16 these analysis.
17 BY HS. LETSCHEa (Resuming) 18 Q Shoreham has already made that judgment that 19 it is not going to add these additional things, is that' 20 right?
21 A (WITNESS KASCSAK) Well, I wouldn 't say the 22 judgment is for ever and ever, but we have concluded at 23 this point in time that it is not necessay.
() 24 (Counsel for Suffolk County conferring.)
25 BY MS. LEISCHE: (Resuming)
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) $54-2345
5090
(} 1 0 Your answers to this concerning the mitigation 2 of ATWS, does that apply to both the automatic standby 3 liquid control and the overpressure vents?
I 4 A (WITNESS KASCSAK) Well, the 5 overpressurization vent would help for all containment 6 overpressurization events and an ATWS or anything else.
7 Q My question was not very artful.
8 Did your sta tements concerning LILCO 's 9 judgment about putting such -- I's.not sure it is a 10 system, but such a change into the LILCO plant, did your 11 answers cover both the ATWS changes and the overpressure 12 vents? .
13 A (WITNESS KASCSAK) Yes.
(/
\
- s. 14 Q Mr. Kas:sak , do you know whether or not 15 Limerick has the same number of diesel generators as 16 Shoreham does?
17 A (WITNESS KASCSAK) I know they do not.
18 0 They have one more, do they not?
19 A (WITNESS KASCSAK) Yes.
20 0 Did LILCO consider that design difference in 21 reviewing the accident sequence probabilities in the 22 Limerick PRA?
23 A (WITNESS KASCSAK) We didn't review the i () 24 accident seg uences in the Limerick PRA to the extent as 25 to how the four diesels interplayed. I think four isn't O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 1
5091
(} 1 necessarily better than three. I mean, it is not as if 2 they are all redundant. There is certainly an amount of 3 redundancy built into the plant. The loads are shared 4 between the diesels. But it doesn't mean because they 5 have an extra diesel that they have more redundancy than -
6 Shoreham. That is not the case. So you would have to 7 get into the real details of the plant to really 8 determine whether or not there is anything better in the G Limerick design than Shoreham. '
10 0 That really wasn't my question.
11 Has LILCO analyzed, whether or not there are 12 . any dif ferences in terms of systems interactions or 13 responses to transients identified in the Limerick PR A -
O
\/ as opposed to those in the Shoreham PRA that arc related 14 15 to the additional diesel in Limerick?
16 A (WITNESS KASCSAK) Only to the extent of 17 looking at how the loss of off-site power affects the --
18 how it affects the overall risk and whether or not there 19 would be*anything unusual about Shoreham compared to the 20 Limerick layout.
21 0 And wha t did you determine?
22 A (WITNESS K ASCS AK) That any differences were 23 vell explained by the systems interconnections, the
() 24 off-site power interconnections and the plant layout.
25 0 But you were just looking at the probabilities O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
5892
(} 1 sssociated with that sccident sequence, the loss of 2 off-site power?
3 A (WITNESS KASCSAK) Well, the probabilities
-)
G 4 plus the design of the plant.
5 0 You weren't looking at systems interactions 6 that might result from that equipment dif f erence ?
7 A (WITNESS KASCSAK) Of course we were. This is 8 all systems interactions, the ability to shut the plant 9 down using off-site power, the ability of the diesels to 10 come on, pick up load, start the proper emergency core 11 cooling equipment, perform decay heat removal,
. 12 containment cooling, all of that is systems 13 interactions. The diesel systems interact with the 14 emergency core cooling systems to perform that 15 function.
16 0 Your determination that there wasn't any 17 unexplained difference between Shoreham and the Limerick 18 PRA results was based upon your review of the dominant 19 sequence probabili5 ties for loss of off-site power, is l 20 that right?
21 A (WITNESS KASCSAK) In part. I mean, that 22 would tell you if there were any large differences, and 23 then you would, if you saw the differences, you would
() 24 try to investigata in detail why the differences existed 25 to see if there was anything peculiar in why those O
l ALDERSON REPORTING COMPANY,INC, l
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
l 1
5893 differences existed. So you would get down into the
[]} 1 2 details of the design by looking at that, by looking at 3 the fact that that was a dominant sequence that led to a O 4 core melt situation.
l l
5 0 Well, did LILCO do that? Did you get down '
6 into the analysis, or did your determination at the ,
7 level of the dominant probabilities enable you to stop 8 your analysis there?
9 A, (WITNESS KASCSAK) Since that review was done 10 by individuals in the electrical engineering department, 11 I can't comment totally on the extent of their review, 12 but they did perform their review to the extent to 13 satisfy themselves that there wasn't anything unusual 14 about Shoreham.
15 JUDGE BRENNER: Ms. Letsche, I was hoping to 16 complete your cross examination before we took the first 17 break, and I thought you were somewhere in item 6 on 18 your handwritten revised plan late yesterday and 'this l
19 aorning, and now I see we have kind of gone back up to 20 parts of item 5.
21 How much more do you have?
22 MS. LETSCHE: Judge Brenner, I have one more 23 question concerning equipment differences a t the two
( 24 plants, and then I have on my piece of paper, I only 25 have three questions that are the wrap-up of No. 6. I O
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W, WASHINGTON, D.C. 20024 (202) 554-2345
m 5894
() 1 cannot tell you how long it is going to take me to get 2 answers to those questions.
3 JUDGE BRENNER: Why don't you ask the one 4 about equipment differences, and then we will come back 5 for the other three.
6 BY HS. LETSCHE: (Resuming) ,
7 Q Mr. Kascsak, are you aware of an equipment i
8 difference between Limerick and Shoreham involving low '
9 pressure pumps? -
10 A (WITNESS KASCSAK) Yes.
11 Q Is it true that at Limerick they have, the 12 Limerick plant has low pressure pumps which do not 13 require a positive backpressure for the operation at 14 high suppression pool temperatures?
15 A (WITNESS BURNS) Can we confer?
16 JUDGE BRENNER4 Why don't we take the break 17 while they confer, but that was two questions already i 18 and not one question.
19 So we are going to take a break until 11:00 20 o' clock, and then when we come back I want the cross l
1 21 examination by the County completed in 15 minutes or l
22 less, hopefully less.
23 We will be back at 11:00.
() 24 (A brief recess was taken.)
25 l
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE, S.W., WASHINGTON. D.C. 20024 (202) 554-2345 L.
1 5895 i
{} 1 JUDGE BRENNER: All right. It is 11:01.
2 Let's go.
3 ES. LETSCHE: Judge Brenner, I would just like O 4 to place an objection on the record to your ruling prior 5 to the break. As I indicated, I do not have extensive 6 additional cross-examination and in fact I have two 7 additional questions. But as I stated before and as has 8 been evident with this panel, it is frequently necessary 9 to ask a question several times or to ask follow-up 10 questions in order to obtain an answer to the one I have 11 asked. And if that is necessary based upon the 12 responses I get, I intend to do so.
13 JUDGE BRENNER: Well, you're taking away from 14 your time, so I won't reply extensively. But your 15 observation is not correct with respect to this 16 morning. It may have been correct at other times. And 17 in fact, while you were being productive and efficient 18 we were more patient and your questioning vent on, even 19 allowing for our approximate interruption of 15 minutes 20 at the most, it went on -- you started questioning at 21 9:15, so that was almost an hour and 15 minutes of your 22 questioning.
23 And the last series started to diverge again.
And one reason you asked the question three or four
( 24 25 times -- and I wouldn't have gone into this except for O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W, WASHINGTON, D.C. 20024 (202) 554 2345
5896 1 your comment -- is you started with a very broad
(~}
2 question about the diesel generator difference and then
- 3 four questions later said, the question I'm asking is 4 about system interaction, when in fact that was the 5 first time you had asked it that way.
6 So let's proceed. ,
7 BY MS. LETSCHEa (Resuming) 8 0 Mr. Kascsak, LILCO has not reviewed Table 9 3.2.1-1 against the results of the Shoreham PRA, has 10 it?
11 A (WITNESS K ASCS AK) I would say Table 3.2.1-1 12 is just a summary table of the systems with the safety 13 systems within the plant. Those systems that I l
O
\- 14 identified have safety functions and those safety 15 functions, the individuals who are performing the 16 reviews are familiar with those safety functions. And 17 in that respect they are using that information in 18 reviewing the Shoreha's PRA.
19 0 Has LILCO reviewed table -- what is set out in 20 Table 3.2.1-1 in terms of the classification of 21 structures, systems and components against the results 22 of the Shoreham PRA? Has such a specific review taken 23 place?
() 24 I think you can answer that question yes or 25 no.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W WASHINGTG:4, D.C. 20024 (202) 554-2345
4 5897 1 A (WITNESS KASCSAK) There is an indirect review
(}
2 in the sense that when -- if for some reason we were to l
3 determine an unacceptable interaction and that l O 4 unacceptable interaction required a reclassification of 5 the system, then we would do that.
6 0 Have you done,that?
7 A (WITNESS KASCSAK) We haven't found any 8 unacceptable interactions that required reclassification 9 of systems.
10 (Pause.)
11 0 Have you used the res21ts of the Shoreham PRA 12 to analyze the adequacy of classification and treatment 13 of structures, systems and componente that are important 14 to safety in the Shoreham plant?
15 A (WITNESS KASCSAK) Yes, in the same sense that 16 safety and non-safety systems, in fact all systems that 17 come into play to mitigate accidents and transients 18 independent of their classification, are evaluated 19 within the PRA and their relationship to success or 20 failure is then produced through the analysis. And in 21 that respect we will evaluate the proper classification 22 of all of those systems, whether they be safety, 23 non-safety or some process in between.
() 24 0 Has that analysis setting out the way a 25 particular structure, system or component is stated in O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
l 5898 1 Table 3.2.1-1 or the importance to safety of that
(
2 structure, system and component been documented anywhere -
gr 3 in terms of a comparison between that and the results of U 4 the PRA?
5 A (WITNESS KASCSAK) The comparison has not been 6 documented. ,
7 0 Are the reviewers of the PRA instructed 8 specifically to make such a comparison?
9 A (WITNESS KASCSAK) The reviewers of the PRA 10 are required to review the results for evaluating 11 whether the plant was properly designed and responds in 12 accordance with its design bases, and in that respect 13 they will see how the systems respond or how they affect l
T 14 the response of the plant.
15 0 Are ther told to review those results against 16 the classifications set out in Table 3.2.1-1?
17 A (WITNESS KASCSAK) Yes, in the sense that we I
18 are' particularly concerned about system and non-safety 19 system interactions, and if we discover a situation 20 where a non-safety system interacted in a way to degrade 21 a safety system we would be particularly concerned about l
22 that. And that is in f act one of the more important 23 things the LILCO reviewers look for, for both the review
() 24 of the PRA results and the review of design documents 25 that are produced by our architect-engineer and the NSSS l
CE)
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE, S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5099
(} 1 vendor.
2 0 Is there any explicit documented instruction 3 to those PRA reviewers to look at the PHA results and 4 compare those results to the classification set out in 5 Table 3.2.1-17 6 ,A (WITNESS KASCSAK) There is no explicit 7 procedure. It is part of what they are told to do as 8 part of the review process, Jut it is not documented.
i 9 0 Thank you.
10 MS. LETSCHE: Judge Brenner, at this point I 11 would like to move the exhibits that we have had marked 12 during this section of Suffolk County's 13 cross-examination into evidence. I have not had an 6' 14 opportunity to discuss that question with Mr. Ellis, but 15 ve are prepared to move them in at this time.
16 JUDGE BRENNERa Does that mean you've 17 Completed all of your questions?
18 MS. LETSCHE Yes, it does.
19 JUDGE BRENNER: Because I didn't want to take 20 avir from your time by doing that. All right.
21 Why don't you tell me what numbers you were 22 thinking of specifically? I suggest that you are 23 probably thinking of 22 through 26?
() 24 MS. LETSCHEs Tes, that is correct.
25 JUDGE BRENNER: Mr. Ellis?
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5900
{} 1 MR. RAWSON: Judge, may I be heard while Mr.
2 Ellis is preparing? .
,e 3 JUDGE BRENNEBs Certainly.
4 MR. RAWSON: The Staff would have no objection 5 to the introduction of those items in evidence to the 6 extent that they are being admitted for the limited 7 purpose of clarifying or impeaching the testimony of 8 these witnesses. He would object to them being admitted 9 f or the truth of all that appears within the documents.
10 JUDGE BRENNER: Why, and which particular 11 parts?
12 MR. RAWSON: Judge, I don't have all of the 13 exhibits in front of me, but there are many segments of 14 this which are segments of much broader documents.
15 There are extensive portions of these documents, I 16 believe, which have not been the subject of specific 17 testimony from these witnesses, and it just seems to the 18 Staff to be improper to be according every word which 19 may appear in these documents into the record at this 20 point. .
21 JUDGE BRENNERs Ms. Letsche, maybe I should 22 come back to you and ach your intent in moving them into 23 evidence.
() 24 MS. LETSCHE Judge Brenner, I think these witnesses have been involved in the preparation of these 25 O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5901 1 documents, and their testimony certainly concerns the
(}
2 documents. And they have had.an opportunity to indicate c$ 3 if there were corrections that needed to be made in the C) 4 things that we have marked as exhibits.
5 And frankly, I don't find any -- I can't think 6 of any basis for the Staff's objections when these 7 witnesses are responsible for these documents.
8 Certainly Mr. Ellis will have the opportunity to put 9 additional information in himself during redirect if he 10 or his witnesses feel it is necessary. ,
11 JUDGE BRENNER: Yes, I think that was the 12 point, and that is one reason the County is moving them 13 in now. And we will hear f rom Mr. Ellis in a moment.
l 14 In general, I have shared the concern you l
15 expressed about just admitting documents into evidence.
16 But these are quite limited pieces. And I also 17 understand your point that the full context isn't 18 there.
19 But part of the reason for moving them in is i 20 because we're going to have further questions on it.
21 LILCO indicated they would and the Staff is also free to 22 pursue that. And one reason I would like the documents 23 in, absent a particular objection, is there are some
() 24 things hanging out there, so to speak, between the 25 questioning and the documents, and that is why.
)
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W., WASHINGTON, D.C. 20024 (202) 554-2345
l
\
5902 -
() 1 If it was just for the purposes you indicated, 2 then we wouldn't have moved them into evidence. We 3 could have left then marked for identification to 4 clarify the questions and answers that were actually 5 asked.
6 Mr. E3 Lis, you indicated the other day you 7 vanted to move them in and I think Ms. Letsche in effect 8 is accommodating that, plus my comment about doing it 9 sooner rather than later, if you were going to do it 10 anyway, to save time.
11 MR. ELLISa Yes. We have no objection to the 12 documents f rom the Shoreham PRA. At the time I made my
_s 13 comment I'm not sure there were any documents introduced
- 14 with respect to the Limerick PRA. We would not have any 15 objection to the admissibility of those, although with 16 the exception of joining with the Staff there for the 17 Limerick PRA, because that wasn't to be the focus of ,
18 this hearing.
19 JUDGE BRENNER: There is very little in those 20 excerpts from the Limerick PRA other than the questions 21 that were asked. And I think I am going to grant the 22 motion to admit Suffolk County Exhib?ts 22 through 26 23 without any limitation, because they are concise,
( 24 limited excerpts. To the extent any clarification is 25 needed, questions can be asked.
O ALDERSON REPCRTING COMPANY,INC, ,
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5903
(} 1 The ruling might have been different had the I 2 documentation been more extensive with respect to the 3 Limerick PRA. So they are admitted at this time.
4 (The documents referred 5 to, previously marked for 6 identification as Suffolk 7 County Exhibit Nos. 22 8 through 26, were received 9 in evidence.)
10 JUDGE BRENNER: I would like to bind each of 11 them in in sequence and make sure copies are provided to 12 the reporter with the exhibit numbers marked, of course, 13 on each copy.
s 14 (The documents referred to, Suffolk County 15 Exhibit Nos. 22 through 26 in evidence, follova) 16 17 18 19 20 -
21 22 23 pl
() 24 25 s
ALDERSON REPORTING COMPANY,INC, ,
400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
.o - - .- . .
bA'I I I bvifolk (a. NMn
,y EEb5-ERE.lffflyy
- 2. 4 PLANT WALKDOWNS A preliminary evaluation of a nuclear power plant can be performed using k ~
- design drawings, specifications, procedures, and FSAR data. However subtle dependencies and interfaces can best be imparted to the analyst during a thorough walkdown of each system.
A comprehensive system walkdown was done as part of the SNPS risk assessment Phase I effort. The purpose of the walkdown was to perform an as-built review of each system modeled in the probabilistic analysis and to consider and document any special features which could affect safety on the capacity of the system to perform in an accident environment. An additional purpose was to familiarize the analysts who would be constructing system level logic models, so that the quality of the model would be improved and the more subtle site-specific system differences could be identified. A final purpose for the system walkdown was to develop sufficient system interaction h and location interaction data for inclusion at a later date in a location dependent model.
Systems were assigned for the walkdown as shown in Table 2.4.1. For the walkdown, each SAI reviewer was teamed with a LILCO engineer, know-ledgable in the particular design and operation of the system. All reviews and walkdownseb' gan in the control room and finished up in the control room.
l At the end of the walkdown, LILC0 operations engineers were questioned on system design particulars and any unresolved issues concerning the system description.
For each system reviewed, a detailed set of worksheets were developed identifying pertinent features of the system and identifying potential l n operational hazards such as sprays, connections, and environmental control b " requirements which could impact system operation. Worksheets, so developed
^were kept by the analysts who could then cross correlate their' observations with their analysis.
i ( i i
l 2-32 i
DRAFT-PRELY NARY ~
{J For interior systems worksheets of the format shown in Table 2.4.2 were used and compared against drawings and the system fault trees. For non-mechenical systems, a set of observations were recorded directly and compared against one line drawings and the fault tree logic. An example of this type of review is given in Table 2.4.3.
The plant walkdowns were quite useful to the probabilistic assessment since Shoreham has a number of unique plant features. Walkdowns reduced the tendency on the part of risk analysts to consider only the gross and generic features of the plant. There is a growing feeling that the truly unique plant contributions to site risk stem not only from site and generic design differences, but from subtle engineering and architectural differences; and that these differences are not easily uncovered unless the analyst is quite familiar with the plant. Included in the system description are valuable information used in the plant analysis which was developed during the detailed system walkdown of the key systems.
n ,
sj r'
V 2-33 i t
R .
i Table 2.4.1 I'.. '..').h[
PLANT WALKDOWN SYSTEM ASSIGNMENTS lh Q _
Service Water E.T. Burns -
- Rx Building Standby Ventilation Chilled Water
- Rx Building Closed Loop Cooling Water (RHR Pump Seal Coolers Only)
A.J. Unione - Electric Power
- ADS A.A. Hughes - Diesels -
- Control Rod Drives /CRD Pumps RHR/LPCI
- Core Spray l
M.T. Drouin - HPCI FW/ Condensate
- RCIC
- Service Water Z.T. Mendoza - Containment Isolation 1
2-34 i
-w A
Tablo 2,4*'w- $w \J PLANT WALKDOWN FORMAT -
f{f_#Eh?tg9y Room Other Environment Generic Hazards Type Connections Controlled Adjoining Lccation Component Type Near By Hazard -Type -How Areas Comp. 1 Valve 1 Hand Yes Spray Test line Yes-Room Yes-Elev. 8-Rx Operated valved out Cooler-H O Hallway ~
Bldg-Nxt, to Rising 2 type - -
Cont. Stera
[n Globe Comp. I Valve 1 Hand Yes Cardox Elev. 8-Rx Operated Bldg-Nxt. to Rising ,
~
Cont. Stern Globe Comp. 2 l .
l l
O 9"
) .
Table 2.4.3 @{F,y'ffps..m.,
LIST OF PLANT WALKDOWN RESOLUTION ITEMS FOR ELECTRIC POWER SYSTEM ITEM ,
DESIGN LAYOUT OPERATION ~ RESOLUTION e 138 kv lines from X FT input data plant are on same checked poles in switchyard e Gas turbine initi- X X Impact on ated to 69 kv lines LOSP frequency is not currently checked capabTe of black start but can be rendered so e 120 VAC instrument X FT logic checked power comes .
only from AC source (no DC inverts) e 138 Kv and 69 Kv lines X X Conmon cause im-cross with 69 Kv line pact on loss of
("cp) undernea th. Single multiple grid 130 Ky failure could connections checked cause 69 Kv failure e Wildwood substation and X X FT input for main-Shoreham switchyard have tenance error mecnanical switches checked locked and 69 Kv circuit breakers unlocked a Reactor not shed from X FT logic for diesel grid during startup of startup signals re-diesels viewed e Diesels started up as X Impact on LOSP fre-safety system power quency checked .
backup for potential expected voltage drops e Motor control centers X FT logic checked are AC controlled e Diesel generators have X Multiple diesel separate cooling for failure data re-
, each unit from ESWS viewed but common header 2-36
~ w SG 64.
bl i [- k h ,ff,,I0] y 43 g- When the large LOC initiator distribution is combined with the probability g distribution of unsuccessful coolant makeup, the accident sequence
/
,, probability distribution is determined to provide the uncertainty bounds for Class III and V. The large LOCA sequence used to characterize Class III and V, has an error factor of 21.
Summary of Secuence Uncertainties In summary, the uncertainty in the frequency of each class of radionuclide release is given in Table 3.8.1. .c e m .;,; y 's Table 3.8.1
'g9 9q._-O Sumary of Sequence Uncertainties in the Frequency -
^II CLASS 015TR15UTICN UNCERTAINTY CE I Tful Log Normal 17-Error Factor
!! T;W Log Normal 14-Error Fac' tor l ,
, !!! AV Log Nomal 21-Error Factor IV TCC Tg2 Log Normal 15-Error Factor l V ses !!! Log Normal 21-Error Factor l
l 3.8.2 Treatment of Uncertainties in Consecuence Calculations To a large extent, uncertainties in the consequence data produced for the l Shoreham PRA result in one or both of the following effects on risk data:
l
- 1. Uncertainty in the probability of a significant release
! given the fact that the core attains a state where it is l vulnerable to meltdown
- 2. Uncertainties in the makeup, magnitude and timing of radio-nuclide releases due to core melt processes taking place.
- & f 3 Some of these effects are addressed qualitatively in Table 3.8.2 which gives
[ the source of uncertainty, its potential impact on SNPS PRA data, severity
- m. .
1 D ~
. 3-172 A
- / . --
- /. _ . . . . . . .
]
jkhf"f,1.l4f)y Table 4.1
(-
<j.
SUMMARY
OF THE FREQUEf1CIES OF CORE VulflERABLE j . .
C0fiDITI0fiS BY ACCIDEflT CLASS cENERALIZED CLASS g3 FREQUENCY OF CCRE VULNERA8LE CLASS (PER REACTOR YEAA)
Loss of Coolant *tateup I 2.7E.5 Loss of containment Heat 11 1.1E.5 Removal ,
, LOCA !!! 3.6E.7 ATVS V/0 Poisen Injection !Y 6.1E-6
,LOCA Outside Containment V 2.0E-8 Total core Vulnerante Frequer.cy 4.4E-5 (Per Ra Tr) 4
. gi , 1 Mean estimate I "T" 5s and 95s l
. 1 Confidence bounds
, 10-* .-
(' E "
D a . '
Gc to's. JL 8 I h e .
N 4 30 ~
5T c >- n
~~
1 5E - --
g s. 30-7 gg o n s w- a .
u --
i .
I a
"~
- M
,.ic
! class class It class sti class tv cuss v 1
I. ,
! 'l .
- Figure 4.1 Sumary of Core Vulnerable Frequencio.c T,ncluding the Uncertainty Characterization e
s, KAN
- 4.4[.speactor fear L
l(A!N.l4 ry) Fo.,ofDCPo-ee LOLn,Selte.t DI w.An - 210'*ntncm stu y; .
i .
s,.s,.. t
, t...,.y....
Tu tv u
ms. ms E . . .
. TQuv ,
~
c3
- . :=2
. Subsequent to WASH-1400, t1RC evaluations of the potential contribution of ATWS to ::"
' core melt (e.g., flUREG-0460) placed the frequency of ATWS is a BWR at nearly twenty times that evaluated in WASH-1400. If this were incorporated into the figure it M*
would be the single dominant contributor to core melt, and would be significantly T larger than the frequency of core melt calculated for Shoreham.
M~
~
r--
- EC.
Figure 3.6.6 Comparison of the Contributing Accident Sequences to the Calculated g:;
' Frequency of Core Melt from WASH-1400 and the Shoreham Analysis . >
~
(Area of " Pie Chart" is proportional to Mean Frequency)
Q D Q . _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ e
SC, DWW4 as
. WASH-1400 basically used one class of accident sequence and five containment failure modes to represent BWRs. Therefore, for the purposes of estimating
(-) the total calculated frequency of a potentially degraded core condition, the Limerick classes should be summed and compared with the value from WASH-1400. The Limerick evaluation produces a total estimate of degraded
(] core conditions smaller than WASH-1400 (see also Figure 3.5.4). Figure 3.5.2 indicates that the events are all of relatively low probability.
30-4 -
kA5H.1400:
Approximate frequency of
- --y2'.2-W'- 2 2*"
in-5 -
/
/V/
j .
, 10-6 ,
\ , -
t: ._
lo*I -
?
l
.s 10'8
~
,. A CLASS I A
CLASS !!
'4 CLAS5 III A
CLASS IV Loss of Coolant Loss of Heat Ant 5 Ants Inventory Nkeup Removal Capablitty from Contaltunent' 1 ..
o .
1 Nl , .
~
- Figure 3.5.2 Sumary of the Accident Sequence Frequencies Leading to Degraded Core Conditions Summed Over ,
All Accident Sequences within a Class. .
- 1 3-96 t . -
Sara C~gbW+ aG of a set of valves which can be operated from the control room to relfeve pressure in the containmtnt (see Appendix B). Since the valves are assumed b$ to be interlocked to higt. radiation monitors, COR can only be utilized for cases where no significant radiation has been released to the drywell.
These cases are generally the Class II and some Class IV types of sequences,
,O '""i"S'""'"'"'"""*"""""'***"*"i"*"'- ' " '" '" **
classes,the reactor core is adequately cooled; the major concern is main-taining the containment intact' and within its pressure capability, while insuring no offsite consequences. (In considering COR, a conservative analysis, using 5% worst meteorology *, a semi-infinite cloud mo' del **, and a realistic noble gas source, showed that offsite doses would be less than one five-hundredth of permissible guidelines (10CFR100), and would result in no offsite consequence, based on exposure of the population.)
WASH-14M (IMERt(K
. MEAN - 3x10*I/ REACTOR YEAR AM - 1.5x10 5/RIACTOR YEAR s, .s, .A s
TQUV
,)
1 ,.
T OU3 F
TW LOSS OF ,
OFF5ITE POWER ATW5
.. ATWS TW g
. 53.52 l
l 7 Figure 3.5.4 Comparison of the Contributing Accident Sequence to the Calculated d .
Frequency of Core Melt from WASH-1400 and the Limerick Analysis (Area of " Pie Chart" is proportional to Mean Frequency) l 1
- Worse conditions exist only 5% of the time
, . ** Conservative by approximately a factor of thre.e. .
b k. .
3-100 e ,
W a
5904
() 1 JUDGE BRENNER: I guess I should note that it 2 is 11:15. So the one time I set a time limit, counsel 3 finished well in advance of it. Maybe I waited too long 4 to set it, I don't know.
5 We will go to the Staff for its questions.
6 MR. RAWSON: Thank you, Judge Brenner. For 7 the Board's inforsation, I intend to be using the 8 memorandum from Mr. Conran to Mr. Thadani of October 20, 9 1981, which is enclosure one to the Dircks memo which is 10 attached to the Staff's prefiled testimony. Counsel has 11 been advised of that and the witnesses provided with 12 additional copies.
13 (Discussion off the record.)
O 14 JUDGE BRENNER: Back on the record.
15 We have got it now. And just for 16 clarification, at the time the Staff moves its testimony 17 into evidence, it is going to include the cover memo 18 f rom Mr. Dircks to Mr. Shewmon, as well as this 19 attachment; is that correct?
20 HR. RAWSON: Yes, that is correct.
21 JUDGE BRENNER: All righ t.
22 MR. RAWSON: May I proceed?
23 JUDGE BRENNER: Yes.
() 24 CROSS-EXAMINATION ON BEHALF OF THE NRC STAFF 25 BY MR. RAWSON:
O ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE, S.W, WASHINGTON. D.C. 20024 (202) 554-2345
5905 e
() 1 Q
~
G e o,t l e m e n , let'me focus you on that_section of 2 your testimony tha t' I want to explore a little bit. The 3 statement is oade at several points in your prefi'ed 4 testimony that a probabilistic risk assessment is the 5 "best method" available for resolution of systems 6 interactions concerns. That statement more or less 7 a ppea rs on page 81 and page 84 and page 97 of your 8 prefiled testimony.
9 Dr. Burns and Dr. Joksimovich, those are your 10 sections of th e testimony, so my questions will be 11 primarily directed to you.
12 Implicit in that statement and explicit 13 throughout LILC3 's testimony is the fact that systems 14 interactions analysis can be done using methodologies 15 other than event tree-fault tree , isn ' t tha t right?
16 A (WITNESS J0KSIMOVICH) There are some 17 so-called methodologies that purportedly intend to 18 accomplish that, although I can 't see in those 19 methodologies any discriminating factors, how we are 20 going to distinguish between something which is 21 important tn tablic safety and not. And I consider some 22 of thevo oe s todologies, as was characterized in one of 23 the L.go;;. sy Lawrence Livermore, as simply enhanced
( 24 QA, or I would tend to say that some of these 25 methodologies, as was, characterized by the member of the O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5906
() 1 ACES, Dr. Okrent, as a brute force procedure for 2 detecting systems in teractions.
3 JUDGE BRENNER: Dr. Joksimovich, what were you O 4 quoting from?
5 WITNESS J0KSIMOVICH: I was quoting f rom a 6 meno.
7 JUDGE BRENNER: Or paraphrasing. I'm not sure 8 if it was a quote.
9 WITNESS J0KSIMOVICHs It was a near quote. .
10 JUDGE BRENNER: That is a paraphrase.
11 WITNESS J0KSIMOVICH: I was quoting from a 12 meno from Dorothy J. Sukor, ACRS fellow, to Dr. David
( ~13 Okrent, ACRS member, 18th of May, 1980, that was part of l
' ([) 14 the package, in going through the package of information 15 that the Staff presented to us from the standpoint of 16 various bits and pieces of documentation on the subject 17 of systems interactions.
18 JUDGE BRENNER: That is not part of what 19 you're going to move in, is it? I haven't seen it, M r.
20 Rawson.
l 21 MR. RAWSON No, sir.
22 BY MR. RAWSON: (Resuming) 23 Q Dr. Jak sim ovich , you mentioned the ACRS. Are 24 you also aware of the statement within the last month by 25 the ACRS concerning probabilistic risk assessment and
)
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5907
() 1 its utility in the licensing process?
)
2 A (WITNESS J0KSIMOVICH) I was an invited 3 presenter at the ACRS Subcommittee meeting last year 4 where a nuater of us, including Dr. Burns, responded to 5 the ACRS invitation to present our views on how PRA 6 should be employed in licensing. And at that session 7 and some other sessions I attended, I.have had ample 8 opportunities to see what the ACRS views are.
9 Q Can you be more specific, Dr. Joksimovich? Do h3 Jou know specifically what has been said by the ACRS 11 after those meetings that you've talked about?
12 JUDGE NORRISa Excuse me just a moment. I 13 thought, Mr. Rawson, you asked whether Dr. Joksimovich.
14 was familiar with a communication from the ACRS in the 15 last month. Am I correct in understanding that -
16 question?
17 NR. RAWSONa Yes, sir, that was my question.
18 JUDGE NORRIS: Did you get an answer to that 19 question?
20 HR. RAWSON No, sir, I don't think I've 1 21 gotten an answer to that question. That is what I'm 22 trying to follow up on now.
23 WITNESS J0KSIMOVICH: Okay, maybe I didn't
) 24 hear it well. Could you repeat it? In the last month?
25 BY MR. RAWSONa (Resuming)
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
5908 ,
() 1 0 Are you aware of a specific statement by the 2 ACRS within the last month on the subject of 3 probabilistic risk assessment and its usefulness in the 4 hearing process?
5 A (WITNESS J0KSIMOVICH) This month ?
G Q I was talking about the last month, 30 days.
7 JUDGE BRENNER: Dr. Joksimovich, we're trying 8 to be efficient here. I think you're being picky.
9 Could you snswer that question in the c.ontext of the 10 immediate past?
11 WITNESS J0KSIMOVICH: If there is no "s" on 12 the end of that, I am familiar with the letter.
. 13 JUDGE BRENNER: Why don't you state what 14 you're familiar with and what time frame, and we will 15 find out if we're talking about the same. And then Mr.
16 Rawson can also be more efficient by pointing to it and 17 asking you if you are familiar with it.
18 In fact, why don't we adopt that, since you 19 think there is an ambiguity in the time frame.
20 WITNESS JOKSIMOVICH: My familiarity with the 21 ACRS --
22 JUDGE BRENNER: Wait a minute.
23 Ask the question the other way, Mr. Rawson.
l
() 24 Tell him what you're thinking about and ask him if he's 25 familiar with it.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W , WASHINGTON, D.C. 20024 (202) 554-2345
5909
() 1 BY HR. RAWSON: (Resuming) 2 0 Dr. Joksimovich, there was a recent statement j 3 by the ACRS in the context of the safety goals that are O 4 being explored by the Commission, and in that letter 5 there were several statements concerning probabilistic 6 risk assessment and the usefulness of PRA in the 7 licensing process. Are you familiar with that 8 ' document?
9 A (WITNESS J0KSIMOVICN) No, I'm not.
10 JUDGE BRENNERa What is the date of that?
11 HR. RAWSON: Judge Brenner, I don't have the 12 date of tha t specific letter with me.
13 HR. ELLIS: Why don't we just show him the O 14 document and see if he knows if.he's familiar with it?
I 15 MR. RAWSON: Judge Brenner, I just want to .
16 mention, I was not planning on getting into this 17 document. I do not have a copy of the document with 18 me. I was surprised by Dr. Joksimovich's comments and 19 decided to explore it at this point.
l 20 JUDGE BRENNER: Well, it sounds interesting to 21 us, too. But the problem is, we don't know what you're 22 talking about yet. Can you identify it again as best 23 you can? And I think Mr. Ellis' suggestion is a good
() 24 one, also. He may be familiar with it and may be 25 misunderstanding your identification.
O ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5910
() 1 MR. RAWSON: Yes, sir. It was a recent letter 2 by the ACRS in which different ACRS members expressed
( 3 views on the safety goals program being pursued by the l
l 4 Commission at this time, and in that letter there were l
5 statements concerning probabilistic risk assessment and 6 its role in the licensing process.
7 JUDGE BRENNER: Who was the letter to, and who 8 signed it, and is there a date on it?
9 MR. RAWSON: The letter was from Dr. Shewmon 10 to the Commission.
11 JUDGE BRENNER: And the ACRS did not rut.a 12 date on the letter? .
13 MR. RAWSON: I'm sure they did, Judge O 14 Brenner. I don't have the data available to me and I 15 don't have the document with me. -
l 16 JUDGE BRENNERs All right, I will let you 17 proceed. It's very difficult to ask questions like that 18 without a dccument around.
19 MR. RAWSON: As I said, Judge, I did not 20 expect to be asking that qur-stion. Dr. Joksimovich has 21 indicated, I think, that he is not aware of that on the 22 basis of my description.
23 JUDGE BRENNERa I 'm not going to give that any 24 weight because it is just not adequate as compared to 25 showing him the document. But proceed.
}
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W. WASHINGTON, D.C. 20024 (202) 554-2345
5911
() 1 MR. RAWSON: I understand, Judge. Thank you.
2 JUDGE BRENNER4 In addition, Mr. Rawson, the 3 reason it's important -- we're not being picky -- he may -
4 well be familiar with the views expressed there, even 5 though not in that form, and that would be important 6 also. So I hope somebody can find this epistle from the 7 ACRS.
8 HR. RAWSON: We will find it, Judge, yes.
9 Thank you. .
10 BY HR. RAWSON: (Resuming) 11 0 Dr . Joksimovich, you're not sugge sting, are 12 you, that systems interactions analysis using .
13 methodologies other than event tree-fault tree is 14 rendered unnecessary by PRA, are you?
15 A (WITNESS J0KSIMOVICH) I am not suggesting l
16 that they are necessarily. I am suggesting that the PRA 17 is the best, the most efficient way of handling the 18 issue, as I stated in the testimony.
19 Q Would you agree with me that qualitative -
20 professional judgment is still an important part of l 21 systems interactions analysis, despite the availability 22 of probabilistic risk assessment techniques?
23 A (WITNESS J0KSIMOVICH) I would never deny 24 that. I think that the judgments are important. I 25 think they provide a high degree of confidence. But I O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5912
() I think the PRA provides tF e highest that we have in 2 1982.
3 0 Gentlemen, let me refer you to page 95 of your 4 prefiled testimony. I would like to have you read for 5 the record the sentence near the bottom of the page, the 6 sixth line from the bottom, beginning "The PRA." Dr.
7 Burns, would you do that for me, please?
8 A (WITNESS BURNS) "The PRA, on the other hand, 9 has a much broader charter, since it must identify to failure modes of systet; and multiple systems, evaluate 11 these qualitatively and quantitatively, and assess 12 public risk associated with such failures. However, in I
l 13 order for the PRA to effectively accomplish its purposes O 14 it must treat the systems interactions issue in a 15 thorough manner." -
16 0 Now, Dr. Burns, I'm interested in knowing 17 better what you meant by the term " failure modes." Does I
18 the term "f ailure modes" that you 've used in tha t 19 sentence include the degradation of the safety function 20 as well as the failure of the safety function?
21 A (WITNESS BURNS) Partial failures are always 22 troublesome to deal with in any methodology that exists 23 now or probably will exist in the future. In the pa s t 24 and in fact in the Shoreham PRA, partial failures are 25 treated, partial f ailures or degradation of systems, are ALDERSON REPORTING COMPANY,INC, 400 VIRG:NIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5913 O 4 2eaer 11r tre ted com=1ete fei1ere - So if enere is l
2 an identified degradation of a system that would impact ;
3 on its ability to perform its function, it is treated as 4 a failure.
5 Examples of tF.at would be instances that have 6 occurred in the industr} where violation of technical 7 specifications have occurred or components have operated 8 far outside their design operating ranges. Those are
. 9 considered failures in the data base that is used to ,
10 characterize the quantifica tion of the fault tree logic 11 models.
12 0 Let me see if I can pursue that a little.
13 Would you refer to page 4 of the Conran memorandum, a O
14 copy of which was provided to you, an extra copy of l
15 which was provided to you yesterday. I would like to 16 have you read for the record the first full sentence 17 after the heading, heading 2.A in the center of the 18 page. Would you do that for me, please?
19 JUDGE BRENNERs Mr. Ranson, it might be a good 20 idea to include the heading context, if you have no 21 problem with that.
22 MR. RAWSON: No problem at all, Judge 23 Brenner.
24 WITNESS BURNSa " Definition of SI and 25 Application of Single Failure Criteria. PASNY and the O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5914
() 1 SI Staff agreed explicitly that the threshold for 2 identification of adverse SI's will be a non-safety
~
3 system or component f ailure that leads to the defeat of 4 one train of a safety system or engineered safety 5 feature . . . even if the remaining trains of the 6 affected ssfety system or ESF could perform the intended 7 safety function."
8 BY MR. RkWSON: (Resuming)
. 9 0 Now, Dr. Burns, you may have gotten close to 10 this with your last answer, but let me ask you in these 11 terms: Does the methodology of the Shoreham 12 probabilistic risk assessment identify systems 13 interactions which only lead to the defeat of one train O 14 of the safety system or an engineered safety feature, 15 even if the remaining trains of the affected safety 16 system or engineered safety feature could perform the 17 intended safety function?
18 A (WITNESS BURNS) Yes, sir.
19 Q In your opinion, does it highligh t such 20 systems interactions if they do not result in core 21 damage or core vulnerability?
22 A (WITNESS BURNS) They are highlighted to 23 myself. Now, what we have used thus far in the 24 development of the Shoreham PRA for evaluation has been 25 the use of a criteria to try to filter out certain O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554 2345
5915
() 1 things that people may consider important or not 2 important. And the criteria that we have used thus far 3 has been twofolds 4 One is a numerical criteria, and that would be 5 in the quantification of fault trees and event trees to 6 highlight those items that are dominant contributors to 7 risk or unusual or important contributors.
8 In addition to that, we have done a 9 q ualita tive analysis of what are the pathways through 10 the sequences, what are the components and systems that 11 do contribute to these failures. In other words, we are 12 summing a large number of failures of many components or 13 many dependencies together to get the sequence, so we O'/ 14 have to break down the sequence. And that qualitative 15 evaluation has thus far been done by SAI, but it is not 16 to date documented to LILCO.
17 So there are two criteria that we have used 18 thus far. If there was another criteria -- in other 19 Words, if the NRC or LILCO or anyone else wished to 20 apply another criteria, about which I am not aware; I 21 don 't know of any other criteria that is available today 22 --
then we could compare our results versus that 23 criteria. But to date the criteria ve have used are 24 those two.
25 0 You said that it highlighted them to you.
O ALDERSCN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554-2345
5916
) 1 Does it highlight them in such a way that preventive or 2 corrective action can be taken with respect to the 3 partial failures that you spoke about, rather than the 4 entire sequence?
5 A (WITNESS BURNS) Well, unfortunately the 6 nuclear power plant is such a complex machine th a t there 7 are virtually millions of these pathways through the 8 sequences. And what we have chosen to do in a 9 qualitative sense is to isolate thpse which have a 10 minimum number of components required to lead to an 11 unacceptable state.
12 . In other words, for those sequences in which f~ 13 possibly ten things have to happen, we choose either to 14 define, is that a high probability tha t's importa nt to 15 us, or if it is a low probability and it is ten things 16 independent tha t have to happen, we choose at the 17 present time to neglect that as not a contributor. If 18 there is either very few items that are required to 19 impact the sequence -- for example, a dependent failure 20 that affects multiple systems and then possibly an 21 additional, one additional failure -- in other words, if 22 I had a double, what is referred to as a double -- that 23 is important irregardless, it is important to me n
(_/ 24 irregardless of what the probability that one would 25 assign to that.
i
(~J1 r
i ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
5917 J
O 1 So that is highlighted in my mind and will be 2 highlighted to LILCO. And we have also looked at 3 summations of events that would ha ve high probability, O 4 even though there are many events in it.
5 i
6 I
7 8
9 ,
10 i
j 11 I 12 ,
13 14 15 16 l 17 18 f
l 19 I
20 21 22 23 O 24
+
25 l
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5918
() 1 0 Do you agree though that there are many 2 potential systems interactions which do not reach the 3 stage cf core damage or core vulnerability and that 4 these are not highlighted as such as the present time?
5 A (WITNESS BURNS) I guess I would have to agree 6 with you that we try to highlight as many as we can and 7 we have a criteria. If we had another criteria that 8 said something dif f e re nt --
9 (Witnesses conferring.)
10 0 I'm sorry, Mr. Burns. I am waiting. Had you 11 completed that thought? I think you said "if I had
. 12 another criteria."
13 A (WITNESS BURNS) If I had another criteria 14 tha t would possibly be much more limiting, I could 15 generate these thousands or millions of potential 16 systems interactions and put those under t'tose 17 criteria. We have not done that, no.
18 JUDGE BRENNER: Mr. Rawson, I am having
! 19
- trouble understaniing what you and the witness mean by
! 20 " highlight" and I don't know if you mean the same thing 21 and even if you mean the same thing, I don't know what 22 that is. I sense, Dr. Burns, that the way you are using 23 it is whether it is included in the written results or 24 not.
25 But you can each tell me what you mean by O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5919
() 1 " highlight" and I will figure out if that changes the 2 answers.
3 WITNESS BURNS Yes, sir. A highlighting to 4 me means that we as an organization make known to LILCO 5 specific items that we have found in our analysis that 6 could be important in their perception of what the risk ,
7 or safety of the plant is.
8 JUDGE BRENNER: And, of course, even within 9 that definition there would be priorities of that which 10 you might emphasize, depending upon your judgment?
11 WITNESS BURNS: Yes, sir.
12 JUDGE BREN NER: Is that what you meant by 13 " highlight?"
- i O 14 MR. RAWSON: I think it is very close, Judge 15 Brenner. There was some testimony yeste rd a y, I think, 16 by Dr. Joksimovich about the probability was 17 sufficiently low to be in the noise band as esposed to 18 those sequences which would stick out, and I use the 19 term " highlight." I an interested in the latter 20 concept -- things that either stick out or can be
! 21 identified through the analysis that has been done.
22 JUDGE BRENNER I will let Dr. Burns comment 23 after I have made a comment. I think you are at a
( 24 dif f erent point in the analysis than Dr. Joksimovich 25 was. I'was asking Dr. Joksimovich about his end point
()
j ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASillNGTON. D.C. 20024 (202) 554-2345
5920 O i end I wes ineu1rine ebout in sumpine to the conc 1 ston 2 of finding ultimate risk outlyers, would you skip over 3 possible systems interaction tha t didn 't show up in that 4 very end point.
5 And he answered my questions along that line.
6 I think what you are now asking about with Dr. Burn,s is 7 the beginning of the analysis. That is -- well, I don't 8 know about the very beginning, but earlier in the 9 analysis where you are making the threshold 10 determination of contributions to risk not necessarily 11 so.significant that they show up in the end result but j 12 sufficient to do the further rigorous analyses as 1
13 distinguished from 1eaving ther aside on a prioritized CN 14 basis.
15 Dr. Burns, is my distinction correct between 16 the other testimony and what we're talking about now?
17 WITNESS BURNS: Yes. I think it is a very 18 good analogy. I don't want to mislead anyone. There l 19 are two aspects. One, we don't throw out items of l
20 importance in the quantification, but there is a cutoff l 21 in the qualitative presentation of those results. I'm 22 not sure if I hava made this clear in the quantification l 23 or the identification of a frequency of potential core 24 melt.
25 We don't actually throw anything away when we ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5921
() 1 present what paths can lead to tha t sta te. We are 2 forced to truncate at some point just to the sheer 3 amount of paper that is involved in the presentation of 4 that amount, so that there are two phases to the program.
5 One is a quantification which we don't 6 truncate, and then one is the qualitative , presentation 7 of the results in which just from the ability to deal 8 with the paper we have to limit the size of the problem.
, 9 JUDGE BRENNER: Do you mean to use your 10 academic example of a system where ten things have to 11 happen in order to lead to a core vulnerable condition, 12 and I assume you meant ten non-dependent things. Is 13 that correct?
- 14 WITNESS BURNS: Yes, sir.
15 JUDGE BRENNER: As to that, you would still 16 include it in the quantification?
17 WITNESS BURNS: Yes, sir.
18 JUDGE BRENNER: All right. That is different, 19 possibly, than some other PRAs, correct?
20 WITNESS BURNS: To my knowledge, IREP does 21 make some judgments about quantification by eliminating 22 some sequences, yes, sir.
23 JUDGE BRENNER: Well, I don't want to get into
) 24 it, but since you are familiar with it, doesn't the 25 Limerick PRA make some threshold judgments also before O
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5922 including things in the quantification?
(]) 1 2 WITNESS BURNSa In the presentation of 3 results, yes, sir.
4 JUDGE BRENNERa I mean in the quantification S and analysis before the presentation of results.
6 WITNESS BURNS Maybe 7 shouldn't be too 7 strong. Maybe I was too strong before. I guess when 8 the probability is very lov those types of things we may 9 not even be able to predict and so in that sense they to may not be included. If there are many, many things 11 that have to happen 'and the probability is very low, 12 those things are probably not included in the analysis.
13 JUDGE BRENNERa For example, would you include 14 a cracking of the pressure vessel?
15 WITNESS BURNS: I would include that as a very 16 important item that should be evaluated, yes, sir.
17 JUDGE BRENNERa Okay. I don't know if that 18 helped.
19 MR. RAWSON: Yes, sir. It helped considerably
- 20 and the point I wanted to explore I think has been made l
l 21 adequately. I have no further cross examination and I l
22 will pursue -- endeavor to track down the document that l
23 we have been talking about.
() 24 I'm sorry, Judge. May I have one minute to 25 consult, please?
O
! ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE . S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5923
() 1 JUDGE BRENNER: Sure.
2 (Counsel for NRC conf erring. )
l 3 MR. RAWSON: Thank you, Judge. I have nothing 4 further.
5 JUDGE BRENNER: Well, I'm surprised the other 6 way this tire. Tha t is, it is certainly shorter cross 7 examination than I would have predicted. There is a 8 lot -- and I emphasize a lot -- of stuff in some of the 9 Staff's documents that at least is apparently not fully to consistent with the emphasis and approach that these 11 witnesses have placed on the use of the PRA relative to 12 the other approach.
13
- I am not asserting there is an inconsistency 14 on the bottom line. I don't know. And I would have 15 expected that to set up the Staff's testimony, so to 16 speak, you would have gotten the views of these 17 witnesses on many of those very similar to what your 18 exploration of the one excerpt from page four -- that 19 there are many statements that are a lot longer than 20 that one that you picked.
21 And one thing on the Board's mind -- and I was 22 hoping cross examination would help set that up -- is 20 what knowledge, if any, the Staff has of the work to
() 24 date on the PRA here, because it is one thing to say we 25 haven't required it, but it is another thing to put O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASH;NGTON. D C. 20024 (202) 554 2345
5924
() 1 yourself in the reality of the schedule of this hearing 2 and the license of this plant when these witnesses -
3 disclose some preliminary results and, for all I know, 4 in the record so far -- and I'm certainly not saying 5 that's the case -- but for all I know, in the record so 6 far, the Staff,is ignorant of those results to date.
7 Whether or not that is the case or not, I 8 don 't know, and whether or not that matters or not, I 9 don't know. But that is the problem.
10 MR. RAWSON: Judge, with respect to that 11 specific point, va have our witnesses ready to go on the 12 stand, and they are available for whatever questioning 13 that the Board wants to put to them, and we would prefer 14 to do it in that fashion.
15 JUDGE BRENNER: I know, but we have been doing i
16 a lot more cross examination here than we had intended, 17 and partially -- well, I don't know what the reason is, 18 and'I guess we will just have to do it again. But it is 19 usually not the normal order in which hearings work.
20 Well, very well, we will consider what 21 questions you want to ask based on the documents and so 22 on.
23 Er. Ellis, redirect?
() 24 MR. REISs Mr. Chairman, I think the Staff's 25 limited cross examination has to be considered in O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5925
() 1 connection with other statements of the Commission and 2 where we are going on probabilistic risk assessment, 3 particularly the statements in the proposed policy 4 statement on proposed goals for nuclear power plants 5 issued on February 17, '82, at 47 Federal Register 7023, 6 and I am placing them in the present context.
l 7 Ihe Staff, by taking this position and not 8 more deeply cross examining, just is following what it 9 believes is the lead there, that you cannot rely on PRA 10 in licensing, and I think the Commission has said it.
11 And for that reason we don't feel we have to go into it.
12 JUDGE BRENNER Well, that is inconsistent 13 with a lot of things I know that the. Commission has 0 14 issued. Tha t's number one. It is even inconsistent 15 with some things the Staff has said elsewhere -- as 16 broad as you have stated it, at any rate.
17 If you wanted to add the word " solely," that 18 would be something else. Aside from that, I can't write 19 a finding because of what you just asserted. The way to 20 get to that finding is to understand the importance in 21 terms of the safety of this plant. I'm not interested 22 in Federal Register notices. We have ruled on what 23 issues are admissible in this proceeding and this issue
) 24 was admitted. We are here litigating this issue.
25 Nobody appealed that action by us and your O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5926
'( ) 1 statement doesn't assist us in litigating the issue and
)
2 I will leave it at that. That might have been a legal ,
i 3 argument six months ago, but it is not a reason here for 4 not joining issue. It's also an argument we would have 5 disagreed with.
,6 I will tell you right now tha t there is a big 7 void in the record as far as the Staff's approach in 8 those written m.esos that the Staff plans on moving into 9 evidence and the views of these witnesses. I am not 10 saying that that void will be a problem for making a 11 decision on the plant because we have a lot of testimony 12 from LILCO and presumably there vill be more in terms of 13 questions.
O 14 But we are here on what everbody knows is a 15 very important issue in this proceeding and I think a 16 lot more questions could have been asked which the County understandably didn't give in their focus and yet
( 17
(
- i 18 the Staf f drops many, many pages of these memos on us 19 and there they sit. And we can't put it together unless 20 questions are asked about it, and the fact that the .
21 Staff's witnesses are going to testify about those 22 documents will be very helpful to get the Staff's 23 witnesses view, but it doesn 't help joint issues when we
() 24 have these experts here available to comment on it also.
25 I am also not sure that the fact that the ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W., WASHINGTON, D.C. 20024 (202) 554 2345 l
5927
() 1 Staff on its own did not ask for a PRA is as pertinent 2 now that we have all of this information available. In 3 addition, I can tell you one of the questions I was 4 planning on asking the Staff's witnesses is what 5 considerations led them not to consider looking at the 6 available results to date of the PRA or even what 7 considerations led them not to request some PRA work, 8 and preliminary to tha t, how close the Staf f came to 9 considering whether or not to request some PRA work for 10 Shorehan similar to Limerick, because there is some 11 history on that also that I expect to get on the record.
12 And the time frame of LILCO's initiating their 13 PRA is arguably consistent with some of that history.
14 In addition, I believe, Hr. Reis -- I don't 15 have the Federal Register notice in front of me, 16 although I am somewhat familiar with it -- I think if 17 you take a look, the context is very precisely drawn to 18 not depend solely on PRAs to make regulations less 19 conservative. That is, not to use, because of the 20 uncertainties in PRA not to use it to relax any safety 21 standards based on deterministic judgments.
l 22 Ihat is wholly different than the context in 23 which you offered your comment and wholly different than l
() 24 the context of my comments -- that is, what use we 25 should make of it to assist in the exploration of l
l
()
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5928
() 1 whether or not the systems interactions and sa.ety l
2 classifications hs ve been properly accounted for. And 3 ve have had a lot of work done that these witnesses are 4 familiar with and I just would have thought the Staff 5 would take advantage of that, both outside the hearing 6 and in the hearing. And the record doesn't reflect 7 whether or not that has been done.
8 HR. REIS: Mr. Chairman, my point was that 9 there are bases to license plants without PRAs at the 10 present time.
11 JUDGE BRENNER4 So what? Nobody is contesting 12 that.
i 13 MR. REISa That is what the Staff has looked O 14 at, and that is what the Staff's review has been. The 15 Staff can only testify as to what their review has been 16 and what they focused on. That was to look at the 17 ordinary determinative factors. listed in Part 50 and the 18 appendices to Part 50. It did not go to things not yet 19 required of them, and that was the basis of my comments.
20 And I think in this policy statement 21 indicating that certain things might change in the 1
l 22 future was the. Commission's opinion that plants may 23 still be licensed on a deterministic basis.
() 24 JUDGE BRENNER: I don't want to belabor it, 25 but your comment was misleading because it took the O
ALDERSON REPORTING COMPANY,INC, ,
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5929 I
() 1 Commission's statement as to not depending uoon PRAs out j l
2 of context, at least the context that I indicated.
3 Further, in admitting this Contention we O 4 issued a written letter that discusses the Commission's 5 history somewhat on PRA and we've got the issue here 6 which we admitted and we want to explore it.
7 Ihe fact that the Staff did not require a PRA 8 and the fact that the Staff now you tell us -- and 9 presumably we will find out from the witnesses -- has 10 looked only at the normal review and has not taken 11 sdvantage of the other information which is available 12 and has been available for some months now and has been 13 highligh ted , if you will, to use the term, in O 14 preparation for this hearing is just somewhat 15 distressing.
16 We have got a contested issue here and we 17 could use the help of all parties in resolving it.
18 MR. REIS Mr. Chairman.
19 20 21 22 23 24 25 O
ALDERSON REPORTING COMPANY,INC, ,
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
I 5930
() 1 MR. BRENNER: I don't think we are assisting 2 the resolution of it by pursuing this. At least I don't 3 plan to pursue it with you any further. But I will give 4 you an opportunity for this further comment if you wish.
5 3R. REIS: Yes. The only one comment the 6 staff wants to make is the staff all the way through 7 this proceeding is concerned with expedition and we 8 don 't intend to put on things that we don' t think are 9 necessary for licensing to move it along. We intend to 10 only stick to what we think is necessary.
11 We are concerned with expedition and we are 12 concerned with the policy statements of the Commission
- 13 on expedition. Are are concerned with the resources of
# We have had here people sitting for a week 14 the staff.
15 now, sitting around waiting to get on the stand and to 16 testify and put the staff's position forward on what we 17 thought were matters that were not as pertinent as 18 others, looking at this present state of the l 19 Commission's regulations.
20 JUDGE BRENNER. We are all in favor of 21 expedition, but that doesn't mean to the point of l 22 leaving large areas that should be explored unexplored, 23 and I think the staff has drawn the line so far the 1 24 other vsy as to this particular matter as to be 25 cou n ter-pro d uc tive to expedition. It is useless to O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGT N D.C. 20024 (202) 554-2345
5931 0 ' expeatte the ae riao aa thea eaa av t the eaa ita aio
- 2 voids in the record, so there is always a tough decision 3 to make, and we will consider what questions we want to 4 ask now.
5 IBoard conferring.)
6 I will leave you with one last comment and 7 then we will take a combination lunch break and cooling 8 off break. If you start out with the proposition that 9 the PRA in this case is not very useful, the way to 10 astablish that is not f or you as counsel, Mr. Reis, to -
11 tell me that. The way to establish that is through the 12 cross examination that I thought would be conducted, and 13 I rill leave it at that.
14 We will come back at 1:15.
- 15 [Whereupon, at 12
- 00 noon, the hearing was 16 recessed, to reconvene at 1415 p.m. the same day.]
17 l 18 19 20 -
21 1
> 22 23 24
! 25 O
ALDERSON REPORTING COMPANY,INC, 400 VIAGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5932 AFTERNOON SESSION
(]) 1
- 2 [1:15 p.m.1 3 JUDGE BRENNER. We are ready to proceed, I 4 guess, with redirect by LILCO.
5 MR. EARLEY: Judge Brenner, if I may add one 6 thing before we get into redirect, in your mailbox is a 7 resolution of Suffolk County Contention 28 A-4, SOC 8 Contention 7A-4, SPDS. When you have had a chance to 9 review that, if you have any questions about it I think 10 Mr. Irwin and Mr. Lanpher were working on that and ther 11 will be available.
12 JUDGE BRENNER4 All right. We won't have 13 anything today on it. We might as soon as tomorrow 14 morning. I don't know. If we are ready this week, we 16 will do it at the beginning of either Thursday or 16 Friday. That is the one that I had settlement pending i 17 noted. Perhaps at the time we handle that one we can 18 get a status report on some of the others also on which 19 there has been a great deal of silence the last two 20 weeks and we haven't asked, but we will when we take up l
21 this one, including Suffolk County 9, ECCS Pump 22 Blockage, which va have put aside vsiting to hear more 23 about, and I guess we will ask about some of the other 24 ones coming up, too.
25 Whereupon, l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5933
() 1 V0 GIN J0KSIHOVICH, 2 EDWARD T. BURNS and 3 ROBERT H. KASCSAK, 4 the witnesses on the stand at the time of recess, 5 resumed the stand and testified further as follows:
6 -REDIRECT EXAMINATION 7 BY HR. ELLIS:
8 0 Dr. Joksimovich, you were asked a number of 9 questions about your review of the Shoreham PHA in your 10 capacity as,a reviewer. Could you give us an estimate 11 of the amount of time you spent as a reviewer reviewing 12 the Shoreham PRA?
13 A (WIINESS J0KSIMOVICH) Yes, I can. My 14 secretary tells se that I have spent over 400 hours0.00463 days <br />0.111 hours <br />6.613757e-4 weeks <br />1.522e-4 months <br /> 15 outside the time dedicated to preparation of the 16 testimony and the time being expended in this 17 courtroom. In addition to that, I was supported by 18 members of my staff who are specialists in various areas 19 of PRA.
20 0 In connection with the testimony that you 21 submitted in this contention, has it been reviewed by 22 any of the other reviewers?
23 A (WIINESS-J0KSIMOVICH) Yes. Professor
() 24 Rasmussen has ' reviewed my 'testi$ony and expressed f ull 25 agreement with the text without any reservations.
~,
l ALDERSON REPORTING COMPANY,INC, 400 V!AGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5934
, () 1 0 Dr. Burns, do you have the transcript? Let me 2 hand you the transcript -- we have one copy of it -- and 3 direct your attention to page 5757 of the transcript, 4 and the answer that begins on line 4 and ends on line 5 10, and I would ask you to read that, if you would.
6 For the Board's information, this was a 7 question and answer that I had some concern about 8 y e s te rd a y , and it was the occasion for my asking the 9 court reporter whether it could be read back, and I 10 never did ask him to mark it. -
11 A (WITNESS BURNS) This was a question regarding 12 s comparison of the Florida utility situation versus 13 LILCO situation in terms of offsite power connections, .
O 14 and I believe that I misspoke when I said W ASH-1400 has 15 multiple lines into the plant. What I meant to say was 16 Shoreham has multiple lines into the plant and they are 17 fed from several sites with gas turbines with black 18 start capability.
19 0 While we are on this subject, Mr. Kascsak, I 20 think you were asked a question concerning LILCO's I 21 interconnect capabilities. Do you hat? additional 22 information on that at this time?
23 A (WITNESS KASCSAK) Yes, 1 do. I would like to
() 24 be a little more complete in response to that question.
25 LILCO has a total capacity of 1390 megawatts O
ALDERSON REPORTING COMPANY. INC, 400 VIRCINI A AVE., S.W . WAShlNGTON. D.C. 20024 (202) 554-2345
5935
() I worth of intertie capability that exists bo th with the 2 Northeastern Power Utility Grid System and tP s New York 3 State Power Pool Grid System.
4 JUDGE JORDAN What was the figure?
5 WITNESS KASCSAKa 1390 megawatts, of which 6 there are four separate interties. We have one intertie 7 at our Northport station, which consists of a 8 300-megawatt intertie to Connecticut to the Norwalk 9 Harbor interconnection, 138 KY, and we have three 10 additional interties through the Con Edison system into 11 the New York State Power Pool Orid, and those three 12 interties, one is located a t Glenwood Landing, which is 13 called the Dunwoodie tie, and interties us directly into O 14 the Westchester Grid independent of New York City. And 15 tha t is a 580-megawatt intertie.
16 Then we have two additional interties, one at 17 Lake Success and one at Valley Stream, both 138 KV. One 18 is 240 megawatts and~the other one is 270 megawatts, 19 both tying in to the Con Ed system, which gives us e 20 total of 1390 mega wa tts of intertie capability.
21 I think that should clarify that point.
22 0 Dr. Burns and Joksimovich, this contention 23 concerns systems interaction, systems classification and 24 the relationship. Let's take each one one at a time.
25 Does the PRA methodology employed in the ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
I 5936
() 1 Shoreham PRA, is it intended to detect and does it 2 detect systems interactions?
3 A (WITNESS J0KSIMOVICH) Yes, it does, as was 4 stated on page 75 of my testimony where I think it is 5 stated that the issue is addressed by virtue of 6 constructing and quantifying the plant event fault 7 trees, which include the treatment of common cause 8 initiating events, intersystes dependencies which could 9 be functionally shared equipment, physical interactions 10 or human interactions and intercomponent dependencies, 11 and Dr. Burns has provided numerous examples in Table 12 3-1, and I will let him explain that.
13 0 Well, without reviewing the testimony we have 14 already gone through, have you described in your 15 prefiled testimony and in your cross examination, Dr.
16 Burns, how the Shoreham PHA is intended to and does 17 detect systems interactions?
l 18 A (WITNESS BURNS) Yes, sir, I believe so.
19 Q All right. Now, given that the Shoreham PRA 20 is intended to and does detect systems interaction, does l
21 this relate in any way to systems classification? Dr.
22 Joksimovich?
23 A (WITNESS J0KSIMOVICH) Yes, it could.
() 24 0 could you explain how?
25 A (WITNESS J0KSIMOVICH) An example that I have O
ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5937
() 1 in my mind is that if we were to have identified what I 2 have termed previously a risk outlier, and that 3 typically would have been an accident sequence, and you 4 can use all sorts of letters, "t," "u," "v," whatever, 5 and if that risk outlier contained a system which had 6 relatively poor reliability, then we would suggest that 7 the reliability of that system be upgraded.
8 And one way of upgrading the system would be 9 to raise it froz, if it was previously a nonsafety to grade; it could become a safety grade. The alternative 11 would be simply to use the reliability engineering 12 techniques and provide enhanced reliability without 13 paying full attention to the specification.
14 0 Well, does the Shoreham PR A methodology look 15 only at systems interactions that are dominant 16 contributors to risk or risk outliers? Are there other 17 systems interactions that the Shoreham PRA looks at 18 other than the risk outliers or dominant contributors to 19 risk?
20 A (WITNESS JOKSIMOVICH) We have focused all of 21 our attention on the dominant risk contributors for a 22 very good reason, because they directly might impact the 23 public safety. Now, there are various other levels that
() 24 one can talk about, and tha t is that in a typical PR A 25 study like Shoreham, we address numerous accident O
i ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5938
() 1 sequences that are in the hundreds and thousands, and 2 each of those are documented in the study.
3 Beyon that, we document every event tree, and 4 as I pointed out, there are 19 of them in the study and 5 there are ten system level fault trees, and in the fault 6 trees we have cut sets which point out numerous systems 7 interactions which are documented for the sake of 8 documentation and completion.
9 0 So the Shoreham PBA then does look at systems 10 interactions apart from those that might be in a 11 sequence that would be a dominant contributor to risk?
12 A (WITNESS J0KSIMOVICH) It does a great deal.
13 It doesn't receive- the.same degree of attention as the O 14 dominant sequences do, and that is why we f ocused on 15 that before.
16 (Counsel for LILCO conferring.]
17 Q Dr. Burns, in connection with Dr.
18 Joksimovich's answer that the Shoreham PRA detects 19 systems interactions other than those that might be 20 dominant contributors to risk, I take it -- well, are 21 these other systems interactions reviewed by your 22 analysts or by you?
23 A (WITNESS BURNS) In the construction of the
() 24 logic models that are used to characterize the plant and 25 represent the plant, we go down to the component level O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5939
() 1 in all of the systems and identify what f un ction al 2 dependencies occur among systems and human interactions 3 among systems and spatial interactions that we can 4 identify among the systems.
- 5 The interactions that we identify are then l 6 reproduced on evaluation both on a system level and on a 7 sequence level. So individual systems are evaluated for i
l 8 potential systems interactions that may occur at the 9 individual systems and the individual trains among 10 within a system.
11 In addition to that, we also evaluate what 12 minimum number of potential interactions could defeat 13 multiple systems in accident sequences. So we look at i
14 both from a safety system reliability stand point and a 15 sequence reliability standpoint what are the systems 16 interactions that do lead to failure or could lead to
- 17 failure of systems and to potential sequences.
18 0 Dr. Burns, you heard Dr. Joksimovich's l 19 testimony concerning the way in which the Shoreham PBA's l
20 detection of systems interactions could be related to i 21 classification. Did you concur with that?
22 A (WITNESS BURNS) Yes, we identify the 23 interactions and then their potential impact on safety.
() 24 And then if there was in fact an unacceptable 25 interaction, then it could be reflected in a change in O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5940 O i the safety =1aseteication of the system.
2 0 Did you or Dr. Joksimovich, either of you in 3 your performance of the PR A or in your review of the 4 PRA, Dr. Joksimovich, did you uncover any unacceptable 5 systems interactions that might affect systems 6 classification?
7 A (WITNESS J0KSIMOVICH) N o, we did not.
8 0 Dr. Burns, is that your answer as well?
9 A (WITNESS BURNS) We did identify systems 10 interactions, but none which would warrant a change in 11 the classification of a system.
12 JUDGE BRENNER: Dr. Burns, on you last answer, 13 how many did you have in mind when you said you O 14 identified some systems interactions but none that would 15 varrant a change? I am not looking for an exact number, 16 just a rough number.
17 WITNESS BURNS: I don't believe I could give 18 you an exact number, but at the urging of the Board, I 19 did go through some examples. I am not sure if that 20 would be halpful for you.
21 JUDGE MORRIS: Can you give just an order of 22 magnitude at this time? Was it a half-dozen or a 23 thousand?
24 WITNESS BURNS: I would say that there are on 25 the order of -- let me clarify exactly what we are O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5941
(') 1 talking about, systems interaction in the sense that a 2 failure of one component may affect another component, 3 so it is a dependentptype f ailure and it can be in 4 something as innocuous as even one train of a safety S system we could be talking about.
6 And from that standpoint, since there are many 7 safety systems in the plant, there are then many systems 8 interactions, so I would say that we have identified on 9 the order of thousands or tens of thousands of 10 interactions of that nature, none of which of themself 11 lead to a serious concern about safety but which are 12 systems interactions as we have discussed them.
. 13 JUDGE BRENNER Did you rank in any way the 1
("%
\- 14 systems interactions that you found that would lead to, 15 relatively speaking, now, among the ones you found the 16 greatest concern, even if in the end you were not 17 concerned?
18 WITNESS BURNS: We ranked them by probability 19 and by order, and by order I mean number of elements 20 within this pathway to potential system failure or l
21 sequence failure.
l l 22 JUDGE BRENNERa Was that a combined ranking so 23 that combining those rankings there would be a list of
() 24 the top ones, or would they be two separate rankings?
25 The two ranking criteria you gave me would not O
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554 2345
i 5942 O i necessari11 he consistent foc each ite . That is hy 1 2 asked.
3 WITNESS BURNS: Right. They are not 4 necessarily identical rankings. If one is interested 5 in -- let me look a t the LPCI system and find out if 6 there are any systen interactions in that system that I 7 want to look at to improve the reliability of that one 8 system. That is one ranking. If I want to look at all 9 sequences in which LPCI occurs as a potential 10 contributor to a core vulnerable situation, then I would 11 get a ranking possibly different than just the 12 individual system reliability.
13 l
O 14 15 l
i 16 17 18 19 20 21 22 23 24 25 O
- ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5943
( 1 JUDGE BRENNER: What I have in mind is whether l 2 or not there would'be an easy way by limited example to 3 illustrate the application of your criteria or
(
4 application of your judgment and criteria on, to use the 5 vernacular, relative closest calls, so to speak, of the 6 interactions you found, and to look at those as examples 7 where you felt that they were not of any further concern 8 and therefore no changes in " classification were 9 warranted.
10 And I don't know how to express the request 11 for examples, but I don 't know if you had the ranking in 12 mind when you considered some examples. Could you help 13 me out? I don ' t wan t to get at the examples right now, g-]
l %-)
14 but I want to know if we can get there at some point.
15 VITNESS BURNSs No, sir, I didn't have that in 16 mind when I generated the examples, and I don't have the 17 rankings with me.
18 JUDGE BRENNER: Wera you going to get into the i
19 examples on redirect and the way in which Dr. Burns l 20 intended to use them?
21 MR. ELLIS: Judge Brenner, yes.- I mean, I 22 would like tr. I'm not sure that I fully understand, so i 23 you will have to bear with me because I'm not sure that I (~h N/ I understand. But I will try to ask the appropriate l 24 l
25 questions.
l l
ALDERSON REPORTING COMPANY. INC.
4C9 VIRGINIA AVE S.W WASHINGTON. D.C. 20024 (202) 554-2345
5944
() 1 JUDGE BRENNER: Well, I didn't mean you have 2 to worry about asking my questions.
l
('t 3 MR. ELLISa Well, I would like to if I can.
4 JUDGE BRENNER: Well, I will stop at this 5 point and allow your redirect to proceed, and then we 6 vill come back at the end after you have : overed what 7 you want to cover.
8 MR. ELLIS's This will probably confirm that I 9 don't understand, but let me ask it anyway.'
10 BY MR. ELLIS (Resuming) 11 0 Dr. Burns, can you give us one or two examples 12 or three examples of the systems interactions that were 13 detected, but not dominant contributors to risk?
O 14 JUDGE MORRIS I think we were focusing on the 15 two criteria he used. One was the order and the other 16 -- you tell me.
17 WITNESS BURNS: The other is the quantitative 18 probability estimate.
19 JUDGE MORRIS s Right. So with those two 20 criteria, can you think of any examples which were high 21 on the list but below the threshold, so to speak?
22 JUDGE BRENNER: While you are considering 23 that, that was my point, M r. Ellis. The reason I didn't 24 express it clearly is I was driving to try to come up 25 with a question that would fit the data he had at hand.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5945
(~)
V 1 I don't wa n t to send him back for more research at this 2 point unless it's absolutely essential.
- 3 It won't help to have an example of one that 4 wasn't even close, and then he tells me tha t he didn't 5 consider that a problem, if it is one that nobody, 6 objectively thinking, would consider a problem. I want 7 to get at the closest ones. And I did ascertain that 8 there was'no one list.
9 And also, the term " closest ones" might be 10 misleading, since we're talking relative rankings of his 11 findings.
12 Maybe we should let you respond to your 13 counsel's question with these comments in mind. You 14 stated you came up with examples at ~ the Board 's urging 15 and I guess I don't have a recollection for that anchor 16 point.
17 WITNESS BURNS: Judge Morris pointed to my 18 testimony as citing that we had identified examples of i
19 systems interactions.
20 JUDGE BRENNER: Do you mean page 104?
21 WITNESS BURNS: Yes, sir.
22 JUDGE BRENNER I guess both of us did that.
23 All right, yes. Since then I think we have had
() 24 testimony that there are literally thoucands, and we are 25 interested in the closest calls, if there is such a way b('s ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5946
{} 1 for you to give us those examples.
2 WITNESS BURNS: I think that I can. My fg 3 examples don't :11 fit that category, but I can Nl 4 certainly make ut attempt to give you a faaling for what 5 I thipi. I don't know if I would characterize them as 6 close calls, but I would characterize them as 7 contributors to those sequences that show up as being 8 po ten ti ally important.
9 JUDGE BRENNER: *About how many examples do you to have?
11 WITNESS BURNS: I have about 25 or 30 12 examples. I can give you two or three of what I would 13 call a higher --
l l 14 JUDGE BRENNER: All right. Let me get out of 15 your way and you pick ones that you think will 16 illustrate applications of your judgment and criteria, 17 or at least describe what you think those examples 18 show. And maybe that will get us closer to the point.
19 WITNESS BURNS 4 One of the items that has
, 20 turned out to be a contributor to the calculated l
21 frequency of core vulnerable condition at Shoreham has l 22 been the availability of DC power supplies, and this 23 arose out of a consideration of the fact that as an
() 24 initiator or a challenge to the systems, it could be a 25 failure of the DC buses, and in evaluating what the C^)
l l ALDERSON REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5947
() 1 consequences of that particular failure would be it 2 turns out that there have been in the past in operating 3 experience some dependent failures of battery systems, 4 even though they are in separate divisions. And this 5 has resulted from either human errors or actual physical 6 interties.
7 And one of the dominant pathways, given this 8 initiater, which would be failure of a DC bus, as 9 characterized by*the operating experience d a ta , when we 10 go through the sequences that could follow that 11 initiater there are many sequences in which the 12 remaining division remains on line and the systems that 13 remain operate as designed.
14 The one that turns out to be the highest 15 contributor in terms of probability and in terms of 16 order -- that is, number of failures that must occur --
17 is the other DC bus fails. And we have characterized 1
18 what that probability is. So those two failures can 19 lead to a problem which could place the core in a core 20 vulnerable condition, and that shows up as one of the 21 sequences that were highlighted, if you will.
l 22 JUDGE JORDAN: Before you leave 1+., then, 23 would you say that no design or operating changes were 24 indicated, or were there any design and operating 25 changes resulting from that? Not a change in
(/)
ALDERSON REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5948 I classification, but a change in either operating
(])
2 procedures, maintenance procedures, because I believe r~ 3 tha t in some of those DC bus f ailures maintenance 4 procedures were involved.
5 WITNESS BURNS: The principal contributor to 6 many of the failures was that there was a physical 7 intertie between the divisions, that was either
~
8 inadvertently operated by the operator or there was some 9 malfunc* tion that does not exist at Shoreham. So there 10 is a design difference in that respect between Shoreham 11 and between some of the older plants which the operating 12 experience characterizes. -
13 JUDGE BRENNERa I would like to stay with tha t i
14 example for a moment, if you are finished with it. Or l
15 were you ready to move on to the next example?
16 WITNESS BURNS: Yes, sir.
17 JUDGE BRENNER: Could you explain to me why 18 the PBA approach and the modeling that goes into your 19 preparation of a PRA assessment would identify that type 20 of potential problem? And I understand your testimony 21 tha t it wasn 't a problem for Shoreham, but identify that 22 type of potential problem as distinguished from the l 23 other, more traditional approaches to analyzing a l fN
(,) 24 plant.
25 What did the PRA approach add to your O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
5949 1 identification of that that not previously identified by
{~ }
2 other approaches and techniques?
3 WITNESS BURNSa It has been previously 4 identified by the NRC, but I believe -- and now my 5 knowledge is -- I can' t testif y that my knowledge is 6 completely accurate. But my belief is that it is 7 outside what people call the single failure criteria.
8 It is not really something that one -- it is an example 1 of multiple failures which is beyond the design basis of 10 the plant.
11 JUDGE BRENNER: But the failures had occurred 12 in the historical operating experience? -
13 WITNESS BURNS: Failures which have occurred, 14 but have not led to a core vulnerable condition.
15 JUDGE BRENNER: Yes. I just meant the 16 occurrence of the dependency. And the answer to that is 17 yes, that it has occurred? -
18 WITNESS BURNS: Yes, sir.
19 JUDGE BRENNER: I guess as you go through your 20 examples, I have the same sort of question, and my 21 framework -- and I will ask you generally now and you 22 can consider it as you give the examples -- is to try to 23 ascertain whether the PRA is merely displaying in some
() 24 convenient form knowledge and conclusions that already 25 exist, or whether the PR A a pproach leads to O
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202; 554 2345
5950
() 1 identification of things that might not have been 2 identified at all or might not have been as readily 3 identified through more traditional deterministic
}
4 techniques.
5 And I will let you now proceed at this point.
6 Mr. Ellis, I guess I should apologize, but we 7 a re interested in this and it seemed like a convenient 8 point. I hope it is not too diverting to your plan of 9 approach.
10 WITNESS BURNS: My second example would be 11 again dealing with power. In this case it would be AC 12 power, which would be initiated by a loss of offsite 13 power. And in this case we have multiple -- we have to 14 postulate multiple failures. However, the history of 15 experience in nuclear power plants, admittedly which is 16 of older plants but nevertheless, that experience 17 indicates that there is indeed losses of offsite pever 18 with what might be considered a relatively high 19 frequency, once in ten years, and that the other methods 20 of onsite power, which are the diesels, may have some 21 coupled dependency among the sources, even though they l 22 are designed to be separate.
23 They are subject to the same maintenance 24 procedures and the same people working on them and the 25 same basic design, and so there is some coupling. And O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
l 5951 1 in the data that is available in the literature, it 2 indicates that when you combine or couple those i
3 failures, that you can end up with an accident in which
/~T I V
4 prolonged loss of AC power to a power plant could result 5 in the case where you have -- you subject the plant to a 6 core vulnerable condition if you can't recover some AC 7 source.
8 And although this probability is very remote, 9 it turns out to be higher than other sequences that have 10 been investigated. So again, this is another one that 11 comes to the top as indicating that for this particular 12 system interaction, which we can say is among the 13 multiple f ailures in divisions due to loss of their 14 power source, in this case the diesels, that systems 15 interaction is on a relative scale among other systems 16 interactions in the plant maybe one of the more 17 important ones that one would like to focus on and see 18 if we have done everything we can to make sure that it 19 is as low as we can make it.
20 Another one that may not be --
21 JUDGE BRENNER: I wonder if I can keep you 22 with that one for a moment, the loss of AC power l 23 combined with a possible common problem to the diesels.
() 24 Now, why was that not something for which any changes, 25 either operational changes or classification changes or O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
1 l
5952 other changes, were considered warranted for Shoreham?
(]) 1 2 Dr. Burns, if you know -- and one reason I'm 3 interested in these is not just as items of what LILCO 4 did, but as illustrations of the application of the 5 judgment of the PRA analysts. And certainly any member 6 of the panel can add anything that they wish after you.
7 WITNESS BURNS: Let ma ask Mr. Kascsak to add 8 to what my perception of it is. My perception is that, 9 again, we have tried to identif y wha t are the major --
to in other words, there is always going to be some .
11 contributor to risk, no matter what level you say you 12 are at. There is going to be something that shows up as 13 contributing to that level of risks in this case, loss 14 of of fsite power, loss of power to the plant over an 15 extended period of time. Maybe I wasn't clear when I ,
16 said tha t. l 17 The Shoreham plant is designed so that it can 18 operate for a relatively long period of time without 19 either offsite or onsite AC power, becauss it does have 20 both HPCI and RCIC turbines which are run off of steam 21 from the primary system, which is available, and DC 22 power from the batteries. So for some period of time 23 until the batteries are drained, which can be a period
() 24 of 4 to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, the Shoreham plant can survive even an 25 extended loss.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20C24 (202) 554-2345
5953 But the determination that I have personally
(]} 1 2 made -- or actually, I haven't made a determination. I 3 have just presented the results to LILCO in the context O 4 that this is one of the sequences, the principal 5 sequences that contribute to the level of risk as we 6 have assessed it at Shoreham on a site-specific basis.
7 So we have not made a judgment as to whether 8 there is improvements required or not, only that this is 9 the way the plant as it is built appears to us on a 10 probabilistic comparison among different sequences..
11 I should ask Mr. Kascsak to answer also.
12 JUDGE BRENNER: Mr. Kascsak?
13 WITNESS KASCSAK: Well, I think the two 14 examples that were given are probably good ones, in the 15 context that they are dependencies that SAI define as 16 something that they discovered through their review of 17 operating experience, and they wanted to input that 18 knowledge both into our thought processes and how we 19 would model the plan t and how we might consider 20 improving the plant to reduce those dependancies.
21 And in both of these cases there was a 22 dialogue tha t existed in reviewing the fault trees and 23 the event trees. We discussed these dependencies and
() 24 evaluated whether or not in our normal design process we 25 did in fact address these.
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5954
{
- } 1 Now, it turns out that the DC system and the 2 DC voltage considera tions have been something that has 3 gotten fairly good attention within the industry, and we 4 have in fact addressed that in some detail, both in the 5 design of the system and in the surveillance and 6 maintenance procedures associated with those systems.
7 So we felt fairly confident tha t we have done what 8 needed to be done to remove those dependencies to the 9 greatest extent we can.
10 But we did still model it that way, and we 11 recognize that it is a dependency that is still 12 significant and we should continue to keep an eye on l
13 it. And in that respect we are sensitized to the fact 14 tha t it is an important system in terms of its 15 contribution to risk. -
16 As far as the unavailability of the diesels .
17 and those common cause dependencies of the three l
l 18 diesels, again, we looked at whether or not there were l 19 any system dependencies and whether or not preper 20 separation and procedures were in place to address 21 that. And again, we felt pretty confident that we had a 22 system that was in place that addressed these.
23 We did, as we've stated before, consider that (n
(,) 24 maybe there should be further consideration to improving 25 availability of an onsite capability other than the l
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTCN. D.C. 20024 (202) 554-2345
5955 emergency diesels using the gas turbine which is
(]) 1 2 available. So that in the sense that we could further 3 improve the availabili ty of AC power, we are considering 4 the addition of black start capabilities on the 5 diesels.
6 So all of that information together was taken 7 into account when that decision was made. So it is that 8 feedback process that we are using here.
9 JUDGE BRENNER: Just to make sure I 10 understand, if you used an onsite gas turbine, that 11 would give you onsite AC power even though the AC power 12 on the grid was down; is that correct?
13 WITNESS KASCSAK4 That is correct.
1
's 14 MR. ELLIS: Judge Brenner, for r.larity's sake, 15 can I just ask tha witness whether that was one of the 16 four examples he gave earlier this morning? I think it 17 was number f our and I just think for the record it would 18 be worth clarifying.
19 BY MR. ELLIS: (Resuming)
I 20 0 Mr. Kas:sak, with respect to the gas turbine 21 and black start capability that is being studied right 22 now, is that one of the four items that you mentioned j 23 earlier today in response to Ms. Letsche's questions?
() 24 A (WITNESS KASCSAK) Yes, it was.
25 JUDGE BRENNERs I think Mr. Kascsak also O
ALDERSON REPORTING COMPANY,INC, 400 VAGINIA AVE, S.W , WASHINGTON. D.C. 20024 (202) 554 2345 i
5956 answered the overall pending question that I meant to
(]) 1 2 apply to each example, and that is as to whether this 3 situation had been identified by non-PR A techniques l
4 also. But if you would like to add anything to that, 5 D r. Burns or Dr. Joksimovich, you can. And beyond that, 6 you might want to find an example or two which in your 7 collective opinions illustrates the usefulness of PRA 8 techniques in highlighting something that might not 9 otherwise have been highlighted, even if the ultimate 10 judgment was that it wasn't of sufficient concern to 11 make a change.
12 WITNESS BURNS: I'm sorry, did Mr. Kascsak say 13 that neither one of those examples --
14 WITNESS KASCSAK I said they were.
15 JUDGE BRENNER: Why don't you ask him your 16 question out loud and then we will get it on the 17 record.
18 WITNESS BURNS: That neither one of those 19 examples were appropriate examples to demonstrate that 20 the PBA methodology could identify or would identify 21 potential contributors to risk other than other 22 methods?
23 JUDGE BRENNER. No. Maybe I should tell you
() 24 my perception, in case it is different than what he 25 intended. Although things were not necessarily modeled ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5957 in the probabilistic sense, the concern over loss of AC
(]) 1 2 power and failure of diesels was something that has been 1
3 highlightei in the industry through other means than l 4 PRA.
5 And I guess that is as f ar as the statement 6 vent. So to that extent, the PRA by itself did not 7 highlight the problem. If you want to add anything on 8 tha t point, you can now.
9 I don't think he went so far as to say that to PRA wasn't useful in thinking about what further steps 11 to take. In fact, I think you can draw the contrary 12 conclusion. But if you want to add to that. !
l 13 WITNESS J0KSIMOVICHs If I may add a 14 clarification to see how you're thinking. The way I 15 interpret what you are thinking is that --
well, le t me )
16 tell you the way I think, and that is that there are two 17 parts of the PRA. One is an actuarial part, and I'm 18 using now the insurance company language, and that is 19 that we take into account the events which did happen on 20 an actuarial basis. And in that respect PRA would be a 21 statistical tool. So that doesn't seem to impress you, 22 that part.
23 JUDGE BBENNEB: No. Well, if you want to know
() 24 what I was thinking to myself, it is that, similar to 25 the insurance industry, if you want to carry the analogy O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
5958 I
1 a little further, where you don't have the actuarial 2 basis you charge enough to cover it. That is, I think 3 you testified last week, or Dr. Burns did, that where he 4 was short on operating experience he attempted to be 5 conservative in the approach.
6 Whether or not people agree with his statement 7 is something else, but I think that is what he said.
8 WITNESS J0KSIMOVICH We tried to get as much 9 actuarial experience as possible. Now, we take it one 10 step beyond th a t , that we are making predictions well 11 beyond the actuarial experience. And if I interpret 12 what you are asking is, you are saying, give us the 13 examples where PR A has predicted something which hasn't 14 occurred in reality at all, it has not been pointed out 15 by any good engineers, and it was just simply the 16 product of the predictive process; is that correct?
17 JUDGE BRENNER That would be a subset of what 18 I'm thinking, but it is more narrow. I am not excluding 19 things that may have occurred in history once, twice, or 20 that some engineer somewhere through deterministic 21 techniques may have identified, however for one reason 22 or another it was not focused on as a check in the 23 design process.
24 So I am not excluding that. It doesn't have 25 to be totally predictable. It can be something that was O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5959 '
perhaps overlooked because of only limited occurrence or
(]) 1 2 judgments made about its not leading to any problems, 3 which judgments may be cast into some doubt by the PRA 4 approach. So you are much narrower than I was 5 thinking.
6 I might back up. The Staff is taking the 7 position -- and any time somebody attempts to summarize 8 something, I'm sure it is misleading and some party will 9 take umbrage at it. But to put it bluntly, as far as to the Staff is concerned you can take your PRA and go 11 home, because they have done their review and they are 12 satisfied with that review without looking at your PRA 13 at all. .
14 You have been asked many questions and have 15 presented a lot of testimony about why you believe the 16 PBA is useful. And to flesh out that testimony a little 17 bit at this point, I thought that some examples might 18 assist everybody's thinking and the record, to attempt 19 to illustrate or support your views, if support can be 20 found in the examples as to the usefulness of 21 identifying potential systems interactions problems 22 through PRA techniques, even though ultimately after the 23 presentation is made judgments are then made that no
() 24 changes need occur for one reason or another.
t 25 WITNESS J0KSIMOVICH: I just might point out
( ()
l ALDERSON REPORTING COMPANY,INC, I
i 400 V.RGINIA AVE., S.W, WASHINGTON, D.C. 20024 (202) 554-2345 l
5960 I that I think that the examples that Dr. Burns presented
(])
2 are very good examples.
3 JUDGE BRENNER: Do you want to tell me a 4 little bit more as to why?
5 WITNESS J0KSIMOVICH: Well, as he pointed out, t
j 6 as a result of his concern LILCO has considered adding 7 the black start ca pability with the oss turbines. And 8 the reason why they have done that is because simply 9 this type of a sequence was surfacing and it turned out 10 to be an important sequence, which was excluded from the 11 design basis approach.
12 JUDGE BRENHERa All right. 'And I think this 1
13 is very productive to help me, at least, as a 14 nontechnical member of the Board. And if you want to go 15 through a few more examples -- you see, I don't know how 16 to suggest which examples to pick, because I don't know 17 what you have. We would like to avoid going through 25 18 of them, but I don't want to cut you short from finding
! 19 illustra tive ones within the examples that would help.
l 20 So I would like to have you proceed with a few 21 more examples.
22 (Board conferrinc.)
23 JUDGE BRENNER: Judge Horris suggests or asks
() 24 whether the examples need be limited to the Shoreham PRA 25 or just the PRA techniques, and I guess for now we would l
l ALDERSON REPORTING COMPANY, INC, l
400 VIRGINIA AVE., S.W., WASitlNGTON, D.C. 20024 (202) 554-2345
5961
() 1 like to limit it to the Shoreham PRA. And I assume that 2 is what you have done, Dr. Burns.
3 WITNESS BURNS 4 I do have some examplec from 4 elsewhere, but I can limit it to the Shoreham PRA.
5 JUDGE BRENNERa It's very difficult for me 6 because I don't know what you have there. I sure wish I 7 had seen this in the direct testimony. Don't feel 8 compelled to limit it if you think it would assis't us.
9 JUDGE JORDANS I just want to say that I acree 10 with the Chairman that these examples are very helpful 11 and I don't think we want to take up 25 examples, but on 12 the other hand we do want you to use your judgment in 13 pointing out examples that have illuminated the record 14 as it has in the two examples that you've given. And if I 15 you have other examples that do that, pa rticula rly tnose l
16 examples tha t have led to design changes, that is jtst l
17 great. We want to hear those.
18 JUDGE BRENNER: I have one suggestion, Mr.
19 Ellis.- See what you think of this. We have to meet 20 with counsel in any event, which I believe weuld not l 21 include you. A r. I correct, Mr. Ellis?
l 22 HR. ELLIS: Yes, sir.
23 JU DGE BR ENN' .: You see, at least one thing
() 24 that has piqued our interest here -- and maybe it would 25 be more efficient to give you an opportunity while we
() '
ALDERSON REPORTING COMPANY,INC, 400 V61GINIA AVE., S.W., W SH!NGTON, D.C. 20024 (202) 554-2345
5962 1 meet briefly for about 15 minutes with counsel, to
(}
2 consider incorporating it in your redirect. I'm not r- 3 saying you would not have, but raybe this would better 4 focus the approach, and tha t way you can have the 5 witnesses confer and consider what examples within the 6 examples might be best used and illustrative of the 7 kinds of things we are interested in.
8 And don't feel compelled to re' strict it to the 9 Shoreham PRA if you think some relevant point can be 10 made by other exasples. But if you do that you should 11 identify that at the time.
12 MR. ELLIS: I appreciate that suggestion.
. 13 That is what I would like to do. And I can proceed with i
14 some questions now that, at least from my nontechnical 15 point of view. would place in perspective the PRA 1e methcdology in the bigger picture, and then come back 17 after the oceak to the examples.
! 18 JUDGE BRENNER: All right. And in the 19 meantime, counsel that we have to meet with can be l
20 prepared to do that, and we will break shortly.
I l 21 JUDGE CARPENTERS In that spirit, Mr. Ellis, 22 when you come back to the broader picture, in addition 23 to the examples I would like to ask if you would think l
1
() 24 about tha following question over the break and see if 25 you can answer it when we come back.
l f~
- La) l ALDERSON REPORTING COMPANY. INC, j 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5963 Last week I asked how many of the documents
{]) 1 2 that were presented in the direct prefiled testimony 1
3 with respect to examples of systems interactions usl studies, exclusive of the material we're talking about 4
5 right now, were used as source material in the PR A . And 6 in that context I would like to ask the following 7 question
~
8 In performing the Shoreham PRA, how many 9 potential systems interactions were looked at that had 10 not been considered before in this other body of 11 material that was presented to us as the systems 12 interactions material that had been used previously? Do 13 you follow what I'm getting at? How much is brand-new?
i . 14 I am not questioning the PR A, but in the sense of how 15 auch is brand-new, and just so we can make clear, how 16 much was studied before, perhaps, but didn't get the 17 spotlights put on them the way they can be with the 18 PRA.
19 JUDGE BRENNER: If you recall, Dr. Burns, that 20 was the list that Judge Carpenter took you through, 21 which were the examples of studies presented in the 22 direct testimony, I think in the fif ties or perhaps into 23 the sixties. I mean page numbers, not years.
- ) 24 WITNESS BURNS
- Yes, sir, I recall.
25 JUDGE BRENNER: Now we will let Mr. Ellis ask (3
~>
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S W., WASHINGTON, D C. 20024 (202) 554-2345
i 5964 questions at this point, leading to the break.
(]) 1 2 BY MR. ELLIS: (Besuming) t 3 0 Dr. Joksimovich or Dr. Burns, are there other 4 --
or can you list the other techniques or methodologies 5 available to avoid unacceptable systess interactions in 6 the design and construction of a nuclear plant? Can you 7 list some of them? Dr. Joksimovich?
8 A ( W IT N ESS J0$SIMOVICH) As has been pointed out 9 in these proceedings, the discipline design ap*proach is 10 one component. The design basis accident approach is 11 another. The defense in depth is another. And various 12 deterministic studies that have been referred to is l
13 another.
14 And at the risk of repeating myself, I will 15 again make a distinction between the design process and 1
16 the assessment process. And I think the design process 17 includes all of these items successfully and provides a l 18 pretty high confidence that a lot of interactions have 19 been identified.
20 Now, when it comes to an assessment process I 21 think that the PRA techniques come into the picture 22 because they provide a diverse way of assessing the 23 importance of various systems interactions.
24 25 O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
i 5965 0 So just to summarize correctly, you are not
(]) 1 2 saying in your prefiled or your cross-examination 3 testimony that the PBA methodology displaces these other 4 techniques?
5 A (WITNESS J0KSIMOVICH) No, I am not. I am 6 saying that as far as an assessment tool, from page 65 7 of my testimony, I explicitly state the PBA as a safety 8 assessment tbol. .
9 Q And you say that the techni* ques of the design to process and the design-basis approach and deterministic 11 studies and the def ense-in-depth implementation of the 12 defense-in-depth concept all give one confidence that a 13 safe design has evolved; is that correct?
14 A (WIINESS J0KSIMOVICH) Yes, that is correct.
15 0 And then you characterize -- how did you 16 characterize the confidence or the assurance?
17 A (WITNESS J0KSIMOVICH) I would characterize 18 that that the design process provides a high level of i
19 confidence and that the PRA, being a diverse tool,
- 20 provides the highest that we can in 1982.
21 0 So they not exclusive, they are 22 complementary?
23 A (WITNESS J0KSIMOVICH) That is correct.
( 24 MR. ELLIS: I think that is what I had in mind 25 in terms of putting it into perspective. And after the (2)
ALDERSON RCPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5966 1 break we will come back with what the Board has 2 requested.
3 JUDGE BRENNER: All right, let us break until 4 2430. And after a five-minute delay, we will meet with 5 cognizant counsel back there.
6 (Whereupon, the afternoon recess was taken, to l
7 reconvene at 2:30 p.m., this same day.)
8 9
10 11 12
' 13 14 15 16 17 18 19 20 21 22 23 24 25 ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5967 JUDGE BRENNER: Let's go back on the record.
(]) 1 2 The break was more extended than we predicted, and we 3 apologize; but counsel and the Board were engaged in 4 discussing matters pertinent to this proceeding, so we 5 think it was time well spent.
6 At this time we will proceed with the redirect.
7 MR. ELLIS: Is the Board 's pleasure for us to 8 go ahead with our examples or to respond to Judge 9 Carpenter's questidns now?
10 JUDGE BRENNER: The idea of taking the break 11 was to allow you to present it in whatever order you 12 would like.
13 BY MB. ELLIS: (Resuming) 14 Q All right, Dr. Burns. Do you have any further 15 examples in the nature of the examples that you were 16 giving that you can offer the Board at this time ?
17 A (WITNESS BURNS) I think I can give three 18 additional examples, and possibly you can stop me if I l 19 have coverad enough ground.
20 The first two examples are sequences that are 21 important in calculating the risk associated with the 22 operation of the Shoceham plant. And the third example i 23 is an example of an interaction that exists in the plant 1
() 24 but may not be important in terms of a risk contributor, 25 but it is important in terms of recognizing that it l
(~8 s/
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5968 1 exists.
J 2 Okay. The first one would be there have been 3 events in which from the operating experience there have p
4 been loss of condenser vacuum events which is in a 5 nonsafety system; it is in the balance of the plant.
6 There have also been events, initiating events as I 7 would characterize them, in which the HSIVs are closed
~
8 during testing, required testing of the HSIVs during 9 operation.*
10 Both of these types of events, one in the 11 balance of the plant and one on a safety isolation 12 system, lead to similar eff ects on the plant in that -
13 they cause a loss of feedwater which is the normal 14 method of injecting water into the plant. They 15 eliminate the turbine bypass as a path for the steam 16 from the reactor into the condenser, and they also close 17 the MSIVs. So from that standpoint we can then start to 18 evsluate what happens, and that is they challenge what 19 remains which is the ECCS system network and the RHR 20 system cepability to both keep the core covered with 21 water and to remove heat from the containment.
22 So what I would say is that the initiators now 23 have -- are all dependent, the initiating event and
( 24 those events which I just described are all dependent 25 functional failures, one leads directly to the others.
n ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5969 Next we look at the safety systems to find out
(]) 1 2 if they are independent of this initiating event and if eS 3 there are any dependencies among them that lead down the 4 line to a possible core vulnerable state. The first 5 attempt by the operators in the automatic systems is to 6 maintain the pressure at near operating pressure and use 7 the high pressure injection systems which are available 8 still, the HPCI and RCIC.
9 And we have identified there.are some 10 potential common cause failures among those systems, and 11 as we talked about on Friday a week ago, we have tried 12 to benchmark the models tha t we are using to mock up the 13 Shoreham design versus the data that exists in the 14 industry. And so, therefore, those common cause 15 fsilures are accounted in the model, and th e re a re a se t 16 of -- cut sets or minimum pathways through those two 17 systems that would define what are the causes that would -
18 lead to failure of both systems due to random 19 independent failures and to systems interaction which
( 20 are common to both systems.
l 21 And in addition to that, if the high pressure 22 systems were unavailable, the next step in line would be 23 can I get to the low pressure ECCS systems that are G
k) 24 available, and those would be through the ADS system or 25 depressurization system at Shoreham. And there are some l
l ALDERSON REPORTINS COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 354-2345
5970 I
1 common cause with dependent systen. interaction failures
({])
2 having to do with the air supply to the ADS valves which 3 are modeled in the Shoreham plant and which could in
-}
ud 4 fact lead to disabling the ADS system. These are very 5 low probablity events, the things we're talking about, 6 but they are in fact identified; they are in fact 7 brought out to the surface and are available for both 8 our information and for LILCO's information.
9 If we succeed in depressurizing the plant, 10 which is a high likelihood event, then the next step in 11 the line is whether the low pressure systems, of which 12 now we have all of the four LPCI pumps and the core 13 spray pumps available to supply water to the core, 14 whether those are also available, and there are some 15 common cause dependent failures among those systems 16 which can disable either multiple pumps within the 17 system or all of the pumps; however, the frequency of 18 those events are also low.
19 I would in this particular sequence, one of 20 the particular changes that was mentioned before on the 21 Shoreham plant was in the change of the RCIC system, the 22 exhaust pressure trip that was raised from 25 psig to 50 23 psig. In this particular sequence the impact of that 24 can be seen from once the HSIVs are closed and the 1
1 25 balance of the plant is no longer available, and the O
ALDERSON REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5971
(} 1 high pressure in the primary system is blowing down into 2 the suppression pool. We are raising the temperature 3 and pressure, the temperature of the suppression pool, a
4 and the pressure in the primary containment. If now 5 RHR, which is a separate system, is unavailable or not 6 initiated by the operator immediately, there can be a 7 pressure built up inside the containment well within the 8 design capability of the containment but yet sufficient 9 in the previous way.that the RCIC was set up to cause a 10 trip of the RCIC and make that unavailable. The change 11 in the set point makes this probability less likely.
12 So in this case we have a sequence or a 13 functional. interaction in a sequence which could 14 adversely affect a system that we could use, and there 15 has been a change in this case to only the set point, 16 but a change to improve the reliability of the RCIC 17 system in cases where containment pressure may be l 18 rising, but for all intents and purposes the RCIC system 19 is still capable of both -- of supplying water to the 20 core. During the sequence that I have talked about. I 21 have given you a rundown on some of the key systems
! 22 interactions that can occur that can lead to the 23 probability of a core vulnerable condition through this 24 pathway.
25 O
1 i
e ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
l I
5972 0 When you say probability, probability
(]) 1 2 generally means more likely than not. Earlier, you said
- 3 for low probability events. Can you clarify that?
4 A (WITNESS BURNS) All of these events that we 5 are talking about do involve multiple f ailures and are 6 beyond the design basis, and are remote sequences, in 7 terms of remote low frequency, very low frequency events.
8 0 Does that mean low probability?
9 A ( WITNESS BURNS) Low frequency events, yes. In 10 order to get at the systems interaction issue, one can .
11 look at this sequence of events and determine what the 12 functional relationship of these components is. In 13 addition to that, one can also go another level below 14 and define what are the multitude of paths through the 15 sequence.
16 In other words, the sequence that we have 17 calculated is presented in event trees as one path 18 through the sequence. However, what makes up that page 19 are multiple paths or multiple contributors f rom fault 20 trees. And those are represented in what we have talked 21 about in terms of cut sets. And I have tried to 22 identify what the major ones are in each of these.
23 The second example I would go through is very
) 24 similar to this, and again, it also is --
25 JUDGE CARPENTER: Before you leave that
()
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5973 1 example, what reduction in risk was associa ted with th a t
(~ ) l 2 change? Just to put a number on the example.
3 WITNESS BURNS In this particular example, l 4 the change in the setpoint of RCIC affects other 5 sequences, also. And I am not sure I could assess 6 that. But the change, that change, as it affects the 7 sequence, is probably a one or two percent eff ect on a 8 very low probability or frequency sequence.
9 JUDGE CARPENTER: Thank you for putting it in 10 perspective. .
11 WITNESS BURNS 4 The other sequence is i 12 virtually the same, except the initiator is different.
13 In this case, we have a turbine trip, and as we see from
[
l l (-)) 14 operating experience, turbine trips can occur very 15 frequently. What they do is they trip the turbine but 16 they do allow the MSIVs to remain open and the feedwater 17 to remt!.n available iIf the water level can be l 18 stabilized in the primary system above the HSIY closure 19 setpoint.
20 And for this particular sequence we have 21 identified the potential that the MSIVs may close more 22 frequently than one would like during this particular 23 transient, and there are two paths that LILCO is
() 24 presently following to, one, verify that our assumptions 25 are correct; and two, to evaluate the possibility of a O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5974
(} 1 change that would improve or reduce the sequence which 2 is already a low frequency -- to reduce that even 3 further.
4 And that would be, first of all, the 5 evaluation of whether it could possibly be a poor 6 assumption. and tha t is, we a re presently using 7 deterministic licensing evaluation analysis done by 8 General Electric for plant response under a turbine 9 trip, and that is how we are characterizing the response 10 of the plant. .
11 During the startup testing, there actually 1 12 will be performed this test on turbine trip at power, 13 and from that test, we will be able to characterize 14 whether the General Electric analysis may be overly 15 c on =e rva ti ve and, therefore, that the plant may be more -
16 resilient; resilient in the sense that the MSIVs can .
17 remain open and the water level stabilized above what is 18 the level 2 MSIV closure trip setpoint.
I 19 If that is the case, then the conditional l
20 probability that we have assigned to keeping the MSIV 21 closures -- the reliability that we have assigned to 1
22 that may be too high in that the reliability may be l l
23 better than tha t. And we will verify that with the 1
( 24 startup testing.
25 The other pathway that LILCO is investigating ALDERSON REPORTING COMPANY,INC, 400 VIRGINtA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5975 at present, as mentioned by Mr. Kascsak, is is there a
(]) 1 2 benefit from reducing the trip point of HSIV closures 1
3 from Level 2 to Level 1, which is nearer to the top of 4 the reactor fuel. But yet, it allows the water level in 5 the reactor vessel to oscillate or to go over a much 6 vider range and yet, still remain the MSIVs open and 7 still maintain the condenser available and still 8 maintain the f eedwa te r available to supply water to the 9 reactor vessel.
10 So there is a big benefit.in these challenges 11 which we know occur in the industry, which we know will 12 occur. They are expected transients; they are 13 anticipated, and we are trying to identify the best way 14 to mitigate or maintain cooling to the core with the 15 normal systems, and not challenge the safety systems.
16 Those are my two principal examples. The.
17 third example is possibly uninteresting, but it is an 18 example of a detail that is in the fault tree' analysis 19 in the logic model. And the particular example I have 20 chosen is the low pressure systems which ultimately we 21 may fall back on the low pressure systems to ultimately 22 preserve core cooling.
23 'Je have four LPCI pumps that inject through
() 24 two injection points, and they are automatically started 25 on both low level in the reactor vessel and high drywell O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554 2345
5976 pressure, so there are a set of accident sequences that
{]}
1 2 can sta r t the pumps.
3 JUDGE BRENNERs You said both; did you mean
)/
4 either, or did you mean both?
5 WITNESS BURNS: I am sorry?
6 JUDGE BRENNER4 You said they would actuate on 7 both signals. Did you mean both?
8 WITNESS BURNSs Yes, both.
9 In addition to that, there are two injection 10 points, so there are two points in which these four 11 pumps come together, and those two points that discharge l
12 into the reactor vessel have discharge valves that are 13 required to open and are interlocked closed. And the 14 only way that those vill open automatically is if they 15 receive a clearing, a clearing of the low pressure 16 interlock inside the reactor. .
17 For the primary system, in the models that we 18 have constructed, we postulated failures of those valves 19 combined, due to both valve failures and to random valve 20 failures, dependent failures and to failures of the l
21 instrumentation, which would be the low pressure signal 22 from the primary system. And this is an example of t
23 human interface with the machine, into which a human l
() 24 error could go undetected and could lead to a common 25 cause or dependent or systems interaction failure of l
ALDERSON REPORTING COMPA'JY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D C. 20024 (202) 554 2345
m 5977
(} 1 both injection points during a challenge.
2 But again, while this is an important fact, 3 because of the redundancy built into the ECCS network, 4 this is one example that does not rise to the top as a 5 primary example of an important risk contributor, if we 6 are using risk as a measure of plant safety, and what is 7 important in systems interaction.
8 BY MR. ELLIS (Resuming):
9 0 Dr. Burns, let me interrupt, if I may. You 10 said both signals, low water and high pressure, and 11 Judge Brenner asked you whether you meant both or 12 either, and I was a little confused. Does the ECCS come 1
13 on in either case, or does it require both to get the 14 ECCS?
15 A (WITNESS BURNS) The low pressure pumps are 16 initiated on lov level signals in.the reactor vessel, 17 and high dryvell pressure. This is the containment; not 18 reactor vessel. The dryvell pressure. Either one of l
19 those. In the one out of two taken twice logic; the BWR l
1
! 20 logic.
l l 21 JUDGE JORDAN So it is either and not both?
22 WITNESS BURNS: Pardon me?
23 JUDGE JORDAN: Either will trip it; it doesn't
() 24 require both?
25 WITNESS BURNS: Oh, I see. Yes, sir, that is O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINI A AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5978
(} 1 correct. Either event will lead to the initiation of 2 the pumps.
3 JUDGE JORDAN: And by failure, you meant 4 failure to open, in this case, is that right?
5 WITNESS BURKS Failure to open?
6 MR. ELLIS: Of those valves.
7 WITNESS BURNS Failure to open of the 8 discha rge valves , yes, sir .
9 JUDGE JORDANS Those valves are normally to 10 protect.the low pressure system from the high pressure, 11 is that righ t?
12 WITNESS BURNSs Yes, sir.
13 JUDGE JORDAN In some instances, aren't those 8 14 stop valves?
15 WITNESS BURNS 4 Those are motor-operated 16 valves. .
17 JUDGE JORDAN These are motor-operated 18 valves. I see. So it is much like a PWR in which the 19 low pressure system is isolated. Okay.
20 WITNESS BURNS: Yet it is like a PWR in that 21 the low pressure systems are isolated. I am not sure if 22 the valve is the same.
23 JUDGE JORDAN 4 If I remember correctly, they
() 24 also have operator valves in addition to the one-way 25 stop valves.
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5979
(~3 1 BY MR. ELLIS (Resuming):
%.)
2 0 Dr. Burns, turning next to Judge Ca r pe n te r 's 3 questions, first, Judge Carpenter asked how many of --
7-)
V 4 JUDGE MORRIS: Excuse me, Hr. Ellis, if I 5 might interrupt. Dr. Burns, do you remember the general 6 question that Judge Brenner wanted to get the answers to 7 these questions; namely, to show to what extent tha t the 8 PRA would identify these instances of importance; 9 whereas, a deterministic analysis would not?
10 JUDGE BRENNER: Or a deterministic analysis 11 would not do it as well. And you can qualify the "as 12 well." And that goes back to the dialogue I had with 13 Dr. Joksimovich.
14 (Pause.)
15 WITNESS BURNS 4 I think in much of this, the 16 deterministic analysis does identify many of the same 17 in te rac tion s . So many of these are included in the 18 design basis, especially those that affect the balance 19 of the plant. Possibly, the ones that are not 20 identified in deterministic analysis are those that 21 involve multiple failures of ECCS equipment, and those 22 would be the ones involving the f ailures in both high 23 pressure systems, at the same time coupled with
) 24 potential f ailures of the automatic depressurization 25 system.
O ALOERSON REPCATING COMPANY,!NC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5980 Specifically, we can look at one particular
(]) 1 2 cut set in the ADS which is the common cause failure of 3 all air supply to the ADS valves to open. This would 3
b 4 require f ailure not only of the instrument nitrogen 5 system that is in the reactor building, but also, the 6 accumulators which are in the reactor building. This 7 could occur f rom isolation of the normal supply, which 8 is the nitrogen system, and then subsequently possibly 9 to back leakage on the check valves to the accumulators,
. 10 causing low air pressure in the accumulators and 11 therefore disabling the ADS valves.
12 Again, these are what we are talking about 13 since in the real world, we only need very few of these 8 14 valves to open. We are talking about a very remote 15 occurrence, but yet, this is beyond the design basis.
16 But it does become within the scope of these multiple 17 failures that we are postulating to s'ee, even in the 18 remote frequency sequences, whether these events could 19 occur and whether on a relative basis they are important.
20 The question of the switch, the low pressure 21 interlock switches, on the valves for the discharge of 22 the LPCI pumps I don't believe is something that is in 23 the deterministic analysis, but I am not familiar with
() 24 everything that is done deterministically.
25 So, I could be way offbase on that, but I l
O l ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5981 don't think it is. It is a multiple failuce in the
{]) 1 2 system.
g- 3 JUDGE M3RRIS: So I guess it is difficult to 4 say that PBA is orders of magnitude better in 5 identifying these initiating events, which might lead to 6 degradation of the system. But it does -- let me ask 7 rou this question. Does it add something to 8 understanding by quantifying the frequency to be 9 e xp ec ted or probability, if you will, in addition to 10 consequences? And therefore, it gives you a better 11 perspective on whether or not remedial measures might be 12 improvements or necessary?
13 WITNESS BURNS: In my judgment, yes, sir, that 14 is where it provides its biggest benefit.
15 BY MR. ELLIS (Resuming):
46 0 D r. Burns, prior to the break, Judge Carpenter 17 listed two questions. The first question was he asked, 18 I believe, how many of the various studies, 19 deterministic and otherwise, were used as source 20 material for the Shoreham PRA. I am not sure whether i 21 there is an ambiguity but I think - oh, I'm sorry.
22 JUDGE CARPENTER: No. I made the point that 23 he had previously testified in response to that
() 24 question. And then I am asking a second question in the 25 context of that pCevious question.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5982
() 1 BY HR. EARLEY (Resuming):
2 0 And I believe the second question was, in
-m 3 doing the Shoreham PRA, how many of the systems 4 interactions that you looked at in the Shoreham PRA had 5 not been considered in the previous studies. Can you 6 answer that, Dr. Burns?
7 A (WITNESS BURNS) That question is extremely 8 difficult, and I am really afraid that I can't do it 9 justica because it would require a detailed comparison to of what they considered and what we considered.
11 JUDGE CARPENTER 4 I thank you for your 12 thoughtful, accurate response. While acknowledging your 13 response, can you give me some feel of whether it -- on l
14 a percentage basis -- 90 percent of the systems 15 interactions that you used had previously been looked at 16 in these other studies, and you came up with 10 17 percent? Is it 1 percent and 99? I really want more a 18 sense of feel than a precise, defensible description.
19 20 21 22 23
( 24 25 l
()
i ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5983
() 1 (The panel of witnesses conferred.)
2 WI"HESS BURNSa I do know that the 3 deterministic studies look at many, many systems 1 O 4 interactions. We have tried to go a little bit beyond 5 the deterministic studies to try and look over the edge 6 and see whether beyond the design basis there are, in in 7 fact, other multiple failures that in the context of 8 systems interaction that could lead to core-vulnerable 9 states.
10 So I think that there are more. I am not sure 11 I can say more. We have looked at diff erent systems 12 interactions possibly, but they are beyond the design
,13 basis. But I am not sure that I would -- really, since 14 I am not intimately familiar with everything that is in 15 the deterministic analysis, I would be afraid of 16 sligh ting those analyses if I tried to say that we did 17 more.
JUDGE CARPENTER: That is why I had tried to 18 19 define the question in terms of the documents that you 20 had already identified as being ones that you looked at, 21 which were previous systems interaction studies, 22 according to the testimony.
23 And I was just trying to get some feel for how 7
24 far you vent beyond based upon other things that lead 25 you to look to systems interaction, like the walkdown, O
AtDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5984 I~h 1 et cetera, the other sources of stimulation, if you V
2 will, in doing the analysis. I am just trying to get 3 some feel. And when you say "a little," I take it it is Ox 4 not 50 percent previously done, 50 percent new, but it 5 is of the order of somewhere between 1 and 10 percent, 6 maybe 90 percent of what you have done had previously 7 been done in the deterministic analysis and something 8 roughly 10 percen tage. And I am not trying ta put words 9 in your mouth.
10 How would you respond, though?
11 WITNESS BURNS: I really feel inadequate to be 12 able to answer. We could go maybe through some some of 13 the deterministic studies. I know that I am not 14 familiar with some of those. Some of those are ongoing, 15 and I have not seen those. Some of the , FEM As that were 16 done that I know about were single failures. So we 17 have, in my judgment, clearly gone far beyond those l 18 FEMAs in determining what system interactions could be l
l 19 important.
I 20 The heavy-loads analysis, we did not use.
21 Their preoperational testing and startup testing has not 22 occurred yet.
23 JUDGE CARPENTER: Dr. Burns, I think I want to i
() 24 thank you for responding to the extent you have. I have ,
25 learned some. Thank you.
O ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
l 5905 BY MR. ELLIS: (Resuming)
(]) 1 2 0 Drs. Burns and Joksimovich, do you have what 3 has been marked Suffolk County Exhibit 19 there? For 4 the record, Suffolk County Exhibit 19 is a NUREG 5 CR-1859, entitled " Systems Interactions State-of-the-Art l
l 6 Review and Methods Evaluation."
l 7 Only portions of that were marked. Do you 8 have that in f ront of you?
9 A (WITNESS BURNS) Yes, sir.
10 0 All right. You were asked questions concernin 11 the techniques of evaluating systems interactions that 12 are listed on Roman XVII, XVIII and XIX. Can you tell 13 se whether each of those techniques or methods have been 14 employed in the Shoreham PRA to detect systems 15 interactions or in connection with other Shoreham 16 studies other than the PRA?
17 A (WITNESS BURNS ) I personally can speak to the 18 question of reviews of operating experience. We have 19 performed those as a part of the PBA. I should also let 20 Dr. Joksimovich and Mr. Kascsak talk to that particular 1
21 item.
22 The analysis by parts, the second item, we 23 have employed methods similar to what ic referred to
( 24 there in terms of constructing the logic models that are 25 input to the PRA. And again I will let Mr. Kascsak l
l O
ALDERSON REPORTING COMPANY,INC, l
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5986
() 1 elaborate sgain on that.
2 The graph-based analyses, we have performed 3 the fault-tree and event-tree analysis that is mentioned O 4 here. So that also we have -- that portion of the items 5 has been incorporated. And on-site inspection in terms 6 of the vsikdowns that we have talked about over these 7 last few days, that issue has also been addressed.
8 Q Mr. Kascsak, do you have anything to add to 9 any of those?
10 A (WITNESS KASCSAK) Well, I think in section 5 11 of the prefiled testimony and previous cross-examination 12 that we have been through in the last couple of weeks we 13 have discussed many of the studies listed in section 5 14 and how they are related to the categories depicted in 15 the study.
16 And I believe previously we had been asked a 17 number of questions as to how these types of techniques 18 were used at Shoreham, and we had characterized many of 19 the deterministic studies into these categories, and I 20 think it would suffice it to say that we have gnerally 21 covered all of these categories in one way or the other 22 in the studies that have been identified in section 5.
23 0 That is all of the techniques listed there
() 24 have been employed in some fashion at Shore, ham to detect 25 systems interactions?
O ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5987 1 A (WITNESS KASCSAK) Yes.
2 0 Dr. Burns, are you aware of other studies that
! 3 discuss systems interactions methodologies?
10 4 A (WITNESS BURNS) Yes, sir.
5 0 Are these done by the other national 6 laboratories?
7 A (WITNESS BURNS) Yes, to my knowledge. There 8 have been reports issued by Sandia, Battelle, 9 Brookhaven, and this one by Lawrence Livermore.
10 Q Is the methodology that you employed in 11 connection with the Shoreham PR A consistent with the 12 recommendations of those studies?
13 A (WITNESS BURNS) My review of the studies from O 14 the national labs, particularly the recommendations that 15 they have made, indicate to me that the methods we have 16 used in the Shoreham PRA to identify systems 17 interactions are consistent with the recommendations 18 that those four national labs have made.
19 '
O Dr. Burns, in response -- or it may have been 20 Dr. Joksimovich -- in response to one of Ms. Letsche's 21 questions, I think you said the difference between the 22 evnent trees and the fault trees at Shoreham and those 23 of WASH-1400 was that the Shoreham event trees and fault 24 trees were more detailed. Is another difference the 25 f act that the Shoreham event trees and f ault trees are O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W WASHINGTON, D.C. 20024 (202) 554 2345
5988 1 plant-specific to Shoreham?
2 A (WITNESS BURNS) Yes, sir.
3 0 Dr. Burns, I would like to hand you, if I may, O
l 4 a page 3-149 from the PRA which was used by Ms. Letsche 5 in asking you severs 1 questions, so that I can ask you 6 some follow-up questions if I may.
7 MS. LETSCHE: Is that part of an exhibit?
8 MR. ELLIS No. You did not make it part of 9 an exhibit.
10 BY MR. ELLIS: (Resuming) 11 0 Dr. Burns, do you remember this particular 12 page from the Shoreham PRA and being asked questions 13 about it?
14 A (WITNESS BURNS) Yes, I do.
15 MR. ELLIS: We would like to have this marked 16 as LILCO Exhibit 13 and admitted into evidence.
t 17 JUDGE BRENNER: Well, hearing no objection --
18 JUDGE MORRIS I think this should be 14.
19 MR. ELLIS: Thank you, Judge Morris.
20 (The document referred to 21 v'as marked LILCO Exhibit 22 No. 14 for
. 23 identification.)
l O 24 avocs 8asaura . We v111 a 1t it 1 ate 25 evidence.
O l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5989 l
Q 1 (The dccument previously 2 , marked LILCO Exhibit No.
~
t
- f. m 3 14 for identification was U 4 -received in evidence.)
5 ,
JUDGE BRENNER. Mr. Ellis, this may not be 6 r equi red . In this particular exauple, it might be
~
. 7 helpfu'l to me, since it was no: previously marked, 60 8 you have an approximate' transcript reference where the 9 questions .were askiPd about this previously, or at leas,.
10 an approxiha te half of whien day's transcript?
11 MR. ELLISt Y e s', sir , i t will just take me a m
12 coment. -~
f .
13 (Pause.)
14 ,
JUDGE BRENNER: Incid en tally, let us, in 15 a d m itting it, for convenience let us bind it into the 16 - transcript a t this point.
17 (The document referred to, Exhibit No. 14, ,i
- 18 followss) ,,
i
- 19. .
a ,.
2b '
i '~ , ,
s i 21 s -
22 23 ,
1 24 ,
~
- 25 4
< ALDERSO'4 REPOPUNG COMPANY. !NC. -
. 400 VIRGINIA AVC., 'i.W., WASriiNGTON. D.C. 20024 (202) 554-2345
w -- ux_ - ._3.w mm- m-wq b k
' ik0f e Class 11 (Figure 3.6.4): Loss of containmer,t heat removal due to a combination of high initiator frequency on the f- randoT. failure of the RHR and the ability to open tfie MSIVs
( over the long time available to provide containtrents* heat reirov a l . p
? e Class Ill (Figure 3.6.5): Inade w ate coo'lant makeup due to a LOCA plus the failure of coolant makeup.
e Class IV (Figure 3.6.5): Containtnent heat recioval proves inadequate due to a failure to insert sufficient negative reactivity. This results primirily from:
1
- 1. The comon-mode f ailure of ell control rods to insert
- 2. The failure of manual operator action to inttiate the one SLC pumps in sufficient time to prevent overpres-surizing the containment.
3.6.3 Comparison yith WASH-1400 I WASH-1400 basically used one class of accident sequence and five contain-
_ ment failure modes to represent BWRs. Therefore, for the purposes of ,
estimating the total calculated frequency of a potentially degraded core condition, the five Shoreham classes should be summed and compared with
! the value from WASH-1400. The Shoreham evaluation produces a total estimate of degraced ccre conditions smaller than WASH-1400.
Figure 3.6.6 provides a comparison of the frequency of core melt from WASH-1400 versus the frequency of core vulnerable condition from the SNPS PRA. As noted previously this comparison is made based renn similar assumptions for the derivation of each frequency. Figure 3.6.L n .o-strates that the total frequency of core melt provided in WASH-1400 approximates that determined in the SNPS FRA to represent the frequency R -
U ' of events which could place the core in jeopardy i.e., vulnerable.
However, Figure 3.6.6 also demonstrates that the composition of the calculated frequency is significantly different although two of the dominant WASH-1400 sequences (TW and TC) remain important in the SNPS evaluation of the frequency of radionuclide release.
2.. u n
5990 HR. ELLIS: It appears to have been right
(]) 1 2 before lunch. I think I am closing in on it here, on 3 Tuesday.
O 4 JUDGE BRENNER: Yesterday?
5 ER. ELLIS: Yes, sir. Yesterday.
6 JUDGE BRENNER: All right, that is 7 sufficient.
8 JUDGE MORRIS: I believe it is page 5749 -- I 9 am sorry, 5759.
10 HR. ELLISs Yes, sir. Thank you.
11 BY HR. ELLIS4 (Resuming) 12 0 Dr. Burns, you were asked in connection with 13 LILCO Exhibit 14 about.the term " degraded-core 14 conditions." Am I correct that you equate that with 15 core melt? Is that right?
i 16 A (WITNESS BURNS) In the context of this 17 summary, the sentence at the bottom of the first 18 paragraph should reads "The Shoreham evaluation 19 produces a total estimate of core melt frequency smaller 20 than WASH-1400."
21 0 So when that sentence was written, the core 22 melt and degraded core were equated, in other words, for 23 that sentence?
24 A (WITNESS BURNS) Yes, sir.
25 0 The sentence, however -- does the sentence ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5991 l
() 1 remain true?
2 A (WITNESS BURNS) Yes. The preliminary, the 3 draf t evaluation tha t we have supplied to LILCO in 4 evaluating the accident frequencies at Shoreham produce 5 a total accumulated frequency of potential core melt 6 tha t is smaller than WASH-1400.
7 0 What is the significance of that with respect 8 to the Shoreham design?
9 A (WITNESS BURNS) In my judgment, it implies 10 that we can characterize the safety of Shoreham to 11 better than or aquivalent to other operating light-water 12 reactors.
13 JUDGE BRENNER: Could I get that answer read 14 back? It is my fault, and I apologize. I missed it.
15 (The reporter read the record as requested.)
16 JUDGE MORRIS: Dr. Burns, would you qualify 17 that within the accuracy of the studies made?
, 18 WITNESS BURNSs I think, yes, sir, I l
19 appreciate that. Definitely, within the uncertainty 20 bands and, in addition, the safety of operating reactors 21 as characterized by WASH-1400 or as characterized by 22 W ASH-1400 and within the uncertainty.
23 JUDGE BRENNER: And I take it what you had in
() 24 mind with respect to your phrase "as characterized by 25 WASH-1400," in part, is that the methodology or
()
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5992
(} 1 a pplica tion of hte methodology in WASH-1400 may va ry 2 somewhat f rom what you did for the Shoreham PRA in a 3 number of respects? And I do not want to list them all, O 4 and you would have to take a look at to whether those 5 differences were material also, given the error bands 6 and other considerations? Correct?
7 WITNESS BURNC: Yes, sir, that is certainly a 8 major part of it also, and the f act that there are a 9 very limited number of other attempts at characterizing .
10 tha t safety, yes, sir.
11 JUDGE BRENNERs Mr. Ellis, as long as we have 12 interrupted you -- and I ask just for a status report --
13 do you have an idea of how much you have on redirect?
14 MR. ELLIS: yes. I think that roughly 30 15 minutes or less.
16 JUDGE BRENNER: Proceed. -
17 (Counsel for LILCO conferred.)
18 BY MR. ELLIS: (Resuming) 19 0 Dr. Joksimovich, did you have anything also to 20 add on that answer?
21 A (WITNESS J0KSIMOVICH) Yes. I would basically 22 like to restrict the statement to WASH-1400 type of 23 assessment. Other assessments have been done of 1
() 24 operating reactors, and what was meant in this statement
! 25 is that it be restricted to the WASH-1400 type of O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5993
() 1 assessment in particular for the BWRs in WASH-1400.
2 We are talking about at Shoreham, that the 3 Shoreham risks are comparable to those assessed in 4 WASH-1400 from the standpoint of a mean core-melt 5 frequency and from the standpoint of uncertainty bands.
6 Q Dr. Burns, in response to Dr. Carpenter's 7 questions, I think you indicated that the size of the 8 pie for Shorehan as compared to Limerick if Shoreham and 9 Limerick had both used core melt instead of Shoreham 10 using core vulnerable and Limerick using core melt, that 11 the pie would have baen smaller for Shoreham. And you-12 gave a range of 5 x 10-6 to 1 x 10-5, indicating you did 13 not remember the exact number. Do you recall or do you
! 0 14 know the point estimate or mean frequency now?
15 A (WITNESS BURNS) The value in the draf t report 16 is 4 x 10-6 per reactor year.
17 0 And so that would then make the pie smaller 18 than the Limerick pie?
19 A (WITNESS BURNS) Yes, sir.
20 0 Dr. Joksimovich, would you look, if you would, 21 please, at page 5831 of yesterday afternoon's 22 transcript, and look, if you would, please, at the 23 question or at the answer and question and then answer 24 from lines 1 through 12?
25 (Pause.)
()
V j ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 1
1 I
5994 ,
l What were the page numbers at the top of that
(]) 1 2 page, please?
3 A (WITNESS J0KSIMOVICH) 5831.
4 0 Dr. Joksimovich, you see there that you 5 indicated in your first answer that certain 6 probabilistic numbers and in the second question you 7 refer to percentages.
8 A (WITNESS JOKSIMOVICH) Right. I didn't mean 9 to imply that the two were linked. I thought that Ms.
10 Letsche had two questions in mind. So my second answer 11 gave my personal view of what do I mean by "important 12 risk contributors." And so 10 percent was my 13 specification for an important risk contributor, O- 14 clearly. So in any risk assessment, when I see 15 contributors in the 10 percent to -- I don ' t know -- 99 16 percent class, those are important risk contributors.
17 When th ey are in the 1 to 10 percent class, I 18 do not dismiss them at all, because I consider the 19 uncertainty bands involved, and hence there are a lot of 20 risk contributors that I would pick up in that class.
21 Below 1 percent I would typically consider them to be in 22 the noise level.
23 (Counsel for LILCO conferred.)
l
() 24 Q And the percentage figures that you gave are 25 not rela ted to the probability figures that you gave in O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (20,2) 554-2345
5995
() 1 the previous answer?
2 A (WITNESS J0KSIMOVICH) That is correct.
3 0 Dr. Burns, you testified in response to Ms.
4 Letsche's questions that the Shoreham PRA considers 5 operator eLcors of omission in the event trees beyond 6 the initiating event and operator omissions, operator 7 errors of commission as initiating events. Is the 8 current s t a t e -o f -t h e -a r t in PR A methodology adequa te to 9 take accurately into account operator errors of to commission beyond the initiating events in the event P
11 trees?
12 A (WITNESS BURNS) In my judgment, that area of l 13 PBA or risk analysis is still under development'and not 14 currently within the state-of-the-art. l l
15 0 Operator errors are modeled, though, in the -
16 event trees, are they not?
17 A (WITNESS BURNS) The human coupling and 18 operator errors in the logic models are included insofar ;
i 19 as operators can violate procedures, can be performing i
20 maintenance, can cause systems to be unavailable due to 21 their actions that are required actions in the plant.
22 And in addition to that, operator errors can 23 -- are included in the context of failing to follow
() 24 procedures in omitting steps, not performing directives, 25 not initiating safety systems.
O l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5996 JUDGE J3RDANs I need a little clarification.
(]) 1 2 I got partly confused, and I partly do not understand.
3 Are you saying that if an operator turns off a safety
}
l 4 system, like a high-pressure injection system, this is 5 an error of commission and is not within the 6 s ta te-o f -th e -a rt? Is that correct?
7 WITNESS BURNSa In my judgment, yes, sir.
8 JUDGE JORDANS All right. So that the 9 TMI-type accident is'.not one that you feel that is in 10 the state-of-the-art because that was an opera tor error, 11 turning off the high-pressure injection system?
12 WITNESS BURNS: That one was complicated by .
13 the fact that his display JLfermation was lost. So in 14 that sense, this methodology does address those issues, 15 issues where the operator becomes confused because of a 16 loss of display information resulting from a transient.
17 So if there is a case where the transient 18 initiator or the initiating event leads to a loss of 19 display or a confusion to the operator, then that is 20 modeled in this PRA, but not his random -- his randon 21 creative actions which may disable multiple systems.
22 JUDGE JORDANS That clears up the problem that 23 I had with the original testimony which occurred, I
() 24 b elie ve , the af ternoon that I was not here. Thank you.
25 BY MR. ELLIS: (Resuming) l O
l ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5997 0 Dr. Burns, Ms. Letsche asked you a number of
(]) 1 2 questions about the walkdown, the walkdowns. Not 3 without a little trepidation do I venture back into that 4 area. But do you have before you Suffolk County Number 5 22? That is the excerpt from the draf t PRA relating to 6 the walkdoens. That is page 2-32.
7 First, let me ask, if I may, Dr. Burns, is 8 there any definite prescription in the PRA guide at the 9 present tise for walkdowns?
10 A (WITNESS BURNS) No, sir.
11 Q Now, with respect to Suf folk County Exhibit 12 22, in the third paragraph --
i 13 JUDGE BRENNER: Your use of the word 14 " prescription" has some possible ambiguity.
l 15 Dr. Burns, did you mean that there is no 16 particular type of walkdown recommended, or did you mean 17 that walkdowns are not recommended a t all where 18 discussed?
19 'dITNESS BURNS : The PBA procedures guide
- 20 recommends that plant-familiarization tours, I believe 21 is the terminology they use, be conducted in conjunction 22 with the performance of a PRA. But the guides, as I 23 have seen them, have not identified what that
( 24 f amiliariza tion tour should consist of.
25 JUDGE BRENNER: Thank you.
O ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
5998
() 1 BY NR. ELLIS: (Resuming) 2 0 In the third paragraph of Suffolk County 3 Exhibit 2 --
4 JUDGE BRENNER4 Do you mean 22?
5 MR. ELLIS: 22, I am sorry.
6 BY MR. ELLIS (Resuming) 7 0 You see in the second sentence it says, "For 8 the walkdown each SAI reviewer was teamed with a LILCO 9 engineer knowledgable in the particular design and 10 operation of the system." Do you see that, Dr. Burns?
11 A (WITNESS BURNS) Yse,. sir.
12 0 And then there was a reference in the last 13 sentence of that paragraph to "LILCO operations 14 engineers." Are those two different engineers, the 15 LILCO engineer in the second sentence and the LILCO 16 operations engineers in the last sentence? Are those 17 two dif ferent individuals or persons?
18 A (WITNESS BURNS) Yes, sir.
19 HS. LETSCHE Excuse me, could I have that -
20 last question and answer read back, please?
21 JUDGE BRENNER: Let me see if I can summarize 22 it for you, and if you do not think that is adequate, we 23 vill get it read back.
24 Mr. Ellis referred to the third paragraph of 25 Suffolk County Exhibit 22, which is in this particular O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
5999
() 1 instance page 2-32. And the paragraph makes reference 2 to both a LILCO engineer who is teamed with each SAI 3 reviewer, and the paragraph also makes reference to 4 LILCO opera tions engineers being questioned at the end 5 of the walkdown. And the question was, "Is that the 6 same person?" And the answer is, "No."
7 BY NR. ELLIS: (Resuming) 8 Q Mr. Kascsak, would you look at page 120 of
. 9 your prefiled testimony, please, sir? ,
10 (Pause.)
11 A (WITNESS KASCSAK) Yes, I have that in front 12 of me. .
13 0 I think I referred tyou to the wrong page.
14 Look at page 122, if you would, please, at the bottom of 15 the page. Do you see a reference there to the sentence 16 that reads, "The report will be assessed by the Nuclear 17 Engineering Department, which acts as the project 18 manager for the analysis and therefore is charged with 19 assessing its findings"? -
20 Is the report assessement by the Nuclear 21 Engineering Department ref e rred to there the -- and the 22 conclusions and recommendations from that assessment 23 that are referred to on the top of page 123 -- is that
() 24 what you were referring to when you answered the Board's 25 questions ea rlier toda y -- or yesterday, it may have O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6000
() 1 been -- concerning LILCO's review of the PRA?
2 JUDGE BRENNER: That was today. It only seems 3 like a long time sgo.
4 (Laughter.)
5 WITNESS KASCSAKa The reference to the report 6 that I had alluded to earlier is the report that is 7 identified on page 123, which is the Nuclear Engineering 8 report dealing with the assessment of the analysis and
. 9 the conclusions and recommendations that we will put, 10 together as a result of the review of the analysis 11 itself.
12 So the report that I believe Judge Morris had 13 asked about earlier is the report that is referenced in 14 this prefiled testimony. .
15 BY MR. ELLIS4 (Resuming) 16 Q Do you have before you, gentlemen, all three 17 of you, the testimor y f rom Friday, page 5651 -- I am 18 sorry, it is June 25th.
l 19 I am sorry, Judge Brenner. I should have 20 alerted you to that.
21 A (WITNESS KASCSAK) What was the page again?
22 0 5651.
23 A (WITNESS KASCSAK) Yes, we have it.
() 24 JUDGE BRENNER: What was the page number 25 again?
()
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6001
() 1 HR. ELLIS: 5651 for June 25.
2 JUDGE BRENNERs Yes, I guess I have misplaced 3 my copy.
4 BY MR. ELLISs (Resuming) 5 0 I would like to direct your attention to the 6 bottom of the page beginning at line 20. Ms. Letsche 7 asked the question whether you gentlemen agree that a 8 combination of th e techniques used would be an effective 9 way of identifying systems interactions. Were fault 10 trees and event trees, they were used in the Shoreham 11 PRA, were they not? That is one of the ones that is 12 mentioned there..
13 A (WITNESS BURNS) Yes, sir.
14 Q And so were walkdowns?
15 A (WITNESS BURNS) Yes, sir.
16 0 Was the FEMA technique or methodology used 17 also in connection with Shoreham?
18 A (WITNESS BURNS) Yes, sir.
19 Q How about dependency matrices?
20 A (WITNESS BURNS) Yes, sir.
21 0 And you are also aware, are you not, that 22 deterministic studies were done?
23 A (WITNESS BURNS) I am aware that they were
() 24 done, yes, sir.
25 0 So all of the techniques that Es. Letsche O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6002 1 mentions in her question have been used in Shoreham?
(}
2 A (WITNESS BURNS) Yes. l 3 (Counsel for LILCO conferred.)
{
4 Q Let me see if we can put into perspective the J
i 5 testimony concerning Limerick PR A giving you a measure, 6 or, I think you said, "some comfort." I take it that, 7 in your opinion -- and I am addressing this to any 8 member of the panel --
in your opinion, a plant-specific 9 PRA would give one greater assuran,ce than one that was 10 not plant-specific? Is that correct?
11 A (WITNESS J0KSIMOVICH) That is correct.
12 0 . Well, to put things in perspective, can you 13 characterize or explain what you mean by "some comfort" 14 from the Limerick PRA? And I would address this to Dr.
15 Burns or Dr. Joksimovich, if they wish.
16 JUDGE BRENNER: Dr. Burns does not lik e th at 17 term. You had better ask Dr. Joksimovich.
18 Dr. Joksimovich?
19 (Laughter.)
20 BY MR. ELLIS: (Resuming) 21 0 or Mr. Kasesak.
22 A (WITNESS J0KSIMOVICH) Since I do not have a l
23 problem with the term, I will respond. I have stated,
() 24 and I will resta te, that when I review a study like 25 Limerick, I taka -- I draw comfort from a number of O
f ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6003 outcomes.
(]) 1 2 One of them is what I have repeatedly stated g- 3 is a risk outlyer. I am very sensitive to risk V 4 outlyers. And one of the first things I do when I look 5 at the results of PBA studies is to look and see whether 6 some particular risk outlyer has been identified. And 7 the reason being is that once upon a time when WASH-1400 8 was issued, there was a belief that one can look at two 9 reactors, being reactors Aike Surrey and Peach Bottom, 10 and one can draw the conclusion for 100 reactors.
11 Well, that was relatively naive, and we have
- 12. lesrned since that the designs are very important. And 13 it has also been repeatedly shown that various -
0 14 combinations of vendors and AEs come up with some unique 15 design solutions and sometimes these unique design 16 solutions, or what I call some idiosyncracies, turn out 17 to create the potential for outlyers.
l 18 And hence, I look for those, simply to see 19 whether there is some particular idiosyncracy. And we 20 have, by the way, seen those in other studies outside l
l 21 the ones we talked about here.
I 22 The second aspect is that when I see an 23 estimate which is in the same ballpark as I was saying
() 24 with WASH-1400 as f ar as the mean core-melt frequency 25 and the uncertainty bands that give me some kind of a l
(3) l ALDERSON REPORTING COMPANY. INC.
400 VIRGINIA AVE., S W., WASHINGTON, D.C. 20024 (202) 554 2345
l l
6004 Q 1 comfort because I believe that WASH-1400 has been so 2 thoroughly reviewed within this country and elsewhere c 3 that by now I think everybody in the business 4 practically knows it by heart, all of the assumptions 5 that are in WASH-1400, the key assumptions.
6 The third thing is that if the risk outlyers 7 have not baen identified, as there is a very good reason 8 to believe, that there are no unique systems 9 interactions of .any significant risk to the public 10 identified.
11 A (WITNESS KASCSAK) I might add just one
. 12 comment to try to put this one to bed once and for all.
13 I think that to the extent that we feel good about the l
l W 14 positive results of the Limerick PBA is really 15 two-fold. It is, one, that to the extent that the 16 plants are similar and that the BWR-IV product line is 17 applicable to Shoreham and the f act that that piece of 18 the analysis showed that there was no unique aspects of l 19 that design that contributed significantly in a unique 1
20 way to plant risk.
21 Secondly, the overall conclusions again lead 22 you to conclude the efficacy of the design process that 23 the design process does, in fact, identify and design 24 properly the plants in such a way that no unacceptable 25 systems interactions exist. And this is another l
O ALDERSON REPORTING COMPANY,'NC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6005 1 verification that, at least for Limerick, that appeared 2 to be the case.
3 And as we said before, that same design O 4 process, the procedures and the personnel in many cases 5 were the same procedures and processes that were used in 6 the design of Shoreham. So those same positive results 7 you would expect to also be embedded into the Shoreham 8 design.
9 ,
10 11 12 13 -
14 15 16 17 18 19 20 21 22 23 24 25 O
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6006 Dr. Burns, a number of questions were asked
(]) 1 Q 2 concerning comparison of accident classes at Shoreham 3 and Limerick, and the Class 4 was listed in both as 4 involving ATWS. Can the differences in the class 5 frequencies, the Class 4 frequencies between Shoreham 6 and Limerick be attributed to the fact that the Lim.erick 7 PRA assumes that in ATWS alternative 3(a) would be 8 installed a t Limerick?
9 A (WITNESS BURNS) There is some contribution to 10 that difference due to the change from alternate 2(a) to 11 3(a), but it is only a partial factor at Limerick in 12 changing the risk between Shoreham and Limerick in what 13 is referred to as Class 4 in bdth plants.
8 14 0 Well, is it a large or a small contributor?
15 (Pause.)
16 A (WITNESS BURNS) I guess I would characterize 17 it as -- it would be easier for me to give you the 18 approximate percentage rather than small or large.
19 0 So ahead.
20 A (WITNESS BURNS) I think that the difference 21 q uo ted was approxisately a f. actor of 30 difference in 22 the Class 4 between Shoreham and Limerick, that due to 23 Class -- that due to the fact that the SLC system is (O) 24 changed from a 2(a) to 3(a) is worth approximately a 25 factor of live because of that.
i l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 l
6007 1 JUDGE BRENNER: M r. Ellis, excuse me. I had a
(}
2 recollection -- and I think I confirmed it by 1 coking
,) 3 back at Suffolk County Exhibits 23 and 25 -- 23 includes G
l 4 page 4-3 from the Shoreham PRA, and 25 is page 3-96 from 5 the Limerick PRA. And I certainly recall testimony 6 yesterday that the classes were defined differently 7 between the two studies. In addition, ATWS is listed as 8 the brief summary caption on the bar graph for both 9 Classes 3 and 4 in Limerick, whereas it's only listed as 10 the short summary title for Class 4 for Shoreham.
11 Were you comparing apples and oranges to some 12 extent, and could you briefly explain, Dr. Burns, why 13 ATWS is listed both times?
14 A (WITNESS BURNS) Limerick has some unique l
l 15 plant characteristics which have resulted in some of the l 16 postulated events of failure to scram coupled with 17 f ailure of poison injection to be more benign possibly 18 than those that are included in Class 4. And those 19 events were incorporated into Class 3 at Limerick 20 because the radionuclide source terms were lower for 21 Class 3, and the results of the accident sequence were 22 closer to what was referred to there as Cla ss 3.
23 However, what turns out to be one of the dominant i 24 contributors to public risk in terms of early fatalities 25 is Class 4 at Limerick, and Class 4 at Shoreham includes O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 l
6008
() 1 approximately the same radionuclide source term as at 2 Limerick, however, and it does include the ATWS 3 sequences that lead to that. However, at Limerick there 4 were or at Shoreham there are additional failures that 5 lead to Class 4 that are included at Shoreham that are 6 not at Limerick, and there are additional things at 7 Limerick such as venting, in this particula r example, 8 which are included at Limerick that are not at 9 Shoreham. Therefore, one cannot compare these two bar 10 graphs directly and get the impact of changing f rom 11 alternate 3(a) to 2( a) only.
12 JUDGE BRENNER: I guess we should clarify that 13 when we talk *about 3(a) and 2(a) we're talking about the 14 alternate proposals in the Commission's proposed 15 rulemaking, correct?
16 WITNESS BURNS: For ATWS mitigation, yes, sir.
17 JUDGE BRENNER4 I don't want to go into it.
18 We will hear about it in the ATWS contention, I'm sure, 19 but what differences from Limerick did you have in mind l 20 that made it more benign so as to fall in Class 3 other 21 than the difference in the proposed regulation?
22 WITNESS BURNS: Well, at Limerick there was --
23 or in this analysis right here this includes containment fh
(_) 24 overpressure relief. That is one issue. The other 25 issue is that there are some set point trips at Limerick l
()
ALDERSON REPORTING COMPANY,INC, l
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6000 1 which are not at Shorehan, and these trips actually 2 terminate flow to the vessel, sacrifice the core prior 3 to containment. So there is a dif ferent philosophy in L
4 how you run the high pressure systems at Limerick.
5 BY MR. ELLIS: (Resuming) 6 0 Dr. Burns, in, answer to Judge Brenner you just 7 indicated that one of the differences was the 8 containment overpressure relief vent. It is true, isn't 9 it, that that particular change, though, as modeled in 10 the Limeri:k PRA say not even be installed at Limerick, 11 or do you know whether it's going to be installed at 12 Limerick?
13 -
A (WITNESS BURNS) My understanding is that the 14 current revision of the PRA which was issued in the 15 spring does not include that, and Philadelphia Electric 16 does not plan to include that design feature.
17 Q So putting to one side all of the other 18 reasons why y'ou indicated that you couldn't compare the 19 two bar graphs, that would still bring the number of 30 20 down, wouldn't it?
21 A (WITNESS BURNS) Yes, sir.
22 0 And even assuming or putting to one side the 23 differences that you indicated precluded comparison of 24 the bar graphs, assuming the difference of 30, as I 25 understand your testimony, approximately 16 percent of O
G ALDERSON REPORTING COMPANY,INC, 400 VIRGINlA AVE., S.W WASHINGTON, D.C. 20024 (202) 554-2345
6010
() 1 it, if you could make that comparison, would be 2 attributable to the 3(a) ATWS alternative. .
3 A (WITNESS BURNS) The 16 percent, how did you 4 arrive at that?
5 JUDGE J3RDANs I thought he said a factor of 6 five. ,
7 WITNESS BURNS: Yes, sir.
8 BY MR. ELLIS: (Resuming) 9 0 Well, would five over 30 not give you 16 10 percent roughly, 16 2/3?
11 A (WITNESS BURNS) I would rather say that 12 instead of a factor of 30 dif'ference changing from 3(a) 13 to 2(a) is a factor of five of that 30.
14 0 Five of that 30. I was just putting five over 15 the 30.
16 JUDGE JORDANS We like the original numbers of 17 Dr. Burns a little better.
18 MR. ELLIS: I just assumed that since there 19 was a reference to percentage that we could get the 20 percentage.
21 JUDGE JORDAN Yes. We heartily feel -- and 22 do you agree, Dr. Burns -- that the 16 percent arrived
! 23 at is hardly applicable in this' case. It is much more
() 24 significant to have the absolute numbers.
25 WITNESS BURNS Yes, sir. I agree with you.
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
6011
(} 1 (Laughter.)
2 (Counsel for LILCO conferring.) -
g-) 3 MR. ELLIS: That completes our redirect.
U 4 (Board conferring.)
5 JUDGE BRENNER: We are going to go to 6 quest,lons from the Board starting with Judge Morris in a 7 moment, but I did want to remind the parties that there 8 were some things put aside that the parties are going to 9 come back to. One is the parties were going to confer 10 and let us know whether they believed thera was some 11 concise portion of the Shoreham PRA and the appendix 12 referred to by Mr. Kascsak ye ste rd a y tha t would be 13 useful to put into evidence to explain page 4-3 which is 14 the second page of Suffolk County Exhibit 23. So I hope 15 you are going to come back to us on that, presumably by 16 tomorrow.
17 And we were going to come back to the GE l 18 report matter, presumably tomorrow, and I certainly
- 1e expect and in fact direct counsel to confer on that in 20 advance. And the guidelines of the Clinton case may be 21 kept in mind in your discussion, ALAB 344, NRC 27. And, 22 in fact, as I recall, the County previously cited tha t 23 case in another context in this proceeding.
() 24 Maybe I didn 't give the f ull title of the 25 case. It's Illinois Power Company in a 1976 case. I O
V ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W WASHINGTON, D.C. 20024 (202) 554 2345
6012
(} 1 don't know if we will have to get into that -- maybe 2 something note informal can be worked out -- but there 3 are some guidelines in that case.
4 (Board conferring.)
5 BOARD EXAMINATION
,6 BY JUDGE MORRIS:
7 Q The witnesses will be glad to hear that quite 8 a few of the questions I had written down have been 9 answered, so at least we won't have to cover that ground 10 again.
11 Gentlemen, in your testimony prefiled on page 12 97 and 98 you state that, "Other independent assessments 13 of possible methodologies have concluded that event l
14 tree / fault tree methods are preferable to others 15 available." And the references, I believe, are to the 16 four national laboratory studies.
17 It is correct that you agree with the 18 conclusions of those studies? You have examined those 19 four references and you agree with the con:1usions l
20 therein? -
21 A (WITNESS BURNS) I agree that it is the best 22 method available today, yes, sir.
l 23
/^
(,)3 24 25 O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6013 0 Well, the best -- so you are equating best to
(]) 1 2 preferable to others available?
3 A (WITNESS BURNS) Yes, sir. For assessing 4 systems interaction.
5 0 On page 104 you state that SAI has identified 6 a number of potential system interfaces and dependencies 7 which were then incorporated into the Shoreham PRA 8 f ramework in a plant-specific manner so tha t they could 9 be reviewed in a consistent framework along with other to contributors to risk. And I think you have probably 11 described examples of that.
12 Would you have anything you would like to add 13 on that subject at this time?
14 A (WITNESS BURNS) No, sir. I could add 15 additional examples.
16 0 Again, on page 104 you state that operating 17 experience must be screened not only for serious event 18 sequences, but also precursors. Are you familiar with a 19 draft report which is NUREG/CR-2497, Volume 1? It is an 20 OCI Analysis / SIC-182, and the title is " Precursors to 21 Potential Severe Core Damage Accidents: 1969 to 1979, A 22 Status Report." The authors are Hinarick and Kukielka.
l
! 23 I recognize this is a draf t report; however, it has been
() 24 referenced in the open literature in Nuclear Safety.
i 25 The March-April issue of 1982.
O ALDERSON REPORTING COMPANY,INC, f 400 VIRGINfA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 f
6014
(} 1 That is a long description, and again, the 2 question is: Are you familiar with that re por t?
3 JUDGE JORDAN: Could I add, I think it was O 4 prepared by AI. I'm sorry, SAI, Dr. Burns's 5 organization.
6 BY JUDGE MORRIS (Resuming):
7 0 Do you recognize those names?
8 A (WITNESS BURNS) No, sir, I am not familiar 9 with that. But I certainly will be.
/
10 ( La ugh te r. )
11 0 I highly recommend it.
12 JUDGE BRENNER: We are going to'show it to you 13 just to make sure your answer stays the same af ter 14 seeing it. And while you are doing thai, there was 15 another pending item, and that is the ACRS letter which 16 I presume somebody will come back to tomorrow, also.
17 BY JUDGE MORRIS (Resuming):
18 0 I take it you are not familiar with that 19 report, is that correct? ,
20 A (WIINESS BURNS) N o, sir, I am not familiar 21 with it.
22 0 I don't wsnt to take the time now to ask you 23 whether you are familiar with the examples that are
() 24 given. Perhaps you have seen them in another context 3 25 and have factored them into your thinking. I don't l
l l
ALDERSON REPORTING COMPANY, WC, l 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (20;f 5 54-2345 l
6015
() 1 know. But perhaps absent familiarity with tha t report, 2 can you give me an example of a precursor that you are 3 aware of that you have f actored into your analysis?
4 A (WITNESS BURNS) The Monticello scram relay 5 problem.
6 Q Any others?
7 (Panel of witnesses conferring.)
8 A (WITNESS BURNS) In the broad context of the 9 word " precursor", taking great liberty with the term, 10 possibly I would include items of loss of off-site power 11 and the loss of DC power as examples of operating 12 experience that we have fed back in. We have also 13 included some events that have occurred such as possible 14 flooding events that have occurred in the industry which 15 can be characterized as precursors of flooding events or 16 actual flooding ev en ts . We fed that in to this 17 calculation. ,
l
~
18 In addition, the Hatch incident really was l 19 used as a confirmatory tool in assessing whether the 1
20 models were correct. That incident involved the loss of 21 high pressure systems at Hatch.
22 Those are examples of what I can think about 23 right now.
3 24 0 Let me back up just a minute and &sk you if s ,/
25 you could characterize the Monticello incident a little l
($)
ALDERSoN REPORTING COMPANY,INC, l ,
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
6016 1
.better so that the record is clear.
2 , - A- (WITNESS BURNS).Gridinally, in the Kahl 3 reactor in Europe, there was a problem with the scram h 4 i elays in which 'they would rot..operatc when demanded.
j - -
5 Aftpr that incident, there p.sd an additional incident at 4
6 Monticello in which prior to operation or during startup 7 t es i.ing --
I am not exactly sure of the time, but -- a 8 similar difficulty was discovered'on their scram rs'.sys 9 which was a common cause tailure. And it' lid not occur 10 during a transient norfduring a demand, but it was an 11 indication of a potential problem.
~
12 Q. T h e.t . 'lyf f i n e .
I.thinkweunJerstand.khich 13 incident you are talking about. .Io_ fact, tid,two 8 14 incidents. One of the reasons I wanted to refer'you to
~ -<
15 this draft report was beenuse it was a survey of the
~
~16 NRC's LER reports over this pe.riod of time. But frca
' ~~
17 that survey, some-numbers-Were derived on the 18; relJ. ability of particular systems, and those numbers I 1? think probably ar3 so m e wh t' t ' dif f e ren t than wa s -
- 20 characteristic of the technology.-
21 I believe Judge Jordan has some examples that 22- have.been taken from the. report and would like to ask
- 23 you'your reaction tc these naarprs -for these sys.ters.
24 BY JUDGE JORDAN ( Gesur.in g ) :
25 0 I had in mind, for.exanole,.that they have o .
ALDERSON REPORifNG COMPANY,INC.
j - 400 viro:NIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
6017 considered the probability of simultaneous failure of
(]) 1
- 2 both the HPCI and the RCIC, which includec, of course, g-) 3 the Hatch event that you mentioned, but there have been V
4 a number of other such events. And they conclude that 5 the simultaneous failure of both would occur with a
-3 6 frequency 5f something like 4 times 10 . That is, l
i 7 per demand.
8 And I wondered if that was anything like the 9 frequency of failures that you assumed in your PRA.
10 A (WITNESS BURNS) That sounds much lower than my -
11 number. My point estimate --
12 0 You assumed that it was more likely than that?
i 13 A (WITNESS BURNS) We assumed a relatively high l .
14 unreliability of those two systems or calculated a 15 relatively high unreliability of those two systems. I
! 16 sean, they are very reliable systems, but in terms of 17 what people normally think about, it is relatively high, 1
18 and higher than this number.
19 0 I see. And perhaps your numbers, then, came 20 also from an observation of precursor events, is that 21 correct? Or was it an independent estimate ? How did l 22 you arrive at your number?
23 A (WITNESS BURNS) The HPCI and RCIC evaluation
() 24 that we performed dates back to af ter the THI incident 25 when General Electric formed a task force to look at the l
ALDERSON, REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6018
() 1 systems. And one of the primary missions was to model 2 these, using the logic diagrams. The specific models 3 that we used were f rom Peach Bottom. We had modeled the 4 Peach Bottom system, which is generally characteristic 5 of other BWR-IVs in most aspects of these two systems.
6 We then used generic data to quantify the i
7 models. In parallel with that, Genersl Electric also 8 performed a survey of the industry at that time to find 9 out if there had been incidents of failures of these 10 systems. And we compared our benchmark, our -
11 calculation, using the model with their survey results.
12 And I guess I would characterize the comparison as -- in l
l 13 terms of at that time, which was 1979 I think.
14 Using the accumulated history to tha t date, we 15 calculated the combined failure of both systems that we 16 would have seen one or two incidents by tha t time -- the 17 accumulated operating experience to date. And at that 18 time, we could identify one incident which was Hatch.
19 And we believe that that represented a fairly accurate 20 comparison of verifying that model that we were 21 developing was in tune with or represented reality; what 22 was happening in the field.
23 0 Thank you. You mentioned Peach Bottom. I
() 24 noted actually one of the precursor events tha t was 25 listed in that report occurred a t Peach Bottom, in which O
ALDERSoN REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6019 l
I 1 there was a lightening stroke that caused a fuse to 2 fail. And that apparently knocked out both the HPCI and 3 the RCIO.
4 Is this a systems interaction that is to be 5 expected? Would that be a kind of interaction that 6 would be turned up in a PRA?
7 A (WITNESS BURNS) I am not familiar with that 8 particular incident. If there was a single f use or a 9 single electrical connection between the two systems, 10 that could impact on both systems. We shou-1d have found 11 that.
12 Q That is the type of thing that a PRA is good 13 for?
14 A jWITNESS BURNS) Yes, sir.
15 JUDGE JORDAN That is all I have.
16 WITNESS J0KSIMOVICH: May I just add that Dr.
-3 17 Burns's estimate at Shoreham is 9 times 10 .
18 JUDGE JORDAN Good. That, I think, is 19 probably not too far off from the reactor safety study 20 estimates.
21 WITNESS J0KSIMOVICH4 It is within a f actor of 22 2 from which you have quoted from the Oak Ridge report.
23 JUDGE JORDANS Yes, I see.
24 ER. ELLISa Judge Jordan, if I might comment, 25 if it would be helpful, at 5683 of the June 25th O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6020 transcript, Dr. Burns was asked what the date base used
(]} 1 2 vac for the failure rates for HPSI and RCIC, and I 3 remember that because he said that was his favorite 4 question and he snswered it at length there.
5 JUDGE BRENNER: I am going to ask you ho. you 6 remembered that. Nobody asked him why it was his 7 favorite question, but I gu ess we can 't answer all 8 questions here.
9 (Laughter.)
to BY JUDGE MORRIS (Resumjig):
11 0 I realize it is unfair to ask you this 12 question, so I won 't ask a question; I will just tell 13 you what the report says, and you can comment if you .
14 wish. But I will read it. "The probabilities of severe 15 core damage associated with the precursors were also 16 used to estimate the frequency of severe core damage per 17 reactor year for the years 69 to 79. The point estimate
-3 -3 18 is between 1.7 times 10 and 4.5 times 10 per 19 reactor year."
20 I think some of us found that rather startling.
21 A (WITNESS J0KSIMOVICH) I am not familiar with 22 the last version of that report. Do they distinguish 23 between severe core damage and core melt?
( 24 25 O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D C. 20024 (202) 554-2345
6021
~T 1 0 My recollection was they did not, tha t they (G
2 don 't distinguish between the two. But don't hold me to 3 that, and I don't want to take time to check it. I 4 wanted to go on, if I might.
5 The reason that they seem to explain these 6 rather high numbers is that they include contributions 7 from three major events. One is the loss of feedvater 8 in stuck-open relief valve at Three Nile Island 2; and 9 two is the loss of non-nuclear instrumentation at Rancho 10 Seco; and three is the fire in the cable-spreading room 11 at Browns Ferry 1.
12 So this leads to a question. Have those 13 accident sequences been thoroughly considered by you and 8 14 taken into account in the Shoreham PRA?
15 (Panel of witnesses conferring.)
16 A (WITNESS BURNS) Before you told me the origin 17 of what they said, I was going to say that many of these 18 incidents were in older plants, I would suspect, and 19 that those numbers that are quoted there are not much 20 different than what I have seen for Big Rock Point, 21 which is a very old plant, designed to early criterion.
22 So I guess I wasn't terribly surprised by that 23 number, except now the reasons that they give are a
() 24 little strange, in that the first two are BEW reactors 25 and B&W reactors I think are significantly different
(
ALDERSON REPORTING COMPANY,INC, I 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345 l
6022 than, certainly than BWR's in their arrangement, and
(]) 1 2 even from the Westinghouse design to the BCW design my 3 perception is that they are significantly different.
,O 4 And I don 't think -- I would not draw any conclusions on 5 the industry in general based upon only BCW incidents.
6 Those are important f acts to know and we do 7 know those, and we have incorporated to the best of our 8 ability those types of events at Shoreham. But there 9 were unique, I think unique features of the BEW plant 10 that led to both of these incidents, contributed to both 11 of these incidents.
12 The third example we have not treated in the 13 Shoreham PRA, it being an external event.. There have 14 been determinis tic studies done on this particular 15 problem that have shown that the Shoreham plant has a
~
16 different arrangement and different characteristics than 17 Browns, Ferry.
18 Many of the recommendations that came out of 19 the incident at Browns Ferry have been incorporated into 20 the Shoreham plant. But I guess I can't speak to the 21 third one other than to say tha t we have not performed 22 that event as an initiater in this PRA.
23 0 I would infer from that that you would not
() 24 sgree with an apparent conclusion in this report, which 25 is as follows: "No apparent differences exist between ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6023 1 plant types and among vendors, architect-engineers, and
(}
2 plant power ratings."
3 A (WITNESS BURNS) I would say that that 4 conclusion, from my knowledge, is probably totally 5 incorrect.
6 0 Yes, this relates to accident sequence 7 precursor highlights.
8 A (WITNESS BURNS) It relates to?
9 Q Accident sequence precursor study highlights, 10 this report being a study of precursors. This is a 11 conclusion in that frama of reference.
12 But from that they conclude that, with respect 13 to the precursors that they looked at, that there were 14 no apparent differences between the plant types, 15 vendors, architect-engineers and power plant ratings.
16 And I guess your disagreement really is that that 17 statement probably doesn't apply to the risks from 18 operating plants.
19 A (WITNESS JOKSIMOVICH) If I may comment, I'm 20 astounded by their conclusions, because when we talk 21 about precursors we talk about precursors for particular 22 plants, and I don't think that Rancho Seco, an event 23 which I have studied at some length, could be in any way
() 24 interpreted as being a precursor for Shoreham.
25 And with regard to TMI, I think we all know O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6024 1 the story there. And now with regard to Browns Ferry, 2 we talk about from the standpoint of fire protection and 1 l
3 cable separation, and we're talking about an obsolete l O 4 plant.
5 So I think their conclusions are really wild.
6 Q Well, let me read you one other conclusion.
7 (Laughter.)
8 A (WITNESS J0KSIMOVICH) Is it as good?
9 (Laughter.)
10 0 Not quite.
11 "No variation with plant age can be 12 demonstrated in the number of significant events." Is 13 it your experience, or do you take into account in your 14 PRA any variation in reliability of the components and 15 systems as a function of age?
16 A (WITNESS BURNS) I would have to say tha t, 17 again, that conclusion is totally inconsistent with 18 o pe ra ting experience, in that there is definitely a 19 trend in plant age, there is definitely indications that 20 there are higher frequencies of incidents in the initial 21 year, even the initial two years of operation, than in 22 subsequent years.
23 Once a plant matures out beyond the first 24 refueling, there has been, historically been, 25 improvement in the plant performance and in the number O
O l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
l 6025 I 1
of incidents, a reduction in the numbar of incidents
(]) 1 2 that occur.
- 3 (Board conferring.)
4 JUDGE BRENNER: Mr. Ellis, in addition to the 5 fact that other parties might have follow-up questions, 6 I also recall that you wanted to have these witnesses 7 here during the questioning on PRA matters of the 8 Staff's panel. So I assume these witnesses are going to 9 be here tomorrow in any event; is that correct?
10 MR. ELLIS: Yes, sir.
11 JUDGE BRENNER: The reason I asked is, we 12 would have considered running later in an attempt to 13 finish up if that would have made a difference. But we l
I 14 vill continue for a little bit, but we will stop at some l
15 point shortly, since it won't make a difference in terms 16 of the witnesses leaving tonight.
17 (Pause.)
l 18 BY JUDGE MORRIS (Resuming) 19 Q Leaving the draft with which you are 20 unfamiliar and coming back to your testimony, with which 21 I assume you are familiar, on page 117, the second 22 pa ra g ra ph , the last sentence says: "The PRA did not 23 reveal any non-safety-related systems, whether used
() 24 normally or in the E0P, that have an unreliability that 25 would jeopardize plant safety."
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W , WASHINGTON, D.C. 20024 (202) 554 2345
6026
(} 1 I'm sorrf, that's "The PRA did not reveal." I 2 misspoke.
3 What sort of criterion did you use to say yes v
4 or no on whether plant safety was jeopardized?
5 MS. LETSCHE: Excuse me, Judge Brenner.
6 Perhaps my copy is incorrect, but my copy of the 7 testimony indicates that particular sentence was
~
8 stricken from the testimony.
9 JUDGE BRENNER: Now that you mention it, so 10 does mine.
11 MR. ELLISa We have no objection to the 12 question. We would be delighted to have it go forward, 13 because we think the inquiry has gone into numbers and W 14 other things and we think it should be permitted now.
15 JUDGE BRENNERa Give me a moment.
16 (Board conferring.)
17 JUDGE BRENNERa It is on the borderline 18 between the distinction we attempted to draw in our 19 ruling with regard to further defining that term, 20 " application of the methodology," but it isn't worth, in 21 Judge Morris' opinion, pushing against the border. So, 22 consistent with having that portion stricken, we will 23 strike the portion of the question asked and get at it
() 24 ano ther way, in fact.
25 JUDGE MORRIS: In a different context, O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6027
{} 1 perhaps.
2 BY JUDGE MORRIS: (Resuming) 3 0 Have you gentlemen had a chance to review the 4 Staff testimony on this contention, particularly the PRA 1
5 part and the systems interactions part?
6 A (WITNESS J0KSIMOVICH) I have done that a l
7 couple of weeks ago.
8 A (WITNESS BURNS) Yes, sir, I have reviewed it, 9 but not recently.
to 0 I read that testimony in summary -- and again, 11 I run the danger of mischaracterizing it -- as 12 ind ica ting that the Staff seems to f avor " dependency 13 methods" as the ptimary approach to use, with fault l 14 trees as a supplement, in identifying safety 15 interactions.
16 Do you agree or disagree with that?
17 A (WITNESS J0KSIMOVICH) I would have l 18 interpreted the Staf f as arriving at that, and there are l
19 probably factions which do that.
l 20 JUDGE BRENNERa Well, we are talking about the 21 testimony that has been presented here.
22 WITNESS J0KSIMOVICH: I will have to refresh I
23 my memory.
() 24 JUDGE BRENNER: Maybe we will come back to In fact, maybe LILCO's counsel i
25 that tomorrow morning.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6028
{} 1 can bring us back to it tomorrow.
2 BY JUDGE MORRIS: (Resuming) 3 0 Is it your opinion that the PRA done for 4 Shoreham would identify system interactions that would 5 aggrava te accident conditions, as well as those which 6 might initiate accident sequences?
7 A (WITNESS BURNS) Yes, sir, I believe that both
~
8 of those would be identified.
9 A (WITNESS J0KSIMOVICH) I agree with that.
to 0 Would you agree that you would identify those 11 situations which would lead to exceeding th e 12 capabilities of the installed safety systems?
13 (Pause.)
14 A (WITNESS BURNS) Exceeding the capability in 15 what sense, in terms of capacity or in terms of whether 16 they are there or not there?
17 0 rheir' ability to perform their function.
18 A (WITNESS BURNS) We definitely evaluate those 19 conditions in which a system would be incapable of 20 performing its function, and in those situations the 21 system is considered to be failed with a probability of 22 one. So there are sequences that safety systems or 23 non-safety systems may be adversely impacted by what is
( 24 going on in the plant and their capabilities, the 25 capabilities to operate or perform are compromised and O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6029 ,
i 1 therefore sre not -- f rom that standpoint are considered
{a}
2 to be failed or not useful in preserving the integrity 3 of the plant.
O 4 Q The background for this question relates to 5 Staff testimony and that part in which it has minutes of 6 a meeting with the Power Authority of the State of New 7 York and a discussion of safety interactions, and there 8 are some comments about what 'the safety interactions 9 study should accomplish, with the implication that 10 perhaps the PRA does not.
11 So that the question that I would lea ve up to 12 you is whether or not, having read that memorandum or 13 the minutes of the meeting which discusses ingredients 14 of a safety interactions study, as they are 15 characterized, whether or not the Shoreham PRA has in 16 fact included those kinds of studies or has not.
17 JUDGE BRENNER: Just for the sake of reference 18 while the witnesses are looking, the document is 19 enclosure one to the memo from Mr. Dircks to Mr.
20 Shewmon, which the Staf f is going to move into evidence 21 as attached to their testimony. And enclosure one 22 consists of an October 20th, 1981, memorandum to Mr.
23 Tha dani , T-h -s -d-s -n -1, f rom Mr . Conra n .
24 ( Pause. )
25 BY JUDGE MORRIS 4 (Resuming)
O l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W, WASHINGTON D.C. 20024 (202) 554-2345
6030 1 Q Maybe, since it's been a couple of weeks since a
2 you've reviewed this, I might just give you one or two 3 more questions and suggest that you refresh your memory 4 and answer in the morning.
5 On page 3 the Staff says that systems 6 interaction has an inherent value completely aside and 7 apart from PRA." My question to you would be, do you 8 agree, and if so why", and if not why not?
1 9 The Staff also says that a safety fnteraction 10 analysis must be regarded logically as a prerequisite to 11 a PRA. And again, do you agree or not, and if so why, 12 and if not why not?
13 And finally, the Staff says it is possible to 8 14 treat " induced operator error." The question is, does 15 the Shoreham PRA do this?
16 JUDGE BRENNER: We are going to break in a 17 moment. I guess beyond those particular questions, it 18 might be helpful for you to become familiar with that 19 attachment, to reacquaint yourself to that attachment to 20 the Staff's testimony and the enclosure. We certainly 21 understand that it is not your document and the attempt 22 is to see if we can focus whatever differences exist.
23 So if at any point in any questioning on the document
( 24 you can't answer because you're not sure you understand 25 what the document is saying, certainly that would be a O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6031
('T 1 perfectly adequate answer.
U 2 2ITNESS J0KSIMOVICH: If there are any other 3 questions from the Board with respect to this document, 4 I guess I would appreciate hearing it today, so that I 5 can take it into account tonight in addition to Judge 6 Morris' questions.
7 JUDGE BRENNER: Well, there may be some other 8 questions,"I'm not sure. So I can't accommodate you 9 precisely. But in general, there is a broad range of 10 sta tements in there that goes to the general point, to 11 put it charitably, uncertainty in at least the author of 12 this memo as to the adequacy of a PRA taking into 13 account systems interactions, which I suggest is not 14 fully consistent with your testimony, again to put it I
I ch a ri ta bly.
15 16 I hate to ask broad questions. We may give 17 you an opportunity tomorrow morning to tell us what you 18 think in terms of particular sections. We won't require 19 it, but if you want to offer it we will accept it.
20 Otherwise we will attempt to go th rough it.
21 It is the kind of thing I thought the Staff 22 was going to do. I don't want to belabor that, but at 23 the time we received those documents and in the
() 24 cross-examination plan we had some perceptions which 25 have turned out to be misperceptions as to what the O
j ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
6032 Staff was going to do with that document.
(]) 1 2 JUDGE MORRIS: To put it another way, you 3 people have put on a positive defense of PRA. Here is a 7)
V 4 challenge to it as a way of doing systems interactions 5 and classifications of systems. And if we lose you as 6 witnesses, we can't come back after the Staff has 7 testified. That is why we are trying to get some
~
8 information from you now.
9 JUDGE BRENNERa I don 't want to preclude that, 10 but we want to avoid it if possible. We have adapted 11 procedures at times and we may do so again. If we did 12 that, we wouldn't wait until the entire Staff testimony, 13 and that is another reason we endorse LILCO's request to 8 14 have the PRA-related matters of the Staff asked first to 15 the fullest extent practicable.
l l
16 I found an anonymous missive in our inbox 17 here, anonymous in the sense that I don't know who left l 18 it there or for what purpose. It is a June 30, 1982, 19 letter to Mr. Allan of the Institute , the IEEE, from one 20 Denwood F. Ross, Jr. Who left it there and why?
21 MR. RAWSON: I left it there, Judge. I 22 apologize. We don't have the secretarial services up 23 here to attach a letter to that.
) 24 That was a responsive letter from the Staff to 25 the IEEE. It is a response to the letter that LILCO O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
6033 1 attached as attachment 4 to its prefiled testimony. It 2 was sent out late last week. I received a copy of it 3 just ye s te rd a y morning and immediately made copies and O 4 distributed it to everyone.
5 We had planned to attach that to the rebuttal 6 testimony of Mr. Conran.
7 JUDGE BRENNER: Just for clarification, I'm 8 sure this was clear to everybody, but Dr. Burns, the 9 Hatch incident that you referred to, was that the June 10 1980 incident that was the subject of one of the 11 county's exhibits for identification? Was it the June 12 1980 incident involving the high pressure core cooling ,
l 13 system malfunction? I think it was Suffolk County ^
l 14 Exhibit 21 for identification. Do you know?
15 Mr. Haguire talked about it.
16 WITNESS BURNS: Yes, sir.
l 17 JUDGE BRENNERs I wanted to make sure Hatch 18 wasn't twice blessed, like Browns Ferry and some other i 19 plants.
l l 20 All right. We will adjourn for the day and we l
21 will resume at 9:00 o' clock tomorrow morning.
22 (Whereupon, at 5:15 p.m., the hearing in the 23 above-entitled matter was recessed, to reconvene at 9:00
() 24 a.m. on Thursday, July 8, 1982.)
25 ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W., WASHINGTON. D.C. 20024 (202) 554-2345
)
O ,
""c^^ "* ""'^' " c ""I' * * "
This is to certify that the attached proceedings before the BEFORE THE ATOMIC SAFETY & LICENSING BOARD _
in the ::2atter of:. Long Island Lighting Company (Shoreham Nuclear Power station)
Da.te of Proceeding: July 7, 1982 Docket s
llumber: 50-322 OL
. Place of Proceeding: Riverhead, New York were held as heroic appears, "and. that this is. the original transc:-1pt thereof for the file of the Coczmission., ,
Ray Heer Officia.T., Esporter (*ryped)
\ -
tt >
Officia M eporter (Signature) e
.u
- y- ;
~
. O'
.f i
I i