ML22216A255
| ML22216A255 | |
| Person / Time | |
|---|---|
| Issue date: | 07/22/2022 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| Wang, W., ACRS | |
| References | |
| NRC-2044 | |
| Download: ML22216A255 (142) | |
Text
Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Reactor Safeguards Radiation Protection and Nuclear Materials Docket Number:
(n/a)
Location:
teleconference Date:
Friday, July 22, 2022 Work Order No.:
NRC-2044 Pages 1-76 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
(ACRS) 5
+ + + + +
6 SUBCOMMITTEE ON RADIATION PROTECTION AND 7
NUCLEAR MATERIALS 8
+ + + + +
9 FRIDAY 10 JULY 22, 2022 11
+ + + + +
12 The Subcommittee met via Teleconference, 13 at 9:30 a.m. EDT, Gregory H. Halnon, Chairman, 14 presiding.
15 COMMITTEE MEMBERS:
16 GREGORY H. HALNON, Chairman 17 RONALD G. BALLINGER, Member 18 VICKI M. BIER, Member 19 CHARLES H. BROWN, JR., Member 20 VESNA B. DIMITRIJEVIC, Member 21 JOSE MARCH-LEUBA, Chairman 22 DAVID A. PETTI, Member 23 JOY L. REMPE, Member 24 MATTHEW W. SUNSERI, Member 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
2 ACRS CONSULTANTS:
1 DENNIS BLEY 2
4 5
DESIGNATED FEDERAL OFFICIAL:
6 WEIDONG WANG 7
8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
3 TABLE OF CONTENTS 1
Page 2
1.
ACRS Chairman Introductory Remarks 3
3 2.
NRC Staff Introductory Remarks 5
4 3.
Overview of Calculating Atmospheric 7
5 Dispersion and How it is Used in the 6
Design of Site Buildings and Systems 7
4.
What is Used Presently for Calculation 30 8
of X/Q for Low Population Zones 9
5.
The Interrelationships of Reg 53 10 Guides 1.23, 1.194, and 1.141 11 6.
The Expanded Use of ARCON96 and 55 12 Limitations on Use 13 7.
Where to Expect this Being Used and 71 14 What Benefits Will Be Realized 15 8.
Public Comments N/A 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
4 P R O C E E D I N G S 1
9:30 a.m.
2 MEMBER HALNON: Good morning, everybody.
3 My name is Greg Halnon. I'm the Subcommittee Chair 4
for this meeting this morning. The meeting will now 5
come to order. This is a meeting of the Radiation 6
Protection and Nuclear Materials Subcommittee of the 7
Advisory Committee on Reactor Safeguards. ACRS 8
members in attendance are Charlie Brown, Davie Petti, 9
Jose March-Leuba, Joy Rempe, Matt Sunseri, Ron 10 Ballinger, Vesna Dimitrijevic, Vicki Bier; and then we 11 have consultants, Steve Schultz and Dennis Bley also 12 present. Weidong Wang of the ACRS staff is the 13 Designated Federal Official for this meeting.
14 Did I miss any of the ACRS or consultants?
15 Okay, good.
16 During today's meeting, the subcommittee 17 will hear a staff informational briefing on the 18 development of Reg Guide 1.249, Revision 0, use of 19 ARCON Methodology for Calculation of Accident-Related 20 Off-Site Atmospheric Dispersion Factors.
The 21 subcommittee will hear presentations by and hold 22 discussions with the NRC staff and other interested 23 persons regarding this matter.
24 The entire meeting will be open to public 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 attendance. This meeting is for an information 1
briefing, and no actions are expected from the ACRS 2
full committee. We asked for this briefing due to the 3
numerous regulatory activity surrounding source term 4
determination, the LPZ sizing, advanced technologies 5
being reviewed, and other code changes.
6 With the interplay of reg guides both in 7
revision, planned for development, and guidance for 8
the new Part 53, it is important for the ACRS to stay 9
aligned with the staff and the applicants on the use 10 of calculational methods, such as this.
11 The rules for participation in all ACRS 12 meetings, including today's, were announced in the 13 Federal Register on June 13th, 2019. The ACRS section 14 of the U.S. NRC public website provides our charter, 15 bylaws, agendas, letter reports, and full transcripts 16 of all full and subcommittee meetings, including 17 slides presented there. The meeting notice and agenda 18 for this meeting were posted there, as well.
19 We have received no written statements or 20 requests to make an oral statement from the public.
21 Today's meeting is being held over Microsoft Teams, 22 which includes a telephone bridge line allowing 23 participation of the public over their computer using 24 Teams or by phone. There will be an opportunity for 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 public comment, and we have set aside some time at the 1
conclusion of the prepared presentations and member 2
discussions for public comment.
3 The transcript of the meeting is being 4
kept, and it is requested that the speakers identify 5
themselves and speak with sufficient clarity and 6
volume so that they can be readily heard.
7 Additionally, participants should mute themselves when 8
not speaking. To mute and unmute on a phone, please 9
push star 6.
10 Do any of the members have any 11 introductory comments?
12 Okay. We will now proceed with the 13 meeting, and I'd like to start by calling on Kevin 14 Quinlan of the NRC staff for opening remarks and 15 presentation of the material. Kevin, it's all yours.
16 MR. QUINLAN: Thank you. And good 17 morning, everybody. I appreciate this opportunity to 18 present to the ACRS on this topic. I will be giving 19 the first portion of the presentation, but I just 20 wanted to let you know that Jason White is on the 21 phone and he will be, we will be co-presenting this, 22 so we will be swapping back and forth a couple of 23 times in the presentation. I'll be controlling the 24 slides from my computer, but Jason and I will each 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
7 giving the presentation.
1 So let me first introduce myself. I'll 2
give Jason the opportunity to introduce himself. So 3
as I said, good morning. My name is Kevin Quinlan.
4 I am a meteorologist in the External Hazards Branch in 5
the Office of Nuclear Reactor Regulation. I've been 6
with the agency since 2008 and have spent most of my 7
time reviewing new reactors, operating reactors, 8
license amendment requests, mostly in licensing space.
9 During the presentation, I will be in 10 presentation mode on my computer; so if you have any 11 questions, you know, feel free to interrupt me. I 12 likely will not be able to see if somebody raises 13 their hand to ask a question or if somebody puts 14 something in the chat. So if somebody could alert me 15 that there's a question, please do and Jason or myself 16 will do our best to answer it fully.
17 I will be using my pointer, which I have 18 changed to the red dot. So, hopefully, everybody can 19 see it clearly. Since this is not in person and I 20 can't talk with my hands, I will be talking with my 21 pointer at times.
22 So, Jason, if you'd like to introduce 23 yourself.
24 MR. WHITE: Yes, good morning. I'm Jason 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 White. I'm also a meteorologist in the External 1
Hazards Branch in NRR, and I've been with the agency 2
since 2009.
3 MR. QUINLAN: All right. Thanks, Jason.
4 So, first, as you can see, the title slide. This is 5
the draft Reg Guide 1.249, the Use of ARCON 6
Methodology for Calculation of Accident-Related Off-7 Site Atmospheric Dispersion Factors. I thought I'd 8
just start with a very brief purpose and need, if you 9
will.
10 Through the evaluation of two different 11 topical reports, the NRC staff had determined that the 12 ARCON computer code is acceptable for modeling 13 exclusion area boundary and low-population zone chi 14 over q values just to add strict dispersion values at 15 relatively short distances, as long as the methods 16 were consistent with the regulatory positions in our 17 existing guidance. The brief history on this is that 18 the ARCON computer code was, until recently, only used 19 to estimate the atmospheric dispersion from an on-site 20 source, such as a vent or a stack or a door, to the 21 control room intake or the technical support center.
22 So there was a bit of a change in the way that the 23 code was used. This draft reg guide aims to provide 24 clear guidance to the nuclear industry on an 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 acceptable method to use the NRC ARCON code or methods 1
to estimate off-site atmospheric concentrations to the 2
EAB and the LPZ.
3 So that's the kind of the impetus that 4
drove the creation of this draft reg guide. There was 5
a change in the way that the ARCON code was used from 6
a control room habitability and control room dose code 7
to using it for off-site does.
8 DR. BLEY: Kevin.
9 MR. QUINLAN: Yes.
10 DR. BLEY: This is Dennis Bley. As I read 11 it, it sounds as if there's been no changes, even with 12 the new version of the code, there's been no changes 13 to the models that are inside it. Were they always 14 available to use this way, or were there some changes 15 to the code to use it for things beyond the control 16 room kind of calculations?
17 MR. QUINLAN: Sure, no, excellent question 18 and perfect timing on kind of the setup. So you're 19 correct. There were no changes to, there were no 20 changes to the code itself. We are using the exact 21 same version of the code.
22 You will see in the reg guide sometimes 23 it's referred to as ARCON96 because that is what the 24 guidance calls it. We recently updated the code to 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 ARCON2. Now, that code, the update to ARCON2 did not 1
change the dispersion algorithms within the code. It 2
really just put a new graphical user interface on the 3
front, making it usable on newer versions of Windows.
4 We were having problems with the previous version and 5
it being used on modern operating systems.
6 So there were no changes to the code. And 7
I suppose, to partially answer your question, there 8
was always the opportunity that an applicant could use 9
this code for purposes outside of the control room 10 habitability, and that's exactly what happened. We 11 had an applicant decide that their exclusionary 12 boundary was fairly short, and we'll get into this in 13 the slides. But there on the order of, you know, 80 14 to a couple of hundred meters, rather than the, you 15 know, 500 meters to a 1,000 meters for the 16 exclusionary boundary that large light-water reactors 17 have.
18 So the smaller designs with their smaller 19 source term decided that they could have smaller site 20 boundaries. And so, therefore, they decided to use a 21 code that was more designed for those shorter 22 distances.
23 DR. BLEY: And being designed for the 24 shorter distances, my memory of this stuff is 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
11 it's able to take account for the impact of building 1
it and building wakes on dispersion. Beyond the site 2
boundary, you don't worry about that so much. Is that 3
correct, or is there something I'm missing here?
4 MR. QUINLAN: No, no, you're correct. The 5
ARCON code does a better job of capturing the effects 6
of buildings and building wake on plumes. So that's 7
why it was always used on-site, going from, you know, 8
one of the support buildings or the buildings on site 9
that may have a release, the turbine building or 10 something like that, to the control room intake.
11 DR. BLEY: Okay, thanks. That helps. I 12 think you said this, but the reason you want the reg 13 guide now is because, with some of the new designs 14 we're expecting, there's going to be more of the same 15 kind of calculation going on, I'd expect.
16 MR. QUINLAN: Correct. So we've now seen 17 it twice. Two different applicants have come in with 18 topical reports that the staff has reviewed. Jason 19 actually has, so he will be able to, he could probably 20 answer questions more specific to applications than I 21 can, but two separate applicants have come in with 22 topical reports to use ARCON for off-site chi over q 23 estimates. So the staff wanted to develop a reg guide 24 to avoid the potential that we're going to have, 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
12 know, say we get 20 applications in so we'd get 20 1
different takes on how to use ARCON for this purpose.
2 So we would like to provide an acceptable method to 3
the staff to try to cut down the number of unique 4
versions of this method. It would save us time, it 5
would save the industry time and money, and so, you 6
know, we thought a reg guide would be a useful tool to 7
get out ahead of where the industry is already going.
8 DR. BLEY: Okay. Thanks a lot. That 9
really helps put this in perspective for me.
10 MR. QUINLAN: Sure. I'm happy to keep on 11 talking. I talk quickly, but I talk a lot, so I'll 12 try to keep us on track here.
13 All right. Moving on. So the general 14 outline will be an overview of atmospheric dispersion, 15 so really not specific to this reg guide but more of 16 a what is atmospheric dispersion; how is it 17 calculated; how is it used in siting and design; the 18 second bullet there; how is our current guidance used; 19 how does our current guidance use chi over q values 20 for the EAB and the LPZ; and then a little bit of 21 discussion on the relationship with other atmospheric 22 transport and dispersion -- that's what ATD stands 23
- for, atmospheric transport and dispersion 24 regulatory guides, how it relates to this new version 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 of Reg Guide 1.249. Then Jason will be talking about 1
the expanded use of the ARCON code for this new 2
purpose, how the new reg guide will be used; and then 3
he'll switch back to me and I will walk you through an 4
example calculation of an acceptable method that the 5
staff sees for how this can be used. And I'll 6
apologize ahead of time and I will apologize later 7
that I will be talking about math on a Friday morning 8
as part of the example calculation, but it is, 9
unfortunately, unavoidable.
10 So moving on. Start with an overview of 11 atmospheric dispersion. So atmospheric transport and 12 dispersion modeling, you can see here it's the top of 13 the food chain here. And it encompasses atmospheric 14 transport or advection. Those words are used 15 interchangeably. Diffusion, so it's a little more of 16 a molecular diffusion that plays a very, very small 17 part. Removal, which can be done through either dry 18 or wet deposition that will not be talked about in 19 this presentation because ARCON does not consider 20 precipitation, nor does the previous code, the PAVAN 21 computer code, which has been historically used for 22 off-site dispersion that also does not consider rain-23 out or washout or precipitation in any way. And then 24 transformation, you know, once it's on the ground, 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 talking about, you know, shine.
1 DR. BLEY: Kevin.
2 MR. QUINLAN: Yes, sir.
3 DR. BLEY: ARCON only does lighter-than-4 air gasses, right? Or can you do both, heavier than 5
air and lighter than air?
6 MR. QUINLAN: So ARCON doesn't actually 7
consider what you are transporting at all. It is 8
strictly a measure of the dispersion in the atmosphere 9
and it only considers the wind speed, the wind 10 direction, and the atmospheric stability. And so it's 11 a measure of the dispersion in the air. That is then 12 used by the dose assessors to determine the resulting 13 dose at a given location. So the dose assessors would 14 basically multiply the chi over q value here, the 15 concentration factor, which is what comes out of 16 ARCON, by the source term. But ARCON itself doesn't, 17 there is no input for a given gas, either lighter or 18 heavier than air, one way or the other.
19 DR. BLEY: Okay. Thanks.
20 MR. QUINLAN: Sure.
21 MEMBER BIER: Another quick question.
22 This is Vicki Bier. So in uncertainty analysis for 23 transport and dispersion, there would, of course, be 24 uncertainty about what the weather conditions would 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
15 and you could run it for numerous different weather 1
conditions. For a given set of weather conditions, is 2
chi over q taken to be a constant and is it well 3
enough understood that the variation around that would 4
be very small?
5 MR. QUINLAN: It's an excellent question.
6 It's slightly outside of what ARCON does. ARCON is 7
not a probabilistic model. It is a deterministic 8
model in the sense that it uses actual recorded 9
meteorological observations for the calculations. So 10 it uses at least one year of data to characterize the 11 dispersion at a given site.
12 MEMBER BIER: Okay. But it does it by 13 kind of averaging or smoothing over all the different 14 weather conditions.
15 MR. QUINLAN: So it considers each hour of 16 data, 8,760 per year, as its own unique chi over q 17 value. And then you can rank them to -- we'll get to 18 it in a couple of slides, but you basically pull out 19 a very conservative value once you rank all of the 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br /> and the individual chi over q values for each of 21 those hours. We pull out either the 95th percent 22 value overall, so for the entire site, or a 99.5 23 percent value, which considers the top, I believe it's 24 top 44 hours5.092593e-4 days <br />0.0122 hours <br />7.275132e-5 weeks <br />1.6742e-5 months <br /> in any given year. So we pull out an 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 extremely conservative measure of the dispersion for 1
a site, and that's what gets used to calculate the 2
resulting dose.
3 MEMBER BIER: Okay. Thank you.
4 MEMBER PETTI: And, Kevin, I was going to 5
say that any sort of experimental uncertainty in those 6
measurements that are done are small relative to this, 7
you know, the way that you're using the data and going 8
out to 95-percent confidence.
9 MR. QUINLAN: Right. So I'm not sure I 10 would use the term 95-percent confidence in this 11 because, again, it's not probabilistic. It is a 12 ranking of the number, ranking of the hours and 13 pulling out a conservative value. And that gives us 14 the confidence that the resulting doses due to the chi 15 over q is used is likely to be the extreme case for 16 any given year.
17 MEMBER PETTI: Right. Okay.
18 MR. QUINLAN: All right. Moving on. So 19 this is cartoon of really all the different factors 20 that go into an atmospheric dispersion assessment. We 21 certainly discussed some of them just a moment ago.
22 This is depicting a stack. Normally, for ARCON runs 23 and PAVAN runs, we consider a ground-level release, 24 but this does a little bit better job of depicting 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 what could potentially go into the transport of 1
radionuclides released to the atmosphere. You have 2
your infection, which is driven by the wind, and 3
that's considered to be constant for each hour. You 4
have your molecular diffusion, which, again, plays a 5
very, very small part. Most of the transport is 6
through the wind. There's the rain-out and washout.
7 Now, again, this is not considered in PAVAN or ARCON, 8
but, certainly, the turbulent eddies. You know, this 9
is a large part of the strength of ARCON. It 10 considers turbulent eddies mainly through the building 11 wake effects, so how that affects the plume. And 12 then, certainly, dry deposition, wet deposition, but 13 ARCON does not consider those.
14 DR. BLEY: So when you indicated the 15 calculations are conservative condition, I guess we're 16 thinking, at least we were thinking, of effects on 17 control rooms such that rain-out would get it out of 18 the air, and so you're not considering that, so that 19 means that that mechanism for removal isn't there.
20 But if we're talking about shine dose to people who 21 are outside of the control room, then not including 22 rain-out seems to go in the other direction.
23 MR. QUINLAN: Yes, I certainly understand 24 your point. That may be a good question for me to run 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 by -- I am not a health physicist. I'm a 1
meteorologist by degree. So once you start getting 2
into doses from shine due to either ground shine or 3
sky shine due to wet deposition or dry deposition, 4
that starts to get a little bit outside of my comfort 5
zone. But I certainly do understand your point, and 6
I think it's a good one.
7 DR. BLEY: So the data input when you run 8
ARCON is primarily a wind data; is that right? Or are 9
there other parts of weather that are coming in?
10 MR. QUINLAN: Sure. So it's an hourly 11 data file, and it consists of wind direction, wind 12 speed, and a measure of the atmospheric stability.
13 DR. BLEY: Okay.
14 MR. QUINLAN: And I'll be getting much 15 more into how we calculate the atmospheric stability 16 because that is, you know, quite important to these 17 calculations.
18 DR. BLEY: If you can, follow up on that 19 one because if we're beginning to use this for, like, 20 dose at the boundaries, then not having the shine dose 21 from washout might really give us optimistic answers.
22 I'm curious about that.
23 MR. QUINLAN: Sure. So I might add now, 24 but it is certainly a good question to follow up, is 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
19 that PAVAN, which has been used since 1982 for off-1 site chi over q calculations to the EAB and the LPZ, 2
also doesn't consider shine in any way, shape, or 3
form. So this is in, use of ARCON is in very close 4
agreement with the assumptions that have been used for 5
40 years now for off-site doses.
6 DR. BLEY: That's interesting. Okay. I'm 7
more familiar with how this stuff --
8 MR. QUINLAN: I don't know if that makes 9
things better or worse.
10 DR. BLEY: -- and, there, they do account 11 for those pieces. Okay, go ahead.
12 MR. QUINLAN: Sure. This is really just 13 another way of looking at the exact same figure. So 14 we really start with the on-site measurements, which 15 are taken by the on-site meteorological measurements 16 program, a meteorological tower on-site. We require 17 one-year minimum from our guidance. We do typically 18 see two years or more for any application, but there's 19 a one-year minimum there. That feeds directly into 20 the transport of the wind, the advection that we 21 talked about; diffusion caused by atmospheric 22 turbulence, so those were the turbulent eddies that we 23 talked about; both thermal eddies and mechanical 24 eddies, and I'll be talking about those probably, 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
20 believe, on the next slide a little bit more.
1 DR. BLEY: So I'm sorry. Right there.
2 MR. QUINLAN: Yes.
3 DR. BLEY: Another input beyond the 4
weather must be the building, the structures --
5 MR. QUINLAN: Correct. And I will show 6
you the input, I'll show you the input screen in a 7
couple of slides from now, which will show you all of 8
the different site-specific inputs, in addition to the 9
weather data that goes into it. So, hopefully, that 10 will answer any questions you might have.
11 DR. SCHULTZ: Kevin, this is Steve 12 Schultz.
13 MR. QUINLAN: Sure.
14 DR. SCHULTZ: Do you have, are you going 15 to be talking more about the one-year minimum and the 16 various ways in which the applicant can provide 17 information related to their measurement, on-site 18 measurements, and so forth? It seemed in the 19 document, as well as in the comments from the public 20 and the resolution of it, that there was a lot of 21 discussion about how this is done and many, many 22 different types of options. It seems like this is one 23 area where the staff and the applicant will need to 24 have some real dialogue in order to gain approval 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
21 their use of the code.
1 MR. QUINLAN: Yes. So Jason will be 2
talking a bit more on his slides about the use for or 3
the use of, I suppose we could call it nontraditional 4
observations. And not to steal Jason's thunder, but 5
the reason that we included that in the regulatory 6
guide is that, historically, we have reviewed large 7
light-water reactors and they have large EAB and LPZs.
8 Every single one has an on-site meteorological tower, 9
and they, you know, if they're looking to expand upon 10 existing units, then they already have the existing 11 tower. If they're looking to build new units at a 12 greenfield or a brownfield site, they'll typically put 13 up a tower a couple of years ahead of time, start 14 collecting the data, and then submit that with their 15 application. It's always been very cut and dry.
16 Lately, we've been seeing a little bit 17 more of a push from the industry to use other data 18 observations, whether it be from the National Weather 19 Service or another source. For example, up at Idaho 20 National Lab, there was some talk about potentially 21 using the DOE, they have a network of weather and 22 observation towers at Idaho National Lab, so that 23 would be DOE data. We've had a site down in Tennessee 24 that wants to use Oakridge National Lab data. There 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 was just a site up in Wisconsin that used National 1
Weather Service data from a nearby airport.
2 So we're starting to see a little bit more 3
of a direction in the industry of wanting to use off-4 site data, if you will, and trying to think ahead for 5
where new small modular reactors or advanced reactors 6
may be deployed in more nontraditional locations.
7 There's been discussions about deploying these, you 8
know, advanced reactors up in Alaska.
9 So these sites don't necessarily have on-10 site met data. It may be very difficult to begin 11 collecting on-site met data. So we wanted to provide 12 some guidance to the industry on how to or what would 13 be an acceptable use of non-on-site data. I still 14 think the most straightforward path for the industry 15 and for any specific application is to use an on-site 16 met data system. It's on-site, so you know it's 17 applicable. It's under their control. We have a reg 18 guide, Reg Guide 1.23, that talks specifically about 19 what goes into an on-site system, so it is a bit of a 20 cookbook. You just follow the recipe, and it will 21 come out. But we thought it was important to include 22 nontraditional observations, considering where the 23 industry seems to want to go.
24 DR. SCHULTZ: Thank you. And I'll wait 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 for Jason's presentation, too. It's obviously a very 1
important feature associated with the overall 2
evaluation.
3 MR. QUINLAN: It is. And so we did spend 4
a good portion of this reg guide discussing it, mainly 5
because it is -- the use of ARCON is not new. It's 6
been used since the mid-90s. But the use of non-on-7 site data, or I guess I could say off-site data, is 8
new. So we wanted to spend time on that and make sure 9
that we got that right and an acceptable path for the 10 industry.
11 So the way that ARCON is being used is 12 new, but the code itself is not. But, you know, the 13 met data is a bit of a new feature.
14 DR. SCHULTZ: Thank you.
15 MR. QUINLAN: Sure.
16 DR. BLEY: And, Kevin, I'll get out of 17 your hair in a minute.
18 MR. QUINLAN: No, no, you're fine.
19 DR. BLEY: One last question from me on 20 this. In the next line, you have the stability, and 21 are we still using the kind of course Pascal 22 categories, or are they doing something more directly 23 measured now for stability?
24 MR. QUINLAN: I would ask you to hold 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
24 thought --
1 DR. BLEY: Okay.
2 MR. QUINLAN: -- for at least three or 3
four more slides, and I will answer it. If I don't 4
answer it, then please just let me know.
5 CHAIRMAN BROWN: Are you finished with 6
this slide? I wanted to ask you a question -- this is 7
Charlie Brown -- before you shift off.
8 MR. QUINLAN: No, go ahead, please.
9 CHAIRMAN BROWN: On your first one, on-10 site measurements, why is it only a one-year minimum?
11 I would think that you would want to see some, at 12 least some variation over a larger period of time in 13 terms of what the atmospheric conditions are. Is 14 there a basis for selecting one year only to do your 15 analysis?
16 MR. QUINLAN: So for certain applications, 17 we require different time periods. And I apologize I 18 don't have the regulatory guide up in front of me.
19 Regulatory Guide 1.23 specifies certain time periods 20 for certain types of applications. So for, like, COL 21 and early-site permit applications, there is a two-22 year minimum.
23 And so what sometimes we will do is, say 24 an applicant won't have two years' worth of data when 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 they submit, so we'll review the application and the 1
other portions while they collect their second year, 2
and then they'll supplement their application with 3
that second year of data. We've done that a few 4
times, and I was just asked that question by a new 5
applicant this week.
6 But other types of applications do require 7
one year minimum because it's not necessarily site-8 specific. So for things like a construction permit --
9 I'm sorry, not a construction permit, a design 10 certification where they don't have a specific site 11 picked out and maybe they're considering a variety of 12 sites that they could potentially deploy to, you know, 13 at that point, one year of data is probably fine, 14 especially if you're considering a number of different 15 sites and you're trying to prove the adequacy of the 16 kind of the proof of concept of the design itself.
17 We've seen plenty of designs, you know. The AP1000, 18 for example, could be sited in a number of different 19 sites. We wouldn't require multiple years of data for 20 every single site that they came in with. You know, 21 we would probably look for one year of data for each 22 site just to prove that it could be sited at a variety 23 of locations.
24 CHAIRMAN BROWN: So, I mean, it would seem 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 to me that we can get over a 40 or 50-year period, you 1
can get a considerable variation just based on general 2
changes that we observe. Nothing is really ever 3
constant. You can look at the normal weather we face.
4 We had big years and faulty years, and it varies over 5
a 10, 15, 20-year period. So that's why I have the 6
question. That's why I think that the dispersion 7
issue for the atmosphere would be of interest to make 8
sure you didn't miss -- kind of like earthquakes, you 9
know. The last big earthquake might have been 300 10 years ago, but you'd like to know at least it was that 11 long ago. So that's why I -- I'm not looking for 300 12 years; that's not the point.
13 MR. QUINLAN: That's good because, when it 14 comes to weather data, you're not going to get it.
15 No, I certainly understand your point, and I think 16 it's a good one. I just looked up the time periods 17 for a construction permit. It's a 12-month period.
18 Operating license, 24 months, so a two-year period.
19 Early site permit, COLs, are 24-month periods.
20 But one of the things that we -- that's 21 not really the last step, if you will. So the period 22 that is collected and submitted with an application 23 for analysis has to be within the last 10 years, so we 24 want to make sure that -- at a site, conditions 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 change, buildings go up, buildings come down, trees 1
grow all the time and they tend to surround the 2
weather station. You know, you're always worried 3
about upwind and downwind impacts on weather 4
observations, so we want to make sure that it's 5
relatively recent, so within ten years is what our 6
guidance says.
7 But then what we do is we go back and we 8
compare the data that is submitted for atmospheric 9
transport and dispersion modeling with a recent, we 10 call it a climatic period, so it's a 30-year data set 11 from a nearby airport or a nearby National Weather 12 Service, FAA, DOE, you know, a reliable quality-13 assured data set that goes back at least 30 years as 14 part of the application. And what we do is we make 15 sure that the observations that are collected on-site 16 and submitted for ATD modeling are representative of 17 the area and other longer-term data sets in that area.
18 DR. BLEY: And if they're not, how do you 19 handle that? Do you have the applicant run a case, 20 say out of that data set that's either very much lower 21 winds or very much higher winds, or what do you with 22 it if this is not, in a sense, representative of all 23 the years in that 30-year period?
24 MR. QUINLAN: Well, we certainly, 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 know, we would take a closer look at the data and 1
perhaps they would need to consider a more limiting 2
chi over q value, a more extreme version. So keep in 3
mind, this is one part of the dose assessment puzzle, 4
so there's the source term, there is the atmospheric 5
transport modeling, there's the receptor. And so you 6
put all those together to come up with a dose at a 7
location over a certain amount of time.
8 So we would certainly compare what they're 9
modeling against the regional data, but also compare 10 it against the dose requirements, you know, the dose 11 limits and say, all right, if it's a 1 rem with a 20 12 rem dose limit, you know, will it make a difference if 13 we use a more conservative assumption in the modeling 14 or, you know, raise the chi over q by some percentage 15 to account for the difference in the weather data.
16 I'll be honest. In the 14 years I've been 17 at the agency, it really hasn't happened. Typically, 18 the on-site data is high quality and is fairly 19 consistent, at least with wind direction, wind speeds.
20 So we do have a measure of confidence that the data 21 collected at the site is representative of the area, 22 but I think we would then need to probably add in some 23 extra conservatisms if it was not.
24 DR. BLEY: That makes sense, and that 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
29 of explains the spot you're in. You haven't seen it.
1 If you do see it, you'll have to figure out what to 2
do.
3 MR. QUINLAN: Much like some of the other 4
areas that we deal with, yes.
5 MEMBER HALNON: Okay. Kevin, we're 6
running behind. We've used up well over 30, almost 40 7
percent of our time, and we're only on slide six. So 8
we need to kind of continue to move on, although the 9
conversation is exactly what I hoped would happen. So 10 let's move on, and we'll pick up a lot of these 11 questions as we go.
12 MR. QUINLAN: Yes, I think we certainly 13 will. So this slide is really another way of 14 discussing what we've been discussing, but it gets a 15 little bit more into specifics of gaussian plume 16 modeling, which ARCON and PAVAN both are. They're 17 both gaussian plume models.
18 It has been historically and widely 19 accepted at the NRC for use in radiological 20 assessments mainly because they're fast, because 21 they're simple, and they only require a few input 22 parameters. And so PAVAN has been used since 1982, 23 ARCON has been used since 1996. They're quick and 24 they're easy, and they provide a very conservative 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
30 estimate of the atmospheric dispersion for use in 1
these assessments.
2 So some of the assumptions that go into 3
these models are that the emission rate is constant.
4 Again, I said earlier we're not considering a specific 5
gas or a specific affluent, but whatever it is it's 6
considered to be constant.
The horizontal 7
meteorological conditions. They're homogenous over 8
the entire space being considered, so this isn't 9
computational fluid dynamics. This is a street line 10 model. And we model one hour at a time, and the 11 average wind speed for that hour is used. There's 12 different ways of averaging, but I won't get into 13 that. We have reg guides that discuss how it should 14 be averaged.
15 We assume the wind direction is constant, 16 again, over that one hour. The atmospheric turbulence 17 or stability is also constant over that hour. There's 18 no wind shear in the horizontal or vertical. The 19 plume is infinite, and there's no plume history.
20 Again, each hour considered on its own merits.
21 Dispersion in the Y and Z directions of a 22 gaussian, it has a gaussian normal distribution. And 23 diffusion is negligible in the downwind again. We're, 24 you know, looking at more transport and dispersion 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 instead of the diffusion, molecular diffusion.
1 And this is a diagram of exactly what I 2
just talked about. And it's similar to the other one, 3
but I wanted to point to here is wind going, like I 4
said, towards the top right of your screen, there's a 5
plume center line. There is the X axis and the Y 6
axis, the gaussian distributions which will create the 7
entire plume. There is no plume rise calculated in 8
ARCON. An operator or whoever is running the code can 9
estimate the plume rise by just really raising the 10 height of the release. So you can kind of see how the 11 plume will spread out in a gaussian fashion both in 12 the X and Y directions.
13 DR. BLEY: A quick question here. If at 14 hour1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> five the real wind actually changed direction, 15 you just assume it's moving in the same direction it 16 was, do you apply different stability factors hour by 17 hour1.967593e-4 days <br />0.00472 hours <br />2.810847e-5 weeks <br />6.4685e-6 months <br />, or when you get a new one at hour five do you 18 recalculate the whole plume?
19 MR. QUINLAN: So every hour will consist 20 of a wind direction, a wind speed, and a measure of 21 the atmospheric stability. So each hour you calculate 22 a chi over q for that hour, and then they get ranked, 23 all those hours, all the 8,760 hours0.0088 days <br />0.211 hours <br />0.00126 weeks <br />2.8918e-4 months <br /> get ranked and we 24 pull out a very conservative 99.5 percent highest 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 value.
1 So you are correct. If at hour five, the 2
wind direction changes and the stability changes, as 3
happens all day long. You look out the window now and 4
it's breezy, it wasn't breezy four hours ago before 5
the sun rose. So every hour is considered on its own 6
and then is ranked on its own, and then we pull out 7
the most conservative values, and that's what we 8
assume is used during a design basis accident release.
9 DR. BLEY: Okay. Thanks.
10 MR. QUINLAN: Absolutely. That's a good 11 question, sir.
12 So getting a little bit more into the 13 atmospheric turbulence or stability, causes primarily 14 are, at least what we consider to be ground heating 15 and cooling, which is what I talked about. The sun 16 comes up, it heats up the ground. You start seeing 17 some breezes. That's why you get those nice afternoon 18 cumulus clouds. It's those turbulent eddies in the 19 atmosphere that are causing that.
20 It can also be done by the mechanical 21 turbulence. Now, this roughness here, in reality, 22 yes, that's due to trees, it's due to buildings, and 23 whatnot. For ARCON purposes, it's buildings. It's 24 the building wake effect of how air moves over 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
33 around a building and what that can do to a plume in 1
the atmospheric.
2 So we have a couple of different ways of 3
measuring the atmospheric turbulence. Indirect 4
measurements, which is the wind speed; solar radiation 5
or sky coverage during the day; net radiation or sky 6
cover at night. This is the old Pasquill-Gifford 7
method where if you don't have two measurement heights 8
for temperature to compare, you can try to do a rough 9
estimate of the stability class based on the sky cover 10 and the wind speeds. So that's what that's talking 11 about.
12 The one that we use in our guidance and 13 that ARCON is based on is the delta T method. It is 14 the delta T standing for the change in temperature.
15 So you'll have a meteorological sense, a thermometer 16 at, typically, 10 meters above the ground and 160 17 meters above the ground. You compare the height, you 18 calculate the difference between the two, and that 19 gives you the measure of how stable or unstable the 20 atmosphere is.
21 There are some other direct measurements 22 that you can make based on horizontal wind speed and 23 how the horizontal and vertical wind speed changes, 24 but our models and our guidance really focuses on 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
34 delta T. And I will get a little bit more into that 1
right here.
2 So this is temperature on the X axis, 3
height on the Y axis, and you can see, you know, as 4
everybody knows, typically, as you go up in the 5
atmosphere, the temperature decreases. As an ideal 6
gas law, as it goes up, the pressure decreases, the 7
partial, if you will, will expand and cool. So this 8
is a way that we look at the changing temperature of 9
the atmosphere. So as it arises, it cools. If you 10 have warm air at the surface and cold air up here, the 11 warm air will rise and that will start, it will go up, 12 and that means air has to come down, and you will 13 start a daytime circulation. Again, those afternoon 14 puffy fair-weather cumulus clouds are exactly this; or 15 if you've ever landed on a plane, you can see that 16 you're about to get down towards the boundary layer by 17 those clouds, and that's when it gets bumpy at 18 landing. But if you have warm air up here and cold 19 air down here, then there's nothing driving that 20 circulation, so the atmosphere is considered to be 21 very stable. Air is not going to be moving up and 22 down, you're not going to get the wind associated with 23 those eddies.
24 So the difference in, the temperature 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 difference between those two levels is what drives the 1
atmospheric stability. We classify them into 2
different categories, A, B, and C being very unstable.
3 So typical conditions, midday, strong sunshine, light 4
wind which will allow you to heat up the surface more.
5 You'll have a heat flux that moves upward, as I just 6
discussed, and you end up with considerable horizontal 7
and vertical turbulence. That's when you see plumes 8
really spreading out, and I'll have some pretty 9
pictures in about two or three slides to show you just 10 that.
11 Mutual stability right in the middle is 12 when it's windy. The atmosphere is very well mixed.
13 There is no temperature difference between the upper 14 and lower boundary layer in the lower atmosphere 15 because it's already mixed due to the wind, or if it's 16 cloudy then you're not going to get that solar 17 insulation, that solar heating.
Very stable 18 conditions happen at nighttime: clear air, light wind.
19 You don't get that vertical motion, so the atmosphere 20 is extremely stable, so if you do have a plume it's 21 not going to rise. It's going to stay near the 22 surface. And so our light wind conditions and G 23 stability, which is extremely stable, are really what 24 drives the highest chi over q values. So when I talk 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 about ranking the chi over q values across, almost all 1
of these get thrown out, to be honest. Our, roughly, 2
one meter per second at G stability, that's what is 3
going to drive the worst dispersion conditions at a 4
site and will lead to the highest doses, so that's 5
what's considered, for the most part, in a design 6
basis accident assessment.
7 This is another way of looking at it. So 8
unstable conditions, you get those turbulent eddies 9
with the plume going up and down and really spreading 10 out in the X and Y directions as it goes down wind.
11 And you can see here temperature, warmer at the 12 surface than it is aloft, so you get the turbulence.
13 The neutral, there's really not that much 14 difference between the surface and the atmosphere 15 right above it, so it spreads out in a nice, pretty 16 gaussian fashion right here, just what you would 17 expect. The extremely stable conditions, E, F, and G, 18 where the ground is cooler than the air aloft, and 19 that plume is not going to spread out much. It's 20 going to stay close to the surface. It's going to 21 stay in a straight line. Whichever wind direction is 22 blowing, that's where it's going to go. So if you 23 picture a ground level release instead of the kind of 24 stack that's depicted here, you can imagine 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
37 that's where you're going to calculate your worst 1
dispersion conditions and that's where you're going to 2
get your highest chi over q values.
3 DR. BLEY: Kevin, that picture looks 4
really familiar to me. There was an old, probably in 5
the early 60s, NRC manual on meteorology. I don't 6
know if it was a NUREG or a watch document. Is that 7
still relevant, or has the science changed a lot since 8
then?
9 MR. QUINLAN: Well, how we model them has 10 changed, but the science is exactly the same. I mean, 11 this is all fluid dynamics, just in the atmosphere.
12 So that would be the same. We still use Skew-T 13 diagrams to explain changes in temperature in the 14 atmosphere.
15 The way that things are calculated are 16 much more complex now, a little bit more advanced.
17 But these same assumptions still hold very true.
18 DR. BLEY: That old handbook has never 19 been updated, has it? Are you even familiar with it?
20 It was --
21 MR. QUINLAN: I have a copy of it myself; 22 I do.
23 DR. BLEY: Okay.
24 MR. QUINLAN: And this probably looks 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 familiar because I'm sure that this came from that, 1
yes.
2 DR. BLEY: That's what I thought. Go 3
ahead.
4 MR. QUINLAN: You're exactly right.
5 MEMBER BIER: I have another question on 6
that previous slide. It gets exactly at something 7
that I was thinking about.
8 MR. QUINLAN: Okay.
9 MEMBER BIER:
If we are modeling 10 dispersion like, I don't know, five miles downstream 11 off-site some place, then it seems really clear that 12 that top unstable condition will have lower dose 13 concentrations just in general, once you get that far 14 downstream. But if you're talking about either on-15 site or very close to on-site, I mean, you show this 16 variability up and down, but I assume there's also 17 variability laterally.
18 For instance, in your top picture, the 19 dose where it looks like the plume is hitting the 20 ground is higher in the unstable than in the neutral 21 and stable, right? Because the instability kind of 22 push the radiation into a pocket that wouldn't 23 otherwise have had high levels of radiation. So I 24 wonder if that's a kind of general issue 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
39 method for short dose predictions.
1 MR. QUINLAN: You are exactly correct, and 2
I'm finding a Section 2.3 of the draft reg guide is 3
called release characteristics, and it talks about 4
exactly that issue that you just raised. I think 5
you've captured it perfectly where, if you have an 6
elevated release, something from a stack, and you have 7
your, let's say, EAB boundary is right here, you're 8
missing the plume. You're not going to, it's going to 9
miss your source.
10 So that is something that needs to be 11 considered by the analysts, and we have a whole 12 section in the reg guide for Section 2.32 is elevated 13 or stack releases and it's how you account for that.
14 Now, our preferred method and our recommended method 15 is to consider a ground level release for exactly that 16 reason. You get a ground level release. Then as it 17 moves down wind under what is likely to be, you know, 18 the driving conditions or worst dispersion conditions, 19 it's going to hit along the ground and stay along the 20 ground.
21 MEMBER BIER: Yes, okay. I guess the 22 question that I have, though, is whether you get also 23 the same phenomenon laterally, like, as the plume goes 24 around a building or whatever, you get a high 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 concentration in an area that would have been 1
predicted under a pure gaussian method to have a low 2
concentration.
3 MR. QUINLAN: So ARCON does consider 4
building wake, and that's actually one of the 5
strengths of the code compared to PAVAN, which you can 6
put in a function of building wake but it's not as 7
advanced.
8 MEMBER BIER: But I guess the whole thing 9
is kind of leading me to wonder whether what you think 10 is conservative is always conservative in taking your 11 kind of extreme percentiles of chi over q. If it puts 12 you in that bottom stable condition, then it might not 13 actually be conservative for all cases.
14 MR. QUINLAN: So, yes, I'm not sure I 15 exactly understand the question. But, yes, building 16 wake certainly plays into it. If your receptor is on 17 the other side of the building and the plume has to go 18 around the building --
19 MEMBER BIER: Right, yes.
20 MR. QUINLAN: -- then you do get some of 21 that turbulence. It's going to mix out. It's going 22 to affect the plume and make it less dispersed.
23 MEMBER BIER: Yes, I should probably just 24 take a look at that chapter in the guidance.
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 MR. QUINLAN: Yes, absolutely.
1 MEMBER BIER: Thanks.
2 MR. QUINLAN: Of course. So some real 3
world examples of atmospheric dispersion have been 4
pulled. You know, there's, of course, plenty of 5
examples out there. Here's one. I believe it's 6
looking across the Delaware River at a fire at some 7
kind of a refinery. But an example of an unstable 8
plume. This looks to be at sunset when the atmosphere 9
is beginning to transition from unstable to a stable 10 on a clear day right now. So you still have a lot of 11 dispersion through the atmosphere, so it comes up and 12 it spreads out. You can kind of see how it's 13 meandering.
14 Down here, a stable plume, no wind. I 15 believe this is from England. But you can see on a 16 day when there's no wind that that hot plume just goes 17 straight up, it does not disperse until it hits some 18 level where it's able to disperse. Now, picture 19 having a light wind pushing that plume to the side and 20 you can kind of get an idea of how -- now, this is 21 from what looked to be an extremely tall stack, but 22 picture a ground level release with a light wind 23 pushing that plume in a single direction and you can 24 kind of get a sense for how trapped into the ground 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 that plume would become.
1 This was from an oil fire, I believe, in 2
Nebraska. And an example of fumigation, which is 3
discussed in a few of our reg guides, but really what 4
it is is the plume being pushed one way and it gets 5
trapped near the ground, and so all the plume near the 6
ground gets mixed together. And so you end up with, 7
you know, a pretty high concentration near the ground 8
overall. So instead of the plume moving up as it is 9
in the other examples, I believe this is right around 10 sunrise, so the atmosphere is still extremely stable 11 and that plume stays near the ground from a ground 12 level source. For some reason, that always helps me 13 picture it a little bit better.
14 And to completely beat a dead horse here, 15 these are the Pasquill-Gifford stability classes A 16 through G, G being extremely stable, and when there's 17 a temperature difference of more than four degrees 18 between the upper and lower measurements on the tower.
19 So you've got an extremely stable atmosphere leading 20 to higher concentrations.
21 Getting a little bit more into how we 22 actually collect the data, the typical on-site 23 meteorological measurement program, this comes out of 24 Regulatory Guide 1.23. The primary tower on each site 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 usually has, and, you know, each site could be a 1
little bit different, but typically you have a 60-2 meter observation level for wind speed direction and 3
temperature, a 10-meter level for wind speed direction 4
and temperature, the delta temperature. The delta T 5
that we use for stability is the measure between the 6
two. And then you've got some other stuff that you 7
measure, of course: dew point; solar radiation at some 8
sites; precipitation is there, as well. There's 9
usually a backup tower in case the primary tower fails 10 or is out of service for some reason, and the sampling 11 rates is usually once per second, but then that gets 12 averaged over eventually to one hour value for that 13 entire hour.
14 So how we actually use these codes in 15 siting a design, there are two NRC models that get 16 used for design basis accidents. There's the PAVAN, 17 which has been traditionally used for EAB and LPZ.
18 There's ARCON which has been used for control room 19 habitability analyses.
20 So for the PAVAN code, it really comes out 21 of the methods laid out in Regulatory Guide 1.145, 22 Atmospheric Dispersion Models for Potential Accident 23 Consequences Assessments. It is captured for the 24 staff in SRP234, Short Term Atmospheric Dispersion 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 Estimates for Accident Releases. And then some of 1
these supporting documents. There's the technical 2
basis and the user's guide for PAVAN.
3 And PAVAN is a gaussian plume model that's 4
used for ground level releases. It does not really 5
consider elevated releasers because, again, it is 6
meant to be quite conservative.
Diffusion 7
coefficients are modified to account for plume meander 8
under low wind speed conditions and building wakes, 9
although ARCON has a more advanced algorithm for 10 building wake effects. It takes the chi over q values 11 for various time average periods, so zero to two hours 12 is really what gets used for the design basis 13 accidents. And so what that means is that those are 14 going to be your worst hours of dispersion; and so if 15 you assume that there's a release for anywhere between 16 zero to two hours, these are the ones that you would 17 use. You can also assume an accident happens for zero 18 to eight hours, 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, three days, 26 days. Those 19 are the periods that are used in accident analyses.
20 And as you go out and you average out over a larger 21 number of chi over q values, your estimates will 22 decrease because you're averaging over a greater 23 number of chi over q values instead of your most 24 extreme.
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 One of the big difference between the 1
codes is that PAVAN uses a
joint frequency 2
distribution, so you're really getting rid of your 3
hourly data and you're concatenating into a big grid 4
that consists of the wind speed classifications and 5
calm winds; wind direction breaking into the 16 6
directions, so you're really removing the granularity 7
of specific wind directions and you're breaking it 8
down just to the compass directions; and then 9
stability class stays the same, still based on delta 10 T.
11 I don't think anything more needs to be 12 said about that.
13 A little bit more on PAVAN. So we talked 14 a couple of times about how you pull out specific chi 15 over q values to represent the zero to two-hour time 16 interval. There's the 0.5 percent maximum sector 17 value or, read differently, 99.5 percent; the two 18 terms are interchangeable. So that is for one 19 specific sector, so that's looking in one direction.
20 You can pull out that value, or you can compare it 21 against the five percent overall. So consider all 16 22 compass directions, what is the 95-percent highest chi 23 over q, how does it compare with one sector which has 24 a more extreme value but you're discounting 15 out 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
46 16 directions at that point. So whichever is higher 1
is then used as the zero to two-hour chi over q value.
2 Even though it is actually a one-hour value, you 3
conservatively assume that it's for two.
4 And this procedure is repeated in PAVAN 5
two times: once for the exclusionary boundary and once 6
for the low population zone. Typically, there will be 7
two rings around a site representing each of those 8
different boundaries.
9 A little bit more about the 0.5 percent 10 or, as used in the alternatively, the 99.5 percent.
11 I just talked about it, but just for each of the 16 12 direction sectors, so it's not an overall site one, 13 it's just for a specific direction. It's calculated 14 in each sector, and then those are placed in order 15 from the greatest to the smallest chi over q values, 16 and then you create a curve out of those values and 17 you pull out, you want to make sure that no other 18 point are plotted above that curve. And from this 19 upper curve, the chi over q value, which is equal or 20 exceeded 0.5 percent of the time, so based on a normal 21 year of data would be the 44th hour, 44th highest 22 hour2.546296e-4 days <br />0.00611 hours <br />3.637566e-5 weeks <br />8.371e-6 months <br />, is obtained. And that ends up being the chi 23 over q value for the 99.5 percent maximum sector.
24 The five percent is a little bit simpler 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 but follows the same type of ranking, same type of 1
approach. But instead of one sector, you're looking 2
at all of the sectors, so all of the wind directions.
3 And for this, you would pull out the 99.5 percent chi 4
over -- I'm sorry, the 95 percent or 5 percent, 5
depending on how you want to look at it, you'd rank 6
them, and you would then pull out that value, which 7
is, again, a typical year at a 438th highest value out 8
of that whole year.
9 Kind of an example of what that would 10 actually look like when you take the chi over q 11 values. The chi over q value is here on the Y axis 12 and the percent frequency with respect to time here on 13 the X axis. Now, this was just shown for two 14 different wind speeds, 1 meter per second versus 0.5 15 meters per second. And if you want to pull out the 16 ones that are circled are the 99.5 percent values for 17 a sector. For the 1 meter per second, the values are, 18 I don't have an exact value here, but they're down 19 here. And then for the half meter per second, they're 20 up here.
21 So the half meter per second, because it's 22 a lower wind speed but it still concentrates the plum 23 in a direction, and you end up with a bit of a higher 24 chi over q value when you extrapolate out. Not every 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 data set is actually going to calculate out that far.
1 So you can calculate for an entire year, 2
so to calculate for a low population zone you maybe 3
want longer values. So in order to get the zero to 4
two hours, you calculate this, but you can also have 5
an overall annual average chi over q, 8,760 hours0.0088 days <br />0.211 hours <br />0.00126 weeks <br />2.8918e-4 months <br />, and 6
you can calculate whatever time period you want in 7
between. There are set ones in our guidance: again, 8
zero to two hours; zero to eight hours; a 16-hour 9
period, a three-day, and then a 26-day period. And so 10 you basically do a linear interpretation between zero 11 to two hour and the annual average to get your chi 12 over q values.
13 I don't think I need to go too terribly 14 deep into the definitions, but, since this guidance is 15 focus and design basis accidents, we are concentrating 16 on the exclusionary boundary, the area surrounding the 17 reactor, and then the low population zone. The area 18 immediately surrounding the exclusionary, which can 19 contain residents but, typically, a small number of 20 residents. And I'm just showing an example of what 21 that can look like in actuality. So this was pulled 22 from an early permit application that ended up not 23 moving forward, but I thought it provided a good 24 example of what an exclusionary area boundary can look 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 like. And I thought it was actually even more 1
helpful, this example, because it was a little more 2
applicable to this guidance document that we're 3
discussing today. But each direction, each of the 4
compass directions, will have a distant out to the 5
EAB. Some sites just do a concentric ring. It's the 6
exact same distance to every area. This site happens 7
to have slightly different values for different 8
directions. And then the low population zone will 9
consist of a ring beyond that.
10 So this is assuming a large light-water 11 reactor placed at that site, so the value is a little 12 bit far. But for a small modular reactor or an 13 advanced reactor, maybe these are in 80 meters or, you 14 know, closer, like 300 meters, I think, was an example 15 that was actually used in real life. So you can see 16 that they're much closer in, so that's why they would 17 consider ARCON instead of PAVAN.
18 A little bit more about ARCON. Also, a 19 gaussian plume model. The diffusion coefficients 20 account for enhanced dispersion under low wind speed 21 conditions and the building wakes, which we discussed 22 at length already this morning. The big difference is 23 that this uses hourly values. Instead of the joint 24 frequency distribution, this uses hourly values 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
50 wind speed, wind direction, and atmospheric stability 1
class. Most other portions of it are similar in how 2
it's treated and how it's calculated. But you, again, 3
can combine estimate concentrations ranging from two 4
hours to 30 days.
5 Cumulative frequency distributions which 6
will become important later on in the presentation are 7
prepared from the chi over q values. And then ARCON 8
calculates the chi over q values that are exceeded no 9
more than five percent of the time for each averaging 10 period, so the 95 percent chi over q values. The 99.5 11 percent requires some additional work, which we will 12 get into.
13 Typically, ARCON, again, is used for 14 control room habitability. So this is an example of 15 the AP1000 and how the ARCON analysis was done for all 16 the sites that used this design. You would have two 17 receptors, one at the control room HVAC right here and 18 one at the annex building access. And then you have 19 eight different sources where you could have a 20 release, and so you would calculate the source to 21 receptor combinations for every single one. And so 22 you really -- AP1000 required doing 16 different runs 23 for ARCON, but you can calculate the direction and 24 then run the met data, and each one would be slightly 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 different because it's a slightly different wind 1
direction, so it uses different wind direction 2
windows. But I just want to show you an example of 3
how you go about setting up an ARCON run.
4 MEMBER HALNON: Hey, Kevin, those are all 5
ground level release sources?
6 MR. QUINLAN: They are considered ground 7
level releases, but each of these will have a unique 8
-- only because there are specific criteria for what 9
is an elevated release. But so each of them are run 10 as a ground level release, I think with the exception 11 of -- well, no, they're all considered point sources, 12 as well, with the exception of number eight here which 13 is the containment shell and that's used as a diffuse 14 area so kind of more of a cloud of radiation, if you 15 will.
16 But, yes, they're all considered ground 17 level, but you do put in, and I'll show you in a 18 couple of slides, the input form. Each of these is 19 going to have a height to it as a source and the 20 receptor will also have its unique height. This one, 21 I believe, is a door, so that would, more or less, 22 three feet, it's assumed.
23 MEMBER HALNON: So the elevation is 24 somewhat considered. Okay, I got it. Thanks.
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 MR. QUINLAN: Yes, absolutely. It is 1
absolutely considered in this.
2 The applicable guidance for ARCON is also 3
captured in SRP234. That's the same SRP section as 4
PAVAN because both of them are considering short-term 5
dispersed estimates for accident releases. The 6
guidance is laid out at Reg Guide 1.194, and I'll be 7
doing a comparison shortly between 1.194 and this 8
guidance, not in-depth but enough to get kind of a 9
flavor for it. And then there's the user's guide and 10 the NUREG which describes the model in much more 11 detail, NUREG/CR-6331.
12 So the relationship to the existing 13 guidance between this reg guide that we're considering 14 today. But, first, a comparison, a very high level 15 comparison, between Reg Guide 1.194 for ARCON and Reg 16 Guide 1.145 for PAVAN. The ARCON reg guide provides 17 the guidance on determining chi over q values for 18 control room habitability assessments, and Reg Guide 19 1.145 provides it for determining chi over q values at 20 the EAB and the LPZ.
21 So in comparing Reg Guide 1.194 versus the 22 current draft reg guide that we're discussing today, 23 they're quite similar, similar in almost every single 24 way except for how it's being used. So they both use 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 hourly met data. The model setup is exactly the same.
1 There's one source receptor pair, which is a distance 2
and direction per run. It can use ground level or 3
elevated release, and it can be modeled as either a 4
point source or a diffuse source. So a point source, 5
you know, picture something coming out of a pipe.
6 Diffuse source, picture more of a cloud.
7 The model results coming out that are 8
discussed in Reg Guide 1.194 are the typical ARCON 9
results which considers the 95th percentile chi over 10 q for each source receptor pair. The draft reg guide 11 will certainly do that, but it also includes a method 12 in its Appendix A of the reg guide for determining the 13 maximum sector 99.5 percent chi over q. That is not 14 a standard output of ARCON. It will take extra 15 calculation for a licensee or an applicant to 16 calculate.
17 So here is the ARCON input display, as 18 promised a couple of times. So here you would enter 19 your input, your direction to the source, so your 20 receptor to the source, I'm sorry, your source to 21 receptor direction. Here is just one degree, so it's 22 basically due north. Your receptor distance, how far 23 it would be. Your release height, so where is it 24 coming out of. For ARCON, typically, out of some 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 portion of the building and it's going into another 1
portion of a different building likely. So you 2
consider the two intake and release heights. A 3
building area, so containment building typically will 4
have, you know, it's a large building, you can use as 5
diffuse source. You could put in how big the building 6
is and any elevation differences. And the building 7
area is the driver between the building for the 8
building wake effect that ARCON considers. So we 9
talked about that a couple of times this morning, but 10 the building wake, that's really where this comes into 11 play.
12 A couple of additional inputs. And I 13 wanted to put this one up because the default is a 90-14 degree wind direction window, so whatever your wind 15 direction is from your source or receptor, say it's 16 due north, so zero degrees, this would consider a 90-17 degree window on either side of that source to 18 receptor window or direction and it would pull all of 19 the wind directions that fall into that 90-degree 20 window. And that is what is considered for the chi 21 over q value. If it's not blowing in that direction, 22 then you really don't care because it's not going to 23 be blowing towards your receptor.
24 So the default is a 90-degree window, 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
55 that comes from Reg Guide 1.194 for use in control and 1
dispersions. Reg Guide 1.249, the one we're 2
discussing today, recommends changing that to a 45-3 degree window because that is more consistent with the 4
PAVAN guidance in Reg Guide 1.145. So we tried to do 5
our best to keep the ARCON, how you set up the ARCON 6
run the same. But since we're not considering control 7
habitability, we're now considering off-site doses, we 8
wanted to make sure that we were more inline with the 9
off-site guidance that we've been using for a long 10 time now. So that's why we recommend changing this to 11 45 degrees. It narrows the window a little bit for 12 the off-site receptors.
13 And then a very high level comparison 14 between the PAVAN Reg Guide 1.145 and this reg guide, 15 so this should look very familiar from a few slides 16 ago. But PAVAN uses joint frequency distribution.
17 ARCON uses the hourly. The model setup is quite 18 different. You run all 16 directions at the same 19 time, and you can run multiple distances all at once 20 versus ARCON which uses one source receptor pair and 21 you really have to do them one at a time. The model 22 results are complicated, but it does calculate both of 23 these values, the 95 and the 99.5 percent chi over q 24 at the same time, versus ARCON which only calculates 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 the 95 percent and then you have to do a little bit of 1
extra work for the 99.5.
2 So at this point, I'm going to turn this 3
portion of the presentation over to Jason. I'll run 4
the slides, so, Jason, just let me know when you want 5
me to move the slide forward. But Jason will now take 6
you into the expanded use of the ARCON code and a 7
little bit more about the reg guide itself.
8 MR. WHITE: Yes. Thanks, Kevin. First, 9
I want to check with Greg. Greg, how are we doing on 10 time? Do we have a hard out? Because we have an 11 example at the end of the presentation that we wanted 12 to leave enough to discuss because it's kind of the 13 heart of what we're trying to accomplish.
14 MEMBER HALNON: Yes, I would, I want to go 15 to public comment at 11:15, so you have about, let's 16 just say 20 minutes of presentation/discussion.
17 MR. WHITE: Okay. Well, I'll try to move 18 through these slides quickly. Kevin did a great job 19 of covering all the basics of the atmospheric 20 dispersion analysis and these basics influence all the 21 aspects of what we are presenting today. A lot of 22 upcoming slides were touched upon earlier by Kevin, 23 but, as I go through them, if there's any additional 24 discussion, we can definitely delve into it again.
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 So in this section of the presentation, we 1
will discuss the expanded use of the ARCON code, which 2
is the basis for this draft Reg Guide 1.249.
3 MEMBER HALNON: Hey, Jason.
4 MR. WHITE: Yes.
5 MEMBER HALNON: Could you move the 6
microphone just a little further away from your mouth.
7 We're getting the puffing.
8 MR. WHITE: Oh, okay. I'm sorry.
9 MEMBER HALNON: Thanks. That sounds 10 better.
11 MR. WHITE: Okay. How is that? Is that 12 still loud enough but better?
13 MEMBER HALNON: Yes, that should be fine.
14 We can adjust.
15 MR. WHITE: Okay. Because I could always 16 adjust it more, if needed.
17 So the ARCON computer code was developed 18 to model shorter distances in the vicinity of 19 buildings typical of control room habitability 20 evaluations. The ARCON dispersion algorithms are 21 based on field measurements taken out to distances of 22 1200 meters. As stated before, large light-water 23 nuclear power plants typically have EAB and LPZ 24 distances that range from 800 to 6,000 meters, while 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 small modular and advanced reactor designs are 1
expected to include EAB and LPZ distances in the range 2
of 80 to 400 meters.
3 Next slide. In this accident source term 4
topical report, NuScale proposed using an ARCON 5
computer code methodology for calculating off-site 6
atmospheric dispersion values at the EAB and LPZ, 7
rather than using the PAVAN computer code. This 8
methodology deviated from NRC guidance. NuScale 9
provided a justification for this use of this 10 methodology and detailed why it was appropriate for 11 its design.
12 During this review for the topical report, 13 the NRC staff conducted an in-person audit of the 14 topical report and the methodologies. The staff 15 reviewed the documentation for how the methodology 16 would be implemented and the staff also performed an 17 independent analysis using the methodology.
18 Next slide. The NRC staff made the 19 determination that the ARCON computer codes acceptable 20 for modeling EAB and LPZ chi over q values at 21 relatively short distances, as long as the methods are 22 consistent with the regulatory positions of Reg Guide 23 1.145 for off-site chi over q values. Therefore, the 24 methodology outlined in this draft Reg Guide 1.249 is 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
59 applicable to the EAB and LPZ distances from source 1
locations out to a distance of 1200 meters.
2 Next slide. In the next section of the 3
presentation, we will discuss how the Reg Guide 1.249 4
will be used.
5 As was stated before earlier in the 6
presentation, this draft reg guide is applicable to 7
sites with EAB and LPZ distances out to 1200 meters.
8 However, certain locations may be affected by 9
transport and diffusion conditions that may be more 10 restrictive than assumed in the contiguous 48 states.
11 For these locations, which may be characterized by 12 extreme and persistent restrictive dispersion 13 conditions, such as in Alaska, the applicability of 14 dispersion algorithms in ARCON may not apply or may 15 require further modification.
16 Next slide. Some of the issues that may 17 arise with the atmospheric dispersion analysis in 18 these locations include different characteristics of 19 accident releases, such as buoyancy and momentum, 20 depending on ambient conditions at the time of the 21 release. Also, transport and diffusion conditions 22 possibly being significantly different or more 23 restrictive than assumed. Seasonal variation of 24 dispersion and meteorological conditions such 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
60 separate modeling approaches and models may be 1
necessary at certain times of the year. These all 2
could possibly lead to the need for field studies to 3
characterize and model the dispersion conditions.
4 Next slide. Kevin went through the 5
rationale on this topic earlier, but, due to the 6
potential of the new or advanced reactor designs being 7
considered in nontraditional locations, there may be 8
a need to seek alternative meteorological sources at 9
one or more of the licensing stages. An applicant may 10 require off-site meteorological data in lieu of or in 11 addition to on-site meteorological data from one or 12 more reputable measurement locations.
13 As Kevin states, some of the examples of 14 these alternate data sources include data from the 15 National Weather Service stations, FAA stations, data 16 from EPA-endorsed measurement programs, and U.S.
17 Department of Defense or Department of Energy 18 facilities.
19 MEMBER HALNON: Jason, this is Greg. Do 20 you have any standards or expectations on what 21 services or what data sources are acceptable? I mean, 22 we've had one applicant use their state, you know, 23 stuff, I guess, regional data. We've looked at the 24 places --
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 MR. WHITE: Yes, we usually take a look at 1
them on a case-by-case basis. And, you know, if you 2
make a determination that it's a reputable source that 3
has a decent history of quality control and, you know, 4
we also do our own independent analysis of the data to 5
check the quality, and if we decide that it's of good 6
quality and appropriate for use, we'll usually accept 7
it.
8 MEMBER HALNON: Okay. So a topic of pre-9 application discussions, I guess, would be more 10 appropriate then, right?
11 MR. WHITE: Yes, yes, usually we have 12 those discussions during the initial stages of contact 13 with the applicants.
14 MEMBER HALNON: Okay, thanks. Go on.
15 MR. WHITE: No problem. Next slide, 16 Kevin.
17 If the data from some of these alternate 18 sources cannot be input directly into ARCON, the 19 applicant will need to perform additional processing 20 before the data can be used. Also, the applicant 21 should identify each meteorological monitoring 22 location to be used and the rationale for the 23 selection. The applicant should then identify and 24 explain any departures between the meteorological 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 monitoring guidance used at each of the off-site 1
locations and the monitoring guidance that's outlined 2
in Reg Guide 1.23.
3 Next slide. As stated in regulatory 4
position two of Reg Guide 1.145 and performed by PAVAN 5
computer code, a 95 percentile overall site chi over 6
q value or a 99.5 percentile chi over q value for each 7
direction of sector should be determined for each 8
analysis. In determining the bounding release 9
receptor combinations, it will be necessary to 10 consider the distance, direction, release mode, and 11 height of the release locations to various EAB and LPZ 12 receptor locations.
13 Next slide. This slide addresses a 14 determination of the source receptor pair distances 15 and directions. The figure shows an example of the 16 limiting and non-limiting distances from buildings 17 within the nuclear island to the EAB and LPZ 18 boundaries.
19 Next slide. The figure depicts an example 20 of how an applicant may use the building locations 21 within the nuclear island area to determine a source 22 receptor pair of the most limiting distance from the 23 edge of the building to the EAB and LPZ. The 24 preferred and most conservative method is to use 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
63 limiting distance from the nearest building edge and 1
apply that distance over all 16 directional sectors 2
and, thus, create a uniform circle around the nuclear 3
island. Using a limiting distance over all 16 4
directional sectors would ensure that the most 5
conservative chi over q value is calculated for each 6
given sector.
7 Next slide. This figure shows the 8
buildings within the nuclear island and each of the 16 9
directional sectors.
10 Next slide. The figure depicts an 11 alternative method for determining the distance for 12 each source receptor pair in each of the 22.5-degree 13 sectors. This method uses the closest point of an 14 applicable building in that sector to the EAB and LPZ.
15 Using the closest point on an applicable building for 16 each sector would create a less conservative but 17 potentially more realistic set of chi over q values 18 than the preferred method we just outlined.
19 Next slide. When determining the bounding 20 release receptor combinations, the ARCON code provides 21 release options that allow an analysis to monitor 22 ground level releases, elevated releases, point source 23 releases, and diffuse source releases. The draft reg 24 guide includes positions that discuss 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
64 preferability of these release types and the 1
limitations on each use, and I think Kevin touched on 2
some of those before. The preferred method is usually 3
the ground level point source release, which is the 4
most conservative.
5 Next slide. This draft reg guide provides 6
a method for applicants and licensees to use ARCON for 7
off-site chi over q estimates. For each of the 16 8
downwind direction sectors, ARCON calculates the 95 9
percentile chi over q values for each source receptor 10 combination, various time errors, periods from two 11 hours1.273148e-4 days <br />0.00306 hours <br />1.818783e-5 weeks <br />4.1855e-6 months <br /> to 30 days. As outlined in regulatory position 12 two of Reg Guide 1.145, the user should also calculate 13 the 99.5 percentile chi over q value for each sector 14 and should select the larger of the two chi over q 15 values. Add to the 99.5 percent maximum sector or the 16 95 percent overall site value to represent the chi 17 over q values for the two hour interval.
18 Since ARCON does not calculate a maximum 19 sector 99.5 percentile, as performed by PAVAN and 20 stated in Reg Guide 1.145, this draft Reg Guide 1.249 21 describes a methodology the user can use to calculate 22 that 99.5 percentile chi over q value for each sector.
23 Next slide. And now I'll turn it over to 24 Kevin for an example of that methodology that we just 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 discussed.
1 MR. QUINLAN: Thank you, Jason. So I said 2
at the beginning of the presentation that I would 3
apologize then and I'm going to apologize again that 4
I'm going to talk about math on a Friday right before 5
lunchtime. But we wanted to provide an example of how 6
you go calculating the 99.5 percentile chi over q 7
value from the ARCON output because this is really 8
where the main differences come between the two codes, 9
and it is certainly one of the important inputs and 10 important methods that licensees have already used and 11 will likely use in the future.
12 So ARCON produces standard output files 13 that can be used to analyze the data and produce chi 14 over q exceedance frequencies that are not typically 15 calculated in ARCON runs. A standard output file 16 includes a frequency file, which is an output file 17 with the extension of a CFD, cumulative frequency 18 distribution. And the file contains the cumulative 19 frequency distributions of the concentrations 20 calculated for ten different averaging intervals and 21 is designed to be imported into a spreadsheet for 22 further data analysis and display.
23 Please stop me if you have any questions, 24 but I think anybody with a basic understanding 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
66 frequencies and linear interpolation, this should be 1
fairly intuitive. There's really nothing terribly 2
complicated about it.
3 This is an example of the output file.
4 This is the cumulative frequency distribution file 5
that ARCON will produce. This one is cut off here 6
with a couple of ellipses just to save some space, but 7
the very first column here is our chi over q values.
8 These are not model output files. These are 9
thresholds that the -- it's basically between the one-10 year average and the one-hour values. I believe it 11 picks specific values, but these are threshold values, 12 which we'll get into; I'll explain that.
13 The second column here represents a number 14 of one-hour averaging intervals, so the number of 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> in the file, the one-year data set in this case, 16 that exceed this threshold value. So in this case, 17 this value of 2.291 times ten to the negative third is 18 exceeded eight times in a one-hour value.
19 The third column is the number of times 20 that it is exceeded over a two-hour average, and you 21 can go on down the line for your different time 22 periods. So depending on what kind of release you 23 want to model, you can choose your different time 24 periods, you can choose your different thresholds, 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
67 so you can calculate for those different release types 1
and different time periods.
2 The one-hour averaging interval should be 3
used, as discussed previously, to drive the zero to 4
two-hour chi over q value. So that's what our 5
guidance calls for, but it does come from a one-hour 6
average and this is consistent with regulatory 7
positions 1.3 and 1.4 in Reg Guide 1.145. And it's 8
also consistent with the ARCON guidance in NUREG/CR-9 6331. So the two codes that are now being used for 10 off-site estimates are consistent with each other.
11 And then for most ground level release 12 scenarios, the one-hour averaging interval is expected 13 to be bounding. We talked about this when going 14 through the general meteorology slides earlier.
15 So the example describes this, but I'll 16 show you how it's done in practice. Taking this first 17 column, these threshold values, as well as the one-18 hour values, and putting them into a spreadsheet. So 19 these are the exact same values, these are the number 20 of times they've been exceeded, and this is the 21 percent of the time that's been exceeded. So this is 22 just an Excel frequency file.
23 Once the columns are delimited, broken out 24 to different columns, the exceedance frequencies can 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 then be identified. So a new column can be created 1
that provides the percentage of the time that a given 2
chi over q is exceeded. So as I discussed earlier, I 3
gave you an example of individual hours. Eight 4
individual hours exceed this threshold value of 2.29 5
times ten to the negative third. It's exceeded 0.09 6
percent of the time. So 99.91 percent of the time 7
this value is not exceeded.
8 So using the exact same file, this can be 9
done by identifying the total number of hours in the 10 log file, which is a separate file that output, but it 11 tells you some of the higher-level statistics about 12 it. So in this case, there were ten missing hours in 13 the data file, so we're left with 8,750 hours0.00868 days <br />0.208 hours <br />0.00124 weeks <br />2.85375e-4 months <br /> of 14 individual hours that can be used. So we use this 15 equation, equation 1A from the document, from the reg 16 guide. It is just a simple equation to calculate the 17 percentage of times something is exceeded. So a total 18 would be the number of total hours, 8,750, minus the 19 number of hours in that averaging interval or 20 exceeded, so, in this case, eight; and then you 21 multiply it by 100 to turn it into a percentage, and 22 then you end up with these values. So these would 23 subtract the number of hours in this averaging 24 interval column, B, for each chi over q threshold 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
69 the total number of hours, 8,750, divide by the total 1
number of hours, multiply by 100, and then you end up 2
with these frequencies right here.
3 As the chi over q values in the first 4
column, A, get smaller, they go from greatest to 5
smallest here, the chi over q becomes more likely to 6
be exceeded in a defined wind direction window. So as 7
discussed, we recommend a 45-degree window. The 8
default ARCON does a 90-degree window, but it doesn't 9
make a difference for this calculation.
10 Therefore, the number of hours above any 11 given threshold equal or increases as the chi over q 12 decreases. So for chi over q values associated with 13 an average interval of, an internal value of zero, 14 sorry, an interval value of zero, this indicates the 15 chi over q is never exceeded; and, therefore, 100 16 percent of the chi over q values are below this 17 threshold value.
18 Same equation as before, nothing changed.
19 Equation 1A up here should be applied to each line of 20 the CFD file until the exceedance frequency of each 21 chi over q in the first column is identified. The 22 99.5 percent value will most likely, and we certainly 23 hope so, most likely be between two different chi over 24 q threshold values. So I apologize on this file. 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
70 should have had a column D here, which subtracts 1
column C from 100. So you can kind of switch it.
2 In this case, the 99.5 percentile chi over 3
q value is bounded by lines 19 and 20 here, 2.09 and 4
1.91. So if you take that 0.4457, subtract it from 5
100, you end up with a 99.55 percentile value. And if 6
you take the next line and subtract it from 100, you 7
end up with a 98.75 percentile value. So between 8
those two chi over q values will be your 99.5 9
percentile value, and that's what we're trying to 10 calculate here.
11 So only a couple of slides left, and the 12 worst of it over; I promise. So using those two 13 values that we just identified on the screen before, 14 the threshold values, 2.09, times ten to the negative 15 third, and 1.91 times ten to the negative third, a 16 simple linear interpolation can be used to determine 17 the 99.5 percentile chi over q with the following 18 equation. I won't go through the horrible details 19 here, but, basically, why is what we're trying to get 20 here, the 99.5 percentile value, why one is the lesser 21 of the two threshold values for chi over q's, why two 22 is the bigger.
23 You can change X right here. Right now, 24 for this example, it is 99.5 because that's what we're 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 trying to get at. But you could change that to 1
whatever value you're trying to get. Environmental 2
reviews use a 50th percentile chi over q value because 3
they're more concerned with a realistic release and a 4
realistic analysis, instead of a design basis most 5
conservative analysis that the safety review is 6
concerned with. So it's two different ways of doing 7
the same analysis, but X here can be changed for 8
whatever value or whatever threshold you're looking to 9
pull out. And then X-1 is the lesser of the 10 exceedance frequencies, and then X-2 is the greater of 11 the two exceedance frequencies.
12 So using the exact same equation with 13 those same values, this is what it actually would look 14 like. So we put a 99.5 here. We subtract out the chi 15 over q's and try to get the difference between those 16 two threshold values. We then take the difference 17 between those exceedance frequencies, you run it 18 through the math, and you end up with a chi over q 19 value of 2.077 times ten to the negative third.
20 So when error-checking, I went back to 21 check and said, all right, does this make physical 22 sense, does this make, you know, sense overall in 23 calculating this. And because this value here, 2.089 24 times ten to the negative third, is quite close 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
72 99.5 percentile value anyway, it's 99.55, you would 1
expect the resulting chi over q to be quite close this 2
one and much farther away from the 98.75, and it is.
3 You can see they're very, very close to each other, so 4
that give us some confidence that this example is 5
working correctly. But it could certainly be modified 6
for different time periods and for different threshold 7
values.
8 And that wraps up the math for my portion.
9 And if there are no questions on that, I will give it 10 back to Jason for our final two or three slides, and 11 then we can open it up to any public questions that 12 may be out there.
13 MR. WHITE: All right. Thanks, Kevin. I 14 think we addressed most of the next points during the 15 course of the meeting, but I'll provide a summary as 16 to why this guidance was needed and the benefits it 17 will provide.
18 NRC regulatory guides provide approaches 19 that the staff considers acceptable for meeting the 20 regulatory requirements addressed by that specific 21 guide. At the same time, an applicant has the 22 flexibility to propose alternate approaches to the 23 guidance, so long as it identifies any differences and 24 demonstrates that the applicable regulatory 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 requirements are still met.
1 NRC staff has received topical reports and 2
license applications proposing the use of an alternate 3
approach to calculate the off-site atmospheric 4
dispersion values of the EAB and the LPZ by using the 5
ARCON computer code.
6 Next slide. Without applicable guidance, 7
the NRC will continue to require justification for the 8
applicability of the alternate approach in each 9
application and continue to review them on a case-by-10 case basis. This could lead to the following:
11 additional staff time to conduct non-standardized 12 reviews and increased number of clarifying questions 13 and RAIs during the staff's review, increased levels 14 of interaction between the staff and applicant during 15 both the pre-application and application review 16 periods, and also a more expensive and lengthy 17 application review at additional cost to the 18 applicant.
19 Next slide. Draft Reg Guide 1.249 20 describes and approach that is acceptable to the NRC 21 staff to meet the NRC requirements for determining the 22 chi over q values and support of modeling on-site 23 releases to outside boundaries using the ARCON code.
24 Draft Reg Guide 1.249 would enhance the efficiency and 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
74 the effectiveness of the review process of licensing 1
applications and topical reports by having the common 2
guidance document as the technical basis for the use 3
of the ARCON methodology.
4 Next slide. And this concludes our 5
presentation this morning, and we will be happy to 6
address any questions that you have. Thank you very 7
much for your attention.
8 MEMBER HALNON: Thank you, Jason and 9
Kevin. Very, very informative, incredibly valuable 10 for us to get back on to this topic since we're seeing 11 it in the background so much right now in the 12 applications we're looking at.
13 Any members or consultants have any 14 further questions or comments? Okay. Given that, I'm 15 going to open up the line to any public comments.
16 Again, recall that, if you're on the phone, star 6 17 will unmute you. If you're on Teams, you can unmute 18 yourself and introduce yourself and affiliation, if 19 appropriate. So at this time, any public comments?
20 Okay. Very good. I understand that 21 there's no public comments. Again, Kevin and Jason, 22 I appreciate your time and efforts in putting this 23 together. The conversation was worthwhile; and, 24 again, because we're seeing this so much 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
75 background of our applications and a lot of diverse 1
places, we appreciate the time and effort you put in 2
to creating this presentation. And I believe this 3
will live for a while on our website for the ACRS and 4
background materials because it's useful to go back 5
and take a look at, especially some of the simplified 6
pictures that you provided us.
7 One last chance for any comments from any 8
members or our consultants.
9 DR. SCHULTZ: Greg, just one comment.
10 This is Steve Schultz. One comment since there were 11 no comments from the public. When this went out for 12 public comments, I just wanted to commend those that 13 did provide comments to the staff. There were a 14 number of substantial comments on the particulars of 15 the reg guide and commenters focused in great detail 16 in ways in which to make the guide clearer and more 17 usable. So a lot of good work was done there by those 18 reviewers, public reviewers.
19 And I'd also like to commend the staff for 20 their response to those comments. It was clear that 21 they were looked at thoroughly and the staff 22 considered them carefully and made changes to the 23 guide where it was appropriate. So it was a nice job 24 in that review and response.
25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1716 14th STREET, N.W., SUITE 200 (202) 234-4433 WASHINGTON, D.C. 20009-4309 www.nealrgross.com
76 MEMBER HALNON: Thank you, Steve. Totally 1
agree. When I read through the comments, I usually 2
gauge a comment from whether or not I can understand 3
the math behind it, and I didn't understand a lot of 4
the math, so they must have been good comments. That 5
works for me.
6 Thank you. Any other members?
7 MEMBER REMPE: This is Joy, and I didn't 8
have a comment pertaining to this session, but I would 9
like members to stay on for a few-minute discussion 10 after you close the record off, Greg. I just didn't 11 want people to hang up and not stay on.
12 MEMBER HALNON: Thank you, thank you, Joy.
13 So, again, just to reiterate, if the members could 14 stay on after I adjourn, I would appreciate it and Joy 15 has a quick message for us.
16 Okay. At this point, then it's about 12 17 after 11 Eastern Time. I will adjourn this meeting, 18 and I appreciate everybody's attendance and the good 19 conversation. The meeting is adjourned.
20 (Whereupon, the above-entitled matter went 21 off the record at 11:12 a.m.)
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
Draft RG 1.249 Use of ARCON Methodology for Calculation of Accident-Related Offsite Atmospheric Dispersion Factors Presentation to ACRS July 22, 2022
Outline
- Overview of Atmospheric Dispersion (/Q)
- Use in Siting and Design
- Current Guidance for /Qs at EAB/LPZ
- Relationship to other ATD RGs
- Expanded use of ARCON code
- How new RG will be used
- Example Calculation
Overview of Atmospheric Dispersion
Atmospheric Transport and Dispersion (ATD) Modeling Atmospheric Dispersion Transport (Advection)
Diffusion Removal Transformation
/Q (sec/m3) - normalized concentration factor Concentration (Bq/m3) divided by release rate Q (Bq/s)
D/Q (1/m2) - normalized deposition factor Deposition Rate D (Bq/m2s) divided by release rate Q (Bq/s)
ATD Modeling 5
ATD Modeling Important meteorological measurements for dispersion calculations Wind direction Wind speed Atmospheric turbulence (stability)
Precipitation Diffusion caused by atmospheric turbulence Thermal, mechanical Transport with the wind (advection)
Onsite measurements One year minimum
Gaussian Plume Model
- Historically widely accepted in NRC radiological assessment activities associated with licensing
- Fast
- Only a few input parameters are necessary Assumptions Emission rate is constant Horizontal meteorological conditions are homogeneous over the space being modeled.
For each hour modeled:
- An average wind speed is used
- Wind direction is constant (straight-line model)
- Atmospheric turbulence (stability) is constant No wind shear in the horizontal or vertical Plume is infinite with no plume history
- Each hour being modeled is independent of the previous hour Dispersion in the crosswind (y direction) and vertical (z direction) has a Gaussian (normal) distribution Diffusion is negligible in the downwind (x) direction
Gaussian Plume Model
Atmospheric Turbulence
- Thermal (ground heating and cooling)
- Mechanical (surface roughness)
Causes
- Indirect Measurements
- Wind speed (insulation method)
- Solar radiation or sky cover during the day
- Net radiation or sky cover at night
- Temperature lapse rate (delta-T)
- Direct Measurements - some combination of:
- Horizontal wind speed
- Horizontal wind speed and/or wind direction fluctuations
- Vertical wind speed and/or wind direction fluctuations Measurement Methods
Atmospheric Turbulence Temperature (deg C)
Height (km)
Atmospheric Turbulence Atmospheric Condition Typical Conditions Heat Flux Nature of Turbulence Unstable (A, B, C)
Mid-Day Clear Sky Light Wind Net Upward (Delta-T << 0)
Considerable Horizontal and Vertical Neutral (D)
Windy or Cloudy or Transition Zero (Delta-T < 0)
Mid-Range Stable (E, F, G)
Nighttime Clear Sky Light Wind Net Downward (Delta-T > 0)
Damps Out Vertical
Atmospheric Turbulence Unstable (A, B, C)
Neutral (D)
Stable (E, F, G)
Real world examples of atmospheric dispersion Fumigation Unstable Plume Stable Plume No wind
Regulatory Guide 1.23 Classification of Atmospheric Stability Stability Classification Pasquill Stability Category Ambient Temperature Change With Height (deg C/100m)
Extremely unstable A
T -1.9 Moderately unstable B
-1.9 < T -1.7 Slightly unstable C
-1.7 < T -1.5 Neutral D
-1.5 < T -0.5 Slightly stable E
-0.5 <T 1.5 Moderately stable F
1.5 < T 4.0 Extremely stable G
T > 4.0
Typical Onsite Meteorological Monitoring Program
- 60-m: wind speed, wind direction, temperature
- 10-m: wind speed, wind direction, temperature
- Between 60-m and 10m: delta-temperature
- 3-m: dew point
- 2.5-m: solar radiation
- Ground-level: precipitation Primary tower
- 10-m: wind speed, wind direction, wind direction standard deviation, temperature Backup tower
- once per second Sampling rate
Use in Siting and Design Current Guidance for EAB and LPZ
Atmospheric Dispersion
- Used in Siting, Licensing, and Design
- Two NRC models used for Design-Basis Accident (DBA) analysis
- PAVAN (EAB / LPZ)
- ARCON (Control Room)
DBA Releases to the EAB and LPZ Applicable NRC Guidance
- SRP 2.3.4: Short-Term Atmospheric Dispersion Estimates for Accident Releases
- RG 1.145: Atmospheric Dispersion Models for Potential Accident Consequence Assessments at Nuclear Power Plants (1983)
- NUREG/CR-2858: PAVAN: An Atmospheric Dispersion Program for Evaluating Design-Basis Accident Releases of Radioactive Materials from Nuclear Power Stations (1982)
- NUREG/CR-2260: Technical Basis for Regulatory Guide 1.145, Atmospheric Dispersion Models for Potential Accident Consequence Assessments at Nuclear Power Plants (1981)
PAVAN Description Gaussian plume model
- for ground-level releases, diffusion coefficients modified to account for plume meander under low wind speed conditions and building wakes Estimates /Q values for various time-averaged periods 2 hrs, 0-8 hrs, 8-24 hrs, 1-4 days, 4-30 days Meteorological input consists of a joint frequency distribution (JFD) of hourly values of:
- wind speed (calms defined as below sensor threshold, historically ~ 1 mph)
- wind direction (16 directions, 22.5 deg sectors, centered on true north)
- atmospheric stability class (preferably based on delta-T)
Building wake impacts on release height
- release points less than 2.5 times the height of adjacent solid structures ground-level releases
- release points more than 2.5 times the height of adjacent solid structures elevated (stack) releases Part-Time fumigation conditions assumed for stack releases
PAVAN Description The larger of the following two calculated /Q values is selected to represent the /Q value for the 0-2 hour time interval.
- 0.5-percent maximum sector value percent overall site value These calculated /Q values are based on 1-hour averaged data but are conservatively assumed to apply for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> This procedure is repeated two times:
- Once for the EAB
- Once for the LPZ
PAVAN Description 0.5-percent Maximum Sector /Q Value For each of the 16 downwind direction sectors (N, NNE, NE, ENE, etc.), /Q values are calculated for each combination of wind speed and atmospheric stability at the appropriate downwind distance The /Q values calculated for each sector are then placed in order from the greatest to the smallest, and an associated cumulative frequency distribution is derived based on the frequency distribution of wind speed and stabilities for each sector.
An upper envelope curve is determined for each sector, based on the derived data (plotted as /Q versus probability of being exceeded), so that no plotted point is above the curve From this upper envelope, the /Q value, which is equaled or exceeded 0.5 percent of the total time (44 hrs) is obtained The maximum 0.5 percent /Q value from the 16 sectors becomes the 0-2 hour 0.5-percent maximum sector /Q value
PAVAN Description 5-percent Overall Site /Q Value
- Using the same approach, all /Q values independent of wind direction are combined into one cumulative frequency distribution for the entire site.
- An upper envelope curve is determined, and the /Q value that equals or exceeds 5.0 percent of the total time (438 hours0.00507 days <br />0.122 hours <br />7.242063e-4 weeks <br />1.66659e-4 months <br />) is selected
PAVAN Description 1.E-05 1.E-04 1.E-03 1.E-02 0
1 2
3 4
5 6
7 8
9 10 X/Q (sec/m^3)
Percent Frequency (with respect to total time)
Min WS: 1.0 m/s Min WS: 0.5 m/s An upper envelope curve
PAVAN Description To determine LPZ /Q values for longer time periods (e.g.,
0-8 hours, 8-24 hours, 1-4 days, and 4-30 days), PAVAN performs a logarithmic interpolation between the 0-2 hour
/Q values and the annual average (8,760 hours0.0088 days <br />0.211 hours <br />0.00126 weeks <br />2.8918e-4 months <br />) /Q values for each of the 16 sectors and the overall site. For each time period, the highest among the 16-sector and overall site /Q values is identified and becomes the short-term site characteristic /Q value for that time period.
1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1
10 100 1,000 10,000
/Q (sec/m3)
Averaging Period (hours)
DBA Releases to the EAB and LPZ Definitions (10 CFR 50.2 and 10 CFR 100.3)
- Exclusion Area
- The area surrounding the reactor, in which the reactor licensee has the authority to determine all activities including exclusion or removal of personnel and property from the area
- Low Population Zone
- The area immediately surrounding the exclusion area which contains residents, the total number and density of which are such that there is a reasonable probability that appropriate protective measures could be taken in their behalf in the event of a serious accident 25
EAB
ARCON Description Gaussian plume model
- diffusion coefficients account for enhanced dispersion under low wind speed conditions and in building wakes
/Q values are estimated for various time-averaged periods 2 hrs, 2-8 hrs, 8-24 hrs, 1-4 days, 4-30 days Meteorological input consists of hourly values of wind speed, wind direction, and atmospheric stability class Hourly meteorological data are used to calculate hourly /Q values
- Hourly /Q values are then combined to estimate concentrations ranging in duration from 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> to 30 days
- Cumulative frequency distributions are prepared from the average /Q values
- /Q values that are exceeded no more than 5 percent of the time for each averaging period are selected
AP1000 DBA Release and Receptor Locations Receptors 1.
Control Room HVAC Intake 2.
Annex Building Access Sources 1.
Plant Vent 2.
PCS Air Diffuser 3.
Fuel Building Blowout Panel 4.
Fuel Building Rail Bay Door 5.
Steam Vent 6.
PORV & Safety Valves 7.
Condenser Air Removal Stack 8.
Containment Shell
DBA Releases to the CR and TSC
- Applicable NRC Guidance
- SRP 2.3.4: Short-Term Atmospheric Dispersion Estimates for Accident Releases
- RG 1.194: Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants (2003)
- NUREG/CR-6331: Atmospheric Relative Concentrations in Building Wakes (1997)
- ARCON96
Draft RG 1.249 Relationship to Existing ATD Guidance
RG 1.194: Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants (2003)
Provides guidance on determining
/Q values in support of design basis control room habitability assessments.
RG 1.145: Atmospheric Dispersion Models for Potential Accident Consequence Assessments at Nuclear Power Plants (1983)
Provides guidance on determining
/Q values for DBA at the EAB and LPZ
- Met Data Input
- Hourly Met Data
- Model Setup
- One source-receptor (SR) pair (distance and direction) per run
- Can use ground-level or elevated releases
- Point or diffuse source
- Model Results
- Determines the 95th-percentile /Q for each SR pair Draft RG 1.249
- Met Data Input
- Hourly Met Data
- Model Setup
- One source-receptor pair (distance and direction) per run
- Can use ground-level or elevated releases
- Point or diffuse source
- Model Results
- Determines the 95th-percentile /Q for each SR pair
- Includes a method to determine maximum sector 99.5th-percentile
ARCON Input Display
ARCON Input Display ARCON default is 90-degree window, per RG 1.194 for use with control room dispersion.
RG 1.249 recommends changing to a 45-degree window for consistency with RG 1.145 for EAB / LPZ calculations.
RG 1.145 (PAVAN)
- Met Data Input
- Joint Frequency Distribution
- Model Setup
- 16 source directions at same time
- Multiple distances at once
- Model Results
- Larger of 95% overall or 99.5% /Q maximum sector value is selected for 0-2-hour timeframe.
Draft RG 1.249
- Met Data Input
- Hourly Met Data
- Model Setup
- One source-receptor pair (distance and direction) per run
- Can use ground-level or elevated releases
- Point or diffuse source
- Model Results
- User can use either 95% /Q within 45-degree window or calculate the 99.5% /Q from within that same window
Expanded Use of ARCON Code
Expanded Use of ARCON Code The ARCON computer code was developed to model shorter distances in the vicinity of buildings typical of control room habitability evaluations.
The ARCON dispersion algorithms are based on field measurements taken out to distances of 1,200 m.
Large light-water nuclear power plants typically have EAB and LPZ distances that range from 800 to 6,000 m.
Small modular and advanced reactor designs are expected to include EAB and LPZ distances in the range of 80 to 400 m.
37
Expanded Use of ARCON Code
- In its Accident Source Term topical report NuScale proposed using the ARCON computer code methodology for calculating offsite atmospheric dispersion values at the EAB and LPZ rather than using the PAVAN computer code
- NRC staff conducted an audit of the topical report
- Staff reviewed the documentation for the methodology
- Staff performed an independent analysis using the methodology 38
Expanded Use of ARCON Code
- The NRC staff determined that the ARCON computer code is acceptable for modeling EAB and LPZ /Q values at relatively short distances as long as the methods are consistent with the Regulatory positions of RG 1.145 for offsite /Q values.
- Therefore, the methodology of draft RG 1.249 is applicable to EAB and LPZ distances from source locations within the nuclear island to a distance of 1,200 m.
39
How will RG 1.249 be Used?
Uses of Draft RG 1.249 Guidance on Locations Draft RG 1.249 is applicable to sites with EAB and LPZ distances from source locations out to a distance of 1,200 m.
Certain locations are affected by atmospheric transport and diffusion conditions that may be more restrictive than assumed in the contiguous 48 states.
For locations characterized by extreme and persistent restrictive dispersion conditions (e.g., in Alaska), the applicability of the dispersion algorithms in ARCON may not apply or may require further modification.
41
Uses of Draft RG 1.249 Guidance on Locations The atmospheric dispersion analysis may be affected by deployment in certain locations. Issues include, but may not be limited to, the following:
Different Characteristics of accident releases (e.g., buoyancy, momentum) depending on ambient conditions at the time of release Transport and diffusion conditions possibly being significantly different or more restrictive than assumed Seasonal variation of dispersion and meteorological conditions such that separate modeling approaches and models may be necessary at certain times of the year The possible need for field studies to characterize and model dispersion conditions 42
Uses of Draft RG 1.249 Guidance on Alternative Meteorological Data Due to the potential of new or advanced reactor designs being considered in nontraditional locations there may be a need to seek alternative meteorological sources at one or more of the licensing stages.
An applicant may acquire offsite (in lieu of, or in addition to, onsite) meteorological data from one or more reputable measurement locations.
Examples include data from National Weather Service (NWS) stations, Federal Aviation Administration stations, U.S. Environmental Protection Agency (EPA)-endorsed measurement programs, U.S. Department of Defense or U.S. Department of Energy facilities, etc.
43
Uses of Draft RG 1.249 Guidance on Alternative Meteorological Data If these data cannot be input directly into ARCON, the applicant will need to perform additional processing before the data can be used.
The applicant should identify each offsite meteorological monitoring location to be used and the rationale for the selection.
The applicant should identify and explain any departures between the meteorological monitoring guidance used at each offsite location and the monitoring guidance in RG 1.23.
44
Uses of Draft RG 1.249 Release Characteristics As stated in Regulatory Position 2 of RG 1.145 and performed by the PAVAN computer code, a 95th-percentile overall site
/Q value, or a 99.5th-percentile /Q value for each directional sector should be determined for each analysis.
In determining the bounding release-receptor combinations, it will be necessary to consider the distance, direction, release mode, and height of the release location(s) to the various EAB and LPZ receptor locations.
45
Uses of Draft RG 1.249 Source-Receptor Pair Distances and Directions Figure shows an example of the limiting and nonlimiting distances from buildings within the nuclear island to the EAB/LPZ boundaries 46
The figure depicts an example of how an applicant may use the building locations within the nuclear island area to determine a source-receptor pair of the most limiting distance from the edge of the building to the EAB/LPZ.
The preferred, and most conservative, method is to use the limiting distance from the nearest building edge and apply that distance over all 16 sectors, thus creating a uniform circle around the nuclear island.
Using the limiting distance over all 16 directional sectors would ensure that the most conservative /Q value is calculated for each given sector.
Uses of Draft RG 1.249 Source-Receptor Pair Distances and Directions Figure shows the buildings within the nuclear island and each of the 16 directional sectors 48
This figure depicts the alternative method for determining the distance for each source-receptor pair in each 22.5-degree sector. This method uses the closest point of an applicable building in that sector to the EAB/LPZ.
Using the closest point on an applicable building for each sector would create a less conservative but potentially more realistic set of /Q values than the preferred method.
Uses of Draft RG 1.249 Release Characteristics The ARCON code provides release options that allow an analyst to model:
-Ground-level releases
-Elevated (stack) releases
-Point source releases
-Diffuse Source The draft RG includes positions that discuss the preferability of these release types and limitations on their use.
50
Uses of Draft RG 1.249 Determining the 95th-Percentile and 99.5th-Percentile /Q Values This Draft RG provides a method for applicants and licensees to use ARCON for offsite /Q estimates. For each of the 16 downwind direction sectors (e.g., N, NNE, NE, ENE), ARCON calculates the 95th-percentile
/Q values for each source-receptor combination for various time-averaged periods ranging from 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> to 30 days.
As outlined in Regulatory Position 2 of RG 1.145. the user should calculate the 99.5th-percentile /Q value for each sector and should select the larger of the two /Q values, either the 99.5-percent maximum sector value or the 95-percent overall site value, to represent the /Q value for the 0-2-hour time interval.
Since ARCON does not calculate a maximum sector 99.5th percentile, as performed by PAVAN and stated in Regulatory Position 2 of RG 1.145, draft RG 1.249 describes a methodology the user can use to calculate the 99.5th-percentile /Q value for each sector.
51
Example Process for Calculating 99.5th-Percentile
/Q ARCON produces standard output files that can be used to analyze the data and produce /Q exceedance frequencies not calculated in typical ARCON runs.
The standard output files include a Frequency File output file with the extension.CFD (cumulative frequency distribution).
The.CFD file contains the cumulative frequency distributions of the concentrations calculated for 10 averaging intervals and is designed to be imported into a spreadsheet for further data analysis and display.
Example Process for Calculating 99.5th-Percentile /Q The second column represents the number of 1-hour averaging intervals that are exceeded by the X/Q thresholds provided in the first column The third column (with a header of
- 2) represents the number of 2-hour averaging intervals that are exceeded by the /Q thresholds.
Example Process for Calculating 99.5th-Percentile /Q The 1-hour averaging interval should be used to derive the 0-2-hour /Q values, as is consistent with Regulatory Positions 1.3 and 1.4 in RG 1.145 Section 3.7 of NUREG/CR-6331 (ARCON), Revision 1, states that the larger of the 1-and 2-hour average relative concentrations should be used for the 0-2-hour period.
For most ground-level release scenarios, the 1-hour averaging interval is expected to be bounding.
Example Process for Calculating 99.5th-Percentile /Q The frequency file is intended to be imported into a spreadsheet and divided into columns.
Once the columns are delimited, the /Q exceedance frequencies can be identified.
A new column can be created that provides the percentage of the time that a given /Q value is exceeded.
Example Process for Calculating 99.5th-Percentile /Q This can be done by identifying the total number of hours in the.LOG file for each averaging interval (8,750 for the 1-hour averaging interval)
For this example, this value is HTotalin Equation A-1.
The user would then subtract the number of hours in the averaging interval column (B) for each /Q threshold (this would be 8 for the /Q threshold of 2.29x10-3 seconds per cubic meter (s/m3), 39 for the /Q threshold of 2.09x103 s/m3) from the total number of hours, divide by the total number of hours, and multiply by 100 to calculate the percentage of the time each /Q is exceeded in the.CFD file.
Example Process for Calculating 99.5th-Percentile /Q As the /Q values in the first column get smaller, the /Q value becomes more likely to be exceeded in the defined wind direction window.
Therefore, the number of hours above any given threshold increases as the /Q decreases.
For /Qs associated with an averaging interval value of 0, this indicates that this /Q limit is never exceeded, and therefore 100 percent of the hourly /Qs are below this threshold.
Example Process for Calculating 99.5th-Percentile /Q Equation A-1 should be applied to each line in the.CFD file until the exceedance frequency of each /Q in the first column is identified.
The 99.5th-percentile /Q value will most likely be between two /Q threshold values.
In this figure, the 99.5th-percentile /Q value is bounded by the 2.09x10-3 s/m3 /Q (99.55th percentile) and the 1.91Ex10-3 s/m3 /Q (98.75th percentile).
Example Process for Calculating 99.5th-Percentile /Q Using the two bounding /Q threshold values from Equation A-1 [2.09x10-3 s/m3 /Q (99.55th percentile) and the 1.91x10-3 s/m3 /Q (98.75th percentile)], a simple linear interpolation can be used to determine the 99.5th-percentile /Q using the following equation:
- where, y
=
the resulting 99.5th-percentile /Q value y1
=
the lesser of the two bounding /Q values y2
=
the greater of the two bounding /Q values x
=
99.5 x1
=
the lesser of the two exceedance frequencies x2
=
the greater of the two exceedance frequencies
Example Process for Calculating 99.5th-Percentile /Q
Benefits of Draft RG 1.249 NRC regulatory guides provide approaches that the staff considers acceptable for meeting the regulatory requirements addressed by the specific guide.
At the same time, an applicant has the flexibility to propose alternate approaches to the guidance so long as it also identifies any differences and demonstrates that the applicable regulatory requirements are still met.
NRC staff has received topical reports and licensing applications proposing the use of an alternate approach to calculating offsite atmospheric dispersion values at the EAB and the LPZ by using the ARCON computer code.
Benefits of Draft RG 1.249 Without applicable guidance, the NRC would continue to require a justification for the applicability of the alternate approach in each application and continue to review them on a case-by-case basis. This could lead to the following:
Additional staff time to conduct non-standardized reviews An increased number of clarifying questions and RAIs during the staffs review Increased levels of interaction between the staff and applicant during both the pre-application and application review periods More expensive and lengthy application reviews at an additional cost to the applicant
Benefits of Draft RG 1.249 Draft RG 1.249 describes an approach that is acceptable to the NRC staff to meet the NRC requirements for determining /Q values in support of modeling onsite releases to offsite boundaries using the ARCON code.
Draft RG 1.249 would enhance the efficiency and effectiveness of the review process of licensing applications and topical reports by having a common guidance document as the technical basis for the use of the ARCON methodology.
Questions?