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