ML16008A689

Agenda - ACRS Subcommittee Meeting, April 20, 2006

ML16008A689
Person / Time
Issue date:	04/20/2006
From:	Advisory Committee on Reactor Safeguards
To:
References
NRC-981
Download: ML16008A689 (316)

Text

Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSIONTitle:Advisory Committee on Reactor Safeguards Reliability and Probabilistic Risk Assessment

SubcommitteeDocket Number:(not applicable)Location:Rockville, Maryland Date:Thursday, April 20, 2006Work Order No.:NRC-981Pages 1-315 NEAL R. GROSS AND CO., INC.

Court Reporters and Transcribers 1323 Rhode Island Avenue, N.W.

Washington, D.C. 20005 (202) 234-4433 1 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 UNITED STATES OF AMERICA 1 NUCLEAR REGULATORY COMMISSION 2+ + + + +3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS (ACRS) 4 SUBCOMMITTEE ON RELIABILITY AND RISK ASSESSMENT 5+ + + + +

6 THURSDAY, 7 APRIL 20, 2006 8+ + + + +9 ROCKVILLE, MARYLAND 10+ + + + +11The subcommittee met at the Nuclear 12Regulatory Commission, Two White Flint North, Room T-132B1, 11545 Rockville Pike, at 8:30 a.m., George E.

14 Apostolakis, Chairman, presiding.

15 COMMITTEE MEMBERS:

16 GEORGE E. APOSTOLAKIS, Chairman 17 J. SAM ARMIJO, Member 18 MARIO V. BONACA, Member 19 RICHARD S. DENNING, Member 20 THOMAS S. KRESS, Member 21 OTTO L. MAYNARD, Member 22 WILLIAM J. SHACK, Member 23 JOHN D. SIEBER, Member 24 GRAHAM B. WALLIS, Member 25 2 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 ACRS/ACNW STAFF:

1 ERIC THORNSBURY, Designated Federal Official 2 PANELISTS:

3 ALAN BEARD, GE 4 SID BHATT, GE 5 DAVID HINDS, GE 6 THEO THEOFANOUS, GE 7 RICK WACHOWIAK, GE 8 NRC STAFF:

9 MARTHA C. BARILLAS, NRR/DNRL 10 SUD BASU 11 AMY CUBBAGE, NRR 12 JIM GASLEVIC 13 LYNN MROWCA, NRR 14 BOB PALLA, NRR 15 LAUREN QUINONES, NRR/DNRL 16 LARRY ROSSBACH 17 NICK SALTOS, NRR 18 19 20 21 22 23 24 25 3 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 C O N T E N T S 1 PAGE 2 Introduction, Chairman Apostolakis .......4 3Introduction to Presentations, Amy Cubbage ...5 4 Presentation by GE 5 Overview: 6 David Hinds ...............8 7 Rick Wachowiak ..............10 8 Internal Events Risk Management and Severe Accident 9 Prevention:

10 Rick Wachowiak ..............47 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 4 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 P R O C E E D I N G S 1 (8:33 a.m.)

2CHAIRMAN APOSTOLAKIS: The meeting will 3 now come to order.

4This is a meeting of the Advisory 5Committee on Reactor Safeguards, Subcommittee on 6 Probabilistic Risk Assessment.

7 I am George Apostolakis, Chairman of the 8Subcommittee. Members in attendance are William Shack, 9Sam Armijo, Mario Bonaca, Rich Denning, Tom Kress, 10 Otto Maynard, Jack Sieber, and Graham Wallis.

11The purpose of the meeting is to begin our 12review of the ESBWR probabilistic risk assessment.

13The Subcommittee will gather information, analyze the 14relevant issues and facts, and formulate proposed 15positions and actions as appropriate for deliberation 16 by the full Committee.

17Eric Thornsbury is the Designated Federal 18 Official for this meeting.

19The rules for participation in today's 20 meeting have been announced as part of the notice of 21this meeting previously published in the Federal 22 Register on April 4, 2006.

23A transcript of the meeting is being kept 24and will be made available as stated in the Federal 25 5 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 Register notice. 1It is requested that speakers first 2identify themselves and speak with sufficient clarity 3 and volume so that it can be readily heard.

4 We have received not written comments or 5requests for time to make oral statements from members 6 of the public regarding today's meeting.

7We will now proceed with the meeting, and 8I call upon Ms. Amy Cubbage, the NRR's project 9 manager, to introduce the presentations.

10MS. CUBBAGE: Good morning. I'd just like 11to give a few opening remarks to set the stage for the 12presentations you'll be hearing from GE today and 13 tomorrow. There will be a staff presentation tomorrow 14 afternoon as well.

15The application for certification is 16submitted in August and then supplemented in 17 September-October. The application was accepted for 18 docketing on December 1st, 2005, and since that time 19we have received Revision 1 of the design control 20 document in three different pieces as listed here.

21 The one piece that has not been submitted 22 yet is Revision 1 of Chapter 19 of the DCD, which is 23 the PRA.24We did provide preliminary requests for 25 6 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433additional information to GE on severe accidents.

1Those were provided to GE in RAI letter number three, 2which was sent to them in December. That should be 3'05, a typo there, and GE is in the process of 4revising the PRA to address these RAIs and also to 5incorporate the changes that were made between 6 Revision 0 and Revision 1 of the DCD.

7 So as you can see here some of the 8chapters of the Revision 1 of the PRA have been 9submitted and have been provided to the committee.

10The additional chapters, I believe some of them are 11 coming today and others will be here within a week or 12 two.13 At that time we'll have a complete 14 Revision 1 of all the PRA documents.

15 Just the overall certification schedule.

16We're currently issuing RAIs to GE, and that will 17proceed through October '06, and then we're expecting 18all of the RAI responses to be received through 19 November '06.

20We're planning to issue the SER with open 21items in October '07, and at that point we'll begin 22the process of closing those open items and issuing 23 supplemental SERs as necessary in assumed 15 months' 24duration to complete that effort, and then we will 25 7 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433start the rulemaking period, which is assumed to last 1 12 months.

2CHAIRMAN APOSTOLAKIS: When you say 15 3 months, starting when?

4 MS. CUBBAGE: Starting with the issuance 5 of the SER with open items. So --

6CHAIRMAN APOSTOLAKIS: October '07, 15 7 months after that?

8MS. CUBBAGE: Right, and that's just an 9 assumption at this point. Until we know the number 10 and scope of open items, we won't be able to establish 11a firm schedule for that. If the number and scope of 12open items is small, we may be able to proceed quicker 13 than that.

14 CHAIRMAN APOSTOLAKIS: So we may go to 15 2009.16 MS. CUBBAGE: That's right.

17CHAIRMAN APOSTOLAKIS: And the ACRS is 18 involved there?

19MS. CUBBAGE: The ACRS would be involved.

20 Right. I would expect a lot of involvement at the SER 21with open item stage, and then as we're issuing the 22 supplements. Of course, if there's any topics of 23interest early on, we could provide more meetings like 24this to provide you with an overview of different 25 8 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 topics.1 So that's all I had.

2MEMBER WALLIS: And you said that there 3 are other presenters tomorrow afternoon?

4 MS. CUBBAGE: Yes, tomorrow afternoon.

5MEMBER WALLIS: We're due to adjourn at 6 12:15. So they may be talking to themselves.

7MS. CUBBAGE: I say afternoon. Mid-8 morning. Sorry. It's noted on your agenda.

9 And what we are doing, briefly, tomorrow 10is just going over the RAIs that we've issued, a 11summary of those, and then Office of Research is going 12to be presenting information on confirmatory severe 13 accident calculations.

14 MEMBER WALLIS: Just a core catcher?

15MS. CUBBAGE: Is Office of Research going 16 to? I don't know. That is a question for GE.

17At this time I'd like to introduce Stephen 18 Hinds to make some remarks for GE.

19MR. HINDS: Good morning. I'm David Hinds 20 from the GE ESBWR Engineering Manager.

21 I'd just like to hurriedly introduce our 22team that we have here today. We have Rick Wachowiak 23over here. He is PRA lead. He will be the main 24 speaker today.

25 9 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 And we also have Sid Bhatt over here, who 1 will also be supporting Rick and making some 2 presentations this afternoon.

3Supported by Alan Beard, basically if you 4 ask some questions that we need to point in his 5 direction supporting the presentation.

6 And then coming in later we'll have 7Theofanous, who will be supporting us on our severe 8 accident analysis.

9And we have a day and a half planned here 10with the focus, overview of our PRA as well as our 11severe accident analysis, and we'll go I suppose as 12 deep as we can within the day and a half time period, 13 and I'm sure we'll be back here to see you again.

14We look forward to sharing information 15with you here today. The PRA with the ESBWR has been 16done in parallel with the design and we're going to 17cover some of that process, but it has been a very 18interesting process using the PRA as a design tool 19such that we can incorporate risk insights into the 20design as we go along. It brings upon certain 21 challenges we're actually closing out and completing 22 in the PRA, but it's a very good design tool and 23useful in our design process, and Rick will cover that 24 in more detail.

25 10 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 So I'll turn it over to Rick.

1MR. WACHOWIAK: Good morning. I guess I'm 2 supposed to sit close to the microphone.

3CHAIRMAN APOSTOLAKIS: Do you mind 4 standing up?

5MR. WACHOWIAK: I don't mind standing up.

6CHAIRMAN APOSTOLAKIS: We want to see your 7 body language.

8MR. WACHOWIAK: I'll go ahead and start 9 from here.

10The first part of the presentation is 11 going to be an overview of what it is we're going to 12do today and tomorrow and talk a little bit about the 13philosophy of how we used the PRA as a design tool, to 14 be able to say. So the agenda for the meeting or at 15least the GE presentation, this is all printed in the 16agenda, but we want to cover an overview of how we use 17risk management. We're going to talk about severe 18accident prevention, which is pretty much the Level 1 19 PRA; severe accident mitigation, which discusses the 20various phenomena of severe accident; containment 21system performance. Once we get beyond the phenomena 22 of severe accidents, what does the containment do as 23 a system itself?

24We'll talk about our off-site consequence 25 11 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433analysis as it relates to a design PRA, a non-site 1 specific design PRA.

2Tomorrow we'll talk about external events, 3shutdown, and then conclude with some of our insights 4and other information about how we'll be proceeding as 5 we go into the future.

6The purpose of the meeting, one, to 7outline the strategy for how we use risk management 8land ESBWR design. We want to be able to demonstrate 9to you the robust nature of the ESBWR as it relates to 10 severe accidents and the way we prevent and mitigate 11 severe accidents.

12 We're also going to talk more about how we 13use the PRA as a design tool for designing and also 14 for licensing nuclear power plants.

15Now, in the DCD phase of this whole 16design, which is what we're discussing now, we have to 17build a PRA that will support the design that goes in 18and is being reviewed for the DCD, and we needed to do 19certain things. We can't do everything at this point 20because we don't know everything at this point, and we 21 may never know everything, but we get closer as time 22 goes by.23What we want to make sure we can do is 24that this PRA needs to be able to demonstrate that we 25 12 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433meet the established goals, risk goals. We want to be 1able to demonstrate that the ESBWR design is actually 2 better than what's currently out there. So not only 3meeting the goals, but we want to meet and exceed the 4 goals. It's hard to say with goals which way is 5 exceed.6Also in this process, we're extending the 7use of defense in depth into the severe accident 8 scenarios themselves, and we'll talk about that on a 9 later slide. We want to be able to identify systems 10that are important to risk and provide a basis for the 11 design reliability assurance program.

12 Those two things are some things that are 13going to be a constant dialogue with the NRC over the 14 DCD process because some of the things that you need 15 in order to identify what goes into these pieces are 16 not necessarily available to go into the analysis at 17 this point.

18 So we have to figure out how we balance 19 what we know at this time in the design versus what we 20 think it's going to be in the future and what controls 21need to be placed on how we address these things in 22the future. I think that's going to be a constant 23 dialogue, and it's not settled business yet.

24 Finally, we want to be able to provide a 25 13 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 framework for the plant specific PRA. In the end as 1we go through all of the iterations for the PRA during 2 the design phase, during the licensing, the ultimate 3output is going to be something that the utilities can 4 use in their operation of the plant. And because it 5has gone through the licensing phase, it will be 6 something that the NRC is familiar with, unlike with 7 the current plant PRAs where there was kind of, you 8 know, the plant guys knew some things, the NRC knew 9 some things, and nobody quite matched up.

10 But we should all be in sync when we get 11 through this process here.

12 MEMBER DENNING: Before you move on from 13this slide, when you talk about demonstrating ESBWR 14meets established risk goals, by that do you mean the 15 quantitative health objectives?

16 MR. WACHOWIAK: Yes, and the CDF and log 17 release frequency goal.

18MEMBER DENNING: Right. Have you 19established goals yourself that are more stringent 20 than those goals or different than those goals?

21MR. WACHOWIAK: In some cases we have, and 22it is kind of built in down here. Demonstrate that 23it's better. Let's take the core damage frequency 24 goal. The subsidiary goal is established at ten to 25 14 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 the minus four per year.

1 Well, the EPRI URD took that down another 2level, ten to the minus five per year. We still don't 3want to be in that range. We're looking at below ten 4 to the minus six for all the things that we know 5 about. We're trying to do as good as we can to be 6 below ten to the minus seven for the things that we 7 know about at this point, and we think we've 8 established that.

9 But those are not -- below the ten to the 10 minus six, it's more of a squishy goal rather than a 11hard goal. We want to get there, but that's how we're 12 using the PRA to drive us toward that range.

13And once again, remember that where we are 14now with knowing what we know at this current phase of 15the design, if our target is below ten to the minus 16seven, as things come up we have room to address them 17 and room to see how we want to proceed with those.

18CHAIRMAN APOSTOLAKIS: Which is exactly 19the point I wanted to raise. I mean, you can't 20 demonstrate that you need to establish goals because 21your PRA is necessarily incomplete, correct? I mean 22you can afford three orders of magnitude below, 23chances are you will meet it, but at this point, I 24mean, we have got knowledge that there are, you know, 25 15 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 many holes in the PRA because you don't have a plan.

1 MR. WACHOWIAK: I would agree with that.

2CHAIRMAN APOSTOLAKIS: I mean, you need a 3fire assessment. Every other sentence says, you know, 4"We don't have information. This is generic. This is 5 generic. We don't have information," which is fine.

6I mean, that's the situation, but we can't really say 7 that we're demonstrating we're meeting the goals. I 8 mean, we're doing what we can with what we have now.

9Of course, if we violate the goals now, we are in 10 trouble.11So do we have the microphone finally? Ah, 12 there you are.

13 (Discussion was held off the record.)

14MR. WACHOWIAK: All right. One of the 15things that was associated with this demonstration, 16 one, you really can't demonstrate until you're done 17 and you know everything that you don't know now, and 18 even when you get to that point, there's still the 19unknown unknowns, and you'll never get all the way 20 down. But we're talking about demonstrating using 21 what we know now.

22 There are also cases that we looked at and 23 we know that we need to know more information to get 24to there, and so what we've done in our process is 25 16 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433we've specified some design requirements that says, 1okay, the analysis is going to assume in the fire area 2it's a fire thing, but it's really in the flood 3 scenario. We're going to specify where some of the 4fire protection piping needs to be in the control 5building because we want to assure an assumption that 6we put into the flooding analysis. So we're providing 7design requirements out of the PRA to address some of 8 these unknowns at this point.

9MEMBER WALLIS: Does your PRA include 10 deliberate human actions in some way?

11 MR. WACHOWIAK: Acts of commission?

12MEMBER WALLIS: Yes. Do you have it so 13 that it's robust in terms of acts of commission?

14MR. WACHOWIAK: The current design phase, 15the current DCD PRA does not include acts of 16 commission.

17MEMBER KRESS: There was some explanation 18 for that, having to do with the fact that no operator 19 actions are required for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> or something.

20MR. WACHOWIAK: No operator actions are 21required for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, but we have to remember that no 22operator actions required doesn't mean no operator 23actions will happen. But the way that our goal is in 24designing the control systems of this passive plant is 25 17 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 that if the operators start to do something and then 1 they stop, the plant should move itself back into the 2 stable state as opposed to where in some of the acts 3of commission and existing plants, where operators 4 start getting into things they send down a different 5 path and it gets kind of unknown.

6 What we're trying to do with this design 7is make it so that if they do something, recognize 8that they're going the wrong way and go hands off 9again, it's supposed to stabilize back into the safe, 10stable state condition. We're not far enough along in 11the design of the control systems to be able to prove 12 that, but that's the goal that we have in mind.

13 The scope of the DCD PRA for internal 14 events at full power, we've got Level 1, Level 2, and 15Level 3, and you have to recognize Level 3 is not a 16real Level 3. It's a Level 3 using imaginary 17 information provided to us in the URD for population 18and things like that. And we really only look out 19about ten miles from the site boundary in addition to 20 that. So it's maybe a three minus.

21Internal events. For shutdown we've done 22Level 1 and in the process of completing a simplified 23 Level 2, which is going to be in one of these 24 submittals here that will come up shortly.

25 18 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433External events. We've done internal 1fires, flood, and high winds. As you said, on what we 2believe is a conservative bounding basis, once again, 3the details to do a detailed analysis of these are not 4 here yet.5Seismic margins on the safety systems 6we've provided, and all of this is associated for the 7 internal events at the Level 1, and we've covered in 8the internal fire and flood both full power and 9shutdown analysis. So that's also the initial Rev. 0 10that you may have seen didn't have the shutdown for 11fire and flood in it. We've completed that analysis, 12and we are in the process of writing that up, and 13 we'll talk about it a little tomorrow when we get to 14 the fire and flood, but you don't have those documents 15 yet.16 Okay. Let's talk a little bit about the 17extended defense in depth. Historically the classical 18design and analysis work that was done for previous 19plants provided defense in depth certainly, but it was 20using the design basis or single failure type of 21 assumptions.

22For an accident you have an accident under 23the parameters and a single failure, and then you make 24sure that you have defense in depth associated with 25 19 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433providing the fuel barrier, providing the reactor 1coolant boundary barrier, providing the containment 2 barrier, still under that whole same framework.

3 Here what we've done is we've moved that 4on into the severe accident arena where we're looking 5at multiple failures of maybe components within the 6same systems or components across barriers, and 7looking at how we can provide defense in depth against 8 things that went beyond what were looked at before.

9And I kind of say this in that the main 10objective is to address common cause type failures.

11 I'll get to the sub-bullets here in a second.

12We also look at defense in depth on the 13 containment side, not only given a degraded core 14that's still in the vessel, which was historically 15done for defense in depth, but now we're looking at 16what kind of protection we have for core in the floor 17 type scenarios, and we'll get to some of those later 18this afternoon and talk about the areas where we've 19 addressed that.

20 Now, one of the places where we're using 21 the PRA as a design tool is in this area of the 22extended defense in depth. How is it that we can 23protect against some of these multiple failure 24 scenarios?

25 20 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Historically when a plant came across some 1common cause failure issue, the only option it really 2 had was to do an augmented QA, if you will, on those 3 components that you may see common cause failures 4 there.5Well, we are in the design process. We 6have the luxury of doing something else in addition to 7that and adding diversity to our systems to try to 8eliminate some of the common cause or eliminate the 9effects, strong effect, of some of the common cause 10failures, and that's something that because we're 11 using the tool this early, we can cost effectively 12 provide that.

13CHAIRMAN APOSTOLAKIS: Are you coming back 14 to this issue later?

15MR. WACHOWIAK: I didn't have any specific 16 bullets on that. So --

17 CHAIRMAN APOSTOLAKIS: Well, it would be 18 nice to see an example.

19MEMBER DENNING: Specifically what were 20 you looking for, George?

21 MR. WACHOWIAK: A specific example?

22 CHAIRMAN APOSTOLAKIS: Yeah. I mean, --

23MR. WACHOWIAK: You know, actually in the 24next presentation I do talk about how we use a 25 21 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433combination of passive and active systems and diverse 1control systems. So I think I have an example there.

2 So we'll get into that.

3MEMBER ARMIJO: But was that an output of 4this process or was that already going in and you had 5 planned to do that? In other words, are you really 6using the PRA to gain insights that will help you 7 create diversity that pays off?

8MR. WACHOWIAK: The answer to that is most 9of the time. Because there are other things that are 10in the we'll say the different requirements documents 11 that are out there that say, well, you've got to look 12 at diversity.

13MEMBER ARMIJO: Right, generally speaking.

14MR. WACHOWIAK: So if we hadn't done a 15 PRA, we probably would have gotten there anyway, but 16in general, those documents to some degree came out of 17 previous risk analysis. So it's kind of in there.

18 However, where we are doing this is when 19we say -- when we look at what the PRA is telling us.

20Here's a common cause failure that we need to address.

21 We go back and we say, "What kind of diversity do we 22 have in the design to address things like that?" 23And especially in the instrument and 24control system area, we did use the PRA to define 25 22 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433which instrument and control systems themselves needed 1to be diverse from the other instrument and control 2 system.3MEMBER KRESS: How did you quantify the 4effect of that diversity? Did you change the beta 5 value?6MR. WACHOWIAK: Basically, that's what we 7 did.8MEMBER KRESS: But you had no way to know 9 what to change it to?

10MR. WACHOWIAK: At this point in the 11design and procurement, yes, it was looking 12 conceptually.

13 MEMBER KRESS: So you use expert opinion 14 or something?

15MR. WACHOWIAK: Expert opinion.

16Conceptually what would the effect of using diverse 17control systems have on this, and conversely, what was 18 the effect of saying that we don't need that diversity 19 requirement here? What would that do to us in terms 20 of our design PRA?

21 Yes.22MEMBER SIEBER: I was going to ask a 23question about diversity in the INC area. My question 24 really goes to the extent to which you use diversity.

25 23 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433For example, in the digital INC scheme, you could have 1diversity in computer, here's Train A, here's Train B, 2here's Train C. But they could all use common 3 software, which sort of defeats the principle of 4diversity because if there's a mistake here and you 5 replicate it over here, a mistake in both places and 6 it will fail both places.

7To what extent have you fleshed out the 8 degree to which diversity would be required not only 9in higher order, but also in software and techniques, 10 databases, et cetera?

11MR. WACHOWIAK: We've looked at basically 12 all of those types of issues. We are specifying the 13 two INC systems need to be diverse. What we mean by 14diverse there is different hardware platform, 15different vendor. I think it would be different 16vendor, different operating system in some case. It's 17going to be different -- I've already covered 18 hardware.19 So we did address those things. Now, is 20it possible that some of the different diverse INC 21systems could have some overlap? And the answer there 22 is yes.23 But the question then is: where is that 24 appropriate? Where we are in the design phase on that 25 24 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433right now is we've got, if you will, a diversity 1matrix that the INC guys and the procurement guys are 2looking at, which is the kind of diversity we want to 3have in this system, and they're in the process of 4evaluating different vendors under a multitude of 5different criteria, including the diversity criteria, 6to try to assign the correct vendor system hardware 7 for each of those different systems, and that's 8 ongoing at this point.

9 MEMBER SIEBER: When you finally certify 10ESBWR, will the INC portion be included in that 11certification, or would that be done at the COL stage?

12 MR. WACHOWIAK: I guess that's --

13 MEMBER SIEBER: Or don't you know?

14MR. WACHOWIAK: I'm going to have to defer 15because those are policy decisions, and I don't get to 16 make those.

17 MEMBER SIEBER: Well, make it, you know.

18 (Laughter.)

19 MR. HINDS: Hi. This is David Hinds.

20As much as possible, the INC system is 21part of the certification, but we are using the DAC 22approach, or design acceptance criteria approach, but 23 we're moving as rapidly as possible to close as much 24of the DAC or design acceptance criteria open issues, 25 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433and it will be flowing through certification and some 1 of it into COL as well.

2But the major issues that affect, I guess, 3the essence of your question and diversity, we intend 4to close that as soon as possible, but we did take the 5 back-up approach. So some of that is going on as we 6 speak.7MEMBER SIEBER: I can see why you would do 8 that, because if you specify today what you would do 9 by tomorrow, it would be obsolete.

10MR. HINDS: That's the reason for the DAC 11 approach. The design acceptance criteria, for anyone 12that's not aware of what I'm speaking of, defining the 13design in the form of a criteria as opposed to just 14the end result of we selected this piece of equipment 15because, as you say, the INC system has become 16obsolete rapidly. So we're defining the criteria, and 17then as rapidly as we can we're filling in details 18that can help us to firmly answer questions such as 19this, the defense in depth, although we have to 20 maintain a certain amount of flexibility due to 21obsolescence of software and hardware, and that's the 22 balancing act we're working with in the INC system.

23 MEMBER SIEBER: Okay. Thank you.

24MEMBER DENNING: Stay there just a second 25 26 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433because you may prefer to answer this question, and 1that is from your perception, what are the regulatory 2 implications of this extension into the control of 3severe accident processes? Specifically, I'm 4wondering about things like as far as the core catcher 5is concerned where there might be a lot of 6phenomenological uncertainty that could affect our 7perception of what the probability of failure that is.

8I mean, it's possible that you could say, 9well, it doesn't need a high confidence or a low 10failure probability because we've got a lot of margin 11in our risk space. Whereas, our perception of safety 12 systems from the conventional view is that they have 13 to have very high likelihood of success.

14When you get into the domain of core 15 catchers and things like that, from a regulatory 16viewpoint, what kind of criterion do you think are in 17front of you? Do you have to really demonstrate with 18high confidence the core catcher will work or is it 19 really just an element of defense in depth?

20 MR. HINDS: Well, I guess I'll start and 21let Rick get into more details, but my view on devices 22 such as that is that it is very much an extension of 23the safety of the plant and taken into another step 24beyond where the current generation of plants are. So 25 27 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433you certainly can make I'll say a somewhat valid 1argument that the reliability of those systems because 2 they're much behind the front line as opposed to the 3typical safety systems which are in the plants today 4front line systems, that the reliability would be 5 different.

6 Rick, if you want to jump in there as far 7as probability and any discussions you have related to 8reliability and probability of the core catcher or the 9 BiMAC.10 MR. WACHOWIAK: Right. At this point in 11 time what we have said is that the BiMAC itself, the 12core catcher for those who haven't seen the future 13presentations here, we believe that it's a non-safety 14 component. At this point it will be treated as a 15written system, which means that we will have some 16kind of reliability controls or availability-17reliability controls on it. That hasn't been defined 18 yet, what needs to be controlled.

19 Now, I think your specific question gets 20to the uncertainty of the phenomena of how this 21device, which effectively nobody has seen before --

22what's the confidence that we have that it's going to 23 work, and how much confidence do we actually need to 24 have to show that it's going to work?

25 28 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 And I'm trying to think if we have this 1 anywhere else in the presentation. I know I don't 2 have it in one of mine. Theo may have it there.

3We want to remember that the BiMAC itself 4was added to the floor of the containment because 5chiefly to address an uncertainty. In the previous 6 ALWR design that GE had, ABWR, we showed that at least 7 at that point in time, we showed that if we could get 8 water on top of the core in the lower dry well and it 9was spread to a large enough degree, that we would be 10able to prevent continued core concrete interaction 11 and prevent the base MAAP penetration by the melt.

12 There have been uncertainties associated 13with that. I don't think that that point has been 14refuted, but it's just not certain whether that's 15 going to happen in all situations.

16So what we've done is we've added the 17 BiMAC as another layer of protection to address that 18kind of uncertainty. So does the BiMAC have to be 19 perfect? Well, it doesn't change the fact that the 20floor and spreading is still there, and we should 21 still in most cases be able to cool pool the corium 22from the overlying pool, but it's there mainly to 23address those areas where we're uncertain if that was 24 going to work.

25 29 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 So to get back to the point, it's not 1there as a replacement for what was done in the past.

2It was there to augment what was done in the past. So 3for that matter I see it as an augmentation, and we 4don't have to be 100 percent certain. We should be 5able to show that within a fairly large band of 6 certainty, this is going to be a good design.

7 Does that answer your question?

8MEMBER DENNING: Not totally, but I 9 understand.

10MEMBER WALLIS: I'm not sure it does 11because you're sort of qualitative, but your PRA says 12it's going to work with 99 percent effectiveness.

13That's a pretty high effectiveness for something 14 that's so unusual 15MR. WACHOWIAK: Okay, and we'll talk about 16that in the presentation after lunch about how we 17determine that 99 percent effectiveness. Based on our 18evaluation and calculations, we think it's better than 1999 percent, but we've backed off on that mainly for 20 the purpose of -

21MEMBER WALLIS: How many tests did you do 22 to verify this?

23 MEMBER SIEBER: Only one.

24 MR. WACHOWIAK: That being said, when we 25 30 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433said it was 99 percent effective, you also notice that 1in there we didn't really even address, you know, what 2if it's not there. How effective would the 3 containment be if it's not there?

4And in one of the upcoming chapters 5 unfortunately that you don't have yet, we will 6 specifically be answering that question.

7 CHAIRMAN APOSTOLAKIS: Can we go on?

8 MR. WACHOWIAK: I think so.

9 Okay. I think we've talked about this 10quite a bit,b ut let me emphasize in using the PRA as 11 a design tool, our thoughts are we want to eliminate 12severe accident vulnerabilities. We want to make sure 13 that these things aren't built into the plant up 14 front. We want to get them out as we see them.

15So this provides us a systematic way of 16doing this, not just guessing at what might be a 17 vulnerability. We actually go through and look for 18 the vulnerabilities and address them in a systematic 19 way.20 MEMBER WALLIS: Now, does it play a more 21important role than DBAs? I mean, could we do away 22 with DBAs if we used PRAs as a design tool?

23 What's your experience?

24MR. WACHOWIAK: I think that we're 25 31 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433addressing things in different ways. We would have to 1do probably more things in the PRA or maybe move some 2 of the same things that we had been doing in the DBA 3 analysis into the PRA if we tried to do that. So at 4this point we start in the PRA with everything we know 5from the DBA analysis, and we have that as a given 6 that it's going to work that way.

7 And it starts us at a good point, good 8starting off point to go and do a robust analysis. If 9we did away with all of the DBA analysis, we wouldn't 10 be starting on as firm a ground with the PRA, and we 11 would have to add a lot of that back in. So I'm not 12 sure from our point of view, from the design point of 13 view, what kind of relaxation that would give.

14On the licensing side, that's up in the 15 air. You know, as long as you see the analysis, then 16maybe you have confidence in what we're doing. So 17 we're not proposing to eliminate the DBA analysis at 18 this point in time.

19We've talked a little bit about this. As 20 a matter of fact, most of the questions this morning 21 have come up. On the effectiveness of using this to 22eliminate vulnerabilities, if we don't know everything 23we need to know to remove vulnerabilities. As anybody 24 who has done PRA knows, the details tend to be where 25 32 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 you find issues that haven't surfaced before because 1 we're not looking at a simple single failure sort of 2 thing. We're looking at multiple failures and 3interactions among multiple components that could 4cause multiple failures, and typically those kinds of 5 things aren't in the details.

6However, we think we've addressed through 7the way we apply common cause and some of our 8sensitivity analyses to identify potentials for these 9failures might be, and we think we have addressed that 10through adding different diversity requirements and 11also other design requirements that come through as we 12 proceed.13 That said, the next bullet makes it very 14attractive to do this because at this point if we can 15identify things before we actually have them designed, 16especially before we have them constructed, it's much 17 easier to correct things that we would determine.

18In the end, an imperfect tool is better 19than no tool at all, better than guessing, and we 20 think that as long as we apply this in a prudent 21manner, we're not going to take things way overboard, 22 but we are going to find a number of vulnerabilities 23 that have been identified without using the tool.

24 On this next page, I just want to give my 25 33 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 perspective of where things are and how we deal with 1 the PRA in a design that is proceeding in parallel.

2On the conceptual design block down in the 3 end, what we're really trying to say is is the design 4 feasible. We don't really have a lot of actual 5 design. We've got concepts of how systems might work.

6 When we're applying a risk assessment in 7that manner, we're really doing it qualitatively.

8What kind of redundancy are we doing to need, do we 9 think we'll need for this system? Should there be 10 diversity applied to some of these things? It's all 11 in a qualitative sense.

12 And we're looking at defense in depth at 13the conceptual level. Pretty much it's based on what 14was found in the past. What problems did we have with 15previous plants, and what don't we want to have 16 problems with now?

17 As we move to the next phase where I 18 believe we are now in the qualitative design base or 19the DCD phase, the questions that we're trying to 20 answer here are can this design be licensed. Okay.

21We've specified most of our major 22 components. We now are at the point where we can do 23 a combination of qualitative and quantitative PRA to 24address specific things, defense in depth between 25 34 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 systems. We can apply common cause factors, but we're 1still in the qualitative range for some things like in 2the fire and flood type of analyses, seismic type 3 analysis.4We think that we can eliminate sequence 5type vulnerabilities, things that would be the big 6 hitters, if you will in the final PRA.

7As we move through the detailed design, 8this question comes in the later part. Will it be 9 licensed? Do we have enough information for this 10 thing to be licensed?

11By then we believe that we'll have the 12components specified. We'll be able to do a 13quantitative PRA, albeit with gaps. We won't have 14detailed evaluation of the humans that aren't trained 15on the systems yet. We won't have plant specific 16 data. We won't have some of the things that are being 17 looked at in the current PRAs.

18 I call this system level vulnerabilities 19eliminated, but it's really just more of a progression 20 till we can do more with it.

21By the time we get to construction, we end 22up with all of our components not actually just being 23specified, but being described. We can do more 24 detailed PRA, and finally get to the point where we 25 35 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 think we've addressed things.

1I get to the hypothetical out in the last 2 column here. The plant is in operation. All of the 3components are described again. They all needed to be 4described before here, but what I'm just mainly trying 5to get at with this next slide is that we are still 6working on the PRA even after the design is done, and 7 so --8MEMBER WALLIS: Well, by described, you 9mean their performance has been quantified.

10"Describe" is a very vague term. You mean you 11 actually got a measure of how they will perform.

12 MR. WACHOWIAK: Yeah, I kind of put that 13into this block, but, yeah, the performance is known.

14MEMBER DENNING: At the construction 15 design level there, where is it to become a site 16specific PRA, and where's the hand off to the utility 17 in your concept here?

18 MR. WACHOWIAK: In my concept, somewhere 19in here is where the COL application occurs and now 20this is being debated, but you know, some say it's 21 here. Some say it's here, but at the COL application, 22 it becomes a site specific PRA. That still has some 23of these issues associated with it. It's not till you 24 get to this construction level where you're actually 25 36 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433saying, "Okay. This is what's there. We've seen it.

1We know how it's going to -- you know, we know what 2 the field routing was. We know what the fragilities 3 are." 4 At that point that's kind of somewhere 5around here. The hand off to the utility we're 6looking at right here, but we're bringing the utility 7people in all along through that whole process so that 8 what we give them meets their needs.

9MEMBER WALLIS: So the PRA that we have 10 seen is where on this picture?

11 MEMBER SIEBER: The second column.

12 MEMBER WALLIS: It's qualitative?

13MR. WACHOWIAK: Qualitative and 14 quantitative.

15MEMBER WALLIS: It tends to be very 16 quantitative. It makes some assumptions and some 17bounding things, but I don't know whether it has much 18of this qualitative. I'm not quite sure what a 19 qualitative barrier is anyway.

20MEMBER SIEBER: Since you don't know what 21 the components are.

22MEMBER WALLIS: Very simplified, but it's 23 still quantitative.

24MR. WACHOWIAK: Yes. That's why I put 25 37 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 down as a combination there of the two.

1CHAIRMAN APOSTOLAKIS: Quantitative means 2 what?3MR. WACHOWIAK: Qualitative means that 4there's judgment applied to major areas. So, for 5 example, in the fire area we've said, okay, we don't 6know where the routing of the cable is. We don't know 7 what the heat loads from a specific cabinet is going 8 to be, things like that, but we do know from past 9 designs that if we confine our cables to where the 10design drawings say they're supposed to go and we put 11in typical types of cabinets that have been used in 12 plans before, we'll get this type of performance.

13And so we qualitatively bound that and 14 used that as an input to the fire risk analysis.

15CHAIRMAN APOSTOLAKIS: Maybe a better word 16would be something along the lines of "significant 17assumptions made" or something. But qualitative is a 18red flag for a log of people. Okay? And it's not 19your fault, and it doesn't really mean anything. Your 20 explanation was really something else, that you have 21 to make major assumptions because you don't know.

22MEMBER WALLIS: Simplify it. Simplify it 23 as much as --

24CHAIRMAN APOSTOLAKIS: Well, actually 25 38 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433significant assumptions I think or something along 1 those lines.

2MEMBER WALLIS: Qualitative to me means 3all waffle, and it's good enough or some sort of vague 4 statement.

5MR. WACHOWIAK: You know, I don't even 6think we're at the it's good enough in the first 7 column. You know, there's other significant judgments 8 and --9 CHAIRMAN APOSTOLAKIS: Major --

10MR. WACHOWIAK: -- thing are made, you 11 know. So I guess maybe it's a way of thinking about 12 it.13CHAIRMAN APOSTOLAKIS: Yeah, I wouldn't 14even call it judgment because judgment is everywhere.

15It's the assumptions. It's the magnitude of the 16assumptions that is different. So we need a better 17 word.18The statement is no defense in depth 19issues is not quite right. You probably mean design, 20a new wall or something, but defense in depth, I mean, 21it could be a problem, right, that is imposed? And 22that problem can be posed even when the plant is in 23 operation. And that's in the name of defense in 24 depth.25 39 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. WACHOWIAK: Right, and --

1CHAIRMAN APOSTOLAKIS: So you mean design 2 defense in depth issues are a result at that point.

3 MR. WACHOWIAK: I'll go with that.

4 CHAIRMAN APOSTOLAKIS: Well, I mean, you 5 don't have to go with that.

6 (Laughter.)

7MR. WACHOWIAK: I agree with that.

8Somewhere along in this phase here we do address which 9programmatic issues we're going to use, but that's not 10 the -- 11MEMBER ARMIJO: The design is frozen. The 12 design is finished.

13MEMBER SIEBER: In the seismic area, there 14is a point where somebody does detailed design of 15 hangers and supports.

16 MR. WACHOWIAK: Yes.

17 MEMBER SIEBER: Where is that in that 18 chart? Matched to the right or there?

19MR. WACHOWIAK: That's in the middle 20 column somewhere, I believe.

21MEMBER SIEBER: So everything is going to 22be precalculated and predesigned and no fit in the 23 field kind of --

24 MR. WACHOWIAK: That's the intent, yes.

25 40 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER SIEBER: Okay. Well, the old 1plants, all the small and medium bore piping was fit 2 in the field kind of.

3 MR. WACHOWIAK: Right.

4MEMBER SIEBER: That's why we went and had 5 700 modifications.

6 MR. WACHOWIAK: The engineering schedule 7that I work from has those activities in those to 8 complete.9 MEMBER SIEBER: That's where it will be, 10 right.11MEMBER ARMIJO: Where do you expect to be 12 for a certified design? When do you think? Is it a 13detailed design? At what point is this thing ready to 14 be certified?

15MR. WACHOWIAK: This is one of these 16things where I'm not sure that anybody has actually 17 settled on that yet, but it's --

18 MEMBER SIEBER: It's going to be between 19 these.20MR. WACHOWIAK: It's between these two 21 columns. If you talk to our friends at NEI, they say 22 the beginning of the second column.

23 MEMBER SIEBER: No, tell them no.

24MR. WACHOWIAK: It's just I think there's 25 41 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433differences of opinion on that, and we're settling in 1where it is. We certainly know that it's going to be 2at least in this column here because that's what we're 3 submitting for certified designs.

4Through the design work, we're getting 5 into this phase now. So it's somewhere in there.

6MEMBER SIEBER: You will be between the 7 two, between the design basis.

8MEMBER KRESS: Tell me. Do you see any 9 value in Level 3 PRA in the design certification 10 stage? Now, be truthful.

11MR. WACHOWIAK: With the order of 12magnitude of frequencies and releases that we're 13 looking at here, no.

14 MEMBER KRESS: It's just not going to 15 happen, is it?

16 MR. WACHOWIAK: We're showing that we're 17 very far away from any types of specified goals.

18MEMBER KRESS: That would be my guess, 19 too.20 MEMBER SIEBER: Ask me.

21MEMBER KRESS: Well, that would have been 22 my opinion. It's a subject we debate sometimes.

23MR. WACHOWIAK: We do the analysis, but we 24would be very surprised if we found that that was 25 42 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 limiting in our design.

1 MEMBER SIEBER: Do you have any estimate 2as to what the uncertainty is at this point in time in 3your analysis? I mean, you can get it to three 4decimals, but if your certainty is four orders of 5 magnitude, you know.

6 MEMBER WALLIS: Well, this is a real 7 mystery slide. I'd love you to explain this one.

8MR. WACHOWIAK: Just to answer the 9 uncertainty question, we have to look at uncertainty 10in several different ways, and so uncertainty itself, 11 I'm not sure can be a number. There are things that 12we can do quantitative, you know, like Monte Carlo 13 type uncertainty and get some information from that.

14We can do sensitivity analyses and we can get other 15 information from that.

16But I guess the question is if we say that 17core damage frequency is three times ten to the minus 18 eight, are we really talking about a three times ten 19 to the minus seven or three times ten to the minus 20 four? Do we know where it falls in that range?

21This would be a qualitative answer. I 22 think that where we are right now is that we probably 23have an order of magnitude span on what we know.

24 However, to address some of that though, some of our 25 43 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 conservative or some of the numbers that we put into 1 the analysis can compensate for some of that because 2 we know we've been on the high side with some of the 3things like initiating event frequencies and things 4 like that.

5 Also in features of the plant that we've 6 chosen not to credit in the analysis at this point.

7 So, yeah, there's some uncertainty, but it's not all 8 uncertainty toward the high end.

9CHAIRMAN APOSTOLAKIS: You say we're going 10to have a discussion of the core damage frequency 11 later?12 MR. WACHOWIAK: Yes.

13 CHAIRMAN APOSTOLAKIS: So let's go on to 14 this.15MR. WACHOWIAK: So the intent of this 16slide is to kind of address some perceptions about 17 what it is that the PRA that we have now is good for, 18 and I'm trying to think of it now in the ASME 19 capability category sort of thing.

20 Where we are now is that for some things 21in the PRA we could do anything, you know, anything up 22to the full capability Category 3. There are other 23things where we're not quite there. So it's really a 24 continuum.

25 44 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 Probably if you went point by point in 1 ASME, you'd find that we had a significant number of 2 holes because we just don't know enough now.

3 CHAIRMAN APOSTOLAKIS: Can you remind us 4 what capability means?

5MR. WACHOWIAK: Capability category, well, 6that's the category where you can use it for, you 7know, the different type of changes, and that's the 8 mindset that I had here when I was creating this.

9CHAIRMAN APOSTOLAKIS: The dark blue means 10 they're more capable?

11 MR. WACHOWIAK: Dark blue means --

12CHAIRMAN APOSTOLAKIS: Higher capability.

13MR. WACHOWIAK: It's probably more 14 weighted toward --

15CHAIRMAN APOSTOLAKIS: -- those where you 16 are.17 MR. WACHOWIAK: Where you are.

18CHAIRMAN APOSTOLAKIS: The more they color 19 them, that's where you are.

20 MR. WACHOWIAK: Yeah, and we see that as 21we move forward, we're going to be striving toward the 22 best or toward the state of the art. We're probably 23 not going to get there till well after operation.

24 We're still going to have some places where we don't 25 45 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433know everything, but the idea is that in the design 1phases, this is the right kind of mix for the DCD 2 phase, and to look at things later, we want to be at 3 the higher end.

4CHAIRMAN APOSTOLAKIS: Is the length of 5the bars an indication of the uncertainties in the 6 PRA?7MR. WACHOWIAK: No. It's an indication of 8 what information is available to apply to the models 9that are said in the standards that we should be 10 applying. So there are certain things that the 11standard says you have to do with your event tree 12 analysis. Okay?

13We've done all of those. I believe we're 14at the high end with that. There are other things 15that it says you need to do with operator actions.

16 We're at the low end for that because we have just a 17 bounding stream analysis.

18MEMBER WALLIS: You're not going to change 19 the structure of that significantly, but you will 20 change the entries. You'll change the numbers.

21 MR. WACHOWIAK: And change the details.

22 MEMBER WALLIS: But I don't think you'll 23change the structure. The PRA we've seen is probably 24going to be about the same throughout. It's just that 25 46 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 your numbers --

1MR. WACHOWIAK: We may add some detail, 2 and there's a potential to make some things clear in 3the future. We may expand the event trees to include 4more specific decision points, but the structure is 5 the same.6MEMBER WALLIS: So it's still capable.

7 It's capable now.

8 CHAIRMAN APOSTOLAKIS: I don't know that 9this kind of slide helps. May we move on? Let's just 10 go on. 11 MR. WACHOWIAK: Okay.

12CHAIRMAN APOSTOLAKIS: Let's go to 13something that we can really -- let's start seeing 14 numbers.15 MR. WACHOWIAK: We think we've got it.

16CHAIRMAN APOSTOLAKIS: So where are we 17 now? We are done with the overview?

18 MR. WACHOWIAK: Yes.

19CHAIRMAN APOSTOLAKIS: Okay. then what's 20 next, the prevention?

21 PARTICIPANT: Internal events.

22CHAIRMAN APOSTOLAKIS: We were talking 23about the qualitative. Now we can move on to 24 something quantitative.

25 47 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. WACHOWIAK: Okay. The intro is the 1 same.2 MEMBER WALLIS: Key features, do we have 3 this? Where are we? What is this? Prevention.

4MR. WACHOWIAK: Internal events risk 5 management.

6 MEMBER WALLIS: We don't have it.

7 MEMBER KRESS: Yes, you do.

8 MEMBER WALLIS: We have mitigation.

9 CHAIRMAN APOSTOLAKIS: I didn't have it.

10 Now I have it. Do you have it?

11MEMBER WALLIS: I haven't got it, by 12 George.13 MR. WACHOWIAK: It looks like this.

14CHAIRMAN APOSTOLAKIS: So we're looking at 15 ESBWR internal events risk management?

16 MR. WACHOWIAK: Yes.

17 CHAIRMAN APOSTOLAKIS: Okay.

18 MR. WACHOWIAK: The features of the plan 19are set out so that we have passive safety systems, 20 active we call asset protection systems, and support 21system diversity. What we try to do for most types of 22 systems is we have the passive function backed up by 23an active function, and then the way that the support 24systems are set up, they tend to support in a diverse 25 48 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 way.1 And this is the kind of target arrangement 2that we look at for each thing at a function by 3 function level. Then we have functions that back up 4 functions. We'll go specifically into some of those.

5CHAIRMAN APOSTOLAKIS: Let's move on.

6 Let's move on.

7 MR. WACHOWIAK: The systems that we have 8 for the different functions. Passive system, you'll 9 see that everything --

10MS. CUBBAGE: Is lined up on the handout.

11MR. WACHOWIAK: -- is lined up on the 12 handout. I think we used a different font on this 13 system.14 So anyway, we have passive systems lined 15in all of the columns, sometimes multiple passive 16 systems. We have active systems to back up all of 17 these. Reactivity control, very important system for 18the plant. We have two essentially passive systems 19 that address reactivity control.

20We have two additional active systems that 21 will provide backups to different aspects of those.

22Pressure control, once again, passive.

23You can see SRV in two columns. There's a passive 24function on it. It lifts on spring pressure here.

25 49 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433It's inactive. You open the valves, and we'll talk 1 about those.

2 Inventory control, which is a little bit 3different in this plan. We'll talk about that later, 4and high pressure. I mean, inventory control, low 5 pressure. Inventory control low pressure. Gravity 6driven cooling system would be the passive system.

7Back that up with fuel and aux. pool cooling system in 8 LPCI mode, and fire water injection.

9CHAIRMAN APOSTOLAKIS: Why do you need the 10 active backup systems? What's the whole idea there?

11 Why not the passive only?

12MEMBER SIEBER: You create an accident to 13 get the --

14 MEMBER KRESS: Asset protection.

15 MEMBER SIEBER: -- stuff to work.

16MEMBER ARMIJO: Well, you've got to 17 operate the plant.

18CHAIRMAN APOSTOLAKIS: I can ask you guys 19 over at -- can we get GE's answer?

20Why do we need active systems? The answer 21 may be simple, but --

22MR. WACHOWIAK: The answer is simple.

23It's recovery from the scenarios. The passive systems 24are extremely reliable, get you very quickly to a very 25 50 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433safe state, but what it takes to recover from that 1 state tends to be expensive.

2CHAIRMAN APOSTOLAKIS: Why is that so? I 3 mean, what do you need? Give me more detail.

4MR. WACHOWIAK: For example, when you open 5the DPVs, you've basically created a steam line break 6inside the containment, and that affects different 7components that are inside the containment. That 8 affects the EQ life of the cabling and solenoids and 9all of the electric components that you have inside 10the containment. It affects stress on things that 11you've evaluated to say that we can take so many of 12 these transients.

13So you may have to reanalyze or replace 14components that are inside the containment. If you 15get into a scenario where you actually have to use the 16passive systems, I think here you're creating a lot of 17stress on equipment that's inside the dry well when 18 you use some of the passive systems.

19 So we have the active systems there that 20we can use to provide the same function and get us to 21a safe, stable state without causing an expensive 22 recovery period.

23CHAIRMAN APOSTOLAKIS: And then the 24opposite question is, you know, if that's the case, 25 51 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433why don't you just use active systems? And is it such 1a big deal to have it declared as non-safety related?

2That's really one of the benefits here. The active 3 systems are --

4 MEMBER SIEBER: Yes, yes, yes.

5MR. WACHOWIAK: It is less expensive to 6 buy and to maintain when they're not safety related.

7CHAIRMAN APOSTOLAKIS: And the reliability 8of these systems expected by most reasonable people to 9be the same as that of safety related systems, right?

10MR. WACHOWIAK: It would be similar to the 11types of things you'd see on oil platforms or in other 12industrial activities where high reliability is 13 required.14So remember these active systems also --

15most of these, main condenser, feedwater, if those 16systems aren't reliable, the plant doesn't make any 17money, and if they're not making any money, then 18 what's the point of building it in the first place.

19CHAIRMAN APOSTOLAKIS: Maybe it's obvious 20 to people. You have the active systems because, you 21 know, they don't create such a mess if you use them, 22 right?23 MEMBER SIEBER: Right.

24CHAIRMAN APOSTOLAKIS: But you still have 25 52 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 the passive systems.

1 MR. WACHOWIAK: Right.

2CHAIRMAN APOSTOLAKIS: Which are the 3 safety related systems.

4 MR. WACHOWIAK: That's correct.

5CHAIRMAN APOSTOLAKIS: Overall you have a 6benefit, right? Compared to a system, a reaction 7 that's only active?

8MEMBER WALLIS: Well, you need the active 9 so that you can talk about a CDF of ten to the minus 10 eight.11CHAIRMAN APOSTOLAKIS: My question is what 12is the ultimate gain of using a combination of the 13 two. What is it that you are gaining from that? Is 14 it dollars? Is it perceptions?

15 MEMBER SIEBER: Yes, yes.

16 CHAIRMAN APOSTOLAKIS: Is it both?

17MR. WACHOWIAK: Well, it is dollars 18because the passive systems are much simpler systems.

19 Okay? So making a system safety related adds some 20exact cost associated with it. If it's a complicated 21system, the cost is more than if it's a simple system.

22If it's a simple system, it doesn't add as much cost.

23So we like our safety systems to be the 24 passive systems.

25 53 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER SIEBER: The passive system is what 1gives you the low PRA numbers. If you didn't have 2 those, you'd be --

3CHAIRMAN APOSTOLAKIS: That's the whole 4 point. They don't give any number.

5 MEMBER SIEBER: Yeah, they do.

6CHAIRMAN APOSTOLAKIS: It's the active 7components that give you the numbers. Do you have any 8 number anywhere that says this is the probability of 9failure of the passive, truly passive system? No.

10 You've assuming --

11MEMBER KRESS: That's called a focused 12 PRA, which I think the staff is asking him to do.

13 CHAIRMAN APOSTOLAKIS: No, no.

14MEMBER SHACK: Let GE answer the question.

15CHAIRMAN APOSTOLAKIS: They assume that 16these active systems are not there. There is an 17explicit statement someplace that says we assume that 18 the passive components do not fail, right?

19In your passive system if you have a check 20 valve that has to open, then you look at the failure 21 rate of the check valve, but you never look at the 22failure of the tank or, you know, what are not coming 23 down.24MEMBER SIEBER: Gravity is in the wrong 25 54 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 direction.

1CHAIRMAN APOSTOLAKIS: Gravity may reverse 2 itself, yes.

3 MEMBER SIEBER: There you go.

4MR. WACHOWIAK: We didn't address the 5gravity reversing itself. What we did look at though 6is there are components in the systems that we call 7 passive. Now, we have to remember here that passive 8is now a defined term. Passive -- a pipe is passive.

9We have pipe breaks in that analysis. We look at pipe 10 breaks. That's a failure of a passive component.

11 MEMBER ARMIJO: As an initiator 12MR. WACHOWIAK: As an initiator, but there 13are passive things that we call passive because 14they're operated only using essentially stored energy.

15 It's not energy that we have to create.

16So these DPVs that are fired using DC 17power from a batter, it's a split valve, DC power, 18 that's been declared to be a passive component.

19 We have the failure rates of those types 20of passive components that need to change state in the 21 PRA. That's where we get the numbers for the passive 22 features.23 MEMBER ARMIJO: What is an ARI?

24 MR. WACHOWIAK: Alternate rod insertion, 25 55 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433and I'll cover that on probably the next slide. I 1 define all of these things.

2 MEMBER ARMIJO: Okay.

3CHAIRMAN APOSTOLAKIS: So the combination 4of passive and safety related, non-safety related, 5overall results in benefits. It's cheaper; they're 6 less expensive to design?

7 MR. WACHOWIAK: Less expensive design.

8 CHAIRMAN APOSTOLAKIS: What else?

9MR. WACHOWIAK: By definition it's adding 10 diversity. So it gets us to the lower -- somebody 11said it gets us to the lower CDF. It does because 12inherently it has to add diversity. If you have an 13active system and a passive system, they don't operate 14the same way. They don't have the same types of 15 components.

16MEMBER SIEBER: Fewer components to fail.

17 MR. WACHOWIAK: In many cases, there are 18fewer components to fail. Some active systems are 19 fairly simple, but in general --

20 MEMBER ARMIJO: Passive systems are also 21 easier to maintain than active systems.

22MEMBER SIEBER: You don't have to do 23 anything.24 MEMBER ARMIJO: You don't have to do 25 56 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 anything.1MEMBER MAYARD: You will have some active 2 systems that will be safety related, I would think.

3 MEMBER SIEBER: No.

4 MEMBER MAYARD: No?

5 PARTICIPANTS: No.

6 MEMBER MAYARD: Nothing? We'll see when 7 you get to operation.

8 (Laughter.)

9 MR. HINDS: This is David Hinds.

10 Just to add just a couple of points just 11 while we're on the topic, one thing to point out is 12 that the column on the right, the active systems, 13they're not enough to license the plant by themselves.

14So there would be additional systems one way or 15another, the safety systems. Then it becomes a choice 16of are those safety systems active or are they 17 passive.18 So we would require those safety systems 19 regardless. So, in essence, the column would not go 20 away. It's just a matter of those systems, do we 21choose to design them as a passive system or as an 22 active system. They would still be necessary.

23And then some of the failure modes of the 24typical active systems that have a large number of 25 57 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433pumps and motor operated valves and things of that 1 nature, we went with the thought process of removing 2as many of those active failure prone components as 3possible, but the system needed to be there regardless 4of whether it was active or passive to perform that 5 safety function.

6 So I don't think we're in a case of 7whether we could remove a large number of systems.

8It's just a matter of whether we choose to design them 9 as active or passive.

10MEMBER WALLIS: The passive aren't 11necessarily more reliable. They may not operate as 12 designed. Your ability to predict how they operate 13may not be as reliable as it is for an active system.

14 MEMBER SIEBER: That's true.

15MEMBER WALLIS: So it's not clear to me 16 that passive is necessarily better.

17MEMBER SIEBER: Well, you're reliant upon 18all of your thermal hydraulic analytical codes, and 19 given what I know about that, I like --

20MR. HINDS: We're reliant upon things such 21as static head of water in a tested integrated system 22as opposed to a conked (phonetic) head of water we 23felt would result in a more reliable configuration as 24 well as there are economics involved as well.

25 58 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER SIEBER: But the differential 1 pressures that derive the flows are small in passive 2systems compared to, you know, a 5,000 or more starter 3 pump.4 MR. HINDS: And another note, too. Many 5of the components in the active category in this slide 6are typical power producing components that are 7 necessary to generate electricity.

8MEMBER WALLIS: But there are ways that 9passive systems can fail. I mean, you can have a pipe 10that's supposed to be full of water. For historical 11 reasons it may have air in it.

12 MEMBER SIEBER: Or steam.

13 MEMBER WALLIS: And may not function the 14 way it's supposed to function.

15We should probably move on, but this whole 16idea that passive is necessarily more reliable I'm not 17 sure is true.

18 MEMBER MAYARD: But those are applied to 19 active components, too.

20 MEMBER WALLIS: That's right.

21MEMBER MAYARD: If you're not meeting your 22 tech specs with water where it's supposed to be --

23MEMBER WALLIS: Have examples of that 24where the pipe that's supposed to be full of water is 25 59 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 full of air. Then your pump can't suck.

1MR. WACHOWIAK: So this is one of those 2examples where we really build on the safety analysis, 3the DBA analysis, because those types of questions, 4will it work, are answered in the DBA analysis for the 5 most part.

6Well, let's talk about the functions here.

7 Reactivity control function. We start with RPS, 8reactor protective system. That's similar to most 9 BWRs. It's a SCRAM function, failsafe I&N. So if it 10gets a signal it SCRAMs a plant. If it loses power --

11 MEMBER WALLIS: This is a case where you 12 don't rely on gravity, right?

13 (Laughter.)

14MEMBER WALLIS: You're pushing against 15 gravity.16MR. WACHOWIAK: We're pushing the rods 17 against gravity, but remember we are using a head of 18water to get them going, and then the flow through the 19core is actually what brings them all the way in. So 20it's against gravity, but it's still the passive 21 direction when it goes that way.

22Often a rod insertion, a question that was 23asked earlier, what does that do? It provides a 24backup to the RPS I&C function. So if for some reason 25 60 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 that passive I&C function doesn't work --

1MEMBER WALLIS: Now, do you credit that in 2 your outsource analysis?

3 MR. WACHOWIAK: Yes.

4 MEMBER WALLIS: In your CDF?

5 MR. WACHOWIAK: Yes.

6MEMBER WALLIS: So it's an active system, 7 but it's credited --

8MR. WACHOWIAK: In the PRA analysis, in 9everything except for the seismic margins analysis, 10 we've credited all of these functions that I will be 11 talking about.

12CHAIRMAN APOSTOLAKIS: Is this a safety 13 related system?

14MR. WACHOWIAK: RPS is safety related.

15 ARI is not safety related.

16 PARTICIPANT: Well it's active.

17 MEMBER WALLIS: Without BSE.

18MEMBER BONACA: It's going to be what tier 19 one says.20 CHAIRMAN APOSTOLAKIS: RPS is not mired?

21 MEMBER BONACA: No, not yet.

22MR. WACHOWIAK: ARI is not safety related.

23 The fine motion control rod drive is also non-safety 24 related. That's the typical way that we would move 25 61 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 the control rods in this plant. It's different than 1what's in the BWR 2 through 6 out there now. It's an 2electrically driven screw arrangement to move the rods 3 in and out of the core for normal power control.

4 MEMBER WALLIS: Excuse me. Now, this 5 inserts different rods.

6 MR. WACHOWIAK: Same rods.

7 MEMBER WALLIS: Oh, the same rods.

8MR. WACHOWIAK: All rods have this 9 function.10 MEMBER WALLIS: No extra rods. It's the 11 same rod, different wave, but --

12 MR. WACHOWIAK: Right. So when we get a 13SCRAM signal, we also tell this fine motion control 14rod to start spinning its screws there. So if for 15some reason the stored energy control rod motion 16doesn't get all the rods, the ones that are back 17behind it, they take a little bit longer, but they 18 also get driven into the core.

19And then finally for the standby liquid 20control system, it's a sodium pentaborate solution 21just like in the existing plants. However, in our 22configuration, we have no pumps here. The solution is 23 in a tank that's pressurized with nitrogen, and when 24 you open the squib valve, the high pressure nitrogen 25 62 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433drives the liquid into the core region, and I think 1 maybe many of you have looked at that analysis.

2MEMBER MAYARD: Is the ARI -- is that a 3 fast insertion?

4MR. WACHOWIAK: It makes the same thing 5happen as the SCRAM function, and so it's just barely, 6barely slower. The SCRAM function individually opens 7up all of the solenoid valves on each hydraulic 8control unit to vent each one. The ARI vents the 9 header. So it, in effect, does the same thing, but 10 it's not, in fact --

11 MEMBER SIEBER: It's not as close.

12 MR. WACHOWIAK: It seconds different.

13MEMBER MAYARD: But you really have three 14systems putting the rods in the normal SCRAM. the ARI 15 is the backup.

16MR. WACHOWIAK: An ARI is the backup to 17the instrument and control portion. The RMCRD is the 18backup to the actual motion of the control rod. So 19 it's really one backup system.

20Once again, this configuration is 21 extremely reliable, and when we look at our numbers, 22ATWS comes out to be less than one percent of total 23 CDF with this configuration.

24Pressure control function. First we have 25 63 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433the main steam system. Obviously -- well, not 1obviously, in this plant it's capable of handling most 2 of the transients, except for the ones where there's 3an isolation of that system for some other reason.

4 It's capable of handling 100 percent of rated steam, 5100 percent bypass capability on this plant. We're 6 not limited by what we can put into the condenser.

7 The isolation condenser system now would 8be the next level of defense here. So this is one of 9these cases where the non-safety system is what we 10 look at first. That's what we want to have. That's 11 our preferred method of removing decay heat. If for 12 some reason that won't work, that doesn't work or it 13 becomes isolated, we move to the isolation condenser 14system or ICS, which provides decay heat removal. The 15key here is if this system goes into operation, we 16never lift any SRVs. So the challenge in the 17containment is eliminated essentially. It removes the 18 heat. The isolation condenser pool is outside the 19 containment.

20 And with this system we can sustain our 21safe shutdown condition for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> with no human 22 actions. With human action we can -- you know, as 23 long as decay heat support it, we can stay there.

24 Finally, if we get to the point where we 25 64 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 don't have either one of those, we do have safety 1 relief valves, basically ASME type valves on the 2 pressure vessel provides the backup, discharges into 3the suppression pool so that we minimize the impact on 4 the containment itself. It really mostly discharges 5to the suppression pool because there are some that 6 can go into the dry well, but those are sequenced on 7 later.8 Even in a transient where we did isolate 9the main steam and isolation condensers don't come on, 10there's several minutes before we pressurize the 11 reactor enough to actually lift the SRVs.

12MEMBER SIEBER: They're spring loaded 13 safety valves?

14 MR. WACHOWIAK: Spring loaded.

15 MEMBER SIEBER: Not pilot operated.

16MR. WACHOWIAK: They are pilot -- they're 17 dual -- no? Alan has --

18 MR. BEARD: This is Alan Beard with GE.

19They are spring loaded safeties when they 20are externally actuated relief valves, but only ten of 21the 18 actually have the external actuation for a 22 relief function.

23 MEMBER SIEBER: Okay.

24MR. BEARD: So eight are pure safeties, 25 65 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 and ten are a combination safety relief valve.

1MEMBER SIEBER: Okay. How many valves 2does it take to comply with the code? One hundred 3 percent flow, all 18 or some fraction?

4MR. BEARD: A limiting situation is 5actually an ATWS event, and we need all 18 valves for 6that case. For the ASME over pressurization with 7 SCRAM, it's significantly less than 18.

8 MEMBER SIEBER: Okay. Thanks.

9MEMBER WALLIS: Now, there's no DPV on 10 this slide?

11MR. WACHOWIAK: No. This is the pressure 12control or over pressure protection on the vessel.

13The DPV is there for allowing the low pressure systems 14to actuate. This is just keeping the vessel intact 15 following a scram or an ATWS.

16 So the DPVs don't play a role in what I'm 17calling pressure control. Pressure control is keeping 18 from over stressing the vessel.

19 MEMBER WALLIS: With regard to filing?

20MR. WACHOWIAK: As a matter of fact, 21because we have time in this plant from when you would 22reach that pressure, if it was not an ATWS, I'm not 23sure of the timing of the ATWS because I really 24 haven't looked at that for actuating DPVs. We would 25 66 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433have -- operators could in that several minute time 1 frame actuate those.

2 We didn't take credit for any operator 3 actions in that short of a time frame.

4MEMBER WALLIS: Another way to 5 depressurize.

6 MR. WACHOWIAK: It would be another way.

7 The result of this type of configuration 8in our analysis is that the vessel over pressurization 9comes out to be a negligible impact. We don't see any 10sequences or at least anything that significantly 11affects the core damage to get there, not in the limit 12 of precision that we're looking at.

13 MEMBER WALLIS: There are sequences that 14 you pursue though.

15 MR. WACHOWIAK: Yes.

16MEMBER WALLIS: Where the vessel pops and 17 it pops the containment.

18 MR. WACHOWIAK: Yes. We have those.

19MEMBER WALLIS: Just the number associated 20 with that is very small.

21MR. WACHOWIAK: They just die out through 22the quantification and don't quite make it to the end.

23 The next thing is the inventory function 24 at high pressure. This one is a little strange 25 67 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 because of an isolation condenser system. We'll get 1 to that in a second.

2Feedwater system, once again, that's what 3we want to use if at all possible. It's available 4 most of the time, a highly reliable system. It does 5require our preferred power system which would either 6be -- you know, it's what has typically been called a 7pass off-site power. We have some other capabilities 8 in this plant, but for now we'll just call that the 9preferred power system there, not diesel generator 10 backed up. It takes the grid type power.

11Capable of handling any transient and 12small LOCAs. We can deal with those, and actually up 13to some fairly significant LOCAs if we can get the 14 system back in line.

15 MEMBER SIEBER: Just keep pumping.

16MR. WACHOWIAK: Just keep pumping until 17 you run out of water basically.

18MEMBER ARMIJO: How big a break would that 19 be?20MR. WACHOWIAK: Essentially we could 21handle any break. The problem is the timing. When 22 does the system isolate and when can you get it back 23in service, and I think in the different LOCA 24scenarios, I think we my have credited it in the 25 68 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433medium LOCA, which is essentially a three inch line 1 break.2The next backup is the isolation condenser 3 system. We saw that in the pressure control. If 4these come into operation, it provides the pressure 5control, but because it's closed loop cooling, it 6condenses all of the reactor steam. We don't need any 7 kind of makeup.

8 Once again, that was key for not lifting 9the SRVs or not losing inventory. So as long as we're 10 not losing inventory, this can keep us in that state 11 for at least 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, potentially forever.

12Finally, the other backup that we have 13that actually starts, comes into service at about the 14same time as the isolation condenser system, is the 15 control rod drive. Here our control rod drive pumps 16are not your father's control rod drive pumps.

17They're 500 GPM each. We have two of them, fairly 18 substantial. Provides backup high pressure injection 19 function that could be used independently of these.

20This is backed by our non-safety diesel 21 generators. So it could be off-site power or on-site 22 power. 23 Handling any transient. When I say here 24"most LOCAs," the flow rates were designed with the 25 69 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 small LOCA in mind, but what we see is if it's a steam 1LOCA also because of where the water level comes out, 2 this 1,000 GPM that we can put in with these two pumps 3 quickly balances decay heat and we can keep the core 4covered with these systems even if the plant 5 depressurizes and there's a bigger LOCA.

6 This combination here, once again, these 7 are all systems that are in the analysis. They help 8maintain the low CDF, and when we see later in the 9 results one of the reasons why this doesn't make it 10negligible with this configuration is what happens 11between the 24 and 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> and what has to happen 12 there.13We're finding the PRA to address that, but 14 we haven't quite addressed it yet.

15 And we've got the low pressure function.

16 We didn't credit in the PRA the condensate 17 function. A lot of existing plants look at condensate 18for providing low pressure injection. When we looked 19at it, we saw that there were so many commonalities 20with the feedwater system that we just previously 21talked about, and the feedwater system was already 22credited in those analyses. We didn't see a lot of 23 extra benefit to adding the condensate system.

24So it's there. It's just not on my list.

25 70 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433We thought it was fairly dependent and it wouldn't 1make much difference in the resolves. So then we get 2to low pressure then. We've got the gravity driven 3 cooling system.

4Here's our passive operation, the tanks 5that you saw inside the vessel there in our sketch on 6the front of each page or each presentation. It's 7inside the containment. All water that we need is 8already there inside the containment. It doesn't need 9to be augmented in any scenario where the containment 10 remains intact.

11 Back up to that would be the fuel 12auxiliary pool cooling system in LPCI mode. LPCI mode 13of operation can transfer suppression pool water into 14 the vessel just like existing LPCIs do.

15 Power, once again, on this one is backed 16by the non-safety diesel generators. We have a third 17method of getting water into the plant through our 18diesel driven fire pump. We have provided a hard 19 connection to put that fire water into the vessel if 20 needed. 21 We don't need any AC power to run this.

22 It's independent.

23So again, this combination along with the 24 high pressure helps maintain the low CDF.

25 71 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Talking about the depressurization 1function, depressurization valves, we call them the 2 DPVs. Passive operation, once again, that means it's 3 stored energy. It's a squib valve. It has got a 4charge on it. You applied power from the batteries 5however it gets there, but power from the batteries 6 that fires these.

7They open. It discharges directly into 8the dry well. A fairly large opening when they all go 9 off.10 It provides complete depressurization, and 11that's the key for the GDCS operation, is that you get 12 the dry well and the reactor at the same pressure so 13 that the head of water in the GDCS tank can allow the 14 system to drain.

15 We do have --

16 MEMBER SIEBER: How long does it take to 17get that equalized pressure? It's a matter of 18 seconds, right?

19MR. WACHOWIAK: Yeah, it's not very long.

20 Do you remember, Alan?

21 MEMBER SIEBER: Ten to 30 seconds?

22MR. WACHOWIAK: It's in the DCD. I can 23 look it up.

24MR. BEARD: Yes, this is Alan Beard again.

25 72 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Actually the sequence is the initial 1depressurization is through the safety relief valves, 2 the relief function.

3 MEMBER SIEBER: Right.

4MR. BEARD: We blow it down to about 20 5 pounds gauge before we'll open up the DPVs to lessen 6the transient on the blow-down in the dry well. So 7overall to get down to that zero differential 8 pressure, it's on the order of 30 seconds.

9MEMBER SIEBER: That's what I figured.

10 Thanks.11MR. BEARD: That's one of these things 12where in the design basis analysis, we look at the 13 sequence a little bit differently. We looked at the 14 DPVs independent from the SRVs when in actuality the 15real sequence is the SRVs open first, and then the 16 DPVs open second, and what we tried to do in the PRA 17is that we don't want to specifically just say you 18have to have both. We look at what kind of redundancy 19 we actually have here.

20For GDCS operation, we need the DPVs. For 21some of the other things, LPCI or fire water, the SRVs 22by themselves are sufficient to operate those systems.

23 MEMBER MAYARD: And these are considered 24 passive valves, DPVs?

25 73 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. WACHOWIAK: The DPVs are considered 1passive, yes. Squib valves that are powered by our 2 batteries, all stored energy devices.

3MEMBER ARMIJO: What actuates those 4 things? What causes the battery to send the signal to 5 this squib valve? How do they work?

6 MR. WACHOWIAK: Essentially what happens 7 is we've got our level control system, and I'll just 8go through the simple case on level control. As in 9 the existing BWRs, there's a Level 1, which will be 10 the ECCS actuation level. That signal then is going 11through the different systems and sends a signal to 12 open the SRVs first and the DPVs and then the GDCS 13 valves. It goes through that system.

14 The I&C is powered by the batteries, and 15 the power that goes to the valves also comes from the 16 batteries. So it's a digital I&C system that's doing 17 that.18 MEMBER ARMIJO: Okay. Thanks.

19 MR. WACHOWIAK: Okay. So again, this is 20a very reliable configuration the way it is. The high 21pressure sequences amount to less than two percent of 22our CDF. So we see if we have a low core damage 23 frequency. Everything tends to be in the low pressure 24 range.25 74 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 We talk about the decay heat removal 1 function here. This really only applies to the Level 2 2 analysis, but we'll look at it here in the list of 3 functions.

4 The main condenser is available. That's 5where we want the heat to go. That's the easiest way 6 to transfer it to the ultimate heat sink. If we get 7into one of these other scenarios, once again, ICS 8will do it by itself, and you start thinking that ICS 9is a pretty important system in this plant. It 10 provides a lot of functionality, a lot of protection 11 in many things.

12We've got the passive containment cooling 13 system which, if there is steam in the dry well, it 14will perform its function. It won't perform its 15function if you don't have steam and the dry well has 16 got to condense the steam.

17This one, again, doesn't need any support 18systems at all for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. If you open the DPVs, 19the passive containment cooling system starts working.

20We say for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> because at somewhere after 24 21 hours2.430556e-4 days <br />0.00583 hours <br />3.472222e-5 weeks <br />7.9905e-6 months <br /> the design requirement in that 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> --

22MEMBER WALLIS: This is on the noted 23 containment?

24 MR. WACHOWIAK: Yes, it is.

25 75 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER WALLIS: And so you've got all of 1 these noncondensables that have to go somewhere.

2 MR. WACHOWIAK: That's correct.

3MEMBER WALLIS: So in order to keep track 4of them in evaluating your effectiveness in 5 condensation, I guess we're going to get into that.

6MR. WACHOWIAK: We'll get into that in the 7presentation this afternoon. I talked specifically 8about how the noncondensables are dealt with in the 9PCCS, but in general, if we start out with a LOCA it's 10 fairly simple.

11Pressure suppression containment works 12like GE pressure suppression containments have in the 13 past. The steam drives the noncondensables through 14the vents in the suppression pool and they're trapped 15 in the suppression pool.

16But the way the PCCS works, it also 17 provides a mechanism for driving the noncondensables 18 in the suppression pool. So in the long run, all of 19the nitrogen is in the suppression pool, and the PCCS 20is a self-regulating device then that can operate 21 indefinitely as long as you have water.

22 And at this 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> point or analytically 23 we show later than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, but at that point you 24need to get more water. We have an automatic means of 25 76 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 opening some valves to automatically bring -- seeing 1 gravity driven more water there, but there are other 2 backups to that.

3MEMBER SIEBER: As far as condensation is 4 concerned, it doesn't make any difference whether it 5 contains it or not.

6 MR. WACHOWIAK: That's correct.

7MEMBER SIEBER: If there's not 8noncondensables in there, then it's the same 9 inventory.

10MR. WACHOWIAK: For a Level 1 analysis, 11whether it's inert or not doesn't make any difference.

12 MEMBER SIEBER: That's right.

13 MR. WACHOWIAK: Another backup system to 14all of this. As long as you have enough inventory in 15the vessel, you've got the reactor water clean-up 16 system. It can operate in a shutdown cooling mode.

17So just like RHR works now, our reactor water clean-up 18has that same RHR function. It can be placed into 19operation back by the non-safety diesels. It does 20require service water and things like that to operate.

21MEMBER SIEBER: It is basically a high 22 pressure system.

23MR. WACHOWIAK: Yeah, it operates at high 24 pressure in reactor water clean-up mode.

25 77 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER SIEBER: Right.

1MR. WACHOWIAK: And then you can switch it 2 into a shutdown cooling mode that can go essentially 3from rated pressure all the way down to cold shutdown.

4MEMBER ARMIJO: But can that by itself 5provide all of the decay heat removal you need or just 6 a fraction of it?

7 MR. WACHOWIAK: Yeah, go ahead.

8MR. BEARD: Yeah, Alan Beard with GE 9 again.10 It is a full pressure rated system. The 11heat removal capacity will not match the decay heat 12 curve for about the first hour. We do need some --

13 MEMBER ARMIJO: Something else.

14 MR. BEARD -- the first hour of heat. In 15about the first hour though we come into the decay 16 heat curve, and the reactor water clean-up system by 17 itself will be able to locate the decay heat.

18 MEMBER ARMIJO: Thank you.

19MR. WACHOWIAK: The one point I want to 20bring out here is if we're looking at the challenge to 21keeping the vessel or the core covered, if we have the 22 injection functions that we've talked about earlier, 23we don't need the containment heat removal function 24for more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. We will talk about that a 25 78 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 little bit more in the Level 2 analysis.

1 But in the first day if we don't activate 2any of these things, we still don't get to a point 3where our active systems are being challenged or where 4the -- it's not the active system. It's where the 5 level in the vessel is being challenged.

6Someone asked a little bit before about 7how some of these diverse control systems were. This 8is our schematic from Chapter 7 in the DCD.

9 Essentially how we get the actuation of the DCD and 10 the GDCS valid.

11MEMBER WALLIS: This is all illegible and 12 proprietary. None of the printing came out in this.

13 MEMBER SIEBER: It works for me.

14MEMBER DENNING: Yeah, it's out of focus, 15 even on the printed page.

16 MEMBER SIEBER: Don't worry about it.

17MR. WACHOWIAK: It's not fuzzed up 18 intentionally. It's a process where you go from a 19 drawing to a PDF back to a drawing to a printing.

20 MEMBER SIEBER: It works for me.

21 MEMBER DENNING: This is PRA.

22MR. WACHOWIAK: But you'll find this 23 drawing in the DCD, Chapter 7.

24 MEMBER SIEBER: There you go.

25 79 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. WACHOWIAK: It's in there.

1 MEMBER SIEBER: As a PDF.

2 MR. WACHOWIAK: Yeah, meeting all of the 3 pixel requirements that were done.

4Essentially we've got two -- for each 5 valve, whether it's a DPV or whether it's a GDCS 6 valve. We've got one valve. On that valve there are 7two drivers or two charges for the squib. So the 8squib needs to fire to open the valve. We've got two 9of them on there. Each one gets a signal from a 10 different train to the system.

11Look at the bottom one here. It's a 12simple one. This is the safety related I&C system.

13Its signals come in from all four divisions. It votes 14 on whether or not we've actually received the signal 15 that we expected to see.

16 It sends a signal to -- independently 17sends a signal to two different load drivers, which 18 allow power to go to that squib and actuate it.

19We've duplicated that from a different 20 division on here, but we've also provided a parallel 21signal in from what we call the diverse protection 22 system to perform the same function.

23Now, what's the diverse protection system?

24This is a separate instrument and control system in 25 80 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433the plant. We call it diverse. So we're pretty much 1saying that it's diverse in manufacturer, hardware, 2 software, that looks at the different ECCS functions 3and provides a backup signal, if you will, to those 4 different functions.

5 So if for some reason we have got some 6failure in the safety system, we have a backup system 7for that. It can fire one of these. The part that's 8common here is the DC power comes from the station 9batteries, and we didn't duplicate the diverse station 10 battery.11 MEMBER ARMIJO: If sensors failed, would 12this system operate? The sensors that say, okay, 13something is wrong; level is wrong. If those sensors 14 failed?15 MR. WACHOWIAK: We would have to fail --

16with this configuration here, we would have to fail at 17 least four sensors, two in each system, where two in 18each system. So two in the safety related system, two 19 in the non-safety system would all have to fail. So 20 if you have any two that work in the safety system and 21any two that work in the non-safety system, this will 22 actuate. So you would have to fail three. I'm sorry.

23 I got my successive failure back.

24 So any two in the safety system that work 25 81 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433or any two in the diverse system that work will get 1 this actuation.

2 MEMBER ARMIJO: Okay. Thanks.

3 CHAIRMAN APOSTOLAKIS: Rick, it seems to 4 me this is a natural place to break.

5 MR. WACHOWIAK: Okay.

6CHAIRMAN APOSTOLAKIS: The next one is an 7 event tree. Question?

8MEMBER BONACA: Just a question generally 9 to do with the reactor safety systems which most of 10 them, they're not safety related.

11 MR. WACHOWIAK: That's correct.

12MEMBER BONACA: Okay. How do you envision 13that that will affect testing? And how do you account 14 for, for example, if you have less test requirements 15 since you impose on the systems because now you have 16the reliance on the passive systems as a major 17 blackout.18 You know, I can see advantages there for 19 the operator. How do you account for those? And do 20you foresee it will be different with panel 21 availability of the system because they're not being 22 tested as frequently?

23MR. WACHOWIAK: One thing about our active 24 systems is that they're not just there to sit and do 25 82 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433nothing while we're waiting for an accident or 1transient to happen. All of those active systems, 2except the SRVs, really have some function to play in 3 the operation of the plant.

4FAPCS is sued for water transfer and pool 5 cooling and pool clean-up, things like that. So all 6of these active systems that we have need to be 7operating, most of them continuously, some of them 8 very periodically in order to do your role operation 9 of the plant.

10 So the list of things that are in standby 11for the active to perform these active portions of the 12 function is a very small list.

13MEMBER BONACA: Very small list. Okay.

14 Thank you.

15CHAIRMAN APOSTOLAKIS: Okay. We'll break 16 until 10:27.

17(Whereupon, the foregoing matter went off 18 the record at 10:13 a.m. and went back on 19 the record at 10:31 a.m.)

20 CHAIRMAN APOSTOLAKIS: So we are back in 21 session please.

22MR. WACHOWIAK: Okay. Now that we're back 23and I'm on, I wanted to go through a couple of the 24event trees here. With the time frame we have, we 25 83 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 couldn't necessarily go through all of them. I want 1to just talk about a couple of representative event 2 trees.3The first one, I think, is the general 4 transient. This is the way we'd like the things to 5 go, well, at least the top part.

6You have some sort of transient with 7successful RPS. Bypass valve opens to the main steam 8 line. We have feedwater available. We're okay.

9That's not so much different than what you see in any 10 BWR.11So as we go through the different systems 12 here, if for some reason we don't have the bypass or 13we don't have the feedwater system, what we end up 14 with is in --

15 MEMBER WALLIS: When I was reading this, 16there's all of the acronyms and things, some of which 17didn't seem to even be defined anywhere. This diagram 18is full of these PRFLs and things. You have to figure 19 out what it means.

20 I have great difficulty even in the list 21 of acronyms finding some of these.

22MR. WACHOWIAK: I'm sorry. In the 23 presentation itself or you are looking in --

24MEMBER WALLIS: In the document, in the 25 84 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 document.1 MEMBER SIEBER: Proprietary.

2 MR. WACHOWIAK: Yeah, that's a --

3MEMBER WALLIS: Maybe they're hidden in 4the text somewhere or something, but anyway it's just 5 a comment on this.

6MR. WACHOWIAK: Yeah, the headings for 7these things should be in Chapter 3 of the PRA, and I 8 thought we had those.

9MEMBER WALLIS: Maybe they're in there 10 somewhere, but they're not gathered together so that 11 you can find --

12CHAIRMAN APOSTOLAKIS: Also in the printed 13version and the electronic version in some of these 14event trees you just can't read your headings, 15especially for the loss of power, which contains a 16sequence that is a dominant sequence, number 44. It's 17 really impossible to read the headings, and I notice 18 you don't have a tree here.

19MR. WACHOWIAK: Not on this presentation, 20and I do have it on my computer. We can talk about 21that one, too, if you wanted to. We'll figure out how 22 to do that in the document. We're somewhat bound up 23by our software and the ability to get the nice 24pictures out of the software and into a document. We 25 85 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433can make it into a stand alone drawing for you and 1send it as a stand alone printed drawing, but to send 2it electronically, it's in the format of the software, 3 and I'm not sure that that's --

4CHAIRMAN APOSTOLAKIS: Do you have a 5bigger figure, you know? You know, print it and send 6 it to Eric.

7MR. WACHOWIAK: We can send hard copies of 8the event tree, and we can send all of those. We 9 could seen, you know, big, 11 by 17 hard copies.

10 MEMBER SIEBER: That would be useful.

11MR. WACHOWIAK: Once again, it loses a 12little bit when you convert it into the PDF, but we'll 13 see what we can do. If it's possible to get that 11 14by 17 scanned into a PDF that's got a high resolution, 15 we can do that.

16But when we go from the software and print 17 to the PDF, it loses it.

18 CHAIRMAN APOSTOLAKIS: You have to go to 19the microphone and identify yourself, please, with 20 sufficient clarity and volume.

21MR. BHATT: My name is Sid Bhatt, and I'm 22 from GE.23 Regarding this form place, you have been 24using only the 11 by eight -- 11 by 17, and it's 25 86 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433easier to read. I agree with you, and we can provide 1you for what Rick has been saying in Section 3.

2Basically that's where all of those trees are defined, 3 and we can give you that set so that you can see T-44 4 and other things too. But that is probably the best 5 way to give it to you, would be a hard copy if it's 6 okay.7 CHAIRMAN APOSTOLAKIS: Yes, sure. We'll 8 take that.

9 MR. WACHOWIAK: Okay, and we'll do that.

10The question about all the different 11headings here being defined somewhere. I know we 12discussed them in Chapter 3, in the original and in 13 Rev. 1, Rev. 0 and Rev. 1, but you're right that there 14 isn't a "here's a list where all of them are." 15CHAIRMAN APOSTOLAKIS: Part of the problem 16is that the actual headings of these three, for 17 example, where it says "I," you're not going to find 18an I because you will find IC in the list of acronyms.

19 You see you have an I there on the fourth 20 column? It's really IC. So if you go through the 21list of acronyms looking for I, you're not going to 22 find it. You're going to find the IC.

23 And for some reason you're using U1CF when 24it's high pressure injection, right? These are the 25 87 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 computer --

1 MR. WACHOWIAK: It's one version of high 2 pressure injection.

3CHAIRMAN APOSTOLAKIS: So these I doubt 4you will find in the list of acronyms because they are 5 just computer acronyms used in the calculations.

6 MR. WACHOWIAK: Basic event --

7CHAIRMAN APOSTOLAKIS: The actual systems 8 are below.

9 MR. WACHOWIAK: Right.

10CHAIRMAN APOSTOLAKIS: Well, you know, the 11 completely scrutable PRA will be produced when there 12 is a really complete PRA, which means never.

13 It's okay. I mean, we don't want you to 14 be shocked.

15MR. WACHOWIAK: I'm just trying to 16 remember where I was.

17 (Laughter.)

18CHAIRMAN APOSTOLAKIS: These things 19happen, but it was interesting that especially that 20transient for the loss of feedwater and loss of 21preferred power, which were really of interest, there 22is no way you can read the headings. It's not a 23matter of finding any of the acronyms. You just can't 24 read them at all.

25 88 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. WACHOWIAK: You can't read it when it 1got into the file. We'll get that fixed. It may take 2 a separate document to do that. Getting it into the 3 one concise document is difficult.

4CHAIRMAN APOSTOLAKIS: Well, today you are 5 giving us an overview of the whole thing.

6 MR. WACHOWIAK: Yes.

7CHAIRMAN APOSTOLAKIS: And then I hope at 8the end of the meeting or maybe at the end of the day 9and at the end of tomorrow we can identify some topics 10on which we would like a more detailed presentation 11 some time in the future.

12MEMBER WALLIS: Well, apparently the Spell 13Checker doesn't work on the PRA either because there's 14 typos all over the place.

15 MR. WACHOWIAK: In these?

16MEMBER WALLIS: "Inyection" and 17"suppresion" and "equilisium." I mean three typos in 18 one chart.

19CHAIRMAN APOSTOLAKIS: I can assure you 20 PRA has nothing to do with it.

21MEMBER SIEBER: Hey, the PRA folks are 22 doing the best they can.

23 CHAIRMAN APOSTOLAKIS: Why don't you go?

24MR. WACHOWIAK: Okay. If we move on then, 25 89 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433we don't have the regular power conversion system, 1 which would be the main steam along with feedwater.

2 We move and we check the isolation condenser. Three 3 of four goes into operation; we're okay.

4I think I heard a question somewhere in 5the audience about passive things that need some kind 6of a signal to actuate. The isolation condenser 7itself is one of those systems where if we lose power 8or lose the signal, the I&C signal goes into operation 9on its own. So it would be activated, and so long as 10 we keep water in the upper pools, it will take us out 11 as long as we need.

12We don't have the isolation condensers.

13We asked do we over pressurize the vessel, and here 14you notice that we've used one of the 18 SRVs. The 15 likely case in the loss of ICs is that we would have 16 some IC capability. We just wouldn't have enough to 17 prevent the actuation of the SRVs.

18After we get into more detail, we can see 19what value should be put there and what we actually 20 need to prevent failure of the vessel versus just in 21 the ASME's range, the stress on the vessel.

22CHAIRMAN APOSTOLAKIS: So these, the 23 second here out of the success criteria --

24MR. WACHOWIAK: It relates to success 25 90 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 criteria.1CHAIRMAN APOSTOLAKIS: So these were 2presumably derived by doing the appropriate thermal 3 hydraulic calculations.

4 MR. WACHOWIAK: Yes.

5CHAIRMAN APOSTOLAKIS: And somebody is 6 checking those, the status of it, I suppose.

7MR. WACHOWIAK: We're checking those, yes.

8 The next one we get to if we do 9 successfully keep the --

10MEMBER WALLIS: Well, what's the 11 probability of 17 of these things failing?

12 CHAIRMAN APOSTOLAKIS: Very low.

13MR. WACHOWIAK: Very low, and it's the 14 same probability as 16 failing and 15 failing and 14 15 failing. So once again, we didn't really get into 16 revision of that number so much since you don't have 17the ability to resolve it down to the difference 18between 11 failing and 18 failing. It's the same 19 thing.20MEMBER SHACK: What calculations do you 21 use for the PRA, what thermal hydraulic code?

22 MR. WACHOWIAK: We can talk about that a 23little bit later, but it's a combination of things.

24 For some things where it's obvious, where we're 25 91 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433looking at things like for that one feedwater pump can 1provide injection since one feedwater pump is capable 2of 45, 50 percent of rated flow, certainly it can take 3 decay heat. We have a hand calculation for that.

4 For other things where we have our design 5 basis analysis, and this particular case here, three 6or four isolation condensers, we've used Track G in 7the design basis calculation to show that that's 8success, and we've just adopted that here in the PRA.

9 Other things that are a little more 10complicated that involve multiple failures. We 11couldn't just lift directly from the safety analysis, 12and we've done calculations with MAAP 4 on those, and 13we're in the process of discussing with the staff how 14 to resolve any sort of uncertainties or other issues 15 associated with using that code.

16MEMBER WALLIS: Well, these are very 17 simplistic. On the idea that isolation can then -- it 18 either works or it doesn't, it's just not quite like 19a pump. I mean, it could get blanketed with no 20condensables and work to some extent. There's a whole 21 lot of these things which can partially work.

22 And the FEPRA says it's there or it isn't 23there, which is very unrealistic for some of these 24 systems.25 92 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. WACHOWIAK: Yeah, we didn't really 1address partial failures of the systems, but if you do 2go into the default tree for the isolation condenser, 3the purge valves for the noncondensables are in there 4 to the extent that we would need those.

5So we didn't really look at saying, well, 6we have two and a third equivalent heat exchangers.

7We've just said does it function the way it's supposed 8 to, just like you would in the active PRA.

9MEMBER WALLIS: And the other thing is if 10you put the uncertainties in the thermal hydraulics 11 into this, then you could be moving from one branch to 12another because of, you know, being on the tail end of 13 some probabilistic distribution of the heat transfer 14coefficient or something, and that's not in here 15 either. 16MR. WACHOWIAK: That's not in here in that 17exact what. What we've done, and we're still 18addressing this, is that when we set the success 19criteria or the threshold for saying success versus 20failure, when we use MAAP what we did was we didn't 21look at actual heat-up of the clad and the onset or 22 the failure of the clad. What we really looked more 23 at was did we uncover the core.

24So where we set our threshold for saying 25 93 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 it's success or not should address things like those 1types of thermal hydraulic uncertainties. The 2question that we're dealing with now is is our method 3for calculating when we get to the top of the core --

4 is that an adequate way of doing it?

5 I think that at least at the preliminary 6stage we have I can't call it agreement yet, but we 7have a conceptual agreement that those thermal 8hydraulic margins could be handled by setting the 9 threshold at the top of fuel rather than doing the 10 detailed calculations.

11So, again, I would think that this is 12something that's appropriate at this DCD phase to use 13that type of a conservative analysis to address 14success criteria and maybe do something more detailed 15as we move forward. But I think there's other 16uncertainties that would be bigger than that 17 particular one when we justify the margin.

18 We're working on that.

19CHAIRMAN APOSTOLAKIS: The way I 20understand it, today's presentation will not get into 21methods for doing things, to quantifying. You're just 22 presenting results and --

23 MR. WACHOWIAK: Okay.

24CHAIRMAN APOSTOLAKIS: -- how it was done.

25 94 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: Can't we ask that? Can't 1 we ask if we wish how did you get something?

2 MR. WACHOWIAK: You can ask.

3 CHAIRMAN APOSTOLAKIS: You can ask.

4 MEMBER WALLIS: But they're not going to 5 reply?6CHAIRMAN APOSTOLAKIS: What I think we 7should do as we go on, we should identify -- let's not 8 wait until the end of the day -- that we would like to 9revisit in more detail at a future time, and here you 10 have -- I mean, when you have 18 lines, then the issue 11of common cause failures, I guess, becomes important, 12 and at some point in the future we'd like to discuss 13 this with you, how you did it.

14You say in the document you used the 15 alpha factor method, and I looked at the table there 16and some of the numbers appear to be low to me, but 17 there may be a good reason for that. So this is one 18 items we have to do in the future.

19 MR. WACHOWIAK: Okay.

20 CHAIRMAN APOSTOLAKIS: Get more into the 21methods for doing things because you have extremely 22 done that in so many places that it drives them out of 23 style. Also the failure of data, that you use the 24 uncertainty analysis that you did.

25 95 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433I guess this meeting is not really methods 1 oriented. It's more this is what we did; this is the 2results, and in the future we will have done this with 3 some of the methods.

4 MR. WACHOWIAK: Okay.

5CHAIRMAN APOSTOLAKIS: Is that agreeable?

6 Great. Thank you.

7MR. WACHOWIAK: And we'll be happy to 8 revisit those.

9 CHAIRMAN APOSTOLAKIS: Great.

10 MR. WACHOWIAK: Or we can arrange it.

11Gather my thoughts again. I'll move into 12the high pressure injection sort of range that is 13 again here, one of two CRD or one of two -- these are 14feedwater trains, I guess, rather than -- it would be 15one of those typos that you are talking about. It 16 could be one of four feedwater pumps.

17The reason we ask that again here is 18because this could have failed because of the steam 19bath and not just the feedwater. We pick up that 20dependence again by looking at feedwater or control 21 rod drive here.

22Once again, the single control rod, if we 23get through this path, the single control rod drive 24pump is sufficient to keep the core covered. Balances 25 96 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 decay head before we get to the top of fuel.

1If we need to go into the low pressure, we 2look at a combination of SRVs and the active system to 3provide that or, conversely, the DPVs and our passive 4systems to provide the injection. The combination of 5 things in the passive systems is addressing a little 6 more than the short-term cooling, is set up to allow 7us to address a long-term cooling there. We do think 8that there may be some conservatism in the way that we 9have addressed this, at least looking at the 10 equalizing lines.

11Finally, when addressing the low pressure 12injection systems, if the DPVs have actuated, have 13 actually actuated, we put this in here again because 14the training and dependence for the operators is 15different between these two. So we asked for that, 16 again, in that scenario, to pick up that dependence.

17In the end, the way we've drawn these 18trees, they look fairly simple. The underlying fault 19 trees that go into these are a little more complex 20that way. It's a tradeoff of how people like to do 21these analyses. Some like to see more detail in 22default trees. Some like to see more detail in the 23event trees. For illustrative purposes, I think it's 24 easier to show in the event trees, but again, it's a 25 97 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 choice for how we address these things.

1 I just mention a couple of other things.

2 The question came up about what happens if we stress 3the vessel, if we take that as a transfer into one of 4our other trees and we analyze that as an initiator 5going into the other tree. Similarly, we do the same 6 thing with ATWS, where we have a separate event tree 7that discusses the sequence of events that would 8 happen in an ATWS.

9MEMBER WALLIS: Now, when you use this for 10 design --11 MR. WACHOWIAK: Okay.

12MEMBER WALLIS: -- do you say that you 13 want something like the same probability in each one 14of these branches or do you say we want a low 15probability at the beginning so that we don't get into 16 some of these sequences later on? How do you decide 17you're going to have a certain number of DPVs, for 18 example?19Presumably it's based upon some kind of 20balancing of the various contributions to the PRA. So 21how do you do that in the design? How do you use 22 something like this for design?

23 MR. WACHOWIAK: The way that we did that 24 was remember the columns from earlier this morning?

25 98 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER WALLIS: But what about quantity.

1 You're going to say we want a certain probability at 2this price, don't you? Therefore, we're going to have 3a certain number DPVs. Is that the sort of thing you 4 do?5 MR. WACHOWIAK: Yes.

6CHAIRMAN APOSTOLAKIS: In other words, if 7 we want higher reliability on the left --

8MR. WACHOWIAK: Right, more reliability on 9 the left.10MR. WACHOWIAK: What we really want to do 11is we want to minimize these high pressure scenarios.

12MEMBER WALLIS: But you can do that by 13 different parts of the --

14MR. WACHOWIAK: You can do it with 15 different parts.

16 MR. WACHOWIAK: Right, right.

17 MR. WACHOWIAK: In the first phase where 18we looked conceptually at what we're going to do when 19we had discussions, based on experience from previous 20plants, the question on SRVs and it wouldn't be 21experience with DPVs, but experience with things like 22SRVs, the question was: what type of redundancy would 23 we like to see in this system?

24 And I said, well, you know, based on what 25 99 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433I've seen before, I think if we had at least an 1additional three, that should give us a low enough 2 probability here that would drive the numbers toward 3the direction we want to have in the low pressure core 4 damage so that the bulk of the core damage frequency 5 would be in the lower pressure scenario.

6So that was what I call a qualitative 7 judgment on that. So we do that, put that into the 8conceptual design. Then we take the conceptual design 9 and put it into actual fault trees and use it that 10 way. So we confirm --

11MEMBER WALLIS: But how did you balance 12 things? I mean, you can change around tremendously 13the importance of different steps in this process, and 14they're in default tree. So 21 SRVs or ten instead of 1518 or I mean, you can change the importance of certain 16 of these branches by design, right?

17 MR. WACHOWIAK: That's right.

18MEMBER WALLIS: How do you decide what to 19 do?20MR. WACHOWIAK: I think the key here is 21that you can't only use this to optimize that because 22 there --23MEMBER WALLIS: So you don't have a 24system, right? You don't have a system. You don't 25 100 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 have an answer that's --

1 MR. WACHOWIAK: We didn't do it that way 2 is what I'm saying.

3MEMBER WALLIS: You don't have an explicit 4--5 MR. WACHOWIAK: We didn't say --

6 MEMBER WALLIS: -- logical --

7MR. WACHOWIAK: -- you have to have a 8 number that's this good here.

9CHAIRMAN APOSTOLAKIS: And from the PRA 10perspective, whether you have a ten or 16 or 14 lines 11 is irrelevant. I mean, PRA cannot distinguish among 12 these. I mean, you bring in the common cause failures 13 after three or four or five at the most redundant 14 lines. Then the number is the same. So you have to 15 use some other argument why you want to go to 17.

16 MR. WACHOWIAK: That's right.

17 CHAIRMAN APOSTOLAKIS: And that's what I 18 think you said, you know, that this is not the only 19way to do this. I mean, actually the issue of common 20 cause failures and their use in design is a real one 21because the methods have been used, you know, for 22existing plants as an assessment tool, and so on, and 23the numbers that you're getting are not very sensitive 24to certain things a designer can do, like having extra 25 101 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433lines or increasing the separation. I mean, it's all 1 a matter of judgment.

2MEMBER BONACA: I say it's look at what --

3 you know, is that you really more than -- I mean, the 4 PRA helped you, but you really used a lot of the BWR 5 experience.

6 CHAIRMAN APOSTOLAKIS: Exactly.

7MEMBER BONACA: And the PWR as a basis.

8I mean, that's where you start from, and I think 9 that's an advantage. You have that advantage. You 10 should use it.

11MR. WACHOWIAK: We want to get as close as 12possible to what the design is going to look like 13 before we even have to go into one of these detailed 14 models.15MEMBER WALLIS: But there's no concept of 16 what sort of the optimum design strategy would be or 17anything like that? It's just what you happen to 18 have? You draw a figure and you take what you've got 19 and you say, "Well, that looks okay." 20MR. WACHOWIAK: Well, no. One of the 21things that we looked at actually got on this 22particular figure, but it would have been on the 23reactor water clean-up line break outside the 24containment, and we went through. We had our 25 102 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 conceptual design of how we wanted that to look. We 1modeled what was there, and when we looked we saw, you 2know, this break outside the containment fraction, the 3 core damage frequency is higher than we wanted it to 4 be. We want it to be negligible. We don't want a 5 break outside the containment leading to core damage 6in a bypass. What can we do so that this is no longer 7 a non-negligible sequence?

8So we went back to the designers and said, 9"What can you do to increase the reliability of the 10 isolation of that system?" 11And we added an extra automatic isolation 12 valve or isolation itself. It's not really just one 13 valve. We added an extra automatic isolation from a 14diverse system into that line which now when we put 15that back into the model, lo and behold, the 16 containment bypass sequences are now negligible.

17So that's really the process we went. We 18 didn't try to say we have a target value for each of 19these branches. We do know that in general we want 20the bypass sequences to be negligible. We want the 21high pressure sequences to be low, and we want the 22overall core damage frequency to be low in terms of 23 what people are used to seeing.

24 CHAIRMAN APOSTOLAKIS: I can see the PRA 25 103 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 being careful because you move down to lower levels, 1 the system and component. Then the value of the PRA 2begins to diminish because the models are not so 3 sensitive by that.

4MR. WACHOWIAK: We can find other things.

5We found places where we identified manual valves that 6are used for maintenance that would need to be 7 instrumented and alarmed because if they are left in 8 a misposition condition after maintenance, it tended 9to drive up the reliability of some of these or under 10 liability of some of these systems.

11 So we go back to the design and say, you 12know, we understand you're going to have component 13 checklists for these things, but let's add something 14else on top of that. We want to make sure that these 15 aren't left in a state where they may not be able to 16 perform their function.

17 So we use it that way rather than trying 18to set a target reliability for each step along the 19 way. Once again, optimizing the entire plant that 20might get us into a -- if we tried to hit targets for 21every individual piece rather than looking at the end, 22 we would get into a problem that might be hyper over 23constrained rather than one that's just over 24 constrained now.

25 104 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER ARMIJO: Rich, but what is the 1requirement for 18 SRVs? You've done this analysis --

2 MR. WACHOWIAK: ATWS.

3MEMBER ARMIJO: ATWS. Okay. For that how 4 many do you need?

5 MR. WACHOWIAK: Eighteen.

6 MEMBER ARMIJO: Eighteen.

7MR. WACHOWIAK: To meet the ASME code for 8ATWS analysis, you need to have 18. So that drives 9 that, and the PRA didn't say you need more than 18.

10MEMBER WALLIS: So 18 barely meets the 11 ASME?12MR. WACHOWIAK: Well, it meets it. I 13 wouldn't say "barely meets it." It meets it.

14MEMBER WALLIS: But you said you had to 15 have 18 to meet the -- does that mean if you had 17 16 you wouldn't meet it?

17MEMBER SHACK: You wouldn't meet the code.

18MR. WACHOWIAK: You couldn't meet the 19 code.20MEMBER WALLIS: You wouldn't meet the 21 code. Okay.

22MR. WACHOWIAK: In the PRA we ran a 100 23percent ATWS case, and looked at things like when 24 feedwater ramping down and reactivity control as the 25 105 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433level came down, and determined that we would not get 1 the vessel to a place where it would fail, and I don't 2remember what the specific number that we used was for 3 that, with nine SRVs open.

4So we looked at the expected scenario for 5an ATWS and looked at how many SRVs did we have to 6have open before we would actually get to the point 7 where we were failing the vessel, not just exceeding 8 code, but failing the vessel.

9CHAIRMAN APOSTOLAKIS: Coming back to the 10issue of having more reliable systems on the left, 11 aren't these the systems that really are involved in 12the design basis accidents so that the conservative 13analyses there indirectly lead you to very reliable 14 systems?15MR. WACHOWIAK: Certainly the design basis 16 analysis uses the reactor protection system.

17 CHAIRMAN APOSTOLAKIS: Right.

18MR. WACHOWIAK: The design basis analysis 19 uses the isolation --

20 CHAIRMAN APOSTOLAKIS: Right.

21MR. WACHOWIAK: It uses these SRVs. It 22uses the DPV and GDCS systems. The FAPCS is probably 23 not included in the design basis, and the fire water 24 injection is not included in the design basis.

25 106 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433As CRD is an injection system for some of 1 the sequences, it isn't analyzed or it isn't used in 2 the design basis, but once again, we set the criteria 3based on an ALWR URD requirement that we do have an 4 active system to mitigate small LOCAs.

5 CHAIRMAN APOSTOLAKIS: Okay.

6MR. WACHOWIAK: Okay. To go to just 7 another example here, the feedwater line break. The 8feedwater line break is what we call large steam LOCA.

9Steam LOCAs depressurize the vessel on their own. So 10once again -- and it's in the dry well. Just showing 11 we don't need to ask things about the DPDs in these 12 scenarios. We can go directly to the low pressure 13 systems.14 MEMBER WALLIS: This is a large LOCA and 15 a feed --16 MR. WACHOWIAK: Feedwater line break.

17 MEMBER WALLIS: Now, Table 5.2, it says, 18"The probability of large steam LOCA . . . train A is 195E to the minus one." It doesn't make any sense to 20 me. The probability of the LOCA is .5? This is Table 21 5.2, 5-2.22 MR. WACHOWIAK: Your notes?

23MEMBER WALLIS: My notes on page 5.5-D.

24 Well, if the probability of a LOCA is .5, I wouldn't 25 107 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433have built a plant at all. I don't understand what 1 that number means. Is that the initiation of this 2 whole --3MR. WACHOWIAK: That would not be the 4 initiation that we --

5 MEMBER WALLIS: It doesn't make sense.

6MR. WACHOWIAK: I'm not sure of the origin 7of that value or the context that it's used off the 8 top of my head.

9MEMBER WALLIS: Well, maybe someone can 10 answer that later in the day.

11 MR. WACHOWIAK: Yes.

12CHAIRMAN APOSTOLAKIS: The large LOCA, 13 Graham?14 MEMBER WALLIS: It says, "Probability of 15LSLOCA" -- large steam LOCA I guess that means -- "in 16 FWTA," FW train A, point -- well, you can look into 17those, but there's some numbers in that table that are 18 really strange, strangely high. It's Table 5-2, and 19 in my version it's page 5.5-D.

20MR. WACHOWIAK: The specific acronym 21 you're saying is -- can you read that off?

22MEMBER WALLIS: It says, "Probably of 23 LSLOCA." 24 MR. WACHOWIAK: LSLOCA.

25 108 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: That means large steam 1 LOCA, right? Does it mean that? In feedwater train 2 A.3 MR. WACHOWIAK: We'll find that out, see 4 what happened there.

5 MEMBER WALLIS: Now, maybe if train B is 6 okay, but train A is in trouble.

7CHAIRMAN APOSTOLAKIS: Now, I think the 8subcommittee would very much like to see a detailed 9 discussion here of the dominant sequences.

10 MR. WACHOWIAK: Okay.

11CHAIRMAN APOSTOLAKIS: Okay? Like this 12sequencing the loss of prepared power and two or three 13 others. So maybe at the next meeting we can do that.

14 MR. WACHOWIAK: Okay.

15CHAIRMAN APOSTOLAKIS: Walk us through it.

16 Tell us what the data used were, where they came from, 17common cause failures, the whole works. That would be 18a useful thing to see. Okay? So that's another item 19 for the future.

20 So initiating events.

21 MR. WACHOWIAK: We'll talk about what we 22 used for initiating events. We covered the spectrum 23 of transients, grouping them as appropriate, various 24loss of coolant accidents. Basically the reason we 25 109 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433split those up is that where the different 1 penetrations come into the vessel makes a difference 2somewhat in how the response is and what the actual 3 outcome is going into the Level 2.

4 MEMBER WALLIS: What does loss of the 5condenser entail? The condenser is three. You don't 6have any water flowing through the coolant side or 7 something?

8MR. WACHOWIAK: It could be several 9 things. You could lose the water on the cooling side.

10You could lose the vacuum so that you get a hole and-11-12 MEMBER SIEBER: Air bound.

13 MR. WACHOWIAK: -- air bound.

14MEMBER WALLIS: It's more likely that you 15partially lost it, isn't it? For some reason the 16vacuum doesn't work very well or something. Again, 17 all of these things are extreme cases. Certainly it 18 isn't there, which seems to be very unlikely.

19 MR. WACHOWIAK: Then in those cases what 20 we have to look at is what I think is on the next 21slide, is going back to how we got those numbers, and 22 if you go into the NUREG, it gives a list of where the 23various numbers came from, and things like partial 24 losses of condenser were either included or excluded 25 110 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433in all those different values, and we summed up the 1 ones where the failure mode is still retained in the 2ESBWR even though we may have augmented the design so 3 that some of these failure modes for transients --

4 MEMBER WALLIS: That's also bothered me.

5 You go back to this NUREG, which is based on past 6 history. You're going to build a much better plant.

7 MR. WACHOWIAK: That's right.

8 MEMBER WALLIS: The condenser won't look 9 quite like all of the old condensers, and yet you're 10 going to use the same number for its failure because 11 that's all you've got? Is that it?

12MR. WACHOWIAK: Well, there's two things.

13There's one, we do have it. That's always a plus.

14It's always good to go with something that you do 15 have.16But our objective here on the PRA is to 17identify things that are associated with the 18mitigating features of the plant. We're not 19 necessarily trying to reduce the CDF just by saying, 20well, we're going to eliminate or reduce the 21initiating event frequencies because remember, once 22 again, initiating events especially on the transient 23side aren't necessarily hardware issues. It's 24hardware and people issues, and what we thought would 25 111 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433be a good representation for this phase of the design 1 is to use the values that were based on operating 2experience at plants, propagate them through the 3analysis, show that the configuration of the plant can 4withstand those, and if in the end we do find out that 5we have a reduced initiator frequency for something 6 because we can prove that the design is better than 7 what's out there, then in a later stage we may use 8 that, or maybe we save that until we get to the 9operational PRA after we actually have some real data 10 with operating these new systems.

11So we made the decision to use the 12existing database for initiating events, and the only 13 place where we really took things out is if something 14in the existing database, that feature or that failure 15mode that was there just isn't there anymore in the 16 ESBWR. We took some of those out. So there are some 17tweaks on the values, but they were fairly consistent.

18 The other thing that we did with this is 19in the LOCA frequencies. Now, you saw we had a whole 20 bunch of different LOCAs there on the previous page, 21and you can't go into any of these documents and find 22 where is the GDCS line break or where is a -- you 23 know, they're based on existing plant type numbers.

24 So what we did to get to our LOCA numbers 25 112 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 was we looked at how it was done, how it was 1apportioned for the existing plants, used the same 2values, but just reapportioned those values associated 3with the piping sizes and classes that we have in this 4 plant.5 So it's essentially the same LOCA values 6 that were used that were found in the other documents, 7 just reapportioned.

8MEMBER WALLIS: So when do you use the 9valves which you're going to stick open? You use 10exactly the same valves in 2020 when you build this 11reactor as were operating experience in 1987 to 1995?

12 Nothing has improved in 25 years?

13 MR. WACHOWIAK: We expect improvement in 1425 years, but the key that we wanted to say is that 15the reason -- we don't want to eliminate consideration 16of I'll call it vulnerabilities, but consideration for 17certain sequences just because we're speculating that 18 20 years from now we're going to have better SRVs.

19CHAIRMAN APOSTOLAKIS: The question is:

20 is it worth the effort to argue with he NRC staff --

21 MEMBER WALLIS: That's it.

22 CHAIRMAN APOSTOLAKIS: -- why you use a 23lower probability distribution when, in fact, it 24 doesn't seem to affect much?

25 113 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: The whole process, the 1whole regulatory process seems to inhibit improvement.

2 CHAIRMAN APOSTOLAKIS: To some exhibit.

3MEMBER WALLIS: Because you know some 4 number that's 40 years out of date, and you know it 5 and it has been approved, we'll stick with it and we 6 won't try to do any better.

7 PARTICIPANT: Or increased margin.

8MEMBER DENNING: But during the operation, 9 they'll --

10 MEMBER WALLIS: You will see that.

11CHAIRMAN APOSTOLAKIS: On the other 12extreme you have people who, you know, and you see 13that mostly in the aerospace business. We change the 14design and, boy, they hit the failure rate by a factor 15of ten or 20, and then of course, nobody believes it.

16 So I think what these guys are doing is much better, 17staying with the numbers even though you know that the 18 distribution will make it have shifted.

19MR. WACHOWIAK: In our optimization of the 20 design, if we find out that one of these assumptions 21for something that we know is going to be better is 22 impacting other parts of the design, like, you know, 23because we did this now we have to have -- I don't 24know -- MSIDs that weigh a million pounds. I don't 25 114 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 know. Something that affects the rest of the time.

1Then we can take a hard look at those and see if 2 there's something that we can do.

3 But for the cut-through that we're doing 4 at this phase, we thought it prudent to look at 5existing operating experience, initiating events that 6come from all sorts of different things, not just from 7looking at a particular design or some component or 8 some system.

9And I think I've covered everything on 10 here.11CHAIRMAN APOSTOLAKIS: Yeah, you've 12 covered it.

13MR. WACHOWIAK: Basic event data. Now, 14this is another one of these places where we had to do 15 something. We needed to use generic data 16CHAIRMAN APOSTOLAKIS: How old is the URD?

17 I mean, that's a long time ago, isn't it?

18 MR. WACHOWIAK: Yes.

19 CHAIRMAN APOSTOLAKIS: How old is it?

20 MR. WACHOWIAK: How old is it?

21 CHAIRMAN APOSTOLAKIS: '80s, late '80s?

22 MR. WACHOWIAK: '80s sounds correct.

23CHAIRMAN APOSTOLAKIS: They assure me 24 there have been PRAs for BWRs all over the place. I 25 115 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 mean, why didn't you use those, or did you check and 1 the numbers were more or less the same?

2Because the later PRAs probably include 3 plant specific information. They are more realistic 4 numbers. I don't think it's a major issue, but I'm 5just curious. I mean, just because it's a document 6 blessed by somebody, we have to stick to it?

7 MEMBER WALLIS: Yes.

8MR. WACHOWIAK: It's more coming from our 9 customers' request that we use the URD as a --

10 CHAIRMAN APOSTOLAKIS: I see.

11 MR. WACHOWIAK: -- guide for our design, 12and the data that's in there is included in that 13 table.14Now, we did look through there and 15compared it to things that we used in the Lungmen 16 plant which we're building now in Taiwan. We've got 17a PRA for that. We have some experience from other 18things factored in, but we've looked at some of these 19 failure rates with respect to the group's experience 20 from looking at operating plants.

21We're not seeing, you know, orders of 22 magnitude difference in these values. So I think at 23this phase of the design, I think the good enough 24 principle applies here.

25 116 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Now, if there's something that is very 1important and we address some of these in the 2sensitivities; we looked at these new squib valves.

3Is the reliability in the URD for squib valves, is 4that appropriate for what we're using here? And we 5tried to see if there was some kind of sensitivity to 6 that. There is some sensitivity, not necessarily 7enough to change our minds on things, but generally 8 that's where we get them from, and we think that it's 9 a conservative way to go.

10 It can be refined in the future, but you 11know, with some of the new equipment, we're not really 12going to know until we operate and start testing some 13 of these things.

14CHAIRMAN APOSTOLAKIS: Speaking of the 15squib valves, I didn't want to raise it, but you did.

16 On Table 4.6-5, list of system common cause failures, 17this is the gravity driven system. There is a 18 probability of the common cause failure of all squib 19 valves equal to three times ten to the minus five.

20And the probability -- oh, no, I'm sorry.

21For two valves, for two squib valves is ten to the 22 minus five, 3.6, ten to the minus five. The 23probability of one valve failing is three, ten to the 24minus three. So if you're going to take the ratio, I 25 117 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 come up with a beta factor of .012.

1Now, the beta factor usually is around ten 2percent, and you are going here with one percent, and 3 I wonder how that came about.

4 MR. WACHOWIAK: Well --

5CHAIRMAN APOSTOLAKIS: One percent is 6 pretty low.

7 MR. WACHOWIAK: Ten percent seems fairly 8 high.9CHAIRMAN APOSTOLAKIS: But that's the 10 number that --

11 PARTICIPANTS: No, no.

12CHAIRMAN APOSTOLAKIS: First of all, 13 that's one of the problems with PRA. I mean, we are 14dealing with all of this as if they were nothing, you 15 know. I don't like that percent. Make it one percent 16 or make it one in 1,000.

17Well, it could be .06, right? It doesn't 18go down by an order, but let me -- no. I mean there's 19 very strong evidence that the beta factors above .1, 20extremely strong, in fact, based on data. In some 21 cases it's close to .2, okay, and I have a figure if 22 you'd like. I'll send it to you, where there is all 23 sorts of information, and the average is about .1.

24MEMBER WALLIS: Is this just for squib 25 118 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 valves or for --

1CHAIRMAN APOSTOLAKIS: No, no, for all 2 kinds of components.

3 MEMBER WALLIS: Everything.

4CHAIRMAN APOSTOLAKIS: Interestingly 5enough, for space systems it's also .1, and people now 6 are scratching their head. What's magical about .1?

7But anyway, but the bigger factor here of 8.012, it seems to me, has to be justified on the basis 9of something, and again, you don't have to answer now, 10but next time, these are the kinds of questions you're 11 going to get.

12 MR. WACHOWIAK: Okay.

13 CHAIRMAN APOSTOLAKIS: It's awfully low.

14It's awfully low, in my view. I mean, there is no 15 basis for it. Okay?

16 Now, for four valves, I understand that.

17 In fact, another thing is for four valves, only four 18squib valves fail to open. It's three, ten to the 19minus five. For two valves, its 3.6, ten to the minus 20 five. I mean, that's incredible accuracy.

21MEMBER WALLIS: Well, there are some more 22 accurate figures in some other tables.

23CHAIRMAN APOSTOLAKIS: And then the CCF 24 for all seven squib valves in the GDCS lines, failure 25 119 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433to open is 1.5, ten to the minus five. That indicates 1sensitivity of the model to the number of valves that 2 I don't believe is there.

3 So all of this is on the transcript now.

4 Next time we discuss this, right?

5MR. WACHOWIAK: Okay. We can discuss how 6 we got those different common causes.

7CHAIRMAN APOSTOLAKIS: Yeah. You say you 8used the alpha factor method, but one of the key 9elements in the methodology that the NRC and EPRI have 10developed is that you go back and look at actual 11common cause failures and you screen out the ones that 12 don't apply to you, and I don't know whether you did 13that, but if you did that, then you probably screened 14 out more than you should have.

15MEMBER WALLIS: Well, I think in order to 16achieve credibility, you have to look at some of 17 these things in the detail that George is looking at 18 it, and you folks have to justify what you did.

19 CHAIRMAN APOSTOLAKIS: Yeah, because you 20 know, the last line there says "low CDF to design 21 rather than data values." Well, I just showed you a 22 data value that may be a driver, in fact, because in 23order of magnitude it's an order of magnitude. You 24know, an order here, an order there. Pretty soon 25 120 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 you're talking about pretty low values.

1 I'm not saying I'm right. I'm saying we 2 need an answer to what I just said.

3 MR. WACHOWIAK: We will discuss that.

4 CHAIRMAN APOSTOLAKIS: Very good.

5MR. WACHOWIAK: I did want to bring up one 6 other point here, is that we do have components that 7 in this plant we don't expect to be tested except on 8a refueling interval basis, and if we use demand data 9from some of these generic sources from that, some 10data are actually based on quarterly type test 11 intervals. So we adjusted those, basically converted 12 the quarterly test interval data into an hourly rate 13 and we applied a longer test interval.

14 CHAIRMAN APOSTOLAKIS: Yeah, the results 15 of the problem there, unfortunately I cannot find it 16now, but you used some formulas to do some things that 17 are not clear to me what they mean, and we definitely 18 need some explanation in the future.

19 MR. WACHOWIAK: Okay.

20 CHAIRMAN APOSTOLAKIS: How you did that.

21You said, you know, using well known formulas this is 22 where we got, and I hope I'll find it and let you know 23 where it is.

24MR. WACHOWIAK: Okay. That would be in --

25 121 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433CHAIRMAN APOSTOLAKIS: Failure rates, you 1 know, how they were --

2MR. WACHOWIAK: In Rev. 1 we included what 3 the formulas were.

4CHAIRMAN APOSTOLAKIS: Oh, I don't know 5which rev. I looked at. Oh, okay. I found it. I 6 found it. It's in Section 5.2 of the PRA. Okay?

7 Component reliability database, 5.2.

8 So if you guys come back later and address 9these issues, for test periods greater than a year, 10this is what we do. For others we do something else.

11For components whose test period is from six months to 12one year it is suggested that the upper bound on 13demand failure probability be used as a computation of 14mean, the median, and then the new mean is that value 15 times the error factor.

16 That's not true. So not that it makes a 17hell of a difference, but we don't want to -- in 18 addition to the typos to have also --

19 MEMBER WALLIS: I think you need to look 20 at it, and the same thing in thermal hydraulics.

21 The devil is often in the details.

22 CHAIRMAN APOSTOLAKIS: Yes.

23MEMBER WALLIS: And you find something 24 which is unjustifiable in the details sometimes.

25 122 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433CHAIRMAN APOSTOLAKIS: In the details, of 1course, what they do is they create an image, a 2 section.3MEMBER WALLIS: Absolutely. You don't 4need many false details to discredit the whole thing.

5CHAIRMAN APOSTOLAKIS: Yeah. So please 6look at that in Chapter 5, 5.2, and then we know we'll 7 talk about it. Okay? Good.

8 Human actions.

9MR. WACHOWIAK: Human actions. This will 10 probably be another one that you're going to want to 11 have --12 CHAIRMAN APOSTOLAKIS: I suspect it will 13 be, yes.14 MR. WACHOWIAK: But at this stage of the 15 game of the design, we did a very simplified version 16of human actions. We looked at two different things, 17pre-accident actions. Basically it was looking for 18places where we expected maintenance to potentially 19 leave systems in an unknown unavailable state.

20Then we looked at what controls were 21 placed on some of these things to see if there was 22 something that we could do with quantification. For 23things where we just relied on check lists or just 24standard things that the plants do, we kept the 25 123 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433standard value. If there were additional controls 1like alarms or indications or things, we would do 2 different things.

3 We took the most credit when it was, you 4know -- and alarmed in the control room stayed on 5 these valves.

6MEMBER WALLIS: I'm just looking at my 7 notes. I wrote on one page here where you were 8 looking at some -- "operator errors are judged to be 9a non-significant contribution." I think you're 10talking here about operation of depressurization 11 valves or something.

12But it's just an assertion. There's no 13explanation of why, and I just wonder how many of 14these sorts of statements are allowed in the PRA. You 15simply say we judge something to be nonsignificant 16without any justification, particularly operator 17 actions. How do you really know what they're going to 18 do unless you've got some basis, how they perform on 19 a simulator or something?

20 So I just picked it out and we should ask 21 why whenever we see statements like that.

22MR. WACHOWIAK: Was that related to the 23 error of commission or was that related to --

24 MEMBER WALLIS: Well, I'm not sure. I'd 25 124 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433have to look at page 2.3-4 to see the context, but 1 this is the kind of thing I pulled out.

2 MR. WACHOWIAK: Two, point, three, dash, 3 four, assuming initiating events.

4MEMBER WALLIS: Right. It's an initiating 5 event, isn't it?

6 MR. WACHOWIAK: Right.

7 MEMBER WALLIS: Yes, okay.

8 MR. WACHOWIAK: So in that context --

9 MEMBER WALLIS: I think it was opening a 10 valve when they shouldn't do it or something 11MR. WACHOWIAK: Yeah, it was causing an 12 initiating event.

13MEMBER WALLIS: Causing an initiating 14 event.15MR. WACHOWIAK: By doing something that 16 they --17MEMBER WALLIS: And you just said that so 18 you assume they won't do it.

19CHAIRMAN APOSTOLAKIS: Yeah, you are 20 right. We will need to have a special session on 21 these things, especially tables --

22MEMBER WALLIS: Well, how long are you 23going to take though? If you go into the details, 24 it's going to take a long time.

25 125 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 CHAIRMAN APOSTOLAKIS: Well, it could be 1a two day meeting. Table 6-1 and 6-2, actually 6-2 is 2 fascinating. You have --

3 MEMBER WALLIS: Where are we here?

4CHAIRMAN APOSTOLAKIS: It's probably --

5the numbers you have there probably come from the EPRI 6 ACR model, right? One called the reliability model.

7 MR. WACHOWIAK: Yes, that sounds right.

8CHAIRMAN APOSTOLAKIS: And if they do, you 9 are probably the only organization at work that's 10using it, and you have are remarkable table here. You 11 are giving us probabilities of failure as a function 12 of time, available time, 30 minutes, 60 minutes --

13 MEMBER WALLIS: Aren't they all one year 14 minus one, one year minus two, one year minus three?

15CHAIRMAN APOSTOLAKIS: Yeah, and also you 16 are classifying them according to the behavior type, 17 skill, rule and knowledge. So this is really a 18 remarkable achievement here.

19MEMBER WALLIS: I say it must be very 20 rough estimates in my notes.

21CHAIRMAN APOSTOLAKIS: But the thing is 22 that this is of great interest to some of us on this 23 committee because we're trying in another context to 24convince the NRC staff that we do need time dependent 25 126 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433-- I mean the distribution for the probability of 1 failure given the time.

2But this is definitely something -- you 3say you are relying on EPRI NUREG CR-1278 and NUREG 4 CR --5 MEMBER WALLIS: Is that Table 6-1 there, 6 too?7CHAIRMAN APOSTOLAKIS: Yes, there is a 6-8 1, but that is pre-initiated.

9MEMBER WALLIS: A detection interval of 10 8,640 --11 CHAIRMAN APOSTOLAKIS: That's one year.

12MEMBER WALLIS: That's a pretty long time.

13CHAIRMAN APOSTOLAKIS: No, they don't mean 14 detection. You mean inspection, I think, human 15 inspections.

16 MR. WACHOWIAK: That's inspection.

17 MEMBER WALLIS: It says detection.

18CHAIRMAN APOSTOLAKIS: It says detection.

19MR. WACHOWIAK: Detection, it's from when 20you operate it this time until when you go back and 21 operate it again.

22MEMBER WALLIS: Oh, it doesn't mean it 23 takes a year to figure out what's going on.

24 CHAIRMAN APOSTOLAKIS: It's the interval 25 127 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 between tests.

1MEMBER WALLIS: I thought it meant that it 2 would take him a year to figure it out. Me, too.

3MR. WACHOWIAK: It is from when we make a 4mistake until we believe there's opportunity to 5 discover it.

6CHAIRMAN APOSTOLAKIS: It's the degree of 7 detection, right?

8 MR. WACHOWIAK: Yes.

9CHAIRMAN APOSTOLAKIS: Between tests.

10 Detection is the wrong word.

11MR. WACHOWIAK: I believe detection is 12correct because it might be between tests or it might 13be between operation. Like let's say it's an FAPCS 14valve and they go to do a full water transfer and they 15say they got water on the wrong place. Oh, we detect 16 it when we're doing this other operation.

17So it's when you can detect it. Sometimes 18it's test interval. Sometimes it's operation 19 intervals.

20CHAIRMAN APOSTOLAKIS: the problem with 21 detection is that it's also used in other contexts.

22Anyway, these are the tables we needed, 6-23 1, 6-2.24MEMBER WALLIS: Almost every table can be 25 128 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 questioned.

1 CHAIRMAN APOSTOLAKIS: The whole Chapter 2 6.3 MR. WACHOWIAK: The tables in Chapter 6, 4 and one of the things that we'll talk about that and 5now it looks like where we'll be for quite some time 6until all of the human factors analysis and all of 7 those things are wrapped up, we're probably going to 8be retaining this type of structure for the next year, 9 year and a half or so before it gets significantly 10 changed in the PRA.

11 So if you would have asked me this six 12months ago, I would have said we're probably going to 13go to something different in the future, but now I 14think that's a good topic because I think that the 15 way the schedule is working out, we'll be using this 16 for some time.

17CHAIRMAN APOSTOLAKIS: Well, yeah, but 18also at the same time you want to do something that's 19 reasonably defensible, right?

20You know, I hear mixed comments regarding 21 this AHCR model, even from the original developers.

22 MR. WACHOWIAK: Okay.

23CHAIRMAN APOSTOLAKIS: I think it would 24behoove you to go and talk to one or two of those 25 129 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 guys. Give them a call. I mean, what's going on with 1 these models?

2You know, they run simulator experiments, 3you know. Then we hear that they were overly 4 enthusiastic in using the results of the experiments 5to produce these numbers. I don't know what to 6 believe myself, and we had a subcommittee meeting on 7human reliability about a year ago, last December, 8well, last December, and some folks from the utilities 9and EPRI presented their calculator, the APRI 10 calculator, which allows you to use -- it's really a 11 problem that allows you to use one of four models.

12 One of the four models is the AHCR, and 13 the guy from the utility told us nobody is using it.

14 Do you remember that. Do you remember that?

15 PARTICIPANT: Yes.

16CHAIRMAN APOSTOLAKIS: So now if nobody is 17 using it and you're the only ones, I'd like to 18understand why. I think I know why. Because it's the 19only model that gives you information like what you 20 have in 6-2, time and probability of failure.

21MR. WACHOWIAK: And it's somewhat 22independent of the variables that we don't know at 23 this point in time.

24CHAIRMAN APOSTOLAKIS: Anyway, I think 25 130 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 it's something to look into.

1MR. WACHOWIAK: Okay. In the end, one 2thing that we didn't include in the PRA is the repair 3and recovery in the base model. We did look at 4recovery of off-site power based on the NUREG curves, 5so the loss of off-site power from 1992 through -- or 6'82 through 2006 I think are the latest one.

7 I want to back up. One --

8MEMBER BONACA: One comment regarding the 9 previous slide.

10 MR. WACHOWIAK: Yeah.

11MEMBER BONACA: I think to me interesting 12is also how do you -- you know, you had 13configurations, and you identified that because for a 14 certain system you have a lot of involvement that is 15maintained and taken out and in. Okay. You could 16really improve the safety of the plant by modifying 17 maybe that system there.

18 Have you had any -- how do you --

19MR. WACHOWIAK: Modifying the which 20 system?21 MEMBER BONACA: I mean the plant.

22 MR. WACHOWIAK: Oh, okay.

23MEMBER BONACA: Well, take the CRDF 24 system. I mean, you have so many valves there, you 25 131 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433know, butanized to a valve being persistent or, you 1 know, each line and then tested and so on and so 2 forth. Is there any better way to do it?

3I mean, I'm trying to understand how do 4you use the PRA to give an input to design. Here 5you're talking about modeling this human actions, but 6it seems to me that you have the opportunity to modify 7the necessary human action at this stage of design, 8and that's what I would like to understand at some 9point, not necessarily today, but at some point I'd 10 like to understand how you came to this ESBWR.

11 How did the PRA contribute to it?

12CHAIRMAN APOSTOLAKIS: I think, Mario, 13part of it was in the original utility requirement 14where they decided that for the first what, 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, 1572 hours0.0182 days <br />0.437 hours <br />0.0026 weeks <br />5.98146e-4 months <br />? The design should not require any operator 16 intervention.

17 MEMBER BONACA: Yeah. No, I understand.

18CHAIRMAN APOSTOLAKIS: So they just 19 followed up, right?

20 MR. WACHOWIAK: The design matches that.

21 In our model here, we look at some of these active 22 systems that can be actuated using operator actions.

23We do look at a sensitivity that says what happens if 24we do analyze it the way the ERD said no operator 25 132 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433actions for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> and take a look at the effect of 1 the overall results on that and the revision to that 2analysis is in the process of being updated before 3 that's done.

4 So we do look at it in that sense, but I 5 think the place where we're more going to use this is 6 as we develop our instrument and control systems for 7 the plant and the layout for the simulator and after 8the simulator, for the control room and for the remote 9shutdown panel, and where different actions need to 10 take place.

11 Where we find in the PRA some of these 12actions to be important actions, we might say, you 13know, maybe you want to put that somehow in the 14automated system or maybe you want to insure that 15 that's in the control room and not out in the field.

16 That would be the way that we would use 17that, but at this point in the overall scheme, all 18we've gotten to is identifying the higher level 19operator actions to the people that are doing the 20human factors analysis. So when they go through their 21process, we'll be factoring this sort of thing into 22 it.23So that's the process for doing it. We've 24identified what's important and then modeling of that 25 133 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433goes into the way that the human interface is put 1 together from the plant, and then we'll be able to 2 come back later and see if we did any good or if it 3didn't make much difference, but I think we can do 4 that. We're just on the front end of it right now.

5 CHAIRMAN APOSTOLAKIS: So in one of your 6sensitivity analyses you assumed that all the human 7 actions --x 8MR. WACHOWIAK: Prevailed. All the post.

9 CHAIRMAN APOSTOLAKIS: Post initiated.

10MR. WACHOWIAK: Post initiated actions are 11 failed, except for the recovery of off-site power.

12That's one where the typical thinking for that was the 13 grid associated loss of off-site power, and it would 14be different people addressing that, but there are 15contributions from the things that are on site. So we 16 may want to relook at how we did that if there's any 17 dependence there on the no post accident operator --

18MEMBER MAYARD: Even though off-site power 19may be restored, the operator still would have some 20action, closing some breakers to get power into the 21 plant.22 MR. WACHOWIAK: Yes, that's one of those 23areas where we made the statement, "Yes, we did that." 24 I'm not quite sure we recognized that those recovery 25 134 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433factors really imply operator actions when we did it.

1And we recognize that. I think that's 2getting into Rev. 1 of that part of the analysis. So 3 it's another thing to verify.

4 Oh, I know why I went back. I left 5something off the slide. Part of the process that we 6did look at is on the back end. You know, you put all 7of these things in your fault tree models, and you 8 could end up with cut sets that have a whole bunch of 9different operator actions in them. We did do an 10evaluation of the cut sets to make sure that we didn't 11 have any highly dependent operator actions there.

12 There were a couple of things that either 13they weren't dependent or they didn't exist or we did 14a judgment call. Really are the two operator actions 15together -- is the value that was used sufficiently 16high that it really would be expected to cover the 17combined action? That would be the case where we 18would have some of those, where there are .2 for 19operator actions, .2 times .2 for the double actions, 20probably like a range where we'd expect anyway. So 21 that process went through on the back end to look at 22 those.23 I think we talked about the success 24criteria a little bit earlier. Hand calculations, 25 135 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433bounding type things for things that we just knew the 1answer to ahead of time. Design basis assumptions for 2things that matched up well with the Track G analysis, 3 and then we used MAAP results for the other things.

4We're in the process now of resolving 5where we should be between that and track. When we 6did look at the success criteria, the way we arranged 7this was we didn't just say this system, what does it 8have to do. We looked at it in the context of the 9sequences where the systems were used, took a look at 10all the sequences, looked at the different attributes 11of those sequences and determined if there were any 12specific limiting sequences to use for that success 13 criteria, and we used it all.

14So in some cases the success criteria 15might be conservative, but we tried to apply the same 16success criteria to the same functions throughout the 17 PRA just to make it simpler for analysis purposes.

18 We're working on a topical for this that 19 we've been discussing with Nick and others.

20CHAIRMAN APOSTOLAKIS: You say all 21sequences reviewed. Was the PRA reviewed by anybody?

22MR. WACHOWIAK: What I meant there was 23when we were determining the success criteria, we 24didn't just look at a system in isolation. We looked 25 136 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 at the system as to how it was used in the sequences 1where it was used in the PRA. So we looked at all of 2 those. All of the sequences of a particular system 3 was credited for success.

4Then what we did was we went through that 5list and said, okay, what are the attributes of these 6 different sequences, and is there any one particular 7sequence or one or more -- actually on some there were 8 two sequences -- that really would make that a more 9limiting success criteria on that particular function?

10And the ones that were the limiting, that 11had the limiting attributes were the ones that we used 12 to determine the success criteria for the system.

13CHAIRMAN APOSTOLAKIS: Yeah, I understand 14 that, but this is a broader threshold now.

15 MR. WACHOWIAK: Okay.

16CHAIRMAN APOSTOLAKIS: Did you have a 17 group reviewing the PRA itself?

18 MR. WACHOWIAK: Outside of this project, 19 no. So we got various contractors and subcontractors 20 looking at different things, but they were all under 21 the task that I am.

22CHAIRMAN APOSTOLAKIS: Because you 23mentioned the ASME standard, and as you know, there is 24a PRA review requirement there. Of course, I mean, in 25 137 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 this design certification business, what is the role 1of the PRA? Because in the ASME standard you are 2 supposed to use it for some real action. So the PRA 3 is very important.

4But here my feeling is that this is really 5a supporting kind of analysis. It's not essential, 6 isn't it? Maybe the staff can answer that.

7 I mean, does the PRA have to be PRA 8 reviewed?9MS. CUBBAGE: I mean, it is a requirement 10in Part 52 that they do submit the PRA, and it is 11 primarily used to insure that the insights have been 12incorporated in any design requirements that were out 13of the PRAare factored into the design.

14 But I guess to some extent you're right.

15It is more of a supportive tool, and it also helps us 16 guide our review to the more risk significant areas.

17CHAIRMAN APOSTOLAKIS: The moment you say 18"insights," it sends a message. Don't do it. Any 19time you use the word "insights," not you personally.

20I think that the word "insights" should be banned from 21 the English language.

22"Insights" means made by the state of the 23art or state of the practice job, but they gained 24 insights, and 52, of course, says that, but there is 25 138 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 not a drive for the PRA to be peer reviewed.

1 If we do a minor change in an existing 2LWR, we demand all sorts of PRA reviews, but from this 3 thing, no.

4 MEMBER BONACA: -- at some point between 5conceptual design and completion of the plan. The PRA 6will be in a situation where, in fact, the peer review 7is worthwhile. I think at this stage I'm not sure 8 that I would consider it worthwhile.

9CHAIRMAN APOSTOLAKIS: Worthwhile and 10required are two different things. If the owner of 11the ESBWR decides not to do anything on a risk 12informed basis, his PRA does not have to be peer 13 reviewed. Only when the owner says, "I'm going to 14 invoke 1174." 15MEMBER BONACA: I'm only talking about --

16 you know, I would expect that this PRA would be much 17 more substantial when we can close up --

18 CHAIRMAN APOSTOLAKIS: Sure.

19 MEMBER BONACA: So at that point I would 20 expect that there would be a higher expectation.

21CHAIRMAN APOSTOLAKIS: No, but as Nick 22said already, that time is running out. I mean, if 23 you have a peer review group that comes back and says, 24"We don't like the HRA," you would say, "I'm not 25 139 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433going to change," unless they say, "Well, then we 1 resign." 2Because, you know, there are certain 3things that are cast in stone, as you are advised.

4Anyway, there is no requirement. You haven't done it, 5 that's fine. Let's move on.

6MR. WACHOWIAK: We do remember though that 7the part of the process that we're talking about here 8and that we talked about this morning is we intend to 9deliver a PRA to the plant that will be operated, and 10they will use that PRA. So somewhere before we get to 11that stage, they've got to have that or else they 12 don't have the complete package.

13 So the question is when, not if.

14MEMBER ARMIJO: In your internal 15procedures, I'm sure you have internal design reviews 16by independent parties, but whether that has to go 17outside of General Electric to some other peer review 18 I don't know, but certainly before you would issue a 19document like that to the utilities, you would have an 20independent design review of the work done to satisfy 21 your management that you've got a quality product.

22 MEMBER SIEBER: QA program.

23MEMBER ARMIJO: Yeah, more, I guess, from 24 QA.25 140 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433CHAIRMAN APOSTOLAKIS: Okay. Let's go on.

1That's the bottom line. Why are you 2 reporting 310 to the minus eight when Chapter 11 you 3 say that the mean value is eight, ten to the minus --

4what is this 310 to the minus eight? It's the median?

5 It must be the median.

6MR. WACHOWIAK: This is the value that you 7get when you use the point estimates for all of the 8 values. Now, what's in Chapter 11 using the 9 simulation code --

10CHAIRMAN APOSTOLAKIS: Yeah, you've got 11 the on site.

12MR. WACHOWIAK: -- the mean looks like it 13 comes out to be a different value.

14CHAIRMAN APOSTOLAKIS: It is eight, ten to 15 the minus eight, and I believe that's the number you 16should be reporting. I mean all of the regulatory 17documents refer to mean times. I mean it's not a big 18 deal. It's just problematic.

19MEMBER WALLIS: What's the worst it can 20 be?21CHAIRMAN APOSTOLAKIS: The 95th percentile 22 is around two, ten to the minus seven.

23MEMBER WALLIS: There's one up at E minus 24 five.25 141 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433CHAIRMAN APOSTOLAKIS: No, no, no, no.

1 The 95th percentile.

2 MEMBER WALLIS: Yeah, I know.

3CHAIRMAN APOSTOLAKIS: Which is 4remarkable, remarkably narrow, right? Think about it.

5 This is the media, 310 to the minus eight, and the 6 upper bound is maybe four times that, a narrow factor 7 of four for a design that has never been built, right?

8MEMBER WALLIS: Why is frequency on log 9 scale? This is not log or is it frequency?

10 CHAIRMAN APOSTOLAKIS: This is log.

11MEMBER WALLIS: Frequency. I had a little 12 trouble.13CHAIRMAN APOSTOLAKIS: No, this is the 14 frequency. This is 1.10, ten to the minus --

15MEMBER WALLIS: No, it depends on the 16 scale.17CHAIRMAN APOSTOLAKIS: No, but this is 18 from the computer probability.

19MEMBER WALLIS: Yeah, but is it 20 probability per unit of logarithmic increment or --

21MEMBER DENNING: It looks like it is.

22 See, these are equal logarithmic --

23CHAIRMAN APOSTOLAKIS: This is the table.

24The table is the result of the simulation. It says 25 142 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 the upper bound to the core damage frequency is 1.8, 1 ten to the minus seven.

2 MEMBER WALLIS: What is the mean?

3CHAIRMAN APOSTOLAKIS: The mean is eight, 4ten to the minus eight. I think you should report the 5 mean.6MEMBER WALLIS: Well, I have a problem.

7 Is it plotted on a log scale? Now, I concluded from 8 their numbers up here that they must be probably for 9unit of frequency, not per unit of log frequency.

10 It's actually different.

11CHAIRMAN APOSTOLAKIS: Forget about the 12 figure. The figure is just for communications. The 13 table is actually from the computer.

14 MEMBER WALLIS: So it's the table.

15 CHAIRMAN APOSTOLAKIS: The table, yes.

16MR. WACHOWIAK: Both are from the computer 17 program.18CHAIRMAN APOSTOLAKIS: The table is the 19 real frequencies.

20 MR. WACHOWIAK: Yes.

21 CHAIRMAN APOSTOLAKIS: Okay.

22MR. WACHOWIAK: One of the difficulties --

23CHAIRMAN APOSTOLAKIS: I mean does it 24 bother you to report eight, ten to the minus eight?

25 143 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 You seem to be --

1 MEMBER SIEBER: He's happy.

2CHAIRMAN APOSTOLAKIS: For heaven's sake, 3 that's low. It's way low actually.

4MEMBER SIEBER: We're falling into the 5 sun.6 MR. WACHOWIAK: To try to compare things 7 on an equal basis then, using that value in the mean 8from that particular computer program would be 9problematic for us because of all the different places 10where we're trying to compare. For the different 11scenarios, the fire, the floor and everything else, 12this number is what's comparable across the different 13 ones.14So I understand. I understand what you're 15 saying there, and we will investigate how to present 16 this. It generates difficulties in talking about 17 things like raw values and things. If you take that 18mean from there when the computer program is 19 calculating all of these other values using this 20 number.21CHAIRMAN APOSTOLAKIS: But it should be 22 using the mean, but that's easy to do.

23Anyway, do you know that the age of the 24 earth's crust is 310 to the ninth years?

25 144 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. WACHOWIAK: Yes.

1CHAIRMAN APOSTOLAKIS: So what you're 2saying here is that if we had a reactor built when 3the earth's crust will start to forming and we had to 4 run it then, then you are just an order of magnitude 5worse than that. It's an incredible number, isn't it?

6 Ten to the minus eight.

7 MR. WACHOWIAK: What we're trying to say 8here is that for just about anything that we could 9think of, we've found a way plus a diverse way of 10dealing with it, and in most cases more than that. So 11 what we're trying to say here, and I think we've used 12this in other presentations before is that we think 13 we've addressed everything that we know.

14MEMBER WALLIS: This is where actually 15permission to comment, I mean, just to talk about 16disgruntled employees rather than any other kind of 17 event, but people doing things to deliberately cause 18 an event really begins to be very important when you 19have numbers like this for the things that you 20 analyze.21 MEMBER MAYARD: Yeah, but one of the 22things that we tend to not take into account from the 23human performance standpoint are the positive 24attributes, like they did not take any repair 25 145 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 activities into account.

1 CHAIRMAN APOSTOLAKIS: Oh, I know that.

2MEMBER MAYARD: And if you don't watch any 3 of your emergency planning scenarios and stuff, what 4the human can do from a maintenance, design, 5modification, a lot of things that they can do that we 6 never take positive credit for in a PRA. So --

7MEMBER WALLIS: The probability of an 8operator going absolutely nuts is probably bigger than 9 ten to the minus eight.

10CHAIRMAN APOSTOLAKIS: I think most of 11 them dominate this culture.

12 MR. WACHOWIAK: It's tentative.

13 MR. WACHOWIAK: Well, I think what we're 14trying to accomplish here is to address the things 15 that we know about and the things that we can know 16about using this methodology. Is that the actual core 17damage frequency? Well, we don't know because there's 18things we don't know about, and maybe there's other 19 tools that are better for doing that.

20But for using this method, we think we've 21addressed just about everything associated with the 22 design of this plant to make the chance of a core 23damage event so remote that we aren't, that it's a 24vulnerability that's been addressed. We don't see 25 146 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 that anymore.

1 So I agree that there might be something 2out there that isn't included that could address core 3damage frequency, and it's probably things that can't 4 be addressed using these methodologies.

5CHAIRMAN APOSTOLAKIS: Let me ask you 6 something else. I mean, we're going now to a 7 different place. It is tempting to me to go back to 8the beginning of the use of PRA now that you ask.

9It's up to you. Look at the numbers we were producing 10at the time, although the reactor safety numbers were 11not that bad, and then see what happens in the 12intervening years, how many times we were surprised 13and knew things happened and so on, and this agency 14 had to promulgate extra rules.

15 Doesn't history apply here? Can I count 16the number of times I was surprised in the past and 17 say, well, gee, maybe in the future I'll be surprised.

18 Therefore, is there anything different here?

19MR. WACHOWIAK: I wouldn't discount being 20 surprised in the future.

21CHAIRMAN APOSTOLAKIS: But is it a 22different situation? I think in some sense it is 23 because --

24MR. WACHOWIAK: But I think it is a 25 147 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 different situation.

1CHAIRMAN APOSTOLAKIS: -- you have just 2said that we are eliminating a lot of the stuff we 3 learn from experience in the NWR.

4 MR. WACHOWIAK: That's right.

5 CHAIRMAN APOSTOLAKIS: So we didn't have 6that benefit at that time, but you know, I remember 7 the first PRA topical meeting in Newport Beach, 8 California, where everybody was reporting for -- you 9 know, we did a fault tree analysis for this system, 10 and the magic number was ten to the minus six, which 11 became ten to the minus four as people came to their 12 senses.13 So I don't know. I don't think you or I 14or anybody has an answer to that, but this is 15something, I mean, when you go to such low numbers and 16you have new designs that have never been billed. You 17really have to worry about these things. That's where 18structure of this defense in depth comes to the 19 rescue.20MEMBER WALLIS: Well, they don't need a 21 containment if they've got ten to the minus eight.

22CHAIRMAN APOSTOLAKIS: We are way over 23 time here.

24MEMBER WALLIS: Are we or not? Is he 25 148 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 going to finish up?

1 CHAIRMAN APOSTOLAKIS: We're supposed to 2 finish at 11:30. He is so slow.

3 (Laughter.)

4 MR. WACHOWIAK: Just a few more pages.

5MEMBER WALLIS: This is a very funny 6 figure to me.

7CHAIRMAN APOSTOLAKIS: Which one is funny?

8 This one?9 MR. WACHOWIAK: This figure here.

10MEMBER WALLIS: I would conclude that you 11way over designed your LOCA response and you way under 12designed your loss of power. I mean, if that's the 13 dominant thing, maybe you should have another diesel 14or something. Maybe you can relax your LOCA defense.

15MR. WACHOWIAK: The actual thing is the 16 loss of feedwater, is what's --

17MEMBER WALLIS: You can do something about 18 that, bring on more pumps or something.

19 MR. WACHOWIAK: The loss of power causes 20 a loss of feedwater.

21CHAIRMAN APOSTOLAKIS: I mean these event 22 trees are awfully similar, aren't they?

23MEMBER MAYARD: We don't want to penalize 24 them though for wearing down the --

25 149 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 (Laughter.)

1MEMBER WALLIS: It's a very funny design.

2 Only susceptible to one major accident.

3 MEMBER KRESS: That's all right.

4MEMBER WALLIS: You can relax your LOCA 5 now. You don't need anything like as much water and 6 all of that stuff because it's --

7 CHAIRMAN APOSTOLAKIS: No, no, no.

8MEMBER WALLIS: That was ten to the minus 9 eight.10CHAIRMAN APOSTOLAKIS: If you relax it, 11 the contributions will change.

12 MEMBER WALLIS: Once the economists look 13at it, they'll say, "Wait. How are we paying for this 14 medium LOCA .8"?

15MEMBER KRESS: We had a concept once that 16tried to allocate the risk contributions to various 17 sequences. It just never went anywhere. It was a 18 bad --19MEMBER WALLIS: No, but there is a --

20there must be an economic penalty to way over 21 designing for LOCA.

22CHAIRMAN APOSTOLAKIS: Well, that's not 23 our business.

24MEMBER SIEBER: Not with passive systems.

25 150 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: That's why I was surprised 1 though.2CHAIRMAN APOSTOLAKIS: Okay. What do you 3 want to tell us?

4MR. WACHOWIAK: I want to say that we 5understand why this is, and we are looking at that 6from other reasons because as you said, over designing 7 for a LOCA versus these, the thing that causes this is 8really more of an operational issue, and we're looking 9at optimizing it because of operations and economics.

10So before we're done, you'll probably see 11a difference went up there, but not because the PRA is 12 driving that.

13We did accomplish what we wanted to do 14 here. Bypass we wanted to be negligible. ATWS we 15 wanted to be negligible. High pressure sequences we 16wanted to be negligible, and the containment can deal 17 with these.

18We think that the design is robust. We 19have put in, as inputs, things that helped us look at 20what the design was capable of doing. We think we 21 came to that. The probability of a severe accident, 22 say, it's remote, and we think that -- well, we know 23 that the use of the PRA as a design tool helped insure 24that because as we went through this process, there 25 151 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433are thousands and thousands of things that need to be 1optimized, and we continue to come back and say, 2"Okay. What is that going to do in the PRA so that we 3 insure that it stays the way that we like it to be?" 4Combination of the passive safety and 5active non-safety systems and the diversity that we 6 built into this thing because really what gets us to 7 the remote chance of a severe accident based on the 8techniques and the things that we know about this 9 design.10MEMBER DENNING: But you didn't put up the 11one that shows now take away the actual systems and 12 what happens.

13CHAIRMAN APOSTOLAKIS: Yeah, the 14 sensitivity analysis.

15MEMBER DENNING: The sensitivity analysis.

16CHAIRMAN APOSTOLAKIS: These are very 17convincing arguments, and you didn't say anything 18 about them.

19 MR. WACHOWIAK: Yeah.

20CHAIRMAN APOSTOLAKIS: We'll get that next 21 time. Okay?

22MR. WACHOWIAK: We could do that next time 23 or we can try to find a slot tomorrow for it.

24MEMBER DENNING: That would be nice to 25 152 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433just have a short discussion of it because it is so 1 interesting.

2 CHAIRMAN APOSTOLAKIS: Seeing all of the 3 active systems.

4 MR. WACHOWIAK: And we can talk about 5 that, but realize when I talk about those, those are 6 all based on our Rev. 0, and we're in the process of 7revising that to Rev. 1. So the numbers may be a 8little bit different when you finally get the whole --

9CHAIRMAN APOSTOLAKIS: We have to discuss 10also this timing business because, you know, if we 11have a meeting in December and we go into the details 12and we don't like something and you say it's took late 13 now and we can't change it, I mean, we have a problem.

14 So have that in mind.

15 MEMBER WALLIS: They have a problem.

16CHAIRMAN APOSTOLAKIS: They have a 17problem, right. They have. That may be fine, but you 18 know, this is not just a formality.

19MR. WACHOWIAK: And I would agree that if 20we had the meeting in December and that was the 21 conclusion --

22CHAIRMAN APOSTOLAKIS: So you guys will 23contribute to the discussion whenever, the next 24 meeting. Okay?

25 153 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. WACHOWIAK: Okay.

1CHAIRMAN APOSTOLAKIS: Okay. Thank you 2 very much. This was very informative.

3 We'll reconvene at one o'clock.

4(Whereupon, at 11:57 a.m., the meeting was 5 recessed for lunch, to reconvene at 1:00 p.m.)

6CHAIRMAN APOSTOLAKIS: On the record.

7 Okay. We're on the record.

8MR. THEOFANOUS: I'll be covering Chapter 9 21 of this 33201 and I hope to, by this coverage, to 10create an opportunity for you to ask questions.

11Obviously, I cannot cover all the details, but I will 12 try to skim over the whole subject.

13 The work was done by myself and Professor 14 Dinh who used to work with me until about a year ago 15and now he is chair of Nuclear Safety at the Stockholm 16Institute of Technology in Sweden. And what do we 17mean by severe accident treatment is that we are 18considering containment integrity threats due to 19 severe accident phenomena. So the part of phenomena 20we're going to cover this afternoon. I'm not going to 21cover it myself, but it will be covered in the 22following discussion, containment integrity due to 23 decay heat removal failures and those failures might 24 occur in the long term. So this is more like a 25 154 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433systems question and that's why we're leaving it to be 1 handled separately in a separate positive PRA.

2Our approach is an interactive assessment 3management approach. This is because this is a new 4 reactor basically we're working on. The reactor was 5just finished in design, we're finishing up some of 6the design. So we had an opportunity to affect the 7design to the interest of forwarding the final touches 8to the reliability of the safety process of this 9 ESBWR.10So we worked on it for about a year and 11 during that time as you will see, we developed a 12number of new procedures and hardware that we think 13improve even better this severe accident rate of this 14 reactor. Because of the nature of the passivity of 15the reactor, because of the very extremely low core 16damage frequency, we felt that the right way of doing 17this severe accident treatment was by placing great 18emphasis on bonding high confidence evaluation. It 19 wouldn't do us any good to say we had a reactor that 20has 10 to the minus umpteen core damage frequency but 21now we have high probabilities where we will know very 22well how the debris may attack the concrete on the 23floor of this reactor should it ever happen to occur.

24And as a result of this high confidence 25 155 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 evaluations that we're ascribing for and emphasis on 1 bonding evaluations, we came up with a number of new 2procedures and hardware that would aim for eliminating 3some of those analyses for which we could not 4 accomplish that goal. So our conclusion now is that 5containment failure is physically unreasonable for all 6severe accident scenarios except the postulated large 7 steel explosions in very deeply flooded low drywell.

8 It's not that we're saying that these 9 kinds of scenarios, hypothetically large explosions, 10we're not saying that they will fail. What we're 11saying is we can not demonstrate with high confidence 12and high reliability the assessment that this will be 13 so.14 MEMBER WALLIS: Are you saying that --

15 MR. THEOFANOUS: It's important to point 16out that it's less than one percent of the core damage 17 frequency falls in that category.

18MEMBER WALLIS: You're saying all 19 scenarios, but there is a scenario where the reactor 20vessel is over pressurized and it pops and that 21popping of the vessel leads to popping of the 22 containment. You're not talking about that kind of --

23MR. THEOFANOUS: No, no. All scenarios 24 that are --

25 156 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: You're talking about core 1 melt scenarios.

2MR. THEOFANOUS: Core melt scenarios, 3 right. And this scenario that you are suggesting is 4such an extreme scenario that it's not even showing 5 anywhere in this core damage frequency.

6 MEMBER WALLIS: It's in the PRA though.

7MR. THEOFANOUS: Yes, it's in the PRA.

8 Right. But it's what you call a residual risk. I'll 9 discuss residual risk in a moment.

10So the people thought the issues in our 11assessment that we had to basically consider and then 12take action on are summarized here. There are just a 13 handful. This is a simplified reactor and we thought 14 it really requires a simplified approach rather than 15 a very complex approach.

16So the first question was do we want to go 17with universal retention. I don't think I need to 18explain to you universal retention, but it's a very 19 popular scheme now since we developed it a long time 20 ago. It does sound that, and actually I have written 21 papers on that, it does help that not only the ESBWR 22but all BWRs are ideally suitable for this concept, 23ideally suitable because they have welded steel and 24they don't suffer from this focusing effect that 25 157 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433created a monetary issue or problem for pressurized 1 water reactors.

2 Why you decided not to pursue that here, 3that's because as you know all the boiling water 4reactors, the lower head is perforated by penetration 5 so that if you really want to make sure that you're 6 going to hold everything inside, you have to support 7this penetration from falling off. So we suggested 8that as a possibility because I was very concerned 9 about making sure that this reactor, in fact, I was 10very concerned about all reactors, we cannot assure 11 the coolability if something should happen.

12So I was concerned at least for these new 13generation reactors that we can assure coolability if 14ever this was to occur. So I said why don't we put a 15plate somewhere on the support guide tubes and weld it 16on there on the outside on the housing so that one 17supports from the other and in fact, such a plate 18would not only be good for holding everything up in 19case a melt came to the lower head, but actually it 20 would be quite beneficial in cutting off the driving 21force behind velocity steam that would come out in the 22case of high pressure which also takes care of the so-23called containment heating problem. But that deal was 24not agreeable to the designers and I was corrected and 25 158 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 I forgot to change it here, not really the designers, 1 but the design managers. So they felt that it would 2be causing a lot of problems with the maintenance, the 3 operation.

4 Then as we were discussing those things, 5 then we came up with another idea that we thought at 6the end actually may be awfully better. So sometimes 7there's a silver lining. Sometimes it's better to 8 find some difficulty or some resistance because then 9you come up with something better. So we came up with 10the ex-vessel core coolability idea that I believe is 11 more robust than even the in-vessel and so we're now 12 with the ex-vessel coolability.

13 The reason that this came to that 14isbecause natural ex-vessel coolability cannot be 15 assured. I don't think I need to explain that to you.

16 You know this very well. It hasn't been possible to 17demonstrate that if you have a melt that is allowed to 18 fall on the floor and you have water before or after, 19 I don't care when you have it, you cannot demonstrate 20 that this thing is coolable.

21In Sweden, the Swedish reactors, as you 22 know, they have also very large pools under the 23 reactors. In fact, they put those pools there, many 24meters, I think it's maybe about ten meters deep.

25 159 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 Professor Becker suggested that they put those pools 1there so that the melt as it comes out, fragments and 2it's supposed to be coolable. Well, that creates 3great news with steam explosions first of all and 4we've done calculations of that and you find that 5indeed the pedestals in that case will not hold it if 6you had a steam explosion and even the coolability 7problem is not right because you have such deep pools 8 that actually will not remain coolable and we have a 9 problem this way.

10 So we have come up within this what you 11call the Boundary-Internal Melt Arrest and 12 Coolability. I think I should use this for pointing, 13Boundary-Internal Melt Arrest and Coolability device, 14 this BiMAC which accomplishes this purpose very well 15 as you will see in a moment. It will accomplish the 16purpose of first of all allowing you to not have water 17there at the time that the melt comes out. That's 18 where the measurable core is going to occur and if 19there is a concern about steam explosion, that will be 20 the time with your concern about steam explosion.

21 The reason that it allows you to start off 22without having the lower drywell flooded is because 23 the moment that water is added to BiMAC and this can 24be done essentially instantaneously after the melting, 25 160 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433BiMAC is effective to operate as an impenetrable 1 device. That is a boundary of which the melt cannot 2 penetrate.

3 The concept is very similar to universal 4retention in the sense that we have still a boundary.

5We have water coolable below. Because of this as long 6 as the thing is actually a nuclear boiling and we've 7 demonstrated that's the case, the temperature on the 8other side of the still is so low that actually the 9fuses melt. And the size and the wall thickness of 10 those pipes are such that so that they have 11 significant integrity so that even a small steam 12 explosion will be no problem to them.

13 Having established this robust coolability 14 posture for this reactor ex-vessel, we won't need to 15be concerned about ex-vessel phenomena because of the 16 nature of this device basically will catch anything 17and everything that comes down. There is no scenario 18 dependence. It's going to come 20 percent first, 30 19 percent later, but you know it's not going to be 100 20percent coming in all at once. But even if it did, 21 that's fine. It's going to be all contained inside.

22 So really that leaves only two more things 23 to be concerned about and that is what happens if we 24have some steam explosions there if for some reason 25 161 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433it's part of the accident and we end up with deep 1water pools especially subcool water pools in the 2lower drywell. And then other one is what if we have 3a high pressure scenario in which the vessel failed at 4high pressure and gave rise to what is known as the 5 direct containment helium.

6So for this problem, we ended up with 7 BiMAC. For this problem, we ended up with some 8procedure changes and some hardware changes that 9minimize the scenarios that gives us all the water in 10 the lower drywell and for this one, we basically did 11 nothing except do a fundamentals based analysis that 12shows that no matter what happens the ESBWR drywell in 13containment will not be overstressed by the direct 14 containment heating.

15So my opinion, the serious issue as far as 16the safety of this reactor as far as your evaluations 17 is this one here. Does it work as we say it's going 18 to work? The other ones, this is one percent of the 19 whole CDF and that's another one of the whole CDF.

20 Actually, it turns out quite interesting 21from a very fundamental and from a technical point of 22 view. So I don't know to what extent you want me to 23 go over those, but I have them here and I will start 24 going. But please if you feel that you want to go 25 162 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 more, spend more time here, I'll be very happy to.

1So the way that I have arranged this 2 presentation is in the same order as in the report 3 which I hope all of you have had with you since last 4August, actually since last November and I hope you 5had a chance to look at it. But here, I'm going to go 6in reverse order. I'm going to first go here and then 7 here and then there.

8All right. Just to summarize then, the 9severe accident threats and failure modes, direct 10 containment heating, because in here, it's a failure 11of a reactor drywell. I remind you that the ABWR 12 assumed that in this scenario would fail the drywall 13and that was one of the reasons that actually the NRC 14staff in their evaluation report, they assigned the 15 maximum possible for conditional containment failure 16 probability. It was because of that support failure 17and they did a very hypothetical analysis actually.

18 I still don't know how they ended up with failure 19because it's not possible to make those reactors fail 20 with suppression pool because of this. You will see 21 in a moment why. So that's one issue.

22The other issue is a much smaller one and 23that is if you don't fail catastrophically in the 24 drywell, can you fail the liner because the liner is 25 163 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433containment boundary and if you did, then you would 1have trouble. The liner, of course, will fail because 2of thermal effects if it fails while the drywell will 3 fail because of pressure overstress.

4Then ex-vessel explosions, the concern 5here is the pedestal of course along with the liner 6failure because of the energetics of the explosion.

7 But here in addition, we have a new twist because of 8 the BiMAC and we need to know if the pipes, those 9 pipes that they put there will survive an explosion.

10And finally on the basemat melt 11penetration, that is the, I guess, the current state-12 of-the-art. I don't know if you want to call it 13state-of-the-art, but the current approach is that if 14 what if we can show that we're not going to penetrate 15the basemat melt in 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> maybe you're okay.

16 That's the 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> rule and then maybe if we can show 17that actually for this reactor you can show that it 18 will survive maybe up to 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, maybe even more 19 than 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />. That's all right.

20But with the BiMAC, we eliminate all that 21because there would be no attack at all and for a BWR, 22 this one, with a small containment, that's very good 23because you don't have to worry about any condensables 24 coming in.

25 164 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Therefore, we have translated the problem 1from sort of hypothetical analysis, basically 2sharpening the pencil, about how long does it take the 3 melt to go through the process and does it go faster 4this way or this way and so on, all this stuff. Here 5we're putting a boundary that cannot be penetrated and 6therefore our concern is to show that what does it 7take to fail this boundary. So our problem is to 8something else, more of an engineering, more tangible 9 and I'll show you in a moment BiMAC can be tested the 10 full scale. So it's a much more, much better domain 11 in which to operate on technical grounds.

12 By the way, I know you have heard, that we 13have large quantities of melt on the floor. People 14are going to say, I'm not going to say because I don't 15 care to, but people will say, "We don't know.

16Actually it's maybe going faster this way than that 17 way." So a big issue these days about that again from 18what I heard. So this is good and this is simple and 19 very easy to apply and I believe one day I hope many 20reactors will use this, not only the ESBWR, very 21 similar things, pipes and downward.

22 So for the BiMAC again we need to worry 23about burnouts, something I call burnout. The burnout 24can occur if the thermal loading locally or in one or 25 165 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 more of those tubes is about the critical heat flash 1of the water boiling on the inside. That's one way of 2 failing.3Another way of failing is that you have so 4much power going into those tubes that actually the 5 two-phase flow actually gets water depleted. So 6actually you get the sense of the water coming in and 7the steam coming out around so in that case there's no 8 water to create boiling.

9Of course, we need to worry also about 10 stability because we want to make sure the flow is 11 reasonably stable going through.

12 And finally, we need to worry about melt 13impingement, melt coming out heating whatever is on 14 top of the pipes making sure that they will not be 15 eroding. This we call it sacrificial layer we put on 16 the top essentially to protect the pipes.

17A couple more, an illustration here, a 18depiction of the three failure modes and please 19forgive me for too many acronyms, but DCH as everybody 20knows, EVE is ex-vessel explosion again is well known, 21and BMP basemat melt penetration. By that we're 22referring to the melt eating up the concrete. So 23 those are our three issues and was read and 24interpreted here. What it shows is that's too 25 166 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433complicated, but mainly I want to show you this is 1 scenario three is one percent of the CDF which is 10 2to the minus 8 roughly. That is the high pressure 3 scenario. Ninety percent of low pressure scenarios so 4 all the high pressure scenarios there which is shown 5 over here. We need to worry about DCH.

6For all the low pressure scenarios over 7 here, scenario one, we need to worry about DCDs. So 8therefore, the question arises in those scenarios, how 9 many of them, what fraction of them, are very deep 10water pool and what it shows here less than one 11percent and that would be a deep water pool. The rest 12 of them are either no water at all or very low water.

13 So in this way, this is what I meant.

14Those are relatively unimportant because that's one 15percent and that's one percent. However all of them, 16 any accident, anything that's going to lead to large 17melting of the core has to be dealt with in a 18coolability point of view. So that's why all 19pervasive features in the accidents is this part here 20which surrounds everything. So that's why I put a lot 21 of emphasis on that.

22Here maybe this looks too busy, but I 23 think I want to point to a few items that are 24important for our analysis and I have marked those 25 167 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433with red so that I remember also not to forget 1 something. First of all, upper drywell/lower drywell, 2 that's our nomenclature. As far as DCH, the volumes 3are important of the lower drywell. That's where the 4mixing is occurring. It's going to lead to 5pressurization of the drywell. Very fundamental to 6BWRs and relative to high pressure scenarios and 7 containment heating is these vents which allow these 8 volumes to vent through the water into the wetwell.

9So that when you pressurize, you're 10 initially releasing. Remember now. We're talking 11about here many hundreds of minutes of speed of the 12steam and how is it coming out of here. Actually, 13 it's quite phenomenal what can occur and we've done, 14you see a tremendously interesting gas flow particles 15occurring over there. So in almost no time at all, 16this pressurizes and the space here is totally out of 17 scale. That's why I emphasized to tell you this is 18not to scale. Over here, there is a restriction 19because of the supporting vessel. So over here, there 20 is like a 70 percent reduction in the flow heading.

21So first, we pressurize that. Then you 22 have to pressurize that and that behaves as a closed 23 volume as long as the vents are not clear just like 24the LOCA. So if you're interested in the integrity of 25 168 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 this thing, you have to be sure you calculate 1 correctly vent clearing.

2 All right. But then once the vents open 3up, then the issue is whether the vent capacity of 4those vents can compensate for the supply energy in 5 the upper drywell. We'll show you what that is.

6 All right. So as far as DCH, those are 7the key components and there's still another one that 8 I want to point out here and that is again another 9 present core well in that there is some skirts over 10here that they are calling refueling skirts and 11basically they are closing off that space. This is a 12metallic head of the drywell. That can become a 13limited component in pressure. We're showing that the 14force is isolated so whatever happens over here, that 15upper head doesn't know it because of that. It has 16holes basically that are communicated from here to 17 here.18 The other important thing that's crucial 19for an integrity point of view is that the head is 20immersed in a whirlpool that is a pool and the heat 21flies from the upper head because there is 22installation here into the drywell head actually is 23very low. It's so low that you won't even cause water 24 to boil. So if you want to do a structural analysis 25 169 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 of this thing, you should be doing it with a cold CAD 1of the upper drywell. That's very important for 2 assessing the fragility of the drywell.

3CHAIRMAN APOSTOLAKIS: What is the name of 4 that pool above the spherical thing?

5 MR. THEOFANOUS: This pool over here?

6 CHAIRMAN APOSTOLAKIS: Yes.

7 MR. THEOFANOUS: Well, this pool is --

8CHAIRMAN APOSTOLAKIS: It's a separate 9 pool. It's not --

10MR. THEOFANOUS: It's separate. It's 11 actually separates the PCCS from the pools.

12 CHAIRMAN APOSTOLAKIS: Yes, PCCS.

13 MR. THEOFANOUS: It's separate from that 14and it is there, I guess, I don't know why that is 15 there. It's just a reactor for the fueling pump.

16 They want to have some space there for refueling.

17About steam explosion, you're talking 18 about this page over here and that is something like 19 maybe seven or eight meters deep and it has about ten 20 meters in the round which is a really big space and 21your concern is that there are doors and those are the 22hatches over here and here through which people are 23coming for the refueling purposes. The pedestal is 24made out of two and a half meters reinforced concrete 25 170 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 and that is not just to take care of spillage. This 1is just because of structural consideration because 2 the reactor is very big and very heavy. That's what 3defines the very robust walls. To my knowledge, those 4 are the most robust walls that are in pressure-5 suppression containments.

6 The thing on the DCH that I forgot to 7mention and I do want to point out is these little 8horizontal lines that go like that and those are 9called lips and I found out. I was suggesting to 10people that we want to put lips there because we will 11most likely, this DCH most likely is going to fail the 12liner here just by splashing about. If it hits, it 13would very likely fail. So I didn't want to hear a 14 communication from the back liner space from here up 15 to these parts of the container boundary.

16 But then I found out and we checks that 17this is part of a normal practice. Every so often, 18 they make lips from the liner that are going to the 19concrete sort of like compartmentalizing the liner.

20 So you could very well, if you failed the liner over 21here, but that doesn't communicate with the back liner 22space over here. So in addition to having those 23anchors into the concrete that hold the liner, you 24 also have those lips.

25 171 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Going back to then steam explosion as you 1will see, I don't believe there's any problem with two 2and a half meter reinforced concrete for steam 3explosions for the pedestal, although if you do a real 4humongous steam explosion like the ones that we can 5actually compute, we find out that we are getting 6there even to the upgraded. We also did structural 7vibrations and we found out that although normally 8people thought the pedestal would take about 100 or 9150 kilo-Pascal seconds, it turns out we're showing 10even 600 kilo-Pascal seconds almost four or five times 11 that with these walls you will begin to just reach a 12 seepage. Also it's very robust from an explosion 13 point of view.

14But it is initially the hatches which are 15 likely to fail if they are overloaded and of course, 16 there is the issue of the BiMAC that we want to make 17sure that we don't run above 600 kilo-Pascal seconds.

18 CHAIRMAN APOSTOLAKIS: I don't quite see 19 where the BiMAC is.

20MR. THEOFANOUS: You'll see that in a 21 moment. I'm coming to that. That's the third item.

22I just finished with the explosion. So the third item 23is the basemat melt penetration which we said we want 24to protect with the BiMAC and I'll come to the next 25 172 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 one to show you the design, but the BiMAC fits right 1in here covering the whole space and the last point to 2make here is that with BiMAC working plus the PCCS we 3have no possibility of long term failure of this. And 4 that's very comforting.

5So here is the BiMAC, and the concept 6basically is make a jacket with pipes and those pipes 7are lined up with some intonation. It is largely a 8two dimensional point. We have chosen this to be ten 9degrees and that ten degrees comes from the idea that 10a ten degree is the critical heat flux because 11remember now these pipes are going to heated from 12 above sort of like that.

13All right. Now as you increase from there 14as you go to a different level, of course, you make it 15possible for the evaporation to go at higher 16velocities and that creates more agitation and most 17 important, we see a wetting of the wall as the vapor 18sluices by and when that happens, you get an increase 19of the heat flux and that increases pretty steeply for 20 about up to about 15 degrees or so. Then it sort of 21levels out a little bit and therefore you go to a very 22high orientation when it goes very high. So our 23 interest here is to try to cut that. So that's why 24it's not five, ten degrees. Then those pipes then 25 173 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433come, here you have a vertical segment. In this way, 1then we can protect the floor. We can protect the 2 walls. We protect even the sumps.

3The other consideration here is to making 4sure that there is enough capacity inside of this 5 dish, if you like, that will catch not only one core 6 but more than one core. We can catch four cores.

7MEMBER WALLIS: This is made of concrete, 8 this brownish stuff or is this --

9MR. THEOFANOUS: This is the normal 10 concrete. This is something that's in there.

11 MEMBER WALLIS: What's this?

12MR. THEOFANOUS: This other stuff here, 13 that is additional concrete. We call it sacrificial 14 material. The one on the top especially will be made 15out of refractory material like zirconia that would 16resist impingement of the melt. From the point of 17 view of, let's say, having a melt on the top of it, I 18don't care whether there is any concrete or not. I 19don't care what concrete is there, but we were 20concerned about possible melt coming out in some 21velocity and coming and hitting it and penetrating 22those pipes. Of course, if you penetrate the pipes at 23 that point, you're --

24MEMBER WALLIS: Again you say parallel 25 174 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 pipes. I presume it's not a cone one there.

1MR. THEOFANOUS: It's a cone that is a two 2dimensional cone. It's like what you see here is a 3 cut through this way.

4MEMBER WALLIS: Okay, but in the other 5 direction --

6MR. THEOFANOUS: The other one is 7 straight.8 MEMBER WALLIS: Okay.

9MR. THEOFANOUS: Okay. So I'll come to a 10--11MEMBER WALLIS: It's almost like a valve.

12MR. THEOFANOUS: Yes, like that. Now it's 13very interesting to point out something that's a 14 little harder to conceive, but I think I can explain 15 it that if you make this cut that is shown over here 16is through a diameter of the drywall that is normal to 17our view. Now if you begin to cut now with additional 18 slices going away or forward from there, then you'll 19find that this dimension is going to get smaller, 20 smaller and smaller.

21MEMBER MAYNARD: The angle would stay the 22 same.23 MR. THEOFANOUS: The angle will stay the 24 same. So that's basically going to get smaller. So 25 175 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433near the end, you're going to end up with channels 1that are very short in this direction, in the incline 2direction and long in the vertical direction. All 3 right. That's important and I'll come back to that.

4 Now it's important from the point of view of thermal 5 loading. I'll explain something that's quite 6 interesting from a thermal loading point of view.

7 That's that. That's one called boundary 8 internal. We are bounding inside. I personally 9believe that, I believe for a long time, that you can 10 have a lot of core on a floor like this like an inner 11reactor on the concrete with lots of water in the tub 12 and I think eventually it will become cooler.

13MEMBER WALLIS: Now this comes out and 14 floods from the top as well.

15MR. THEOFANOUS: I'll explain that in a 16moment, but let me finish my thought here which is 17eventually it will be cooled, but we can't demonstrate 18 that. That's the problem. So it's very possible that 19 this BiMAC never comes into play even if you --

20MEMBER WALLIS: When does it switch on?

21 Do you wait --

22 MR. THEOFANOUS: I'll come to that.

23 MEMBER WALLIS: Do you wait until you --

24MR. THEOFANOUS: You're very impatient.

25 176 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 You're very impatient. Just wait a minute.

1MEMBER WALLIS: Well, you've spoken all 2 the time. I can speak --

3MR. THEOFANOUS: You can ask a question if 4 you like, but don't talk to me for the future of 5 what's coming up.

6CHAIRMAN APOSTOLAKIS: Can you point to 7the certainly BiMAC itself? The BiMAC consists of 8 what?9MR. THEOFANOUS: The BiMAC consists of 10this dish and then you're right that there is an 11 integral piece of that which is the lines that are 12coming from GDCS, the lines which are coming in and 13I'll come to that in a moment. I want to give you 14 sort of a global view, not too much of the technical 15 details because I think more detail plots later I'll 16 show you.17 I'd also like to answer this very 18 question. What I said before, BiMAC works right away 19and the reason is that it is connected to the GDCS.

20So the moment you turn on the valve, that valve 21supplies this central part that goes that way. So 22it's filled up right away and the flow is running out 23 of all of those pipes which means that it's 24 essentially it's immediately effective for cooling.

25 177 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433However at some point the GDCS is going to 1 run out of water and especially --

2MEMBER WALLIS: I'm not quite sure of 3 that. We have an event. We haven't gone through the 4 vessel yet. Do you switch this thing on before?

5 MR. THEOFANOUS: No, no. We said --

6MEMBER WALLIS: When do you switch it on?

7 MR. THEOFANOUS: I'll get to it.

8MEMBER WALLIS: No, I'm with you. I want 9 to know what's happening.

10MR. THEOFANOUS: I said before but you 11weren't thinking. You were a little bit paying 12 attention to something else.

13 MEMBER WALLIS: Okay.

14MR. THEOFANOUS: I said earlier we are 15switching it on after the initial core of the melt.

16 We don't want to have water there when the first core 17 occurs because it will give us steam explosions.

18MEMBER WALLIS: The first core. That's 19 right.20MR. THEOFANOUS: Okay. So we're switching 21 it on after the first core and then after that 22happens, then we have plenty of water there and it 23 continues.

24MEMBER ARMIJO: When do you know that you 25 178 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 actually have melt on the floor? How do you know?

1 MR. THEOFANOUS: By temperatures. Again 2 we'll come to those issues in a moment. I think now 3 let's first just get -- make sure that we understand 4 that and then we'll find some of those.

5MEMBER WALLIS: What is on the lid? What 6 is the lid?

7MR. THEOFANOUS: Yes. If you let me 8 explain, I was trying to get there.

9CHAIRMAN APOSTOLAKIS: Okay. Let him 10 talk.11MR. THEOFANOUS: So we want this to be 12very basic in the concrete. I got into this stuff 13because I was telling you that this may never even 14come into play even if you have a problem. Okay.

15It's better than here. So what we have here, we have 16 a grate with support poles which are not illustrated 17here which are basically holding a plate also.

18Actually you don't know there's anything there. It's 19 just a steel plate that is thin.

20It's like two millimeters thick on the 21steel plate and the reason why it's thin is because I 22want it. That's not an important part of the 23consideration but I wanted it. It is a high pressure 24 melt injection, I have a melt jet coming out at high 25 179 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 velocity. I want this to melt right away rather than 1 splattering. I want to melt locally right away and 2 then the melt is going to flow right in here and I say 3I think most of it is going to be captured, because 4there is a high velocity steam. By the time, you have 5 it reach here, this high velocity steam has expanded 6 to about 20 times the area.

7Therefore, that stagnation pressure holds 8the plate down and therefore, this plate is quite 9 resilient as long as you have enough support for it 10 and it will just stay there and the melt will catch.

11Some melts will come out. That is discussed in the 12 report. I don't want to take time for this now. So 13small amounts will come out, but it's good to have it 14 there.15So now, I'm also showing that the 16operation of this initially will be with water coming 17 in from here and then going out from the vertical 18pipes, all the vertical pipes. Okay. Now later on 19 when the water finishes coming in here, then we have 20other downcomers which are, now this time the pool is 21already filled up and our water is coming in from 22those channels that are not heated or from the 23 downcomers.

24So I will talk about assessing elements 25 180 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433and how again we say it's very important for this and 1to inquire in this report the very high level ability 2for switching on the water coming in, high level 3ability also for not switching it wrongly on. All 4 right. It was not our job to design completely the 5sensing elements, however, just talking with designers 6 we have some ideas we can use, for example, 7thermocouples that are embedded in here so the moment 8that something came in, you know it's high 9 temperature.

10You can use also spring actuated nitrogen 11bottles which hold some pressure so that when the 12temperature goes high, some detector melts and then 13opens up and opens up the valve. I like basically to 14make this spositively activated based on very high 15 temperatures in here.

16MEMBER WALLIS: And the brown stuff there 17 is?18MR. THEOFANOUS: The sacrificial material 19like I said before. The important part is at the top.

20The top should be something that is very resistant 21 like a refractory material, zirconia, something like 22 that and it would be like 20 centimeters. You don't 23 need very much there.

24All right. So now it is from the point of 25 181 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433view of evaluating. I'm now going to the three topics 1and again please use your prerogative as a committee.

2 You can tell me that there's too much detail here.

3 Let's go to BiMAC. But my idea is to at least touch 4 on these issues to then finally come to BiMAC.

5So the direct containment heating, I'm 6going to cover a number of items. One item is 7containers depressurization. Is it possible you have 8 sitting there a vessel sitting there all buttoned up 9with very high temperatures for such a long time. We 10 asked that question first. Then the parameter range 11covered, also whole range parameters and results.

12 Then the thermal loads. And that finishes the 13 catastrophic part.

14Then over here is thermal loads to liner.

15And then we want to compare to fragility. The 16fragility we have nothing with do with the fragility.

17This was taken from one other chapter of 33201 and 18 summary of bounding approaches, we'll conclude, just 19 like finish here.

20 First, this potential for this container 21 depressurization, I should remind that you for PWRs 22 the DCH depressurizes.

23So I asked the question initially after 24can this happen and there are three possibilities.

25 182 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 There is the possibility of the isolation condenser, 1the pipe that goes into the isolation condenser.

2 There is the, here it explains in detail, main steam 3 line and then SRV that hangs off from there and so we 4 have three places.

5But actually the one that is important is 6this because here is closed first and it operates 7continuously and therefore you get especially into 8 high temperature or element in the element. You get 9 hydrogen produced and now you have also good thermal 10 conductivity material that is found out here.

11 The other thing to point out is this is 12high pressure source steam and hydrogen can carry all 13the heat. That's why you saw exactly that it's not 14convection in the steam environment. Normally, you 15 wouldn't have expected it. It was in fact it's high 16 pressure steam and high density, so it can carry all 17 the heat around.

18 So the question is will that fail and we 19took the typical materials for the construction 20materials and what is showing here is showing that 21this is the count of the material strength and this is 22the temperature and here is the main steam line, here 23is the isolation condenser and here is the SRV. What 24that means is that the main steam line should be 25 183 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 between 1,000 and 100 degrees to fail, the isolation 1projection in this range and the SRV in this range.

2 The SRV is made much more substantial because it has 3to take loads and that's why it has more strength.

4 That's why it can take higher temperature.

5 So the question now basically is can you 6 actually achieve this kind of temperature in the MSL 7 which as I showed you the MSL is heated by the flow 8going to the SLV. Just happens the MSL was a pipe, a 9 thinner one, so that's why it's here. The typical 10core transient, it doesn't really depend on what core 11you use. You will find that you get a lot of 12oxidation and get a lot of snowballing effect and you 13 get temperatures of 3,000 degrees for two and a half 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br />. So all you have there is like a quicken path 15with 3,000 degrees over here. Gas is naturally 16 conducting.

17And you ask the question will you ever 18reach 1,000 degrees? I think you will. But I didn't 19 want to just arrive just to that and say okay, we 20don't have DCH problem. It was kind of fun to work 21through the dynamics of the DCH as well. So that's 22 what we've done. That's what I want to show you how 23 that works.

24First point because I've seen analysis of 25 184 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433people not so BWRs. Several of my old NRC friends 1 wanted to see a real good BWR, DCH analysis. So I 2think this is going to do that. But I've seen people 3that have done interpretations of experiments as well 4as PWR calculations trying to get the fallout and then 5they average out over the whole cross-section area of 6 the space which is fine.

7Actually you see that we've done CRD 8 simulations and I think I'll get into a problem here 9 if somebody knows how to do that. You find out that 10 you get a supersonic jet out here. It's something 11 like 600 meters per second, this fantastic speed you 12 get and this jet comes and hits the bottom floor and 13is diverted and becomes a wall jet around the floor 14and then it's diverted again and becomes a vertical 15 jet with hundreds of meters per second out here. If 16you average it out, it just like a fizzle, well not 17 quite a fizzle but it is much lower of course because 18 this is a big area.

19So I believe that this as well as any 20other reactor that I have done for PWR before this 21process here is tremendously intense. In fact, I was 22so curious about this that I even did a few 23 experiments in my lab with a scale to that. We did 24some experiments and it's an amazing force. So 25 185 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433there's very little doubt in my mind that we really 1 have a high pressure melt ejection unit right here.

2And what the other fundamental physics 3here is that you have a liquid mass. Liquid masses 4are microscopic in inertia. So therefore it is not so 5easy to accelerate those masses and get them out 6before they fragment and mix with steam. So the 7reason we have this very fine fermentation is because 8 of the melting velocity and the instabilities which 9created basically an atomizing mechanism that is very 10 fine.11MEMBER WALLIS: Why does this high 12velocity, very hot jet, why does it get diverted, not 13 simply drill a hole?

14 MR. THEOFANOUS: Why does it do what?

15MEMBER WALLIS: The high velocity jets can 16 drill holes as well as get splashed. They can drill 17holes in things. Why doesn't it drill a hole right 18 through the base?

19MR. THEOFANOUS: Because like we're 20saying, we're protecting that with the refractory 21 material.22 MEMBER WALLIS: Oh. Well --

23MR. THEOFANOUS: Before you can actually, 24there's five meters of concrete on the floor and it's 25 186 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 on top --1MEMBER WALLIS: Five meters, that stuff is 2 five meters thick, that brown stuff.

3MR. THEOFANOUS: No, but it's all 4 connected together. It is sitting on the top of the 5--6MEMBER WALLIS: I would think it would 7 destroy some of your tubes.

8MR. THEOFANOUS: Well, you might think so, 9 but --10MEMBER WALLIS: Well, I'm just asking.

11 Does it destroy?

12 (Several speaking at once.)

13MR. THEOFANOUS: No, I'll just tell you 14 now. It doesn't destroy the tube.

15MEMBER WALLIS: You just told us it had 16tremendous force and all this stuff. So I would think 17 you --18MR. THEOFANOUS: Yes, we know exactly what 19 the force is. We know. We've --

20 MEMBER WALLIS: So you have analyzed the 21 survival of the tubes.

22MR. THEOFANOUS: Sure. In the report. In 23the report, you will find the stagnation pressures for 24 the --25 187 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: You will find all that 1 stuff. Okay.

2MR. THEOFANOUS: So now the problem we 3want to solve here is therefore steam coming out at 4high velocities, mixing up very intensely reducing 5very fine automatization of the melt and especially it 6is zirconia that's there and oxidating it and 7releasing the oxidation energy from the point of view 8of most gas coming out. It doesn't make much 9difference because you have one more hydrogen or more 10steam, the same thing so it's all there. We have 11another containment here. So it doesn't really matter 12 whether it's hydrogen or steam.

13However, there is extra energy that is in 14this initial oxidation and that heats up the gases and 15that's important. Before the gases go through this 16operational pool, the temperature is very important 17because that really generates the peak of the pressure 18and there you need to account correctly for all that.

19So we have them not coming out. Then 20steam after that. Automatization, oxidation fine 21 scale. The stuff is blown out into the space over 22here and there it separates some of the bigger pieces.

23They fall off. The velocities are very low by the way 24 out here, very low. But this volume is pressurizing 25 188 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433now quickly and will continue to pressurize like as if 1 it was a closed volume until the vent's clear. That 2 is a process that can take like a second. So it's 3very intricate and you want to calculate that 4 correctly.

5 Now I want to contrast that a little bit 6with what we did for issue resolution for pressurized 7water reactors. This was our probabilistic framework 8 that we used and you'll see here that there is a lot 9of detail like there is amassed how much O 2 you have 10or how much zirconia and how much was steel and then 11how you get pressurization as a result of these 12compositions and then something we call the coherence 13 (PH) ratio which has to do with how much of the steam 14 is in to see how much of the melt here in this process 15and all this was happening in the closed because there 16 was a lot of static containment. It was in a closed 17 volume. It couldn't go anywhere. So in fact, in this 18case, the dynamics were not so important. It was 19 what's important was the maximum pressure and that's 20why also Marty Pilch who worked with me together to do 21this serious problem. He used what's called a two 22 cell equilibrium model which basically does the same 23 thing that my model did except that one was just like 24an equilibrium thermal dynamics. So take that and put 25 189 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 it equal and put it equal.

1So here for this reactor, it's not enough 2to get the final pressure because you have an optimum 3volume pretty sure we can -- So we want to get the 4dynamics so the full -- is needed. So we use the 5 same. We call it convection limited containment 6 heating (CLCH). We use that same model but now in a 7full transient model. The model assumes basically 8 that the steam and the melt come to what -- the 9perimeter at some rate in which the melt is being 10carried out as fine particles to go out. So that 11defines the rates of contact, the rate of containment 12 and the steam going down. Okay?

13Then basically that's what we did and the 14 reason I put this up here because I want to show you 15 that the evaluation for PWRs, thanks to the presence 16of this suppression pool and the venting from that 17volume from the drywell to the wetwell actually is 18totally insensitive to essentially all that stuff. So 19you can assume the whole mass and even more, almost 20 anything you can do, you can not overpressurize this 21 area. 22 So what you've done here, that's new, is 23you've extended the model to make it transient and 24then we coupled to event clearing model and then each 25 190 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433one of those models were verified in the transient --

1 And here is to illustrate for you the 2facilities which were used, the data we used, came 3 from. This is IET series. It's called integral 4effect tests that were run in counterpart, two series 5(PH) SND at 1/10th scale. That's the South Sea 6facility at 1/10th scale and then at 1/20th scale, at 7I think it's called core exit facility -- used real 8materials and they used -- Pretty significant sized 9 experiments. That's what we used every time for very 10fine my model and Marty Pilch's model and we could get 11 done the job for the PWRs.

12 I'm using the same data here but now also 13paying a lot of attention to the transient itself and 14 I'll show you in a moment the results.

15MEMBER WALLIS: I have no idea what you're 16 modeling here.

17 MR. THEOFANOUS: I'm sorry.

18MEMBER WALLIS: I have no idea what you're 19 modeling. What is this supposed to be modeling?

20MR. THEOFANOUS: What the experiment is 21 modeling?22 MEMBER WALLIS: Yes.

23MR. THEOFANOUS: It is modeling the 24 process I described to you.

25 191 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER WALLIS: It's modeling the entire 1 containment with the venting and everything.

2 MR. THEOFANOUS: No. I said this is for 3 PWRs. Okay. This is for PWRs, pressurized water 4 reactors.5MEMBER WALLIS: So what is it modeling 6 then?7MR. THEOFANOUS: So it's modeling their 8containment heating processes in what is dry container 9 which means there is a reactor, there is a reactor 10 cavity, there is a --

11MEMBER WALLIS: You're squirting something 12 into this containment and --

13MR. THEOFANOUS: They are squirting 14 something into the cavity of a PWR and then you have 15a containment which is this one here which is like 16what is dry containment to find how much pressure you 17 get. In this case for PWR, you're also going to know 18 how much hydrogen you get and you're going to know 19whether that hydrogen is going to combust or not. So 20it was a real challenge here to find also the hydrogen 21produced and the combustible hydrogen because that 22 evolved into the final pressure here.

23Now in our case, we are interested in the 24-- And CLCH model was found to work as well actually 25 192 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433in telling the hydrogen and final pressure and 1 everything. Here we are interested in reactor 2 containments. So some of those tests were done with 3nitrogen only in the large volume here. So we used 4those obviously because those are the ones that are 5relevant for the present comparison. That's that.

6 The other one -- So that's for the DCH phenomenon 7 itself.8Over here, we have the vent clearing, what 9I was telling you before, and those are the PSTF 10 experiments. Those were actually done when I was a 11 little child. A long time ago. I remember those 12 tests. Actually I've been inside those facilities at 13 the time I was a consultant and we were looking over 14 those tests. I was sitting on the other side of the 15test and those are full scale actually. Those are 16 full scale and that full scale is the same full scale 17as we have in the ESBWR. Actually it's exactly the 18 same.19MEMBER DENNING: But that's the easy part 20 of the problem. Right? That's just acceleration of 21 the slug and it's just verified how long it takes to 22 accelerate.

23MR. THEOFANOUS: Yes. But you'd better do 24it though because -- So I'll show you in a moment.

25 193 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 It's a very interesting dynamic because of that. So 1what happens here you get a supply of vapor going into 2 some of the models of drywell. That pressurizes and 3that pushes through the down carbon and through the 4vents and pushes it right out and pressurizes this and 5 this. So that's the dynamics we're interested in 6 doing.7So I want to show the verification of the 8reports and they're coming together to do the full 9 DCH. Here is an example of the IET DCH test model 10 experiments. This is typical comparison with the vent 11 clearing. The vent clearing, you like to know you 12catch the peak and also the time of the clearing. In 13the report, you'll find more details about both of 14 those things. It's interesting to point out that in 15this test here and this is for the significant one 16 notice that we are much even in the long term and the 17reason is that there was such a big facility here the 18velocities were negligible. If you look at the Argun 19 test, you find that the experimental data, they show 20the decline even in times like this and the reason is 21you have heat losses which of course you don't care 22 about.23All right. So now the dynamics, there are 24three regimes that I've identified for quantifying the 25 194 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433log. So we just go on and run some calculations and 1show further pressure and velocity. We wouldn't 2 understand what's going on and what drives it. So 3what drives it here is that Regime I you hold 4hypothetical because it is a very humongous area of 5failure of the lower head. There is something around 6one meter in diameter all at once forming. If that 7was to occur, you'd have pressurization that is so 8strong because of DCH that actually the pressure in 9 the lower drywell which is this exceeds the pressure 10 in the upper drywell very significantly just because 11 of that restriction I was telling you about.

12 All right. So we form a structure first 13 and then it reaches a maximum, but then of course as 14 it goes to high enough pressures, it is able to vent 15 faster. So you have a decrease. Then it cuts --with 16it. So from that point on, the pressure is essentially 17made the same and then at that point, it finishes the 18 blowdown. At this point again, this cools and the 19wetwell gradually arises as the contents of the 20 drywell atmosphere vents into the water.

21MEMBER DENNING: Those are results of your 22 analytic tool?

23 MR. THEOFANOUS: Yes.

24MEMBER DENNING: And in this case you have 25 195 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 basically two separated volumes.

1 MR. THEOFANOUS: Yes.

2 MEMBER DENNING: Whereas in the PWR, you 3 had only one line.

4MR. THEOFANOUS: One and it wasn't even 5 passing. The PWR, it was like we only looked at the 6 final result.

7All right. So that's hypothetical.

8 That's Regime I. That's a very extreme regime, but 9even that one doesn't fail. By the way, you'll see in 10a moment that the containment is beginning to be 11 challenged around this pressure. So around 11 to 12 12 bar.13MEMBER DENNING: Is that what the 14 fragility is?

15MR. THEOFANOUS: Yes, that's the fragility 16for the drywell. I'll show you in a moment. Then the 17Regime I is if you took an extreme of the case we have 18used for a creep rupture in pressurized water reactor 19 the one we had during DCH and for that only, we used 200.55 meters in diameter. So we have given a 21probability distribution of the possible sizes. Where 22 11 narrow I was reminding myself the other day is a 23narrow distribution but the very upper outside end of 24 that was 0.55.

25 196 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 This calculation here was around 0.5. I 1wanted to be exactly literally correct when I said we 2 took the upper limits. I just had the calculation 3around yesterday of 0.55 and of course, that's a very 4 slight difference from that.

5So in effect, what the source here is that 6this initial difference in pressure between the lower 7and upper drywell is limited to a very short time.

8 Instead of a peak here, you get an inflection point.

9 They join together. Again there is a peak and there 10is some panning order, finally catches up here with 11the wetwell. Eventually from there, it goes out like 12 that.13 I want to tell you that it takes something 14like about 30 seconds, 40 seconds, to do the full 15blowdown, but the main part is during the time that 16 you put in the melt out and I'll show you how one is 17to do that and the shorter you make that melted 18 premium time the more big piles you make over here 19because it happened before the event is cleared. The 20longer you make the melted premium time the more 21you're spreading out the energy into the steam. So 22now from one point of view, that helps you oxidize 23 more. It helps you more contact, more energy comes 24out, but on the other hand, the cooling spreads it out 25 197 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433brings in this suppression pool after the clearing.

1So from one hand, that helps on one hand. On the 2 other, no matter what you do you can't get into -- So 3that is Regime II. That's the upper end of the 4category range for PWRs and those would be PWRs. The 5reason we did that by the way is because some PWRs 6have no penetration to the lower head. That means it 7will suffer from creep rupture and then we have to 8 know how big the area is. In this reactor as well as 9in some other PWRs where there are penetrations you 10essentially never expect to have a creep rupture 11 scenario.12 And then finally Regime III is the most 13 likely scene for a boiling water reactor and that is 14if one or more of the penetrations fail and it doesn't 15really matter whether it's one or two or three or four 16because if you fail more penetrations then the melt 17comes out sooner. It doesn't not bleed so much so 18that the final area is not so different from having 19one and you let it -- and in the process of melt 20coming through the hole is un plated, un plated, un 21plated, and eventually comes out to something like in 22this case about 30 centimeters. That is a huge hole.

23So therefore the relevant area for getting 24 steam out is 30 centimeters diameter hole. In that 25 198 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433case, Regime III as you see the dynamics are much more 1 benign and this is the steam that is falling out the 2melt. That is again a creep rupture-like scenario like 3 the one I showed you.

4 So here is the coverage. We've done 50, 5100, 300 cones. Three hundred cones is basically more 6than what you have there even if you accounted for 7 everything. The diameters, these are typical of creep 8 rupture. These are smaller. I'm sorry. Penetration 9failure. These are smaller than those because a 10smaller amount of mass therefore less oblation. Then 11 0.5 is for the creep rupture I was telling you.

12The temperature inside the vessel was 13taken to was taken to be 100, 150, 100. Actually, the 14higher the temperature is the less density of the 15steam inside and therefore the less potential for 16 oxidizing. So actually you make it more severe by 17using a lower temperature and that's why you use lower 18 temperature.

19 The t m needs some explanation. That is 20 the --21MEMBER WALLIS: What is the temperature of 22 the core area?

23 MR. THEOFANOUS: The core is around 2500 24 degrees, maybe higher.

25 199 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. THEOFANOUS: The t m is what you call 1the mixing time or melted varying time and that is we 2have this formulation for calculating that in the 3pressurized water reactor case and use the same one 4because basically there is steam and -- like things 5going out but we played what we call metrics and as I 6told you that's a matter of use. But typically for 7 these kinds of situations we have about seven to ten 8 seconds of time for this to come out.

9If you make this melt time, of course, 10given an area if blowdown of the steam is fixed by the 11area and the pressure so now if you said that I'm 12 going to take this t m to be very short, what that 13 means is that you're going to allow a lot of the steam 14 in the area up high during which the melt is coming 15out basically to be not useable because it -- and 16there's nothing more to oxidize. So by making that 17too short, you're making higher the defaults, you try 18to make higher the defaults, however you are -- That's 19 why we call it convection conductivity. It's really 20limited by that process. But we've done 3.6, ten 21 seconds, so a whole range of different choices here.

22 And here are the results. By the way an 23important parameter that we call the DCH scale 24 expresses that coherence between the melt coming out 25 200 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433and the steam coming out. Those are two 1characteristic times for that process. This ratio, 2it's very important because when that ratio is less 3than one, as I said before, the process is still 4limited and it's less than what it is the elastic 5pressure again. So for example in a dry container 6 even, if that's very small you get much more pressure 7 than compared to if it is but one. So actually in a 8dry container, if you plotted the pressure increase 9versus this coherent ratio you find out another steep 10increase up to one and up to one should have straight 11 up. 12 So what you see here is what we have and 13our cases are anywhere from as low as 0.104 to as high 14as 1.3. So we have covered the whole spectrum of 15likely contacts between the steam and the -- I should 16point out the pressures however. The first peak is 17very modest in relation to the fragility. The second 18peak is also very modest. As you're going to this, 19these are creep rupture scenarios you get about six 20 bar. Then the temperatures, I'll show you in a moment 21how the temperature does. It looks like it goes up 22and then it goes down and eventually settles in about 23 a minute settle to some value and that value is around 24 one thousand degrees.

25 201 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER DENNING: You're not showing that 1first kind that you first showed us because that's not 2 considered credible.

3MR. THEOFANOUS: I wanted to do that 4 because I wanted to provide a backdrop against which 5you can see where the conservative case is and then 6 the more likely case is. So we also did run outside 7of the report outside of Chapter 21 additional 8sensitivities about condensation and dust cooling, 9oxidation efficiency composition and drywell 10 atmosphere and so on and basically same results. So 11here then putting it together, here is the upper bound 12 is six bars I was showing you before, upper bound of 13the loading that you could have and it really doesn't 14decept the fragility. I don't want to get into any 15 games about saying so much of that is like this and so 16much of that this. Just I used here the complimentary 17cumulative distribution. So everything is below that.

18Then for the fragility which is as I said 19we got from another chapter of the VDOT report is 20initially here. You see that for the 50 percent 21values about 16 bars. This value here around 11 or 12 22is running two percent only. Over here is 10

-5. So 23 it's really just there's no intersection whatsoever.

24 So that's the story for DCH and I don't know if I want 25 202 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433to belabor that anymore. This is a conservative area 1as it is as in the PWR case the creep rupture is on 2 the upper limit of the size and upper bound of 3 available materials participating and no new section 4 at all.5The temperatures now, coming to 6temperatures, here is a typical behavior we see. We 7see a very high pressure pulse -- temperature pulse in 8the lower drywell. Of course, it makes sense because 9not only have you got 2,000 degrees in the melt but 10now you get the oxidation area and you have a 11tremendous energy machine for using there. So you can 12 reach another 1,000 degrees when you cover it.

13Now why does it go so steeply down is 14because after the melt gets out of there then 15basically it washes out with the cool steam and 16 hydrogen that come out from the vessel expanding and 17cooling down. So that's the issue I intend to show 18 now. But keep in mind this temperature on the 1,000 19degrees because that's really a benchmark against 20 which to say now if I have this for a few minutes on 21the upper drywell what will happen with the liner and 22the liner started sagging. Obviously if the liner had 23no anchors, you would see the liner sort of falling 24off by its own weight because it's really that one 25 203 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 that's stripping off.

1However, in the case of which we are 2covered so well by those anchors, the way that has to 3 be self supported for each, let's say, cell, that is 4so small that it just doesn't do anything. In fact 5 creep helps us because it helps relieve the stresses 6 so there is no cracking.

7MEMBER ARMIJO: Why doesn't it buckle and 8 pull away?

9MR. THEOFANOUS: No, it will do some 10 buckling. In fact, in the report, I didn't know if I 11 had time here, but in the report you'll find pictures 12in which you see the full buckle. It makes like a 13wavy structure and that's why I'm saying it helps you 14because it can creep without peering, without creating 15cracks because of the high temperature. And then 16 again to mention the lips again on the -- We make no 17claim by the way for wire integrity in the lower 18drywell not in light of these temperatures I wouldn't 19and not in light of the fact that there's all kinds of 20 melts flashing all over the place.

21Ex-Vessel Explosions and BiMAC pipe 22 crushing and the pedestal failure, what we're saying 23 here is that we are saying that if we had a deep pool 24 and if we had pools of melt that are tens per second 25 204 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433which can't be excluded, you know, people usually will 1 use and I've seen people use tens of kilograms per 2 second and rarely you see hundreds of kilograms per 3second, but this is very heavy material. If you have 4 a core there, who is going to tell you you're not 5 going to get a few hundreds at least.

6So we used the 700 kilograms per second in 7our calculations and we found that in doing the 8impulses on the form that it can be significant. With 9these kinds of pools, we find that because the 10pedestal is quite far away and because especially 11 shower pools can vent (PH) the energy we find that the 12 impulses are rather low.

13The impulses by the way are the figure 14here because these are millisecond scale pressure 15 pulses which show the detail of the pulse, the detail 16of the pressure transient, is not important but rather 17the integral on the code. So using impulses to measure 18 explosive release energy and then we use the impulse 19 to measure fragility.

20MEMBER WALLIS: Now in the PRAs it says 21that the probability of an EVE is zero for depths less 22than 0.7 meters. Then it becomes one when you get up 23 to 1.5.24 MR. THEOFANOUS: Where are you now?

25 205 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: I'm reading Section 8.3-4.

1 The notes from the PRA that we're reviewing here.

2 MR. THEOFANOUS: From the PRA.

3MEMBER WALLIS: It says that the 4probability of an EVE is insignificant for water 5 levels less than -- Is this something you did or 6 something they did?

7 MR. THEOFANOUS: The rupture scenario --

8MEMBER WALLIS: It says that when water up 9 to here you have no EVE and when you have water up to 10 here, it's a probability of one.

11 MR. THEOFANOUS: He did that.

12 MEMBER WALLIS: I'm just wondering. Can 13you really predict with that precision that nothing 14 will happen when it's up to here and it's inevitable 15when it's up to here? Can you really predict with 16 that precision?

17MR. WACHOWIAK: This is Rick Wachowiak 18from General Electric. That's a calculational tool if 19 you will. What we're saying is when it's --

20MEMBER WALLIS: It's not a modeling of the 21 physics.22MR. WACHOWIAK: When it's below, what Theo 23is going to show you in a minute is when it's below 24 the lower threshold there is no way that we're going 25 206 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 to have a steam explosion that's going to affect any 1of the structures or any of the equipment. When we 2get to the deeper subcooled pools what he's saying is 3 that we can't rule out that there may be some damage.

4 So when we did the calculation --

5 MEMBER WALLIS: You took it as one.

6MR. WACHOWIAK: -- we said when it's high 7 we assume that. We'll just take the worst case.

8MR. THEOFANOUS: No, he's not asking that.

9 MEMBER WALLIS: So he has to --

10MR. THEOFANOUS: He's asking how do you 11 know that what fraction of scenarios are for shallow 12 pools, what fractions are for --

13 MEMBER WALLIS: I'm also asking how well 14can you really say that it's zero for a certain height 15 and then it suddenly becomes one.

16MR. WACHOWIAK: And what I think he's 17 going to show you is that even with the one meter or 18two meter that it really shouldn't be one. It should 19 be --20MEMBER WALLIS: It's very unlikely. He's 21 going to show it.

22 MR. WACHOWIAK: -- some small fraction.

23MEMBER WALLIS: So we have to listen to 24 them all.

25 207 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. WACHOWIAK: Yes.

1 MR. THEOFANOUS: Unfortunately.

2 MEMBER WALLIS: Okay. That's all right.

3 You'll get to it.

4MR. WACHOWIAK: It's just a calculational 5 tool that we use.

6MR. THEOFANOUS: I thought you were asking 7 about the fraction of scenarios that --

8MEMBER WALLIS: No, I was asking about the 9 probability of an EVE depending on pool depth.

10 MR. THEOFANOUS: Oh. Then let's go on.

11MEMBER WALLIS: You're going to get to 12 that? Okay.

13MR. THEOFANOUS: I'll get to that. All 14 right. So what you said is that there was a --

15prohibiting information of such pool but design 16 changes -- they really are. So --

17 MEMBER WALLIS: So you mustn't switch it 18 on too soon.

19 MR. THEOFANOUS: Yes. As usual.

20 MEMBER WALLIS: All right.

21MR. THEOFANOUS: I don't have it here, but 22 I put that --

23MEMBER WALLIS: So more water isn't 24 necessarily better. It could be worse.

25 208 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. THEOFANOUS: Of course. That's why I 1 don't want to have water there.

2MEMBER WALLIS: Not yet. Not until you 3 need it.4 MR. THEOFANOUS: And I don't need the 5water there when the -- in the reactor vessel. Now 6what we mean by prohibiting, you have to prohibit 7 that, just make it less likely for having water there, 8that means that there was a GDCS overflow for example 9if we had let it when revising the original design it 10would basically almost virtually guarantee you're 11 going to lots of water down there.

12There was another one that would allow 13overflow the suppression pool which again would almost 14guarantee that you're going to get into some flooding 15 situation. That was taken care of. So that's what we 16mean by containment layouts and systems and then in 17addition to that as I explained already in the case of 18BiMAC we want to make sure that we require the 19reliability of, I don't know, the reliability of 10

-3 20for failing to supply the water when needed and the 21same reliability of 10

-3 for supplying the water too 22 early. So that's a systems question. So we are going 23to get down to a shouting match about how we're going 24 to assure this 10

-3 but that's a systems question 25 209 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433which we believe is more properly a problem for the 1 COS stage in the license.

2According to bounding estimates and 3impulses the conclusion is here. Fragility is the 4additional margin even for subpooling. So here the 5 real picture, that's the basemat. There is a BiMAC, 6basically a concrete structure with pipes and with 7 some cover of -- on the top of it and there on the 8floor, there is the grating and these are the two and 9a half meters thick pedestal wall. These are the 10hatches I mentioned before. So if you want to know 11why for example we keep that value for about two 12 meters it will be lineated again what we foresee the 13deep pool or fire pool it's because if it was more 14than about two meters above this floor it would be 15exposing the hatch door to the explosion. So the 16 issue here is one in which we have differing levels 17 of water up here and that comes out at about ten (PH) 18 ton per second.

19 What kind of pulses here can I get here 20and at the BiMAR? That's the question. Then will the 21structure survive this pulse? So already I mentioned 22the release rate and we did calculations for the one, 23 two and five meter deep pools. We considered such a 24rate in subpool water and what we're finding is at the 25 210 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433floor it's about 100 kilo Pascal seconds pulse up here 1if some cooled pool. If we have saturated water pool, 2we do nothing. And then, for the side walls because of 3the distance and because of the venting you get about 440 to 50 kilowatts Pascal seconds, but also in the 5 fragility mode.

6 Now this is new. This is -- that these 7actually DYNA3D I think it's called -- is the 8 commercial version which is operating for commercial 9 purposes. This is something that's used for national 10security issues and of course is exercised a lot with 11high explosives. Now high explosives may give you 12assorted pulses, however, most of these is for 13cracking purposes. However, for our purposes here, 14one or two millisecond pressure pulses are also pretty 15 steep. So we believe that's very appropriate in terms 16of the natural frequency structures. So that would be 17 if there's a real disaster these days.

18And I referenced in the Chapter 21, I 19referenced a rather extensive document when this part 20was published from Livermore and we tried a lot of 21 compiled data of this --

22MEMBER WALLIS: There was just a one shoot 23bang or does it bang and then bang again? When you're 24pouring the stuff into the pool, you have an explosion 25 211 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433and everything comes together again. You're still 1 putting stuff in. Does it explode again? Several 2 times?3MR. THEOFANOUS: Yes. Sure. That is one 4 of the issues that arises if you have very deep pools.

5 MEMBER WALLIS: Right.

6 MR. THEOFANOUS: That's why I don't want 7to say much about very deep pools. That's why I tried 8 to stay away from deep pools because if I have a one 9meter pool and I have an explosion in there and the 10 water goes all over the place, you're not going to 11 have a pool anymore.

12 MEMBER WALLIS: Well, it falls back down 13 again.14 MR. THEOFANOUS: Yeah, but how long will 15 it take for the water pressure for the --

16So the calculations actually were very 17detailed with millions of notes and a very detailed 18representation of the -- By the way, those are 19symmetry planes and that means in a symmetry plane the 20 thing is not allowed to move normally, but it's free 21 to move this way and a very detailed presentation of 22 all the rebar, the concrete, the -- bar, the -- bar, 23the mercury bar, everything is there in these 24 calculations.

25 212 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433And here is actually a very interesting 1 movies to show how the -- Of course, that's highly 2 exaggerated. This is the 600 kilos Pascal second 3welding (PH) that we put into this loading as well and 4what happens in this case is you begin to have -- This 5 is illustrated here by the yielding of the rebar and 6by the crashing of the concrete which is shown by this 7red area. So basically in the area where the concrete 8 is, you crack it and the rebar yields and you have 9 failing. It takes that kind of energy to be put in to 10 create failure.

11This is represented schematically here.

12We've done calculations with the obstacles here, here 13and here and there was no failure. Over here is what 14you just saw, some failure. So I just draw just 15 schematically. That's why this dotted line, some kind 16of a cumulative salable probability that starts 17 arising between here and here.

18As I mentioned before, what was wrong 19before about failures of those structures was actually 20 a paper that I did many years ago. At that time, we 21 considered one and a half meters concrete with rebar 22and that was failing right around here, at around 100, 23150 kilos Pascal seconds. Because of this paper, I 24think most people, you go out there and you ask people 25 213 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433that know about this problem how much does it take to 1 fail the model within 150 kilo Pascal seconds. So 2actually we were very pleasantly, we were anticipating 3some increase but that is a very significant decrease 4 in fragility because of the size and the concrete.

5Nothing special on the concrete by the 6 way. This is just a normal 5,000 psi concrete. You 7 can get that. If you have 10,000 psi concrete, it's 8 going to be even better.

9 MEMBER DENNING: On the DYNA3D when you 10 run that analysis, do you actually put in, you don't 11put in just the kilo Pascal seconds. You put in a 12 certain --

13 MR. THEOFANOUS: The pulse, yes.

14 MEMBER DENNING: The pulse, right.

15MR. THEOFANOUS: Any report you'll see all 16 kinds of pulses. For example, it will be like twice 17the maximum pressure, half of the width of the pulse.

18You'll see a pulse here in the report. So what you 19are showing here then is that for the pedestal in the 20report you will find a number of compilations that 21 will show you get in the report only about 100 kilo 22Pascal seconds. So it's a huge margin. I believe 23 when we have pools like that, one, two meter pools you 24 cannot fail the pedestal by a --

25 214 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Because of what Graham was saying before, 1 I don't want to say to defend an eight meter pool and 2what happens with that. So therefore we decided we 3 don't want to have pools like that and we managed to 4 do this by not flooding into --

5 MEMBER WALLIS: Theo, you have this pool 6and you have an explosion in it. Is the explosion in 7 the middle of the pool or is it near the wall?

8 Doesn't it make a difference where it is?

9 MR. THEOFANOUS: Of course.

10MEMBER WALLIS: Because it attenuates 11 there.12MR. THEOFANOUS: Yes. Of course it makes 13 some difference.

14MEMBER WALLIS: So you can blow a hole in 15 one side of it or near that side.

16MR. THEOFANOUS: No, actually what we have 17done here to account for that kind of thing here 18similarly we have proceeded, if you look at the report 19again, you'll find that the radial, actually symmetric 20operation basically with the diameter of ten meters to 21 a diameter of only about four meters. So that means 22we put the explosion close enough to the wall as if it 23was coming from the edge of the reactor vessel and 24 what you would see again is sort of very conservative 25 215 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433but it gives you an idea that what we've done is a 1 conservative picture.

2MEMBER WALLIS: But the stuff that is 3 coming out, you said it's in a jet, a high velocity 4 jet.5 MR. THEOFANOUS: It's in a jet.

6MEMBER WALLIS: So it could go way off to 7one side and it could actually go very, very close to 8 the pedestal wall, couldn't it, before --

9 MEMBER SIEBER: No.

10MR. THEOFANOUS: No, it couldn't do that.

11There is no reason to do that because that stuff is 12 heavy and it's not --

13MR. THEOFANOUS: But it's driven by its 14 own gravity. It's not at high pressure.

15MEMBER WALLIS: So it's not high pressure 16 anymore.17 MR. THEOFANOUS: No.

18 MEMBER WALLIS: That's a low one.

19MEMBER DENNING: That's a different 20 scenario.21MR. THEOFANOUS: It's a different scenario 22 now.23 (Several speaking at once.)

24MEMBER WALLIS: So it's just oozing out 25 216 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 and falling out.

1 MR. THEOFANOUS: Yes.

2MEMBER DENNING: This is low pressure 3 scenario.4MR. THEOFANOUS: It's a different 5 accident.6MEMBER WALLIS: This is low pressure 7 scenario. Okay.

8MR. THEOFANOUS: It's a different 9 accident. We started high pressure scenarios for 10those that we do DCH. Low pressure scenarios the 11 issue is not DCH but these explosives.

12MEMBER WALLIS: And there's nothing in 13 between that could be both?

14MR. THEOFANOUS: There is nothing in 15 between unfortunately.

16 MEMBER WALLIS: Okay.

17MEMBER SHACK: It's estimated to 90 18 percent.19MR. WACHOWIAK: This is Rick Wachowiak 20 again. There's not any way really to get water in the 21lower drywell in the high pressure scenario. So 22 that's the main reason why we don't have to consider 23the combined effect. There's just no high pressure 24 scenarios we can find where --

25 217 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: You can't drain the pool.

1 MR. THEOFANOUS: What?

2MEMBER WALLIS: You can't drain the pool.

3 Is that it?

4 MR. THEOFANOUS: No. All right. Next.

5CHAIRMAN APOSTOLAKIS: Wait, wait.

6Comrade Theofanous. Is this a good time to take a 7 break?8MR. THEOFANOUS: Excellent time because we 9 are changing subjects.

10 MEMBER WALLIS: He presents the stats.

11CHAIRMAN APOSTOLAKIS: One other question.

12There's a lot of slides in your handout. Are these 13part of the severe accident mitigation or do they 14 include the containment systems performance?

15 MR. WACHOWIAK: They do not.

16 MEMBER DENNING: They do not.

17CHAIRMAN APOSTOLAKIS: Well, there was an 18 hour2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> and a half.

19MEMBER DENNING: Are you sure you want to 20 take a break at this time?

21MR. THEOFANOUS: Yes. There is enough for 22 two and a half hours according to the agenda.

23CHAIRMAN APOSTOLAKIS: You had 12:30 p.m.

24 to 3:00 p.m. Yes, you're right.

25 218 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. THEOFANOUS: That's two and a half 1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. Right?

2CHAIRMAN APOSTOLAKIS: Two and a half 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />. So let's take a break.

4 MR. THEOFANOUS: We are midway now.

5CHAIRMAN APOSTOLAKIS: We'll be back at 6 2:45 p.m. Off the record.

7(Whereupon, the foregoing matter went off 8 the record at 2:33 p.m. and went back on the record at 9 2:51 p.m.)

10CHAIRMAN APOSTOLAKIS: Back on the record.

11MR. THEOFANOUS: So we are in to steam 12 explosions. We're now going to look at the BiMAC 13 itself. From a structural point of view, the BiMAC is 14supported by concrete which itself is similarly top of 15 basemat. The pipes are schedule 80 pipes. That means 16one centimeter. We would have pretty significant 17 figures basically for structural purposes. They are 1810 centimeters in diameter and embedded into this 19 sacrificial layer which is like 27 meters.

20Now the question initially is if you have 21an explosion here what does it take to crash those 22 pipes. Obviously, if we are sitting and those pipes 23are right below the explosion and there is enough 24impulse to crash them, then at least in that location 25 219 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 you're not going to be able to shoot the water that 1 you need in order to prevent melt completing.

2I do want to make a general remark and put 3this into perspective. We have like 100 square meters 4floor area. You have an explosion that is sitting 5someplace with impulses right under it, the localized 6 impulse. So that's going to be like being hit by a 7 truck. Actually, I don't think it's going to mean 8very much for the whole function of the device, but I 9still nevertheless would like to know what an 10 explosion will do to those pipes.

11Again, analyze them with DYNA3D and see 12 that they support to each other. We found planes of 13symmetry so that we could analyze this for extreme 14detail, representing both the pipe, the wall thickness 15and the concrete above and below it. The results 16tells us how the quality of the metal yields and 17whether the concrete cracks. This was for 220 kilo 18Pascal per second welding and you see a significant 19 crack in the concrete.

20I do want to say that this cracking of the 21material which is especially important for high 22pressure material itself, I mean the material is 23important for basically resisting any oblation in the 24 pipes after you pour in the crack a little bit it 25 220 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 carries over.

1MEMBER WALLIS: So the pipe's intact, but 2the concrete is cracked. Is that what you're saying?

3MR. THEOFANOUS: I didn't say anything 4 yet. 5 MEMBER WALLIS: Well, I know.

6 MR. THEOFANOUS: I haven't said anything 7 yet. I'm trying to put into perspective.

8MEMBER WALLIS: I'm just trying to 9interpret your last figure, the last figure you showed 10 us.11MR. THEOFANOUS: Oh, the last figure, that 12 last figure was to show --

13MEMBER WALLIS: You said that concrete was 14 cracked. Now what about the pipe? Is the pipe okay?

15MR. THEOFANOUS: For this kind of a 16loading the pipe is some narrow -- oh, I understand 17 your question. I beg your pardon. In some location 18where the pipe is incorporated with the other pipe, 19that is in those similar things, they begin to yield.

20 You take that --

21MEMBER WALLIS: But it's intact. It's 22 intact.23MR. THEOFANOUS: It's intact, yes. But we 24take that to be the beginning of failure of the pipes.

25 221 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Once it starts yielding significantly even in a narrow 1 area, then that's the beginning of the crashing.

2So if we put these results in a 3probability plot again, for 99 percent of the 4scenarios we have essentially no explosions. So that 5is covering for most of it. This is for what we call 6 the low level.

7For the one to two meter levels, our 8results show that you can have a hundred. You can 9 have even more, maybe up to about 150 kilo Pascal per 10 second. So that shows schematically here that there 11is some distribution that we don't know what it is but 12that is what is shown on the dotted line. And also 13it's shown here that somewhere around 200 kilo Pascal 14seconds or maybe about that you begin to get 15 significant yielding of the pipe.

16 So that's why then the CFP starts rising 17 over here and the whole intent of this is to show that 18for the scenarios that we played we have integrity.

19But there is just no comparison, not even anywhere 20 near. The purpose of that is show that even if by 21 chance you had some small depth like one or two 22meters, you could begin to interfere with the 23 integrity of the pipes.

24 MEMBER KRESS: What steam exposed to the 25 222 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 model would be used to get this dots?

1MR. THEOFANOUS: We used PM alpha which we 2used before and to me that is the state of the art.

3The way it works is you get the melt into the water 4 and the PM alpha which is the mixing core.

5 MEMBER KRESS: Premixing.

6 MR. THEOFANOUS: The Premixing core, the 7PM alpha, it basically tell you what are the possible 8ranges of special space and time distribution so if 9melt fractions and steam void fraction. Then we take 10 that --11 MEMBER KRESS: That takes care of --

12MR. THEOFANOUS: Oh, if you like, we have 13it in the back of the report. We have the whole 14 evaluation basis of those cores.

15MEMBER KRESS: Just one question. What 16sort of triggering do you have there? Does it trigger 17--18MR. THEOFANOUS: We use significant 19 triggering. Significant triggering means that once 20 you get the premixture we can put a trigger in.

21MEMBER KRESS: The trigger time occurs 22 after you get this premixing volume?

23MR. THEOFANOUS: Yes. Right. Any time 24 you have premixing. In other words, anytime --

25 223 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER KRESS: You can trigger any time.

1 MR. THEOFANOUS: Right. So the equation 2you take is -- since you don't know. Triggering is a 3kind of a spontaneous event that you don't know how to 4 predict it.

5 MEMBER KRESS: Yeah.

6 MR. THEOFANOUS: So you are saying that 7for all the evolutions with the premixture we are 8looking for cases where and you know how to drive the 9quality of the premixture from the point of 10 explosivity. So we're finding the worse premixtures.

11 The way you create a trigger is by taking one cell 12 and mixing the fuel that's there with the water very 13 rapidly. That creates a pulse.

14 MEMBER KRESS: And that expands.

15MR. THEOFANOUS: And that expands and then 16this calculation is done with M which is an explosion 17 point which also we have fully documented and viewed 18it and all that and I have in an appendix to the 19Chapter 21 you will find all the verification basis 20for the PM alpha but because this was done extensively 21before I didn't want to bore you with that stuff. So 22 I didn't include it here.

23MEMBER KRESS: Some of the members have 24 had the privilege of hearing that before.

25 224 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. THEOFANOUS: Yes.

1MEMBER KRESS: Another question I have is 2 you're pouring at a certain rate.

3 MR. THEOFANOUS: High rate, yeah.

4MEMBER KRESS: The very high rate. Can 5 you delay your trigger until you get it all in?

6MR. THEOFANOUS: Yes, we can, but at that 7 point what happens is --

8MEMBER KRESS: You have too much melt for 9 the water.

10MR. THEOFANOUS: Yes, exactly. We are 11 getting into the physics now of the explosion. What 12happens here is if we have too much melt and we don't 13 have enough water then the melt --

14MEMBER KRESS: So somewhere in there 15 there's --

16MR. THEOFANOUS: That's what I was saying 17 before.18MEMBER KRESS: Okay. Now I understand 19 what you're referring to. Thank you.

20MR. THEOFANOUS: All right. So I think 21now we are switching to the last topic which is the 22basemat melt penetration and this is to illustrate the 23scope of the work and what's all the different loading 24mechanisms that we have and the different criteria 25 225 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433that we have to consider and then we have to challenge 1 if it has integrity or no integrity.

2 What you see here is there is a thermal 3 loading on the jet impingement. All right. So this 4is we have local peaking right here because of the 5oblation depth and you'll find an extensive discussion 6of that in the report. So I'm not going to go through 7 that here. It's just to show you that you just can 8kind of impact that layer. And that's why we went 9 into a refractory material so that we can be pretty 10 sure that we're going to pack it.

11 The second item has to do with thermally 12loading from imagine now we have this -- which is full 13of melt and it is a natural circulation and now that's 14 going to produce a thermal loading to the bottom and 15 to the sides and now we want to show that this thermal 16loading would be possible to be accommodated by 17loading on the other side that so that it will over 18 here. If this is categorized by decay heat flux, this 19is a local criteria and this job here is done by 20taking into account any possibility of the local 21 peaking of the heat flux.

22MEMBER WALLIS: So are you doing a thermal 23 shock analysis of this stuff?

24 MR. THEOFANOUS: Thermal shock?

25 226 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: Yes, the sudden thermal 1 loading.2MR. THEOFANOUS: It's not the sudden 3thermal loading. Yes, we can discuss it if you want.

4MEMBER WALLIS: The sudden thermal --

5 6MR. THEOFANOUS: But it's not a sudden 7thermal loading. First of all, even if it was, it 8would have no impact in this kind of situation. So 9 two answers. Do you want me to elaborate?

10MEMBER WALLIS: So you have or have not 11 done a thermal shock analysis.

12 MR. THEOFANOUS: Huh?

13MEMBER WALLIS: I'm just trying to find 14 out if you did a thermal shock analysis.

15MR. THEOFANOUS: No, I didn't do a thermal 16 analysis. I think it's irrelevant to this problem of 17 thermal shock analysis. So if you disagree with me, 18 we can discuss it.

19 The point I'm trying to make is that this 20 evaluation involves local peaking. So it's not 21 sufficient to say I have this on the floor or I have 22 anomalous heat flux. My heat flux is less than this 23 average. That's not good. You have to make sure that 24watery you always are below. The water is always 25 227 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433below the heat flux. So that is a more sticky 1evaluation because you're looking for all the peaking 2of the flux and not so conventionally know that there 3can be all kinds of distributions. So we need to get 4 to those distributions.

5The second topic however has to do with 6the possibility in the pipe of basically depleting of 7water as it's boiling out. So that's what defines the 8 size of the pipe. That defines in fact this 9consideration and this consideration you find the 10 size. You can see very easy, in fact, these are very 11small pipes which in some ways would be desirable from 12 a structure integrity point of view because they are 13 kind of small and have very, very thick walls.

14Basically it would be indestructible. But if I did 15that then I would be susceptible to both this and 16 this. So that's why the ball park stands in the 17middle because we say we want to optimize that because 18we were doing testing on that for the COL and we want 19 to optimize the test.

20 But this one here, that has to do with 21 depletion of the water. It doesn't care if the 22profile is like this or like that. It really cares 23about the total thermal power it's putting on the 24 pipe. Of course with that sensitivity, it also 25 228 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 demonstrates it doesn't go back the shape.

1So two things you're looking for there in 2 this. You're looking actually not two, but three 3 things. We're looking for critical heat fluxes down 4 from the horizontal pipe and on the vertical segment 5No. 1. No. 2 we're looking for the average like a 6bounding average heat flux I can have again on the 7 horizontal and on the vertical and this is for this 8problem and then I'm also looking for the local 9peaking that they have and this is for that. Then in 10addition, of course we have the explosions which we 11just talked about. So those are the topics I want to 12 cover now.

13Again, the same picture as before, but now 14a little bit more detail, I think I'm going to give 15you more detail about how this thing looks. So now if 16 I look at it from the top, this is what I was telling 17you before. As you take slices this way, this pipe 18gets shorter and they get longer. Okay? And we have 19a main distributor here that the distributor is sized 20and the downcomers are sized. The downcomers are 21distinct because they are sized in a way that they 22will provide no significant frictional resistance 23compared to the frictional resistance of the two-phase 24flow over here. So there is no starvation of the 25 229 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 flow.1MEMBER ARMIJO: I guess I don't understand 2 that drawing. Are they all pipes or is this --

3 MEMBER SIEBER: Yes.

4 MEMBER ARMIJO: So there's pipe that --

5 MEMBER SIEBER: It just goes everywhere.

6 MR. THEOFANOUS: Yes.

7 (Several speaking at once.)

8 MR. THEOFANOUS: Maybe too --

9 MEMBER ARMIJO: -- How did that work?

10 MR. THEOFANOUS: And also presenting the 11sumps which are by the way not always very well 12protected in the plants, in previous plants. We want 13 to protect the sumps too and the sumps are important 14to have there for operational purposes. But you don't 15 want to have bypass of the BiMAC by getting the melt 16 from here to here and then going out into -- this 17would be a tremendous for the point of view into 18basemat because you lose a lot of the concrete. So in 19 the two near the edge there, we then worked with the 20 people in the design and made them to be hiding the 21wall as much as possible. Hiding the wall means 22increase this dimension, decrease this dimension so 23they can be covered just like the wall of the pedestal 24 by the pipes.

25 230 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: Can we go back again? The 1sequence of events, when do you turn on the water for 2 this and when does the hot melt come out and impinge 3 on this?4MR. THEOFANOUS: Well, we wait until melt 5 comes out.

6MEMBER WALLIS: So it's not water when the 7 melt first comes out?

8MR. THEOFANOUS: There is no water from 9 when the first melts come out. The moment the first 10 melt comes out the water is initiated.

11 MEMBER WALLIS: The moment --

12MR. THEOFANOUS: We don't want water 13 there. We have these pipes, the downcome is from the 14 GDCS.15MEMBER WALLIS: So the initial thing is 16just to heat up of the refractory by the melt. Is 17 that what's going on?

18MR. THEOFANOUS: Yes. We are making these 19 pipes to be large enough so that when they open they 20will flood this pretty quickly. So you don't want to 21 really have the water starting earlier than before.

22 MEMBER SIEBER: A quick question. Maybe 23 you could, if you melted the entire core and some 24 surrounding structures --

25 231 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. THEOFANOUS: Yes.

1 MEMBER SIEBER: What level of melt would 2 you get in that sump?

3 MR. THEOFANOUS: The next slide.

4 MEMBER SIEBER: Well, I'm looking at the 5 slides.6 MR. THEOFANOUS: Maybe the next to that.

7MEMBER WALLIS: The next after the table.

8MR. THEOFANOUS: It would be better to 9 show you in numbers rather than give you --

10 MEMBER SIEBER: Well --

11 MR. THEOFANOUS: Just for example either 12you can wait or you don't hardly wait. So we're going 13 to come to the next one. What is this here is BiMAC 14as a fraction of melt pool height resulting in average 15heat flux. That's the question you just asked.

16 Correct?17 MEMBER SIEBER: Yes.

18MR. THEOFANOUS: Okay. So here now we 19 have a table that says here is the height of the melt 20 and this is in meters 0.2, 0.4, 0.6, 0.8, all in 21 meters. That's the volume now of the melt. We are 22converting that volume with the typical density of two 23tons and then you can see therefore that a typical 24whole pool with floating melt in it would be about 300 25 232 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 tons. So what you see here is that you have such an 1 amount of melt in the BiMAC, it would be somewhere 2between 0.8 and one meter of height the melt would be 3 in there.4MEMBER SIEBER: And where does that place 5it on the side wall? Does it get to the side wall or 6 go back to --

7MR. THEOFANOUS: That would be all the 8 space that would be inside of that --

9MEMBER SIEBER: Up to where the point is?

10 MEMBER WALLIS: Where is a meter on that 11 map?12MR. THEOFANOUS: All the way, it would be 13 essentially I think up to about here.

14 MEMBER SIEBER: Okay.

15MR. THEOFANOUS: Now an important point to 16make here is that remember we're in a low pressure 17 scenario. That means the melt that comes out first 18 would be the melt that is molten at the time and 19 suddenly you would not wait until 100 percent of the 20melt melts before it fails. It will come out some 21time before. It will be a fraction of this 300 points 22 that it comes out. So that one is going to come out 23 as one lump in a way.

24 MEMBER SIEBER: Right.

25 233 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. THEOFANOUS: Any material that comes 1 after out is going to be coming out at the rate in 2which is melting which is going to be dribbling down.

3MEMBER SIEBER: Yes, a dribble. But it 4 could start that way too.

5 MR. THEOFANOUS: It could also.

6MEMBER SIEBER: Also you'll lose the 7control drive of mechanism penetration and you'll 8 dribble out and then all of a sudden the bottom will 9 come out and you'll dump a load and then from then on 10 it's dribbling out.

11MR. THEOFANOUS: Exactly. So then as far 12as heat fluxes the important thing to remember is that 13not all material comes together as a melt. So that 14comes out and you have this and you fill up to some 15 time. Now additional material that is dribbling is 16 going to see a water pool, a cold water pool, and it's 17going to solidify and it's going to solidify there and 18it's going to make debris then which however is a 19 fraction of this 300 tons which will not participate 20in the energy balance of the melt that is loading the 21BiMAC to the bottom because the BiMAC can be loaded 22downwards only by the melt, not by the debris that is 23 cooled.24MEMBER DENNING: But potentially it may or 25 234 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 may not cooled the debris bed.

1MR. THEOFANOUS: If it is not cool that's 2all right. But you know my own -- the significant 3fraction of it is going to be somewhere and it's going 4to be coolable because there's no reason for it to 5remelt because it is all cool from the bottom anyway.

6 So it's not really into a dry quicker but it's a wet 7 quicker if you like.

8MEMBER SIEBER: The part that comes out as 9 a lump --10MR. THEOFANOUS: No. We said it can be a 11 pool.12MEMBER SIEBER: A pool. It's going to be 13 still molten while this other stuff is solidified.

14 MR. THEOFANOUS: That's right.

15MEMBER SIEBER: And it's going to be very 16 difficult to remove heat from this molten pool in my 17view compared to what it would be. The stuff that 18 dribbles and drips down, that's pretty easy.

19MR. THEOFANOUS: Well, exactly. That's 20why you're putting BiMAC there because if it was easy 21 to remove the heat, then we wouldn't need to put the 22 BiMAC.23 MEMBER SIEBER: Even with BiMAC --

24MR. THEOFANOUS: Well, then you have to 25 235 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 tell me in a minute how you're going to fail the 1 BiMAC. That's where we're going through in the 2 analysis.3 MEMBER SIEBER: Right.

4MR. THEOFANOUS: You could always 5legislate of course that it will fail, but I think the 6 idea here is to --

7MEMBER WALLIS: After a while what happens 8 this GDCS pool keeps pouring water into this thing?

9MR. THEOFANOUS: That's what's going to 10 happen, the emptying is going to stop and then you 11 have natural convection.

12 MEMBER WALLIS: But then you have no 13 cooling underneath.

14 MR. THEOFANOUS: No cooling where?

15 MEMBER WALLIS: No flow in the pipes 16 anymore.17MR. THEOFANOUS: Natural convection 18 because in the pipes --

19MEMBER SIEBER: They don't crush the 20 pipes.21MR. THEOFANOUS: -- are in the water pool.

22 So the water would be coming through the pipes.

23MEMBER WALLIS: Oh, so it keeps on running 24 itself.25 236 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. THEOFANOUS: Of course. If it didn't 1 cool down, it wouldn't do any good.

2MEMBER WALLIS: Well, I was just wondering 3 about that.

4MR. THEOFANOUS: Of course. Okay. So we 5 have then the torus here and then taking the --

6MEMBER WALLIS: Doesn't entrain stop when 7it goes and recycles around? Is it pure water that 8 goes around? Is there junk in the water?

9 MEMBER SIEBER: There will be sooner or 10 later.11MR. THEOFANOUS: This is natural 12 convection. It's not forced pumping.

13 MEMBER WALLIS: No.

14MR. THEOFANOUS: -- sump there is suction.

15 MEMBER WALLIS: The water is on the pool 16 sitting onto of the molten core.

17 MR. THEOFANOUS: Yes.

18MEMBER WALLIS: And there's nothing going 19 on that is putting stuff into the water. It always 20seems to be so placid just sitting there being cooled.

21MR. THEOFANOUS: Yeah. Then we have --

22 these areas --

23MEMBER WALLIS: And you call it cert city.

24 A cert city is not very placid, is it?

25 237 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. THEOFANOUS: Who?

1 MEMBER WALLIS: This Icelandic lava pool 2that goes into sea. Isn't that cert city? You called 3 one of your --

4 MR. THEOFANOUS: That's very placid.

5 MR. WACHOWIAK: That was an experiment.

6 MR. THEOFANOUS: Let's look at the steam 7explosion for a while. That is all but placid. So we 8take the areas and from the material that's there and 9 from the decay power and the decay power can either be 10tacked to the material at the time we essentially have 11all the core, but in respect to the total and then it 12doesn't change anymore. So you see here the decay 13power increases because the material increases and in 14here it reached already all the core, all the fuel.

15So there is no more than whatever decay here is and 16 this decay heat is taking some conservative value 17 appropriate to the timing of these things, typically 18a few hours and now you have removed about 35 or 36 of 19the most megawatts. We take that then and we say how 20was it removed. It was removed downwards, upwards, 21 and sideways.

22 What are the fluxes for doing that or the 23 other? The fluxes are as you see here for the upward 24they go 45 to 100 to 205 to 271. So those are the 25 238 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433upward fluxes. The downward fluxes 15, 43, 74, 100 1and up, about almost 100. Side flux in this case did 2not even have any side. It was only on the conical 3 part. After that, 300, 320, 350. So those are 4 average fluxes.

5Please keep those in mind because now what 6 I want to do is take these other fluxes and then we'll 7going to apply to them a peaking factor so we can also 8find what the local fluxes will be. Now we've done 9this job with the concrete fluid dynamics basically 10calculating natural convection and this is actually a 11 very accurate simulation. Those are based on what's 12called Lusardi simulation. That means they account 13 for all the random movements --

14MEMBER WALLIS: Excuse me. This is in the 15 core again?

16MR. THEOFANOUS: That's the core. That's 17the melt. The situation is holding and the important 18 things are that in this high value we get tubal mixing 19in the main part of the pool. We have stable 20 stratification at the very bottom and you have 21descending cool layers along the walls because the 22 walls are cool. You have the BiMAC there, remember?

23 So this is cool. So it does that.

24The important thing to remember is that in 25 239 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433all those problems we have a constant temperature 1bundle condition because all these things are 2 surrounded by crusts because it is cooled here, here 3and everywhere. So just crusts. So it's an actual 4thermal bundle condition. That's why they put another 5 calculation. That's what you find. This is the 6velocity distribution. It's again tubal over here and 7 it is a nice sliding layer over here.

8 Now to point out since I have the picture 9up there, when I have the near-edge channels with the 10vertical pipes over here, I'm going to have, remember 11those channels are also shorter in the incline and 12along that way and what this does is it creates a 13whole layer on the vertical side that floats along and 14 impinges right in that corner where the incline 15 begins. That can locally load and you want to know 16about that. They can locally load higher heat flux 17because of that impingement in natural convection.

18That's all natural convection and then it surrounds it 19just like that. So that's what it's stating over 20 here. That can be quite significant. It can be three 21times the other heat flux locally and you get that 22only near the edge channels. You don't get that in 23 the other channels.

24 Okay. So here then is kind of a summary 25 240 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433of all the results and basically we're having a number 1of scenarios which are defined in terms of what you 2 want of the BiMAC, near the edge, near the center, 3 different -- we find a number of things.

4 MEMBER WALLIS: What's in the --

5 MR. THEOFANOUS: I'm sorry.

6MEMBER WALLIS: The corium contains the 7 control rods and everything like that?

8 MR. THEOFANOUS: Yes, everything.

9MEMBER WALLIS: And it all stays in there.

10There is none of it which is evaporates or anything 11 like that. It all stays in there?

12 MR. THEOFANOUS: Only volatile --

13 MEMBER WALLIS: Homogeneously distributed.

14MR. THEOFANOUS: Only volatile fission 15 products will vaporize from this side of the vessel.

16MEMBER WALLIS: Right. They are slowly --

17 MR. THEOFANOUS: So what we find here is 18 that the up to down, what's important, those are the 19 fluxes, up, down and on the sides and of course, 20there's no vertical for the near-edge samples because 21you see there is no vertical segment. The core, it's 22 applicable because there's a vertical segment.

23Those are average fluxes and then we take 24 here the ratio of q up to q down and you find that in 25 241 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433the -- oh, the important thing is ABC and MNO those 1are 2D simulations. Two D is much cheaper to do 2 because basically you're assuming that there is no 3movement in the direction normal to the slides that 4you are calculating. In a way what that does is it 5restricts the turbulence. It restricts natural 6convection, it can only rotate this way, but it cannot 7go over that way and that has a restricting effect on 8 turbulence.

9 So as a result of that, the q up to q down 10is about two in a 3D simulation which is one of those 11cases C and M basically repeating Case C but in 3D and 12repeating M in 3D, this ratio is more than three, 3.4, 13 3.5. And in Chapter 21, there is one calculation of 14 each. This is taken from Chapter 21. Since that 15time, essentially we had nothing else to do. So we 16 had a lot of time to calculate in between. So we've 17done lots of those 3D calculations since that time 18 which are very laborious and very computer intensive 19because now you end up with millions of notes 20 especially on fine grid.

21 MEMBER DENNING: And those are DNS.

22MR. THEOFANOUS: Those are DNS, yeah.

23Large simulations so you solve in all directions. But 24 any way, we confirmed these values of about 3.4, 3.5 25 242 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433and we're going to probably make an addendum for that 1but we're going to publish these results in the 2 literature. So in that publication, we're going to 3 add these results.

4 MEMBER DENNING: Going back to the NES.

5 MR. THEOFANOUS: NES.

6 MEMBER DENNING: So this is NES.

7 MR. THEOFANOUS: Yes.

8MEMBER WALLIS: Is the Agency going to 9 accept a design which is only verified by CFD?

10 MR. THEOFANOUS: I think that is for you 11 to decide.

12MEMBER WALLIS: Not me. I was just 13 wondering about the Agency.

14MR. THEOFANOUS: Well, first we have to 15explain to them what CFD is to the Agency and then 16 they have to decide if they are going to accept it or 17 not.18 That's why I give you a few more results 19 here so you can get a handle on what we mean by CFD.

20 What's possible to do at CFD?

21MEMBER WALLIS: So how confident? What's 22 the probability that you're right?

23 MR. THEOFANOUS: I'm going to explain to 24 you.25 243 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: The CFD. How do you 1 assess that?

2MR. THEOFANOUS: I'm going to explain what 3 CFD is.4MEMBER WALLIS: I understand what CFD is.

5MR. THEOFANOUS: Let me explain to you 6what CFD is and then I'll tell you how confident I am.

7MEMBER WALLIS: CFD isn't very good for 8natural convection, is it? The turbulence model.

9 It's just any simulation.

10 MR. THEOFANOUS: This is --

11 (Several speaking at once.)

12MR. THEOFANOUS: In fact, we are not going 13 in this area. If you want to talk more, we can talk 14 more. You tried to do with a certain model for 15 example. A CFD will total the results. If you do 16 Lusardi simulation, you get wonderful results. Some 17 of that is in the report.

18MEMBER WALLIS: Wonderful, full of wonder?

19MR. THEOFANOUS: No, full of wonderful 20 results. Now I knew you were going to be a little 21 skeptical about it so I picked that one.

22 MEMBER WALLIS: Okay.

23MR. THEOFANOUS: You might like that. So 24one question one might ask, exactly, how good CFD is.

25 244 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433So we are for example of course prepared because of 1our experiment that we did, the appropriate experiment 2that we did for the industrial retention for 3Westinghouse, those are half scale experiments, so 4 half scale natural convection experiments.

5We have interpreted that with the CFD. We 6adjusted the parameters. We have interpreted smaller 7experiments with other people before us, new 8experiments, but this one is a big experiment and it's 9 part of the typical hydro-dynamics.

10Somebody might ask and we did ask 11ourselves a more fundamental question. How can we 12actually predict the stability? When you start 13something going off, you are going to develop a 14pattern of rolls of fluid that rises and falls and 15some very interesting things happen there and we 16happened to observe them quite coincidentally because 17 we had an experiment that we used for this, for this 18 one.19But we have an experiment that we'll call 20it the "better experiment" in which we were interested 21all done up and was interested to know what makes 22burnout in nuclear boil. We can go into that if you'd 23 like but it's essentially one about burnout. It's a 24previous -- but we'll come to that by the way in a 25 245 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 moment because we care about burnout with BiMAC.

1So it's interesting to know what makes 2burnout and people will tell this all is a super idea 3 and there is interference with the water coming down 4 and the steam going up and it's none of that, nothing 5 of the sort actually. It has nothing to do with the 6 burnout. So we had that experiment which we were 7 doing for NASA.

8MEMBER WALLIS: It may have been improved 9 by the NRC.

10MR. THEOFANOUS: Yes. They're still 11 around. So we have here this better experiment which 12 was developed for NASA and what this is a 100 micron 13 thickness glass which has on the top of it about 100 14 nanometers of titanium very good deposited. So it's 15 very, very smooth and it is very almost optimistically 16 smooth. But there are thin, but eventually these are 17 thin. It serves as an instantaneous temperature 18locally over that whole surface if you can observe it 19 with a infrared high speed camera and that's what we 20have here. So you have 100 nanometer and this is two 21 by four centimeters, 20 by 40 millimeters.

22 And we can now see fluxes that are three 23 times equal to it in here even with an anoscopically 24smooth surface. That's another story. But then when 25 246 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 you started off and observe you get some interesting 1patterns forming. Ah, maybe let's see if maybe we can 2 predict those patterns.

3So this experiment was underway. We have 4laying on the top of this nano-film. Then we have 5 here a glass mirror, a gold mirror, which are to see 6that whole area with a high speed infrared camera 7which heightens in high speed. So we run into 8 thousands of frames per second and the resolution is 9really at some microns. So it's really a very 10 accurate measurement and each pixel will tell us the 11 temperature instantaneously there.

12 So then this is the moon is very beautiful 13but again, I didn't tell the space to do it again. So 14 it shows you here the experiment, the development of 15this runny -- this cellular structure, how it's 16starts. This is tremendable time and this is what the 17CFD will give us color coded. So to me, that's really 18 remarkable to catch that. With the velocity and the 19stability and the development of the cellular 20 structure, you can do very well.

21Now I'm going into more mundane things 22 then. The central samples were decided from the table 23that a bounding downward flash on the horizontal is 24100 kilowatt per square meter. By the way, I point 25 247 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433out that in these slides this is wrong because the 1computer played a weird game on me when I was pasting 2it and this line by mistake was pasted up here. So 3 just put an arrow that shows it on the --

4So for the central samples we have 100 and 5 we have applying to local peaking which by the way I 6 didn't point out the amounts of local peaking in the 7 previous. Here is the peaking over here. This is the 8old 1.25 because you apply to the 100 and get the 125.

9Here is the peaking on the incline and here is the 10 peaking on the vertical. So applying those peaking 11factors there give us near-edge samples 100 and 300, 12 that's the factor of three, and the radial channels 13 it's is 320, 450 and that's --

14 MEMBER WALLIS: What's the BTUs per hour 15 per square foot?

16 MR. THEOFANOUS: What?

17 MEMBER WALLIS: What is that in BTUs per 18 hour2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> per square foot?

19 MR. THEOFANOUS: Okay. Let me see. If 20 you could tell me how much is square foot --

21MEMBER WALLIS: Is it 300,000 or something 22 like that?

23MR. THEOFANOUS: Okay. So if it's 300,000 24then this would be one-third of that. So it would 25 248 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 100,000. I'm just rescaling because everybody knows 1 that --2MEMBER WALLIS: And I know it's completely 3 wrong.4MR. THEOFANOUS: All of those could be 5 wrong. It's very convenient because it's thousands.

6 So now we are running into another interesting topic 7and that is how much thermal loading will those pipes 8 take. At this time I address that question was for 9 universal intention and people were asking me. Some 10 very skeptical people were saying the bottom of that 11lower head and very, very bottom is so flushed that in 12 theory you should take zero critical heat flux. But 13of course, you don't because even however so slight 14 the inclination that we have actually creates lenses 15 and those lenses of water they escape and periodically 16this happens and as the boiling occurs there, you have 17a micro-layer forming on the surface and as long as 18the lenses escape and the flattening of the water 19happens, we think the time interval is that is less 20than what it takes to dry that micro-layer, you're 21 fine.22And we demonstrated that this is so by 23experiment which this was later incarnation of that 24experiment and the very first experiment. Don't have 25 249 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433the picture because it's not a nice clear picture. It 1was a computer data based, but what it is is basically 2 a slight incline just like models the very bottom of 3a reactor vessel but in a channel geometry. That's 4where it's actually relevant to the BiMAC and we have 5 pipes and we filled it up to some point and we let it 6go in natural convection and we heated it from the top 7and what you find is those lenses form and escape, and 8then we've got critical heat fluxes in the very, very 9 bottom of the pool of over 300 kilowatts per square 10 meter.11 We went to this channel here because for 12the standard it was interesting to see if we could get 13more not for the bottom. Nobody cares for the 14 bottom. It's for the sides because for PWRs you get 15this focusing effect and we put a channel so that 16natural convection hopefully would create a smooth 17 current and we decreased the critical heat flux here 18and indeed it increases it. So here we have the 19channel geometry of whirlpool configuration of four 20 and this was done for Westinghouse.

21Here I showed you the real facilities and 22it's pretty large. So it was full scale flash of the 23lower head and it goes into a riser and in the back 24here, there is a downcomer and there is a condensation 25 250 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 time. So you get boiling here with the heat boiling 1from the top. The width by the way of this, the width 2of this box which creates the heat is 15 centimeters.

3So the heat is going downwards, boils. The two-phase 4 flow goes to the riser and we see all kinds of 5interesting instability phenomenon that occurs like 6geysering and stuff like that's important for other 7 things for boiling water reactors.

8 And then up here, the steam condenses in 9 the coil and then we have the downcomer. So it goes 10like that. So in a full, when we say running in full 11natural simulation mode, we're running it so that the 12 water is enough to create a continuous flow. But we 13have also been running in a pool of boiling water 14 which the water is so low here that it doesn't close 15 the loop. So that's what it is.

16 And here we have, this is a 300 power 17diesel generator with 400 kilowatts power coming in 18 and this is controlling the power surge so that we 19 could have any power surge we want and so all this 20 good stuff. And here, this guy is a big guy. So it 21 gives you the idea of the size of this.

22All right. So Configuration 1 was the one 23that was natural convection. It was only the very 24bottom part with a very slight inclination that I was 25 251 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433describing before. And this is the critical heat flux 1 expressed as a function of the angle is like that.

2The characteristic dimensions of that channel was very 3similar to BiMAC with about 14 or 15 centimeters by 10 4centimeters, something like that about 10 centimeters.

5Then the Configuration 4 also in the pool 6 volume. That means putting the power over the whole 7thing in Configuration 4. You see 90 degrees you get 8about a megawatt which magically, Graham, that's your 9magic number even though it's vertical and then all 10those points are what we did for Configuration 4. So 11 for us, that will give you an idea.

12For the incline part of our BiMAC, we are 13about here. So we would expect about 400 kilowatts 14per square meter for the vertical part. For the 15vertical pipe we would expect about a megawatt as 16 limits. So that is represented over here. Critical 17heat flux. This is for the incline section. This is 18for the vertical section. So this plot is made of two 19 parts. One part is the incline and the other part is 20the vertical and over here is the heat flux and the 21black is near central channels. So near central 22channels with high goes to a maximum but a very small 23 maximum and not strong. That was more near the edge 24 of the channels and then it falls off. Then for the 25 252 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433near edge channels, the blue, that it goes to that 1local peaking that I was describing before. That's 2 because of the descending layer.

3 MEMBER WALLIS: Now heat flux is defined 4 based on the flattened area or the --

5 MR. THEOFANOUS: Based on the flat area.

6 MEMBER WALLIS: Or the two -- the flat.

7 MR. THEOFANOUS: The flat area, yes. An 8equivalent flat area and then over here, what you see 9 is the thermal loading on the vertical wall.

10MEMBER WALLIS: So an equivalent flat 11area, don't you mean the actual flat area? You don't 12 know --13MR. THEOFANOUS: Well, the calculation, in 14the calculation you don't make the boundary like that 15 in a calculation of getting from a wall.

16MEMBER WALLIS: You use the superficial 17 area.18MR. THEOFANOUS: Yes. And to put that in 19terms of the margins are defined in this way and we 20find margins of course but this is a departure from 21 one in this ratio. So you find the minimun and even 22that is about 60 percent margin and also near the top 23of that. Actually, when we run these experiments, we 24 find that this for the BiMAC you find out that this 25 253 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 flux here near the end of the incline is going to be 1much higher I believe because also you have natural 2convection there. Remember this one also based on 3full boiling and also I think the other part on the 4 vertical side. So that's what that is.

5 And so at this point, this is the vertical 6transposition of thermal loading to alpha G. Critical 7 heat flux is alpha G. Thermal loading is what comes 8out from the peaking of natural convection. Where the 9trouble is that you have 60 percent margin to failure.

10This needs to be remembered to put in context and 11 that's really tough of being extremely conservative 12on the thermal loading and reasonably conservative for 13the critical heat flux. So in a way, again there's no 14intersection between load and fragility and we see the 15 failure of this thing is physically unreasonable.

16 So for someone then again going back to 17the question that Jack was asking for the survival of 18that, that's how we decide those things. We find the 19 loading. You find what is day-to-day failure, compare 20the two and say okay, you'll fail with that. This is 21pending of course information because I'll be the 22 first one to say that for -- and that is quite 23different from the CRD question that Graham asked 24 before.25 254 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433When you actually want to make use of 1 something of that is of an empirical nature which is 2the critical heat flux is empirical, what you make 3sure is, let me finish, you want to make sure that 4 your experiment is really representative of the real 5 condition.

6 MEMBER WALLIS: Now, Theo --

7MR. THEOFANOUS: So therefore I would say 8 that BiMAC as far as critical heat flux is concerned 9needs to be confirmed with real experiments and we can 10 go into that. Yes.

11MEMBER WALLIS: This is my stuff. But you 12have this corium and sitting on this layer which I 13 thought you said was sacrificial.

14 MR. THEOFANOUS: Yes.

15MEMBER WALLIS: When it's gone, don't the 16 pipes seal the corium?

17 MR. THEOFANOUS: Of course.

18MEMBER WALLIS: Corium interacts with 19 steel.20 MR. THEOFANOUS: Yes. Before --

21MEMBER WALLIS: Does corium eat the pipe?

22 MR. THEOFANOUS: No.

23 MEMBER WALLIS: Why not?

24 MR. THEOFANOUS: For the same reason, it 25 255 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 doesn't hit the pipe --

1MEMBER WALLIS: There are all kinds of 2 Utechics and stuff.

3 MR. THEOFANOUS: The same reason that it 4 doesn't do that for the --

5 MEMBER WALLIS: So it's cold.

6 MR. THEOFANOUS: No, for the same reason 7 it doesn't do it for --

8 MEMBER WALLIS: Does it crust the --

9MR. THEOFANOUS: Because it crusted it.

10 Right.11 MEMBER WALLIS: Okay.

12MR. THEOFANOUS: Because now corium cannot 13 exist at temperatures of --

14MEMBER WALLIS: So the crust protects the 15 pipes, although the sacrificial layer is gone.

16MR. THEOFANOUS: Yes. Basically what 17 happens is that it's a self-adjusting situation. If 18the thermal conduction resistance is more than what 19the thermal loading is, there is going to be a little 20 bit more until now it's just as much as the thermal 21loading to the cooling. But it will never eat more 22 than that.

23I show just for engineering purposes, I 24emphasize that because in CFD when you know what 25 256 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433you're doing, you actually are making predictions 1based on basic physics and we talked about that. For 2 that one if you check your calculation and make sure 3that all the physics are presented, then you're fine.

4You don't need ULPU experimentation. Even then we 5 have a lot of comparisons as I mentioned before, but 6 this one is totally empirically based. I claim that 7 we cannot really predict critical heat flux yet 8correctly even on a horizontal pool boiling facing 9 upwards. So I certainly don't want to tell that you 10 can predict it facing downwards or inclined.

11So that's why I went through very special 12 pains here actually to show you that on the basis of 13principles this BiMAC is a good concept and that is 14 principal evaluations. It just so turns out that we 15 were lucky in that we had channel data for ULPU that 16are quite applicable to both dimensions as well as 17orientation of interest here. So that gave us a very 18 good idea of what we can expect when we do full scale 19 experiments to BiMAC which in fact you can do full 20 scale. We can actually make full scale without any 21 big deal and we plan to do this.

22 All right. That is all for the critical 23heat flux. But we're not finished yet because we said 24 we also want to make sure that there is enough water 25 257 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 depletion so that near the end of the channels I end 1up with a 70 percent void fraction and 70 percent void 2fraction, I don't know where the liquid is. Most 3 likely liquid is on the bottom but not at the top and 4I want to make sure they have liquid everywhere to 5keep the wetting the walls because that's underlined, 6this rewetting of the walls, to actually very 7 interesting because nuclear boiling in fact is a 8misnomer here and as it is, even in nuclear boiling in 9 misnomer even on the flood plate faces upwards.

10The reason it is a misnomer is by the time 11you go to near critical heat flux levels actually the 12whole surface is covered by vapor film and all the 13cooling is happening with the micro-layer that is 14 hidden underneath that film. So the only difference 15between a plate facing upwards and the plate facing 16downwards is in the renewal process of that film maybe 17thinning and thickening again. What you don't want to 18 do is you don't want to have that film go to zero even 19 for a short time because that's going to be burnout, 20 although for vessel retention for vessels as well as 21for BiMAC when you have significant wall thickness 22there is enough thermal inertia and the fluxes are low 23 enough so that even if you dried out temporarily 24 you're not going to go to very high temperatures and 25 258 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433before the temperatures get to the first point you're 1going to still be able to rewet and recover the film.

2 Okay. So that's -- and we're going to go 3to that and we're perplexed about that because what do 4you use it for getting the natural convection and 5incline in the pipe like that. There is no literature 6 for measure. So I remember that many years ago when 7 I was doing the retention and we were doing the ULPU 8the French decided to sort of a similar experiment, 9 but they wanted to go more fundamental and they did 10 the SULTAN facility more fundament than us.

11We tried to mock up the reactor because I 12believe that the right way of doing critical heat flux 13at least at this time is by mocking up the real 14 situation. The French thought they could build a 15 straight channel that is facing downwards, so 15 16centimeters just like whirlpool, four meters long, 17facing downwards. They put it on the platform so they 18 could orient it from vertical to near horizontal and 19they thought they supplied forced flow through that 20 and they figured that -- they measured pressure drop 21and they measures critical heat flux. So the idea was 22 that take fundamental data presumably which then can 23be used in some codes whatever to predict critical 24 heat flux and of course this never happened.

25 259 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433So at the end of this experiment as far as 1contributing to data for the critical heat fluxes, 2they contributed zero because at the end then they 3started using my data from whirlpool. However it did 4contribute us now because I remember that they 5 measured pressure drop and I said now I can see if I 6can calculate correctly pressure drops on incline 7channels of the size of kilometers and there is no 8other data anywhere to find on that, so sort of 9 sitting there getting resolved.

10 So we have this nice set of data, very 11 appropriate distances like four meters. We were 12 interested in about four meters or five meters. The 13dimensions there 10 degree inclination was included in 14the data. The characteristic length was 15 15 centimeters. They also got 15 centimeters. The 16channel length four meters. The pressures were all 17the way from one atmosphere to I think five or ten 18 atmospheres. I forget now. Power levels accounted to 19 kilowatt per square meter. They get detail pressure 20 drop data and again here from the top.

21 So we took this and we made a boiling 22model which was basically an equilibrium model in 23equilibrium boiling using LOCA Martinelli for the 24pressure drop modified by as far as the void fraction 25 260 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433modified by something that is a function of the 1inclination and this came out from an obscure paper 2that nobody knows about. It was published in Thermal 3 Engineering or something in Russia back in the `70s or 4`60s which they actually did exactly that thing. They 5 took LOCA Martinelli and they found how they want to 6correct LOCA Martinelli for orientations other than 7horizontal and by using that, we got actually very 8nice interpretation of the shorter experiments. So by 9 having this kind of basis, I can say that we are 10calculating correctly pressure drops through the 11channels under all kinds of fluxes that will fall even 12 well beyond fluxes I'm interested in.

13 Having said that, now all I need to do 14simply check and find what is the gravity imbalance I 15get in those channels match it against my pressure 16 drop and then I get my natural convection. Simple as 17 that. So now having that, I get this. Here is the 18heat for different heat flux levels. They must 19formulate natural convection of course increases as 20you increase the flux, reaches a maximum and the 21gradually decreases and that's because of two phase 22 friction up here.

23 So remember the point of interest for us 24is from here to here, somewhere inside here and the 25 261 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433flow in this situation is very stable. The flow is 1such that it's actually self exhausting because any 2increase of void fraction the net change in gravity 3here is more than the change of friction in the range 4of interest and that is the definition of having a 5 stable flow.

6The next question, the more interesting 7question, is what is the void fraction. As I said, I 8 didn't want to see here at this kind of flux, I didn't 9want to see 70 percent void fraction there and 10fortunately I don't. I see, like in the upper limit, 11 I see 40 percent void fracture. So in the most I'm 12going to be in some kind of a slight -- because I knew 13anyway which means bubbles are forming, they are going 14fast and then very high frequencies of wetting and 15 rewetting in the sense of the micro-layer. So that's 16 the story for this.

17So BiMAC then, so besides the point I made 18already which says that the BiMAC needs to be verified 19by experiments and what I visualize here is full scale 20 experiments. That means the full dimension, full 21pipe, full length, vertical downcomer with real power, 22power shape, whatever I want to do that so I can 23define the local critical heat flux, No. 1. No. 2, 24 also I want to run experiments which are going to be 25 262 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433subscale, maybe half scale or quarter scale, in which 1 I have many pipes, many pipes, which are maybe loaded 2 differently. So I want to see the actions between the 3 channels between the pipes and whether that can have 4any -- I cannot conceive of any detrimental effect of 5 that, but it's good to really have that and it's not 6 a big deal to get that.

7So in addition to those conditions, we say 8 that BiMAC needs to be at least RTNSS and that implies 9a qualification of function in its design state and 10this is shown now in terms of principle in 11 development. So this is really the experiments we're 12talking about in the COL and then in addition the 13identification of continuing ability to function as 14design throughout the operating life and that means 15 this will require simply testing of this orientation 16of control which goes back to the probability of 17 actuating this, measuring and actuating.

18MEMBER WALLIS: So it just sits there 19after an accident for the next ten years or something 20 so percolating away?

21 MR. THEOFANOUS: I'm sorry?

22MEMBER WALLIS: After the accident, it 23just sits there and it percolates away for the next --

24 forever.25 263 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. THEOFANOUS: Well, no because you --

1MEMBER WALLIS: Until you get a solid lump 2 and it closes up.

3 MR. THEOFANOUS: Yes, because decay heat 4 slowly goes away. Yes.

5MEMBER WALLIS: But there is quite a long 6 time this thing has to sit there and function.

7MR. THEOFANOUS: Well, you don't have much 8choice, do you? You have it inside the vessel, inside 9 the lower head.

10 MEMBER WALLIS: It has to be somewhere.

11MR. THEOFANOUS: Inside the lower head.

12It's going to be sitting there the same length of 13 time. But it is better to have it sitting somewhere 14 percolating rather than going through the concrete I 15 think.16MEMBER KRESS: And it's eventually 17 solidified in the radiation.

18MR. THEOFANOUS: Yes. Sure. Like I say, 19I expect that in reality there is so much water there 20I believe that BiMAC actually will not really be 21 needed. But you want to make sure that you say that's 22a boundary that just cannot be penetrated. That's the 23intent of the BiMAC. It can be demonstrated to be 24 true.25 264 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER KRESS: When you did your CFD 1calculations, the heat flux, did you assume a uniform 2 mixture of the core and the metal and --

3 MR. THEOFANOUS: Yeah.

4 MEMBER DENNING: Now is that a good 5 assumption? Can there be separation of --

6 MR. THEOFANOUS: You can get separation, 7but really not very much at all to anything. If 8anything, the separation was actually pursued by some 9 people. I believe not rightly so after our work for 10the Agency standard but for the purpose of finding you 11get more heat going upwards than downwards. So for 12upward heat flux we get separation they go more 13 upwards.14 MEMBER WALLIS: So you're --

15MR. THEOFANOUS: I worry about downwards.

16MEMBER WALLIS: So your water after awhile 17gets saturated with cesium iodide and stuff like that.

18MR. THEOFANOUS: There's a lot of water 19 there.20 MEMBER WALLIS: Presumably it does.

21MR. THEOFANOUS: There's a huge amount of 22 water.23MEMBER DENNING: When you said saturated, 24 did you mean literally saturated?

25 265 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER WALLIS: I mean just --

1MEMBER DENNING: Or means gets a lot it 2 in.3MEMBER WALLIS: Gets a lot of it in. I 4don't know what saturated means. Eventually, 5presumably dissolving fission products get in the 6 water and it keeps go round and round.

7 MEMBER DENNING: Sure.

8MEMBER WALLIS: So then you get chemistry 9going on and stuff. There's a lot of term analysis to 10 be done of what it is that you have there that's 11 cooling this debris. It's not pure water.

12MR. THEOFANOUS: There's a huge amount of 13 water. Huge amount.

14MEMBER WALLIS: That's a qualitative 15 statement.

16MR. THEOFANOUS: I can tell you exactly 17 how much it is.

18MEMBER WALLIS: No, but I know. I'm 19saying that there has to be some analysis of what's in 20 the water after a period of time.

21MR. THEOFANOUS: That would be a good 22 question to ask --

23 MEMBER WALLIS: Huge or not.

24 MR. THEOFANOUS: Then we can --

25 266 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: The fouling of your tubes, 1 your tubes foul after awhile.

2 MR. THEOFANOUS: That's an --

3 MEMBER WALLIS: Foul after awhile.

4MEMBER DENNING: Well, is there debris of 5 some sort in character?

6MEMBER WALLIS: Through the precipitation 7 or something.

8MEMBER DENNING: Precipitating out of 9 boiling boundary.

10 MEMBER WALLIS: Right.

11MR. THEOFANOUS: Actually the fouling 12improves critical heat flux interesting enough as you 13 know.14 (Several speaking at once.)

15MR. THEOFANOUS: Right. In fact real 16 cores that's going to be fouled and they might have a 17higher margin so you can up the power. All right. So 18 pulling it all together now and that leads us to the 19 end, we have three conclusions or three concluding 20 slides. Conclusion 1 is for the low pressure 21scenarios and here is a containment phenomena event 22 three, a CPET and what is shown is the major decision 23 points one has to make in order to decide at the end 24this position of those scenarios. So we have here 25 267 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 that's okay.

1 We have low pressure core melt and water 2 level. We ask the first question. What's the water 3level in the pedestal because that makes an impact on 4 steam explosion potential failure. So we have three 5levels defined, already explained the rationale for 6them and it turns out that this is by far the much 7 more likely. This is like one percent of the cases.

8 This is much less than even one percent because that 9situation simply we have no other one or we have lots 10of water. What you have in between is not very 11 likely.12Then we follow this branch and already we 13said that if we take this branch here, the pedestal 14damage cannot be excluded. And the question then that 15we next ask is is the pedestal intact? Okay. We say 16 no. Then the next is are we supplying the BiMAC with 17 water? Are the flooding the lower drywell? And of 18 course, in this case, it's already flooded. So it's 19 yes and then debris successfully cooled and again we 20need to ask that question that related to BiMAC 21 function and as we demonstrated here on the basis of 22principles, the BiMAC function would be good and it 23will be coolable but you put a start to indicate that 24failure or rather the nonfailure BiMAC function needs 25 268 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 to be confirmed experimentally.

1So here then we have however a containment 2failure already because we basically destroyed the 3 pedestal. If we are able to destroy the pedestal we 4destroyed the BiMAC for sure and therefore that means 5 in all those cases you assume containment failure.

6MEMBER WALLIS: The pedestal supports 7 something, doesn't it?

8MR. THEOFANOUS: Yes, of course, that's 9 why it's called a pedestal.

10 MEMBER WALLIS: I know. So what happens 11 when it fails?

12MR. THEOFANOUS: Well, I don't think very 13 much actually except failing the containment because 14this thing as you very well find out is to get a 15failure of the pedestal by steam explosion. If you 16 fail, you fail locally. You will not jeopardize the 17structural integrity of the pedestal function.

18However we cannot count on containment at that point.

19That's why this is known as assumed containment 20 failure. That's one percent of the accidents.

21 For all the other cases, we have no damage 22 here. Don't even ask the question. No damage and 23then here yes, again with very high probability based 24on requirements we have for these two control and 25 269 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433actuation systems. And then here again, pending 1verification, the BiMAC function correct. So we have 2 for here physically unreasonable in all these cases.

3 For those two cases, we're saying that we are 4transferring to CSETs, containment system event trees, 5because now even though everything is fine here, we 6need to know what happens after five days. Is the 7 PCCS pool replaced with water? What does it take to 8 not have containment heat anymore at that point? So 9all these systems affect in the next presentation. So 10 that's what that means. This takes us to that.

11And this one is the high pressure CPET.

12The first question of course is is the reactor cooling 13bundle intact and already I showed you that natural 14convection is very likely for this reactor as with all 15reactors because with the high pressure vessel 16convection of the steam. However we didn't want to 17 come to that on that basis. So we used that in what 18we're calling our jargon. In Rome we call it splinter 19 scenario. That means since you don't, we can't 20 guarantee that that's what is going to happen, we're 21going to assume that either that or that happens. So 22 that means we take that as if it was to be the case 23which means it doesn't fail and that's why this is 24 written in the way that's the ES branch.

25 270 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433And then the DCH containment failure no 1 damage we demonstrated. It won't fail. This is 2physically unreasonable and we demonstrated this 3branch with the physically unreasonable. Then we have 4 the flooding and the function. So again if it's all 5 yes, yes, yes, it goes to CSET.

6So the conclusion three is a summary of 7 containment threats and mitigative mechanisms on the 8 systems, all the systems in place. So here is like a 9capturing of them together for the three threats that 10we addressed and this is the failure marker. Already 11 we covered that.

12And here is pretty more crisply what is it 13that we are putting in place to deal with that. So for 14example for the DCH we have pressure suppression 15 vents. That's the principal mechanism and we have 16reinforced confidence support. That allows us to use 17the high fragility and the events allows to have a 18 limit on how much can be pressurized.

19 Then on the liner thermal failure is the 20liner anchoring system. On the lower drywell is also 21the separation by the lips as I mentioned before. On 22 the explosions the pedestal liner failure here again 23is the dimensions of the wall and the enforcement what 24holds it together. The BiMAC failure is the pipe the 25 271 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433size of the thickness. So it is structurally very 1 robust and backed up by a lot of concrete.

2 Then on the BMP, BiMAC activation 3functions essentially actuation is through rotation 4but it would be specifically designed to very high 5 standards of reliability and the diverse and I would 6 like to see passive valve action so to make sure all 7 those scenarios we can flood the lower drywell.

8Local burnout, natural circulation and the 9 inclination of the pipes, that's what it takes. Next 10the case for BiMAC, water depletion again, that's 11natural circulation and inclination and low boil 12fractions and actually lower heat fluxes also. As we 13have seen in the local melt-through is the refractory.

14 And I think I have a bunch of back-up slides in case 15 you want to ask me more questions about CFD.

16CHAIRMAN APOSTOLAKIS: Any questions from 17 the members.

18MEMBER SHACK: If you don't credit the 19BiMAC, is the melt spreading and heat flux you get for 20 this comparable to the ABWR?

21 MR. THEOFANOUS: Yes. In fact, more.

22 MEMBER SHACK: More.

23MR. THEOFANOUS: And in fact like I said, 24we could have easily have taken, not easily, but we 25 272 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433could have taken the approach to say EPRI criteria and 1 just like ABWR and then just argue that we can take a 272 hours0.00315 days <br />0.0756 hours <br />4.497354e-4 weeks <br />1.03496e-4 months <br /> or more to eat through the concrete. That 3 was again sort of the traditional approach, but it's 4not good. I think maybe we should make sure these 5reactors are having something that we are sure that 6 this will not penetrate.

7MEMBER WALLIS: Well, it's a very 8impressive story. I'm just wondering what you have to 9do to convince the very skeptical agency, the very 10conservative regulatory body, that they can accept 11 this with a lot of confidence.

12MR. THEOFANOUS: I think that what we need 13 to do is that we need to for sure make BiMAC 14experiment which as I said before we can do it full 15 scale. That's why it's convenient.

16MEMBER WALLIS: But you're going to 17simulate that corium heating electrically. We're not 18 going to have real corium --

19 MR. THEOFANOUS: Of course.

20MEMBER WALLIS: So there are always going 21 to be questions about --

22MR. THEOFANOUS: I'm sorry. I'm sorry, 23 Graham. A kilowatt is a kilowatt and a meter is a 24 meter. Now if you want to be conservative so you can 25 273 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433begin to realize the units of thermal power then I 1 throw my hands up.

2MEMBER WALLIS: I'm not saying what I 3 want. I'm just asking questions. That's all.

4MR. THEOFANOUS: I wish it was about 5 safety rules in that category.

6CHAIRMAN APOSTOLAKIS: Any other 7 questions? Okay. Thank you very much. Let's take a 8 few minutes because we have another hour and a half, 9 guys. Ten minutes. Off the record.

10(Whereupon, the foregoing matter went off 11 the record at 4:01 p.m. and went back on the record at 12 4:15 p.m.)

13CHAIRMAN APOSTOLAKIS: On the record.

14 Okay. Next subject is Containment Systems.

15 MR. WACHOWIAK: Containment systems. So 16this is the continuation on now from the CPET into the 17 CSET. It think there is quickly two things, at least 18 one thing I want to answer from before. I think the 19question came up peripherally and I'm not sure it was 20 answered, how did we decide which things went, which 21sequences went, into the high, medium and low water 22level categories. Basically, what we did was we 23 looked at the scenarios that got us to core damage.

24The low pressure scenarios or all the scenarios in 25 274 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 fact, but especially the low pressure scenarios that 1 got the core damage.

2If it was something that was putting steam 3 into the containment and we were condensing steam on 4the walls and it was condensate from off the walls 5just getting down into the lower drywell, then we 6 showed that we're only going to get a few centimeters 7of water down in the buyback. So we called all of 8those low. If there was a break in like a drain line 9 or a large break in the reactor called liquid type 10breaks, if it was a large break down low, not a steam 11 break, but a liquid break, enough liquid from those 12breaks put a lot of water down in there and it got 13 above that value.

14 Now we did look at some other things where 15some of the breaks were kind of in between and it 16depended on whether or not you had any injection 17systems working or not like if it was just a break and 18the water came out, it would be in the medium 19 category. But if a CRD pump was running, it would 20have moved it up to the high category, so maybe not 21 quite enough to cover the core, but enough to add a 22little bit of water. And so it was kind of in between 23 and there were some of those things in the medium 24 category.25 275 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 In the end when we went back and we looked 1 at all the different scenarios that we had to try to 2see where things fell, like in our Section 7, 3everything in the top set of cut sets that we 4described, nothing fell into the medium category. It 5 was either high or low.

6In this last round when we did the 7 quantification for the containment system event tree 8 we took another look at that with all the sequences 9 that were above the truncation value and there might 10be one or two sequences that are at the 10

-13 level 11 that could fall into there. So what we did was for 12the purpose of the analysis, we just took some from 13 the low water level and we just put it into the medium 14 level. So the low level came out to be like 0.991 and 15 we made it 0.99 and we put .001 in the medium category 16just to cover those scenarios that might be just 17beyond our truncation limit. So that's how we 18assigned all of those by looking at what specific 19 scenarios got us to the severe accident.

20 Just to be clear on it, the high pressure 21sequences, we didn't see anything in the high pressure 22sequences that would have fallen into a medium or high 23 water level category. Those were all low and in the 24ATWS sequences once again, those all looked like they 25 276 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433were low water level. There were a couple right at 1the truncation limit that may fall into some of the 2 other categories.

3MEMBER DENNING: I think that in some 4scenarios you carefully limit the amount of water 5addition to prevent overflow into the cavity. Is that 6 true? Isn't that true?

7MR. WACHOWIAK: What we did -- Let me 8answer it this way. We altered the design so that the 9 design itself will limit the amount of water flow into 10the cavity. When you get the steam environment in the 11drywell, the steam as we'll see in a minute there goes 12 into the PCCS, condenses and then goes into the GDCS 13 pool. What we've done is we've designed the GDCS pool 14so that if it overflows, the overflow water goes into 15the suppression pool. It doesn't go into the drywell.

16Things that condense on the wall though 17 will still run down the wall and go down into the 18water drywall. We've also added to our emergency 19procedure guidelines instructions that say don't spray 20the containment unless you are either absolutely 21positively sure that you're not going to lose core 22cooling or you know that the core is on the floor. So 23 those are our emergency procedure guidelines because 24that would be the other way is operator doing 25 277 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 something that would --

1 Okay. And I can't think of the other 2question right now that I heard on the periphery that 3we may not have answered, but we'll see when we get 4 through these.

5 I'm going to talk now about the 6containment systems. We looked at for the last couple 7 of hours, we looked at the basemat melt penetration, 8 EVE, DCH and the robust design that we have for those 9 scenarios. What we haven't necessarily looked at here 10are the containment bypass and containment 11 overpressurization and the systems that are involved 12 in addressing these particular things.

13 So let's start out with the simple one, 14the containment bypass. How can you get a containment 15 bypass? You have big penetration that's open to the 16containment at the time that you have the severe 17 accident. So we went through our list of penetrations 18that are in the design. They are all listed in the 19 Chapter 6 of the DCD and we did an evaluation. They 20are all either normally closed during operation, 21connected to a close system inside the containment, 22connected to a close system outside the containment or 23 have already been addressed in our break outside the 24 containment evaluation in the Level 1 analysis.

25 278 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433Our conclusion is then that we really 1don't have a credible bypass scenario here. There are 2a couple caveats on that. No. 1 is there's a bunch of 3little penetrations that haven't made it in the 4detailed design phase that we talked about in the 5third column here this morning. We don't really know 6what those would be. We're pretty sure how they would 7 come out, but we just don't know yet.

8 Then also some of these that are connected 9to some of these other systems may be periodically 10operated during the operation of the plant a very 11small fraction of the time. But there is a chance 12that they'd be there. So we retained in our 13containment system event tree structure the 14possibility of having the containment bypass from one 15of the penetrations being open. And the way we 16addressed that was we looked at what is the likelihood 17 that we're going to have a severe accident where the 18 control systems for these isolation valves would not 19be available and that's how we kind of assigned the 20 value there.

21Containment isolation valves tend to be 22 failsafe. They fail closed when they lose power.

23They go in the right direction that we want them 24 passively. So the control system is really the key 25 279 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 factor there.

1MEMBER SIEBER: Did you make any 2 assumptions about vents and drains?

3MR. WACHOWIAK: Vents and drains are in 4the detailed design phase that we don't have that 5 information and support.

6MEMBER SIEBER: I mean they are usually 7pretty small. On the other hand, it's an opportunity 8 to have a bypass.

9MR. WACHOWIAK: Pretty small. Now we know 10what we did in the ABWR analysis for the total of 11 containment bypass, but we really needed to know the 12detailed information on those small penetrations to 13figure the aggregate of all those. That will be done 14again just like that in a later phase. But once 15again, we did retain this here trying to make sure 16 that we capture the phenomena.

17Now overpressure protection, our function 18for overpressure protection is provided by the passive 19containment cooling system and it can also be provided 20by the fuel and aux pool cooling system and then 21finally, if there's a, if we get into a really bad 22situation now, we could go and do a controlled manual 23event of the containment through the suppression pool 24 to the elevated release point.

25 280 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 Just like at Level 1 like we talked 1 before, we have a passive function backed up by an 2active function backed up by a redundant active 3 function. So the robust nature of how we deal with 4the containment overpressurization. Let's talk about 5 some of the individual pieces of this.

6The PCCS operation during a severe 7 accident. In the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, there is nothing 8that has to happen. It's completely passive. As a 9matter of fact, we have some analysis that shows it's 10significantly longer than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. It gets out 11 toward a two-day period, but the design spec now has 12 to be for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. There's enough water there. So 13 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> in, nothing has to happen.

14Steam in the drywells condensed return to 15the drywell. It's a closed system. Now in the 16scenarios where we're looking at this in the severe 17accident scenario, remember from the containment 18phenomena of entries, we've already passed through the 19 question did the deluge line to the BiMAC work. Did 20 those lines open? So even though the PCCS goes back 21to the, sorry, the GDCS pools, those lines are open 22 from the GDCS pools to get it back down to the BiMAC 23 again. So it is a closed system here.

24There is some residual risk if you will 25 281 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433that has already been addressed in our quantification 1of looking at the lower branches where the deluge 2lines have failed. For the CSET, we always have those 3 lines open.

4The aerosols that are generated and get up 5through the water, that's an interesting question 6about what's ultimately retained in that pool of water 7there, but what we've seen is that they're carried up 8with the steam, condense in the PCCS and the aerosols 9are actually not deposited inside the PCCS heat 10exchangers themselves. It's carried with the 11condensate back down into the mixture of water that's 12 back in the containment.

13 The only real issue that we have here is 14how much non-condensable gets up into and held up into 15the PCCS. If we do have non-condensables there, it 16 reduces the effectiveness of the system. There is a 17vent line that's provided and in a couple minutes 18 here, I'm not sure exactly where the slide is, but 19we'll explain exactly how that works. It does have 20 this. It requires our vacuum breakers, suppression 21 pool to drywell vacuum breakers, to remain seeded in 22 order to make the thing work.

23So the situation here is on the drywell 24side the steam gets into the heat exchanger. The pool 25 282 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433outside boils. The condensate comes down through 1 these lines here and comes back to the GDCS.

2MEMBER WALLIS: There must be some kind of 3way of separating the condensate from the non-4condensables which has not been very clear in this 5 picture.6MR. WACHOWIAK: Yes, it's difficult to see 7in this picture. I agree with you, but these are 8really a pipe within a pipe kind of arrangement to 9minimize penetrations. I think that's how it was 10described to me. The condensate comes from the bottom 11 of these.12MEMBER WALLIS: So the non-condensable 13line goes up inside the other pipe. There's a 14 different --

15MR. WACHOWIAK: And the non-condensable 16vent line goes up inside so that it's at the top of 17these end bell tanks so that the condensate -- And 18because this system condenses faster or at the same 19rate or faster than it's being supplied, I'm sorry.

20It condenses at the same rate it is being supplied and 21 the drain goes out faster than it's being supplied.

22 In order for this to work, the drains have to be open 23just like you're taking a shower. All the water goes 24down into the drain there. It's coming out of the 25 283 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433showerhead and going down in the bottom. A similar 1sort of thing, it's condensing in here and falling out 2 through the bottom and so this tank here is mostly 3 empty. The top of the tank has non-condensables and 4 steam mixture there.

5This is piped directly into the 6 suppression pool and it shows a sparger here, but it 7could be an open pipe. It really doesn't matter.

8All that really matters is that the submergence of 9this pipe is less than the submergence of these 10 events. Then you always have a differential pressure 11 between the drywell or the inside of these valves is 12the same as the drywell and then the drywell pressure 13is higher than -- I'm sorry. The differential 14pressure between the end of this pipe and in here, 15that column of water, is going to be the difference 16 between --

17MEMBER WALLIS: Where is the level in the 18 vent pipe reaching and where is the level in the --

19 MR. WACHOWIAK: The level is in the vent 20 pipe is always going to be the same as the level --

21 MEMBER WALLIS: The same as inside.

22MR. WACHOWIAK: Yes. A small difference.

23 MEMBER DENNING: No wait a second.

24 MEMBER WALLIS: No, it's not.

25 284 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER DENNING: No, but when you have 1 pressure in the drywell, then it's going to drive it 2down to the submergence and that's where you get your 3 head to drive.

4MR. WACHOWIAK: Oh, that was your 5 question.6 MEMBER DENNING: Yes.

7 MEMBER WALLIS: Yes. It's down there.

8MR. WACHOWIAK: Water in the vent will be, 9 in the vertical vents, will be here.

10MEMBER WALLIS: But it will be driven down 11 eventually to the vent, won't they?

12 MR. WACHOWIAK: No, because the flow tap 13is through the PCCS into here. So it equalizes out 14 around here.

15Now it does fluctuate some going in and 16 out. What the TRACG analysis kind of shows is that 17 this will tend to burp if you will as it builds up 18some non-condensables. The heat transfer is a little 19bit less effective. The pressure goes up. It drives 20 the water column down and pushed the non-condensables 21 out and they kind of equalize out there. So this is 22 one of these what again is one of these self-limiting 23 processes such that the only heat that it can remove 24 is how much steam is going into it.

25 285 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 So in the end, we end up with a constant 1pressure that stays at that constant pressure 2 essentially forever. The only way that the pressure 3 in the containment goes down is due to heat transfer 4 through the side walls and outside that way. But no 5 excess heat is transferred out from the PCCS.

6MEMBER WALLIS: Well, conceivably, the 7 pressure could get to be less in the drywell than it 8 is in the wetwell.

9 MR. WACHOWIAK: If that happens for some 10 reason --11MEMBER WALLIS: Can you open the vent 12 valve?13 MR. WACHOWIAK: There's a vacuum breaker 14 here. Now this is from two separate drawings, but 15it's meant to show that here's the suppression pool 16 here. Air space of the suppression pool, we have this 17device here that's a vacuum breaker and there is three 18 of them.19MEMBER WALLIS: You send some non-20 condensables back out again.

21 MR. WACHOWIAK: You can send some non-22 condensables back out again and the whole process 23recycles or it doesn't. It's one of these things that 24you just can't tell for sure whether they're going to 25 286 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433open and reclose or not. The way that this is 1 designed is it's really like a garbage can lid.

2 MEMBER WALLIS: But it has to close to 3 make the sparger work, doesn't it?

4MR. WACHOWIAK: It has to be reclosed to 5make the sparger work. The way that it's designed 6though is that there's really no way, it's not like 7the vacuum breakers on MARK 1 that are kind of like a 8 hanging check valve. It's a positive direct action 9seating by gravity of this. It's a total vertical and 10it's arranged such that the failure mode is very 11unlikely for reseeding that vacuum breaker. However, 12this seeding surface is instrumented and if for some 13reason it's detected that it hasn't seeded right, 14there's a butterfly valve that is inside this thing 15here that can switch positions and isolate that vacuum 16breaker so that if it's leaking enough it's isolated 17on its own. If the containment pressure starts to go 18up again an indication of something gone wrong 19possibly with these, we would have procedures that 20would tell the operators to cycle through and try to 21close those to see if that's the problem. So we do 22 have a way of isolating the failed backing breaker.

23MEMBER SIEBER: You built the prototype to 24 test this, right?

25 287 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MR. WACHOWIAK: I believe so. We did.

1MEMBER SIEBER: I've seen pictures of it.

2 MEMBER WALLIS: Stay in the suppression 3 pool. Above the suppression pool.

4MEMBER ARMIJO: Are there a number of 5 those?6 MEMBER DENNING: It actually cools down.

7MR. WACHOWIAK: Three vacuum breakers, six 8 PCCS heat exchangers.

9MEMBER WALLIS: So all that gas is to go 10 in there.11 MR. WACHOWIAK: So that is an integral 12part of the containment system. We consider these a 13passive type component. Gravity is holding them in 14 place. It's a positive indication that it's the way 15that it's supposed to be. I kind of went through this 16 and it can be isolated.

17Let's look at the PCCS itself. There's 18really no way of failing this thing in the first 24 19 hours2.199074e-4 days <br />0.00528 hours <br />3.141534e-5 weeks <br />7.2295e-6 months <br />. It's open. It provides the heat transfer.

20The physical arrangement is what makes it work. So 21outside of the vacuum breakers there's really not much 22 in the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> that can happen here.

23However after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, somewhere before 2472, we need to have more water added in. There is --

25 288 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: Now wait a minute. There 1is no way for it to fail? Presumably there is some 2debris which can be carried around with the steam and 3get into this thing and block up the non-condensable 4 lines for instance.

5 MEMBER SIEBER: Twenty-four hours.

6MEMBER WALLIS: Lock up the condensate 7 drain with some debris which flies around and can get 8 up there.9MR. WACHOWIAK: I think debris was 10 addressed in the testing of the PCCS.

11 MEMBER WALLIS: Only fine debris would 12 probably block up that condensate line, wouldn't it?

13MEMBER SIEBER: That line is a pretty big 14 line, right?

15 MR. WACHOWIAK: Yeah.

16 CHAIRMAN APOSTOLAKIS: A big pipe.

17MEMBER WALLIS: But you said it's 18unreasonable to consider. Am I doing something 19 unreasonable?

20 MEMBER SIEBER: Again.

21 MEMBER WALLIS: Taboo?

22MR. WACHOWIAK: I guess maybe I choose my 23words improperly there. Maybe not unreasonable to 24 consider but --

25 289 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER WALLIS: If there were flying 1 debris you could in fact conceivably block something 2 that is essential to the operation of the PCCS.

3 MEMBER SIEBER: It could.

4MR. WACHOWIAK: Like I said, the aerosols 5were looked at in the test program for the PCCS and 6 that wasn't determined to be a failure.

7MEMBER WALLIS: It would have to be 8 particulates of some sort.

9 MEMBER MAYNARD: I think you created a 10 challenge with that statement there.

11MEMBER SIEBER: They were called HU, 12 highly unlikely.

13 MEMBER DENNING: What about molten --

14 MEMBER WALLIS: Are they all reasonable?

15 MEMBER SIEBER: Yes, right.

16MEMBER DENNING: What about molten 17material during the high pressure? No, that's later.

18MEMBER WALLIS: Well, it's latent debris.

19 Someone just left something around the containment.

20MR. WACHOWIAK: We have looked at debris 21like that, insulation and things like that, and I 22believe in the design there is a guard there to keep 23flying material in the LOCA situation like insulation 24and other things that would be expected during a 25 290 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433blowdown that could affect that. So that's been 1 addressed. It's the particulate fission products that 2I believe we'll find the answer to that in the test 3 report for the PCCS.

4We do have an automatic way considered.

5It's considered in our analysis. We do have an 6automatic makeup. The pool reactor for that refueling 7cavity in the laydown area that's for the steam dryer 8and separator for refueling purposes, that's all 9filled with water. Somewhere after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> before 72 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, those valves will open up pretty much based on 11level in the PCCS ICS pools providing enough water for 1272 hours0.0147 days <br />0.353 hours <br />0.0021 weeks <br />4.83996e-4 months <br /> worth of operation. Beyond that, we still 13have a connection to the firewater system that could 14add water there. FAPCS can add water. We could even 15make a connection to a hose station outside the 16reactor building and have a fire truck put more water 17 in there.18MEMBER SHACK: So considered in this case 19 means possible.

20MR. WACHOWIAK: Its automatic makeup.

21When I said considered here, I really mean what did we 22put in the fault trees when we did this analysis. So 23 we put this in. We really didn't put that in.

24 MEMBER WALLIS: You didn't put what in?

25 291 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 MEMBER DENNING: PCCS.

1MR. WACHOWIAK: Spontaneous failure of the 2 PCCS.3CHAIRMAN APOSTOLAKIS: Is there manual 4 action in the venting?

5 MR. WACHOWIAK: It's in there.

6 CHAIRMAN APOSTOLAKIS: And you said that 7 it was because -- when you do sensitivity analysis?

8MR. WACHOWIAK: In Revision 0 of the PRA 9we did not do that. In Revision 1 that we're 10 finishing up part of that chapter as we speak now, 11that's one of the considerations that we're doing in 12 there. We recognized that we missed that in --

13CHAIRMAN APOSTOLAKIS: So which one do we 14 have, Rick?

15MR. WACHOWIAK: You have Rev 0. We did 16 not give you Rev 1 of Chapter 11.

17 CHAIRMAN APOSTOLAKIS: Okay.

18MEMBER DENNING: How do we test the system 19and how frequently is it tested and how do you test it 20to make sure that it would operate, you know, that 21there isn't something that's happened during normal 22 operation that it's led to corrosion?

23MR. WACHOWIAK: During the outages, these 24 are part of the inspection program.

25 292 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER DENNING: They are inspected, but 1there's no testing possible. Is that true? Or how do 2 you -- You can't test them for function.

3MR. WACHOWIAK: No. At least not the 4 installed ones.

5MEMBER SIEBER: Sort of an inactive 6 passive system.

7MEMBER DENNING: This requires heat 8 condensed to make it really work.

9 MR. WACHOWIAK: That's correct.

10MEMBER WALLIS: Well presumably if there's 11any moisture in the drywell in normal operations and 12 it would very slowly set this thing off.

13MR. WACHOWIAK: Well, not really because 14there's an active drywell cooling system that provides 15much more steam condensation effect than this would be 16 subject to. So you wouldn't see it there either.

17 So when we go through the analysis, we 18find that the PCCS failure including the vacuum 19breaker portion of that is unlikely in 99 percent of 20 the core damage sequences.

21 MEMBER WALLIS: What does unlikely mean?

22 Is that 10

-5 or something?

23 MR. WACHOWIAK: In 99 percent.

24MEMBER WALLIS: The term unlikely doesn't 25 293 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 mean anything to me.

1 CHAIRMAN APOSTOLAKIS: No, in 99 percent 2 it's extremely unlikely.

3MR. WACHOWIAK: I've done the reverse on 4this one. It's not going to fail in 99 percent of the 5 cases.6MEMBER WALLIS: So essentially it's zero.

7 You mean it's essentially zero.

8MR. WACHOWIAK: There's a 0.1 failure rate 9 or 99 percent reliability.

10CHAIRMAN APOSTOLAKIS: No, no. It's not 11the same thing. In 99 percent of the sequences it's 12 extremely unlikely. That's what that means.

13MEMBER DENNING: That's what that means, 14 but is that what he means?

15 CHAIRMAN APOSTOLAKIS: Is that what you 16 mean? One percent is not extremely unlikely. That's 17 not -- You cannot mean that.

18MR. WACHOWIAK: Let me get to my next 19slide if it is what I think it is. It's this picture 20 here and I'll explain what I meant by that because 21the statement was accurate and I think we're all 22probably saying the same thing. So let's make sure we 23get to there. The way we quantified this containment 24 system of entry, remember we're coming in after 25 294 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433asking the deluge line and all the rest of those 1 things. For each of the different accident subclasses 2 that we would have that would affect things like vapor 3suppression function, this includes FAPCS also in case 4 there is an issue there.

5Things that would affect these, what the 6conditional failure probabilities of these headings 7would be, we made different accident subclasses and we 8take all the cut sets upon the sequences and add to 9those different accident subclasses and append these 10 functions and calculate what the subclass specific 11split fraction would be for each of these functions.

12So in 99 percent of our core damage sequences, these 13 numbers are like 10

-6 , 10-4 , 10-8.14CHAIRMAN APOSTOLAKIS: That's interesting.

15MEMBER WALLIS: That's what you're saying.

16MR. WACHOWIAK: Yes, in one percent of the 17 sequences this note here is about 0.7. 0.6, 0.7.

18MR. WACHOWIAK: So that's extremely likely 19 then.20MR. WACHOWIAK: And that's why I said in 2199 percent of the sequences it's extremely unlikely.

22In one percent of the sequences, we're probably going 23to get to a containment event. So what we would say 24there is that the reliability to overpressure 25 295 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 protection is about 99 percent.

1CHAIRMAN APOSTOLAKIS: And tomorrow you 2 will talk to us about the seismic effects.

3 MR. WACHOWIAK: Very briefly, yes.

4CHAIRMAN APOSTOLAKIS: But these numbers 5 don't change when you consider earthquake.

6 MR. WACHOWIAK: What we did for seismic 7was a seismic margins analysis and we only considered 8the safety related systems. So what we were attempting 9to prove with that is that all of our safety related 10functions would remain operable up to I think it was 11two times SSE or 2.4 times SSE, something to that. So 12 we really didn't get into what the degraded 13reliability of these systems would be in a seismic 14 event. So if that was your question, we didn't do 15that in the analysis. I wasn't really going to talk 16a lot about seismic. It's fairly -- It's a simple 17 margin.18MEMBER DENNING: In this one percent, what 19is it that makes them vulnerable? Is there some 20obvious aspect of that one percent of them that means 21 that you're --

22MR. WACHOWIAK: They're in high pressure 23 sequences. The reason you would end up having a high 24pressure sequence is basically because all of your DC 25 296 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 power has failed and amongst other things too, but 1 mainly they all involve no DC power. If you don't 2have DC power, we're relying on the operator action to 3provide that extra water to the PCCS. That's why we 4end up with a very high conditional failure 5 probability there.

6 MEMBER DENNING: We have 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to do 7 it.8MR. WACHOWIAK: Once again, we tried to do 9a screening analysis and we're trying not to overly 10rely on operator actions, but that tends to be what it 11 is and even the operator action that we have, we're 12not at the point yet in the design that we're sure 13that you can do that operator action in all cases with 14no DC power available because you have to get up into 15-- To locally operate that valve, you have to be 16somewhere that may not be a very nice place due to 17radiation to be to manually operate those valves if 18you're in that kind of a cinder accident. So we 19 really aren't taking much credit for the manual action 20 when you don't have all your DC power systems.

21Just to go through it, we solved all these 22for the different subclasses, some things off on the 23end, and that's where we come up with our input for 24the release rates or for the source terms. But in 25 297 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433general, jumping ahead of myself here, for the 1 containment failure probability due to 2overpressurization, it really comes down to that one 3 percent.4 Now one of the things that we want to talk 5about is what happens in the case where you lose the 6ability for the PCCS to operate. What happens if you 7lose containment heat removal? Just to get an idea of 8when the containment is going to be vented or when the 9containment is going to fail, we hypothetically said 10let's not have any containment heat removal from time 11 zero. We don't have any scenarios that get us there 12with any significant probability, but let's just look 13 at what happens if we start there.

14 We're seeing that it's more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> 15before you get to the point where the operators are 16going to consider that they would need to vent. Now 17let's move that into our scenario that we had was a 18one percent that was on the long term failure of the 19 containment heat removal. That failure is not going 20to create release here and we think it's more like out 21 here. So you still have another 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after that.

22So we're talking about a 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> before you really 23 have to vent and that's time that you have to figure 24out how to get more water up there and do something 25 298 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 else. So it's really a long term scenario in the 1containment overpressurization. It's not something 2 where we're going to get a failure right away.

3 Now what we did for calculating source 4terms, we took a much more conservative approach than 5that and looked at things earlier. The code for 6calculating the Level 3 doesn't really deal with those 7 long type of scenarios, so we added some of the 8 hypothetical on that side.

9So here are the results we come up with.

10 Bypass we believe is negligible. Overpressurization 11within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> is negligible. Overpressure later 12than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> can occur. Some high pressure sequences 13once again about one percent. There is mitigation 14 there. It would be a filtered release, but as we 15agreed up front on this project that we're just going 16 to call those releases.

17MEMBER DENNING: I'm sorry. Did you say 18that we're just going to call those releases? You're 19telling me that you would not take credit for removal 20 of iodine and things like that?

21 MR. WACHOWIAK: When we used it to 22 calculate the source term for the level three.

23 MEMBER DENNING: Yes.

24MR. WACHOWIAK: We factored in the vent 25 299 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433through the suppression pool. So we took the reduced 1source term, but what we're saying what's the 2reliability of the containment. We added that in to 3the one where it says we're going to have a release.

4 Not a big one, but --

5 Okay. Any questions on this?

6 CHAIRMAN APOSTOLAKIS: Thank you.

7MR. WACHOWIAK: Next we're going to have 8 Sid Bhatt talk about the offsite consequences.

9CHAIRMAN APOSTOLAKIS: You have to do this 10 for design certification?

11MR. BHATT: We did because I thought we 12wanted to get an idea of the thought process all the 13way and see what happens to the final situation and 14 how to --15 CHAIRMAN APOSTOLAKIS: Do they have to 16 submit a Level 3 PRA?

17MEMBER DENNING: This isn't the Level 3 18 PRA. It's a consequence analysis.

19 CHAIRMAN APOSTOLAKIS: What is Level 3 20 then?21 MEMBER DENNING: Well, site specific and 22 things like that.

23CHAIRMAN APOSTOLAKIS: No, but I'm 24 curious. I don't think it's required.

25 300 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: Are you going to prevent 1 him from presenting?

2 MEMBER DENNING: But it looks good.

3MS. CUBBAGE: I'd have to get back to you 4 on that.5MEMBER DENNING: But the results are 6 fantastic. That's the point they're going to make. So 7 why not make them?

8 COMM. MEMBER BRADY: You actually are 9 driven from the goal.

10MS. CUBBAGE: Someone had just mentioned 11that the severe litigation design alternative review, 12 this factors into that.

13MR. BHATT: Traditionally, whenever we had 14once upon a design like ABWR, we used to carry this 15 all the way to the end to see level one, level two and 16then probably get resuming certain code as you can see 17and resuming some numbers for the containment 18phenomenalogy event tree like CPET, what of that, 19serial accident phenomenon that you want to analyze 20and then also look into the systems, containment 21system and suppose they fail, how they all converge 22and they provide essentially some kind of a key 23information from the fault tree on the right hand 24side, some lump end states like bypass, like how to go 25 301 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 through those kind of categories.

1As you do the fanning process, it's 2important to kind of figure it out and say where are 3we are going to go from the offsite consequences for 4a review for a generic path that we do not have yet a 5site specific. So this is an attempt. So we created 6-- I will go through three parts, goals, what kind of 7process which we have been going through, it's nothing 8new, what are the results and how does it compare to 9 the goals we tried to look for.

10So we created three kind of goals which 11traditionally we have been using. One is the 12 individual risk and again we are looking near the 13vicinity of the power plant and we used the reference 14which is given from the National Safety Council 15 essentially defining some kind of a goal --

16 MEMBER KRESS: Is it basically the QHOs?

17MR. BHATT: Yes. So the second part of 18 this, it is also similar to that.

19MEMBER KRESS: Who are you going to go to 20 societal leaks? It doesn't fit my -- of society.

21MR. BHATT: Yes. Understood That's the 22 reason why I cannot put it in any other designation.

23 The to-debt context is comparable.

24MEMBER KRESS: It's still an individual 25 302 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 risk.1MR. BHATT: And this is an individual 2 risk.3CHAIRMAN APOSTOLAKIS: It's called 4 societal risk.

5 (Several speaking at once.)

6MEMBER KRESS: That's the reason why I 7 quit calling it that.

8MEMBER WALLIS: When it comes to something 9like less than one in a million for the societal 10 risks, less than that, isn't it?

11 MR. BHATT: Yes.

12 MEMBER SIEBER: Tom wants it to be --

13 MEMBER WALLIS: Less than 10

-6.14 MEMBER SIEBER: - less than 10

-6.15 (Several speaking at once.)

16MR. BHATT: And the third one is to create 17certain sources as you meet certain failures have 18occurred that's caused the core melt to come out. Now 19you do have sufficient productivity (PH) scenarios and 20then if it's released out from the plant in different 21situations, one way certain things are still there but 22under technical specification, it allows you to have 23 some kind of a controlled release.

24MEMBER KRESS: Where does that third goal 25 303 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 show up in the regulations?

1 MR. WACHOWIAK: It shows up in the URD.

2MEMBER KRESS: Oh, that's a URD provision.

3 We don't have it in the regulations.

4CHAIRMAN APOSTOLAKIS: What is it that we 5 don't have now?

6 MEMBER KRESS: That third goal.

7MEMBER SHACK: We just calculated in the 8 Environmental Impact Statement though.

9 MEMBER SIEBER: Yes.

10MR. WACHOWIAK: Rick Wachowiak from GE.

11 I believe the customers use it in their site --

12CHAIRMAN APOSTOLAKIS: Yes, but it's not 13 part of the QHO.

14MR. BHATT: So the whole process is trying 15to get an idea about what's the variation risk. When 16 you look into it from the point of view of the boxes 17are intended to kind of get a focus on what the 18synthesis is all about to kind of get an assessment 19and kind of gives you a sanity check. The inaccuracy 20or accuracy of the probabilistic risk assessment 21numbers will depend upon the upfront like CDF, CSCD, 22 things like that. So they are filtered in.

23 Also you would have to look into what kind 24of fuel was loaded into the core and for example what 25 304 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433kind of a cycle you are using. If you expose the fuel 1for a longer time, you have bigger fission inventory, 2things like that. So that one is calculated by the 3core entry point of view for ESPWR but at this point 4in the presentation, you say they are going to running 5 for a 24 month cycle.

6 In terms of the upper lefthand part, we 7already talked to you about the Level 1 PRA. We are 8 calculating CDF, looking into the cut sets and 9creating bins, defining what is the containment event 103 and the Level 2 type of probabilistic risk number.

11 So all that part provides a certain kind of release 12 frequency for those kind of release categories.

13Now if you know the release categories, 14then you say how are we going to calculate the detail 15fission product release to create a source term and 16then synthesize source term and release frequencies 17and using a computer code which has traditionally been 18 used to calculate the consequences.

19 MEMBER KRESS: Is that the EPRI version?

20MR. BHATT: Yes. So what happens is that 21curricularly was modified to actually look into the 22ESPWR essay (PH) features and was benchmarked as the 23track to have comparisons for the design base 24 accidents so that you can say when the accident 25 305 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433starts, at least the initial point also is okay.

1There is a separate report which we have provided. I 2think EPRI provided to NRC. Right? And you have 3 that.4So essentially it can also be done by some 5 other code track, mel code, etc. the release 6 fractions. So if you propagate this synthesis process 7there essentially you do have a source term associated 8with these different release categories. In this 9analysis we have 11 of them and for each end state of 10 the CETS for example or the release categories, the 11radionuclides were lumped into 12 different groups and 12then we looked into the consequences at the end of the 13 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and at the end of 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.

14MEMBER WALLIS: Well, the worst 15 consequences seem to be when the BiMAC system fails.

16 MR. BHATT: Yes.

17MEMBER WALLIS: And it makes a tremendous 18 difference.

19 MR. BHATT: Yes. Which one is that?

20MEMBER WALLIS: It makes a tremendous 21difference whether or not the deluge system in the 22 BiMAC works.

23MR. BHATT: Which slide are you looking 24 at?25 306 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER WALLIS: I'm just reading from my 1notes from reading the PRA document. I'm not looking 2at your slides at all. You're talking about release 3fractions and I said I noticed when I read the PRA 4document that they depended very much on whether or 5 not the deluge system in the BiMAC worked or not.

6 MR. THEOFANOUS: May I say something?

7 MR. BHATT: Yes, go ahead.

8 MR. THEOFANOUS: Of course it works.

9 MEMBER WALLIS: Yes, of course.

10MR. THEOFANOUS: That's why you put BiMAC 11 in.12MEMBER WALLIS: I know, but I notice how 13 important it is. It's extraordinarily important.

14 MR. WACHOWIAK: This is Rick Wachowiak 15 again. Just remember how we did this calculation.

16We said if the BiMAC fails, then we will have that 17 release. We did not try to say if the BiMAC fails 18what's the chance that we're going to have core 19retention on the floor without the BiMAC. That 20question wasn't asked and it wasn't answered. So you 21--22 MEMBER WALLIS: Bring to the surface.

23MR. WACHOWIAK: You can't necessarily 24infer that if BiMAC fails then the release is much 25 307 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 higher.1 MEMBER WALLIS: Right. I've just looked 2at the sequence and it says if BiMAC fails or BiMAC 3doesn't fail. The difference is so when does it 4 matter to any release.

5MEMBER DENNING: Does the BiMAC failure 6imply from this assumption that you don't get to 7 scrubbing the suppression pool?

8MR. WACHOWIAK: I'm not sure how that came 9out in Revision 0. In Revision 1, it makes it clear 10which ones with the releases from with the deluge 11lines are successful so we scrub versus the deluge 12 lines fail. So it's unscrubbed. So there is the 13distinction that's made. Once again, they are 14containment failures and the probabilities of those 15are low enough that they're really not driving this 16 answer. But once again, there is a difference there.

17MR. BHATT: So Division 1 has the complete 18story what we have gone through and also of the Level 19 2 which we used this for the OP and bypass scenarios.

20It also has the CPETS and the CSETS synthesis done and 21it goes through the end states which are considered 22here. There are 11 categories and I will go over those 23 quickly here too.

24But this is a short story. Then we can 25 308 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433come back to the point about the release frequencies.

1The processes where we again did use for the ALWR URD 2 times. In some cases, we are talking about generic 3one-part law so we had to use certain databases. So 4 population we used for the Sandia report which was -

5MEMBER KRESS: The Sandia side, they 6looked at a lot of sides. Did you chose one of those 7 or what?8MR. BHATT: One other thing is the 9 population density which was on a more convoluted 10 side. So here we kind of make things compounded from 11 the point of view of what might go bad and things like 12 that. It may not be realistic. We also for example 13I assumed there was no evacuation which again is 14pushing the limit. Then we did say that all this 15release are going to be happening at the ground level, 16not at the top level. One of the reasons why is 17because we are near the vicinity of the harbor. Now in 18case of a plume was released also as if it had no heat 19 content. This is kind of my field.

20MEMBER WALLIS: You believe there is 21 caloric theory.

22 MR. BHATT: No, this was --

23MEMBER DENNING: Based on what your 24 concern was.

25 309 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. BHATT: What it says is that, yes, we 1could push out a number. So basically when a generic 2calculation like that, we did zero, zero, a million, 3 things like that, just to kind of get an idea.

4Essentially what it does is that the plume is released 5 at the higher level and with the higher heat content 6it can propogate further and then you are trying to 7analyze some goals which are near the vicinity of the 8font then in that situation so this again is pushing 9 the limit.

10Then essentially the whole dose at a half 11of mile is a probably direct sentence (PH). The top 12 line one 10

-6 is a kind of a goal.

13MEMBER KRESS: That's for the atmosphere.

14MR. BHATT: That is a goal which we have 15 and the plots, there are two plots on this one, the 72 16hours and 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Essentially they are theoretical 17scale. The calculation numbers kind of has significant 18 margin.19MEMBER KRESS: Before you leave, the .25 20 sieverts, is that the 50 percent lethal dose?

21 MR. BHATT: No.

22 MEMBER DENNING: Oh, no. That's 25 rem 23 and this gets barely up to the point of health 24 effects.25 310 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MEMBER SIEBER: It's the first level of 1 detectability. So back on cell change.

2 MEMBER DENNING: Nobody's going to --

3 (Several speaking at once.)

4MR. BHATT: This again are the 5 requirements I am saying there.

6 MEMBER DENNING: Before you get off of 7 that, I think the place that goes into the coordinate 8 there, that's the core damage frequency of the 9component. Recognize that because what we're basically 10 looking at are things that are down to 1/30th of the 11core damage frequency. That's kind of where we're 12 going here.

13 MR. BHATT: Yes.

14 (Several speaking at once.)

15MEMBER KRESS: This is the SC curves that 16 we're talking about.

17MEMBER WALLIS: This is cumulative 18 probability consequence.

19 MEMBER KRESS: Yes.

20MR. BHATT: In terms of how, if you look 21at the bottom list, probably to what decimal numbers, 22but you throw out a basis and say this is what it is 23and then you try to compare them. Then the comparison 24 says that this is the goal which we set for the 25 311 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433example for the variation dose which is 10

-6 and the 24 1hour period case, the 72 hour8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> case. We do meet the 2goal but politically we say you can say yes. In terms 3of decimal number, I think it's kind of not that 4significant because we really do not know with that 5 decimal number.

6MEMBER SIEBER: Does that include iodine?

7 MR. BHATT: Yes. The 12 groups.

8 MEMBER SIEBER: Right.

9MEMBER WALLIS: These are individuals 10 risks. So even if there are a million people affected 11 you would still in some cases --

12 MR. BHATT: So for the site specific --

13MEMBER WALLIS: It's a cycle. You have a 14million people. Multiplied by a million, you still 15need more. So it's pretty close to a million people.

16 MEMBER SIEBER: It's an accumulated dose 17 as opposed to a health impact.

18CHAIRMAN APOSTOLAKIS: The number of 19people is a pattern because it's expressed in terms of 20 the individual.

21 MEMBER WALLIS: Yes, that is individual.

22 (Several speaking at once.)

23 MEMBER WALLIS: Even if it is that you 24modify by a hundred thousand, you would still be 25 312 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 within the goal.

1 MEMBER DENNING: The nice thing is you 2 don't kill anybody.

3MEMBER SIEBER: You never get the levels 4at that distance that are sufficient to cause cellular 5 change. Now it's below the so-called emergency dose 6 that radiation workers are allowed to get.

7MEMBER WALLIS: So you can put this in the 8 middle of a city?

9 MEMBER SIEBER: Not my city.

10 MR. BHATT: It's not impossible.

11 (Several speaking at once.)

12MR. BHATT: Now when you see the site 13specific application in the PRA where you would see a 14 certain case like there could be in Washington, D.C.

15 or New York City and there is a plant and what kind of 16the detail whatever, at that time probably this thing 17should be revisited. For example, one of our customers 18is already doing that. In those situations, we would 19 probably get the more realistic.

20CHAIRMAN APOSTOLAKIS: So how is it? Did 21you do any uncertainty analysis here? What are we 22 talking about?

23 MR. BHATT: Uncertainty analysis --

24CHAIRMAN APOSTOLAKIS: You did it for the 25 313 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433 core damage.

1 MEMBER SIEBER: It's just a number.

2MR. BHATT: This is just a number. For 3the core damage frequency, the numbers would be from 4 one PRA and --

5 MEMBER WALLIS: So what's the number --

6CHAIRMAN APOSTOLAKIS: There are no 7 uncertainties after that.

8 MEMBER DENNING: Max.

9 MR. BHATT: We have propagated the 10 uncertainty. You are right. We have --

11 CHAIRMAN APOSTOLAKIS: So this 3.7 10

-11 , 12 how high could it be?

13 MR. WACHOWIAK: This is Rick Wachowiak 14with GE. Let me try to answer that in the best way we 15can because No. 1 we did not try to propagate any 16 uncertainty. So the Level 1 input is point estimate.

17But if you remember how we did the Level 2, we looked 18at bounding parameters to get us to the different 19release bins. We think we're on the upper edge for 20calculating the frequency, translating the Level 1 21 frequency into the release bin frequencies.

22 Then when we took the representative 23source term, we really looked at what would be the 24upper limit source. I don't want to say bounding 25 314 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433because if we have two cases, one was 10

-10 and one was 1 10-13, we tended to look at the 10

-13 case. But we 2intended to use more bounding values to get the actual 3source terms. We put them in here and then max'ed 4those as Monte Carlo stuff for all the rest of the 5 things.6 So did we specifically do an uncertainty 7 analysis? The answer is no. What uncertainty 8analysis would be applicable to this? It tends to be 9 more on the Level 1 feeding into the Level 2 that 10would get us there and then we'll use bounding beyond 11 that. So it's an interesting question. I'm not sure 12that if we think about the Level 1 uncertainty of 13knowing one order of magnitude at higher infrequencies 14and propagating that to here that it would really 15 change much of the answer.

16MEMBER MAYNARD: Well, you probably had 17most of the uncertainties there covered by the 18 conservatism that you get built into the parameter 19analysis like the assumption that you allow your 20 container to contain things that you don't have.

21 MR. WACHOWIAK: That would be -- in fact, 22 you could probably make --

23MEMBER MAYNARD: Or significantly delayed.

24 MR. WACHOWIAK: So it's a mixture.

25 315 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.(202) 234-4433WASHINGTON, D.C. 20005-3701(202) 234-4433MR. BHATT: So essentially this whole 1story with tremendously surprising results bartered 2across the -- missed frequencies where coming low and 3then as you add this to reach down the slow sterns 4helps. But that's partly the purpose of setting some 5goals for the Level 1 PRA and Level 2 PRA and trying 6to come out. So essentially this shows tha PRA tests 7 help.8 (Several speaking at once.)

9 CHAIRMAN APOSTOLAKIS: Is this it?

10 MR. BHATT: I think so.

11CHAIRMAN APOSTOLAKIS: Any other 12 questions? Okay. This concludes the day's 13 presentations. I would like to thank the speakers. It 14 was very informative. So we'll see some of you 15 tomorrow morning. Thank you. Off the record.

16(Whereupon, at 5:10 p.m., the above-17 entitled matter was concluded.)

18 19 20 21 22 23 24 25