ML12089A272
| ML12089A272 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 03/29/2012 |
| From: | Atomic Safety and Licensing Board Panel |
| To: | |
| SECY RAS | |
| Shared Package | |
| ML12089A267 | List: |
| References | |
| RAS 22113, 50-247-LR, 50-286-LR, ASLBP 07-858-03-LR-BD01 | |
| Download: ML12089A272 (35) | |
Text
ENT000369 Submitted: March 29, 2012 Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION
Title:
Entergy Nuclear Vermont Yankee
,Docket Number: 50-271-LR; ASLBP No. 06-849-03-LR Location: Newfane, Vermont DOCKETED USNRC August 15, 2008 (4:00pm)
Date: Monday, July 21, 2008 OFFICE OF SECRETARY RULEMAKINGS AND ADJUDICATIONS STAFF Work Order No.: NRC-2296 Pages 694-879 NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers 1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005 (202) 234-4433
800 1 JUDGE KARLIN: -- so people can hear.
2 MR. STEVENS: I've done quite a few 3 calculations with 6909 going back a few years as part 4 of the review process included. And we did the - we 5 calculated CUFens for the VY, all nine locations that 6 were evaluated for environmental fatigue.
7 And the CUFens using 6909 are lower than 8 those reported by Entergy in the testimony for all 9 nine locations. And that would support what -- Mr.
10 Fair's prior testimony or the ACRS comment that he i1 made earlier.
12 JUDGE REED: Mr. Stevens, in fact, the 13 only change was the Fens, not the CUFs, is that 14 correct?
15 MR. STEVENS: No, sir. We used --
16 JUDGE REED: Did you also change the CUFs 17 in this --
18 MR. STEVENS: Yes. We used 6909 in its 19 entirety, so we used the curves, calculated Fens in 20 accordance with that document, and also CUFens as a 21 product of the two.
22 JUDGE REED: Okay. So we are looking at 23 a result that was attained by changes both in the 24 environmental effects as well as how you did the basic 25 CUF analysis.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
801 1 MR. STEVENS: Yes, sir.
2 JUDGE REED: Okay.
3 JUDGE WARDWELL: How long does that 4 analysis take you to do?
5 MR. STEVENS: Approximately four hours.
6 JUDGE WARDWELL: Thank you.
7 JUDGE KARLIN: And let's just clarify.
8 Dr. Reed's questioh I think was reasonably -- was well 9 taken, and I think the clarification is you're saying 10 NUREG-6909 involves more than just changing the Fen 11 calculation, is that correct?
12 MR. STEVENS: That's correct.
13 JUDGE KARLIN: And it involves what other 14 changes?
15 MR. STEVENS: It involves calculating 16 fatigue with revised fatigue curves, compared to 17 those --
18 JUDGE KARLIN: In air.
19 MR. STEVENS: Correct.
20 JUDGE KARLIN: Right.
21 MR. STEVENS: Compared to the ASME code 22 fatigue curve.
23 JUDGE KARLIN: And what -- maybe we can 24 just ask what the values you derived --. what were the 25 values for the nine -- what, nine locations, six NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
802 1 locations, nine pieces of equipment?
2 MR. STEVENS: Okay. The nine items -- and 3 I'll report the values using 6909, and these are 4 CUFens. Item 1 was the RPV shell bottom head 5 location, CUFen .0263. Item 2 is the RPV shell at 6- shroud support location. CUFen was .2637. Item 3 was 7 the feedwater nozzle blend radius. CUFen was .2175.
8 Item 4 was the recirculation RHR Class 1 9 piping return T' CUFen .4151. Item 5 was the 10 recirculation inlet nozzle forging. CUFen .1921.
11 Item 6, recirculation inlet nozzle safe-in, CUFen 12 .0152.
13 Item 7, recirculation outlet nozzle 14 forging, CUFen .0278. Item 8, core spray nozzle 15 forging blend radius, CUFen .0524. Item 9, feedwater 16 Class 1 piping, CUFen .1350.
17 JUDGE KARLIN: And what was the highest 18 CUFen calculated by that -- in thatanalysis?
19 MR. STEVENS: That would be Item 4, 20 recirculation, RHR Class 1 piping return T of .4151.
21 JUDGE KARLIN: Okay.
22 JUDGE WARDWELL: How did this analysis --
23 how was this analysis performed in regards to the 24 three sets of analyses that were presented as part of 25 the testimony in regards to the basic initial NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
803 1 calculation performed in the application, those that 2 were done for the refined analysis, and those that 3 were done for the confirmatory analysis? Analyses, I 4 should say.
5 MR. STEVENS: We started with the -- all 6 of the calculations that Entergy has done over the 7 past year. And for the three nozzles that are a part 8 of *the testimony -- the recirculation outlet nozzle, 9 the core spray nozzle, and the feedwater nozzle -- we 10 used the refined analyses.
11 And the analysis -- the calculations we 12 did here started with the stress results that went 13 into the fatigue calculation. So we used -- you know, 14 the majority of the analysis remained unchanged. We 15 just took the stresses that fed into the fatigue 16 analysis. We replaced the fatigue curve and 17 recalculated Fens. The very tail end of the analysis 18 was reperformed.
19 JUDGE WARDWELL: So let me make sure I 20 understand this correctly. You used the stress 21 analyses that -- the most recent stress analyses for 22 a given component based on whether or not the -- it 23 was -- the most recent was done as either a -- the 24 original analyses, refined analyses, or confirmatory 25 analyses.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
804 1 MR. STEVENS: We did not use the 2 confirmatory calculation for the feedwater nozzle, our 3 reason being to be consistent with these comparisons.
4 JUDGE WARDWELL: Strictly the refined for 5 the three nozzles and then they're basically --
6 MR. STEVENS: That's correct.
7 JUDGE KARLIN: Does that mean that the 8 Green's function issue was eliminated, or still -- as 9 I understand it, we have three analyses essentially 10 that you submitted, that Entergy submitted. First, 11 it's the initial analyses, right? With the 12 application. Is that correct?
13 MR. STEVENS: That's correct.
14 JUDGE KARLIN: Then, there was the 15 reanalysis in September and December of '07, right?
16 MR. STEVENS: Correct.
17 JUDGE KARLIN: And that was for all seven 18 locations or pieces of equipment -- all nine, I'm 19 sorry. And then, there was a confirmatory analysis 20 with regard to the feedwater nozzle, right?
21 MR. STEVENS: Correct.
22 JUDGE KARLIN: So that's the terminology 23 I'm going to use -- the initial analysis, the 24 reanalysis for all of them, and then the confirmatory 25 analysis for the feedwater nozzle. And as I hear what NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
805 1 you're saying, you used the approach of the 2 reanalysis, and then you applied-NUREG-6909 to that, 3 is that right?
4 MR. STEVENS: That's right.
5 JUDGE KARLIN: Now -- well, then, there 6 may be a concern about that because the whole reason 7 that the confirmatory analysis was done was because, 8 as I understand it, the staff raised concerns about 9 the simplification of being caused by using simplified 10 data in the Green's function. Are you with me?
11 MR. STEVENS: Yes, sir.
12 JUDGE KARLIN: Was that -- is that problem 13 still inherent in your reanalysis of these?
14 MR. STEVENS: Well, I'd just like to 15 clarify that I don't consider that to be a problem 16 with those analyses. But the presence of the Green's 17 function is in the three evaluations of the nozzles in 18 these results I reported to you.
19 JUDGE WARDWELL: And just for 20 clarification, is it fair to say that the.reason you 21 did that is just to compare it to using 6909 versus 22 5783 -- what's the other NUREG number -- approach?
23 You did it for comparison purposes rather than any 24 final calculation of record.
25 MR. STEVENS: Yes, sir. I believe at NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
806 1 least two of the parties -- that would be Entergy and 2 NRC Staff -- have made some statements that the newer 3 regulatory guide and associate NUREG generally provide 4 lower results on a CUFen standpoint than the ones used 5 for license renewal. And so these were done to 6 support those statements further.
7 JUDGE WARDWELL: Okay.
8 JUDGE REED: Mr. Stevens, I need to 9 understand better how these calculations are actually 10 performed. We've kind of jumped into the middle here, 11 and I need to come back more to the beginning. You 12 work for SIA, is that correct?
13 MR. STEVENS: That's correct.
14 JUDGE REED: And you personally did all of 15 these calculations.
16 MR. STEVENS: Not -- no, sir. I was -- I 17 did a few of them, and I supervised the staff that did 18 the calculations.
19 JUDGE REED: But you are intimately 20 familiar with the methodology that goes into 21 calculating a CUFen number for a particular component.
22 MR. STEVENS: Yes, sir.
23 JUDGE REED: So you know from beginning to 24 end exactly how that calculation proceeds and what 25 assumptions are made.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
807 1 MR. STEVENS: Yes, sir.
2 JUDGE REED: Could you give us a brief 3 discourse on how that's done? I don't want a half-4 hour tutorial. I want a five-minute description of --
5 take a particular transient or a set of transients 6 that are analyzed and tell us exactly how you 7 calculate and arrive at a CUFen number.
8 MR. STEVENS: Okay.
9 (Pause.)
10 I'm trying -- I'm going to look my 11 testimony up, so I can keep this as brief as possible.
12 I apologize for the delay.
13 JUDGE REED: No, that's fine. You might 14 start by talking about the kinds of transients that 15 affect, say, a feedwater nozzle and how those lead to 1.6 stresses, and then how you calculate the stresses, 17 then how you determine what the maximum allowable 18 stresses are, so that you -- and I don't know how to 19 do these calculat-ions, so I don't want to lead you too 20 far.
21 MR. STEVENS: Okay.
22 JUDGE REED: But --
23 MR. STEVENS: I'll try and be brief but 24 descriptive here. So we have to -- we have to collect 25 all of the loadings for a particular component we're NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
808 1 evaluating, and those would come from --
2 JUDGE REED: Can you speak just a little 3 louder?
4 MR. STEVENS: Okay. Sorry. We have to 5 collect all of the loadings for a particular component 6 in order to analyze it from a stress and fatigue 7 standpoint. We would typically collect those loadings 8 from the design specifications for the component by 9 the vendor or manufacturer. So we would collect 10 geometry and we would construct a model that would 11 allow us to calculate stresses for all loadings. In 12 this case, a finite element model is an industry 13 standard, so we would model that component geometry --
14 JUDGE REED: So if I may interrupt, these 15 are static loadings on a particular component. You 16 calculate a stress tensor for that or a stress field?
17 MR. STEVENS: Correct.
18 JUDGE REED: Okay. So the issue of 19 transients has not yet come into it.
20 MR.. STEVENS: Not yet.
21 JUDGE REED: Okay.
22 MR. STEVENS: As a part of the loadings, 23 though, there are thermal transients which are loads 24 that vary with time, and we would also use the finite 25 element model to evaluate stresses as a function of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
809 1 time for those loadings.
2 JUDGE REED: So you take a series of 3 snapshots during a transient, and all of which have 4 different loadings is --
5 MR; STEVENS: That's --
6 JUDGE REED: -- and calculate the 7 stresses?
8 MR. STEVENS: Yes.
9 JUDGE REED: Okay. I'm with you so far.
10 MR. STEVENS: So during a thermal 11 transient, we would have a stress history versus time 12 for that transient. And then, knowing how pressure 13 and other loadings vary during that transient, we 14 would calculate those stresses and combine them. This 15 is all a linear, elastic analysis, so we would combine 16 them by superposition. And we would get a stress --
17 a total stress history for the component from those 18 analyses.
19 JUDGE REED: For a particular transient, 20 you develop a stress history, meaning the stress as a 21 function of time.
22 MR. STEVENS: That's correct.
23 JUDGE REED: Okay.
24 MR. STEVENS: And then, we have 20 25 transients, we would repeat that process 20 times and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
810 1 come up with 20 stress.histories.
2 JUDGE REED: Now, when you say "20 3 transients," you don't mean the same thing happening 4 again and again. You mean 20 different kinds of --
5 the plant responding to something happening.
6 MR. STEVENS: That's correct.
7 Different --
8 JUDGE WARDWELL: Are these plant-specific?
9 MR. STEVENS: Yes. So you are correct.
10 There would be 20 different transients of a different 11 type. Each of those transients would occur a 12 different number of times. So the quantity of those 13 20 transients would be specified and would be 14 different.
15 JUDGE REED: And some of the transients 16 are more severe than others?
17 MR. STEVENS: That's correct.
18 JUDGE REED: And would lead to larger 19 stresses?
20 MR. STEVENS: That's correct.
21 JUDGE REED: But they may occur fewer 22 times, so you may have a smaller transient that occurs 23 much more frequently that is going to contribute 24 relatively more to the ultimate answer, is that 25 correct?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
811
- 1. MR. STEVENS: That's correct-.
2 JUDGE REED: Okay.
3 MR. STEVENS: So now, once we have the 4 stress history, and we would feed that into a fatigue 5 analysis, and the fatigue analysis first tells us how 6 to take those stress histories and link them together 7 or combine them in such a way that we can count number 8 of stress fluctuations on a component. Stress 9 fluctuations are what lead to fatigue.
10 JUDGE REED: So it's not stress, but 11 fluctuations in stress.
12 MR. STEVENS: That's correct. It requires 13 a stress fluctuation to contribute any kind of 14 fatigue.
15 JUDGE REED: Okay.
16 MR. STEVENS: So, and the ASME code gives 17 us guidelines and methodology for doing all of this 18 analysis. And, in particular, counting cycles, how 19 that's done in a conservative fashion, because in a 20 fatigue analysis of the component we don't necessarily 21 know the order of occurrence of the events. So the 22 methodology assumes the worst possible occurrence by 23 pairing the extreme stresses together.
24 So when we go through this counting 25 process, we take the highest extreme, the highest peak NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
812 1 stress with the lowest low stress and pair-these off 2 to get these stress fluctuations.
3 JUDGE REED: So that's a conservative 4 assumption you are making? Is that the point?
5 MR. STEVENS: Yes, sir.
6 JUDGE REED: So you don't take the actual 7 stress history of the transients. You somehow 'take 8 the high stress and the low stress and pair them up 9 to -
10 MR. STEVENS: Well, when we when we --
11 that's correct.
12 JUDGE KARLIN: May I ask a question in 13 this -- on any given transient, you are talking about 14 the stress load that occurs during a transient, right?
15 MR. STEVENS: Yes.
16 JUDGE KARLIN: Do you have monitors that 17 are telling you what the stress load is on this 18 particular valve or nozzle or outlet? How do you know 19 where are you getting the data that tells you how 20 much stress they have?
21 MR. STEVENS: The stress comes from the 22 finite element analysis of that component.
23 JUDGE REED: And what code is being used 24 to do that analysis?
25 MR. STEVENS: We use the ANSYS finite NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
813 1 element code.
2 JUDGE REED: And that is a commercially-3 available code?
4 MR. STEVENS: Yes, sir.
5 JUDGE REED: And ANSYS stands for?
6 MR. STEVENS: I don't recollect.
7 JUDGE KARLIN: Could you spell the 8 acronym?
9 MR. STEVENS: A-N-S-Y-S.
10 JUDGE KARLIN: Okay.
11 JUDGE REED: And that code is in wide use 12 or very narrow use throughout the industry or --
13 MR. STEVENS: It's widely used in the 14 nuclear industry for finite element analysis.
15 JUDGE REED: And what can you say about 16 the accuracy of the code, the benchmarking of the 17 code? How confident are you that it is giving you 18 correct numbers?
19 MR. STEVENS: The code is extensively 20 benchmarked. It comes with verification manuals that 21 -- where analyses are run and compared to theoretical 22 or hand solutions and checked for accuracy. It is all 23 controlled and developed under a 10 CFR 50 Appendix B 24 quality assurance program.
25 And when we bring that code into our house NEAL R.. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
814 1 we also have to adopt it as part of that program and 2 do the extensive checking and verification in 3 accordance with those verification manuals, and then 4 ensure that our use is consistent with those -- the QA 5 program and those checks.
6 JUDGE REED: So when you generate a finite 7 element mesh for ANSYS, do you do a mesh refinement 8 study?
9 MR. STEVENS: Generally not on a case-by-10 case basis. This is one of the areas where it does 11 require analyst judgment and experience to do it 12 properly. We have done those as a part of our company 13 extensively for other reasons, but'we don't do it on 14 a case-by-case basis.
15 JUDGE REED: Would it not be possible that 16 as you refined the mesh you would find a little local 17 area where the stress is much-higher than it would be 18 calculated with a much coarser finite element grid?
19 MR. STEVENS: That's possible. I don't 20 think that that's -- I would not agree with that 21 assessment for the models used for Vermont Yankee.
22 JUDGE REED: But you.can't rule it out, 23 since you haven't done the mesh refinement study.
24 MR. STEVENS: No. We can only rule it out 25 based on our experience.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
815 1 JUDGE REED: So you can't actually be
.2 certain that higher stresses might not be generated if 3 you refined the meshes in your calculations.
4 MR. STEVENS: Generally, that kind of an 5 artifact. would come near discontinuities.
6 JUDGE REED: Yes.
7 MR. STEVENS: And the locations that we're 8 evaluating here don't have those kihds of 9 discontinuities.
10 JUDGE REED: Okay.
11 MR. STEVENS: So I'm confident that you 12 -would not experience a significant change in stress 13 with your suggestion.
14 JUDGE REED: So is it your point that you 15 believe that you have resolved the stress field 16 reasonably accurately with the finite element meshes 17 you're using in your calculations?
18 MR. STEVENS: Yes, sir.
19 JUDGE REED: Okay. I hope we can proceed 20 with how you do the calculation now. And I'm sorry 21 that 22- (Judges confer.)
23 JUDGE WARDWELL: Dr. Hopenfeld, do you 24 have any major objections or arguments with the way 25 Mr. Stevens just described the approach that they use?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com v
816 1 DR. HOPENFELD: Yes.
2 JUDGE WARDWELL: Is that your 3 understanding of hw.0 the approach was used?
4 DR. HOPENFELD: Yes. Yes, Your Honor. I 5 understand exactly what he's talking about.
6 JUDGE WARDWELL: And is that your 7 understanding of how it was performed when you 8 supplied your testimony?
9 DR. HOPENFELD: Yes. Yes, it is.
10 JUDGE WARDWELL: Okay. That's all I 11 needed to know. We'll get to you in a moment. Just 12 wanted to make sure we're all in agreement.
13 DR. HOPENFELD: Well, make sure you pick 14 me up.
15 JUDGE KARLIN: No. But I'm trying to --
16 when you said yes, do you have disagreements with what 17 he just said, or are you on board?
18 DR. HOPENFELD: I have no disagreement 19 with what he said.
20 JUDGE KARLIN: Okay.
21 DR. HOPENFELD: But that's only part of 22 it.
23 JUDGE KARLIN: Yes, right. We'll get to 24 that.
25 DR. HOPENFELD: That's not the whole NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
817 1 story.
2 JUDGE KARLIN: Okay. Fine. I just want 3 to be clear on that. We'll get the rest of the story.
4 DR. HOPENFELD: Okay.
5 JUDGE WARDWELL: Go ahead. I just wanted 6 to --
7 JUDGE REED: Well, we were partway through 8 this issue of exactly how you take a series of 9 transients and calculate a CUFen number.
10 MR. STEVENS: Correct. So once I have all 11 of these different stress histories for, arbitrarily 12 talking, 20 transients here, the ASME code tells me 13 how to start to pair these together to get fluctuating 14 stresses that would contribute to fatigue. So I have 15 to combine these transients. As I said, it does it in 16 a conservative way that it takes the most -- the 17 extreme stresses and pairs those and uses those 18 fluctuating stresses into a fatigue analysis.
19 And very, very simplistically, if I took 20 the highest stress and the lowest stress that would be 21 a stress range, and alter -- a possible alternating 22 stress range that that component may have been exposed 23 to --
24 JUDGE REED: For that one kind of 25 transient.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
818 1 MR. STEVENS: For that one pairing of two 2 extremes, which could be from two different 3 transients. Remember, I'm going after extremes to get 4 the most conservative fluctuation of stress that 5 component would have seen.
6 JUDGE REED: I'm failing completely to 7 understand why you picked the high stress from one 8 transient and the low stress from a completely 9 different transient that presumably occurs at some 10 other time in the plant's history.
11 MR. STEVENS: Well, because that the 12 component would see the stresses caused by both of 13 those transients at some point in its life. So in 14 terms of a fluctuating stress --
15 JUDGE REED: These could be days, months
- 16. apart, though.
17 MR. STEVENS: Could be years apart. But 18 you see that a typical design analysis does not know 19 the order of events. So if it's analyzing 20 20 transients, and it has no particular knowledge on the 21 order those events could occur, and if it puts them 22 next to each other as if they occurred minutes apart, 23 that would be the most conservative.
24 JUDGE REED: And that's what you do.
25 MR. STEVENS: That's what we do. And it's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
819 1 a conservative way of stating the extreme stress 2 fluctuations that that component will go through, not 3 knowing the order of events ahead of time.
4 JUDGE REED: I'm struggling to understand.
5 I want to ask you at some point -- you've got 20 6 different kinds of transients.
7 MR. STEVENS: Correct.
8 JUDGE REED: Presumably all sorts of 9 things that happen to the plant -- changes in power 10 level, turbine trips, whatever., And these are all 11 stressing this particular component, and they are all 12 occurring with different frequencies, some occurring 13 once a year, some occurring every several years, some 14 occurring every decade. Am I right, in this --
15 MR. STEVENS: That's correct.
16 JUDGE REED: So what I want to understand 17 is how you ultimately take all of those various 18 transients, calculate some sort of utilization 19 factors, and then cumulate them up into a CUF.
20 MR. STEVENS: Let me try again. Let's 21 start with the stress history for one transient. So 22 I have one transient that has temperature during the 23 time, and I have calculated stresses for that 24 transient. So I have a stress versus time plot I can 25 make for that one transient.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
820 1 JUDGE REED: Okay.
2 MR. STEVENS: Now I have 19 other 3 transients that I do the same thing. And what I'm 4 going to do is I'm going to take the stress history 5 for each of those other 19 transients, I'm going to 6 tag it on to the end of the stress history for 7 transient number 1. Now what I've got is this very, 8 very long stress history versus time that has all 20 9 transients attached together to each other.
10 JUDGE WARDWELL: These aren't 20 actual 11 transients that occurred at Vermont Yankee. These are 12 20 different types of transients that are generically 13 experienced at a boiling water reactor similar to 14 might be at Vermont Yankee.
15 MR. STEVENS: These are specified by the 16 designer for the plant.
17 JUDGE WARDWELL: Okay.
18 MR. STEVENS: So I now have a long stress 19 history, stress versus time, that represents 20 20 transients attached together, one after another.
21 Now, when I put that history together, I 22 made a big assumption on the order of those 23 transients. If I would have done transient 1, 24 transient 2, transient 3, then that stress history 25 represents that the order of occurrence of those 20 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
821 1 transients was in that order -- time order I put them 2 in.
3 JUDGE WARDWELL: And the time is not a 4 real time, it is the linkage of all of the transients, 5 and they are attached based on the link that the 6 transient occurs. Then, you immediately start the 7 next transient after that-one is over with, is that 8 correct?
9 MR. STEVENS: Yes, that's correct.
10 JUDGE REED: So you are putting the plant 11 through all of its paces basically one right after the 12 other with no intervening period of six months of 13 stable operation, is that --
14 MR. STEVENS: That's correct.
15 JUDGE REED: Okay.
16 JUDGE KARLIN: Let me just -- may I ask 17 basic questions, where we're going here. I was 18 concerned by the answer you gave to Dr. Wardwell's 19 question. Are these -- when you were talking about 20 the 20 transients, are these actual transients that 21 occurred at Vermont Yankee that you were basically 22 saying, okay, there are 20 transients and here is how 23 much stress -- transient 1 gave X stress, transient 2 24 gave Y stress.
25 We're going to add them all together, and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
822 1 here's how much stress Vermont Yankee has experienced 2 over the 35 years it has been operating. Is that what 3 is going on, or are these some sort of assumed 4 stresses that are theoretical?
5 MR. STEVENS: These are -- the transients 6 that we're calculating these stresses for are 7 specified by the designer, they are part of the design 8 basis, and they have been shown to be conservative --
9 very conservative definitions, compared to the actual i0 transients that Vermont Yankee --
11 JUDGE KARLIN: Yes, because here's where 12 I'm going. Let me just be real basic. Metal fatigue, 13 I'm focusing on the NUREGs and the CUFens and that 14 sort of thing, metal fatigue. As I understand it, 15 metal fatigue -- you take a paper clip, it's made of 16 metal. You bend it once, twice, you bend it 100 times 17 and then it breaks. And you do this experiment a 18 dozen times and it breaks at a hundred.
19 With the hundredth bend, 90-degree bend, 20 it breaks So now you know that that piece of metal 21 is going to break at the -hundredth twist of 90 22 degrees. That's my simplistic way of thinking about 23 this. This is the ASME curve that says at a hundred 24 breaks in air -- bends at 90 degrees in air this paper 25 clip will break.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
823 1 Now you go and say, "All right. I've got 2 a paper clip here. I bent it 37 times. And I bent it 3 90 degrees some of those times, and I bent it 45 4 degrees some of those times. I bent it 200 degrees 5 some of -- I bent it a lot of different ways." And L 6 total up those 37 bends, and I say, "Okay. I have 7 used up this amount of stress." And if I -- and now 8 I know how long it is going to take for me to get to 9 the hundredth time when it breaks. Is that what's 10 going on?
11 MR. STEVENS: That's simplistically what 12 is going on.
13 JUDGE KARLIN: And so when we are talking 14 about the calculations of the 20 transients, which are 15 bends, the times that there was stress imposed upon' 16 this, are you talking about the -- how we get to the 17 hundredth, the theoretical one hundred, or how do we 18 get to the actual history of this plant?
- 19. MR. STEVENS: The transient severity is 20 analogous to your how far did you bend your paper 21 clip?
22 JUDGE KARLIN: Right.
23 MR. STEVENS: Because if you bent your 24 paper clip 90 degrees and got 100 times until failure, 25 you would get a different number if you only bent your NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
824 1 paper clip 45 degrees.
2 JUDGE KARLIN: Right.
3 MR. STEVENS: It's a measure of the load 4 on that component. These transients .all have 5 different severity. One transient may have a 45-6 degree bend analogy, another one a 90-degree bend 7 analogy, another one a 10-degree bend analogy.
8 JUDGE KARLIN: Right.
9 MR. STEVENS: They all have different i0 severities. They are specified by the designer to be 11 very conservative, so that the designer can do a 12 fatigue design of the plant prior to construction.
13 And those have been shown through numerous studies 14 throughout the industry over the history to be 15 conservative, and those are the definitions that were 16 used in the analyses for Vermont Yankee.
17 JUDGE KARLIN: So the American Society for 18 Mechanical Engineers and the people who designed this 19 particular boiling water reactor figured out for each 20 type of metal in that reactor how many bends, how many 21 stresses it would take before it broke. Is that 22 right?
23 MR. STEVENS: Not quite.
24 JUDGE KARLIN: Okay.
25 MR. STEVENS: What they would do is an NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
825 1 analysis in accordance with ASME code to demonstrate 2 that for those loadings the component will not show 3 unacceptable results. And in this particular case 4 that's a usage factor less than one.
5 JUDGE KARLIN: Right.
6 MR. STEVENS: That does not necessarily 7 mean failure.
8 JUDGE KARLIN: Yes. So you have a hundred 9 -- if you had -- so the calculation is this thing will 10 you can bend it a hundred times 90 degrees and it 11 will break. Now, you've actually only bent it 70 12 times at 90 degrees, so your CUFen factor is .7, 13 right?
14 MR. STEVENS: That is correct.
15 JUDGE KARLIN: And you've got 30. bends 16 left before it is going to break or before the 17 calculation says it will break, right?
18 MR. STEVENS: Well, again, the calculation 19 isn't indicating breakage. It's --
20 JUDGE KARLIN: Failure of some kind.
21 MR. STEVENS: -- the acceptance criteria 22 that's used, and there is margin on that acceptance 23 criteria.
24 JUDGE KARLIN: Okay. I understand.
25 MR. STEVENS: We work with safety factors NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
826 1 and other things that prevent breakage.
2 JUDGE KARLIN: Okay. Sorry for the 3 digression. Go ahead.
4 JUDGE REED: Please continue.
5 MR. STEVENS: Back to our -- we now have 6 our 20 transients linked together. I have made an 7 assumption in how those transients link together, the 8 order they occur. And what the ASME code tells us to 9 do as well, in order to capture the most conservative 10 order, we're going to go through this history and 11 we're going to take the largest extremes and say that 12 they define a stress fluctuation that the component 13 will go through.
14 So perhaps of this entire stress history 15 let's say the highest stress experienced was 16 transient 1, and the lowest stress was transient 18.
17 i would start by pairing those, and that would define 18 a stress range, the maximum stress range that could be 19 conceivable for that component to see throughout that 20 stress history.
21 When I make that choice it is like taking 22 transient 18 and moving it next to transient 1, 23 assuming that order, adjacent order.
24 JUDGE REED: Well, aren't you, 25 furthermore, assuming that these two transients occur NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.Gom
827 1 with the same frequency? Suppose one-is an annual 2 thing and one is a weekly thing. How do you justify 3 putting them together in your analysis when you know 4 that one occurs 50'times more often than the other?
5 MR. STEVENS: That would be something I 6 would like to take advantage of, because -- but this 7 way is .conservative and it's making that assumption.
8 It is very bounding.
9 JUDGE REED: Okay. I understand that.
10 JUDGE WARDWELL: This is just defining the 11 angle at which you're bending the paper clip.
12 MR. STEVENS: It's trying to -- back to 13 your paper clip analogy, it is trying to just say that 14 based on the stress history, and if I could reorder 15 things in any order I wanted to, the worst extreme 16 your paper clip may go through is 90 degrees, so I 17 want you to take that as the top cycle, the most 18 severe cycle to evaluate. But it's --
19 JUDGE WARDWELL: The right paper clip 20 would -- the fact that one transient occurs now and 21 another one occurs a week later isn't necessarily -
22 I'm trying to find a word that -- with a paper clip 23 you could bend it once, and then a week later bend it, 24 it's still going to know -- it's got a memory -- it's 25 going to know you have bent it. So it's not NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
828 1 completely unreasonable to move 18, because this is an 2 arbitrary number, up next to the other one, because 3 that range of stresses has been felt by that metal.
4 MR. STEVENS: That's correct. But it 5 ignores that there could have been events in between 6 those two. But in principle what you're saying is 7 correct.
8 JUDGE KARLIN: Why does it make any 9 difference, I mean, whether you put them together --
10 you have had 37 events at this -- on this paper clip, 11 you know, of bending at different angles. I don't 12 care whether they are all done, what order they are 13 done, at the end of the day 37 events have occurred,
- 14. this much stress has occurred, and you have used up 15 this amount of your CUFen before you get to 1, right?
16 MR. STEVENS: If transient 15 actually 17 occurred after transient 1, and had a much smaller 18 stress associated with it than transient 18, then the 19 cycle the component saw on that day is much smaller 20 than the one I assumed by putting transient 18 next to 21 1.
22 JUDGE -KARLIN: So you're assuming a 23 conservative assumption, the worst case as it were.
24 MR. STEVENS: It's the worst possible --
25 JUDGE KARLIN: Higher stress. So it would NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com v
829 1 be more consumption of the-- what's available 2 fatigue, what available fatigue there is before it 3 breaks or fails.
4 JUDGE WARDWELL: I hate to beat this, but 5 180 degrees as opposed to 90 degrees is --
6 MR. STEVENS: That's a very good analogy, 7 yes.
8 JUDGE REED: So if we move on now, you've 9 taken and done this pairing, what do you do with all 10 of the other transients?
11 MR. STEVENS: Well, I go -- okay.
12 JUDGE REED: You pick the worst two, put 13 together, and you put those together --
14 MR. STEVENS: One occurrence of the worst 15 two.
16 JUDGE REED: One occurrence.
17 MR. STEVENS: If you can imagine now, 18 remember in the beginning I told you that each 19 transient has a specific number of occurrences 20 associated with it.
21 JUDGE REED: Yes.
22 MR. STEVENS: So, if you will, each peak 23 and valley of this long stress history --
24 JUDGE REED: Yes.
25 MR. STEVENS: -- has a number of cycles NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
830 1 associated with it that's equal to the number of 2 transients. If I take the portion that's transient 1, 3 and if my designer said 100 of those could occur 4 during the life of the plant, then that stress history 5 -- every point on it-- could occur 100 times, almost 6 analogous to now I've got this third dimension kind of 7 on this graph. I've got a stress versus time history, 8 and each point has a different number of occurrences 9 associated with it -- the number of transients that 10 were specified by the designer.
11 So if I take transient 1 and transient 18, 12 I have now taken one occurrence of each of those 13 events. If I had 100 occurrences of transient 1, I 14 have 99 of those points left to deal with. So you can
- 15. imagine, if I go through this history and take the 16 highest and the lowest, and cross one site -- one 17 occurrence of each of those off, and take the next 18 highest or the next lowest, and I repeat.that process 19 until all occurrences of all transients have been 20 consumed, I'd get a nice stress array right from 21 largest stress range to lowest with the number of 22 occurrences next to it.
23 JUDGE REED: I see that.
24 JUDGE WARDWELL: So far, the three NUREGs 25 of interest haven't even come into play yet, is that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
831 1 correct?
2 MR. STEVENS: Yes.
3 JUDGE WARDWELL: This is all stress 4 analysis.
5 MR. STEVENS: That's correct.
6 JUDGE KARLIN: So in that respect it's the 7 ASME code as opposed to the Fen part of the analysis?
8 MR. STEVENS: That's'correct.
9 JUDGE KARLIN: Okay.
10 MR. STEVENS: So now I have a spectrum of 11 stress range and cycles applied to that stress range.
12 So now I have what's required to calculate fatigue.
13 I'll stop here for one --
14 JUDGE WARDWELL: To calculate what?
15 MR. STEVENS: Fatigue usage.
16 JUDGE WARDWELL: Fatigue.
18 JUDGE WARDWELL: CUF.
19 MR. STEVENS: Now, this represents 90, 95 20 percent of the work.
21 JUDGE WARDWELL: How long does it take to 22 do this -- person-hours, labor-hours?
23 MR. STEVENS: Just as an example, the 24 confirmatory calculations for Vermont Yankee feedwater 25 nozzle took about three weeks dedicated. approximately NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
832 1 three individuals full-time.
2 JUDGE' KARLIN: So let me understand. The 3 confirmatory analysis for one nozzle, the feedwater 4 nozzle, took three people three weeks of work?
5 MR. STEVENS: That's correct.
6 JUDGE KARLIN: It took nine weeks of time, 7 nine --
8 MR; STEVENS: Nine man-weeks.
9 JUDGE KARLIN: Nine man-weeks.
10 MR. STEVENS: You must understand that 11 that includes all the quality assurance checks and 12 documentation proper filing.
13 JUDGE KARLIN: Well, how did you do all of 14 them in four hours?
15 MR. STEVENS: No.
16 (Laughter.)
17 MR. STEVENS: This is the point I was 18 going to make. What's left took us four hours. So 19 I'll describe what's left next.
20 JUDGE WARDWELL: Before you do that, Dr.
21 Hopenfeld, is everything Mr. Stevens has said to date 22 your understanding of how they did it when you did 23 your review and filed your testimony?
24 DR. HOPENFELD: With respect to the 25 specific numerical analysis, I am not an expert in NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
833 1 stress numerical analysis.
2 JUDGE WARDWELL: I'm not asking you any 3 opinion. I'm asking you whether or not, when you were 4 reviewing this for NEC and filed your testimony, is 5 everything that Mr. Stevens said consistent with what 6 you thought they had done?
7 DR. HOPENFELD: Yes, yes.
8 JUDGE WARDWELL: That's all we need right 9 now.
10 DR. HOPENFELD: Oh, yes. I will --
11 JUDGE WARDWELL: Thank you. Just want to 12 make sure he hasn't told any 13 DR. HOPENFELD: Even now, I don't have any 14 issue with what he said.
15 JUDGE WARDWELL: And there isn't a 16 misunderstanding 17 DR. HOPENFELD: No, no, no. I have no --
18 I am absolutely in agreement with it.
19 JUDGE WARDWELL: So you don't contest --
20 DR. HOPENFELD: No.
21 JUDGE WARDWELL: -- the basic approach 22 that they have said they did.
23 JUDGE KARLIN: And the answer to that is 24 -- Dr. Hopenfeld, you don't contest that approach.
25 DR. HOPENFELD: I have -- no, this has NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com