ML22027A333
| ML22027A333 | |
| Person / Time | |
|---|---|
| Issue date: | 11/17/2021 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| Abdullahi, Z, ACRS | |
| References | |
| NRC-1751 | |
| Download: ML22027A333 (158) | |
Text
Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Reactor Safeguards Thermal-Hydraulic Phenomena Subcommittee Docket Number: (n/a)
Location: teleconference Date: Wednesday, November 17, 2021 Work Order No.: NRC-1751 Pages 1-76 NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers 1716 14th Street, N.W., Suite 200 Washington, D.C. 20009 (202) 234-4433
1 1
2 3
4 DISCLAIMER 5
6 7 UNITED STATES NUCLEAR REGULATORY COMMISSIONS 8 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 9
10 11 The contents of this transcript of the 12 proceeding of the United States Nuclear Regulatory 13 Commission Advisory Committee on Reactor Safeguards, 14 as reported herein, is a record of the discussions 15 recorded at the meeting.
16 17 This transcript has not been reviewed, 18 corrected, and edited, and it may contain 19 inaccuracies.
20 21 22 23 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
1 1 UNITED STATES OF AMERICA 2 NUCLEAR REGULATORY COMMISSION 3 + + + + +
4 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 5 (ACRS) 6 + + + + +
7 ACCIDENT ANALYSIS: THERMAL HYDRAULICS SUBCOMMITTEE 8 + + + + +
9 WEDNESDAY 10 NOVEMBER 17, 2021 11 + + + + +
12 The Subcommittee met via Video 13 Teleconference, at 9:30 a.m. EST, Jose March-Leuba, 14 Chairman, presiding.
15 COMMITTEE MEMBERS:
16 JOSE MARCH-LEUBA, Chair 17 RONALD G. BALLINGER, Member 18 VICKI BIER, Member 19 DENNIS BLEY, Member 20 CHARLES H. BROWN, JR. Member 21 VESNA B. DIMITRIJEVIC, Member 22 GREG HALNON, Member 23 DAVID PETTI, Member 24 JOY L. REMPE, Member 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
2 1 ACRS CONSULTANT:
4 DESIGNATED FEDERAL OFFICIAL:
5 ZENA ABDULLAHI 6
7 ALSO PRESENT:
8 STEVE M. BAJOREK, RES 9 PETER J. YARSKY, RES 10 TAREK ZAKI, RES 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
3 1 C O N T E N T S 2
3 NRC Staff Introductory Remarks (Kim Webber) 7 4 TRACE Assessment Against KATHY Test Data 5 (Pete Yarsky) 8 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
4 1 P R O C E E D I N G S 2 9:32 a.m.
3 CHAIR MARCH-LEUBA: The meeting will now 4 come to order. This is a meeting of the ACRS Accident 5 Analysis: Thermal Hydraulics Subcommittee. I am Jose 6 March-Leuba, Chairman of the Subcommittee.
7 Because of COVID-19 concerns, this meeting 8 is being conducted remotely. Members in attendance 9 are Vicki Bier, Vesna Dimitrijevic, Greg Halnon, Joy 10 Rempe, and I know we're having technical difficulties 11 with Dave Petti and Dennis Bley, which -- they will be 12 joining us shortly.
13 (Simultaneous speaking.)
14 MEMBER BIER: Ron Ballinger looks like 15 he's on.
16 MEMBER BALLINGER: Yeah, I'm here as well.
17 CHAIR MARCH-LEUBA: Oh, I -- sorry. Yeah.
18 So Ron Ballinger is, and Greg Halnon is. So today's 19 topic is a presentation by NRC staff documenting a 20 series of oscillatory flow tests performed on a 21 prototypical BWR fuel bundle simulated under stable 22 conditions.
23 The staff will present the results of 24 their analysis and how to incorporate those results in 25 future trace code calculations. We have not received NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
5 1 any requests for public comments, but we will have an 2 opportunity for spur-of-the-moment comments at the end 3 of the meeting. If you are using the phone line link, 4 you will need to use star-6 to unmute yourself.
5 The ACRS was established by a statute and 6 is governed by the Federal Advisory Committee Act, 7 FACA. As such, the Committee can only speak through 8 its published letter reports. Any comments raised by 9 members today are their individual opinions.
10 The ACRS section of the U.S. NRC public 11 website provides our charter, bylaws, agendas, letter 12 reports, and full transcripts from the open portions 13 of all full and Subcommittee meetings, including the 14 slides presented there.
15 The Designated Federal Official today is 16 Zena Abdullahi. A transcript of the meeting is being 17 kept. Therefore, speak clearly and state your name 18 for the benefit of the court recorder. Please keep 19 the microphone on mute when not in use, and don't use 20 video feed to minimize bandwidth problems -- except 21 maybe the presenter.
22 I apologize, but I have a personal 23 conflict that I cannot get out of at around 12:00 24 p.m., when we are supposed to finish this meeting.
25 However, if we extend beyond that time, Member Rempe NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
6 1 has agreed to serve as the Subcommittee Chair and 2 close the meeting.
3 At this point, let's request NRC staff to 4 start with a presentation, with introductory remarks 5 from Kim Webber.
6 Kim, you have the floor.
7 MS. WEBBER: All right. Great. Thank you 8 so much.
9 Good morning, Chairman March-Leuba and 10 ACRS members. My name is Kim Webber. I'm the 11 Director of the Division of Systems Analysis. I'm 12 really pleased to be here to talk with you today about 13 a topic that we believe has a lot of interest by the 14 Committee.
15 Over the last several years, NRR sought 16 assistance from the Office of Nuclear Regulatory 17 Research, or RES, in studying important phenomena 18 surrounding BWR instability to support the licensing 19 of plants seeking to operate within the MELLA+
20 expanded domain.
21 Accordingly, my staff's performed numerous 22 studies, analyses, and conducted experiments to better 23 understand how postulated transience without scram in 24 the MELLA+ domain could lead to adverse fuel 25 conditions. They've presented results of those NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
7 1 analyses before the ACRS Power Uprate Subcommittee and 2 full Committee in the past.
3 Today, Dr. Pete Yarsky will present some 4 of his background information, along with the staff's 5 thorough assessment of TRACE against full-scale 6 integral test data from the KATHY Facility. Following 7 his presentation, Dr. Steve Bajorek will provide some 8 information regarding a nondimensionalized approach to 9 analyze the KATHY data.
10 And now I'd like to turn the presentation 11 over to Pete.
12 MR. YARSKY: Thank you, Kim.
13 I'm Dr. Pete Yarsky from the research 14 staff, and I'll be presenting on the TRACE assessment 15 against the KATHY test data. But to start, I would 16 like to give a little bit of background about the 17 motivation behind the KATHY tests and to describe the 18 KATHY test facility and sort of the two variety of 19 tests that we conducted.
20 For many members of the Subcommittee, this 21 will be a bit of a review. And the new material to 22 present today is the comparison of these data to TRACE 23 calculations, and we'll wrap up with the conclusions.
24 So, first --
25 (Simultaneous speaking.)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
8 1 MEMBER REMPE: Peter, this is Joy. I hate 2 to interrupt you, but I tried to wait till the end of 3 the slide. I had a question that's kind of a holdover 4 from prior discussions on this topic.
5 There has been a lot of interest by 6 vendors who are not involved in the actual KATHY test, 7 as well as university professors, in having access to 8 the data obtained from these tests for their own 9 research or their own desire to try and show that 10 their models work well. What is the status of the 11 data?
12 MR. YARSKY: Well, some of the data are 13 considered sensitive and are classified under SUNSI 14 data. So we have not been able to release 100 percent 15 of the data to the public. To the extent possible, we 16 have published the data and our findings in a series 17 of papers and a publicly available NUREG report.
18 And I think at the end of this 19 presentation, we'll be talking about a nondimensional 20 approach that would provide yet a further avenue to 21 make our analyses and our data more publicly 22 available.
23 MEMBER REMPE: Great. Thank you.
24 MR. YARSKY: Yeah. Thank you.
25 And again, just for some background NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
9 1 information, the MELLA+ expanded operating domain 2 represents an expansion of allowable power and flow 3 operating conditions and is shown by the MELLA+
4 boundary on this power/flow map. And what this 5 expanded operating domain would allow is operation of 6 a BWR plant at a higher power level and a lower flow 7 rate or a higher power-to-flow ratio.
8 With this expanded operating domain, what 9 it would allow is, under certain conditions, if there 10 is a dual recirculation pump trip, the plant 11 trajectory following such trip would result in 12 conditions of very high power-to-flow ratio as the 13 plant achieves a natural circulation condition.
14 If we postulated an ATWS event, such as a 15 turbine trip without bypass, this turbine trip without 16 bypass would result in a pressure pulse and a trip of 17 the recirculation pumps and a loss of extraction steam 18 for the feed water heater cascade.
19 Such an ATWS could be expected to yield 20 unstable conditions with large-amplitude power 21 instability and that this could occur early, and the 22 operators will react to this condition by activating 23 the standby liquid control system to inject boron into 24 the vessel, as well as by lowering the reactor water 25 level to adjust not only the gross core power level NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
10 1 but also the inlet core subcooling to try and 2 stabilize the reactor.
3 The research staff has performed extensive 4 analyses of these kinds of events, and this is showing 5 a typical transient reactor power response to a 6 turbine trip without bypass ATWS instability. And, as 7 shown early in the event, there's a power pulse in 8 response to the turbine trip, which creates the back 9 pressure, and the void collapse resulting from that 10 leads to the power pulse shown here.
11 After that pulse, there then is a steady 12 increase in the reactor power as the reactor is 13 responding to an increase in inlet subcooling because 14 the extraction steam has been isolated from the feed 15 water heater cascade.
16 Eventually, the reactor becomes unstable, 17 and oscillations begin to grow. They grow to large 18 amplitude, and there's an unstable phase before 19 operator actions become sufficient to damp those 20 oscillations and ultimately shut down the reactor.
21 During these oscillations, it's possible 22 for the reactor power oscillation itself to become 23 quite complex. Shown here is a map of 3D bundle 24 powers, indicating in one case that we analyzed a 25 bimodal oscillation where the reactor power is not NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
11 1 only oscillating in terms of its gross magnitude, but 2 also, it's shifting from one side of the core to the 3 other side of the core in a regional-type oscillation.
4 What this implies is that it's possible 5 for there to be very large-amplitude power 6 oscillations in the individual bundle, and those 7 oscillations could be larger than what's implied by 8 the oscillation in the core power itself.
9 In our TRACE analysis --
10 MEMBER REMPE: Peter, this is Joy again.
11 I'm sorry to bother you, but apparently there's been 12 some issues, still, Jose, with Dennis Bley getting in.
13 I see that the public line -- apparently, Corradini 14 just got in -- has the problem with -- I mean, Dennis 15 is texting Mike, Scott --
16 (Simultaneous speaking.)
17 CHAIR MARCH-LEUBA: Dennis is in.
18 MEMBER REMPE: Okay. Great. I just 19 wanted to make sure that -- especially --
20 (Simultaneous speaking.)
21 MEMBER REMPE: -- line wasn't open.
22 CHAIR MARCH-LEUBA: Yeah. I don't know 23 with Petti. I think Petti's in, too. So --
24 MEMBER REMPE: Yeah. Apparently, Dave 25 also said that Mike Corradini and the public line NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
12 1 folks got in. So that was my biggest concern, too.
2 We were having trouble with the public line.
3 So, anyway, go ahead, Peter. Sorry.
4 MR. YARSKY: Okay. No problem.
5 What I'm showing here is a figure from 6 that same TRACE calculation. And what we predicted 7 using TRACE was that there would be a period of cyclic 8 dryout/rewet followed by a temperature excursion. And 9 that temperature excursion would be exacerbated by the 10 bimodal nature of the oscillations. And in this case, 11 we predicted that the temperature could exceed 2,200 12 F, indicated on this figure with a line at 1,478 13 Kelvin.
14 So, to go sort of through this fuel heat-15 up and temperature excursion, using TRACE, we 16 predicted this mechanism. The oscillation magnitude 17 will increase during the transient, and the fuel will 18 initially undergo cyclic dryout/rewet during the 19 oscillation period.
20 But as that oscillation magnitude grows, 21 the rewet period of the cycle becomes insufficient to 22 remove all of that heat. So you'll dry out and you'll 23 heat up a little bit, but when you rewet, that rewet 24 portion isn't enough to bring you all the way back 25 down to where you started in temperature. And, as a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
13 1 result, the temperature will slowly ratchet up during 2 the cyclic dryout/rewet period.
3 And we predicted that, eventually, the 4 temperature will reach the minimum stable film boiling 5 temperature, and the cladding surface will lock into 6 film boiling, and once there, there'll be a subsequent 7 temperature excursion, and that this mechanism could 8 lead to potential fuel damage.
9 So, in short, MELLA+ operation exacerbates 10 the consequences of ATWS. And we predicted that under 11 ATWS-I conditions, the cladding surface may fail to 12 rewet, leading to fuel damage, and that in practice, 13 it's difficult to ensure core coolability if some 14 portion of the fuel may become damaged, particularly 15 in cases where you're looking at complex power 16 oscillation contours.
17 For instance, you may have a rotating mode 18 of the instability, which could subject a large 19 fraction of the fuel to these kinds of conditions. So 20 this predicted heat-up mechanism was very important 21 for us to study and was the subject of an experimental 22 program at the KATHY Facility.
23 The KATHY Facility is a full-scale bundle 24 test facility primarily used for critical heat flux 25 testing. It's full reactor pressure, and it tests a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
14 1 full-scale electrically heated fuel bundle. One of 2 the features that makes the KATHY Facility unique is 3 that it can be reconfigured in such a way as to supply 4 either forced circulation, or a downcomer can be, 5 essentially, valved in to allow the facility to 6 operate in natural circulation mode.
7 And in addition to having that capability 8 to operate in natural circulation mode, the KATHY 9 Facility also incorporates a control module called 10 SINAN. And the SINAN module performs calculations 11 based on certain sensed conditions, such as inlet flow 12 rate, and can calculate void fraction and use that in 13 a reactivity kinetics program to simulate void 14 reactivity and Doppler feedback, allowing the test 15 loop to serve as a simulated reactor where the 16 feedback is simulated by this external control system 17 that can adjust the voltage on the electrically heated 18 components.
19 We selected this facility for those unique 20 capabilities, and we wanted to design a heater rod 21 assembly that would be representative of the operating 22 fleet. And here in the NRC fuel bundle that we 23 designed, we wanted to make sure it had all of the 24 features of modern BWR fuel. So this includes part-25 length rods, water rods, and spacers that are typical NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
15 1 of modern BWR fuel assemblies.
2 We have a somewhat typical radial power 3 distribution where we can use rods that have different 4 mechanical designs that allow different power 5 peakings. And in this experiment, we conducted our 6 tests with bottom-skewed axial power shape. This is 7 generally more limiting from a stability perspective.
8 So, in our experimental work, we first 9 wanted to, of course, verify or study the fuel heat-up 10 mechanism that we predicted during oscillatory power 11 and flow conditions. And we wanted to do so under 12 conditions that were most typical of ATWS-I scenarios.
13 A follow-up to that, of course, was to 14 assess and validate TRACE against these data. We 15 performed extensive testing in the failure to rewet 16 conditions in the Kathy test loop in December of 2016.
17 When we conducted these tests, there were two 18 varieties of tests that we conducted.
19 The first variety are tests that we did 20 without using the SINAN simulated feedback. And in 21 these tests, the way they were conducted was to 22 steadily and slowly increase the bundle power till we 23 were first able to induce instability, and then we 24 kept going. We kept increasing the power until we 25 were able to go through the phase of cyclic NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
16 1 dryout/rewet and ultimately until we drove the 2 assembly to the point of fuel heat-up.
3 We conducted a second type of test where 4 we used the SINAN feedback module to simulate 5 reactivity feedback. And in this case, instead of 6 adjusting the power, we slowly increased the strength 7 of the feedback to induce instability and to 8 subsequently induce heat-up.
9 So here's a sample of one of those tests 10 without feedback. And here, we're showing that the 11 bundle power is slowly increased in steps until a 12 failure to rewet is observed. So this is showing the 13 power. What we are showing here is the flow rate so 14 that even once we induce the instability, we're able 15 to grow the amplitude of the flow oscillation by 16 increasing the power.
17 And, eventually, we drive the assembly 18 into a failure-to-rewet condition. So this is what 19 sets up the fuel temperature excursion. In the TRACE 20 calculations, the predictive mechanism dictates the 21 fuel heat-up occurs once the cladding surface fails to 22 rewet. And this failure to rewet occurs in the TRACE 23 predictions once the temperature exceeds the minimum 24 stable film boiling temperature.
25 What we did in sort of a preliminary NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
17 1 analysis based on the data was record the maximum 2 thermocouple temperature observed prior to any failure 3 to rewet. And this would be an indication of the 4 highest temperature achieved before the cladding 5 surface fails to rewet.
6 So he was giving an idea. This is the 7 same test, the Test 407.1. And, first, this is 8 showing on the left-hand side of the figure successive 9 dryout/rewet cycles. Of course, this is fully 10 consistent with what TRACE predicts. But, also, you 11 can observe the ratcheting mechanism that TRACE also 12 predicted and now we're observing in the tests.
13 Eventually, there is the temperature 14 excursion. This is what we say is the failure to 15 rewet. Instead of allowing this temperature excursion 16 to occur, though, if the temperature increase is 17 observed over a full period of the oscillation, that's 18 how we diagnose a failure-to-rewet condition.
19 And to protect the experiment, the reactor 20 -- not the reactor power. The bundle power is reduced 21 to protect the experiment. And so that's why you see 22 the temperature turn around in this figure. But we're 23 able to show that we have the temperature excursion 24 and the failure to rewet.
25 CHAIR MARCH-LEUBA: So, Pete, can you go NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
18 1 back to that slide? At time 1095, we see a plateau of 2 the temperature, and then a little bit later, you see 3 a completely -- that clearly is the shutdown. What 4 happened at 1095 again?
5 MR. YARSKY: So, around 1095, we do a step 6 reduction in the bundle power. So it's not as though 7 the entire bundle power is reduced to zero. We step 8 it down by a significant fraction --
9 (Simultaneous speaking.)
10 CHAIR MARCH-LEUBA: Like roughly 10 11 percent or 20 or --
12 MR. YARSKY: I would say roughly 20 13 percent, in that neighborhood.
14 CHAIR MARCH-LEUBA: Yeah. What that 15 shows, in my mind, is that whatever is happening 16 thermohydraulically on the film is maintained even 17 with lower power.
18 MR. YARSKY: Yeah, for a short period.
19 And then there is the rewetting.
20 (Simultaneous speaking.)
21 MR. YARSKY: You can see around 1098, 22 right? But there is that period where it's still in 23 film boiling before it rewets, and that's why it's 24 relatively flat.
25 I just wanted to take a second to say, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
19 1 instead of fully dropping the power down to zero, the 2 reason why it's only reduced a bit is so that we can 3 keep the loop hot for the next test.
4 MEMBER HALNON: So this is Greg Halnon.
5 Without the reduction, we would expect the temperature 6 to continue to increase?
7 MR. YARSKY: Yes. Without the power 8 reduction, the excursion would continue. But it would 9 likely continue to the point of damaging the test.
10 MEMBER HALNON: Yeah. So is there a 11 threshold that you know of -- I mean that you drop the 12 power down below a known threshold, or is it just 13 enough to get it to start decreasing?
14 MR. YARSKY: A lot of this is sort of 15 based on the experience of the owners to know, maybe 16 we do it by 20 percent or 25 percent or 10 percent.
17 MEMBER HALNON: Okay.
18 (Simultaneous speaking.)
19 MEMBER HALNON: -- fuel for what's 20 required to protect their equipment.
21 MR. YARSKY: Exactly.
22 MEMBER HALNON: Yeah. Okay. I got it.
23 Thanks.
24 MR. YARSKY: Yeah. But for our purpose, 25 we wanted to study this heat-up mechanism. And if the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
20 1 excursion persists over a full period, then we know 2 that the cladding surface is failing to rewet. We 3 know that we're on this trajectory. So the experiment 4 is still confirming the mechanism even though we're 5 not driving it all the way to fuel damage. We know 6 that it will head in that direction. We know it's on 7 that course.
8 MEMBER HALNON: Well, it's a steep line 9 either way.
10 MR. YARSKY: Right.
11 MEMBER HALNON: I mean, we're talking five 12 seconds or less than five seconds.
13 MR. YARSKY: We know that we're in a bad 14 neighborhood, right? So we terminate the test. But 15 I think that we've learned from the test the condition 16 under which we've entered this trajectory.
17 So here's -- for a separate test. This is 18 Test 20703 on Rod 87 -- is that the peak for highest 19 temperature achieved before the excursion may occur 20 several periods before the excursion itself. And so, 21 in our preliminary analysis, we wanted to pick up what 22 the highest temperature was for which we observed 23 rewetting.
24 And this could be an indication of the 25 temperature at which the surface would fail to rewet, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
21 1 given the nature of how we conducted these tests with 2 sort of a slow, step-wise approach in either power or 3 reactivity feedback.
4 Here's showing a zoom of that same data 5 showing that the peak temperature may occur a few 6 periods before the excursion.
7 I want to now show a sample test with 8 feedback. This is showing the dynamic power during 9 Test 801.02. And you can see the power oscillating.
10 This power oscillation not only shows how the 11 amplitude increases as we strengthen the simulated 12 reactivity feedback but also how the average power 13 increases as the bundle is destabilized.
14 In our initial assessment of the data, the 15 first thing we did, as I said, was to record these 16 failure-to-rewet temperatures. And we found that the 17 maximum was around 700k; the minimum's around 600k, 18 with an average in the neighborhood of 650k.
19 This temperature corresponds relatively 20 well with the homogeneous nucleation plus contact 21 temperature. Now, this homogeneous nucleation 22 temperature is the lowest temperature, theoretically, 23 at which liquid will spontaneously nucleate into 24 vapor. So it represents a sort of theoretical minimum 25 for the equivalent of the minimum stable film boiling NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
22 1 temperature.
2 We wanted to compare our measured failure-3 to-rewet temperatures to this theoretical temperature 4 and just see how they line up. When we did this, we 5 also compared the failure-to-rewet temperature to a 6 variety of other models of Tmin, such as Groeneveld-7 Stewart, Henry Shumway, and Peterson and Bajorek.
8 And what we found are that those models 9 that use the homogeneous nucleation or homogeneous 10 nucleation plus contact temperature really are right 11 in the neighborhood of our observations for the 12 failure-to-rewet temperature.
13 When we did a preliminary comparison of 14 TRACE to sort of post-process data, we found that if 15 we were to dial in the minimum stable film boiling 16 temperature in a value range between 600 and 638 17 Kelvin, that we were able to get fairly good agreement 18 between the experimental results and the TRACE 19 results.
20 However, dialing that minimum stable film 21 boiling temperature in the TRACE model up even a 22 little bit more into, say, the 630 or 640 Kelvin range 23 would completely preclude the occurrence of the 24 failure to rewet.
25 So the analysis of the fuel consequences NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
23 1 can be quite sensitive to assumptions about the 2 minimum stable film boiling temperature as it relates 3 to the predictive fuel heat-up mechanism.
4 So, as an interim approach based on our 5 initial analysis of the data and as a conservative 6 approximation, we recommended, and in our own analyses 7 have adopted, the approach of setting the minimum 8 stable film boiling temperature equal to the 9 homogeneous nucleation temperature plus contact 10 temperature.
11 And we've been successfully using this 12 approach to analyze ATWS-I from MELLA+ BWRs to support 13 licensing determinations. However, we have taken the 14 opportunity now to perform a more detailed assessment 15 of TRACE through a more thorough study of the KATHY 16 experimental results.
17 CHAIR MARCH-LEUBA: Oh, yeah, I wanted to 18 ask, this Tmin homogeneous plus contact -- have you 19 benchmarked it against all of the experiments, and 20 does it fit all of them? Or are there any outliers?
21 MR. YARSKY: Oh, in terms -- okay. So 22 what we have done in the TRACE assessment activity is 23 to run TRACE calculations using this approach and 24 comparing it to the data. And I think that that will 25 maybe address your question more thoroughly.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
24 1 I was stepping through sort of what we've 2 been doing since we conducted the tests, but before we 3 completed the TRACE assessment, was to utilize this 4 Tmin equals homogeneous nucleation plus contact 5 temperature approach because our initial look at the 6 data and the theoretical basis for homogeneous 7 nucleation plus contact temperature means that it 8 should be a conservative approach. But it does seem 9 to line up with the data in our initial look.
10 In what I'll be presenting next, we've 11 done a number of TRACE calculations using homogeneous 12 nucleation plus contact temperature, as well as other 13 Tmin models, and compared the results against the 14 data. And I think that will show which particular 15 tests were outliers, which particular tests agreed 16 well.
17 Is that sort of what you're getting at?
18 CHAIR MARCH-LEUBA: Yeah. I wanted to 19 know your feelings -- you and Steve have been looking 20 at this data with a fine comb -- if you had a feeling 21 that you had just got lucky, or does it really hold 22 for the majority of the test data? And I think you're 23 saying that it holds.
24 MR. YARSKY: Yes. So, when we -- I mean, 25 this is a spoiler. We will show how it -- it gives a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
25 1 reasonable prediction. However, it's consistently on 2 the conservative side. But, as we showed in this 3 slide, even if you're just a little bit on the non-4 conservative side -- so say the true minimum stable 5 film boiling temperature is 620. You'll predict the 6 heat-up.
7 But if you're off by 20 or 20 degrees, it 8 completely changes your prediction of the 9 consequences. So it's good to be reasonable, but in 10 this case, if you're slightly conservative, that's 11 probably the best place to land.
12 CHAIR MARCH-LEUBA: Yeah. I -- go ahead.
13 MR. YARSKY: So in our Stage 1 TRACE 14 assessment, this is conducted in three steps. There 15 are two stages to the overall TRACE assessment 16 activity. In Stage 1, we just kind of wanted to see 17 where we were. And in Stage 2, we wanted to get a 18 better understanding of what next steps we might want 19 to take in terms of improving TRACE models.
20 So Stage 1 can be thought of as a more 21 traditional TRACE assessment, which is just run the 22 code, compare to the experimental results, and see, 23 are you reasonable? Are you conservative? Are you in 24 sufficient agreement? What is the degree of agreement 25 between the TRACE calculations and the experimental NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
26 1 results?
2 Again, this took us three steps. The 3 first was to use steady-state experimental results to 4 adjust empirical parameters, and in particular the 5 spacer loss coefficients and critical heat flux 6 multipliers. This is a very standard approach.
7 In Step 2, we defined figures of merit 8 associated with key phenomenology and phases of the 9 transient. So, for those members who aren't familiar 10 with the MDAP process, you may have one event, but 11 that one event may be characterized by discrete 12 phases.
13 And what we wanted to do is we wanted to 14 have figures of merit that captured the different 15 phases that are represented by the experiment. And 16 so, while the experiments themselves were conducted 17 primarily to study the fuel heat-up process that 18 occurs when there's a failure to rewet, the test 19 itself represents a number of phases.
20 There's a phase of the experiment where 21 the bundle was stable. There's a phase when it 22 becomes unstable, but the oscillations are small 23 amplitude. There's a phase where the oscillations 24 have grown to the point where the bundle is undergoing 25 cyclic dryout/rewet phases. So we wanted to, for each NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
27 1 of those phases, have a variety of figures of merit 2 that we could compare.
3 And then, of course, in Step 3, we wanted 4 to compare the TRACE results to the experimental 5 figures of merit using both the default TRACE model as 6 well as our interim approach. So this means using a 7 Tmin value based on Groeneveld-Stewart, which is the 8 default, and a Tmin value based on homogeneous 9 nucleation plus contact temperature, which is the 10 interim approach that we just described.
11 So the figures of merit that we looked at 12 -- of course, there are temperature-related figures of 13 merit, and there are pressure-drop-related figures of 14 merit. We also looked at the TRACE-calculated void 15 fractions, but this was not a figure of merit that we 16 defined as part of the comparison process, but just a 17 figure of merit that we defined to help us analyze the 18 experiment and the result so the system would get a 19 better physical picture of what was going on in the 20 analysis part.
21 But we looked at the distribution of 22 pressure drops as well as maximum temperature and the 23 temperature of a specific location.
24 So this is just giving a sample of the 25 comparisons because we conducted very many tests, and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
28 1 there are very many figures of merit. I just wanted 2 to give a representative sample here that we could 3 discuss. We have a very long report with a large 4 number of these kinds of figures, and they kind of all 5 show a very similar result.
6 So this experiment is SINAN 12301T01. So 7 this is a very interesting test because in this 8 particular test, we started the test off by initiating 9 what would be similar to a loss of feedwater heater 10 AOO is how we initiated the test. And this was done 11 by dropping the feedwater temperature in the test.
12 What we show here in this comparison is if 13 we use TRACE with the default option for Tmin, which 14 is the Groeneveld-Stewart, is that we do not observe 15 any failure to rewet in the test when we simulate it.
16 However, in the test, we do observe a failure to rewet 17 towards the end.
18 However, if we run TRACE with the Tmin 19 equals homogeneous nucleation plus contact 20 temperature, we do observe the failure to rewet and 21 the temperature excursion. However, it occurs a 22 little bit earlier in the TRACE calculation relative 23 to the test.
24 Now, in the way we conducted the 25 calculation, the reduction in power occurs at the same NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
29 1 clock time. It doesn't occur at the same time 2 following failure to rewet. So that's why the 3 temperature remains high in the TRACE calculation 4 until the same point in time when the power reduction 5 takes place.
6 And this is fairly consistent across all 7 of the different tests is that when we use the Tmin 8 equals homogeneous nucleation plus contact 9 temperature, the TRACE will predict the start of the 10 temperature excursion a little bit early. But 11 overall, because it's capturing the same behavior, we 12 consider this to be relatively reasonable but on the 13 slightly conservative side.
14 So, if we summarize how we address all of 15 the different phenomena that we wanted to capture with 16 all these different figures of merit, using our 17 standard categorization, overall, when we use Tmin set 18 to homogeneous nucleation plus contact temperature, 19 we're able to garner reasonable, slightly conservative 20 levels of agreement for all the figures of merit over 21 all the phases. This includes, also, the -- when we 22 look at the pressure drops, which is giving us an 23 indication of, are we capturing the instability 24 itself?
25 However, when we used the default model, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
30 1 the Groeneveld-Stewart, the Tmin temperature just 2 appears to be too high, and we very rarely predict a 3 failure to rewet when using that option even though 4 all of the tests were driven to the point of failure 5 to rewet with temperature excursion.
6 So now, in Stage 2, we talked a lot about 7 how the TRACE calculations are sensitive to the 8 minimum stable film boiling temperature. But we 9 wanted to do a more in-depth study using statistical 10 analysis techniques to study the impact of different 11 constitutive models on the assessment in such a way as 12 to identify candidate constitutive models for possible 13 improvements.
14 This relies on a two-step process, which 15 relies on using uncertainty parameters that are in 16 TRACE that are used for uncertainty quantification.
17 We can use that same tool and that same mechanism to 18 drive sensitivity calculations and ultimately use a 19 Morris screening technique to determine which 20 constitutive models in TRACE have the biggest impact 21 on affecting the agreement between TRACE and the 22 experiments.
23 What this will ultimately do is not only 24 give us an idea of what are the important constitutive 25 models, but a relative ranking of how those models NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
31 1 impact the degree of agreement between TRACE and the 2 experiment.
3 What we wanted to do here, of course, was 4 to verify or confirm our understanding of the 5 importance of the selected option for minimum stable 6 film boiling temperature, but also understand if there 7 are other constitutive models that are playing an 8 equivalently important role or are perhaps interacting 9 in such a way that we don't want to miss anything.
10 So we defined new figures of merit. We 11 have three new figures of merit for this purpose, and 12 they're illustrated on this figure. I'm going to take 13 a little bit of time to sort of explain what we're 14 looking at.
15 So, instead of characterizing the 16 consequences, which is what we normally do with TRACE, 17 we want to characterize, how well does TRACE agree 18 with the test? And so we've come up with three 19 different numbers.
20 The first is looking at a temperature-21 related figure of merit, and this compares the maximum 22 temperature that TRACE predicts to the maximum 23 temperature measured in the tests. And then we take 24 that difference, and that difference ideally should be 25 small, but that difference can characterize how well NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
32 1 TRACE and the experiment agree in terms of peak 2 temperature.
3 Similarly, we can define the time of the 4 failure to rewet that is measured in the test as well 5 as what's predicted in TRACE, and we can calculate the 6 difference in that timing.
7 And, lastly, not just capturing time and 8 temperature, we can look at the integral of the 9 temperature difference over that whole phase of time.
10 And that integral difference gives us a third figure 11 of merit to characterize the agreement between TRACE 12 and the test.
13 So the first figure of merit is sort of 14 the difference between the green line and the blue 15 line. The second figure of merit is the difference in 16 time represented by the blue shaded area. And the 17 third figure of merit is the blue shaded area plus the 18 beige shaded area -- and then averaged over time.
19 So now, in the second stage, we follow a 20 two-step process. So I know it's a little confusing 21 because we've had two stages and several steps, but 22 here, we have a two-step process in Stage 2. And the 23 first step, we do what's called a single parameter 24 variation, or one-at-a-time variation.
25 And we study, by conducting a series of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
33 1 sensitivity calculations, how much each of these 2 figures of merit change when we vary just one 3 parameter and one constitutive model at a time. The 4 purpose for doing this is to sort of cut a slice 5 through all of the possible variety of available 6 constitutive model parameters that can be adjusted to 7 come up with a smaller subset that we can analyze in 8 a second step.
9 In the second step, we do multiparameter 10 variation. So this allows us, in a Monte Carlo 11 approach, to shift the values of several parameters at 12 once, run many, many, many calculations, determine 13 sensitivity coefficients, and use a Morris screening 14 technique to establish, what are the key parameters 15 that are dictating the difference between TRACE and 16 the experiment?
17 So this is showing the result of the first 18 step individual parameter variation on the integral 19 FOM for six no-feedback tests and 17 feedback tests.
20 And what we can do is we can draw this red line to 21 scoop up which parameters we want to keep for the 22 second phase.
23 Clearly, there are a number of parameters 24 that have very weak influence on the FOMs, and there 25 are some that have substantially more impact. So here NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
34 1 we're showing that we've drawn a line to pick up, say, 2 these top-five parameters that we want to preserve 3 going into the next step when we look at 4 multiparameter variation.
5 If we look at the time-based FOM, we're 6 slicing and picking up these top eight. And when we 7 look at the temperature FOM, we're selecting the top 8 15. Now, you may say, wait. It looks like some 9 parameters in FOM temperature might still be 10 important, but you're not capturing them with the 11 slice.
12 And it turns out we selected these slices 13 kind of judiciously because some of these parameters 14 it looks like we're leaving out with this FOM 15 temperature, we actually picked them up in the other 16 two cases already. So, if we look at FOM integral, we 17 pick the top five.
18 But when we consider all three FOMs, we're 19 picking up the top ten because five other ones were 20 considered important according to the other FOMs. And 21 with FOM temperature, we're picking up the top 19 at 22 the end with how we've done the slicing.
23 And so this is the same series of figures, 24 but when we consider the duplication of parameters, it 25 shows how much we're actually capturing. So this is NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
35 1 for FOM integral, for FOM time, and for FOM 2 temperature. And so that tells us which parameters we 3 want to vary.
4 The next important thing is, what is 5 Morris screening? So, now that we're going to do 6 several-at-a-time variation, we're going to do a 7 number of sensitivity calculations. We're going to 8 vary all of these parameters, and we'll sample one 9 case, with all the parameters kind of being randomly 10 distributed over a range.
11 We'll sample one of those cases and then 12 calculate the derivative of the FOMs with respect to 13 a single parameter. But we'll do multiple samples 14 over the full-analysis face phase. And this will 15 allow us to calculate a distribution of the 16 sensitivity of the FOMs to the input parameter, but 17 over a wide range of values of the other parameters.
18 So we have now a distribution of the 19 sensitivity coefficient. And if we plot the mean and 20 the standard deviation of that distribution, that 21 gives us an idea of how important that phenomenon 22 represented by that particular constitutive model is.
23 If it has a very high mean value, which would be the 24 Mu*STAR, that means its sensitivity coefficient on 25 average is very large.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
36 1 That means that the physical process 2 represented by that constitutive model is very 3 influential on the figure of merit. If at the same 4 time, it has a high mean value but a high standard 5 deviation, that means that not only is it very 6 influential, but that it has a nonlinear effect or it 7 has a strong coupling, or a strong interaction, with 8 one of the other physical processes.
9 And, conversely, if it has a low mean and 10 a low standard deviation, that just means it's not 11 important. It's not having a strong influence. So, 12 if we do -- when we go through that process, we can 13 calculate these values of the sensitivity coefficient 14 and its standard deviation for all of the parameters 15 that we've captured in the second step.
16 So there are 19 parameters that we're 17 capturing. We're crunching the numbers, and then this 18 is showing the composite results for all six of the 19 tests without feedback as well as showing one single 20 experiment for one of the FOMs plotting the standard 21 deviation and the sensitivity coefficient.
22 What's important to capture here, what we 23 want to look at in the composite figure, are values of 24 the mean sensitivity coefficients that are large 25 coincident with large standard deviation. And those NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
37 1 are telling us not only the most influential 2 parameters but also the ones that have -- potentially 3 are strongly coupled to other phenomena or can really 4 change the results in terms of the figure of merit in 5 a nonlinear way.
6 MEMBER BLEY: Hey, Peter? Dennis Bley.
7 I must have been down a rabbit hole for 30 years.
8 I've never run across Morris screening before. That 9 looks pretty interesting to -- you're getting a range 10 of change with respect to each of the figures of merit 11 to --
12 (Simultaneous speaking.)
13 MR. YARSKY: Yeah. It's like it's taking 14 the sensitivity coefficient, but what you're going to 15 end up doing is -- that sensitivity coefficient, 16 you're sampling it over your full range of 17 uncertainty. So we're using some techniques that --
18 we're using these techniques in an off-label way.
19 Many of these techniques were developed 20 for uncertainty quantification purposes. But we're 21 using them -- instead of trying to quantify 22 uncertainty, we want to see how much influence does 23 this parameter have on average over the full phase 24 space?
25 MEMBER BLEY: Well, and that's part of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
38 1 quantifying uncertainty. That's --
2 (Simultaneous speaking.)
3 MR. YARSKY: Right. Right.
4 MEMBER BLEY: So --
5 (Simultaneous speaking.)
6 MEMBER BLEY: -- that a little bit.
7 MR. YARSKY: We're using it off label, but 8 we're using machinery and techniques that have been 9 developed for a more well established purpose.
10 MEMBER PETTI: Peter, just a question. On 11 the bottom right, the little dots --
12 MR. YARSKY: So these little boxes 13 represent if there's an outlier. So the -- you're 14 going to be testing my memory. The way the bands work 15 I think is the line inside the band is the mean. And 16 then I think it's two Sigma for the thick part and I 17 think three Sigma for the line. And then the dot is 18 like if it's outside of that. That's the outlier.
19 MEMBER PETTI: Okay.
20 MR. YARSKY: I mean it might not be 21 exactly that, but I'd have to look it up in the 22 report, but when there's a dot, it really means that 23 this is really far off. There's one case that's 24 really outside of the mean. So I hope that answers 25 your question. If you need more detail, we can back NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
39 1 open the report and I can get the real numbers out for 2 you.
3 MEMBER PETTI: Thanks.
4 MR. YARSKY: And then here are the results 5 with feedback. And I think what you'll see in the two 6 tests here is that the standout constitutive model is 7 parameter 10-10, which is the minimum stable film 8 boiling temperature.
9 And so now that we've done this detailed 10 statistical evaluation the importance I'm making of 11 the key constitutive models is not surprising. The 12 one that had the most impact and the most variation is 13 minimum stable film boiling temperature, which is what 14 we went in suspecting, and this analysis has confirmed 15 that.
16 The runners-up for next importance are 17 annular-mist flow interfacial drag. This is also not 18 -- it's not as strongly influencing as minimum stable 19 film boiling temperature, but it is -- also not 20 surprising is this is this is really affecting void 21 fraction.
22 And then the next one is critical heat 23 flux, which also makes sense because this is affecting 24 the dryout/rewet phase. I know that these second two 25 parameters, while important, are less influential and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
40 1 nonlinear than the minimum stable film boiling 2 temperature. And to a certain extent these two are 3 affected by the upfront normalization process where 4 you adjust things like space or losses to match 5 pressure drop and where we adjust the critical heat 6 flux multipliers to get the right critical power. And 7 then there are additional phenomena there identified 8 here but are not nearly as important as these ones.
9 In conclusion we found that TRACE produces 10 reasonable but slightly conservative predictions. The 11 fuel heat up during postulated ATWIS-I when the Tmin 12 option is set to homogenous nucleation plus contact 13 temperature. However, that slight conservatism we 14 found that the consequences can be really very 15 sensitive and non-linearly so to whatever is assumed 16 in your calculation with respect to the failure to 17 rewet temperature. And so we think it -- our interim 18 approach we think is still reasonable and will remain 19 our standard practice even though it is slightly 20 conservative, but overall we think it's pretty 21 reasonable.
22 When we conducted our detailed assessment 23 of the key models using advanced statistical 24 techniques, we found that the TRACE predictions are 25 consistent with our expectation, and we even took NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
41 1 these numerical values and compared it to our PIRT.
2 The PIRT is generally developed in a non-quantitative 3 and a more qualitative way based on expert panel 4 opinion, but we can compare it to the quantitative 5 results from our statistical analysis and our Monte 6 Carlo exploration of the phase space. And we found 7 that they're in agreement. So this exercise has 8 really also served as a quantitative approach to 9 confirming the conclusions of our PIRT.
10 And lastly, while we do think that better 11 agreement could be garnered with improvement to the 12 minimum stable film boiling model at this stage we 13 don't think it is necessary because we're doing a 14 pretty good job, though slightly conservative. But I 15 think that the -- it remains prudent to stay slightly 16 conservative, but our analysis does indicate that we 17 could do a better job of agreeing with the data if we 18 fine-tuned a little bit more what we with minimum 19 stable film boiling temperature.
20 And so with that I'd like to conclude the 21 first presentation and address any of your outstanding 22 questions.
23 MEMBER PETTI: Just a question on -- you 24 basically talk about this non-linear (audio 25 interference) that need to be on the conservative NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
42 1 side. What's the uncertainty in the measurements?
2 Ten degrees is not an awful lot of deviation to get 3 such a large change. So --
4 MR. YARSKY: Right. No, this is a very 5 excellent.
6 MEMBER PETTI: -- (audio interference).
7 MR. YARSKY: Yes, this is a very excellent 8 point because the thermocouples are not perfect 9 instruments. There is some measurement uncertainty in 10 the temperature. I believe that I'd have to go back 11 and look at the NUREG to figure out what that 12 temperature is, but I think it's about 5K, which is 13 substantial when you look at the sensitivity because 14 when we talk about non-linear effect, it's really do 15 you experience failure to rewet or do you not? It's 16 kind of like a binary switch that you hit or you don't 17 hit, right?
18 So like the -- if the -- all of the 19 experiments go to failure to rewet. If your 20 calculation does not go to failure to rewet, you're 21 going to have a really substantial difference in your 22 figure of merit, right? So it's going to say the 23 agreement is very poor. But if you ratchet down that 24 minimum stable film boiling temperature say by 5 or 10 25 kelvin, all of a sudden your calculation shows failure NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
43 1 to rewet. Now your agreement is quite good. And so 2 that's the non-linearity aspect of it is that you 3 either -- is that you have this bifurcation where you 4 either hit failure to rewet or you don't in the 5 calculation that creates this non-linear sensitivity.
6 That's why we think it is rather prudent 7 to sort of look at your failure to rewet temperature 8 data and try and bound it as opposed to try and 9 characterize it with like a mean and standard 10 deviation.
11 CHAIR MARCH-LEUBA: I agree with that 12 because from the practical point of view what you are 13 trying to identify is whether it rewets or not.
14 MR. YARSKY: Right.
15 CHAIR MARCH-LEUBA: And the actual 16 temperature it reaches is not important. To start 17 with these are very low frequency, very low 18 probability events and while this is -- this 19 calculation is driving you to evaluate the operator 20 actions that they would have to do to prevent from 21 failure to rewet from happening. So I think your 22 approach is very good and I like it.
23 I have a question, and I know you cannot 24 speak for NRR, but while you were (unintelligible) we 25 were always requiring vendors to use homogeneous NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
44 1 nucleation temperature plus contact on their 2 calculations using their code instead of their 3 correlation. Is that correct?
4 MR. YARSKY: Yes, that is by my 5 recollection when --
6 CHAIR MARCH-LEUBA: Yes.
7 MR. YARSKY: Not necessarily when I was at 8 NRR, but I was engaged in performing confirmatory 9 analysis to support licensing actions. And with my 10 recollection for those specific projects I recalled 11 the vendor calculations being performed in a 12 consistent way with the homogeneous nucleation 13 temperature.
14 CHAIR MARCH-LEUBA: And we don't foresee 15 today any new MELLLA++ that would require these type 16 of calculations, but if they do happen my expectation 17 will be that NRR will know -- even the vendors will 18 know to use homogeneous plus contact approach.
19 MR. YARSKY: Right. So this is -- we 20 released this finding of the homogeneous nucleation 21 plus contact temperature approach relatively soon 22 after conducting the tests. And our latest more in-23 depth thorough analysis indicates we still want to 24 keep doing this. So it's really not a change.
25 There's really no change in our guidance or our NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
45 1 process or our approach.
2 CHAIR MARCH-LEUBA: Okay. Thank you.
3 MR. YARSKY: It's just I think that our 4 more thorough analysis has sort of indicated that --
5 if you want to call it lucky or insightful, is that 6 we've landed on the right approach in our interim 7 analysis.
8 CHAIR MARCH-LEUBA: And it makes some 9 sense, some physical --
10 MR. YARSKY: Right.
11 CHAIR MARCH-LEUBA: -- sense.
12 MR. YARSKY: Right.
13 CHAIR MARCH-LEUBA: So it's not completely 14 empirical.
15 MR. YARSKY: There is like a theoretical 16 basis for what we suggested at first and I think what 17 we're observing now.
18 CHAIR MARCH-LEUBA: Okay. Thank you.
19 Members, you have any more questions for 20 Pete?
21 (No audible response.)
22 CHAIR MARCH-LEUBA: So we are a little bit 23 ahead of the schedule and we are going to get into the 24 non-dimensional analysis. Do you want to take a 10, 25 15-minute biological break, because it's early in the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
46 1 morning for some people?
2 (No audible response.)
3 CHAIR MARCH-LEUBA: So let's take a 15-4 minute break and come back at 10:50, 10 minutes before 5 the hour. Okay? And by time we'll have changed 6 presenters and have the new slides on the screen. So 7 we are on break, on recess.
8 (Whereupon, the above-entitled matter went 9 off the record at 10:32 a.m. and resumed at 10:49 10 a.m.)
11 CHAIR MARCH-LEUBA: So let's go back in 12 session. We're still in open session. You are not 13 going to present anything proprietary, correct, Steve?
14 MR. BAJOREK: That is correct?
15 CHAIR MARCH-LEUBA: Okay. So go ahead and 16 do your magic. Maximize the -- yes, perfect.
17 MR. BAJOREK: Okay. There we go. Well, 18 good morning, everyone. My name is Steve Bajorek.
19 I'm from the same Division of Safety Analysis as Pete 20 Yarsky and Tarek. And what I'd like to do next is 21 talk about some of the non-dimensionalization of the 22 KATHY data.
23 In the first presentation today, 24 especially towards the end when we started to take a 25 look at TRACE and how it could predict some of these NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
47 1 phenomena, you could see that there was actually a 2 fairly large number of physical processes that could 3 influence and affect this failure to rewet.
4 So what I want to do next is kind of take 5 a step back and simplify things by looking at the 6 experimental data and trying to see what does it tell 7 us about the parameters and the things which are most 8 important and can we come up with a way of showing 9 when we'll have this failure to rewet in terms of 10 several of the more important quantities that affect 11 the data?
12 Now part of this originated because of 13 some concerns on keeping the data proprietary. There 14 was a lot of sensitivity from the experimentalists on 15 whether someone might be able to take the data and 16 back out some information that they considered 17 economically vital to them. So we wanted to be very 18 protective of the data. We understand the importance 19 of protecting the information.
20 So as this discussion of what's 21 proprietary and is not proprietary, we said well let's 22 take a look at how we can present the data but not do 23 so in a way that would give away those things which 24 are very important to the experimental organization.
25 We also recognized that the data was NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
48 1 showing that the conditions could lead to this failure 2 to rewet, so there were some safety implications that 3 we needed to show in a way to others that hey, there 4 are conditions that will certainly lead to this. And 5 we wanted to try to come up with a way that the NRC, 6 our reviewers and our own evaluation, come up with 7 quick screening method by which we could point to 8 conditions that would give us failure to rewet and 9 possibly lead to damage.
10 So I went back and looked at probably the 11 more fundamental works in flow instability which had 12 been done by Ishii in the mid-'702. Now most of the 13 work that he did was looking at density wave 14 oscillations, but his work was fairly general and 15 could cover a large number of situations. He took the 16 mass/momentum/energy equations, non-dimensionalized 17 those and derived several dimensionalist parameters 18 that were indicative of when you should have certain 19 stabilities and how those stabilities could develop.
20 This isn't all of those, but I've listed some of the 21 ones which were more important including the Froude, 22 the Reynolds, the subcooling number, a phase-change 23 number and the density ratio.
24 Now of these the subcooling number and the 25 phase-change number were attributed to be the most NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
49 1 influential. Others became very -- were also 2 important. The Reynolds number. A couple of them I 3 left off. Actually it was kind of great to see in the 4 models that were becoming very important to TRACE.
5 The wall drag and the interfacial friction. Those 6 were two other parameters that turned out to be 7 important in sensitivities in Ishii's studies. But it 8 was the subcooling and the phase-change number that 9 tended to be the most influential.
10 The reason for that is the subcooling 11 number tended to be more stabilizing. The larger that 12 was, that more -- it kept your fluid in a subcooled 13 condition, prevented oscillating from occurring.
14 However on the other hand the phase-change 15 number was such that as it became larger and as your 16 exit quality became large, this was more 17 destabilizing. And looking at a large number of 18 situations Ishii was able to develop a map which sort 19 of has a little bit of a horseshoe curve to it. And 20 his instabilities would originate in this crosshatched 21 region as you had a sufficiently large phase-change 22 number and a subcooling number which was moderately 23 low.
24 So we thought well, let's take a look at 25 the KATHY data in terms of these two numbers. In one NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
50 1 way it's kind of very interesting because we talked a 2 lot about Tmin, the minimum stable film boiling 3 temperature. And in one way you see something very 4 similar here. In the stability you have subcooling, 5 your ability to cool the liquid, versus generation of 6 the voids, whereas Tmin is essentially a balance 7 between getting energy away from a surface as opposed 8 to getting energy conducted to the surface. So it's 9 kind of similar in one way when you think of it that.
10 So the first thing we did is let's take a 11 look at all of the tests in general, the tests with 12 SINAN, the non-SINAN tests, and what we're really 13 seeing in those. And very typically over the first 14 100, couple of 100 seconds of the test everything is 15 stable. And then depending on whether you're driving 16 it with the SINAN component, with the power, or you're 17 stepping up the power in the non-SINAN test, you start 18 to get oscillations. Many times this occurred first 19 in the flow rate which is shown in the purple and sort 20 of in the middle one of this slide. Eventually the 21 power would start to become oscillatory and more and 22 more unstable.
23 Down at the bottom you see the temperature 24 in this case for one of the rods that went through a 25 failure to rewet. Now well after you started to get NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
51 1 oscillations in the mass flow rate and the power 2 oscillations in the temperature this ratchetting of 3 dryout, going back towards saturation, starts to creep 4 in into those temperature oscillations until the power 5 and the flow becomes such that you go through the 6 failure to rewet.
7 So I said well, let's try to characterize 8 these -- really these four periods that we see where 9 there's oscillations or steady behavior in the flow.
10 And in some of the nomenclature that we'll have in a 11 couple of the figures we'll have a three-letter index 12 to tell us whether the temperature mass flow rate and 13 the power is either stable, oscillatory; I used a U 14 for unstable for that; oscillatory is probably a 15 better term, or if when it's very clearly a failure to 16 rewet.
17 And by this we'll be able to take a look 18 at the KATHY data and characterize the four periods 19 generally such that we see everything's stable, period 20 1. Another period, period 2, where temperature is 21 stable but the flow or the power starts to become 22 oscillatory. And then we start getting close to the 23 failure to rewet. The rod temperatures begin to 24 oscillate. The power is oscillating. The flow rate's 25 oscillating. There's a return to rewet, but we know NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
52 1 that by -- from many of the tests that if we continue 2 to ratchet up the power or reduce the flow eventually 3 we're going to get to this clear failure to rewet 4 while the other two parameters are oscillating, and 5 that's sort of a very short period 4 because we want 6 to protect the equipment.
7 MEMBER HALNON: Okay. So --
8 MR. BAJOREK: So if you go back and you 9 take a look -- I'm -- yes?
10 MEMBER HALNON: Steve, this is Greg 11 Halnon. Just real quick on that period 3 it looks 12 more of a transition period to me where -- they were 13 transitioning to film boiling, but can you sustain 14 sort of a period 3 without it going to a period 4?
15 MR. BAJOREK: As long as you remain below 16 your minimum film boiling temperature. Okay. As long 17 as that -- the power and the flow is such that you 18 keep that rod in transition boiling. Okay?
19 (Audio interference) 20 MR. BAJOREK: And that's sort of a hybrid 21 so you can stay in there and the temperature won't get 22 away from you, but once you get above Tmin, that 23 temperature is so high, you no longer have that good 24 liquid-to-wall contact. And then all of a sudden 25 you're cooling is largely by radiation.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
53 1 MEMBER HALNON: Okay. So as long as you 2 don't increase the power to where that Tmin is going 3 to increase --
4 MR. BAJOREK: That's correct.
5 MEMBER HALNON: -- you can sustain it?
6 Okay.
7 MR. YARSKY: And, Greg, if you look at 8 Steve's slide, he's going to show you we sit -- in the 9 experiments we can sit in that period 3 for a long 10 time.
11 MEMBER HALNON: Okay.
12 MR. BAJOREK: Right. Right.
13 MEMBER HALNON: So that was my question, 14 whether or not it was just a transition period that 15 was going to occur -- go beyond that anyway or if it 16 was sustainable and --
17 (Simultaneous speaking.)
18 MR. BAJOREK: Yes, it's sustainable under 19 the right conditions, but I think when we start to see 20 the temperature, the flow rate and the power 21 oscillating -- and the power, the mean power slowly 22 increasing, I think we realized that that is a 23 condition you really don't want to be in because 24 you're sort of on the knife edge. As conditions get 25 just a little bit worse that is going to send you NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
54 1 through that failure to rewet. So as you run the 2 tests, once you've collected enough data and you're 3 not quite there, you don't want to push you luck too 4 much.
5 MEMBER HALNON: Okay. Got it. Thanks.
6 MR. BAJOREK: Okay. So we go back and we 7 take a look at the flow and the power as we did 8 earlier. And now we'll start to characterize the four 9 periods in this case. And up until 800 seconds 10 everything is stable. Then we get this stable 11 temperature, but oscillatory power, oscillatory flow 12 that starts to grow until about 1,600 seconds in this 13 test. And then this period 3 takes over where 14 everything is oscillatory and approaching -- which in 15 this test did give us a failure to rewet. So we kind 16 of see all four of these periods as we characterize 17 this.
18 So what I did is I -- we took all of the 19 tests that were available to us -- and this just shows 20 when the temperature starts to go oscillatory, go 21 through all of those tests. And in the middle of each 22 of those periods: steady, steady with -- steady 23 temperature, unsteady other parameters, to completely 24 unsteady -- went to the middle of those and said well, 25 let's pick out the conditions that we get from the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
55 1 experimental data, determine what is the subcooling 2 number, the Reynolds number, and also the phase-change 3 number.
4 And we did this -- and this figure kind of 5 shows everything, and that's kind of where the 6 nomenclature is. Everything where you -- the 7 temperatures were steady occurred to the left of the 8 phase-change number of 20. And you kind of see two 9 different groupings here because the tests were run at 10 two different subcoolings. And I think that's the 11 upper distribution and the lower distribution.
12 Once you exceed a phase-change number of 13 20, okay, now everything was oscillatory and in some 14 cases we go through this failure to rewet. And can 15 you -- I think you can see my cursor.
16 CHAIR MARCH-LEUBA: Yes, we can see it.
17 MR. BAJOREK: Okay. Good. Yes, you can 18 see that. And you can see these kind of brown 19 downward-facing triangles. Those are your failure to 20 rewet cases. They're sort of mixed in with others 21 where everything is uncertain -- not uncertain, excuse 22 me -- un-oscillatory, which is why we sort of look at 23 those data as being very, very close to achieving this 24 failure to rewet. But every time we saw that failure 25 to rewet it was at the phase-change number of 20 or NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
56 1 greater.
2 Now as we look at this idea of cooling, 3 keeping the rod surface cool versus generating void, 4 one of the other parameters that Ishii found that was 5 important to influence that boundary had been the 6 Reynolds number. Now we don't have enough data to 7 really map out the curve as had been done in some of 8 the Ishii studies because we only had two subcoolings 9 and we only had two pressures. You have to vary that 10 over a much broader range in order to really map 11 everything out.
12 However, Reynolds number should also 13 influence this because it's another way of helping --
14 telling you that you have good or poor convective heat 15 transfer. So I said let's define another 16 dimensionless parameter, the product of the subcooling 17 and in the Reynolds number, and recast the same data 18 in that manner. And now we sort of get a different 19 type of curve. Failure to rewet is greater than --
20 nominally greater than a phase-change number of 20.
21 And it was always when you had relatively 22 low subcooling and lesser convective heat transfer 23 with a lower Reynolds number as you did not have a 24 large phase-change or you had good cooling or high 25 subcooling up in that kind of quadrant in the upper NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
57 1 left. No, you never go to that failure rewet. But 2 down here in this lower quadrant, okay, this was the 3 -- these were the conditions in the tests that led to 4 oscillatory behavior and that failure to rewet.
5 CHAIR MARCH-LEUBA: If these -- finish.
6 Let me know when you're finished and I'll ask the 7 question.
8 MR. BAJOREK: Let me see. Okay. Well, go 9 ahead and ask your question, please.
10 CHAIR MARCH-LEUBA: All right. Can you go 11 back to the previous slide, slide 9?
12 Okay. What I see here is a large -- very 13 little dependent on any subcooling. When you double 14 it, you go from 1 to 2, it's not that much.
15 MR. BAJOREK: Yes.
16 CHAIR MARCH-LEUBA: There is not much 17 change of the MPCH of 20. Right?
18 MR. BAJOREK: Right.
19 CHAIR MARCH-LEUBA: However, you would --
20 if you triple it, you will get numbers higher. And 21 now you go back to 10, to slide 10, the next one. If 22 you have an subcooling greater than 3, for example, 23 you will be outside that square above the line of 24 6,500 -- 65,000.
25 MR. BAJOREK: Yes.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
58 1 CHAIR MARCH-LEUBA: Do you see what I 2 mean?
3 MR. BAJOREK: Right. Right. Yes, I --
4 CHAIR MARCH-LEUBA: So I kind of like the 5 MPCH 20. I'm not sure I can justify the subcooling 6 line.
7 MR. BAJOREK: I think I agree with that 8 because -- we'll take a look at another slide that I 9 have coming up. This is what does the code say about 10 this? And we'll get to that to the second.
11 But where we're at right now is the data 12 is suggesting a criteria, okay, that you have this 13 failure to rewet when you have the phase-change number 14 of 20. And that might be it. Okay? You're also 15 subject when you're down at these relatively low 16 Reynolds subcooling number, because what -- I think 17 what that says is that the next ones you might want to 18 be concerned with are these over here. Okay>?
19 So you may have conditions where this 20 phase-change number may give you failure to rewet, a 21 lower value, if you were to drop these numbers over 22 here slightly. So that's just sort of a preliminary 23 boundary of where we see the -- what the data is 24 showing us right now. Now --
25 MEMBER BLEY: Steve? This is Dennis Bley.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
59 1 MR. BAJOREK: Yes.
2 MEMBER BLEY: I kind of agree with Jose on 3 this. I don't see much -- unless you've got something 4 more that's -- than what's showing up on these charts, 5 that 65,000 just doesn't seem to be a real trigger for 6 anything.
7 MR. BAJOREK: I think so. It was 8 something that showed up in that second one and -- I'm 9 going to jump ahead a couple, three here, sort of a 10 backup slide -- and we did some preliminary work, try 11 to say well -- we took one of the MELLLA+ calculations 12 that had been done at the time, and we did this about 13 three years ago -- what was TRACE showing us?
14 And we had two different transients. One 15 of them showed a failure to rewet. I think that was 16 the one that Pete Yarsky showed in his earlier figures 17 that gave us the high temperatures. Phase-change 18 number greater than 20. Okay? The other one did not 19 go through a failure to rewet. And that's over here.
20 Again a phase-change number less than 20. So it may 21 well be that it's all wrapped up in the phase-change 22 number. And where we're at in terms of the subcooling 23 and what the convective heat transfer is may be 24 playing a second or a third role.
25 Now this was all done as kind of a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
60 1 preliminary screening because realize we still have 2 work to do on this. We did this at a time where we 3 didn't have all of the KATHY electronic data, and this 4 was picking that data right off of the figures in the 5 middle of these periods. Okay? It was also done by 6 using a couple of assumptions where -- what will lead 7 to this condition would be overpower with 8 undercooling. So in calculating this phase-change 9 number and the subcooling number we're looking at 10 minimum flow, maximum power during this period.
11 That might not be the best way of doing 12 it. We might really want to look at the mean power or 13 perhaps a square root sum of the squares, okay, to 14 come up with that. And since that time we have 15 awarded a couple of grants to universities to take the 16 experimental data and now take it the next part of 17 this to re-look at it, use all of the data that's 18 available to them and see whether these parameters are 19 the most appropriate ones, whether it's a phase-change 20 number of 20 that is sort of the criteria, or how that 21 would vary if we went up in pressure, down in 22 pressure, or had higher or lower subcoolings.
23 And secondly, examine the code simulations 24 when we have this failure to rewet to see are the 25 codes predicting that on about the same basis.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
61 1 So we wanted to show you this just as a 2 means that, one, we think we can take the data, non-3 dimensionalize it in a fashion that we can use it as 4 a screening device, give it -- give some of this to 5 universities, other vendors as a way of indicating 6 where you think failure to rewet can and will occur, 7 but without giving away all of the experimental data.
8 MEMBER REMPE: So, Steve, let me quiz you 9 about something you just said before this last 10 comment. You said you've made some grants to several 11 universities, and I believe the words you used were to 12 -- and allowed them access to all the data for further 13 evaluations. Are you indeed giving them access to all 14 the data or are you giving -- tying one hand behind 15 their back by only giving them part of the data?
16 MR. BAJOREK: I believe we've given them 17 some of the data but not necessarily all of that.
18 MR. YARSKY: Well, Steve, for the grant 19 that you're talking about we did give them all of the 20 data but the researchers are under an NDA for the 21 portions --
22 MR. BAJOREK: Right.
23 MR. YARSKY: -- that are sensitive.
24 MEMBER REMPE: Okay. That makes sense.
25 Thank you.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
62 1 MR. BAJOREK: Yes, it's difficult dealing 2 with some of the grantees because officially we aren't 3 allowed to interact with them now. The grants have to 4 be conducted by the principal investigator without 5 dialog between them and the staff.
6 CHAIR MARCH-LEUBA: Hey, Steve, is this a 7 new requirement or is something -- there's not -- no 8 contact with your contractors?
9 MR. BAJOREK: It is not a contractor.
10 It's the difference between a contract and a grant.
11 CHAIR MARCH-LEUBA: Oh.
12 MR. BAJOREK: A contract yes, but if it's 13 a grant, we are forbidden to give any kind of 14 direction. And the instructions to us has been we 15 can't even talk to them.
16 CHAIR MARCH-LEUBA: And is this coming 17 from the NRC lawyers?
18 MR. BAJOREK: That's -- I believe that 19 is --
20 CHAIR MARCH-LEUBA: Is there something 21 that can be changed, because it makes no sense.
22 MR. BAJOREK: I wholeheartedly agree, but 23 I believe that's the message.
24 CHAIR MARCH-LEUBA: Kim wants to say 25 something.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
63 1 MS. WEBBER: Yes, Jose. So the Integrated 2 University Program, of which this is one of the 3 recipients of funding from that program, statutorily 4 we cannot direct the research. In this particular 5 case the staff reviewed their grant proposal along 6 with many others, and many of the proposals were 7 accepted. And for any of those grant recipients we 8 cannot direct their research. And that is something 9 that Office of Research worked very closely with OGC, 10 our lawyers, to figure out what we could and couldn't 11 do relative to communications with grant recipients.
12 And so --
13 CHAIR MARCH-LEUBA: So these --
14 MS. WEBBER: -- it's very clear. So 15 it's --
16 CHAIR MARCH-LEUBA: Well, I hear you say 17 that it doesn't come from our lawyers; it comes from 18 Congress.
19 MS. WEBBER: Well so even the --
20 CHAIR MARCH-LEUBA: And they created the 21 program?
22 MS. WEBBER: Yes, it's my understanding --
23 and I haven't read the legislation recently, but the 24 interpretation of the legislation is that the purpose 25 of the Integrated University Program is to support NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
64 1 nuclear engineering and nuclear fields to get not only 2 research completed, but to support nuclear programs in 3 general that then -- the resources used to support 4 faculty appointments and fellowships and even this 5 research go to prepare the pipeline to put students 6 and folks into the nuclear field. And so that --
7 CHAIR MARCH-LEUBA: This makes a lot more 8 sense. NRC is a conduit to the distribute the money.
9 MS. WEBBER: Correct.
10 CHAIR MARCH-LEUBA: It's not really --
11 yes, okay. I retract my comment.
12 MEMBER REMPE: Well, actually --
13 CHAIR MARCH-LEUBA: Okay. Joy, you 14 have --
15 MEMBER REMPE: -- we'll hear about -- it 16 used to be called the Integrated University Programs, 17 but we heard from Ray it's called the University 18 Nuclear Leadership Program. But we're going to be 19 hearing about the projects in an upcoming meeting and 20 we can discuss this more at that time, too.
21 MS. WEBBER: Correct. Thanks, Joy. The 22 only thing that I was -- the only additional thing 23 that I was going to say is that I think we're in --
24 coming up on year three in this new area where some of 25 the $16 million that we fund these opportunities --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
65 1 some of that money is used to fund research and 2 development, whereas in the past prior to three years 3 all of that funding went towards faculty development 4 and fellowships and trade schools and other 5 organizations more for the educational side of it.
6 But in the last two years, going on three years, we've 7 been working with OGC to help us carve out some 8 funding that will actually promote research in areas 9 of interest to the NRC.
10 So that's what this -- this is one grant 11 recipient who is actually doing some research, as I 12 think Pete and/or Steve pointed out, that are doing 13 something of interest to us. So that's where it's a 14 little bit of a sticky situation given the context, 15 but I think, Joy, to your point, at the future 16 presentation there will be people much more 17 knowledgeable about it than probably I am.
18 MEMBER REMPE: Yes, we can explore it a 19 bit more. And again, I'm still kind of stumbling over 20 this wall that precludes any sort of interaction.
21 Since the research is of interest to NRC it might be 22 good to be a little bit more flexible about that, but 23 maybe we can explore that a bit more.
24 CHAIR MARCH-LEUBA: Steve, I assume you're 25 down to your last slide?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
66 1 MR. BAJOREK: I am done now unless there 2 are additional questions.
3 CHAIR MARCH-LEUBA: Any questions from the 4 members for both Steve, Pete or Kim?
5 (No audible response.)
6 CHAIR MARCH-LEUBA: I heard something 7 yesterday about members -- five seconds for members' 8 comments and doughnuts. We'll assume that there are 9 no more questions, so I'm going to open the floor to 10 members of the public.
11 If anybody wants to make a comment, place 12 it on the record, you can do it now. Please state 13 your name. And if you are using the phone line, you 14 need to un-mute yourself using *6.
15 (No audible response.)
16 CHAIR MARCH-LEUBA: Okay. Five-second 17 rule again. There are no comments.
18 At this point I just wanted to say I'm 19 really interested. I mean it's a great thing that NRC 20 performed these tests because it has shown us 21 something that we didn't expect, right? And it of 22 value and it is of value to continue to analyze it and 23 try to ensure that both NRR test confirmatory 24 calculations or performed reviews of vendors or 25 applicants -- that we keep this data in mind. Okay?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
67 1 So with that said, thank you very much.
2 Excellent job.
3 Anybody has an additional comment?
4 MEMBER BLEY: Yes, just a question for 5 Peter. Going back to the Morris comparison stuff. Is 6 that something you guys have been using for a while?
7 And if it's not, I think there's pretty broad 8 application for that in a white paper for folks in 9 other areas around the NRC. It might be very useful.
10 MR. YARSKY: Yes, Dennis, that's a very 11 excellent point. This is, as far as I can tell the 12 first deployment of this technique at the NRC. It's 13 being utilized within a larger framework at the 14 University of Illinois for the objective of 15 uncertainty quantification, but we're using it here 16 just for a more limited purpose. And on the report I 17 made a point of including kind of like a white-paper-18 level description of the technique to make it more 19 accessible to the more thermohydraulically-inclined of 20 -- and of the staff that we interact with so that 21 hopefully it can gain some more traction.
22 MEMBER BLEY: Well, it's not just that.
23 I mean, they'll read this for interest anyway, but 24 folks in PRA reliablity, for example, and many other 25 areas who try to do uncertainty quantification and who NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
68 1 do sensitivity studies, this gives a pretty good tool 2 for clarifying the important factors involved in those 3 kinds of cases. And a separate white paper would be 4 nice because it would open it up to people who won't 5 be working through the details here associated with 6 thermohydraulics.
7 Anyway, nice piece of work. I liked it.
8 MR. YARSKY: Thank you.
9 CHAIR MARCH-LEUBA: I think Vicki is in 10 line.
11 MEMBER BIER: Yes, I just have a quick 12 follow-up on Dennis' comment. Like Dennis I had no 13 idea about this method until today, so I'm probably 14 bringing coals to Newcastle or whatever. I'm sure you 15 know much more about this than I do, but just from my 16 quick kind of Google research during the meeting this 17 morning it looks like the original Morris method had 18 problems when there were negative effects or negative 19 coefficients or whatever, and there's a modern version 20 that corrects for that. I assume you guys are aware 21 of that distinction and dealing with that, et cetera?
22 MR. YARSKY: Yes, Vicki, that's a really 23 great question. That's very true. What we do -- if 24 you go back to the -- I guess I can go back to the 25 slide and show this, if I can share again. Just give NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
69 1 me one moment and I'll bring this back.
2 (Pause.)
3 MR. YARSKY: Yes, this is a problem. And 4 so a very astute observer would notice that -- I 5 mention here the Mu and the Mu*STAR and that later on 6 we are presenting results from Mu*STAR.
7 So this is the more modern technique that 8 addresses the --
9 MEMBER BIER: Okay. Got it.
10 MR. YARSKY: -- (audio interference) of 11 coefficients. Yes, so when you have to do it 12 sometimes you get negative coefficients and when you 13 start averaging you end up erasing the sensitivity.
14 So you have to kind of absolute value the negative 15 ones before you do the averaging. So it's not 16 capturing -- so it becomes a quantitative measure of 17 the sensitivity but not necessarily the direction of 18 the sensitivity.
19 MEMBER BIER: Right. No, I understand 20 that because I know other things that I've dealt with 21 like coefficient of variation becomes sort of 22 meaningless if the quantity is centered around zero, 23 et cetera.
24 MR. YARSKY: Yes.
25 MEMBER BIER: So there's a lot of things NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
70 1 that have that problem.
2 MR. YARSKY: Yes.
3 MEMBER BIER: So thanks for the 4 explanation.
5 MR. YARSKY: But it's -- you shouldn't 6 think of the mean value here as being necessarily the 7 true mean value of the sensitivity because it will 8 positive value the negative coefficients. So it's 9 giving an idea of is this sensitive or not without 10 explicitly quantifying that average because sometimes 11 that average might be around zero. Right? So it 12 would erase itself out. But that's in fact right --
13 MEMBER BIER: Thank you.
14 MR. YARSKY: -- and that's what we have to 15 take into account. So that's the (audio interference) 16 --STAR is how that's (audio interference) in the 17 nomenclature.
18 CHAIR MARCH-LEUBA: Joy, I believe you 19 want to say something?
20 MEMBER REMPE: Yes, I also really 21 appreciate Peter and Steve taking the time to give 22 these presentations today and updating us.
23 I guess I have a question for you, Jose.
24 I know we do this sometimes for reg guides, but I 25 think this is a good activity that's been done by the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
71 1 research organization that's very practical and 2 getting good results. Do you think it might be 3 worthwhile just having an item in the upcoming P&P for 4 December? So it wouldn't require any more 5 presentations, but just saying the subcommittee did 6 review this work, they had a very favorable impression 7 that this is a worthwhile effort that's produced -- or 8 is producing some very useful results, or something to 9 acknowledge what --
10 CHAIR MARCH-LEUBA: I would love to do 11 that. I'll --
12 MEMBER REMPE: -- (audio interference) 13 have it issued and the meeting minutes?
14 CHAIR MARCH-LEUBA: I would love to have 15 a trial balloon of what I'm proposing to do for these 16 situations. So I'll take the action item of writing 17 a couple of paragraphs similar to what we do for reg 18 guides.
19 MEMBER REMPE: That's what I'm thinking 20 exactly.
21 CHAIR MARCH-LEUBA: Yes. Write two or 22 three paragraphs from what I've seen what the topic is 23 and then P&P accepts it, or rejects it, and we ask 24 Scott to send a letter saying thank you for coming to 25 this meeting. We loved it. We don't have any follow-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
72 1 ups. And attach the paragraph.
2 Yes, I think we should add it to P&P if we 3 can.
4 MEMBER REMPE: Okay. Thanks.
5 CHAIR MARCH-LEUBA: Kim?
6 MS. WEBBER: Yes, I'm not familiar with 7 that process because we don't do the reg guides, so 8 I'm just curious to better understand what that 9 process entails.
10 CHAIR MARCH-LEUBA: With reg guides one 11 members gets assigned a new reg guide --
12 MS. WEBBER: Okay.
13 CHAIR MARCH-LEUBA: -- typically an 14 update. That members reviews it on paper and then the 15 member makes a proposal to the Committee whether we 16 will have a Committee -- a presentation and a letter 17 on it or not.
18 MS. WEBBER: At the Full Committee meeting 19 or --
20 CHAIR MARCH-LEUBA: This is done during 21 the Full Committee in what we call the P&P, the 22 Process and Procedures.
23 MS. WEBBER: Okay.
24 MEMBER REMPE: But is part of the Full 25 Committee. It's not prior to -- sometimes we have a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
73 1 P&P Subcommittee during Full Committee week. This is 2 part of the Full Committee meeting activities.
3 CHAIR MARCH-LEUBA: So my concern as that 4 we're talking about this, Kim, is that sometimes we 5 have an item like this where we don't have any issue 6 whatsoever with it. I mean we love it. And therefore 7 we just drop it and don't write a letter on it.
8 MS. WEBBER: Okay.
9 CHAIR MARCH-LEUBA: And I mean we only 10 write letters when we have problems and we have 11 recommendations.
12 MS. WEBBER: Okay.
13 CHAIR MARCH-LEUBA: So I've been saying 14 there has to be a way to -- so our executive director 15 can send Kim a letter saying we reviewed it in the 16 subcommittee; we don't have any problems.
17 MS. WEBBER: Oh, okay.
18 CHAIR MARCH-LEUBA: Thank you very much.
19 MS. WEBBER: And so does that mitigate a 20 presentation at a Full Committee meeting or --
21 MEMBER REMPE: Yes.
22 CHAIR MARCH-LEUBA: Yes. Yes.
23 MEMBER REMPE: There would be no 24 presentation required by you guys. It's just a -- in 25 this case a way of saying we liked what we heard and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
74 1 it would be something that would be approved by the 2 Full Committee.
3 MS. WEBBER: Okay.
4 MEMBER REMPE: Unless of course someone 5 raises their hand and doesn't like it, but that's what 6 we are proposing.
7 CHAIR MARCH-LEUBA: Yes, it does not 8 involve any work on your part, on research's part.
9 MS. WEBBER: Okay. Yes, I just wasn't 10 familiar with that process.
11 CHAIR MARCH-LEUBA: Yes.
12 MS. WEBBER: But I think it would be 13 helpful to have your statement that you loved this 14 work and that you don't need to write a letter because 15 you feel that this work is very technically mature and 16 all the good words that you might want to say.
17 CHAIR MARCH-LEUBA: It's not our 18 recommendation from the Committee, but a sign of 19 approval of what you're already doing. So it doesn't 20 raise to the level of a letter.
21 MS. WEBBER: Okay. Okay. Thanks for 22 explaining that.
23 CHAIR MARCH-LEUBA: Okay. Any further 24 comments or questions?
25 (No audible response.)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
75 1 CHAIR MARCH-LEUBA: If not, we're going to 2 adjourn, but there is an administrative issue off the 3 record.
4 Zena, can you confirm this afternoon's 5 meeting? Does it start at 1:00 or at 2:00?
6 MEMBER BALLINGER: This is Ron. I've been 7 trying to get an answer to that and I have been 8 unsuccessful. I'm assuming it's at 2:00. It's my 9 meeting.
10 CHAIR MARCH-LEUBA: The agenda says 1:00, 11 so --
12 MS. ABDULLAHI: Let me explain. So when 13 the meeting was scheduled it was scheduled for 1:00, 14 however, there seems to be a general plan that we 15 should take -- to it 2:00 to 6:00. But since this was 16 scheduled from 1:00 and the agenda says 1:00, we will 17 start at 1:00.
18 Larry, are you on?
19 MR. BURKHART: Yes, I --
20 CHAIR MARCH-LEUBA: We can do something 21 similar to what was done here. There was some 22 confusion whether it was 9:00 or 9:30. And if 23 somebody's on the line, if somebody shows up at 1:00, 24 we tell them come back at 2:00. It is up to Ron to 25 tell us what he wants to do, I guess.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
76 1 MR. BURKHART: So this is Larry. The 2 agenda that's on the website says 1:00, so we should 3 start at 1:00.
4 MEMBER BALLINGER: So let it be written, 5 so let it be done.
6 CHAIR MARCH-LEUBA: Okay. So the meeting 7 was already adjourned. Thank you everybody for your 8 comments.
9 MS. WEBBER: Thank you very much.
10 MEMBER CHAIR REMPE: REMPE: Yes, thank 11 you.
12 (Whereupon, the above-entitled matter went 13 off the record at 10:43 a.m.)
14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
Non-Dimensionalization of KATHY Data Stephen M. Bajorek, Ph.D.
Office of Nuclear Regulatory Research United States Nuclear Regulatory Commission Ph.: (301) 415-2345 / Stephen.Bajorek@nrc.gov ACRS Thermal Hydraulics Subcommittee Meeting November 17, 2021
Introduction & Background
- After the KATHY data was obtained, there has been discussion on what is Proprietary and what is not Proprietary.
- Non-dimensionalization proposed as a way to:
- 1. Protect data that may be considered as Proprietary.
- 2. Provide the NRC with a quick means to evaluate licensee analyses for the possibility of severe consequences during a postulated ATWS-I.
2
Two-Phase Flow Instabilities
- Work initially performed by Ishii helps: [Ishii, M., STUDY ON FLOW INSTABILITIES IN TWO-PHASE MIXTURES, ANL-76-23, March 1976]
3
Stability Map Stabilizing Nsub Npch Destabilizing 4
Test Periods Q
m T
5
Data Characterization
- For each test, characterize each of four periods based on main parameters (rod temperature, flow rate, power) & condition.
- Nomenclature:
T m Q
- Conditions: S = stable U = oscillatory with F = FTR for rod temp.
6
Power and Flow rate 7
Rod Temperature 8
Nsub vs Npch 9
NreNsub vs Npch 10
Failure to Rewet Possible when:
and 11
Future Directions Slide 12
Recommendations Utilize all of the KATHY (electronic) data to obtain or estimate non-dimensionless parameters using local conditions.
Examine the non-dimensional parameters and assumptions on flow rate and power in their calculation.
Examine code simulations of FTR (i.e are codes predicting this trend?)
Slide 13
14 Where is TRACE ?
15
Presentation to the ACRS Thermal-Hydraulics Subcommittee on TRACE Assessment against KATHY Test Data P. Yarsky, T. Zaki and S. Bajorek RES/DSA 11/17/2021 1
Outline
- Background and Motivation
- KATHY Test Facility Overview
- Description of Tests
- Failure-to-Rewet Temperature
- Comparison of Data to Models
- Conclusions 11/17/2021 2
MELLLA+ Power/Flow Map 11/17/2021 3
Operating Domain and 2RPT 11/17/2021 4
Overview of ATWS-I
- ATWS event considered is a turbine trip event with turbine bypass capability (TTWBP).
- The TTWBP results in a pressure pulse, a trip of the recirculation pumps, and a loss of extraction steam to the feedwater heater cascade.
- The TTWBP ATWS is expected to yield unstable conditions and large amplitude power instability.
- Operators control reactor water level and inject boron through the standby liquid control system (SLCS) to mitigate the event.
11/17/2021 5
Transient Reactor Power Oscillation growth begins around 100s.
Oscillations are damped by operator actions. They are indiscernible after
~230s.
11/17/2021 6
Bi-modal Oscillations 11/17/2021 7
Peak Cladding Temperature Results 1478K Cyclic dryout/rewet PCT exceeds 2200°F after onset of non-linearity. High PCT observed before that point.
11/17/2021 8
Predicted Fuel Heat-up Mechanism
- Oscillation magnitude increases and the fuel undergoes periodic dryout/rewet cycling.
- As oscillation magnitude continues to grow, the rewet period of the cycle becomes insufficient to remove all of the energy accumulated in the fuel during the dryout period. This is accompanied by a ratcheting of the fuel temperature upwards after each dryout/rewet cycle.
- Once temperature ratchets up to the minimum stable film boiling temperature, the cladding surface locks into film boiling heat transfer.
- Once locked in film boiling, and while reactor power is high, fuel temperature excursion occurs.
11/17/2021 9
Fuel Heat-up and Coolability
- MELLLA+ operation exacerbates the consequences of ATWS and the NRC staff predicts that under ATWS-I conditions the cladding surface may fail to rewet leading to fuel damage.
- It is difficult in practice to ensure core coolability if the fuel might become damaged.
- The heat-up mechanism was the subject of an experimental program at KATHY.
11/17/2021 10
KATHY Facility
- Full scale bundle test facility.
- Full reactor pressure.
- Capability for natural circulation flow rate to perform stability and instability tests.
- Instrumented to measure temperature for CHF tests, adequate for indicating the failure to rewet phenomenon.
- Implements a module called SINAN that simulates reactivity feedback.
11/17/2021 11
Heater Rod Bundle Heater rods have a bottom skewed axial power shape.
The assembly includes part length rods, water rods, and spacers typical of modern BWR fuel assemblies.
Type 1 rods have higher radial peaking factor than Type 2 and so-on.
11/17/2021 12
Experimental Work
- Experimental study of fuel heat-up mechanisms during power/flow oscillatory conditions typical of ATWS-I scenarios.
- Assessment and validation of TRACE to analyze fuel heat-up during ATWS-I.
- Performed extensive testing of failure to rewet conditions at the KATHY test loop in December 2016.
11/17/2021 13
Tests with and without Feedback
- Two types of tests were conducted.
- Tests without feedback did not utilize the SINAN module and power was steadily increased to induce instability and subsequent heat-up.
- Tests with feedback used SINAN and increased the feedback coefficient to induce instability and subsequent heat-up.
11/17/2021 14
Sample Test without Feedback (407.01)
Bundle Power during 407.01 Normalized Power 1.03 1.02 1.01 Relative Power (-)
1.00 0.99 Power is slowly increased in steps until 0.98 the failure to rewet is observed.
0.97 0 200 400 600 800 1,000 1,200 Time (seconds) 11/17/2021 15
Sample Test without Feedback (407.01)
Bundle Inlet Flow during 407.01 Flowi kg/s 10 9 As power increases, the oscillation 8 amplitude of the inlet flow increases.
Bundle Inlet Flow (kg/sec) 7 6
5 4
3 2
1 0
800 850 900 950 1,000 1,050 1,100 1,150 Time (seconds) 11/17/2021 16
Identification of Failure-to-Rewet Temperature (TFTR)
- The predicted fuel heat-up mechanism dictates that fuel heat-up occurs once the cladding surface fails to rewet.
This failure to rewet (FTR) occurs in the TRACE predictions once the temperature of the surface exceeds the minimum stable film boiling temperature and the cladding becomes locked in a film-boiling heat transfer regime.
- For each instability test, the staff recorded the maximum thermocouple temperature observed prior to the observed failure to rewet (i.e., the maximum temperature from which the cladding surface was observed to rewet).
- The highest rewet temperature observed in the test should correspond to the TFTR assuming that the rewetting process is purely dictated by the surface temperature.
11/17/2021 17
Failure to Rewet Test 407.01 Thermocouple Data Indicating Failure to Rewet 79/5Z 59/5Z 87/6Z 525 In successive 500 dryout/rewet cycles Thermocouple Measurement (Celsius) 475 the peak and average 450 temperatures increase. The failure to rewet is observed when the 425 temperature increase 400 persists over at least a 375 full period.
350 325 300 1,080 1,085 1,090 1,095 1,100 1,105 Time (seconds) 11/17/2021 18
Thermocouple Measurements for TEST 207.03 ROD 87 FTR @790s Peak Temperature before FTR @780s 11/17/2021 19
Failure to Rewet Test 207.03 Thermocouple Data in Test Run 207.03 on Hot Rod at Two Axial Elevations 87/5Z 87/6Z 87/5X 800 775 Thermocouple Measurement (Kelvin) 750 The highest temperature prior to 725 the FTR occurs a few periods prior 700 to the FTR itself in this specific 675 case.
650 625 600 575 550 730 740 750 760 770 780 790 800 Time (seconds) 11/17/2021 20
Sample Test with Feedback (A01.02)
Dynamic Power during A01.02 Normalized Power 3
Adjusting the feedback in SINAN increases 2.5 the average power and the oscillation magnitude until FTR is observed.
2 Relative Power (-)
1.5 1
0.5 0
3,500 4,000 4,500 5,000 5,500 Time (seconds) 11/17/2021 21
Failure to Rewet Temperature Results
- TFTR Maximum is ~700 K
[800 F]
- TFTR Minimum is ~600 K
[620 F]
- TFTR Average is ~650 K
[710 F]
11/17/2021 22
Homogeneous Nucleation Plus Contact Temperature (THN+CT)
The staff compared the measured TFTR to the THN+CT because the homogeneous nucleation temperature is the lowest temperature at which liquid will spontaneous nucleate into vapor. Since homogeneous nucleation is the lowest such temperature, it can be expected to bound the TFTR regardless of any local processes affecting film boiling or rewetting.
+ = +
11/17/2021 23
Various Models of Tmin 11/17/2021 24
Preliminary TRACE/INTEMP Comparison Dialing in Tmin to a value between 600 and 620 K yields TRACE results that predict similar heat-ups with the KATHY measurements.
If Tmin is too high, TRACE does not predict significant heat-up.
11/17/2021 25
Interim Approach
- As a conservative approximation, fuel heat-up can be modeled assuming:
Tmin = THN+CT
- TRACE has been used successfully to perform confirmatory analyses of ATWS-I events for MELLLA+ BWRs.
- The NRC staff has performed a more detailed assessment of TRACE through a more thorough study of the KATHY experimental results.
11/17/2021 26
Stage 1 TRACE Assessment
- Step 1: Use steady-state experimental results to adjust empirical parameters (i.e., spacer loss coefficients and critical heat flux multipliers)
- Step 2: Define figures of merit (FOMs) associated with key phenomenology and phases of the transient
- Step 3: Compare TRACE results to experimental results for FOMs using default and interim approaches for Tmin.
11/17/2021 27
Figures of Merit 11/17/2021 28
Sample TRACE Comparison -
Feedwater Temperature Transient
- TRACE with Tmin =
THN+CT shows slightly earlier heat-up and, consequently, higher temperatures, but timing and rate are in reasonable agreement.
- Higher Tmin values preclude TRACE prediction of the observed heat-up.
11/17/2021 29
Summary of TRACE Assessment
- TRACE in reasonable to excellent agreement for all FOMs when Tmin is set to the THN+CT (i.e.,
Option 6).
- When the default option is used (i.e., Groeneveld-Stewart or Option 0) the temperature related FOMs are in minimal agreement.
11/17/2021 30
Stage 2 TRACE Assessment
- Use statistical methods to study impact of constitutive models on the assessment to identify candidate constitutive models for possible improvements.
- Relies on a two-step process with uncertainty parameters in TRACE, Monte Carlo driven sensitivity calculations, and Morris screening to determine constitutive models with the biggest impact.
11/17/2021 31
New FOMs Temp TRACE Test tFTR TRACE tFTR Test Time
- FOM-Temperature - difference in max cladding inner temperature (~PCT)
- FOM-Time - difference in a time of FTR
- FOM-Integral - average difference in max cladding inner temperature 11/17/2021 32
Two Step Process
- Single Parameter Variation
- Rank impact of the single parameter on the three FOMs.
- Slice the important parameters for the second step.
- Multiple Parameter Variation
- Determine ranges for the important parameters identified in during single parameter variation.
- Use Monte Carlo techniques and Morris Screening to establish the key parameters.
11/17/2021 33
FOM-Integral [K/%]
STS (no feedback) 6 experiments Blue SIN (feedback) 17 experiments Select top 5 Orange parameters 11/17/2021 34
STS (no feedback) Select top 8 FOM-Time [seconds/%]
6 experiments Blue SIN (feedback) 17 experiments Orange parameters 11/17/2021 35
FOM-Temperature [K/%]
STS (no feedback) 6 experiments Blue SIN (feedback) Select top 15 17 experiments Orange parameters 11/17/2021 36
Important Parameters Rank Integral Temperature Time
- FOM-Integral 1 2
1010 1012 1031 1010 1010 1030 3 1022 kfac 1014
- Top 5 10 parameters 4 5
1013 1014 1036 1009 1028 1022 6 1028 1037 1001 7 1029 1001 1041 8 1041 1029 1012
- FOM-Time 9 10 1042 1030 1042 1041 1008 1042 11 1003 1011 1033
- Top 8 10 parameters 12 1004 1008 1015 13 1001 1030 1005 14 1037 1002 1036 15 1009 1015 kfac
- FOM-Temperature 16 1039 1028 1029 17 1005 1014 1002 18 1031 1022 1003
- Top 15 19 parameters 19 20 1033 1000 1013 1000 1037 1006 11/17/2021 37
FOM-Integral [K/%]
STS (no feedback) 6 experiments Blue Extended to 10 because of duplication SIN (feedback) 17 experiments Select top 5 Orange parameters 11/17/2021 38
STS (no feedback) Select top 8 FOM-Time [seconds/%]
Extended to 10 6 experiments because of duplication Blue SIN (feedback) 17 experiments Orange parameters 11/17/2021 39
FOM-Temperature [K/%]
STS (no feedback) 6 experiments Blue Extended to 19 because of duplication SIN (feedback) Select top 15 17 experiments Orange parameters 11/17/2021 40
Morris screening
- Qualitative measure of importance (µ/µ*) and interaction ()
of input parameters on the output FOM 11/17/2021 41
6 Experiments without Feedback Single experiment Composite of all experiments 11/17/2021 42
17 Experiments with Feedback Single experiment Composite of all experiments 11/17/2021 43
Key Constitutive Models
- The importance and ranking of the key constitutive models is not surprising
- 1010 - Minimum Stable Film Boiling
- 1030 - Annular/Mist Flow Interfacial Drag
- 1041 - Critical Heat Flux
- 1012 - Subcooled Boiling Heat Transfer
- 1022 - Wall Drag 11/17/2021 44
Conclusions
- TRACE produces reasonable, but slightly conservative, predictions of fuel heat-up during postulated ATWS-I when the Tmin option is set to the THN+CT.
- This approach will remain our standard practice.
- Detailed assessment indicates that the key models affecting TRACE predictions are consistent with our expectations and Phenomenon Identification and Ranking Table (PIRT).
- Better agreement could be garnered with improvements to the minimum stable film boiling model, but it is not deemed necessary at this stage.
11/17/2021 45
- Questions?
11/17/2021 46
BACKUP SLIDES 11/17/2021 47
Regulatory Purpose
- To provide confirmatory analysis of Anticipated Transients Without SCRAM (ATWS) events for boiling water reactors (BWRs) operating with a maximum extended load line limit analysis plus (MELLLA+) expanded operating domain.
- It is the intent to use TRAC/RELAP Advanced Computational Engine (TRACE) to simulate postulated MELLLA+ ATWS events to study plant transient response, consequences, and effectiveness of mitigating actions.
- The Office of Nuclear Reactor Regulation (NRR) uses the confirmatory analysis results to guide the review. These analyses improve the efficiency of the overall review effort by focusing staff RAIs on key issues and, in certain cases, eliminating the need for some RAIs.
11/17/2021 48
MELLLA+ Domain
- MELLLA+ is an expanded BWR operating domain allowing high thermal power (120% of rated thermal power (RTP)) at low flow (80%
of rated core flow (RCF)).
- MELLLA+ operation introduces new aspects to the progression of ATWS events.
11/17/2021 49
Safety Significance of the FCW
- During ATWS events, the reactor power is decreased by a trip of the recirculation pumps (2RPT).
- The power and flow decrease as the pumps run down.
- Power then increases due to a decrease in feedwater temperature.
- When the flow rate is low (80 %RCF), the 2RPT becomes less effective in the reduction of gross core power.
11/17/2021 50
Operating Domain and RPT 11/17/2021 51
MELLLA+ Benefits
- The flow control window (FCW) allows
- Global reactivity changes without control blade motion
- Spectral shift operation
- Reduces incidence of Fuel-Clad Interaction (FCI) fuel failure
- Reduced control blade pattern swaps
- Reduces incidence of pellet-clad interaction fuel failure
- Low-flow depletion (spectral shift)
- Improves fuel cycle economics 11/17/2021 52
Feedback and Instability Higher thermal power following 2RPT greatly increases the chances that the reactor will undergo unstable power oscillations. ATWS leading to instability is ATWS-I 11/17/2021 53
ATWS-I Results
- Representative Case:
- Generic BWR/5 model
- TTWBP with 100% bypass capacity
- Initial core flow rate is 85% rated
- Initial power is 120% of originally licensed thermal power (OLTP)
- Core exposure is peak-hot-excess (PHE)
- Operators attempt to control reactor water level to top of active fuel (TAF) starting 110 seconds into event
- Operators initiate SLCS at 120 seconds into event 11/17/2021 54
PHE ATWS-I Case - Sequence of Events 11/17/2021 55
Base Case Conclusions
- Point in cycle studies confirm that PHE is the most limiting state-point
- Large amplitude regional power oscillations develop (modal coupling with frequency doubling).
- High amplitude power oscillations (local) results in calculation of high PCT (~1700 K [2600 °F]).
- Operator action to reduce level
- Effective in reducing FW flow, limiting increase in core inlet subcooling and eventually eliminating inlet subcooling.
- Effective in suppressing power oscillations and reducing core power level.
11/17/2021 56
Methodology Overview: Codes
- TRACE
- TRACE simulates the thermal-hydraulic response of the plant and core
- PARCS
- PARCS simulates the neutron kinetics in three-dimensions
- SCALE/Polaris
- The Polaris sequence calculates parameterized nuclear data
- Coupled gamma/neutron transport calculations with MCNP establishes direct energy deposition factors
- FAST
- FAST simulates fuel thermo-physical behavior over exposure and is used to calculate initial gas gap properties and other related thermal-mechanical conditions 11/17/2021 57
Methodology Overview: Tools
- Scripting Tools
- Scripting tools are used for automatic generation of core inputs for TRACE
- SNAP
- Visualization tool used for generating TRACE thermal-hydraulic, control system, and heat structure models
- GenPMAXS
- Code that converts SCALE output into PMAX files for use in PARCS 11/17/2021 58
Methods: Fuel Properties Input Data Calculations Output Data PMAX Nuclear SCALE GenPMAXS Data Files MCNP DED Factors Lattice Designs Functionalized Gap Rod FAST Pressure and Design Composition 11/17/2021 59
Methods: Systems Analysis 11/17/2021 60
Comparison of Tmin to TFTR Data Cladding temperature measured at inside surface, so outside surface expected to be
~10K higher, so a bias on this order is expected. THN+CT models compare well.
11/17/2021 61
KATHY Test Indices 11/17/2021 62
Uncertainty Ranges 11/17/2021 63
Film Boiling Heat Transfer Coefficient
- Average from Experiment: ~2.0 kW/m2-K
- Average from TRACE:
~1.3 kW/m2-K
- TRACE in good agreement, slightly conservative 11/17/2021 64
Test 11 - Regional Coupling Test 11/17/2021 65