ML25237A266
| ML25237A266 | |
| Person / Time | |
|---|---|
| Issue date: | 08/25/2025 |
| From: | Derek Widmayer Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| NRC-0376 | |
| Download: ML25237A266 (1) | |
Text
Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Reactor Safeguards Docket Number:
(n/a)
Location:
Rockville, Maryland Date:
Wednesday, July 9, 2025 Work Order No.:
NRC-0376 Pages 1-132 NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers 1716 14th Street, N.W.
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
+ + + + +
5 WEDNESDAY, 6
JULY 9, 2025 7
+ + + + +
8 The meeting was convened at Two White 9
Flint
- North, 11545 Rockville
- Pike, Rockville, 10 Maryland, and via videoconference, at 8:30 a.m. EDT, 11 Walter L. Kirchner, Chair, presiding.
12 13 PRESENT:
14 WALT KIRCHNER, Chair, ACRS 15 GREG HALNON, Vice Chair, ACRS 16 DAVE PETTI, Member-at-Large, ACRS 17 RON BALLINGER, Member, ACRS 18 VICKI BIER, Member, ACRS 19 VESNA DIMITRIJEVIC, Member, ACRS
- 20 CRAIG HARRINGTON, Member, ACRS 21 ROBERT MARTIN, Member, ACRS 22 SCOTT PALMTAG, Member, ACRS 23 TOM ROBERTS, Member, ACRS 24 MATT SUNSERI, Member, ACRS 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 DESIGNATED FEDERAL OFFICIAL:
1 DEREK WIDMAYER 2
3 NRC STAFF PRESENT:
4 THOMAS DASHIELL, PMDA, ACRS 5
TIM DRZEWIECKI, NRR, DANU, UTB2 6
DENISE McGOVERN, NRR, DANU, UAL2 7
TRAVIS TATE, NRR, DANU, UTB1 8
9 ALSO PRESENT:
10 BRIAN FROESE, X-energy 11 MILAN HANUS, X-energy 12 DREW NIGH, X-energy 13 MATT THOMAS, X-energy 14 JAMES TOMPKINS, X-energy 15 16
- Present via Teams 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 CONTENTS 1
2 Call to Order and Opening Remarks by the 5
3 ACRS Chairman 4
Walt Kirchner, Chair 5
Comments Regarding Member Ron Ballinger's....
8 6
Service on the ACRS 7
Walt Kirchner, Chair 8
X-Energy Topical Report on Mechanistic
..... 10 9
Source Term Approach 10 Remarks from Subcommittee Chairman
...... 10 11 Robert Martin, Member 12 Presentation by Applicant........... 11 13 Milan Hanus, X-energy 14 Presentation by NRC Staff........... 83 15 Tim Drzewiecki 16 Opportunity for Public Comment (None)...... 98 17 Committee Deliberation on Letter Report..... 98 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
8:30 a.m.
2 CHAIR KIRCHNER: Good morning. The 3
meeting will now come to order. This is the first day 4
of the 727th meeting of the Advisory Committee on 5
Reactor Safeguards, ACRS.
6 I'm Walt Kirchner, Chairman of the ACRS.
7 ACRS members in attendance in-person are Ron 8
Ballinger, Vicki Bier, Greg Halnon, Craig Harrington, 9
Robert Martin, Scott Palmtag, Dave Petti, Thomas 10 Roberts, and Matt Sunseri. ACRS Member Vesna 11 Dimitrijevic is participating virtually via Teams.
12 If I've missed anyone, either ACRS members 13 or consultants, please speak up now.
14 Derek Widmayer of the ACRS staff is the 15 Designated Federal Officer for this morning's full 16 Committee meeting.
17 No member conflicts of interest were 18 identified. And I note that we have a quorum.
19 The ACRS was established by statute and is 20 governed by the Federal Advisory Committee Act, or 21 FACA. The NRC implements FACA in accordance with our 22 regulations.
23 Additionally, in accordance with Sections 24 29 and 182(b) of the Atomic Energy Act, the Advisory 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 Committee on Reactor Safeguards shall advise the 1
Commission with regard to hazards of proposed or 2
existing facilities and the adequacy of proposed 3
safety standards.
4 In addition, the ACRS is implementing 5
Executive Order 14300, ordering the reform of the 6
Nuclear Regulatory Commission, dated May 23, 2025.
7 Section 4(b) of the Executive Order states, in part, 8
that the "functions of the Advisory Committee on 9
Reactor Safeguards shall be reduced to the minimum 10 necessary to fulfill ACRS's statutory obligations,"
11 and that review by ACRS shall focus on issues that are 12 unique, novel, and noteworthy.
13 Reviewing/reporting on new reactor 14 facilities and proposed safety standards are the 15 minimum statutory functions of the ACRS under Sections 16 29 and 182(b) of the Atomic Energy Act. The 17 Commission may refer additional duties to the ACRS in 18 accordance with the Act.
19 Per these regulations and the Committee's 20 bylaws, the ACRS speaks only through its published 21 Letter Reports. All member comments, therefore, 22 should be regarded as only the individual opinion of 23 that member and not a Committee position.
24 All relevant information related to ACRS 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 activities, such as letters, rules for meeting 1
participation, and transcripts, are located on the NRC 2
public website and can be easily found by typing 3
"About Us ACRS" in the search field on the NRC's home 4
page.
5 The ACRS, consistent with the agency's 6
value of public transparency and regulation of nuclear 7
facilities, provides opportunity for public input and 8
comment during our proceedings. For this full 9
committee meeting, we have received no written 10 statements. However, written statements may be 11 forwarded to today's Designated Federal Officer. We 12 have also set aside time at the end of this meeting 13 for public comments.
14 A transcript of the meeting is being kept 15 and will be posted on our website. When addressing 16 the Committee, the participants should first identify 17 themselves and speak with sufficient clarity and 18 volume, so that they may be readily heard. If you are 19 not speaking, please mute your computer on Teams. If 20 you are participating by phone, press star-6 to mute 21 your phone and star-5 to raise your hand on Teams.
22 The Teams chat feature will not be 23 available for use during the meeting.
24 For everyone in the room, please put your 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 electronic devices in silent mode and mute your laptop 1
microphone and speakers. In addition, please keep 2
sidebar discussions in the room to a minimum, since 3
the ceiling microphones are live.
4 For presenters, we remind you that these 5
table microphones are unidirectional and you'll need 6
to speak directly into the front of the microphone to 7
be heard online.
8 Finally, if you have any feedback for the 9
ACRS about today's meeting, we encourage you to fill 10 out the public meeting feedback form on the NRC's 11 website.
12 During today's meeting, we will consider 13 the topic of X-energy's Topical Report on Mechanistic 14 Source Term. But before I pass the microphone to Bob 15 Martin, our Subcommittee Chair for today's meeting, I 16 want to note that Dr. Ballinger completes his third 17 term with the Committee in August. Ron joined us on 18 August 4th of 2013. So he's our ancient mariner on 19 the Committee. And I want to note that, in addition 20 to being a professor emeritus at MIT, notably, Ron has 21 led several major projects for the Committee.
22 The APR-1400 review, that was the C-E 80+
23 that Korea Hydro and Nuclear Power brought to the NRC 24 for review by both the agency and the Committee. He 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 also led the SHINE Medical Isotope Project. That was 1
an interesting one because it was a fusion-fission 2
hybrid approach to medical isotope production. And, 3
most recently, he's led our review of increasing 4
enrichment rulemaking activities.
5 So, we thank you for that, Ron. I want to 6
note Ron will continue as an ACRS consultant after his 7
term ends. And so, on behalf of the Committee, I want 8
to thank you for your service, your expertise, your 9
collegiality, and your random acts of kindness, which 10 I was the beneficiary of, as well as others.
11 So, Ron, thank you very much. And would 12 you like to make any comments, Ron?
13 MEMBER BALLINGER: No.
14 (Laughter.)
15 PARTICIPANT: Another random act of 16 kindness.
17 (Laughter.)
18 CHAIR KIRCHNER: Okay. So, with that, let 19 me turn to Bob Martin, who will lead us in the 20 morning's topic. Go ahead, Bob.
21 MEMBER MARTIN: Okay. Thank you, 22 Chairman.
23 As Walt
- noted, we are discussing 24 X-energy's Mechanistic Source Term Topical Report. We 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 are reviewing this particular report because it 1
presents foundational methods that directly support 2
the offsite dose calculations required to demonstrate 3
compliance with 10 CFR 50.34, "Contents of 4
Applications," for the Xe-100 high-temperature gas 5
reactor.
6 Reports its novel aspects include the 7
reliance on TRISO functional containment, event-8 specific mechanistic source term modeling, and the 9
applicant's in-house XSTERM codes. Early examination 10 of these features ensure they align with regulatory 11 expectations and provide a robust and performance-12 based safety case for good licensing.
13 Of course, as you noted, we had a 14 Subcommittee meeting last month where we had 15 presentations from X-energy and, of course, our own 16 staff. And we will be doing letter deliberation after 17 these formalities.
18 But right now, we have X-energy -- unless 19 anybody from the staff wanted to say anything to 20 introduce anything. But other than that, X-energy, 21 you can proceed with your presentation.
22 Please introduce yourself.
23 MR. HANUS: Thank you. Thank you very 24 much for having us here and for your offer for coming 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 here. I heard it was that a tape from an interview.
1 So, we really appreciate you coming here and being 2
here today.
3 My name is Milan Hanus. I am the software 4
Engineering Manager at X-energy and, also, developed 5
the mechanistic source term model and the code XSTERM, 6
as mentioned. And today, I'll be presenting the 7
mechanistic source term approach, which we provided to 8
the NRC for U.S. licensing Topical Report.
9 So I will first introduce some generic 10 efforts that X-energy has taken to model the source 11 terms for the Xe-100 paper, and then the actual models 12 with how we are doing the system generation and 13 propagation through the system, and at the end there 14 will be the time for focused questions and answers.
15 16 In the Topical Report that we presented to 17 the NRC, we included the description of the 18 mechanistic system with us that is used to determine 19 the radionuclide transport phenomena for the 20 preliminary analysis and deliberation of the Xe-100 21 and for establishing the safety case.
22 What is not included in this Topical 23 Report, in this Topical Report, is the actual 24 implementation details and the evaluation of the cases 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 and outputs, which we are planning to include in a 1
future Topical Report on the goal to XSTERM which 2
implements these models.
3 As a basis for the Topical Report and for 4
the whole mechanistic system modeling, we used some 5
documents which are listed here:
6 The Risk-informed Performance-based 7
Licensing Basis Approach which provides the link to 8
NEI 18-04 methodology that we adopted for the system 9
and engineered the safety analysis modeling and 10 application.
11 The Topical Report on the Transient and 12 Safety Analysis Methodology, which was also presented 13 last time, the December report, which describes the 14 approach that X-energy takes or took to elevate the 15 DBAs and to basically evaluate the transients and 16 safety of the reactor.
17 Principal Design Criteria and the TRISO-X 18 Pebble Fuel Qualification Methodology provides the 19 data, the parameters, for the models.
20 And the last document defines the 21 dispersion factors that we used to calculate the final 22 dose of the system.
23 CHAIR KIRCHNER: Milan?
24 MR. HANUS: Uh-hum?
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 CHAIR KIRCHNER: Would you pull your 1
microphone closer to you? You're soft-spoken.
2 MR. HANUS: Yes. I will hold it like 3
this.
4 CHAIR KIRCHNER: Okay. Perfect.
5 MR. HANUS: Sorry about that, yes.
6 So, X-energy looked at and model the 7
mechanistic source terms. It's risk-informed, 8
performance-based, and we had the guidances in 9
Regulatory Guidance SECY-93-092, in which we justified 10 the approach by attempting to model the deterministic 11 generation of the source terms in such detail using 12 sufficient data and adequate events.
13 The systems are different from the usual 14 systems which are based on severe core damage and 15 accidents. They are event-specific, determined 16 mechanistically using models of the fission product 17 generation and transport and account for the inherent 18 and passive design features of the reactor design and 19 all product release barriers that constitute the 20 functional containment.
21 Here, on the left, you can see the fuel 22 element that is used in Xe-100, the pebbles with the 23 TRISO particles within them. The container itself, 24 the layers of the TRISO particles, the matrix of 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
13 pebbles from the first radionuclide release barriers.
1 Out of these barriers we have the helium pressure 2
- boundary, which forms another
- level, another 3
functional containment layer.
4 And even in the case of the break in the 5
pressure boundary, if the helium tries to escape to 6
the reactor building, which technically forms another 7
release barrier for the radionuclides, but we looked 8
at it, the reactor building in the safety 9
calculations, as we have enough safety margin on the 10 first barriers actually for the TRISO fuel, you know, 11 the fission product capabilities.
12 So the fuel is represented by this sphere, 13 and it shows the different TRISO layers, as well as 14 the mechanisms by which the radionuclides can 15 potentially escape or get past those layers.
16 So the fission products are generated by 17 fission of the -- but also to the fission of the heavy 18 metal contamination in the pyro carbon layers in the 19 matrix. They can potentially diffuse out of the 20 pebbles under high temperatures and get released to 21 the core with the helium that, you know, is flowing 22 around the pebbles, and get transported --
23 Another mechanistic source term, the model 24 and the events that can happen in the fission boundary 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 regarding the fission product transport. They go to 1
the fission boundary surfaces. They lift off by the 2
helium from those surfaces, actually, the helium 3
steam. And also, the postulation and attachment or 4
absorption of the helium plates to the dust and 5
transport to the fission boundary.
6 As I mentioned previously, we do not 7
credit the reactor building. So here, I'll not speak 8
about that reactor building in this presentation and 9
sort of looking for the --
10 MEMBER MARTIN: Milan?
11 MR. HANUS: Mm-hmm?
12 MEMBER MARTIN: At our Subcommittee 13 meeting, the staff noted the lack of a PIRT, but, you 14 know, to do mechanistic source term, you've obviously 15 identified a lot of the PIRT. You know, there's kind 16 of a formality involving the application of Reg Guide 17 1.203 to have that exercise where you bring in the 18 experts to determine what's important, and then, of 19 course, downstream it usually feeds the design of 20 these codes.
21 So I'm certainly going to give you credit 22 because you can't do a mechanistic source term without 23 some insight onto the phenomena, but at the same time, 24 do you have a plan now in place to kind of complete 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 that formality of the PIRT document.
1 MR. HANUS: Yes. So, we already went 2
through one iteration of the PIRT process.
3 MEMBER MARTIN: Okay.
4 MR. HANUS: So you have a table for 5
phenomena identification and ranking. And we are 6
planning to revise it, as per the suggestions of the 7
Committee as well. You know, some of the phenomena 8
might be assessed as important, especially those 9
related to the fuel particle -- the particle barrier 10 operations. But if you have that --
11 MEMBER MARTIN: So it's a process you 12 started already?
13 MR. HANUS: Yes, yes.
14 MEMBER MARTIN: And, certainly, by the 15 time we get to the operating license activities, we'll 16 be seeing that document?
17 MR. HANUS: Yes. And that document is 18 also used -- I will just focus to this site. We, 19 yes, will continue the process of validation and 20 verification of the code that implements these 21 methods. And that validation, the case is selected 22 for the validation now based on the PIRT. So, you 23 know, the specific phenomena.
24 MEMBER MARTIN: Okay. 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
16 MR. THOMAS: Matt Thomas, Licensing 1
Manager at X-energy.
2 So the PIRT that we're referring to is the 3
one that's included as part of the TSAM Topical, and 4
that's where this will be documented at.
5 MEMBER MARTIN: Okay. Okay.
6 MR. THOMAS: Is that right, Milan?
7 MR. HANUS: Mm-hmm. Yes, yes.
8 MEMBER PETTI: Milan, just a question.
9 And I
can't remember from when we had the 10 Subcommittee. Do you account for vessel breathing as 11 a barrier? Do you know what I mean by "vessel 12 breathing"?
13 MR. HANUS: Vessel bleeding (sic) --
14 MEMBER PETTI: Vessel breathing. You 15 know, when you have the break, the helium goes out; 16 the temperatures go up. Then, eventually, the 17 temperatures will turn around and you'll pull gas back 18 into --
19 MR. HANUS: Oh.
20 MEMBER PETTI: -- the reactor. There's 21 no --
22 MR. HANUS: Air exchange?
23 MEMBER PETTI: Air exchange, and you call 24 it -- "vessel breathing" is the short term vernacular 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 of the gas reactor we're talking about.
1 Is that part of -- if the model says that, 2
you take that -- hydraulically?
3 MR. HANUS: No. Yes, we don't do it 4
either way.
5 MEMBER PETTI: Okay.
6 MEMBER MARTIN: Well, that might just fall 7
out of the codes that you're using, right? I mean, if 8
it depressurized, you know, there's a momentum, terms 9
that going to have that vessel or the system below the 10 pressure of whatever you have modeled outside. And 11 that, it should suck in, just kind of inherent with 12 the governing equations and standard constituents --
13 MR. HANUS: Yes, you know, that's a 14 question for --
15 (Simultaneous speaking.)
16 MR. HANUS: In terms of the mechanistic 17 data, the XSTERM models, they get the flow rates from 18
- outside, from performance operations.
Those 19 operations are taking this into account and ensure the 20 task (audio interference).
21 MR. NIGH: Yes, this is Drew Nigh, Manager 22 of Risk-Informed Safety Analysis. I can speak to 23 that.
24 Dave, yes, we do take credit for, 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
18 take credit for it, and we do account for vessel 1
breathing as part of the mechanistic source term 2
calculations. So we reduce the source term -- well, 3
we quantify the source term that escapes from fuel 4
pebbles, and then we only release a fraction of that 5
source term, based on what percentage is released or 6
what portion is released while there's a driving 7
force, like pressurized helium, to carry it out of the 8
vessel. And then, afterwards, we do release a little 9
bit more, based on continued heat-up of the helium and 10 air in the vessel.
11 MEMBER PETTI: Okay.
12 MEMBER MARTIN: For points driven in a 13 later term and --
14 MEMBER PETTI: Yes, thanks.
15 MR. HANUS: All right. So thanks. Thank 16 you for saying that.
17 And now, I'll get into the overall 18 picture, a summary of the mechanistic system models 19 that are part of the MST approach of X-energy.
20 Those models are detailed in the Topical 21 Report in appendices A through G. And the first, FPM, 22 the Fuel Performance Model. The second, THM, 23 Thermodynamics Model. SOLM, the time-dependent 24 radionuclide release and diffusion. And GASM is 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
19 steady-state gaseous fission product release and 1
transport. And DUSTM is the Dust Production Model.
2 HPBM models the helium fission boundary. CORRM is the 3
corrosion model that we can use for the moisture and 4
gas calculations.
5 Other models and methodology are 6
documented in the code, as well as the case 7
statements, but they are not (indiscernible due to 8
accent) the mechanistic source term methodology 9
(indiscernible due to accent) that are mentioned in 10 the introductory section of the Topical Report.
11 And, basically, this simulates the reactor 12 power operation, the transient, and orchestrates the 13 other modules to provide a source term picture in the 14 fully indicated calculation. These are used mainly to 15 establish the basis for the transient calculations by 16 generating the (indiscernible due to accent) in the 17 plant. And it's used, also, as a basis for the DTM 18 (phonetic) calculations by the feedback model, but 19 they are outside the scope of this Topical Report.
20 This version and those models are based on 21 the Topical Report that we included previously. That 22 was the previous version and dose calculation 23 methodology, which defines the appropriate dose 24 conversion factors combined with the plant operations.
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 In this table, the (indiscernible due to 1
accent), I included the different codes. There are 2
many more (indiscernible due to accent) know about, 3
but they (indiscernible due to accent) particle bonds 4
of the system, generation, propagation, and transport.
5 Our approach is to run coherent code and coherent 6
methodology that contains all the models, and 7
therefore, there are a couple of ways we can easily 8
modify and adjust these models, as a result of being 9
in contact with the authors of the codes that might 10 not be available anymore.
11 MEMBER MARTIN: Milan, so one of the 12 points you made in the Topical Report is that you're 13 doing mechanistic source term for each licensing basis 14 event. Now, the events will be characterized by the 15 phenomena, of course, associated and there will be 16 different phenomena and they'll be weighted different.
17 I'm going to kind of lead the witness here 18 a little bit. I feel like that you're not going to be 19 turning anything off with an XSTERM, right, as far as 20 phenomena is concerned to capture it? You're really 21 just using the thermal hydraulic boundary conditions 22 from, say, the Flownex or the GOTHIC, and that mass 23 flow rates and such are really what's driving the 24 transport?
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 MR. HANUS: Yes, to --
1 MEMBER MARTIN: But it's XSTERM is the 2
same XSTERM in every case? There's no tweaking of 3
phenomena, phenomena models, constituent packages?
4 MR. HANUS: No, the code is the same. The 5
methodology is the same, and the selection of the 6
options in the code is the same. V&V changed a bit 7
the options, for example, for the validation, which is 8
a different -- you know, it's not Xe-100. So, that's 9
why the scope of the code is bigger, is larger than is 10 presented here with the same model. But probably the 11 phenomena that are patented for that, defined by the 12 code, use the same code.
13 MEMBER MARTIN: It does kind of create a 14 V&V challenge. I mean, it's a classic challenge of 15 doing the V&V for the separate effects, as well as the 16 integral. And maybe you're going to be talking about 17 this here in a few slides, about, you know, the 18 approach to V&V where you do target separate effects, 19 and then, also, have some integral tests that can 20 verify or validate the relative contributions of the 21 different phenomena.
22 MEMBER HALNON: This is Greg.
23 Yes, that kind of asked my question.
24 Maybe we'll see it in a few minutes. But I'm not 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
22 code guy, but I'm an operator. So, when I look at a 1
lot of procedures and I look at pieces being put 2
together, I get concerned about the bounding -- the 3
assumptions for each procedure and relative to the 4
applicability of that to the final product, whether or 5
not they're all in sync and applicable.
6 So, when I see all these codes, it makes 7
me wonder if all the input assumptions and the 8
assumptions that the code uses to say these are the 9
boundary conditions that I'm working on -- how do you 10 fit all that together and make sure that everything is 11 copacetic or in sync, so that your final product 12 doesn't have some inappropriability [sic] because 13 there was some bounding assumption for one code that 14 doesn't get met down the path?
15 I don't know if I asked that question --
16 I asked it in an operator way, trying to mix it? But 17 do you see what I'm trying to --
18 MEMBER MARTIN: Sure, sure. It's a 19 similar question. I mean, there is just XSTERM, 20 right? I mean, obviously, this slide has codes 21 implementing similar capability. So those codes are 22 really -- I mean, I don't know if you can do a code-23 to-code comparison as part of some V&V exercise.
24 Right?
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 MR. HANUS: Yes. Yes, we do.
1 MEMBER MARTIN: That certainly gets to 2
your question a lot when you're trying to --
3 MEMBER HALNON: Yes, but I need you to 4
translate.
5 MEMBER MARTIN: It makes sense of all the 6
individual models and how those two codes could relate 7
to each other. That is its own challenge. And, of 8
course, we can't really review that in this setting.
9 MEMBER HALNON: No, no. But at the end, 10 if things all fit together, that's great. And if they 11 hand off appropriately, they're handing off, also, 12 their input assumptions as part of that for the codes.
13 I want to make sure that there's not something being 14 invalidated downstream, based on the fact that you 15 didn't put assumptions for something downstream that's 16 different.
17 MEMBER MARTIN: So I'm going to answer 18 your question. So XSTERM is basically going to have 19 all this capability. There will be, say, lower-scale 20 phenomena that is happening, of course, beginning in 21 the core. And it's going to be kind of feeding in the 22 same way that you might feed these other codes, you 23 know, if you had stacked them together, a couple of 24 them together. So it's all being taken care of within 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 XSTERM term. So there really is no human touch 1
involved in packaging all these different phenomenon 2
into one package.
3 So, does that help?
4 MEMBER HALNON: Yes.
5 (Laughter.)
6 MR. HANUS: I don't know if that answered 7
the question. The question, the fundamental question 8
is, you know, it is a difficult question because each 9
of those models has its own uncertainties within it, 10 which can be obtained from the original authors of the 11 models, and we try our best estimate the different 12 answers and to quantify these answers as well.
13 And the combination of the uncertainties, 14 they can stick up, obviously. What I can say is that, 15 in the decision of making the code parameters, the 16 parameters for the code, we generally adopt a 17 conservative approach. When there is an uncertainty, 18 we use what is conservative. (Unintelligible.) It's 19 like more and more conservative, eventually.
20 MEMBER HALNON: Okay. So you've biased 21 everything to the conservative side to make sure that 22 the end product is also conservative?
23 MR. HANUS: Yes.
24 MEMBER MARTIN: You biased it within 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
25 code itself or as inputs to the code? Or both?
1 MR. HANUS: It's both. I would say it's 2
both, but -- well, (indiscernible due to accent) the 3
code balances calculations and if there's a choice 4
between using two different calculations, we use the 5
more conservative calculation.
6 MEMBER MARTIN: Okay. So, then we have 7
within XSTERM a way to do best estimate as well as, 8
say, evaluation model type also?
9 MR. HANUS: Yes, yes. So the code has 10 different options. For the purposes of the safety 11 analysis, we used the most conservative ones. For the 12 purpose of scoping and design, we use the --
13 MEMBER MARTIN: So, like when you present 14 your V&V, I mean, would you turn on -- would you run 15 XSTERM twice, once with kind of the conservative EM 16 approach, and then, one with a more best estimate 17 selection?
18 MR. HANUS: We use the more conservative 19 assumption. And, you know, we can do both. The 20 calculator now (indiscernible due to accent) is 21 focused on the safety analysis, so we use the more 22 conservative pathway.
23 MEMBER MARTIN: Well, I might suggest, 24 again, going down the road -- this is just 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
26 beginning -- that when you present that information, 1
since you have that capability in the code to select, 2
there be more best estimate models; that when you 3
present your V&V, you run it twice. And so, you know, 4
for your external reviewers, they can very easily see 5
those deltas, you know, between the data that you 6
might have and the best estimate models. And when you 7
put it in sequence, to have the more conservative 8
- models, do you know immediately where those 9
conservatives are and the effect of them?
10 MR. HANUS: Mm-hmm.
11 MEMBER HALNON: So it's very informative 12 to do that twice, and you have that capability.
13 MR. HANUS: Yes, and thank you. Thank you 14 for this suggestion. It sounds --
15 (Simultaneous speaking.)
16 MEMBER BIER: One other question. This is 17 Vicki Bier. Following up on Greg's point, or I think 18 Greg's point, about that you may be conglomerating 19 different models or submodels with slightly different 20 assumptions, and whatever, this is kind of abstract.
21 And my point may not even apply to what you're doing.
22 But is it possible that what seems to be conservative 23 in one submodel may then turn out not to be 24 conservative in another? And you can have, you 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
27 conservatism throughout that ends up being physically 1
unrealistic or impossible. Or, you know, have you 2
thought about that? Have you encountered that in what 3
you're doing?
4 MR. HANUS: Yeah, I would say yes. We 5
encountered that. You know, there is an example 6
(indiscernible due to accent) in some DTM calculations 7
that, because the wall thickness of the steam 8
generator tubes, the wall base, the less conservative 9
-- or the more important variation -- the less thick 10 the tube is, the more (indiscernible due to accent) 11 out of the tubes. But it has the opposite effect on 12 the (indiscernible due to accent).
13 That's why we need to study the source 14 term for each event and take into account for the 15 final --
16 MEMBER MARTIN: Well, certainly in that 17 particular example, you'll think about a case where 18 you look at how rapidly the temperature of the tubes 19 will change. Right? There's oftentimes about 100 F, 20 or whatever, in it. And clearly, if you had a thicker 21
-- if you assumed a conservative from a heat transfer 22 standpoint, that would slow it down. If you used, of 23 course, a thin one, you know, or thinner than 24 realistic, it would be too conservative. So 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
28 really would want a best estimate in that particular 1
case, but it would be counter to maybe other 2
scenarios.
3 So that kind of gets back to the question 4
of event-specific -- not only event-specific, but 5
figures of merit-specific. You know, I think with a 6
lot of these advanced reactors, we're always concerned 7
with the fuel, but I think we're also concerned with 8
every other thing that could possibly break, including 9
generator tubes are probably pretty close to the top 10 of that list.
11 And so, I think you might find that staff 12 or a body like ours will be very sensitive to those 13 kind of assumptions. Because, to Vicki's point, on 14 one setting, it's conservative.
15 MR. HANUS: Yes.
16 MEMBER MARTIN: And at another setting, 17 it's not.
18 MR. HANUS: Yes.
19 MEMBER MARTIN: You know, it's not just 20 we're looking at fuel temperature. We'll be looking 21
-- you know, you've really got to look at anything 22 that is certainly a pressure boundary barrier or a 23 safety system, you know, in those particular cases.
24 So we're sensitive to that. But, good question.
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 MR. HANUS: Okay. So now, I'll talk about 1
the code itself, which the code is the foundation of 2
the models. So it is part of the whole safety 3
analysis evaluation model which is planned in the TSAM 4
report. And we developed it under our qualifications 5
program with the goal to (indiscernible due to 6
accent), for which you always need to perform 7
extensive validation and EQ. And (indiscernible due 8
to accent) is being performed after we staged the 9
validation of phases, depending on the (indiscernible 10 due to accent) are covered by each phase.
11 And, as mentioned previously, we used the 12 PIRT to determine the events to be evaluated in each 13 of those phases. We try to get some specific effect 14 test as well, which is difficult to get that.
15 Validation is also data for that. But it's all 16 planned in the validation plans we are using.
17 We also performed the verification of the 18 code by line-by-line comparison with the defenses.
19 And we saw adequate solutions which (indiscernible due 20 to accent) change the code, basically. So, that's the 21 one slide on the validation. Again, the validation, 22 the validation process is ongoing. We will present 23 the Topical Report on the code itself (indiscernible 24 due to accent) when we are presenting the methodology 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 itself.
1 MR. THOMAS: Hold on. Just one point here 2
I wanted to make. So again, Matt Thomas, X-energy 3
Licensing Manager.
4 I
appreciate the discussion, the 5
suggestions, and whatnot, from the members. And while 6
Milan is talking much about the code itself and V&V of 7
the code itself, the Topical Report that we have in 8
front of the staff right now is for the MST theory 9
only. So, you know, maybe one day these questions and 10 suggestions will be relevant to, like, XSTERM topical 11 code -- or an XSTERM code topical. But for the topic 12 at hand right now, we're just really focused on that 13 theory.
14 Thank you.
15 MEMBER MARTIN: Of course. And our 16 comments -- obviously, it's very, very early and we're 17 telling you what we're sensitive to. So that, when 18 you get down to those things, you go, you know, I 19 remember that meeting. So this is for your benefit, 20 too. Okay?
21 I mean, I'm a little preachy. So I 22 recognize that. But at the same time we're hoping 23 that you're listening and we're helping you prepare 24 for the next time.
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 MR. THOMAS: Great. Thank you.
1 MR. HANUS: Yes, and I appreciate these 2
comments as well, this guidance. I appreciate this 3
time and your guidance, in fact, as well. So the 4
earlier, the better. Thank you.
5 MEMBER PALMTAG: This is Scott Palmtag.
6 I had a comment on your previous slide.
7 MR. HANUS: Oh, yes.
8 MEMBER PALMTAG: I have some questions on 9
NQA-1 qualifications as the goal. So it's not NQA-1 10 right now?
11 MR. HANUS: It is not, because we don't --
12 we do not have the fully validated code. It is 13 developed using NQA. We suggest we need a 14 recommendation. That is needed. We follow the rules.
15 But it's not quite there yet because we did not finish 16 the whole thing yet. So we do not have the full set 17 of recommendations in this data that is finished.
18 MEMBER PALMTAG: Because NQA-1 isn't 19 something you do at the end.
20 MR. HANUS: No.
21 MEMBER PALMTAG: It's not something you 22 develop code and then you -- I mean, it's a process on 23 which you develop the code.
24 MR. HANUS: 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
32 MEMBER PALMTAG: You actually don't have 1
to have validation and verification done for NQA-1 2
because NQA-1 is a process on how to get there. So 3
you are following the NQA procedures?
4 MR. HANUS: Okay. Yes.
5 MEMBER PALMTAG: Okay.
6 MR. HANUS: Yes.
7 MEMBER PALMTAG: So it's not something 8
you're just at the very end?
9 MR. HANUS: No, no. It's not that.
10 Regarding the procedures, there's the NQA-1 -- we 11 recommend based on that. But, as I said, it's a work-12 in-progress.
13 MEMBER PALMTAG: You might want to check 14 because it should be NQA-1 as you develop it, not --
15 MR. HANUS: I would say it's --
16 MEMBER PALMTAG: It's not something you do 17 at the end.
18 MEMBER MARTIN: At the places I've been, 19 we have not called it an NQA-1 code until we completed 20 it all. But, yes, we apply the quality program.
21 MEMBER PALMTAG: But you should have the 22 quality program under the procedures.
23 MEMBER MARTIN: Well, so you might call 24 NQA-1 as an open item, so that you can do your safety 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 analysis and call it "safety" with an open item.
1 I assume you have a whole open item. It's 2
all in the QA space.
3 CHAIR KIRCHNER: Yes, the other thing is 4
it's qualification under NQA-1 approvals.
5 MEMBER PALMTAG: So that would be NQA-1, 6
not --
7 MEMBER MARTIN: I would say, some people 8
say it's terminology. As long as, like I say, if you 9
have an open item and you're tracking, I guess you 10 could call it that, but I'm kind of with Milan, that 11 I would just say it's not quite there yet. It's not 12 done.
13 MEMBER PALMTAG: I would say NQA-1 is a 14 process in which you develop --
15 MEMBER MARTIN: Yes, if you say that in 16 front of QA people, they figure you're done. And it's 17 very dangerous.
18 MEMBER PALMTAG: You might want to just 19 check it and check your terminology on that, whether 20 you're NQA-1 -- so you don't have to be done to do 21 NQA-1 qualification, but you should be doing your 22 development.
23 The other piece of the validation or 24 verification and validation, it seems like you're kind 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 of late in the process for doing this, because you're 1
doing design calculations. Right? You're doing 2
design calculations. You're calculating your release 3
rates. But yet, your codes haven't been verified and 4
validated. So the timing seems strange to me.
5 I mean, I would think, when you're 6
starting to do these source term calculations, these 7
design term calculations, design calculations, you 8
will want your verification and validation to be done, 9
so you know that your results are going to be good.
10 The danger, of course, is you design the 11 system, and then you do your verification and 12 validation and you find out, oh, no, your source term 13 is going to be much higher. And then there's sort of 14 a -- there's a lot of political thing to kind of make 15 your answers match. It just seems like the V&V should 16 be done before you do the design calculations.
17 MR. HANUS: Yes, so the V&V has been 18 ongoing since before at X-energy, but the code 19 capabilities that we developed have been added as part 20 of the development process. So, for example, the PIRT 21 table identified some of the things that the code, the 22 methodology was not able to capture originally. So 23 that was added. And that added a few additional 24 requirements for validation.
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 And so, the process is ongoing. The code 1
is not fully validated. It's verified to a big 2
extent. Then we added the features that have not been 3
verified yet.
4 MEMBER PALMTAG: But it sounds like you 5
have enough V&V done to have confidence in your 6
results?
7 MR. HANUS: Yes.
8 MEMBER PALMTAG: You just haven't quite, 9
you know --
10 MR. HANUS: Yes, and, you know, my 11 purposes at presentations at the NRC in 2019 about 12 V&V. I have identified now that that presentation is 13 in the design already. But I'm sure even before I 14 joined X-energy.
15 So the code has been verified/validated 16 and the design has changed, and the plan, to continue 17 the requirements which, again --
18 MEMBER PALMTAG: Oh, okay, I understand.
19 I don't want to cut you off, but --
20 MR. HANUS: Yes. So that's how I tell you 21 that.
22 MEMBER PALMTAG: Okay. Thank you.
23 MEMBER SUNSERI: This is Matt. Just two 24 easy questions for the Regulatory Affairs Director.
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 One is, do you have a quality assurance 1
program that follows NQA-1?
2 MR. HANUS: Yes.
3 MEMBER SUNSERI: Okay. And are you using 4
that program in the validation of this code?
5 MR. HANUS: Yes.
6 MEMBER SUNSERI: Okay. I wanted those two 7
answers clearly stated --
8 MR. HANUS: Yes.
9 MEMBER SUNSERI: -- so at least I know 10 where it is. Thank you.
11 MEMBER PALMTAG: You sound like you do 12 know.
13 MEMBER MARTIN: I don't want to belabor 14 it, but it sounds like you do have NQA-1 answers.
15 Okay.
16 MEMBER SUNSERI: I'm sympathetic, Bob.
17 Having been in your shoes, sometimes it's the 18 messaging in different audiences, you know. Like I 19 said, you're saying it the way I used to say it. But 20 other people --
21 MEMBER MARTIN: Matt and I are both I 22 think are on the same page.
23 MEMBER SUNSERI: Exactly.
24 MEMBER MARTIN: You either have a program 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 that's approved or you don't.
1 MR. HANUS: Right, right.
2 MEMBER SUNSERI: Usually, the QA topicals 3
are the first thing that's like approved. Right? And 4
we don't see that typically; that's all done.
5 MR.
HANUS:
So with NQA-1
- code, 6
(indiscernible due to accent) NQA-1. But the main 7
topic here is not the code itself, actually. So 8
that's the methodology and the models, but we 9
obviously do describe the source term. And at least 10 for the purpose of models, as I mentioned previously, 11 form a system, an integrated system, that there are 12 second iterations in between each other, each of those 13 models.
It's (indiscernible due to accent) 14 schematically, but I will go into (indiscernible due 15 to accent) and kind zoom in on this diagram.
16 Here, I just want to mention that we do 17 have a small amount of input from (indiscernible due 18 to accent) models or codes, (indiscernible due to 19 accent) for the neutronics and Flownex for the 20 (indiscernible due to accent) and the temperatures.
21 So these are the external codes and the discussion on 22 those, again, is using dose conversion factors that 23 have been developed separately. They have been 24 represented in a separate Topical Report. 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
38 basically, it's the input from the source term, the 1
source term release, and converted into a dose.
2 So, the thermodynamics model is the basis 3
of most of the calculations, because it provides the 4
temperatures in Flownex and provides them to the other 5
source system models. It is based on the geometry of 6
the pebble -- or, I should say, on the particle of the 7
pebble in the reactor.
8 And we also include a model for the 9
(indiscernible due to accent) compact for the 10 validation purposes, so that you can simulate them, 11 the AGR experiments themselves.
12 The basic thermal model, as expected, are 13 the heat transfer phenomena, the conduction, 14 convection, radiation, and the heat sources from, 15 let's say, decay heat and the fission and gamma 16 sources. And by using the traditional, let's say, 17 (indiscernible due to accent) for the pebble bed 18 convection and the (indiscernible due to accent) for 19 the conductive heat transfer. We raised the 20 temperatures in the old system going from the particle 21 all the way to the reactor nodes. In case we simulate 22 only the pebble, we can use the pebble temperatures to 23 be shown by assuming the heat conduction, the decay 24 heat conduction to the pebble using the temperatures 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 above (indiscernible due to accent).
1 Those calculations are (indiscernible due 2
to accent) from the heat conduction. It depends on 3
the temperature. We use the (indiscernible due to 4
accent) calculations, presented in terms of 5
gas,(indiscernible due to accent) into (indiscernible 6
due to accent) that are used within the model.
7 And as I mentioned previously, we input 8
for the THM input data from the VSOP and Flownex.
9 From the VSOP input, we perform a transformation, I 10 think, of the (indiscernible due to accent), using 11 (indiscernible due to accent). We choose to 12 (indiscernible due to accent) nodes to each other zone 13 (indiscernible due to accent) to simplify the 14 temperature calculations. So we first perform 15 (indiscernible due to accent) in the VSOP to the 16 source term, the source term model, and then perform 17 the calculations on the grid.
18 The method that is employed is 19 (indiscernible due to accent)
- simple, the 20 (indiscernible due to accent) method, and we use it 21 for the (indiscernible due to accent) calculations and 22 to match towards the steady state, just to be the 23 basis for the transient calculations.
24 The Particle Failure Probability 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
40 yes?
1 MEMBER PETTI: Just a question on the 2
pebble trajectories.
3 MR. HANUS: Mm-hmm.
4 MEMBER PETTI: How many trajectories do 5
you run to get the source term? You know, a pebble 6
can come in; it can come into any of however many 7
radial nodes you've developed. Do you do some 8
bounding trajectories or do you try to do a map 9
probabilistically? How do you do that? How does that 10 fit into the model?
11 MR. HANUS: Mm-hmm. Yes. So we look 12 probabilistically. I will not mention the actual 13 numbers. This is proprietary information, how many 14 runs we do. But the method distributes the pebble 15 towards the channel. The pebble goes through the 16 channel. Then it's put through a random channel 17 again, a different one or potentially the same even.
18 And once again, until it emulates the target burn-up 19 and generate (indiscernible due to accent) the pebble, 20 the time that it stays in the core, and then it's 21 discharged.
22 We do this many times, many times over, to 23 get statistically significant information and collect 24 the pebble information, and inventory, the burn-up 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 effluents and all this, as the pebble moves through 1
the system.
2 And that forms a library that is used in 3
(indiscernible due to accent) at the beginning. And 4
then we use that as the basis for other subsequent 5
calculations. So in the subsequent calculations, in 6
each transient operation, we can look at that pebble 7
at a given location in the core at a given path.
8 Because the pebbles are going on so many paths through 9
the core. And by this approach, we can retain the 10 full history of the pebble, as opposed to using the 11 RESO (phonetic) data. The RESO data has that 12 information as well, but it averages the pebbles at 13 each path. We would not get the information, the 14 different information about the history of a given 15 pebble at the given path that we need for the 16 inventory.
17 MEMBER PETTI: I just wonder about how you 18 validate that. Well, it's also very complicated. And 19 I know you've got a lot of margin here. So it might 20 be worth just looking at some bounding things, like 21 worst-case trajectory, average trajectory, least harsh 22 trajectory sort of thing, and see how big a difference 23 it really makes. Because that's a lot of information 24 that 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
42 MR. HANUS: It is.
1 MEMBER PETTI: And again, helpful probably 2
in the design sense, but I do worry about it in 3
validation sets. That could be a heavy lift.
4 MR. HANUS: I agree. Thank you for this 5
comment, because, you know, currently, the code is 6
more capable than it probably needs to be, good code 7
capable, of course, pending the validation. But the 8
capabilities are there. So we can actually quantify 9
all the defense, how much we actually are needing, the 10 detail that we need. And these are our data points 11 that --
12 MEMBER PETTI: Yeah, I mean, if we were 13 talking UO2, and historically, this was a huge area of 14 questioning by people because UO2 had hard limits on 15 burnup, temperature, and the like. So knowing 16 trajectories is really important. UCO doesn't have 17 those limits. You know, it's a much bigger design 18 window. And so, again, more capability than 19 necessarily needed for the UCO than were it for the 20 UO2.
21 But, you know, there used to be questions 22 about pebbles sticking in the core, you know, the old 23 German designs. And what if it goes above the burnup 24 because of that? Those are not really relevant today 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 and to your design, from what I can tell.
1 MR. HANUS: Okay. Thanks.
2 CHAIR KIRCHNER: How many cycles does a 3
pebble make before it's removed? And what's the 4
burnup target that you're looking to?
5 MR. HANUS: I don't want to disclose 6
something that might be proprietary. Anything on the 7
code knows that the burn-up is public information.
8 There's a number of passes. It's six on average. As 9
I mentioned, as the pebbles go through different 10 channels, it may be less. We can imagine the pebble 11 going through hottest
- channel, you
- know, 12 neutronically, all the time. But it can be less. But 13 that would be the bounding pebble. But, yeah --
14 CHAIR KIRCHNER: From a design standpoint, 15 when you use this, how many passes can you make, 16 assuming that the pebble goes down the hot -- the 17 equivalent of what would be a hot channel? It gets 18 the most burnup. It's also running at the highest 19 temperature.
20 MEMBER PETTI: Well, I mean, you can do 21 the math. It's not that difficult. Six passes always 22 in the hot channel. You can calculate that pretty 23 easily. It's pretty small, a small number.
24 So I'll just say, my guess is that 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
44 burnup is, quote, high.
1 CHAIR KIRCHNER: High?
2 MEMBER PETTI: Typical of pebble beds, they 3
get excellent fuel utilization because of it.
4 MR. HANUS: Okay. So the Particle Failure 5
Probability Model, this model is designed to provide 6
the information about the detection of particles that 7
could undergo failure.
8 Another model, (indiscernible due to 9
accent) means that the silicon-carbide layer fails, 10 the most important layer of the containment, so the 11 (indiscernible due to accent) abilities. And we do 12 models that show effects of phenomena due to which the 13 particle can fail, which are listed here. And 14 (indiscernible due to accent) phenomena (indiscernible 15 due to accent). We still model them, we do actually 16 see a low particle failures due to some of these, like 17 the kernel migration, for example, the Amoeba.
18 We still include these models because 19 there -- you know, partially, because we also want to 20 validate other models using old UO2 designs which are 21 certainly well-documented. And for these, we want to 22 match the UO2 performance, even though it's not part 23 of the UCO. But, for example, the (indiscernible due 24 to accent) through the matrix, through the speed 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
45 the pebble, is what you can, if you can validate why 1
it is you get compact particles debris.
2 And so, I don't know if I'm going into 3
details much here, because we model the manufacturing 4
effects, which are, you know, non-operating conditions 5
that are defining particle failure effects, the other 6
limited ones.
7 MEMBER MARTIN: Yes, go ahead, Scott.
8 MEMBER PALMTAG: I just had a question 9
about these failures. I know some of this may be 10 proprietary. So I'll keep it at a high level.
11 But you are going to have some pressure 12 vessel failure rate and some manufacturing defects in 13 the system. How well do you know these? And is your 14 fuel pellet very different from operating experience?
15 Is it close to it? Is this going to be something that 16 you're going to calculate? How well do you think you 17 know these failure rates? How are you going to 18 determine the failure -- how do you know if your 19 failure rates are -- have confidence in your failure 20 rates?
21 MR. HANUS: Basically, the question of 22 validating those models.
23 MEMBER PALMTAG: Okay. So you have a fuel 24 pellet. Is your fuel pellet very different from 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 what's been used before? Or is it similar or?
1 There's measured values out there. How applicable are 2
they to your fuel?
3 MR. HANUS: It is similar. So I would 4
have to ask the fuels lead which would be James.
5 MEMBER PALMTAG: Okay.
6 MR. HANUS: This is his end. He would 7
probably be able to answer the question as to how 8
different we are from the other --
9 MEMBER PALMTAG: Okay. Yes, my question 10 is, how different are -- how much are you going to 11 rely on the measured data that's out there and how 12 much you're going to rely on your calculational 13 capabilities?
14 MR. TOMPKINS: Hi. James Tompkins, 15 X-energy ARDP VP, Nuclear Fuel Lead.
16 Yes, we are performing a qualification 17 test at INL. And as part of that, we've fabricated 18 test fuel. So we will have some data to validate 19 fabrication failures, and then, in-tile performance 20 indications from -- it's a test that's fairly similar 21 to AGR, to lead out with short-lived fission product 22 gas monitoring.
23 So we will have some data that we can use 24 to point to demonstrate how similar we are in 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 performance to AGR and other data that we plan to use 1
for validation.
2 MEMBER PALMTAG: I guess my question, I'm 3
looking for some confidence that you know these values 4
well. I mean, you're doing the testing in Idaho.
5 That's good. But these failure rates are going to be 6
relatively small. So, if you only test a few 7
particles, how do you estimate a larger, you know, 8
PPM-type?
9 MR. TOMPKINS: We're testing over several 10 hundred thousand particles. So I guess it depends on 11 what you consider statistically relevant. Right?
12 MEMBER PALMTAG: No, that's good. That's 13 what I wanted to hear. So you are testing a large 14 amount of particles. That's going to give you some 15 confidence in these failure rates. Okay.
16 All right. Thank you.
17 MEMBER PETTI: And, Scott, they should 18 note the manufacturing stuff is going to vary batch by 19 batch.
20 MEMBER PALMTAG: Manufactured, but I was 21 thinking more of the question of failures.
22 MEMBER MARTIN: Yes, they won't get any 23 pressure vessel failure. It's designed not to fail.
24 MEMBER PALMTAG: There haven't been 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
48 past, right?
1 MEMBER PETTI: No, it is the most often, 2
but incredibly rarely found failure records. That's 3
because you can analyze it, that we know it.
4 MEMBER PALMTAG: So there's not going to 5
be any --
6 MEMBER PETTI: The silicon-carbide should 7
be in compression. So it shouldn't fail. That's the, 8
like, No. 1 design rule for TRISO fuel, is that you 9
make sure that design -- that you don't push it so 10 that you go into tension. If you go into tension, you 11 will get failure.
12 MEMBER PALMTAG: Okay. So you shouldn't 13 expect any pressure vessel failures?
14 MEMBER PETTI: It would be like 10 to the 15 minus 12. It would be a number so low that you can't 16 validate.
17 CHAIR KIRCHNER: No, that's the real 18 problem.
19 MEMBER PETTI: The real problem is that a 20 lot of these failure mechanisms, that they are so 21 low --
22 CHAIR KIRCHNER: So low.
23 MEMBER PETTI: -- that it is very 24 difficult --
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 MEMBER PALMTAG: Right. That was my -- I 1
guess that was my question.
2 MEMBER PETTI: So that's why it's so low.
3 You'll see. But the defects, they're going to move 4
around batch by batch, block by block.
5 MEMBER PALMTAG: I understand that, yes.
6 MEMBER PETTI: But there will be a spec, 7
and then there will probably be in the safety analysis 8
a margin of that spec.
9 MEMBER PALMTAG: Yes. No, I understand 10 that.
11 MEMBER PETTI: Yes.
12 CHAIR KIRCHNER: So are you, in your 13 testing at Idaho, are you going to intentionally drive 14 to failure, so you can try and see some of these 15 effects? Otherwise, I don't think you're going to be 16 able to do any validation. Right?
17 MR. TOMPKINS: Yes. James --
18 CHAIR KIRCHNER: Individual mechanisms, 19 Jim?
20 MR. TOMPKINS: Yes. So we have -- it's a 21 full, I mean, irradiation test campaign. We're 22 essentially performing the irradiation to demonstrate 23 in-path performance, and then, as soon as -- this is 24 kind of the condition of the fuel, and then, once we 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 get it out, we will do post-irradiation heating, 1
oxidation testing, and, you know, attempt to draw 2
bounds around the design and safety basis numbers.
3 So, yeah, the idea would be less to drive 4
to failure and more to demonstrate performance under 5
the operating envelope of, you know, selected LBEs, 6
then, specifically to see how difficult it is to fail 7
the fuel.
8 MEMBER PETTI: So, James, I tend to think 9
of it as a proof test. Because what's been made in 10 the past is a compact. This is a pebble. It's 11 different. How you make it is different. The forces 12 that are imparted potentially to the particles are 13 different. So, you've got to irradiate it to convince 14 yourself that you haven't done something, introduced 15 a new mechanism or something in that, you know, in 16 this new fuel form. No, it's just like German fuel.
17 So, you know, it's not anticipated, but you have to go 18 through that step.
19 CHAIR KIRCHNER: And I know testing is 20 going to be of pebbles, not just loose particles.
21 MR. TOMPKINS: Yes, that's correct, 16 22 pebbles.
23 CHAIR KIRCHNER: Okay.
24 MEMBER PETTI: Sixteen? Wow.
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 MR. TOMPKINS: Well, we had to put them on 1
a diet to put it an ATR, so they won't be full-scale, 2
but close --
3 MEMBER PETTI: Aw, okay.
4 MR. TOMPKINS: -- close enough.
5 MEMBER PETTI: Yeah, the Germans had to do 6
the same thing. These are big fuel elements for test 7
reactors.
8 MR. TOMPKINS: Yeah. The Chinese HTR-PM 9
or HTR-10 validated fuel as well. I think they had to 10 reduce the diameter.
11 MEMBER PETTI: Pretty common.
12 MEMBER MARTIN: Some of my question is, 13 this is really taken off of something Dave is going to 14 have in our letter, or is put in our draft letter, 15 related to the use of UO2, you know, for UCO fuel. Or 16 maybe I put words in your mouth. I think that was 17 unfair.
18 I guess my question is, in XSTERM, in 19 back-to-back, in models that are able to say that the 20 failure mechanisms, are they generic for the fuel form 21 in this case or do you say, all right, well, I'm 22 looking at UO2 data, so I'm going to run this code 23 with a UO2 model? I'm asking your --
24 MR. HANUS: Yes, and is most generic. I 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 would have to really look into each model separately.
1 I believe we would have some models that use the UO2 2
data because they still use UO. But --
3 MEMBER MARTIN: But okay. So the next 4
obvious question was, what's the value of the UO2 if 5
you used UO2-specific -- unless, of course, you are in 6
some way planning to have UO2 pebbles. Or, I mean, 7
no, you're not, of course. The failure rates are 8
higher.
9 If the code itself is distinguishing 10 through its constituent package a UO2 fuel form, but 11 you're not planning to have UO2 fuel, what would be 12 the value of including that in any validation package?
13 MR. HANUS: We use the UO2 calculations, 14 specifically the UCO2, for the similar thing that the 15 German -- the old German experiments with that. And 16 it was done before we started this through 17 qualification, as it was the only sphere of fuel that 18 had the validation data.
19 But the AGR (phonetic) compacts, they are 20 the UCO kernels, our kernels, but not our geometry.
21 And the AGR irradiated pebbles, they are using the 22 kernel, but they are not pebbles. And so, you know, 23 when validating, when using the data validation, they 24 focused on the heavy-weight transporter to 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
53 (indiscernible due to accent) to the pebble. And so, 1
we had to validate, to some extent, the geometry 2
effects. But we need to compare it with the results 3
that were generated by --
4 (Simultaneous speaking.)
5 MEMBER MARTIN: And I'll put some words in 6
your mouth again. So, interval test data is limited.
7 Right?
8 MR. HANUS: Yes.
9 MEMBER MARTIN: Some of it is really old, 10 where they used UO2. Right? And I do know that 11 XSTERM is at least 10 years old, if not longer. I 12 don't know if you've ever said that in our meetings.
13 MR. HANUS: Yeah, probably in 2016.
14 MEMBER MARTIN:
Okay.
So, since 15 developed, it would have been actually ongoing.
16 MR. TOMPKINS: Yeah.
17 MEMBER MARTIN: I just remember the first 18 paper I think I looked at was, like, a 2016 paper.
19 MR. HANUS: Yes, yes.
20 MEMBER MARTIN: It was probably line an 21 ANS conference or something.
22 So part of your answer is this is a bit 23 legacy because that's what you had back then.
24 MR. TOMPKINS: Right. 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
54 MEMBER MARTIN: So you started with that.
1 As the data from AGR was coming out, okay, and you're 2
pivoting. You brought in the UCO models. You may 3
also have some needs related to just the limited data 4
out there. So it allows you to do some, say, interval 5
validation. I mean, it might not necessarily be 6
directly applicable to Xe-100 down the line, but it 7
covers the package as a whole. So there certainly is 8
value in there.
9 But I'm trying to understand, it's really 10 about the data is limited out there. The timeframe in 11 which you started all this work and what was available 12 at the time. So you've carried it on, even though, 13 when it's all said and done, the code, as far as the 14 NQA question, is applicable only to UCO? You're not 15 going to make a claim that --
16 MR. HANUS: Yes. Yes.
17 MEMBER MARTIN: -- in some ways, it's 18 safety-related for UO2. It will never happen. So 19 what's the point?
20 MR. HANUS: Right, right. Yes. And 21 that's also, you know, there are more validations we 22 don't include. We include those only that are 23 (indiscernible due to accent) data and which are 24 needed. We could select (indiscernible due to accent) 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 different correlations for something related just to 1
UO2, but we just use one because it's not the main 2
focus of the code.
3 (Simultaneous speaking.)
4 MEMBER MARTIN: So the one failure mode 5
you'd be worried about would be if a user accidentally 6
put, say, UO2 in the input file. You would hope they 7
would always have the --
8 MR. HANUS: Yes, yes. Yes.
9 MEMBER MARTIN: So you'll get a different 10 package.
11 MR. HANUS: No, that's right. Yes, that's 12 very correct. We would in that case.
13 MEMBER PETTI: Let me just note that we 14 are falling behind schedule, Mr. Chairman.
15 (Laughter.)
16 MEMBER PETTI: So let's try to pick up the 17 pace.
18 MEMBER MARTIN: Okay. Just general 19 advice. You're using the same presentation as you had 20 at the SC. That, of course, is not necessary. Just 21 for future reference, you can abbreviate that for full 22 Committee. Because this is the danger of doing the 23 same slide set. We rehash the --
24 MEMBER ROBERTS: Yes, this is Tom. I have 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 one question. It might be for Dave. But I wanted to 1
ask it while the applicant was still here.
2 If I
understood your assessment, 3
basically, all the phenomena listed here are designed 4
out to UCO fuel, but there's a phenomena not listed 5
here that is not designed to use UCO fuel and it's 6
very difficult to model. And currently, the INL is 7
working on how to model it.
8 So, as I understood what your testimony, 9
it is that, basically, they're monitoring more 10 phenomena, and what they're modeling is not a problem 11 for UCO fuel and what they're not modeling tended --
12 do I have that right? And I'd be interested to get 13 X-energy's perspective on that.
14 MEMBER PETTI: No, I think that's a fair 15 assessment. And again, it's easily accounted for, 16 like, pressure vessel failure. Yes, there will be 17 fission gas pressure, but it's always at the bottom of 18 the list in terms of what's important.
19 But, yes, all these -- the U.S. approach 20 to fuel, particle fuel design, was always to design it 21 away. You either change the particle design, you 22 change the temperature run-up effluents. You limited 23 it so that you didn't have -- it's like you cut the 24 grass lower and lower and lower, and that was their 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 theory. So there's nothing left.
1 And the failure that was observed is very 2
difficult to model. But again, it occurs at about 10 3
to the minus 5. And so, that's really difficult to 4
model. What it means is the DOE modeling focus has 5
taken it out. As a challenge plot, it's so difficult 6
to model.
7 MEMBER ROBERTS: I'm not aware of X-energy 8
agreeing with that characterization that, with the 9
testing of a large number of pebbles and a lot of 10 particles coming up, it may be your long-term view is 11 to refine this model to more closely resemble what it 12 is you're actually producing? Is that right? Or is 13 this the way you plan to go into the operating 14 license?
15 MR. HANUS: Yes, that's generally right.
16 Like we always look at the results that we will get 17 from the qualification. You know, it's very nice.
18 And, you know, we haven't made any decision yet on the 19 models that showed this effect, for example. We might 20 as well use just the facts that comes from the 21 experiments that is beyond what we can pull out with 22 the count for this elegant mechanistic model. We will 23 need to first see the results and we will try to 24 quantify the needs for this.
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 MEMBER ROBERTS: Okay. Thank you.
1 MR. HANUS: Let me continue quickly with 2
the SOLM, the Fission Product Transport Model. It's 3
based on basically the fission equation, the time-4 dependent fission equation. That's been between the 5
production and removal of the radionuclides from the 6
particle (indiscernible due to accent) and into the 7
pebble graphite.
8 And we will have a set of radionuclides 9
that have been determined previously to be 10 (indiscernible due to accent) TGR systems. And the 11 scope of that SOLM model is either single sphere or, 12 again, for all core meshes. This goes back to the 13 question of taking the pebble as use of the core. So, 14 in that type of calculation, that's when a model is 15 used to accumulate the other isotopes in the pebble as 16 it moves towards the core. And the output of the SOLM 17 model is (indiscernible due to accent) release over 18 birth, basically the release ratios (indiscernible due 19 to accent).
20 The model is based on the diffusion 21 coefficients, dependent on
- time, through this 22 (indiscernible due to accent) law. The diffusion 23 coefficients are mostly taken from the (indiscernible 24 due to accent) the IAEA (indiscernible due to accent) 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 performed a study of the more recent papers and 1
(indiscernible due to accent) determination of the 2
diffusion coefficients of (indiscernible due to 3
accent) energies from multiple different papers. They 4
do not infer as much bias. So, we use these diffusion 5
coefficients to close the model, basically.
6 MEMBER ROBERTS: So that document that I 7
recall is -- you know, there's not data for every 8
radionuclide. Right? I mean, they get the big ones.
9 So how do you kind of fill in the gaps for all the 10 other radionuclides that you're considering?
11 MR. HANUS: So we look at the different 12 properties (indiscernible due to accent) there are 13 similar properties. And then we use the same 14 diffusion coefficient. So we use, you know, examples 15 of the iodine and (indiscernible due to accent).
16 MEMBER ROBERTS: I mean, everybody has the 17 same problem. Right? The data is limited in this 18 particular case. Do you have a strategy to be 19 conservative in this, you know, with this?
20 Again, I understand the challenge that 21 you've had. I've dealt with it myself. But my goal 22 wasn't to be necessarily conservative; just as 23 accurate as possible, which, of course, introduces 24 uncertainty, which I never got around to actually 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 quantifying.
1 MR. HANUS: Yeah, we base the selection on 2
the diffusion corrections on the properties of the 3
atoms (indiscernible due to accent) selecting at 4
higher -- you know, a higher diffusion coefficient, 5
let's say, or lower, that needs to occur for 6
conservatism.
7 It's hard to quantify what is conservative 8
here. (Indiscernible due to accent) gets more dust 9
abated and the dust includes all (indiscernible due to 10 accent) in the pebble. In this case, you would want 11 the juncture of chemical and physical properties.
12 MEMBER ROBERTS: Right, right.
13 (Simultaneous speaking.)
14 MEMBER PETTI: It's just a point in a 15 subsequent report. Almost every applicant we see has 16 a table that says, here's the fission product where we 17 have data and it represents the following elements.
18 You know, you're going to assume iodine behaves like 19 noble gas. What are you going to assume about 2RM 20 (phonetic)? You know, that mapping is pretty 21 standard, so that everybody understands sort of the 22 rules of the road at the very beginning, how you're 23 going to fill that in.
24 It's not in the document; make a note 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
61 it. We'd love to have that table. You guys have it 1
in your head and in your coat, I'm sure.
2 MR. HANUS: Yeah, mm-hmm.
3 MEMBER PETTI: It's just making it 4
explicit.
5 MR. HANUS: Okay. Yeah, thanks.
6 CHAIR KIRCHNER: The table that you have 7
on slide 20, which are the dominant actors and where 8
do you have the most uncertainty? And how do you 9
bound that?
10 MR. HANUS: At the moment, the cesium-137, 11 the silver-110.
In terms of uncertainty, 12 (indiscernible due to accent) which has the most 13 uncertainty in the data.
14 CHAIR KIRCHNER: So, with your current 15 model, which are the dominant radionuclides in terms 16 of the contribution to eventual source term?
17 MR. HANUS: Yeah. That is the cesium.
18 CHAIR KIRCHNER: The cesium.
19 MR. HANUS: The 137.
20 (Simultaneous speaking.)
21 MR. HANUS: Yes, so, (indiscernible due to 22 accent) a solution (indiscernible due to accent). You 23 can choose. But the solution of the equation start 24 being solved eventually (indiscernible due to accent) 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 methodology. Right?
1 And next we'll do -- or next model is the 2
Gaseous Fission Product Transport Model, which is 3
similar calculations -- this is similar output, again, 4
(indiscernible due to accent) but for the gaseous 5
radionuclides. Compared to the SOLM model, this is a 6
(indiscernible due to accent) model. It's much 7
simpler. It doesn't solve any (indiscernible due to 8
accent) equations. So it's basically a solution 9
(indiscernible due to accent). It was based on the 10 (indiscernible due to accent) solution of the 11 (indiscernible due to accent). It takes into account 12 a
(indiscernible due to accent) basically, 13 (indiscernible due to accent) of
- areas, of 14 capabilities of other types of areas.
15 But we use this model mainly to inform the 16 SOLM model, to generic diffusion coefficients for the 17 gases that -- the effective coefficients for the noble 18 gases, which were not included in the defenses that we 19 have available. So we basically use this model to 20 calibrate the SOLM for the gaseous radionuclides.
21 Optionally, we can use the GASM model as 22 well in the main calculations (indiscernible due to 23 accent) for the noble gases with the fast or slow 24 half-life, except for the few which are long-lived.
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 The Dust Production Model is based on the 1
distribution of dust particle sizes, which comes from 2
the previous AVR data that is available. And we 3
modeled the pebble-pebble and pebble-reflector 4
abrasion, the pebble abrasion (indiscernible due to 5
accent), and the control rod abrasion as the controls 6
in the shaft with operating period.
7 And the in-core dust production is 8
proportional to a geometry-dependent dust production 9
rate parameter, which is calibrated against the AVR 10 data. So, it is too much dust is produced in AVR, 11 scaled to our design and get that dust production 12 parameter.
13 And the dust production in the fuel 14 handling system is based on a dust generation rate per 15 meter of the movement comes from the reactor model 16 requirements, design requirements document.
17 (Indiscernible due to accent) for the RCS 18 for how much dust about to be generated in the 19 reactivity control system. And, all together, we have 20 the dust production parameter that is important to the 21 helium pressure boundary module to obtain the dust 22 production distribution throughout the plant, through 23 the helium pressure boundary, as a function of time, 24 which is then used in the statistical operations.
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 MEMBER PETTI: Milan?
1 MR. HANUS: Mm-hmm?
2 MEMBER PETTI: Do you guys have access --
3 have you seen anything from the Chinese HTR-PM?
4 Because everyone has always wondered. They've got 5
them operating. Because, you know, there's a lot of 6
controversy about how much dust you really get. And 7
it's actual operating reactor. Have you guys seen 8
anything from them that you'll be able to --
9 MR. HANUS: Yes. So we (indiscernible due 10 to accent) some documents from the HTR-PM. Yeah, I 11 haven't gone through the (indiscernible due to accent) 12 looked into these. There was no information -- there 13 was very little information from HTR-10.
14 (Simultaneous speaking.)
15 MR. HANUS: Yeah, there's no dust there, 16 almost. So, yeah, it's (indiscernible due to accent) 17 what we can get from the Chinese. I need to get some 18 data (indiscernible due to accent).
19 MEMBER PETTI: Yes, if you could, I mean, 20 certainly bringing operating data to the party here 21 would be really valuable.
22 MEMBER MARTIN: Of course, you calculate 23 the dust source term. So, during a depressurized loss 24 of forced circulation, do you just assume 100 percent 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 of that gets out? Or do you use some fraction of 1
that?
2 MR. HANUS: Essentially. We calculate 3
that depressurization. So we calculate the release of 4
the dust, along with (indiscernible due to accent).
5 Yeah, so we do calculate (indiscernible due to accent) 6 everything out.
7 MEMBER MARTIN: Okay. So that would be 8
very conservative.
9 MR. HANUS: Then, it only depends on the 10 reactor without a vent. We are looking at -- in some 11 cases, we do increase everything out and it's almost 12 conservative. But we do have calculations that 13 actually track the dust to the HPV.
14 MEMBER MARTIN: Okay. And so you have to 15 make a separate transport model for the dust outside 16 of the reactor or?
17 MR. HANUS: No. So that adds to the --
18 yeah --
19 (Simultaneous speaking.)
20 MEMBER MARTIN: And it's really the source 21 term, but the overall plume itself --
22 MR. HANUS: Yeah.
23 MEMBER MARTIN: I think, as I recall, like 24 with AVR data, I mean some assessment of dust, as I 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 recall, it just said that the dust got all over the 1
nooks and crannies. It was highly unlikely to really 2
get a significant liftoff at a relatively -- some 3
missed this, but, again, you never know until you 4
know. But if you use 100 percent, it probably could 5
be as much as five times or more, you know, a 6
conservative in that particular case.
7 But that's what I wanted to hear. You're 8
using a conservative assumption on dust. Okay.
9 MR. HANUS: Okay. The last barrier that 10 we take into account in the reactor building is the 11 Helium Pressure Boundary. It is also probably the 12 most complicated model because it takes into account 13 phenomena that can happen in the pressure boundary.
14 And it's listed here. We model the migration of the 15 radionuclides in the multiphase flow, taking into 16 account actions, the mass exchanges between the dust 17 phase and the radionuclide. We model the gas 18 propagation through the system.
19 It's simplified into, essentially, a 1D 20 pipe kind of geometry, although we take into account 21 the multiple inputs. So, one component can get flow 22 from multiple different components, not just from the 23 upstream. There can be multiple (indiscernible due to 24 accent) data for a given component.
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 The phenomena I'll skip for the sake of 1
time. And, here, this slide shows a schematic of 2
those phenomena in a given node or movement of the 3
pressure boundaries. And as you can see, it takes 4
into account the radionuclides, deposition, dust 5
liftoff, absorption, and graphite de-sorption back 6
into the helium. And, again, we calculate the 7
standard method of putting the (indiscernible due to 8
accent).
9 And the Corrosion Model at the end can be 10 used for other (indiscernible due to accent) 11 calculations.
Basically, the increase of 12 (indiscernible due to accent) due to oxidation. And 13 this one is based on, also, the General Atomics fuel 14 manual, actually, for the calculation (indiscernible 15 due to accent) models for MHGTR and based on 16 (indiscernible due to accent).
17 All right. So that's the end. Again, 18 that is the final slide. It shows the connections of 19 all those models together. And we are going to 20 simulate the source term transport as a whole, 21 mechanistically as much as possible. Of course, there 22 are many assumptions involved which we have 23 (indiscernible due to accent) to be on the 24 conservative side. Optionally, we have some design 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 models as well that you can use for the scoping 1
calculation and for the safety. We are biased towards 2
the conservative side.
3 MEMBER HARRINGTON:
This is Craig 4
Harrington. Just as a practical matter, this 5
flowchart, basically, on the screen, when you execute 6
this code, is it one -- I mean, do you execute it as 7
in its entirety all at once? Or are each of these 8
major modeling elements executed individually and 9
there's manual handoffs between the pieces? Do you do 10 both? Can you speak to that?
11 MR. HANUS: It depends on the event and 12 the part of the operation. And the code execution, 13 again, is to get more detail in the XSTERM Topical 14 Report. It's modeled in the methodology. But in the 15 (indiscernible due to accent) methodology we describe 16 how the source term methodology is used to do things 17 for certain different events.
18 For example, for the steady state and the 19 helium pressure boundary, we use the core integrated 20 model to calculate everything from release to 21 transport to the fission boundary.
22 As part of the Flownex criteria, from the 23 previous presentation, that technology involves 24 calculation of individual pebbles under those specific 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 conditions, temperature and burnup and extreme 1
conditions, in which case we will go to HPBM. The 2
DUSTM and all these models are inactive. They are not 3
used because we are looking at one single pebble, 4
given conditions. So we only execute the THM, FPM, 5
SOLM.
6 It's the same code practically, you know.
7 The (indiscernible due to accent) code is the same 8
code. It just uses different paths, the old code.
9 And here, the (indiscernible due to accent) are using 10 all the different (indiscernible due to accent) 11 separate models with all them.
12 MEMBER MARTIN: So it's an integrated 13 execution, but may or may not use all pieces at any 14 given --
15 MR. HANUS: For practical reasons. But 16 that's all defined in the safety analysis methodology, 17 how we use the XSTERM code or these models in 18 conjunction with the safety analysis. So, each safety 19 analysis for a given event is using this methodology.
20 And it differs. It differs by event and it's driven 21 by the practicality.
22 In essence, we could use the full 23 integrated code all the time, but it would not be 24 necessarily the most time-efficient, considering 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
70 (indiscernible due to accent). And it would be too 1
much, too time-consuming, for example.
2 MEMBER MARTIN: And so, I assume, from a 3
V&V standpoint, you have a combination of V&V for 4
individual models, as well as an integrated V&V 5
package? Is that the plan?
6 MR. HANUS: Yes, that's right. That's why 7
we are focusing on the individual models, because they 8
all feed into a package as well. So, yes, the 9
validation proceeds step by step, essentially, in 10 those phases. So that we don't have to repeat in 11 order to do the validation. Because it's for the full 12 period and pace of operation.
13 MEMBER MARTIN: Okay.
14 MEMBER HALNON: This is Greg. I've got 15 one question that I asked during the Subcommittee 16 meeting, but I asked the NRC staff. So I wanted to 17 ask you all, anticipating that they're going to come 18 back and say: lots of good, but we can make no 19 conclusions based on all this work.
20 What are you going to use this Topical 21 Report for?
22 MR. HANUS: We use this as a specification 23 for the code. Obviously, the actual -- what we have 24 on this is a code. The code is to implement 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
71 The verification of the code basically looks at those 1
models and checks that the code implements these 2
models.
3 This Topical Report provides the 4
justification file that goes with us, right, for our 5
design and safety case. And if we implement those 6
models correctly, then we get confident in the safety 7
case.
8 MEMBER HALNON: But as you go forward, the 9
NRC has not approved. I mean, they can make no 10 conclusions. So, are you proceeding at risk, hoping 11 that you will get more technical review down the road, 12 that this XSTERM is going to be acceptable to use?
13 Yes, conceptually, yeah, it sounds like a great idea, 14 but we can make no conclusions. So I don't know how 15 you're going to use this going forward as a technical 16 justification to use in your continued evaluation of 17 safety.
18 CHAIR KIRCHNER: Or is it part of the V&V 19 package in general?
20 MR. HANUS: Well, we will provide the 21 XSTERM Topical Report which asks NRC to make 22 conclusions about this methodology. And so now, if 23 now we get -- you know, you can provide a conclusion, 24 but it is information for us. And we'll take notice 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 of the suggestions.
1 But this should be used in something like 2
ODAD -- But we see the need of the future XSTERM 3
application in this regard in this particular domain, 4
so that the application will be, hopefully, much 5
easier to approve, since it's already been seen 6
before. The methodology has been seen before.
7 I don't know how identical it is --
8 MR. THOMAS: Yes, I'll just -- again, 9
Matt Thomas, Licensing Manager at X-energy. So I'll 10 just add onto that.
11 So, you know, based on the recommendations 12 of the Subcommittee or the suggestions in the 13 Subcommittee, the ultimate goal is to present the 14 XSTERM Topical Report to the NRC and get approval on 15 it, but we're also looking at, you know, a contingency 16 in case -- right? And kind of in parallel, looking at 17 a more validating type of approach that can be used 18 maybe as verification or just another method to 19 maintain our current schedules and stuff for our 20 project.
21 MEMBER HALNON: I guess, normally, when we 22 get these Topical Reports, the methodologies, we get 23 NRC's buy-in on it and we can reference and just use 24 it going forward. This one, it doesn't seem like --
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 it's more of a white paper when the NRC reads it and 1
agrees, "Yeah, okay, it looks great to us, but you've 2
got a lot of work to do." It kind of feels more like 3
a white paper than a Topical Report.
4 And I just want to make sure that, if we 5
see this referenced in a construction application down 6
the road, we understand what we're seeing. And what 7
we're seeing for this Topical Report especially is 8
it's not able to be referenced as an approved 9
methodology.
10 And so, unless you do all the work that 11 you need to go through -- that the NRC is planning on 12 you in an SE. So it just feels preliminary. It feels 13 like we've not wasted time, but we've spent a lot of 14 time kind of agreeing to agree that this looks good.
15 (Laughter.)
16 But I'm struggling with the endpoint of 17 where we're at with this.
18 MR. THOMAS: It's here to support the CPA, 19 right?
20 MEMBER HALNON: But it's a non-conclusion.
21 You can't support it with --
22 MR. THOMAS: Well, I mean, it's kind of 23 what the staff had touched on their -
24 MEMBER HALNON: But they finally came out 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 and said, to put words in your mouth, that it's 1
probably good enough. But we'll see it again.
2 (Laughter.)
3 CHAIR KIRCHNER: May I ask a specific 4
question? You've got in process your irradiating 5
pebbles up in Idaho. I don't know the timeframe for 6
that and subsequent trying to tie thermal testing to 7
trying to release rates and such, but this activity up 8
at Idaho will give you data that will validate the 9
THM, FPM, and SOLM modules, right? And maybe then the 10 gas transport?
11 MR. THOMAS: Yes.
12 CHAIR KIRCHNER: So what you're working on 13 up at Idaho will allow you to do V&V for the four 14 boxes in the middle of this diagram?
15 MR. HANUS: Yes, that's right.
16 CHAIR KIRCHNER: And then, presuming that 17 you get pretty high quality -- you have high quality 18 particles that you're irradiating, you're going to see 19 low release rates and such, as was the case with the 20 AVR experiment. So, what's your sense of where are 21 the big gaps? You'll complete that. You'll be able 22 to validate the middle models on this diagram. Then, 23 you'll go forward with the dust model as you have it 24 now.
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 I'm trying to just kind of get a sense for 1
how much V&V effort is needed to have this package in 2
a state where it's validated and you know what -- you 3
can then do an assessment where the biggest 4
uncertainty is and use it effectively as a licensing 5
tool.
6 Right now, you're using it to make 7
bounding estimates. As you go into the CP 8
application, it is my assumption -- but when do you 9
expect to have it validated and how does that fit into 10 your overall timeline, your schedule?
11 MR. HANUS: The validation completion out 12 of the full validation can only happen after we obtain 13 the data from INL, which is a couple of years from 14 now. James can say a bit about that schedule and 15 share this with the public.
16 But that will take time. We have this 17 phased validation approach here which we do, the fact 18 that it is being validated at the same time as we have 19 data available from the previous experience, and this 20 provides validation for the other models as well.
21 But the fuel qualification is now ongoing, 22 and right now it's where the best validation that you 23 can get for the fuel. But we will not get any new 24 data. We are looking for the verification, especially 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 for the dust model and for the HPBM, to get the data 1
from the Chinese. But I can't say when we are getting 2
that data. But this continued validation is waiting 3
for verification.
4 MEMBER BIER: Thank you.
5 MEMBER SUNSERI: This is Matt again. And 6
perhaps I'm a little bit more optimistic about where 7
we are in the process and the state of this thing. It 8
seems to me, with the fact that we're tying this to 9
functional containment, which you guys are going to be 10 one of the first out of the door on that approach, 11 having a mechanistic source term approach for the 12 development.
13 That is, I'll stop short of having it 14 approved by now with the regulator, but at least 15 having agreement on where it's going to be, based on 16 the level of effort it's going to take to push this 17 over the finish line. It seems prudent and important 18 at this stage of where you're at as far as, you know, 19 to build and operate.
20 I think the effort right now seems 21 worthwhile and useful, but that's my opinion.
22 MEMBER PALMTAG: This is Scott. I'm just 23 trying to think this through.
24 But my understanding is this is 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
77 procedure you're going to have to determine your 1
source terms going out of it, but you're not going to 2
have validation. So it sounds like we're not going to 3
see this XSTERM for years, right? At least a couple 4
of years? So you're saying, "This is the path we're 5
taking for the validation." You must have some number 6
for what you think your source term is going to be, 7
and then, that's going to be enough to go ahead with 8
construction permits.
9 So it's kind of like what Greg says; 10 there's some risk there.
11 MR. HANUS: Yes.
12 MEMBER PALMTAG: If you have some number 13 for your construction permit, and then all this 14 validation comes and it's higher, then you're at risk.
15 But it sounds like --
16 MR. HANUS: Yes.
17 MEMBER PALMTAG: Is that kind of the plan?
18 MR. HANUS: Yes. Again, you know, the 19 validation is not now waiting; I mean, it's ongoing.
20 The qualification of gamma is not the only validation 21 data that we have for these models, these first few.
22 So here we are with these validations.
23 They are looking good, but, of course, they are not 24 the other fuel. So we are now getting back to 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
78 you know, we have the AGR results. We have the AGR 1
results of the pebble. And all these, all these 2
results are -- we have confidence that we are not 3
doing something wrong, but to get the actual 4
validation, to create something that we could -- we 5
need to perform for fuel realistic.
6 MEMBER PALMTAG: Sure. So, in my personal 7
opinion, it sounds like this is the right track.
8 You've give us kind of what you plan to do, but just 9
to acknowledge what Greg said, there is risk here, 10 right, that the validation won't --
11 MR. HANUS: There is risk later, yes, that 12 the qualification uncovers something that we have 13 taken into account.
14 MEMBER PALMTAG: Yes, I think this LTR is 15 very important because it tells what's going to be 16 coming in the XSTERM in a couple of years.
17 MEMBER HALNON: Well, it is typically in 18 a white paper to get kind of conceptual agreement on 19 how they're going to develop a methodology.
20 MEMBER PALMTAG: Even for a CP?
21 MEMBER HALNON: Well, they're not in the 22 CP yet. They don't have a CP.
23 MEMBER PALMTAG: Yes.
24 MEMBER HALNON: We don't know what the CP 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
79 is going to look like and how much detail is going to 1
be.
2 MEMBER PETTI: The source term is going to 3
be low. I could on one piece of paper calculate what 4
their source term is, because I've done this for a 5
while, you know. You don't need all these codes.
6 What I worry about is not that -- I worry 7
that, for instance, the irradiation testing in Idaho 8
will not be able to unravel, because it's an integral 9
test, what you need to validate each of the models.
10 This happens all the time. Everyone who says, "We're 11 going to use AGR," I say, "Good luck." Go look at 12 these papers. We've tried this. Idaho tried this.
13 It's not that simple. These are very complicated, 14 because it's a very complicated fuel.
15 But the benefit is, you know, you get this 16 benefit. So that's why I keep saying it's nice to 17 have a strategy in your back pocket of something 18 simple, that you either compare it to that, and if, in 19 fact, you can't get there, you have a simple thing.
20 And you can do this. It's not difficult, because 21 there's lots of margin here.
22 You know, if you're right against the dose 23 limit, you're going to need the sharpest pencil you 24 have. That's not where they are. And so, they can be 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
80 more, quote, "cavalier," if you will, in terms of the 1
model. It could be a lot simpler and still show they 2
meet all things.
3 CHAIR KIRCHNER: I concur with Dave's 4
assessment because my sense is that you're going to 5
need to have a really low source term for almost all 6
your transients, except one, and that will be when you 7
have intrusion of H2O.
8 So, we have a steam generator tube 9
rupture; that is probably going to bound things. And 10 as Dave put it, I mean, all the other detail in these 11 other models, that kind of event will mask everything 12 that you'll see during normal transients and operation 13 of this particle fuel, assuming you get good fuel from 14 your vendor.
15 So that box in the lower right there, I 16 presume that's your mechanistic model for steam 17 interacting with graphite. It's probably your 18 dominant worst-case source term.
19 MEMBER PETTI: It was in the MHTGR.
20 CHAIR KIRCHNER: So I know this is a 21 design detail, but assuming that that module works the 22 first order, you can run your projected worst-case 23 steam generator tube rupture kind of events and see 24 what your bounding source term is and use that, Matt, 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
81 for the CP application, as appropriate.
1 My sense is that lower right-hand box will 2
dominate.
3 CHAIR KIRCHNER: So it's good enough at 4
this point?
5 MEMBER SUNSERI: Yes, I mean, look, I 6
can't speak for the Committee, but it almost seems 7
like we're being overly critical of the effort they're 8
putting in. They're refining the calculation. Maybe 9
they could do a bounding or whatever, but what do we 10 care what their motivation is? They have their 11 motivation. If they want to get increased operational 12 margin on whatever it is through this effort, then 13 they should be able to pursue that. Is there anything 14 from a safety significance that we should be concerned 15 about now?
16 So we're having a pretty extensive debate 17 over something that you're acknowledging has marginal 18 safety implications.
19 MEMBER PETTI: It's the matter of proving 20 it. This is what it always is. And how difficult 21 that can be when you get into the details.
22 I had a question for Matt. You raised the 23 steam question. Is there a dump system off the steam 24 generator in the event that there's a leak? Some gas 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
82 reactor designs have that to limit how much steam can 1
get into the systems, how much moisture can get into 2
the system.
3 I don't remember in any of the meetings 4
we've had whether that was discussed. Do you --
5 MR. THOMAS: Yes, Brian Froese on the 6
phone, do you happen to have a response to that?
7 MR. FROESE: Yes. This is Brian Froese, 8
X-energy, Project Manager, Analysis Integration.
9 Apologies I'm not there in person again this time, but 10 this has been a really interesting discussion.
11 The Xe-100 does have a dump system. It's 12 not safety-related.
13 And one point of clarification: our 14 limiting accident right now is our long-term DLOFC 15 right now instead of our tube rupture. We've done a 16 little bit more analysis on tube rupture. Our 17 oxidation of graphite is fairly low and the amount of 18 radionuclides, assuming that we get the fuel that 19 we're hoping to get, the amount of radionuclides 20 throughout the primary system is -- we're seeing it is 21 fairly benign.
22 So the driver for our limiting dose 23 accident is long-term DLOFC and those peak fuel 24 temperatures and associated diffusion out of the fuel 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
83 from that.
1 MEMBER MARTIN: We think we're done here 2
with X-energy.
3 MR. HANUS: Yes.
4 MEMBER MARTIN: Yes, I think we should 5
take a break.
6 We do have a presentation by the staff, 7
about nine slides, I think, plus their title and their 8
9 So, 10 minutes? Fifteen minutes? Okay.
10 CHAIR KIRCHNER: Let's take a break until 11 10:40 local time. Okay? So we are in recess.
12 (Whereupon, the above-entitled matter went 13 off the record at 10:26 a.m. and resumed at 10:42 14 a.m.)
15 CHAIR KIRCHNER: Okay, we're back in 16 session.
17 I'll turn it back to Bob Martin.
18 MEMBER MARTIN: Okay. We've wrapped up 19 X-energy's presentation and we're going to move over 20 to the staff's presentation. I believe it's being 21 done remotely; I'm not sure. Do we know who's going 22 to be doing that?
23 Speak into the microphone and just 24 introduce your colleagues.
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
84 MR. DRZEWIECKI: Yes, this is Tim 1
Drzewiecki from the staff.
2 Oh, sorry, was that for me?
3 (Laughter.)
4 MEMBER MARTIN:
You can introduce 5
yourself. It is perfectly all right.
6 MR. DRZEWIECKI: Yes. So, yes. I'm Tim 7
Drzewiecki. I'm the lead tech reviewer for the Xe-100 8
which also extends to the Long Mott Generating Station 9
CPA, which is currently in-house.
10 With me is my Branch Chief Travis Tate, as 11 well as the Project Manager for this review. That's 12 Denise McGovern.
13 Did I miss anybody?
14 (No response.)
15 MR. DRZEWIECKI: Okay. I do want to give 16 a bit of context to this, and that's to basically 17 state that this Topical Report is referenced in the 18 current Long Mott Generating Station CPA, which is in-19 house now. And so, there are comments in this 20 presentation about doing a detailed review of those 21 models and we do plan to look at those in more detail, 22 and are planning to do that and started to do that now 23 actually. So I just wanted to give that just as 24 context.
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
85 By the way, just as a sound check, I just 1
want to make sure that I'm being heard clearly.
2 CHAIR KIRCHNER: Yes, we hear you well.
3 Thank you.
4 MR. DRZEWIECKI: Thank you.
5 Okay. So, as far as some background, this 6
Topical Report, it was submitted in May of 2024 with 7
an update earlier this year. That update, it includes 8
updates of Section 1.5 and 7.1 to clarify that it 9
applies only to preliminary analysis. There were some 10 typos that were corrected, as well as an update to 11 Appendix H, which shows all the model interfaces. And 12 that was shown a few times in the previous 13 presentation.
14 Also for context, there were no RAIs 15 issued on this topical. Everything was handled as 16 part of the audit.
17 Just for some background information, 18 there is a lot of information that is out there that 19 is relevant to gas reactor source term analyses. This 20 is a subset of them that staff did not bring to bear 21 during this review, but are relying on these as part 22 of the current review that we're doing for these 23 models as part of the PSAR.
24 Also, I just want to highlight the fact 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
86 that there were several references that were brought 1
up at the Subcommittee meeting that have been brought 2
to our attention. We're so grateful for that feedback 3
from the Committee.
4 As far as the regulatory basis for this, 5
there are several.
6 50.34(a)(1), it does require an evaluation 7
of a postulated fission product release to evaluate 8
offsite radiological consequences.
9 50.34(a)(4) does require a preliminary 10 analysis and evaluation of SSCs, and the ones that are 11 most germane to this area include PDC 10, which is 12 SARRDLs for reactor design; RFDC 16, that's your 13 functional containment, and PDC 19, that's your 14 controllable.
15 Also, 50.43(a)(8), this is something that 16 is pretty unique to a construction permit application 17 under Part 50. And this is having the need for an R&D 18 program to resolve any safety questions that would 19 need to be done before you complete construction of 20 your plant.
21 Relevant to this -- and this was discussed 22 at the Subcommittee meeting -- is this 50.43(e). This 23 is the requirement that you have data to assess the 24 tools that you use in your plant analysis. This 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
87 requirement does not apply to a construction permit, 1
but it does apply to an operating license, pretty much 2
everything under Part 52. We had some discussion at 3
the Subcommittee about how this impacts this review, 4
as well as CPA reviews.
5 Scope of staff review, Section 4.2 of this 6
topical, it describes that the models that are in 7
XSTERM code are used to calculate dose consequences 8
for licensing basis events, as well as deterministic 9
evaluation of design basis accidents.
10 Staff reviewed the MST modeling approach 11 to address radionuclide transport phenomena to support 12 preliminary analysis of the Xe-100. However, staff's 13 review is limited to and focused on high-level 14 physical phenomena of interest and whether the 15 approach and methods can reasonably support future 16 licensing applications or actions. That is because 17 this design is preliminary.
18 The development and assessment of the 19 methods were in progress or planned, and therefore, 20 you know, the evaluation of models within XSTERM for 21 acceptability will be conducted during the review of 22 an application that relies on the results of XSTERM 23 evaluations. And that includes of current CPA that 24 they're looking at now.
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
88 Just for some more background, you know, 1
I was looking at the functional containment. It's 2
largely in the fuel kernel, your SiC and your PyC 3
layers, the fuel matrix, and the fuel-free zone of 4
fuel pebble within the helium pressure boundary.
5 There is a reactor building which is 6
actually present and would hold radionuclides, but it 7
is not credited in this methodology.
8 As far as the XSTERM itself, I don't want 9
to go into these models too much, but I want to 10 highlight these three here as far as a thermodynamics 11 calculation model.
12 This is one that we think is going to be 13 of a high importance because a lot of the radionuclide 14 release from the fuel, we expect it to be diffusion 15 dominant, and so, therefore, having a model that can 16 get that. So this is a model that we plan to look at 17 in more detail.
18 And the Point Kinetics Simulation Model, 19 the way it was described inside of this Topical 20 Report, it looked like it was a novel approach to 21 point kinetics. And so, that's one thing that we want 22 to make sure that we understand it and see how it's 23 used, and how it's different from what we consider, 24 like, a more traditional-type point kinetics approach.
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
89 And as far as the Tritium Model, it was 1
still under development at the time of this Topical 2
Report. So there was no -- there is no information on 3
this model.
4 There are six other models that were 5
discussed -- the Failure Probability Model, the Solids 6
Products Transport Calculation Model, Steady-State 7
Gaseous Fission Products Transport, dust generation, 8
Helium Pressure Boundary, and the Core Corrosion 9
Model.
10 Again, not looked at in detail, but staff 11 had determined that these models address phenomena 12 needed to predict source term to support preliminary 13 analyses. And that is based on the fact that these 14 models rely on previous modeling and operational 15 experience from gas-cooled reactors, such as the AVR, 16 and then, based on staff's experience with the light 17 water reactor and non-LWR source term analysis, there 18 were not significant gaps that were seen in these 19 models.
20 Section 4.2 does state that the source 21 term modeling as described may be revised. So staff 22 did not perform a detailed technical review of the 23 models as described in this Topical Report and made no 24 conclusions regarding the acceptability 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
90 models.
1 MEMBER MARTIN: Excuse me. It's Bob 2
Martin.
3 Just some question here. X-energy, of 4
course, acknowledges that there are inputs that come 5
in from other codes, like Flownex and VSOP. So we 6
reviewed Flownex as part of our last Subcommittee.
7 When it comes to VSOP, you know, what it is, the Very 8
Superior Old Program, where does that get reviewed by 9
the staff? Is it part of this? Is it part of 10 something else? I mean, I would say Flownex is part 11 of something else.
12 MR. DRZEWIECKI: Yes. So, yes, at the 13 Subcommittee meeting, this Topical Report, it was 14 submitted as a group of four. So it was source term.
15 It was safety analysis methods.
It was 16 GOTHIC/Flownex, and it was the core design 17 methodology.
18 MEMBER MARTIN: Okay. That's our fault 19 because we chose not to --
20 MR. DRZEWIECKI: Exactly. Yes, yes. It 21 was just the core design methodology was, yes, yes, 22 that was not subject to our previous meeting.
23 MEMBER MARTIN: Okay. Thanks for the 24 reminder. 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
91 MR. DRZEWIECKI: Yes. Okay. Yes.
1 So, as far as the code assessment plans or 2
the V&V, Section 6 of this Topical Report, it does 3
state, you know, yes, that the effort is underway to 4
ensure the XSTERM is qualified to support final safety 5
analyses. Validation plans are developed to cover 6
high-and medium-ranked phenomena that are identified 7
through the PIRT process. The phenomena modeled by 8
XSTERM are extracted from an earlier version of the 9
PIRT.
10 So, just to kind of clarify that, it does 11 state that a PIRT was done and that the V&V efforts 12 are informed by the PIRT. So, based on that, staff 13 determined that this process is acceptable because of 14 the identification of code assessment requirements 15 through the PIRT process. That's an established 16 approach. It's called out in 1.203.
17 However, we were unable to assess the 18 adequacy of the V&V plan because the validation plan 19 is not based on the latest PIRT information, even 20 though it's based on some PIRT information, not the 21 latest. And we had asked the information would 22 describe the knowledge level of the phenomena 23 identified in the PIRT. Just again, it's important 24 for us to know generally how important the phenomena 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
92 is, but how well understood it is, so that we have an 1
idea of what testing is needed. Also, the plan is 2
preliminary and subject to change.
3 As far as staff's conclusions, we 4
concluded that this Topical Report provides a 5
reasonable plan for the development of the source term 6
methodology. And that's because several of the models 7
appear to cover phenomena needed to predict source 8
term to support preliminary analysis and evaluation of 9
the Xe-100 design, and it describes an acceptable 10 approach to V&V, which is largely of informing what is 11 needed for code assessment based on a PIRT.
12 However, staff makes no conclusions 13 regarding acceptability of the models in XSTERM for 14 source term analysis because the models are still 15 under development. A detailed review of the models 16 was not completed, in part, because of that. Details 17 regarding key phenomena identification of associated 18 knowledge levels are not provided in this Topical 19 Report, and the models and associated validation plans 20 are preliminary and subject to change.
21 Additionally, this last bullet here, the 22 staff expects that a detailed technical review of 23 XSTERM model applicability to the Xe-100 reactor will 24 be addressed as part of the review of a licensing 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
93 application that makes reference to this Topical 1
Report.
2 To that last point, it is currently 3
referenced, again, in the Long Mott Generating Station 4
construction permit application, and we are to look at 5
these models in more detail now. So we just started 6
doing that, and I do expect that we will have further 7
engagement on this topic as part of that review.
8 So I think that's my last slide --
9 MEMBER PALMTAG: This is Scott Palmtag.
10 I just had a question on that.
11 So, from what we heard earlier, I thought 12 the XSTERM was going to be -- it was going to be 13 several years before that was finished because they 14 were waiting on validation from the INL. But it seems 15 like you are saying that this is going to be reviewed 16 in a construction permit?
17 MR. DRZEWIECKI: So, I will say, yes -- so 18 what we do need to make a finding on is the fact that 19 there are consequence evaluations performed in the 20 construction permit application that appear to make 21 use of XSTERM. And so, we do need to come to a 22 reasonable assurance or conclusion in terms of the 23 outcome of those calculations. Or, basically, to get 24 to some reasonable assurance finding of the fact that, 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
94 you know, the consequence analysis is acceptable.
1 Now, we understand that there are risks 2
associated with that. And so, we do have plans in 3
place to manage that risk, and that includes us doing 4
our own independent confirmatory calculations. I 5
believe this was discussed at the Subcommittee 6
meeting. But, as well, we just want to get a better 7
understanding of the capabilities and details of these 8
models, because they do appear to be used to support 9
findings in the CPA.
10 MEMBER HALNON: This is Greg. I just 11 looked at the CPA, since I wasn't aware that it is in 12 a different folder. But it does say that a separate 13 licensing Topical Report for XSTERM will be submitted 14 for the OLA. So we should see a specific Topical 15 Report for the XSTERM that would sort the rest of the 16 references that he was talking about.
17 MEMBER PALMTAG: At the OL?
18 MEMBER HALNON: At the OL stage, yeah.
19 MEMBER PALMTAG: I think that kind of 20 agrees with our first discussion.
21 MEMBER HALNON: Yeah.
22 MEMBER PALMTAG: It's consensual that 23 these models are --
24 MEMBER HALNON: Right, the Topical Report 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
95 will come before the OLA. So that's what's in 1
progress I think. I'm just reading into what this 2
paragraph says in the CPA.
3 MEMBER MARTIN: This is Bob Martin. I'll 4
throw out -- just to get a personal experience.
5 My last employer, implementation of NQA-1, 6
computer software development, we had a two-stage 7
qualification approach where the first feed of V&V was 8
taken care of. We called it commercial grade software 9
verification, which is oftentimes much easier to do 10 than gathering up all the data.
11 And I'm kind of looking over at X-energy.
12 Do you have anything like that where you can kind of 13 get that queue. You know, that's our designation. We 14 would put the queue, and then, down the road, we would 15 get the safety, where the second "V" would come in.
16 Do you all have any -- it's not, of course, explicitly 17 required by NQA-1, but it did allow you to have kind 18 of a demonstration of progress earlier in the 19 development than waiting until the end for everything.
20 I don't know, Milan, from that, if you 21 could speak on that?
22 And I don't know, Tim, if whether you've 23 ever seen anything like that by an applicant, where 24 they had kind of a two-step process with computer 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
96 software.
1 MR. DRZEWIECKI: Well, yeah, so I can 2
speak to this. And so, it's going to be somewhat 3
scattered and I could talk to you about, okay, so as 4
far as my background, I was NRC for a while. I was 5
Duke Energy for a while. About eight or nine months 6
ago, I was employed at Kairos Power. So, I'm familiar 7
with what they had done there. I was the Safety 8
Analysis Manager.
9 A lot of my experience at NRC was largely 10 in Part 52. And so, those methods were more further 11 along. They were supporting -- you know, because they 12 had to meet the 50.43(e) requirements.
13 MEMBER MARTIN: Right.
14 MR. DRZEWIECKI: I will say for the test 15 reactor, like what was done for Hermes, some of the 16 methodologies were further along. Source term was.
17 They had a Source Term Topical Report that was already 18 reviewed and approved and it was a little further 19 along, well ahead of their CPA.
20 However, their safety analysis methodology 21 was not. That was submitted as a tech report that was 22 supporting the Hermes CPA. And, in fact, I believe 23 you'll probably be seeing the safety analysis Topical 24 Report later this year.
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
97 So, it's been a mixed bag. So, in terms 1
of that, have I seen it staged like that? I've seen 2
people come in with preliminary information certainly 3
under a construction permit.
4 However, I do think source term is one of 5
those items, when it comes to like the siting 6
evaluation, where I think it's a little more important 7
to get that a little earlier.
8 MEMBER MARTIN: That's only fair.
9 Milan, if you want to say something, come 10 up underneath the green light.
11 MR. HANUS: Yeah, it's not -- X-energy.
12 We have the validation. So it's important. So we 13 have some preliminary. It's also very -- I wouldn't 14 call it a staged approach by design where it 15 designates stage one, stage two.
16 We can provide some information already on 17 the validation results. It has been done. But, yeah, 18 it's preliminary.
19 MEMBER MARTIN: I guess, as Tim said, 20 they've seen a lot of different examples. I guess, as 21 we kind of feel through the CP application review 22 process, you know, I think in the moment, are we 23 letting a lot of -- are we allowing for that 24 ambiguity? We find it acceptable?
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
98 But something else to consider is, at 1
least from traceability, since that first "V" is a lot 2
easier, it might help in-house to have that specified 3
in your procedures to have it earlier. It also helps 4
with, like, if you're having turnover and you can kind 5
of put a stamp on it, and then, pick it up again. But 6
that's another story.
7 Anyway, all right. Thanks for that 8
clarification.
9 CHAIR KIRCHNER: Tim, did we let you 10 complete your presentation?
11 MR. DRZEWIECKI: Yes, yes, yes. That was 12 all that I had.
13 CHAIR KIRCHNER: Okay. Thank you.
14 MR. DRZEWIECKI: Yeah, yeah, outside of 15 the acronym slide.
16 CHAIR KIRCHNER: Okay. So, members, 17 questions for the NRC staff?
18 (No response.)
19 CHAIR KIRCHNER: Okay, Bob. I think 20 that's it for the presentations.
21 MEMBER MARTIN: Yes, good.
22 CHAIR KIRCHNER: At this point, should we 23 take public comments?
24 MEMBER MARTIN: 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
99 CHAIR KIRCHNER: So, for any members of 1
the public online, if you wish to make a comment, just 2
raise your hand or open up your mic and speak out.
3 And state your affiliation, if appropriate, and your 4
comment.
5 (No response.)
6 CHAIR KIRCHNER: Hearing or seeing none, 7
okay, thank you. With that, then, Bob, I think, and 8
Dave, we're ready to entertain a letter.
9 Okay. So we may need to take a short 10 pause break here, and then we'll bring up a draft 11 letter we've worked for reading into the record.
12 And then, with that, we can excuse the 13 court reporter, I believe, for the rest of the 14 meeting.
15 So just hold for a bit while we bring up 16 our letter report.
17 CHAIR KIRCHNER: While we're doing that, 18 I'll thank both X-energy and the staff for their 19 presentations today.
20 (Pause.)
21 For those online, we're just taking a 22 short break here. We'll resume within a couple of 23 minutes.
24 (Whereupon, at 11:06 a.m., the foregoing 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
100 matter went off the record and went back on the record 1
at 11:08 a.m.)
2 CHAIR KIRCHNER: Okay, I think we're ready 3
to go.
4 And I'll turn now to Dave Petti, who is 5
our lead author on this letter report. Go ahead, 6
Dave.
7 MEMBER PETTI: Okay. This is the X-energy 8
Topical Report Mechanistic Source Term Approach.
9
Dear Mr. King:
10 "During the 727th meeting of the Advisory 11 Committee on Reactor Safeguards, held from July 8th 12 through 11th" -- that should be "9th" -- "2025, we 13 completed our review of the X-energy Topical Report 14 entitled, "Mechanistic Source Term Approach," and the 15 associated Draft Safety Evaluation.
16 "Our X-energy Subcommittee also reviewed 17 this matter on June 3, 2025. During these meetings, 18 discussions with the Nuclear Regulatory Commission 19 staff and X-energy were beneficial, as were the 20 methods documents.
21 "Conclusions and Recommendations.
22 "One, X-energy is developing a
23 sequence-specific mechanistic source term through the 24 use of a functional containment concept for the Xe-100 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
101 pebble bed reactor. This functional containment 1
consists of the tristructural isotropic fuel kernel 2
and coatings, the pebble bed matrix, and the helium 3
pressure boundary.
4 "Two, we recommend caution regarding the 5
verification and validation of the models in the 6
mechanistic source term code. This process will be 7
challenging due to the numerous models and submodels 8
- involved, their complexity, the absence of a
9 historical database, the lack of applicability of 10 uranium dioxide TRISO fuel performance models to 11 describe uranium oxycarbide TRISO fuel performance, 12 and residual uncertainties associated with the 13 constitutive models and material properties.
14 "A simpler semi-empirical approach, tied 15 more directly to the statistically significant failure 16 data from the U.S. Advanced Gas Reactor UCO TRISO 17 testing program with appropriate margins would 18 overcome these shortcomings and better represent the 19 performance of the fuel that will be fabricated for 20 the Xe-100 and the resulting source term.
21 "Three, the Topical Report should be 22 issued and the concerns mentioned in the letter should 23 be considered for the future licensing application.
24 "Functional Containment Strategy.
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
102 X-energy is implementing a functional containment 1
strategy to retain fission products within their 2
design. The functional containment consists of the 3
fuel kernel, the silicon carbide and pyrocarbon layers 4
of the TRISO coating, the fuel matrix surrounding the 5
particles and the pebble, the fuel-free zone on the 6
outside of the pebble, and the helium pressure 7
boundary. No credit is given for fission product 8
retention in the reactor building.
9 "The release of fission products is 10 calculated based on the as-manufactured heavy metal 11 contamination and silicon carbide defects, in-service 12 failures under irradiation, and incremental failures 13 during licensing basis events.
For certain 14 radionuclides, silver diffusion through intact 15 coatings is also considered. Additionally, plate-out 16 on the surfaces of the helium pressure boundary and 17 dust in the system is accounted for, as well as 18 resuspension of" -- quote -- "`liftoff' of the dust 19 during licensing basis events.
20 "Evaluation Model Development.
21 "The X-energy evaluation model known as 22 XSTERM comprises several computer models and submodels 23 to describe the generation, release, and transport of 24 radionuclides from the fuel to the environment. 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
103 evaluation model includes codes to calculate the 1
thermal hydraulic response of the fuel pebbles in the 2
core under normal operation and licensing basis 3
events; the production, decay, and transmutation of 4
radionuclides in the pebble and their transport to the 5
coolant, including the effects of dust; the transport 6
to the reactor
- building, and evaluations of 7
radiological dose to compare to regulatory limits.
8 "The Topical Report does not present a 9
Phenomena Identification and Ranking Table (PIRT) to 10 identify key phenomena, as required by Regulatory 11 Guide 1.203, as an early step in the development of an 12 evaluation model.
13 "Key submodels in XSTERM include the 14 Particle Failure Model, which considers pressure 15 vessel failure, kernel migration, fission product 16 corrosion, thermal decomposition, inner pyrolytic 17 carbon layer cracking, and manufacturing defects.
18 "Solid fission product release and 19 transport is based on diffusion through the kernel, 20 coating layers, and fuel matrix using detailed 21 nodalization of each pebble in the analysis.
22 "The thermodynamics module calculates 23 temperatures in all the reactor components necessary 24 to support a mechanistic source term.
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
104 "Steady-state fission gas release is 1
calculated using two different, German and U.S.,
2 release models that account for (a) diffusion through 3
the kernel and buffer porosity for failed and 4
defective fuel, and (b) through the matrix for the 5
initial heavy metal contamination.
6 "Dust production is estimated using the 7
measured particle size and estimated generation rates 8
for the German pebble bed AVR scaled to Xe-100.
9 "Fission product transport, deposition, 10 and liftoff behavior in the helium pressure boundary 11 is based on models for absorption of fission products 12 on dust, plate-out of the dust, and condensable 13 fission products -- for example, cesium and strontium 14
-- during normal operation and subsequent liftoff 15 under licensing basis events.
16 "Core corrosion models describe the 17 response of the core to oxidation events, including 18 models for mass transport and chemical reactions 19 during an air or steam ingress event, using data and 20 air and stream oxidation behavior of graphite, and 21 data from industrial chemical synthesis technologies.
22 "The tritium behavior module tracks the 23 production, decay, permeation, and absorption of 24 graphite release upon oxidant ingress into the core, 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
105 separately from other fission products.
1 "The Reactor Kinetics Model uses 2
two-dimensional kinetics for steady-state and steam 3
generator tube rupture events.
4 "Atmospheric dose calculations are based 5
on traditional Gaussian plume dispersion and appear to 6
be consistent with NRC dose models for use in siting 7
and control room habitability assessments.
8 "The isotope listing used in XSTERM is 9
consistent with previous high-temperature gas-cooled 10 reactors safety assessments.
11 "Staff Safety Evaluation.
12 "The staff safety evaluation concentrated 13 their review on the evaluation methodology. They 14 found that the overall plan appears reasonable.
15 However, they noted that the verification and 16 validation plan is not based on a PIRT. The details 17 of the individual models of XSTERM were not examined 18 and no conclusions were made about the acceptability 19 of the models, as they are still under development.
20 A detailed review is expected as part of the licensing 21 application.
22 "Discussion. V&V Plans.
23 "X-energy intends to verify and validate 24 the modules of XSTERM using a combination 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
106 historical German and Chinese data related to UO2 1
TRISO fuel with more modern U.S. UCO TRISO fuel and 2
analytical benchmark problems from HTR-10. In some 3
cases, separate effects data are utilized to validate 4
individual submodels, such as SANA pebble bed heat 5
transfer testing, the German VAMPYR plate-out data, 6
the Chinese Lifting Line Platform Facility data on 7
dust generation.
8 "The following sections outline our 9
assessment of these plans, their limitations, and some 10 recommendations to enhance the validation process.
11 Staff is now looking to additional data not mentioned 12 in the Topical Report.
13 "Defects, Failure Fractions, and 14 Performance Envelopes.
15 "The performance of TRISO fuel in a 16 high-temperature gas-cooled reactor is a strong 17 function of the initial level of manufacturing 18 defects, contamination, and silicon carbide defects; 19 the in-service failure through irradiation, and the 20 incremental failures during accidents.
21 "Because of the importance to the overall 22 functional containment strategy, the initial fuel 23 failure fractions used in the X-energy safety analysis 24 approach must be demonstrated with actual 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
107 manufacturing data to be consistent with or better 1
than the AGR UCO data. The values used for these 2
fractions should align with the results from the AGR 3
UCO TRISO testing, along with any additional fuel 4
qualification testing planned by X-energy.
5 "Finally, the Service Conditions.
6 "Burnup, temperature, fast fluence 7
experienced during normal operation and postulated 8
licensing basis events should remain inside the 9
testing envelopes associated with the U.S. AGR UCO 10 database, as supplemented by any additional testing 11 planned on UCO by X-energy.
12 "Fission product groupings.
13 "The database on fission product behavior 14 in TRISO fuel focuses on measurements of noble gases, 15 cesium, strontium, europium, silver, and in some 16 limited cases, iodine. Consequently, grouping of 17 fission products into classes based on chemical or 18 volatility considerations is essential for estimating 19 source terms for HTGRs, similar to other reactor 20 technologies.
21 "For HTGRs, iodine and tellurium are 22 recommended to be modeled as noble gases, while 23 europium should be treated similarly to strontium.
24 Lower-volume fission products such as lithium 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
108 cerium are not expected to be released from UCO TRISO 1
fuel, according to measurements from the AGR program.
2 These groupings align with previous HTGR safety and 3
source term assessments.
4 "Steady-State Fission Gas Release.
5 "The report did not mention steady-state 6
fission gas release data collected over a large range 7
of temperature runup from the U.S. UCO TRISO program 8
(AGR-3/4) that contained failed fuel. These data have 9
been published (Reference 2) and can be used directly 10 for fission gas release from exposed kernels or used 11 to validate models for this part of the source term.
12 "Fuel Performance.
13 "The fuel performance validation plan has 14 three major shortcomings.
15
"(a) The models predominantly describe UO2 16 rather than UCO TRISO fuel performance.
17
"(b) The failure mechanisms observed in 18 the AGR UCO TRISO irradiations and heating tests was 19 not accounted for in the evaluation model.
20
"© The uncertainties of material 21 properties required to describe fuel behavior are 22 significant enough to make validation challenging.
23 "A.
UO2 versus UCO TRISO Fuel 24 Performance.
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
109 "X-energy plans to validate the fuel 1
performance models using a combination of irradiation 2
and heating or safety data from UO2 (HFR K5 and K6) 3 and UCO TRISO fuel (AGR-1 and AGR-2). While some of 4
this data is acceptable, the UO2 data does not fully 5
represent modern U.S. TRISO fuel. Many, if not all, 6
of the failure mechanisms models in XSTERM have been 7
engineered out of UCO TRISO fuel. This fuel 8
development approach was adopted in the U.S. decades 9
ago by modifying particle design, altering of 10 fabrication conditions, and limiting reactor service 11 conditions. These mechanisms are not optimal in UCO 12 TRISO. It raises questions about the effectiveness of 13 use such a validation effort for XSTERM fuel 14 performance models.
15 "B. Failure Mechanism in AGR Testing.
16 "The applicant plans to use the results of 17 AGR-1 and AGR-2 in their validation efforts. However, 18 it is important to note that the failure mechanism 19 noted for USO TRISO in AGR-1 and AGR-2 (Reference 3) 20 is not modeled in any particle fuel performance code 21 due to uncertainties in the material properties 22 necessary to support the model. Advanced multiscale 23 modeling efforts are underway in the Department of 24 Energy's Nuclear Energy Advanced Modeling 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
110 Simulation Program as a challenge problem because of 1
its difficulty.
2 "C. Material Property Uncertainties.
3 "From a
fission product release 4
perspective, attempts to validate the diffusion of 5
fission product release models using data from AGR-1 6
and AGR-2 have had limited success. This is due to 7
the low level of releases that were measured and the 8
potential multiple sources of fission products --
9 contamination, exposed kernels, or through intact 10 particles -- and the uncertainties of the underlying 11 diffusion coefficient database.
12 "The report also does not address many of 13 the well-known uncertainties in the thermomechanical 14 properties that significantly influence model 15 traditions. For example, the silicon carbide strength 16 data in the Topical Report is based on German TRISO 17 fuel and shows considerably greater strengths than the 18 more recent U.S. UCO TRISO fuel. This difference is 19 likely due to differences in the microstructure of the 20 silicon carbide layer.
21 "Additionally, the pyrolytic carbon 22 shrinkage rate data used in XSTERM that determined the 23 survivability of the PyC layers are simple 24 fluence-based estimates from old German testing. They 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
111 do not represent the more contemporary data used in a 1
recent International Atomic Energy Agency Code 2
benchmark that are temperature-influence-dependent.
3 These data are based on a compendium of historical 4
U.S. and international data and are considered to be 5
more representative for UCO TRISO fuel behavior.
6 "The uncertainties in these fundamental 7
material properties are impacted even by the small 8
scale of the samples and no limitations in testing.
9 These thermomechanical properties -- the silicon 10 carbide and pyrocarbon strength data and the 11 pyrocarbon shrinkage and creep data -- drive failure 12 probability predictions, and the uncertainties are 13 large enough to make validation of failure 14 probabilities challenging.
15 "Overall, considering the significant 16 performance margin in the Xe-100 core and the concerns 17 mentioned above, it may be more beneficial to directly 18 use the measured statistically significant failure 19 data from AGR-1 and AGR-2 with appropriate margin as 20 an estimate for the fuel failure fraction in source 21 term calculations.
22 "Fuel Behavior Under Reactivity Events.
23 "Fuel performance under reactivity events 24 is not mentioned in the Topical Report. Did it exist 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
112 on such fuel performance. (See Reference 9.) This 1
may not be a problem for the Xe-100 design, because of 2
the low access reactivity in the core in their 3
reactivity control strategy, but some mention of fuel 4
behavior is worthwhile from a
completeness 5
perspective.
6 "Dust.
7 "The models for dust generation, 8
transport, deposition, plate-out, and resuspension of 9
liftoff are highly complex and challenging to 10 determine if they are conservative. Sensitivity 11 studies:
12 "A. Examine the timing of dust liftoff 13 relative to fission product release during licensing 14 basis events.
15 "B. Varying the dust generation rate.
16 "C. Performing calculations with and 17 without dust are recommended to help establish the 18 overall role of dust on fission product transport in 19 the Xe-100 during postulated events and provide more 20 confidence in the predictions.
21 "Beyond the VAMPYR plate-out data in 22 Germany, the report does not mention the extensive 23 testing done in the COMEDIE facility in the 1990s (See 24 Reference 10) to examine deposition and subsequent 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
113 liftoff under various break sizes.
1 "Additionally, the applicant appears 2
unaware of the large amount of data on the suspension 3
of metallic aerosols with dust in the aerosol 4
literature (See References 11, 12, and 13) that could 5
be useful for validation.
6 "Ultimately, validation through 7
measurements of gaseous and metallic fission products 8
during operation will be required to assure specified 9
acceptable radionuclide design release limits, or 10 SARRDLs, are being met.
11 "Core Corrosion/Oxidation.
12 "The existing database on the response of 13 TRISO fuel to aerosteam ingress is quite limited.
14 Some in-trial testing has been performed for short 15 duration at specific temperatures, focusing primarily 16 on fission gas rather than fission metals that tend to 17 dominate the radiological dose. There are some 18 limited data on particle testing in air that show high 19 failure rates. (See Reference 9.) However, there are 20 no data on the effects of steam ingress on UCO TRISO 21 coated fuel particles, as would be expected in a steam 22 generator tube rupture. The assumption index that has 23 determined that steam will not result in particle 24 failure is unsubstantiated.
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
114 "During discussions in our Subcommittee 1
meeting, the applicant mentioned that in their 2
simulations the steam never reaches the particles due 3
to the fuel-free zone on the surface of the pebble.
4 This behavior will need to be validated.
5 Nevertheless, given the importance of these events, as 6
highlighted by the results of the MHGT, modular 7
high-temperature gas-cooled reactor probabilistic risk 8
assessment, the U.S. AGR program depends on testing 9
fuel compacts in steam and air under a wide range of 10 temperatures and partial pressures of oxygen. (See 11 Reference 14.) The testing will measure fission gas 12 and metallic fission product release during such 13 exposures.
14 "Historically, such testing of this nature 15 has been planned in the U.S. TRISO fuel qualification 16 programs for decades. However, it was never carried 17 out, per the requirements of special furnaces and 18 fission collect detection systems. These systems have 19 only recently been developed under the AGR program.
20 "In summary, X-energy is developing a 21 sequence-specific mechanistic source term using a 22 functional containment concept for their Xe-100 pebble 23 bed reactor. The functional containment consists of 24 the TRISO fuel coatings, a pebble bed matrix, and 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
115 helium pressure boundary.
1 "We recommend caution regarding the 2
verification and validation of the models in the 3
mechanistic source term code. This process will be 4
challenging due to the numerous models and submodels 5
involved, their complexity, gaps in the historical 6
database, the lack of applicability of UO@ TRISO fuel 7
performance models to describe UCO TRISO fuel 8
performance, and residual uncertainties associated 9
with the constitutive models and materials properties.
10 "A simpler, semi-empirical approach, tied 11 more directly to the statistically significant failure 12 data from the U.S. AGR USO TRISO testing program with 13 appropriate margins, would overcome these shortcomings 14 and better represent the performance of the fuel that 15 will be fabricated for the Xe-100 and the resulting 16 source term.
17 "The Topical Report should be issued and 18 the concerns mentioned in the letter should be 19 considered for the future licensing application.
20 "We are not requesting a response to this 21 letter."
22 CHAIR KIRCHNER: Thank you, Dave.
23 Any comments, Members?
24 MEMBER MARTIN: In that category, just 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
116 as it relates to the EO. The EO expects this new, 1
novel, and unique and maybe in the statement, maybe in 2
the beginning of the background, that captures that 3
thought? It might be a P&P topic, too.
4 CHAIR KIRCHNER: It might be a P&P topic.
5 MEMBER PETTI: Yes, I developed most of 6
the letter before the EO came out.
7 MEMBER MARTIN: Right.
8 MEMBER HALNON: Yes, that's what I was 9
going to say. I think we have some things that we 10 need to finish up, regardless, and we just can't chop 11 off what we started. But I think your point is good 12 that we need to mention that or something to that 13 effect.
14 MEMBER MARTIN: Well, my suggestion was 15 I'd just take the words that I used today -- well, 16 I'll hand it off. I guess the message that came to me 17 prior to the meeting was that we would say something 18 on the "novel and unique." And so, I had the three or 19 four sentences.
20 MEMBER HALNON: Okay. We'll look into it, 21 maybe put those in.
22 MEMBER MARTIN: For your consideration.
23 No, no, that's fine.
24 MEMBER HALNON: Yes, I had just 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
117 recommendation. I just think that Recommendation No.
1 2 can be shortened and take credit for what you have 2
in the body of the letter. I mean, we can point to 3
it.
4 MEMBER MARTIN: Well, my thought with 2 is 5
that there's actually two points, and that you split 6
it with the two.
7 MEMBER PETTI: Yes, I'm open to that.
8 MEMBER HALNON: We can do that, too. I 9
mean, I just think it's laborious reading through.
10 You kind of lose the -- but I think you do a really 11 good in the body describing the shortcomings that need 12 to be checked-out as we go forward.
13 MEMBER PETTI: It was hard to think of it.
14 MEMBER HALNON: Yes, I think you had a 15 statement in there that said that it was not based on 16 a current PIRT. What I thought I heard was -- I mean, 17 I've heard, what I thought I heard was that there was 18 a PIRT, but it needs to be updated, and the updated 19 PIRT is not reflected in this. I just want to make 20 sure that we talk a little --
21 MEMBER PETTI: It was not in the -- they 22 have done it, but we haven't seen it.
23 MEMBER MARTIN: Exactly. Yes, I had --
24 MEMBER MARTIN: I'm happy to take it --
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
118 MEMBER ROBERTS: Greg, the only think I 1
suggest is taking out that paragraph in our discussion 2
because it's in the staff discussion. I don't think 3
we need to talk about PIRTs at all 4
MEMBER HALNON: Yes, that could be the 5
answer, but my point was --
6 MEMBER ROBERTS: It kind of breaks up the 7
flow of the background anyway.
8 MEMBER HALNON: -- that it wasn't based on 9
a PIRT, but it really sounds like, from what the staff 10 said, that the PIRT was done, just that it was 11 preliminary and they've got to update it.
12 MEMBER MARTIN: An alternative that I 13 pitched to Dave yesterday --
14 MEMBER PETTI: Which I haven't seen yet.
15 MEMBER MARTIN: Oh, okay. Which was 16 basically give them credit for that big laundry list.
17 Right?
- Because, effectively, they identified 18 phenomena. Right? And they incorporated XSTERM. And 19 so, I have words to that effect about, you know --
20 well, I don't really mind --
21 MEMBER PETTI: I'm willing to take it out 22 because if we're going to put in stuff on new and 23 novel, this letter is going to get really long.
24 MEMBER MARTIN: Right, but it diminished 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
119 the PIRT point and just elevated what they did do.
1 MEMBER PETTI: I noticed little things.
2 First of all, X-energy looked at it. XSTERM is not in 3
an evaluation model. So I need to know -- but it's a 4
conglomeration of codes together. And I'm trying to 5
-- I need a term for it, Bob, as the evaluation model.
6 But it's a piece of the evaluation model? Is that how 7
it should be? I mean, how do we describe it? And are 8
those are modules inside? Is that a way to --
9 MEMBER MARTIN: "Evaluation model" has 10 this motherhood kind of definition, and the codes 11 themselves are just a piece. It's really the 12 responsibility of X-energy to give us the right word 13 for it. Look over there.
14 MEMBER PALMTAG: But it is a code. It's 15 not --
16 MEMBER PETTI: Yes, it is not the 17 evaluation model. I describe it as the evaluation 18 model. But the evaluation model, this is one piece.
19 MEMBER MARTIN: Yes, but it's not like a 20 whole bunch of codes. It's just one code.
21 MEMBER PETTI: Correct. Yes. So we can 22 just call it a computer code. That is part of the 23 evaluation model. That's what I was thinking.
24 MEMBER BALLINGER: I come away with this 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
120 thinking: you think these guys are fools. It's very 1
damning. It may be correct. It may be correct, but 2
it's not very complimentary letter. It's not a 3
complimentary letter at all.
4 The implication is these guys should have 5
known about this. They didn't or they did know about 6
it and decided not to deal with it. But there's 7
something very negative here and I'm wondering whether 8
or not that's -- if that's the message you want to 9
send, okay, but it comes across to me as quite 10 negative.
11 MEMBER PETTI: It surprised me when I read 12 it, that I had to reference what I consider all the 13 contemporary stuff there is about TRISO fuel.
14 MEMBER BALLINGER: So you are intending 15 for it to be negative?
16 MEMBER PETTI: Well, I'm trying to 17 highlight to the staff that, okay, here's their view 18 of the world, but, no, there are these other 10 19 references that are absolutely seminal if you're going 20 to do a proper review.
21 CHAIR KIRCHNER: And they all have many 22 failures.
23 (Laughter.)
24 CHAIR KIRCHNER: I had a couple 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
121 comments along this line as well. You know, there's 1
a statement towards the back and near the summary, or 2
something, that says, that it's summarized on the 3
bounding approach, however we say it -- it will 4
provide a result that is more representative of fuel 5
performance. That seems pejorative to me. That's 6
saying that the work they're doing is going to produce 7
a result that is so wildly inaccurate that, if you 8
just do a back-of-the-envelope calculation, it would 9
be better than that.
10 So I think this is kind of what Ron is 11 pointing at in his comment.
12 MEMBER PETTI: My point was, their models 13 sit on UO2, not on UCO. So what they're doing is 14 calculating to a fuel system that's not going to be 15 used in the reactor. If you go back to the data that 16 is going to be used in the reactor, you get a source 17 term that's representative of that. That's all I was 18 trying to --
19 MEMBER MARTIN: You're saying they're 20 basically false, right?
21 MEMBER PETTI: There's very little, 22 though, that's UCO. There's one model that they put 23 in, but, again, that was engineered away. It was put 24 in, and then it was engineered away.
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
122 MEMBER SUNSERI: And then, a second 1
comment or point, and this will express probably some 2
ignorance on my part. We go into a lot of depth and 3
discussion about failure of the fuel and, you know, 4
the data and all that stuff. I am under the 5
impression -- well, first off, when we reference fuel 6
like that, are we talking about the particle, the 7
TRISO particle, which I think is the fuel, right?
8 MEMBER PETTI: Mm-hmm.
9 MEMBER SUNSERI: So, I mean, we've 10 approved a couple of Topical Reports on how to build 11 a particle and do it in a quality manner, and, I mean, 12 all of the release and failure mechanisms and all 13 that. Why are we dragging all this stuff into that?
14 Build the TRISO particle in accordance with 15 established guidelines. That's what they need to do.
16 Now, if you want to compose the particles 17 into a pebble, and then, if we've got some concerns 18 about that dust, or whatever, then that's another 19 story. But it just seems to me we're going into an 20 area that's already been reviewed and approved when we 21 talk about failure mechanisms of the particle.
22 Because a couple of years ago, we did 23 approve a Topical Report for the quality -- and this 24 really is about a particle.
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
123 MEMBER PETTI: Yes, and this report does 1
not leverage -- it does not truly leverage what's in 2
there. That's what I'm saying: guys, here's this 3
report. If you read it and you really understand what 4
it says, this is what it means. It says: don't use 5
German models on UO2 that aren't appropriate for UCO.
6 MEMBER MARTIN: When do we expect, though, 7
the UO2 performance to be worse, meaning you would get 8
higher source terms? And can you make an argument 9
that these are conservatively applied to UCO? I mean, 10 obviously, they have to prove it, but, I mean, if your 11 whole point was bringing in UCO because it eliminates 12 certain failure mechanisms, this should be -- and if 13 they could live with UO2, and then that --
14 MEMBER PETTI: You can make a pebble bed 15 with UO2. It may not be economic, but you can make 16 one. That's why UCO exists.
17 MEMBER MARTIN: But if you're calculating 18 a source term with UO2 data, would you expect it to be 19 worse than UCO? And that's an argument for --
20 MEMBER PETTI: I don't have any problem 21 with the calculation of the source term.
22 MEMBER MARTIN: Right.
23 MEMBER PETTI: I didn't say anything --
24 the diffusion coefficient, that's all completely fine, 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
124 even the dust. I think they can -- we've talked about 1
it in the Subcommittee. I put some recommendations in 2
there, so the staff can think about, you know, if 3
they're going to stand up confirmatories, hey, here's 4
some sensitivity studies you guys ought to do because 5
I'm sure the applicant's done it. And then, you'll be 6
in a good place.
7 But using models from the 1970s for a fuel 8
system that isn't this fuel system, it just raises the 9
question. If the staff asks, well, why is this valid?
10 MEMBER MARTIN: Right. Well, this goes 11 back to -- I mean, you can't stand on just UO2 alone.
12 If you want to make an argument that --
13 MEMBER PETTI: One thing that I think that 14 it may not be is they're using strength data to make 15 the product super strong. The copies are very strong.
16 They would make them super strong. You would never 17 calculate a failure.
18 Some of the data I saw really jumped out 19 me because it didn't -- it's not consistent.
20 MEMBER MARTIN: They talked about that 21 they have a best estimate mode. Now, if this best 22 estimate brought in more UCO data, or something like 23 that, you might be able to make -- again, I'm 24 advocating a little bit too much probably, you 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
125 just for the sake of a discussion.
1 MEMBER PETTI: I also just think that, you 2
know, there's already plenty of use of UO2. Every 3
diffusion coefficient we know --
4 MEMBER MARTIN: Right.
5 MEMBER PETTI: -- is UO2. There's nothing 6
on UCO.
7 MEMBER MARTIN: That's IAEA import data.
8 MEMBER PETTI: Right, that's all UO2. But 9
when you get into the failure stuff, you get into just 10 a really messy, messy area, where you can say 11 something fails. You take the uncertainty band on a 12 parameter. I could tell you what they're like. And 13 it goes from everything's okay to you get a large 14 failure fraction. So the uncertainty is right in the 15 wrong spot. You can prove it.
16 But, yes, I mean none of the testing done 17 in France on liftoff was mentioned, which was 18 surprising to me. Those were multimillion experiments 19 that they got. So there were just a lot of things 20 that I was, like, well, gee, I wonder why that wasn't 21 22 MEMBER MARTIN: They assume 100 percent 23 gets out on liftout.
24 MEMBER PETTI: They won't. They've gotten 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
126 enough correlations. They're not going to -- no.
1 CHAIR KIRCHNER: I think I heard Vesna 2
trying to say something.
3 MEMBER MARTIN: Yes.
4 MEMBER DIMITRIJEVIC: Yes.
5 CHAIR KIRCHNER: Go ahead, Vesna.
6 MEMBER DIMITRIJEVIC: Yes. So I am really 7
confused with this letter because I cannot really 8
define what this letter is. Because how can we write 9
such a long letter on just this preliminary? This is 10 just mostly planned how to develop the methodology.
11 It's not the methodology. All right. So this plan 12 may have its shortcomings, you know, but, very 13 clearly, they need the models, or the validation plans 14 are just preliminary and they're subject to change.
15 So, then, I was thinking, okay, maybe we 16 are giving them advice, what should they consider or 17 to do. But are those already our concerns? I mean, 18 it just happened that Dave has really a big experience 19 of that and he can be helpful in there, you know, 20 telling them to make sure not to miss that; to 21 consider this data, and that this can be an issue.
22 But this is just a plan. This is not 23 methodology. So, why are we going into so many 24 details about it? Can we just phrase it differently, 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
127 you know? Just say, in the development of 1
methodology, as you are finalizing this, these should 2
be concerns, blah, blah, blah, blah.
3 I mean, they identify phenomena, right?
4 I mean, you know, I don't really know what is the --
5 it is not criticism of what is done, because nothing 6
is done. It is just a preliminary plan. And so, is 7
it our advice? What is it? How do you see this 8
letter, Dave?
9 MEMBER PETTI: Well --
10 MEMBER HALNON: Dave, let me referee this 11 a little bit because I've been trying to reconcile the 12 same question that Vesna has in my mind, but I was 13 doing it with the Topical Report, when I called it 14 kind of "white paperish."
15 We're certainly in the pre-application 16 stage for the operating license at this point and this 17 is a little -- I don't know. It's actually that we're 18 expecting the full Topical Report for this in the 19 operating license.
20 So we're in this mode of it's a little bit 21 unique from the standpoint that it was -- could we 22 call it "preliminary"? But just like the staff in 23 their SER is providing some assurances and some 24 agreements on approaches on how to do this, and it 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
128 looks like it's going to be good, and it's a similar 1
mode of providing some assurance and some additional 2
information -- you could call it "advice," if you want 3
-- on what the Topical Report for XSTERM actually 4
needs to look at when we approve it for use to 5
actually calculate a number that would show acceptable 6
release rates and source term calculations for 7
emergency plans and other licensing basis events.
8 So I think, in my mind at least, we are in 9
a mode of here's a lot of information that's already 10 out there. It is a little bit damning, from what Ron 11 was saying, the fact that you should have already 12 included all this in our discussions. But part of the 13 strength of the Committee is that we have depth of 14 experience, that we need to make sure that it's not 15 "Bring me a rock and I'll let you know if you got it 16 right." It's "Here's what the rock needs to look 17 like," or at least consider going forward.
18 So that's how I've reconciled that in my 19 mind this is actually a pretty good letter from the 20 standpoint of providing another level of information 21 that can be used. So, whenever we see the final 22 Topical Report, we won't have these same questions.
23 MEMBER PETTI: So my concern when I put 24 this together was thinking about when the OL is going 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
129 to be here. I won't be here when the OL is here. So, 1
I thought I would give the staff the bread crumbs. As 2
we've talked about in other letters, these are things 3
you ought to be looking for.
4 I predict the OL is at least five years 5
away. There's no way they can execute the experiments 6
in Idaho and post-irradiation testing in a couple of 7
years. It's at least two years in ATR. It could be 8
even closer to two and a half.
9 Then, cool it down. Bring it into the hot 10 cell. Stick it in the furnaces. You know, that takes 11 time.
12 That's why I was kind of surprised when I 13 heard that they were actually relying on that for 14 validation. Relying on it as a proof test is a 15 different thing than relying for validation.
16 But, you know, just fast-forward. There's 17 only going to be a handful of us on the Committee in 18 five years. So again, this is kind of a --
19 MEMBER SUNSERI: So, why don't we -- I 20 mean, an approach would be to write a shorter base 21 letter that strikes right at the point that Vesna and 22 Greg just outlined, and then, put an appendix with all 23 this technical "for your future consideration" stuff 24 in the back. Because, quite frankly, all 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
130 discussion, in my mind, just distracts really. I 1
don't even know what we're saying, as far as the 2
Topical Report goes, until you get to the very end.
3 But that's just me.
4 Once again, it would be clearer if there 5
was a very succinct letter that says: go forward with 6
the model or develop the model. But refer to the 7
appendix portion for some potential shortfalls, and 8
whatever.
9 MEMBER HALNON: I like the letter. I 10 think it's appropriate. Remember, this TR is asking 11 for a roadmap. You know, they're coming in. This is 12 the roadmap we're going to follow for our licensing.
13 And this is where we found some holes in it. There's 14 some stuff missing. So I think it's very appropriate 15 to have it here instead of waiting until later, until 16 we actually see this XSTERM, and then, pointing out 17 the holes. It's better to have the information early.
18 You're right that the full XSTERM TR is 19 going to be the operating license, but the NRC 20 presentation talked about a lot of the results are 21 going to be needed for the construction license, too.
22 So it wasn't really clear, but it sounds like they're 23 not going to have the full XSTERM now, but they're 24 definitely going to have results at the CP 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
131 So they are going to need this stuff 1
sooner than later. So I like the tone of the letter.
2 MEMBER SUNSERI: Can we kind of make this 3
kind of a simple change on the letter by -- Sandra, if 4
you can go to 3? There's just a few lines there.
5 Just kind of put that last conclusion in a more 6
specific context about this being adequate for review 7
of Xe-100 PSARs, which is basically what the staff 8
says.
9 But the way it reads, it's just kind of 10 one of our generic topicals. It's fine. And if we 11 narrow it down to, okay, well, it has a limited 12 application, acknowledged by the staff, go ahead and 13 issue it, and then we'll see.
14 I think your point about the concerns kind 15 of addresses what we're kind of talking about. It's 16 that we're presenting some concerns. We hope to see 17 those addressed down the road.
18 CHAIR KIRCHNER: I just would make two 19 points.
20 One, that two of these points were 21 actually in the staff's conclusions. One is that it's 22 limited to preliminary analyses, and two, that it's a 23 reasonable plan for development of an MST methodology.
24 And I think somehow we ought to capture that.
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
132 I want to stop us here. We have the court 1
reporter with us and we usually don't transcribe our 2
letter-writing activities.
3 So, at this point, we can let you go for 4
the rest of this meeting. So thank you for your 5
services. And for the rest of the meeting today and 6
tomorrow, we will not require a transcription. Thank 7
you.
8 (Whereupon, at 11:47 a.m., the verbatim 9
reporting of the proceedings was concluded.)
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
© 2025 X Energy, LLC, all rights reserved 1
© 2025 X Energy LLC, all rights reserved 1
Xe-100 Licensing Topical Report Mechanistic Source Term Approach Milan Hanus July 9, 2025
© 2025 X Energy, LLC, all rights reserved 2
Department of Energy Acknowledgement and Disclaimer This material is based upon work supported by the Department of Energy under Award Number DE-NE0009040.
This presentation was prepared as an account of work sponsored by an agency of the United States Government.
Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
© 2025 X Energy, LLC, all rights reserved 3
X-energys Mechanistic Source Term (MST) Approach Mechanistic Source Term Models Q&A Outline
© 2025 X Energy, LLC, all rights reserved 4
Mechanistic Source Term (MST) Approach
- Scope:
Included: Description of MST models used to determine radionuclide transport phenomena and estimate mechanistic source terms used to support the preliminary analysis and evaluation of the Xe-100 Excluded: Actual implementation evaluation cases and outputs (included in a future XSTERM Topical Report)
- Interfacing Documents:
Risk-Informed Performance-Based Licensing Basis Approach for the Xe-100 Reactor (ML21196A071)
Xe-100 Licensing Topical Report: Transient and Safety Analysis Methodology (ML25077A285)
Xe-100 Principal Design Criteria Licensing Topical Report (ML24047A310)
Xe-100 Licensing Topical Report TRISO-X Pebble Fuel Qualification Methodology (ML22216A179)
Xe-100 Licensing Topical Report Atmospheric Dispersion and Dose Calculation Methodology (ML23268A456)
© 2025 X Energy, LLC, all rights reserved 5
Mechanistic Source Term (MST) Approach
- Xe-100 MST methodology is part of the implementation of a risk-informed, performance-based design and licensing basis according to the Nuclear Energy Institute (NEI) 18-04 and Nuclear Regulatory Commission (NRC) Regulatory Guide (RG) 1.233
- Regulatory Guidance (SECY-93-092):
Reactor and fuel performance under normal and off-normal operating conditions is sufficiently well understood to permit a mechanistic analysis.
Sufficient data should exist on the reactor and fuel performance through the research, development, and testing programs to provide the adequate confidence in the mechanistic approach.
Transport of fission products can be adequately modelled for all barriers and pathways to the environs, including specific consideration of containment design. The calculations should be as realistic as possible so that the values and limitations of any mechanism or barrier are not obscured.
Events considered in the analyses to develop the set of source terms for each design are selected to bound severe accidents and design-dependent uncertainties.
© 2025 X Energy, LLC, all rights reserved 6
Mechanistic Source Term (MST) Approach Xe-100 source terms are:
- Event-specific
- Determined mechanistically using models of fission product (FP) generation and transport
- Accounting for the reactors inherent and passive design features and the performance of FP release barriers that constitute the functional containment
- Inclusive of the quantities, timing, physical and chemical forms, and thermal energy of the release
- Different from light water reactor source terms based on severe core damage event(s)
© 2025 X Energy, LLC, all rights reserved 7
Radionuclide Release Barriers Mechanistically Modelled Radionuclide Transport Phenomena Mechanistic Source Term (MST) Approach
© 2025 X Energy, LLC, all rights reserved 8
MST Models LTR Appendix Model Scope Codes implementing similar capability A
FPM Fuel Performance (particle failure fractions)
PARFUME (INL), PANAMA, STACY B
THM Thermodynamics VSOP-THERMIX-KONVEK, AGREE (UMich), STAR-CCM+ (Siemens)
C SOLM Time-dependent radionuclide production, transport & release from fuel elements PARFUME (INL), GETTER, FRESCO, STACY D
GASM Steady-state gaseous fission product release PARFUME (INL), NOBLEG, STACY E
DUSTM Graphite / metallic dust production N/A F
HPBM Dust, fission and activation product behavior in primary circuit DAMD (PBMR), PADLOC (GA), RADAX, SPATRA, RADC (GA), MELCOR (Sandia)
G CORRM Air/water Ingress, fuel materials corrosion rates OXIDE-4 (GA), TINTE, GRACE (FE), STAR-CCM+ (Siemens), Fluent (ANSYS)
KSIM Plant simulator using point-kinetics core model with spatial and thermo-dynamics coupling MGT / TINTE, AGREE (UMich), RELAP-7 (INL), Flownex (M-Tech)
TRITM Tritium plant mass balance TRITGO (GA), THYTAN (JAEA), TPAC (INL), TMAP (INL), ORIGEN-S (ORNL)
Dispersion/Dose Off-Site doses MACCS (Sandia), HotSpot (LLNL), RASCAL (NRC), ADDAM (AECL)
US-DOE Code Other Commercial NQA-1 Code German (Legacy) Code XSTERM Xe-100 Licensing Topical Report, Atmospheric Dispersion and Dose Calculation Methodology (ML23268A456)
© 2025 X Energy, LLC, all rights reserved 9
XSTERM Code Evaluation Model for the quantification of Xe-100 source terms and dose calculations Implementation of the MST models presented in this TR, part of X-energy safety analysis code suite Developed under X-energy Quality Assurance Program Goal: NQA-1 qualification Verification and Validation, Uncertainty Quantification in progress Line-by-line code verification, comparison with analytical solutions, automated testing in the continuous-integration framework Phased validation - each phase aims at validating a set of medium and high ranked phenomena from the Xe-100 Safety Analysis PIRT Phase 1: Activity Release and Transport (FPM, GASM, SOLM, HPBM)
Phase 2: Reactor Temperature and Power (TDM, KSIM)
Phase 3: Dust Production (DUSTM)
Phase 4: Exposure to Oxidating Environments (CORRM, TRITM)
© 2025 X Energy, LLC, all rights reserved 10 Core MST Models Relationship
© 2025 X Energy, LLC, all rights reserved 11 THM: Thermodynamics Model
© 2025 X Energy, LLC, all rights reserved 12 THM: Application Domain Xe-100 Single-Pebble / Core Calculations Fuel Compact Calculations (Validation)
© 2025 X Energy, LLC, all rights reserved 13 THM: Phenomena Modelled Conductive Heat Transfer: Fourier law of heat conduction with space-, temperature-and dose-dependent conduction coefficients Zehner-Bauer-Schlünder model in the pebble bed
= + +
Reflector graphite model based on research of G. Haag Convective Heat Transfer: Kugeler-Schulten correlation Radiative Heat Transfer: Stefan-Boltzmann law Heat Sources: fission, gamma, decay heat (DIN-25485 standard)
Pebble temperatures: 1D-radial heat conduction through pebble mesh zones with core node temperature as boundary condition - semi-analytic Particle temperatures: 1D-radial heat conduction through particle mesh zones with pebble zone temperature as boundary condition - semi-analytic Compact temperatures: 2D-axisymmetric heat conduction through compact zones with prescribed outer temperature - finite-difference discretization +
successive-over-relaxation iterative method
© 2025 X Energy, LLC, all rights reserved 14 THM: Material Properties Explicit correlations or data fits Generally temperature-dependent => iterations
© 2025 X Energy, LLC, all rights reserved 15 THM: Input Data and Boundary Conditions Second order least-squares mapping from VSOP grid to simplify heat transfer calculations Xe-100 Reactor VSOP Model MST Model Total core mass flow rate (Flownex)
Disclaimer: Values shown above only for illustrative purposes
© 2025 X Energy, LLC, all rights reserved 16 THM: Time-Stepping Algorithm
- Backward Implicit/Explicit Iterative Method
- Iterative calculation to converge reactor and coolant temperatures in each time-step
- Transient simulations
- Establishing steady state
© 2025 X Energy, LLC, all rights reserved 17 FPM (Particle Failure Probability Model)
© 2025 X Energy, LLC, all rights reserved 18 FPM: Phenomena Modelled Pressure vessel failure of TRISO particles effects of pyro-carbon (PyC) irradiation-induced creep on the effective stress of the silicon-carbide (SiC) layer Irradiation-Induced dimensional change of PyC layer Fission gas pressure Kernel irradiation swelling Kernel thermal expansion Kernel migration (Amoeba)
Fission product corrosion SiC thermal decomposition Manufacturing defects Exposed kernel (i.e., defect of all coating layers)
SiC defect (i.e., defect of the SiC layer with at least one other coating layer intact)
Inner PyC layer defect Dispersed heavy metal fraction
© 2025 X Energy, LLC, all rights reserved 19 SOLM (Fission Product Transport Model)
© 2025 X Energy, LLC, all rights reserved 20 SOLM: Phenomena Modelled
- Fission product production by direct fission in kernels, recoil from kernels to the buffer layer, decay and activation
- Fission product removal by means of decay and activation
- Transport and release of fission products from particles and fuel elements by means of diffusion
- Effects on isotope transport and retention from as-manufactured particle defects, contamination and particle failures that may occur during operation
© 2025 X Energy, LLC, all rights reserved 21 SOLM: Modelling Scope
- Calculate isotope concentrations in fuel elements and particles
- for all core meshes, following the multi-pass re-loading scheme to reach desired burnup to create an isotope library with the concentration distributions (initial condition for steady-state inventory and transient calculations)
- for a single sphere/compact to model irradiation and safety (annealing) experiments (validation) and isolated fuel element calculations
- Calculate the release over birth (R/B) ratios for steady-state irradiations and fractional releases for transients
© 2025 X Energy, LLC, all rights reserved 22 SOLM: Methodology
- Time-dependent heat conduction equation with temperature-dependent effective diffusion coefficients
- 1D radial (pebbles) or 2D axi-symmetric (compacts)
- Boundary conditions: Zero surface concentration or sorption transfer via iso-thermic exchange between pebble surface and thin gas layer
= 11
+ 22
© 2025 X Energy, LLC, all rights reserved 23 SOLM: Numerical Solution
- Implicit Euler method for time-stepping for pebbles and compacts, or backward-difference method (BDF2) for compacts
- At each time step, a system of linear algebraic equations arising from the finite difference discretization of the diffusion and source terms is solved:
3-diagonal matrix system for pebble geometries
- solved by Gaussian elimination 5-diagonal matrix system for compact geometries
- solved by the Gauss-Seidel iterations
© 2025 X Energy, LLC, all rights reserved 24 GASM (Gaseous Fission Product Transport Model)
© 2025 X Energy, LLC, all rights reserved 25 GASM: Phenomena Modelled Steady-state gaseous fission product (FP) release from particles and pebbles (R/B ratios)
Short half-lives of the gas isotopes => transport from the fuel kernel through intact coatings can be neglected FP sources:
heavy metal (HM) contamination of matrix/outer PyC layers of particles failed particles Two models dynamically switching based on fuel temperature: Rllig Model and Richards Model Failed Particles HM Contamination
© 2025 X Energy, LLC, all rights reserved 26 DUSTM: Dust Production
© 2025 X Energy, LLC, all rights reserved 27 DUSTM: Phenomena Modelled
- Graphite and metallic dust production from pebble-pebble and pebble-reflector abrasion pebble abrasion during its transport through the fuel handling system (FHS) piping control rod abrasion during its movements over the operating period
- Dust particle size spectrum lumped into bins, based on the historical measurements from the German pebble-bed reactor AVR
© 2025 X Energy, LLC, all rights reserved 28 DUSTM: Methodology In-Core dust production Wear from friction based on the theory of asperity contact Total dust production proportional to geometry-dependent and frictional force Frictional force proportional to temperature-dependent friction coefficient and height-dependent load pressure on the pebbles Load pressure on the pebbles computed by modified Janssens silo pressure formula including the effects of pressure drop Dust production in fuel handling system (FHS)
Proportional to empirically determined dust generation rate per meter of pebble movement in the FHS, number of fuel passes and length of the FHS pipe Dust production in the reactivity control system (RCS)
Proportional to empirically determined dust generation rate per meter of RCS rod motion and the total RCS rod distance travelled during the operation time Dust production rate parameter determined by applying the model on the AVR reactor operation data and adjusting the parameter to yield the measured total lifetime dust in AVR Graphite/metallic dust ratio obtained from the Vampyr II experiment data dust production rate parameter dust production rate parameter
© 2025 X Energy, LLC, all rights reserved 29 HPBM: Helium Pressure Boundary Model
© 2025 X Energy, LLC, all rights reserved 30 HPBM: Phenomena Modelled radionuclide (RN) release from pebbles, deposition on component surfaces (RN, dust),
plate-out on dust (RN) re-entrainment into circulating He (RN, dust),
intra-and inter-component transport (RN, dust),
RN transmutation through activation and radioactive decay RN removal through radioactive decay RN sorption into the graphite dust and de-sorption into circulating helium
© 2025 X Energy, LLC, all rights reserved 31 HPBM: Phenomena Modelled DUSTM SOLM
© 2025 X Energy, LLC, all rights reserved 32 HPBM: Particle Mass Transfer (1)
Liftoff of elemental radionuclide from plated-out dust (2)
Liftoff of plated-out dust from component surface (3)
Sorption of elemental radionuclide from plated-out dust into helium (4)
Plateout of entrained dust onto component surface (5)
Plateout of entrained elemental radionuclide onto plated-out dust (6)
Sorption of entrained elemental radionuclide into plated-out dust (7)
Liftoff of elemental radionuclide from component surface (8)
Plateout of elemental radionuclide onto component surface
© 2025 X Energy, LLC, all rights reserved 33 HPBM: Numerical Solution Coupled set of 2D partial differential equations for multi-phase flow and mass balance Phases:
- 1. Helium gas
- 2. Circulating dust
- 3. Deposited dust
- 4. Circulating isotopes
- 5. Deposited isotopes Finite volume + Backward Implicit-Explicit Method (IMEX) discretization
© 2025 X Energy, LLC, all rights reserved 34 CORRM: Core Corrosion Model
© 2025 X Energy, LLC, all rights reserved 35 CORRM: Methodology Corrosion Rate = (Kinetic Factor)(Driving Force)
(Adsorption Term) with temperature-dependency of kinetic factor and adsorption terms modelled by Arrhenius-type correlation Applied to fuel element graphite materials to determine the radionuclide release enhancement due to corrosion Based on correlation for H-451 fuel element graphite investigated at General Atomics Material State Atmosphere UCO kernel Normal Trace H2O DLOFC Air/He PyC coating H2O ingress H2O/He US A3-3 matrix Normal Trace H2O DLOFC Air/He
© 2025 X Energy, LLC, all rights reserved 36 Overall Mechanistic Source Term Calculation
© 2025 X Energy, LLC, all rights reserved 37 Phone: 301.358.5600 530 Gaither Rd., Rockville, MD 20850 x-energy.com @xenergynuclear X Energy, LLC Milan Hanus mhanus@x-energy.com Q & A
© 2025 X Energy, LLC, all rights reserved 38 List of Acronyms CORRM Corrosion model PyC Pyrolytic-carbon DLOFC Depressurized loss of forced cooling Q&A Questions & answers DOE Department of Energy R/B Release to birth ratio DUSTM Dust production model RCS Reactivity control system FHS Fuel handling system RG Regulatory Guide FP Fission product RN Radionuclide GASM Gaseous FP release model SiC Silicon-Carbide HM Heavy metal SOLM RN diffusion and release model HPBM Helium pressure boundary model THM Thermo-hydraulics model HTGR High Temperature Gas-cooled Reactor TR Topical Report KSIM Neutron kinetics & plant simulation model TRISO Tristructural-Isotropic MST Mechanistic source term TRITM Tritium release model NQA Nuclear Quality Assurance UCO Uranium Oxycarbide PIRT Phenomena identification and ranking table XSTERM X-energys mechanistic source term code suite
NRC Staff Review of Topical Report (TR) 000632 Xe-100 Licensing Topical Report Mechanistic Source Term Approach (MST)
ACRS Full Committee Meeting July 9, 2025 https://www.nrc.gov/reactors/new-reactors/advanced.html
=
Background===
- MST TR originally submitted on May 10, 2024 (ML24131A146) with updated submittal on March 14, 2025 (ML25073A093)
- Updated submittal includes updates identified during regulatory audit
- Updates to MST TR sections 1.5 and 7.1 clarify that applicability is limited to preliminary analyses
- Correction of typos
- Updates to Appendix H showing MST model interfaces 2
=
Background===
3
Regulatory Basis Title 10 of the Code of Federal Regulations (10 CFR) 50.34(a)(1)(ii)(D) requires, in part, that an applicant for a construction permit (CP) perform an evaluation and analysis of a postulated fission product release to evaluate the offsite radiological consequences.
Under 10 CFR 50.34(a)(4) an applicant for a CP must perform a preliminary analysis and evaluation of the design and performance of structures, systems, and components with the objective of assessing the risk to public health and safety resulting from the operation of the facility and including the determination of margin of safety during normal operations and transient conditions anticipated during the life of the facility.
- Staff identified relevant Principal Design Criteria (PDC): Xe-100 PDC 10, RFDC 16, PDC 19 Under 10 CFR 50.34(a)(8) an applicant for a CP must identify the systems, structures or components of the facility, if any, which require research and development to confirm the adequacy of their design and describe the research program that will be conducted to resolve any safety questions associated with such systems, structures, or components. Such research and development may include obtaining sufficient data regarding the safety features of the design to assess the analytical tools used for safety analysis in accordance with 10 CFR 50.43(e)(1)(iii).
4
Scope of NRC Staff Review
- MST TR section 4.2 describes that MST models, implemented in the XSTERM code, are used to calculate dose consequences for licensing basis events, including the deterministic evaluation of design basis accidents
- NRC staff reviewed MST modeling approach to address radionuclide transport phenomena to support preliminary analysis of the Xe-100
- Review is limited to and focused on high-level physical phenomena of interest and whether the analysis approach and methods can reasonably support future licensing actions
- Design is preliminary
- Development and assessment of methods are in progress or planned
- Evaluation of models within XSTERM for acceptability will be conducted during the review of an application that relies on the results of XSTERM evaluations 5
Barriers to Radionuclide Release Fuel
- Fuel particle kernel (Uranium Oxycarbide (UCO)) within the TRISO fuel particles
- Silicon Carbide and Pyrolytic Carbon coatings applied to the fuel kernel
- Fuel matrix and fuel free zone of the fuel pebble Helium Pressure Boundary Reactor Building (Not credited) 6
XSTERM Models
- Thermodynamics Calculation Model (THM)
- NRC staff identifies this model to be of high importance because radionuclide release is expected to be diffusion dominant (temperature-dependent).
- Use of THM for analyses supporting a Xe-100 licensing application requires justification by the applicant.
- Point Kinetics Core Simulation Model (KSIM)
- The description of point kinetics appears to be different than standard point kinetics approaches (i.e.,
0D, single eigenvalue, lack diffusion coupling). Use of KSIM for analyses supporting a Xe-100 CP application requires justification by the applicant.
- Tritium Production and Transport Model (TRITM)
7
XSTERM Models MST TR describes nine models in XSTERM (cont):
- TRISO Particle Failure Probability Model (FPM)
- Solids Product Transport Calculations Model (SOLM)
- Steady-State Gaseous Fission Products Transport Calculations Model (GASM)
- Dust Production Rate Calculations Model (DUSTM)
- Helium Pressure Boundary Model (HPBM)
- Core Corrosion Model (CORRM)
NRC staff determined that FPM, SOLM, GASM, DUSTM, HPBM, and CORRM address phenomena needed to predict MST to support preliminary analysis:
- Models rely on previous modeling and operational experience from gas-cooled reactors such as Arbeitsgemeinschaft Versuchs Reaktor (AVR)
- Based on the NRC staffs experience with light water reactor (LWR) and non-LWR source term analysis, the NRC staff did not identify significant gaps in the MST models.
- MST TR section 4.2 states that the source term modeling described may be revised
- NRC make no conclusions regarding the acceptability of these models 8
Assessment Plans (Verification and Validation (V&V))
1.
V&V effort is underway to ensure that XSTERM is qualified to support final safety analyses 2.
Validation plans are developed to cover high and medium ranked phenomena that are identified through a Phenomena Identification and Ranking Table (PIRT) process 3.
The phenomena modeled by XSTERM were extracted from an earlier version of the PIRT
- NRC staff determined that the assessment process is acceptable because the identification of code assessment requirements through the PIRT process is an established approach (see RG 1.203)
- NRC staff are unable to assess the adequacy of the V&V plan:
- Validation plan is not based on the latest PIRT information
- MST TR does not contain information describing the knowledge level of the phenomena identified in the PIRT
- The plan is preliminary and subject to change 9
Conclusions The NRC staff concludes that X-energys TR 000632, Xe-100 Licensing Topical Report Mechanistic Source Term Approach, Revision 3, provides a reasonable plan for the development of the MST methodology.
- The FPM, SOLM, GASM, DUSTM, HPBM, CORRM models in XSTERM appear to cover the phenomena needed to predict the MST to support the preliminary analysis and evaluation of the Xe-100 design
- The TR describes an acceptable approach to V&V NRC staff make no conclusions regarding the acceptability of the models in XSTERM for the MST analyses of the Xe-100 because:
- Models within XSTERM are still under development
- A detailed technical review of the individual models was not completed
- Details regarding key phenomena identification and associated knowledge levels are not provided in MST TR
- The models and associated validation plans are preliminary and subject to change The NRC staff expects that a detailed technical review of XSTERM model applicability to the Xe-100 reactor will be addressed as part of the review of a licensing application that references MST TR.
10
11 Acronyms AVR Arbeitsgemeinschaft Versuchs Reaktor CORRM Core Corrosion Model CP Construction Permit DUSTM Dust Production Rate Calculations Model FPM Failure Probability Model GASM Steady-State Gaseous Fission Products Transport Calculations Model HPBM Helium Pressure Boundary Model KSIM Point Kinetics Core Simulation Model LWR Light Water Reactor MST Mechanistic Source Term PDC Principal Design Criteria PIRT Phenomena Identification and Ranking Table SOLM Solids Product Transport Calculations Model THM Thermodynamics Calculation Model TR Topical Report TRISO Triple Coated Isotropic Particle TRITM Tritium Production and Transport Model UCO Uranium Oxycarbide V&V Verification and Validation