NUREG/CR-6331, Discusses 970627 Meeting Between W & NRC Re Assessment of Atmospheric Dispersion Re W AP600 Control Room Habitability. Believes W Should Consider Use of Alternative Methodology, Described in Rev 1 to NUREG/CR-6331
| ML20149J235 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 07/23/1997 |
| From: | Quay T NRC (Affiliation Not Assigned) |
| To: | Liparulo N WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| References | |
| RTR-NUREG-CR-6331 NUDOCS 9707280126 | |
| Download: ML20149J235 (45) | |
Text
{{#Wiki_filter:. __ . _ _ . . July 23,1997' i^ Mr. Nichol:s J. Liparulo, Manager .
~ Nuclear Safety ~and Regulatory Activities I Nuclear and Advanced Technology Division ,
Westinghouse Electric Corporation, l i P.O. Box 355 ' l Pittsburgh, PA 15230 . 1
SUBJECT:
ASSESSMENT OF ATMOSPHERIC DISPERSION RELATED TO WESTINGHOUSE AP600 CONTROL ROOM HABITABILITY
Dear Mr. Liparulo:
1 During a June 27, 1997 meeting, Westinghouse personnel and NRC staff discussed i , the atmospheric dispersion assumptions for control room habitability design basis accident (DBA) assessments for the AP600. As stated in its response to NRC RAI 470.3, Westinghouse developed its model based on the methodology described in NUREG/CR-5055, " Atmospheric Diffusion for Control Room Habitability Assessments." As a result of the peer review conducted in 1994 , and the availability of the ARCON96 methodology, the NRC staff has determined j that the NUREG/CR-5055 methodology should no longer be used for licensing ' applications. Therefore, Westinghouse should consider use of an alternative l methodology such as that described in Revision 1 to NUREG/CR-6331, l , " Atmospheric Relative Concentrations in Building Wakes," (ARCON96) or the Murphy-Campe methodology identified in Section 6.4 of the Standard Review Plan (NUREG-0800). The information in Enclosures 1, 2, and 3 provide background 2 , and information regarding the 1994 peer review that may be helpful in understanding the staff's determination regarding NUREG/CR-5055. Any use of l the ARCON96 methodology for the AP600 control room habitability assessment should include consideration of the matters discussed in Enclosure 4. If you have any questions, please call Thomas Kenyon at (301) 415-1120. Sincerely, original signed by: Theodore R. Quay, Director Standardization Project Directorate L Division of Reactor Program Management Office of Nuclear Reactor Regulation Docket No. 52-003 n
Enclosures:
- 1. NUREG/CR-5055 and Atmospheric Dispersion Related to Control Room Habitability 0lgn. 2. May 20, 1994 letter from J. Ramsdell, PNNL b 3. August 23, 1994 letter from J. Ramsdell, PNNL y 4. Considerations on Inputs to ARCON96 for Westinghouse AP600 Design Basis Accident Assessments I f cc: See next page I h<
m DISTRIBUTION Docket File CMiller JSebrosky hbb l R PUBLIC REmch BZalcman Rg ACRS (11) TKenyon JKudrick o a.< PDST Reading File WHuffman MSnodderly PERB Reading File JWilson RPalla MSlosson TTMartin WDean, 0-17 G21 JMoore, 0-10 B18 LBrown JLee Ng M M BCEr TRQuay DJackson DOCUMENT NAME: G:\ WEST 2TRQ Ts receive a copy of this documeset insecate in the boa: "C" = Copy without attachment / enclosure *E' = Copy with attachment / enclosure "N" = No copy OFFICE POST:NRR/ W l BC:PER8:NRR l4 BC:PDST:NRR l l l NAME TKenyon(/l CMILler & m Touay dlM DATE 07/ M /1/ 07/ JA /97 07/ 'L3 /97 4vVU0J A(6 0FFICLAL RECORD COPY
I Mr. Nicholas J. Liparulo Docket No. 52-003 i- Westinghouse Electric Corporation 4 AP600 cc: Mr. B. A. McIntyre Mr. Ronald Simard, Director Advanced Plant Safety & Licensing Advanced Reactor Programs Westinghouse Electric Corporation Nuclear Energy Institute i Energy Systems Business Unit 1776 Eye Street, N.W. P.O. Box 355 Suite 300 Pittsburgh,_PA 15230 Washingtn , DC 20006-3706 $ Mr. Cindy L. Haag Ms. Lynn Connce l Advanced Plant Safety & Licensing Doc-Search Associates i Westinghouse Electric Corporation Post Office Box 34 Energy Systems Business Unit Cabin John, MD 20818 i Box 355 Pittsburgh, PA 15230 Dr. Craig D. Sawyer, Manager Advanced Reactor Programs
- Mr. S. M. Modro GE Nuclear Energy i Nuclear Systems Analysis Technologies 175 Curtner Avenue, MC-754 i Lockheed Idaho Technologies Company San Jose, CA 95125 Post Office Box 1625
, Idaho Falls, ID 83415 Mr. Robert H. Buchholz GE Nuclear Energy Mr. Sterling Franks 175 Curtner Avenue, MC-781 U.S. Department of Energy San Jose, CA 95125
' NE-50 19901 Germantown Road Barton Z. Cowan, Esq.
1 Germantown, MD 20874 } Ecker: Seamans Cherin & Mellott 600 Grant Street 42nd Floor 'j .Mr. Frank A. Ross Pittsburgh, PA 15219 U.S. Department of Energy, NE-42 i Office of LWR Safety and Technology Mr. Ed Rodwell, Manager 19901 Germantown Road PWR Design Certification
- Germantown, MD 20874 i Electric Power Research Institute I 3412 Hillview Avenue Mr. Charles Thompson, Nuclear Engineer Palo Alto, CA 94303 j.
AP600 Certification NE-50 19901 Germantown Road j_ Germantown, MD 20874 } t
i i . l. j NUREG/CR-5055 AND ATMOSPHERIC DISPERSION {. RELATED TO CONTROL R00N HABITABILITY l 4 i
- General Design Criteria 19 of 10 CFR Part 50 requires that nuclear power plant control rooms be designed to provide adequate protection from radioactive i
1 effluents under routine and accident conditions. Standard Review Plan 6.4 of NUREG-0800 references the 1974 Murphy-Campe paper as an acceptable methodology for calculating some of the inputs needed in the design basis accident (DBA) j assessment, includirg relative concentration values (X/Qs). This methodology l was intended to be a bounding-type calculation requiring little site-specific j information. l By the mid-1980s, after a number of atmospheric dispersion field tests were conducted within building complexes, it became apparent that the Murphy-Campe methodology tended to overestimate relative concentrations under some atmospheric conditions. The NRC decided to consider the feasibility of l identifying or developing a more robust methodology for its use that would be
- capable of better describing atmospheric dispersion near buildings. The
- initial effort performed for the NRC by Pacific Northwest National Laboratory
- -(PNNL) resulted in a statistical multiple regression analysis "best-fit" of
! meteorological and available field measurement data. NUREG/CR-5055, j " Atmospheric Diffusion for Control Room Habitability Assessments," describing j that work, was published in 1988. Using detailed site-specific meteorological
- field data, the methodology showed improved performance in predicting the
! effects of different atmospheric conditions on maximum effluent concentrations
- in building wakes, particularly under light wind conditions. The NRC staff j took no action to endorse the work for generic applicability.
- In 1994, PNNL seated a peer review group for the NUREG/CR-5055 methodology and follow-on work completed by PNNL. Several comments resulted from the peer i review, including that the turbulence generated by buildings should be assumed
- to be proportional to the wind speed in conformance with accepted theory and i physical reasoning, and that the effects of meander during low wind speed i conditions should be treated explicitly in the model, but separate from the j treatment of building wakes. PNNL revised the methodology to consider the comments and developed the ARCON95 computer code described in NUREG/CR-6331,
" Atmospheric Relative Concentrations in Building Wakes," (1995). Subsequent review of the ARCON95 methodology by the NRC staff and PNNL code developer
- resulted in recommendations for enhanecent and several additional l modifications. In May 1997, Revision 1 to NUREG/CR-6331, " Atmospheric
! Relative Concentrations in Building Wakes," was pu61ished and, in July 1997, l the ARCON96 computer code was made available on disk. l As a result of the shortcomings identified by the peer review group and the i availability of the ARCON96 methodology, the staff determined that the
- NUREG/CR-5055 methodology should no longer be used in licensing actions. In a i May 30, 1997 letter, the NRC staff provided information concerning this
- determination to an operating reactor licensee. NRC is in the process of
} getting ARCON96 approved on a generic basis. ,
j Enclosure 1 5 4-
! :o - i i : QBattelle i Pacific Northwcu taboratories sai e4 swa,e ess.m - adhiand. Wa h.nciv MM.' h
- **P *** 8s00376-8626 j May 20,1994 1
i Mr. Jay Y. I.se l Radiation Protection Branch j Office of Nuclear Reactor Regulation j US Nuclear Regulatory Commission
- Washington, DC 20555 0001
Dear Jay:
Peer review of the updated buildma. wake model was held in the Banelk officas at iWashington DC on May 10,1994 The peer review panel consisted of J individuals selected from the list of potential revice iirs verbally approved by you on April 8.1994. The panel memb
- 1. Spickler, R. P. Hosker, J. F. Sagendorf, A. H. Huber, and W. 5. Petersen. The review 5 specific recommendations related to the model. The recommendations are listed at th letter. In my assessment, these recornmendations are reasonsble. Implementation of the recommendations research effort. does not involve shber a major revision of the alternate diffusion sno This letter describes the essence of the peer review. The agenda for the review included e
introductory remarks by Tom Easig e description and evaluation of building wake models in current NRC regulatory guidan e description and evaluation of the updated building waks snodel e description of applications of the updated buildirs uns model
- discussion of building.waka models in general e formulation of panel recommendations Building. wake dispersion models are found in NRC Regulatory Guides 1.3.1.4 and 1.145 and in Murphy.Campe procedure referenced in Sr.andard Review Phn 6 4 (NUREG 0800). W exception these models estimate concentrations assuming that the receptor (or air intake)is at cente of a plume represented by straight line Gaussian model. These models account for. bu thclosure 2 8.
1T)3cLW 0 - . .
_ _ _ _ - - - - - - - - - - - - - ~
\ modifying the diffusion c4 + :lents. Regulatory practice generally limits the re concentradora in building wakes to a factor of 3 or less. The s basic equation IN*
=I,I,u
'b building wake effects, and U is the wind speed at e or10 m.where The exceptien to the Gaussian straight line modeling approach is fo - outward unifor%y in a!! directions whh a speed equ
, calculated with this model are assurned to cover both intakes simuhaneou minutes each day.
2 Regulatory Guide 1.145 adds a second correction to NRC dispersion models / concentrations to account for plume meander when wind speeds are less than 6 comparr.d with concentrations calculated using the is assumed for replatory applications.
, Plume centerline concentrations predicteo by the various NRo building wake e compared i
Asse'.sments". The result of this work, published a ta i ty Control Room Habitability Assessments showed thatithe Ar on or models did not pre yx concemtrations related to changes in building area and .atmospheric It also conditions showed that the models significantly over$redict concentrations at low wind The definition ofI, isThe updated building wake model, wh provides new definitions for I, and I,. , I,=re).o}.P/8 diffusion in building wakes. The increased diffusion in o',e?"fus:r
- u. .
turbulence caused by the building: A, s is the cross section a turbulence
.. scaling velocity in the atmosphere that is related to wind speed, atmospheric 2
_ _ _ _ . _ ~ _ - _ _ _ _ _____ _ _ _ _ . _ _ _ _ . _ _ _ _ _ . ~ ( ] roughness; WEF is and WEF is a wde capansion factor that is a functionef distance fr f i ser = (2 - (2 . p )exp g )) i { i
- The value of a is a function of E,.osp,ei;c stability and su atmospheric stability, she value of a is about 0.09. Sim!!ar expressions, with explicit atmospheric stability, were derived for E.
- Two assumptions related to Zl, and R were made in the Armorpheric Epironment ar j . these assumptions was that any of the standard sets of diffusion coefficient algorithms c i to estimate diffusion in the abseru: of the wakes (ncrmst diffusion). The second assum
} e, is independent of wind speed. With these assumptions and the parameterization for nornu i diffusion coefficients found in most NRC computer codes, the revised building wake mode I shown to result in better predictions of centerline concentrations in wakes than the mode i due largely to better prediction of concentrations at low wiad i 1 ! Commenting on the Atmospheric DWronment article Briggs, et al. (1992), pointed out thatI j , increase in turbulence associated with building waku should be a function of wind s! pointed out that the improved predictions of the updated model at low wind speed be related to better treatment of meander than of building wakes. ] l In preparation for the peer review meedng, the updated wake model was tested using d ' j assumptions. First e was assumed to be proportional to the wind speed. The result was a decre ! in predictive skill of the model because concentrations were over predicted during l! conditions. However, the model was still better than previous models. The updated' wake mod j j { also tested using a more state of the art parameterization for normal diffusion coefficients in 1 to assuming e proportional to wind speed. The results were improved but still not as go i were the peer when the model was applied with the original assumptions. This information was presen reviewers. t l Presentation of material on the updated wake model ended with a discussion of application of i snodel in three computer codes: EXTRAN, R/, SCAL, and ARCON. The discussions of the ! EXTRAN and Implementation wasRASCAL more detailed.implementations ofite model were brief; the discussion of the j During the meeting the peer review panel capressed two related concerns. They were concern ; , about the use of straight line Gaussian models in the immediate vicinity of a release point la a .! . building complex. They were also concemed that most (> 805) of data used for building-wak! anodel evaluation were farther downwind of thk release points than distances typleally of interest in i control room habitability assessmenta. The peer review panel recommendations are: I J a i I i
y ,. '" '
!'e , .
) . I . .*{ 1) The turbulence increment generated by buildings should be assumed to proportional to the j i wind speed in accordance with accepted theory and physical reasoning. .
; 2) j The effects of mean6r durire low wind speed conditions should be treated explicitly in the model, but the treatmeru should be separate from the treatment of building wakes.
- 3) Appropriate subsets of the available data should be used to evaluate the model sher the
- suggested changes have been made.
- 4) An approach to determining concentrations other than straight.line Gaussian models should be l considered when releases are from a building and receptors are on or near the building. A distance equal to three building beighu was offered as a reasonable lower limit for the
( j application of Gaussian models. A paper by Wilson and Britter (1932), and unpublished papers by Wilson and Chul, and Wilson and Lamb offer alternative approaches. j 5) Appropriate subsets of the available data should be used to evaluate approaches for deternuning the concentrations on or near buildings. The current work statement for Task Order No. 3 'Atnespheric Relative Concentrdons in Buildmg Wakes
- under " Environmental Licensing and Regulatory Support' JCN J 2028 does not cover the
! changes to the model or the model evaluations recemrnended by the peer review panel. However. I have received a request for a proposal to modify the Task Order 3 work statement to cover these { changes. I will submit a proposal in response to that request. ~; ' if you have any questions, please call sne. i Si erely, I James V. Ramsdell, Jr.
~
! Senior Research Scientist i l Multimedia Exposure Assessmera Group I EARTH AND ENVIRONMENTAL SCIENCES CENTER ! IVR/Is
~
4 I > i a , $, l. . t 1 4 ) i
.. s. , .
3 . 4 f . j r . OBa11elle j' . Petific Northwest iaboratoriel i Saneliv Soulesard~ P.O. Son 9H i Adland. Washington M332 s
. Selepheae OM) 3/6-8626 j u st n . 2,..
\
1 4 Jay Y. Lee i i ! Division of Radiation safety and Safeguards
- Office of Nuclear Refulatory Research
- U.S.
Washington Nuclsar Regulatory Comunission D.C. 20555 !; , Dear Jays i t ! Attached is a draft letter report covering changes to my 199C buildi j diffusion model that have been made in response to the May review. , 199410 ng-wake peer in the immediate vicinity of the release point using concentrations satinated using minimum-dilutions models.
~ with the revise i
2 believe that the changes to the model are fully respon 4 i recommandations and concerne tepressed by the reviewers.sive to the ' , comparison of revised model concentration predictionsInwith addition, the data and predictions of models indicates concentrations in the 1sumediate vicinity of the release point. that the revised model may be used to tre A comment.copy of this draft report will be sent to each of the peer revi i and J. Fairobeat Copies forofconsnant. the report will also be sent to F. Sifford, Dewers for The final letter report wi . J. teilson, I reviewer's conunants on this draf t along with my response.ll sentain the
- If you have any questions, please call me.
- j. Sin erely, " '
1 i
- James V. Ramsdell
! -Senior Research Scientist
- Nultimedia Exposure Assessment Group
' EARTE AND DrVIRONMENTAL SCIENCE 5 CEN'43R { .TVR/13 i i D
- t I
! D1 closure 3 [ 4 Ar*- ,aC A o DD/,c--h. g .e g n.
e ?, 'I'. - . 'c - - l-a QBa11elle Pacific Northwest Laboratories tattelle Bowlevard P.O. Son 999
- . RecMand. Washington 99352 i .
teuphone (509) 376-8626
- August 23, 1994 l
i A. H. Huber ' USEPA/ AREAL ! MD-56 ResearchTrianlef' ark,NC27711 . i )
Dear uber:
p , i i
- Attached is the draft letter report to Jay Lee that describes the revisions to ;
my 1990 building wake diffusion model in response to the peer review panel recomendations. low and high wind speed conditions.The revisions explicitly treat increased diffusi In addition to covering the revisions to ! i , the model, the letter report covers the peer review panel concerns related to the use cf a Gaussian-based diffusion model for estimating concentrations at , receptors on and adjacent to building surfaces. ) i The Nuclear Regulatory Comission staff would like to use the letter report to ! support a change in its regulatory position related to control room i habitability for submission to the Advisory Comittee on Reactor Safeguards in October. Please review the letter report to ensure that it adequately dispositions all of your concerns and provide written confinnation that your concerns have been addressed. In order to make any final revisions to the model that are necessary prior to the meeting, I need to have the confirmations and any coments by September 12, 1994. Sine rely, James V. Ramsdell Senior Research Scientist Multimedia Exposure Assessment Group EARTH AND ENY1RONMENTAL SCIENCES CENTER
~
JVR/lg - Attachment ec: J. Fairobent F.A. Gifford
. J.v. tee '
D.J. Wilson- ' Peer Group I
, h '* s 't 4 i .. . .-
5/2-i/ s v {* * ' 4 DISPERSION ESTIMATES IN THE VICINITY OF BUl!. DINGS 1 4 i J. V. Ramsdell, Jr 4 C. J. Fosmire l l 4 1 ! l l 4 i August 1994 ' i i Prepared for the Division of Radiation Safety and Safeguards Office of Nuclear Reactor Regulation U.S. Nucleat Regulatory Commission Washington, DC 20555
- NRC JCN P2028 l
i Pacific Northwest Laboratory Richland, Washington 99352 i - 4 ( . i QV1c, v nm n nn_ il ^(
' " % ._ s -
4
, I g~ ,
j CONTENTS INTR ODU CTION . . . . . . . . . . . . . . . . ................. .
. ................ 1 1990 Model . . . . . . . .
- Pee r Re vie w . . . . . .......
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . .
Experimental Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3 REVISED M ODEL . . . . . . . . . . . . . . . .............. . . . . . . . . .5 . . . . . General Form for Diffusion increments . . . . . i Low Wind Speed increment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. s High Wind Speed increment . . . . . . . . . . . . ............... 6 3 Comp'et e M ode l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 M odel Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 i
..................................... 11 4
i ALTERNATIVE MODELS . . . . . . . . . . . . ............... . . . . . . . . . . . .13. . . . . . { Wilson Chu1 Model . . . . . . . . . . . . . .
- Wils on. Lamb M odel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14. . . . . . .
4 Comparison of the Revised Model with the Attematives . . . . . . 15 ......... i ......... 17 1 i NEAR FIELD CONCENTRATION ESTIMATES ............. . . . .19. . . . . . CO N C LUSIONS . . . . . . . . . . . . . . . . . . ............ . . . . . . . .21. . . . . . i REFER EN C ES . . . . . . . . . . . . . . . . . . . ........... . . . . . . . .24. . . . . . . i t j i i i k 4 I . l - 2, e o I 1 I J . J k
- w. 3, s
^
1
).
.i ' ' FIGURES s Figure 1. Increase in lateral turbulence as function of wind spee'd . . . . . . . . Figure 2. . . . . .. 8 Increase in vertical turbulence as func6on of wind speed ....... Figure 3. . . .. 9 . Comparison of re WWue . . . . . . . .vis&d model concentration predictions w*th observed i Figure 4. Bias in Murphy Cam 11 wind speed . . . . . . pe model concentration predictions as a function of i j Figure 5. ........................................ 12
- j Blas in Regulatory Guide 1.145 model c 1 function of wind speed . . . . . . . . . . . .oncentration predictions as a i Figure 6. . . . . . . . . . . . . . . . . . . . . . . . . . . 12
- Blas in revised model i
apsed . . . . . . . . . . . . concentration prediction as a function of wind { Figure 7 ......................................13 i Comparison of cumulative frequency distributions of predicted to i
- observed concentration ratios for the Murphy Campe, Regulatory 1
i Figure 8. 1.145 and revised models . . . . . . . . . . . . . . . . .. 14 . . . . . . ....... ! Comparison of Wilson Chui m < observed values . . . . . . . . . .odel concentration predictions with Figure 9. ................................. 16 s Comparison of Wilson Lamb observed values . . . . . . . . .model concentration predictions with l 4 Figure 10. ................................. 17 < Comparison of revised model concentration estimates with values in the building surface data set . . . . . . . . . . . . . . . observed Figure 11
. . . . . . . . . . . . 18 Comparison of cumulative frequency efistributions for the ratios of -
predicted to observed concemrations for the Wilson.Chul, Wilson Lam and revised models based for the building surface data set 18 Figure 12. Ratios of predicted to observed concentrations for the Wil model as a function of normalized distance . . . .. .. .. . .. .. .. ... .. .20 . son Chul Figure 13. Ratios of predicted to observed concentrations for the Wil model as a function cf normalized distance . . ... .. .. ... .. .. ... .. . 20 . son Lamb Figure 14.
- Ratios of predicted to observed concentration
. a function of normalized distance . . . . . . . . .s for the revised mo
. . . . . . . . . . . . . . . . . . . . . . 21 Figure 15.
Cumulative frequency distributions of the ratios between predicted an observed concentrations for the Wilson Chul, Wilsor Mmb model for all data near the release point . . . . . . . . . . . . . .,.and revised
........... ~22 l
E m41
l' , ' ' l
- s j j
i
-( -
DISPERSION ESTIMATES IN THE YlfdNfTY OF BUILDINQS *
- INTRODUCTION j
Control room habitability assessments and the avaluation of the consequence { j design basis accidents invoNo estimating dispersion of emuents from building.t teve' stacks. Nuclear Regulatory Commission guidance to staff and I;oensees d i several acceptable methods for estimating dispersion in building wakes, for exai , Regulatory Guide 1.145, and the Murphy Campe procedure (Murphy and soferenced in Standard Review Plan 6.4 (NUREG4800). These methods attem the concentration at the conter of the plume down'vind of the release point u based on the straight line Gaussian model. { { According to the Gaussian model, the concentration at center of a piume is 1 j x/O - - u e,e,U (1} f where x is the concentration, O is the release rate, o, and a, are horizontal and vertic 1 i diffusion coefficients respectively, and U is the wind speed. Diffusion coefficien
- .' the effects of turbulence and are generally estimated on the basis of atmos distance using empirical relationships derived from axperimental data. Building j
diffusion models have the same form but represent the effects of the wake by diffusion coefficients. These models are typically written as z/O - ' ' sI,2,U (2) j where I, and I, are diffusion coefficients corrected for building wake effects. , l Plume centerline concentrations predicted by the various NRC bu!! ding wake m j were compared with experimental data under NRC'JCN B2970 ' Atmospheric Diffusion f Control Room Habitability Assessments'. The resutts of this work, publishe l 5055 Atmospheric Dttfusion for ControlRoom HsNtaWitty Assessments (Rams , ahowed that the models did not predict variations in concentrations rela'ted to c building area and atmospheric conditions very well. It also showed that the models k s algnificantly overpredict concentrations at low wind speeds, J I i $ L i'
- e. 1 l
l - l
0 >*- i- . I
- t .
1990 Model { t -
- 2,. TheM updated definition ofIy is building wAs model (Ramsdell 1990) provides ne ,
l 4 } 2,=(c'+Ao,')* i N) a i where o, describes the diffusion in the absence of ap describes buildingthewake, and i increased diffusion in building wakes. The increasedy diffusion i A o,' = ** ! u.' (1 -(1 X+."*-)exp( *VX )) (4) j where k is a constant with a value of about 1; Ao { m is the increase in the horizontal
- oomponent of turbulence caused by the building; A is the cross section I building; u is a turbulence scaling velocity in the atmosphere atmospheric stability, and surface roughness. ,
{ { The expression in brackets on the right side of (4) controls the e } plume as x, the distance from the release point increases. M { to 0. For x less than 0.2a/A", the expression increasesnapproximate , x is I greater than Sa/A" the expression has reached its maximum value is a proportionality constant between u. and U and is therefore stability and surface roughness. For near neutra! stmospheric stabi about 0.09. Similar expressions, whh explicit dependence on atm derived for I,.
- Two assumptions of notey related to I and I, were made in dev the 1990 model. The first of these was that any of thee cstandard ent sets algortthms could be used to estimate diffusion in the absence of the w diffusion). The second, and more critical, assumption was that e,ls speed. With these assumptions and the parameterization for normal tund in most NRC computer codes, the 1990 model resufts in better p centerline concentrations in wakes than the modsis described in cu .
Commenting on the 1990 model, Briggs /et al. (1992), point out t turbulence associated wfth building wakes should be a function o . i t
{ .i. s?. ).- *
- i -
[;.
* . p.t.1 ett! . hat the improved predictions of the updated model at low wind speeds ar
- ,(.
t +' ? 'n rotated to better treatment of meander than of buildig w.'cas. 3
. %r Review j '
A panel wd cor:vened in May 1994 to conduct a peer review of the T;ie primary 9 commendations of the peer review panel were: . 1) j The turbulence increment generated by buildings should be assumed to prop i to the wind speed in accordance with accepted theory and physical reaso ! 2) ,
- The effects of meander during low wind speed conditions should be tre d
in the model, but the treatment should be separate from the treatment of wakes. . ~ 3) An approach to determining concentrations other than straight line Gaussian mod should be considered when releases are from a building and receptors are on or nea the building. 4) Appropriate subsets of the available data should be used to evaluate the model afte the suggested changes have be6n made. This letter repert describes the disposition of the recommendations of the peer pawl. In addition to this introduction, the report consists of three sections that direct address the peer review panel recommendations. These sections describe the model rettsions and compare the revised concentrations predicted by the revised model with measured concentrations and with concentrations predicted by attomative models. The firs , of these sections discusses revisions to the model to separate the effects of low a wind speed phenomena on diffusion. It is followed by a section that discusses the issue attomative models in the for situations in which receptors are on or adjacent to the struc trom which the release occurs. Both sections include discussion of model performance which model predictions are compared with observed data. The last model evaluation section compares concentrations predicted by the revised model and two attomative mode with measured data that are particular'y appropriate for evaluating models used for control room habitability assessments. Ernerimental Data . The two data sets used in evaluating model performance contain data collected in' field experiments at seven different reactors. Three of the seven reactors- the Materials Tes Reactor-Engineering Test Reactor (MTR) (Isittzer 1965), the Experimental Breeder Reactor Il
. 3
J ,. , i . a e
- I-
- l.-
{ (EBR 2) (Dickson et al.1989), arxi the Experimental Organically Cool i (Start et al.,1980) are at the Idaho National Engineering 1.aboratory. Th j are the Duane Amold Energy Center (DAEC) in Iowa (Thulliier and Mancuso
- 1982), Diablo Canyon Power Plant (DC) (Thulliier 1988) and Rancho Seco
{ Station (RS) (Start et al.,1978) b Califomia, and the Three Mile Island Nu (TMI) (GPUSC 1972)in Penemytvania. - j I The Srst of these data sets, referred to as the ground 4evel data set, consisl { . l concentration measurwments made at regular Intervals on sampling arcs from ' 3 i tom ground 4evel release points. The maximum concentration on each a t k the best approximation of concentration at the conter of the plume as R cr i all, there are 379 data points taken during 131 release periods in the groh4 { { Meteorological conditions during the releases ranged from extreme class G) to extremely unstable (stabilNy class A), and wind speeds range { ! meter per second to greater than 10 meters por,second. Of the 379 dats po 3 253 represent measurements made with wind speeds less than 4 m/t,208 d represent measurements made during stable s'tmospheric conditions, and 138 i represent measurements in low wind speed, stable atmospheric conditions. Conce predicted for low wind speed, stable atmospheric conditions are generally used f I of consequences of accidental releases in control room habhability assessme site boundary. In some cases the actual release point was offset from the cente) ' sampling arc. As a result, the range of distances of the data points in the ground set is 8 to 1200 m. The ground 4evel data set is described in earlier publications (Ramsde The data have been used as reported except that stability classes have been
' tow cases where 1he stability class determined by the NRC detta T method wa with other reported data, for example wind speed or season and time of day. Thl <
modifications typically involved changing extremely unstable or extremely s; classes to more nearly neutral stability classes. Neutral stability was as,sumed experiments in which the wind speed exceeded 6 m/s. i The second data set, referred to as the building surface data set, consists of 26! concentration measurements made at locations on and adjacent to buildings at Rancho Seco, the Duane Amold Energy Center, and the EOCR. Data from both ground lev . 4 1
j 8 - ! t t {' h elevated rele'ase poirge +.G% *od in tM dsta set. Meteoro l . cover the full tangrist std:2Gy Md wint speeds from less tha More than haff of the fata poir#s represent measuroments in lo j , stable i conditions. All of the measurements were made less than 1 ,
. I The samplers were not arranged in pattoms that e ood would en !
j that the concentratien at the center of the plume was captuted at s ance. As a result, l these data are not appropriate for use in developing er ne a model to ! concentrations. However, the data can be used to evaluate rm . centerfine concentrations predicted by a model should tend measured concentrations. REVISED MODEL. The first two recommendations of the peer review group ha 1 y revision of the 1990 building wake model. The revisions
~
rectly add a d related wind to wind speed and modify the existing increment speed conditions.- ' ow to be Revision of the model starts by redefinition now definition ofI, is ,y of dWusion ,. The coeffi 2, = (o,'
- A o p' + A oye)w s (5) .
where op represents diffusion from a point source under normal g represents conditions, A an increment to diffusion associated with low wind speed phenomen y, represents an increment to dWusion associated with high wind speeds. ,. To maintain continutty with existing regulatory guidance, th define the diffusion coefficients in the NRC sumed toPAVAN be a rpplicable for a, and o,. The relationships were developed ir9t (1968) and Tadmor and Gur (1969) as approximations usion to the Pasq coefficic:1t curves. The relationships have been extended wng to i the guidance in the February 1983 reissue of Revision u e 1.145. 1 of NRC R General Form for Diffusion incremgf113 . Derivation of expressions for Ap '
- p. Aoye, and the corresponding increments to vertical diffusion generafly follows the derivation of thea wake diffusion increm
. t 5
. . . , - , _m,
. . i '
' i
- i. -
t
... > del. The derivation starts by .asuming that some phenomenon or combination of I
' phnnomena causes an liusase 1.1 turbulence above the turbulence implicitly assumed in th j
normal diffuslen coefficients. At low wind speeds, meander and poss'bly uneven heatin building surfaces may be responsible L; increased diffusion. At high wind speeds, the mechanical turbulence associated wtth wakes is responsible. The senond assumption is th i i the effect of the turbulence increment on diffusion decreases expone^tially as a function o j time relative to an appropriate time scale. f Whh these assumptions, a horizontal diffusion coefficient increment is defined by " i l Ae,'=2f[Ao,8axp(-t/T)R(t)dedt (6)
^ where Ao, is the .ncrease in the lateral component of the turbulence, t is the time since release, T is the time scale, and R(t) is the autocorrelation function for the turbulence. .lf f is assumed to have a constant value, k, near the release point, the double integral in (6) j the solution l
4
~
l A o,a = 2k4 o,'Tagj _(j ),xp{. )) , M i .' 4 l The first part of the expression on the right side of (7),2kAo,8 T 8, determines the maximum l increment to the diffusion coefficient. The second part of the expression, which is in brackets, determines the fraction of the maximum increment that is applied as the time (or [ distance) increases. Note that the terms in brackets in (4) and (7) have the sa , } term is zero at the release point and asymptotically approaches one as the distance { increases. Two consequences of this behavior are 1) that the model does not predict instantaneous diffusion at the release point, and 2) that the I, approaches e, at large j distances, it is not necessary to apply arbitrary limfts to the model to svold unrealistic
. asymptotic behavior either near the source or at large distances.
Low Wind Somed increment The relationship in (T) is general and may be used to define Ao,,s, 3,y,s,3,g,s, and u Ao ' provided the appropriate turbulence increments and time sesfes are used in each case. For the low wind speeds, a time scale of 1000 s has been chosen for the horizontal turbulence increment. This time scale,'which is larger tha. . o 5 :sfes usually associated with mechanical turbulence near the ground, is consistent w Ume scales associated with i 6
^
g , t the transition from diffusion coefficients proportional to time to dif i aquare root of time. proportional to the ' ) Two time scales have been chosen for light ent during the vert winds. The time scale for stable conditions is 100 s. This y onger than the inverse of the Brunt.Valsala frequency for the temperature l apse rate separating stability ! classes D and E as defined in Regulatory Guide cyis 1.23. T undefined for neutral and unstable condnions. ns is T assumed to be zero. As a consequence, the low wind s { aero for neutral and unstable conditions. Hioh Wind Rosed increment For high wind speed conditions, the time scale for decay
~
1ncrement is a urbulence i i T.E u, (8)
, where A is the cross sectional area of the estructure ur ulence, and generati u* is the friction velocity upwind of the structure.
turbulence increment is The time e vertical T.- A u,(2 +a/L) where z is the wind speed measurement height and L is the Mo ength. In unstable conditions, the time scales for decay ofcethe horizonta incromonts are about equal, while in stable conditions the time sca or decay of the vertical turbulence increment is about a factor r ecay of 2.5 less than of the horizontalincrement. This stability correction is discussed in mo n e description of the originalmodel(Ramsde!!1990). Turbulence data published by latitzer (1965), , wa and Dickson e Oikawa (1982) have been used to develop parameterizations for e ncrements during high wind speeds. Changes in o, and p, were computed as ' 7 O
. ** a1
. \
1
-1 .
t { ( ao = fo/-of . (10) where the u and.d subscripts represent upwind and downwind,/espectively. 4 Instances, the subtraction resulted in negative dmorences. Thes; differenc } considered to be unreal an were set to zero. The resulting turbu'ence I against the square of the upwind wind speed in Figures 1 and 2. As i, review panel, these incrernents are functioris of wind speed. The between the turbulence incrernents and the square of the upwind wind ' 0.01 for aon and Ao n . respectively. These noemeients are dirnensional with units of seconds / meter. The corrstation coemeient for the relationship /4nwn in F for the relationship shown in Figure 2, it is 0.81. 1 2.5 1 . - 2.0 ' j -
*/ --
e j s j 1.5
/ - -
*1 -
30 . 0.5 , : 7 e - Dickson, et al.1969
*h" t b - Ogawa & Oikawa 1982 n t c -inirtret 1965
- .. ! n '
0.0 . . 0.0 20.0 40.0 60.0 80.0 100.0 120.0 Wind Speed Squared (m/s)^2 Figure 1. Incrosse in lateral turbulence as function of wind speed. e m g 8
~ '
i . . . 1 ! ,1 - '2' l ! 1.00 , $ f o.so 1 . g, ..
- f. 0.80 o.40 *- .
4 m
- i 020 ^
, a- , et a!.1969 i , ~
, b- Ogawa and 04swa 1982 i
*
- c-latteer le55 l
;* .;e
- o.co .I . 3 '
c.o 20.0 40.0 80.0 30.0 { 100.0 120.0 Wind Speed Squared (m/s)^2 - Figure 2. Increase in vertical turbulence as function of wind speed. With these relationships, and the assumption g that Ao and ao,,, are equal, remaining unknowns in the model g are k, Ao . Values for these parameters we ' by minimizing the function P = EWJiog( )g-log ( )J (11) where K/Q),,,,is the normalized concentration predicted by the model, K/0) is the normalized concentration measured in the ground-level data set for the same W, is a weighting function. The weighting function is equal to 1.0 when the model overpredicts the measured concentration, and equal to 2.0 when the model und Optimum values of kgand Ao could not be determined almuttaneously. H search for an optimum value of Aog converged rapidly when a value was specified Optimum values gof Ao were determined for k equalto 0.25,0.5,0.75, and 1.0. T of P were not particularfy sensitive to the value of k, but they were sonsttive the parameters that appears in the low wirid speed diffusion increments. The op for the product is approximately 0.5 with lower values of P being associated with
, 9 ~ . - - , , -, . . - r -
7 .: ' i.
-
- I
!' of k. Uttimately, k was set to 0.5, and n the optimum valu The. search ulgorithm may have talled to converge on an opt . . ecause there I 4 are Camelete Model there are relatively few data in the data set that represent ons. [ dop a, gn 3 aelectiona
, gg y,s. and Apg aare:
of values g for k and'Ao completes the . model rev ons for ! Aop i 8 = (2)(0.5)(0.7)'(1000)sgg .($ .1),,(_ t ))
- 1000 1000
! * * (12)
= 5.0x10'(1 -(1 + -
- 10000)exp( 1000U))
I A no ' = (2)(0.5)(0.7)'(100)*[1)exp( -(1 + }) g hW-
- 5.0 x 108(1 -(1 + -
1 )
. Ao n
; t ye = (2)(0.5)(0.02U8)' u **
4/u. A {1-(1 +1)exp( 8/u.)) (14)
- 4.0 x 10db(1 -(1 + E)axp( ~"*))
aa g g . and Ae a A n = (2)(0.5)(0.01U8)' - u.a(2+A)'(1-(1+ u m+3)t),,pg-u.G+AM)) 4 4 1.0 x10d A (15) as (2+2/L/[1-(1+bO)exp( 2+&)x)y 4 % , respectively. .
= 0.5 and Figure Ao 3 compares the predicted and measured concentration withk n = 0.70. The median ratio between predicted and maaiured 1.25, and alightly more than 85% of the predicted concentrations are the measured values. .
actor of 10'of t 10 m - - . ..,,.g.... . . . - - - , . . . ~ - , , . _ . - - -
1 . i ) - , ... 1E+02 j ('- 1E+00 I e { 1E 02 \ y M'/. ' -K 1E.o4 '.- 1 fa@iiFd C l ] ,,, j ^ f l i
- 1E 08
/
I j 1E.10 . 1E.08 1E 07 1E 06 1E45 1E 04 1E 03 1E 02 { Observed X/Q 1E-01 l Rgure3. } Comparison vabes. of revised model concentration predictions with obs , i Medal Evaluation Early evaluation of the current NRC wake models indicated that the mod
- significantly overpredicted concentrations during light winds (R j is shown in Rgures 4 and 5. NRC guidance related to the use of thes the models should be used do determine X/O values th the time. Typically these highest values are associated with wind s .
Mgures 4 and 5 show that the current models almost always ov an order of magnitude when the wind speed is 1 m/s or less and that predict concentrations that are about 2 orders of magnhude too high. The revised wake modelincludes corrections to the diffusion coe addressed to improving model performance at low wind speeds. Mgure 6 variation of ratio between predicted and observed concentrations of the r
- function of wind speed. Compared with the current NRC models, the r t
11
. . 1E* 04
; -- - x-4 m...__.
7__ m s - pg____ _, ,
. i . ,
1E*C3 1 ;"=.,3 ,_ _ , ,._m,37 73.h. - = ,ug g =r ,.m s.. l
=4 -
_n r t mg ;,
- p. .
9, ,, g.%
. 1E+02 .. ._ - . . . j
-mA= 8 ----..--a
=_=; .:.e.= _=. 3.
, _ + ~-g._ p.c.tn.% % .g--. .
, _-....;- - . : :. -. :::.:. e_,a . .
\
. .~.
1E+01 * * .
, s = =p,m . -- g.s'
;-4e..., *e-4 . -- ,
1E+00
* '.- - #'2' r
_ ggI *.7 - q- -- _' -
. > = =
^
i
.n n .
g77 .- , 1E 01 * '
. n._
----_...= . _m m
-._= ='. _ . . .
E F m,= , 1E 02 m_ =m x--. ____.=y __;_.___, _ m.t . y m_. n% 1E43 . _ . . . ._- _ . . .. . . . . 4 0 2 4
' 6 8 10
- 12 Wind Speed (m/s)
Fi2ure 4. Biasspeed. wind in Murphy Campe model concentration predictions as a function 4 4 ,. 1E+04 ---_ m.m . 1
- _ _ - _ _ - m-c- ---
- - m ._m_..
1E+03 4. m., .y ._ m . . . _ _ m .% j.- m3 _wg g . _.,
..........,y
=
y sa 1E+02 , . -. .._ gg_.._.. . - . . . _u
._,, - m._ _ _gg _2: mg gg .
# .c
, 1E+01 '" ^ *^*#*
m n.:_~.?A fkg+.tlIM.c?M::::t =m- L_ik.=ai1;Le_-tsusat WL=miMi-g;.
* .:. Jg; m_ .
c.
- i # -
..** t.__a
- 1E+00 .
s mpg _qq - % _. ; _%.-.--w-
.n-
_.e w_.2
, __%.-- e,_,
1E41 _ - - - . . -s-
--b _ _ , , - -.*7--___.. -[ ---
Y ma: .gmg a: 1E42 . - - w .=- .w._ aw:_ m . m_rm --
-_ 2
%- . -,.: : ..~.--": .
am= . 1E43 _. ... 0 2 4 4 8 10 12 Wind Speed (m/s) Figure 5. Blas in Regulatory Guide 1.145 model concentration predictions as a function of wind speed. 12 w -
. ,o .
5
)
, , 1Eo64 ~
) .
j
= z @ - =s_; ~ ~ =-c
- -.-
- s y 3. w.__
{ 1E+03 i _4 l 4
. .n
= -
- ~ .
=
{ ' 9}1E+02= l I i
~
~#. .ra, A_ a ek
-~. .-
=-- - - - - m W=rsm= m.r, usi
. _ - p 7.
i
- _
1E+01 ^#" '# * -' '
=== a __:u e iw; c n,, - gr3 =-
-_x w--
. _ w: .4 5wm .
3
,g.go
" ? ~~1 3 # . a = u_*. i'm . . ' ^
~
^
, M.>jpfsPM-97 _.- R- c7 -~ ! _v- : , g.
=
i@r5 1E-01 * -, , 2__,,_ i e e- =.= =w p? _ - - -
^
w~u
**T
- g "_ w* =
mm4 i . _ j 1E42 pn= - j ; 1 y _ = m._.a _. 1E ...
-~:-~;r~. 1 :
c 2 j 4 8 8 10 1 12 Wind Speed (m/s) i Figure 6. i Blas apsed, in revised model concentration prediction as a fun ! t j j tendency to overpredict concentrations at low wind . oes appear that speed However i
- there is still a slight tendency toward overprediction for wind speeds l 1 ss than 3 m/s. The
{ improvement in model performance is further litustrated y in t j ' distributions shown in Figure 7. The median ratios between conce predicted by current NRC models and the maximum observed concentration . j addition, concentrations predicted by the current models e of ar the observed concentrations only about 60% of ythe . time Figure 7 clear shows'that the a ', Improvement in model performance is gained by reducing mode i algnificantly increasing underpredictions, ons without 1 ALTERNATIVE MODELR { The third recommendation of the peer review group e was to [, a Gaussian plume model for estimating concentrations
. wo in the vic minimum dilution models were identified as potential altamat t
I 4 13 1 i 4 A t 4 4
,- ,,n..,
yms . , ~
- 2
. g 3.j y ::.ut:3.i T.T:M:U;;3.i:i:C" ." *n i;; n.
j,".j. : : ."
- y. ..;;.;] .; =i ;;;;;dib.. . .. g I
- (., 1E+c3 - ' ' '
pig.;e.s :yie : .ms " - - --- - = m.em ' 'i 4 , _:il:esurw.e _an n=n=a: i n =u :- =a. g 1E+02 indugg; I I ' l I
--N.
I 4
== .imra 1;;.Reguistory
- - - = - ! :s.:.mumn: ;;;g;.i.e .'
.. .e a==
a.auy==.1:ci - Guide 1.145 ! s< 1E+01 t =a'ce.s s=:e'un:ue w'as a: q" p_:' ' i an +g pc 2
=.+.3 =a=g ,,,,,-=.'=g.'
' /4 '
! t
-uwn 77 1E+00 [*5diihiid
.g n #g7 #
i uniestii:liiii l ' ' a mamm .jgg.i:inig
,.y -ay ;;migagg gg Mig 3
n . s . .
~-
1E 01 p g.'~'.mm #
. mam'=r;i.w2es "~
"" rCampe I ' '
-a 1 3
- m. __,
y= - g=ia u.- s = .- i , , , i 1E 02 j g am ,;gyv .[y_m bg m;d;= g;,j;es ,d. . ._ j_ m-._m._ j I 1E 03 I I ' ! ' i I ' I ' 0% 10% 20% 30%
- 40% S0% S0% 70%
30% 90% 100 % , Cumulatve Fregency
- Figure 7.
i Comparison of cumulative frequency distributions of predicted to i . observed concentration ratios for the Murphy Campe, Regulatory Guide ! 1.145 and revised models. i { model. Minimum dilution models attempt to provide a lower bound to dilution. When
- i minimum dilution models are used to estimate the concentration, they should estima j
upper bound on the concentration. This is in contrast with typical diffr%n models which attempt to estimate the average concentration at a position in the plume, j The first attamative model (Wilson and Chul 1994) predicts maximum concentrat ) as a function of wind speed, building area, and downwind distance, and the escond m (Wilson and Lamb 1994) predicts maximum concentrations as a function of wind direction tiuctuations, and stretched string' distance. Stretched string distance is the m } distance between the release point and receptor without passing through the structur models have a correction term for initial dilution for releases through stpeks and vents. ! Wilson-Chul Model - The Wilson Chui model was developed and tested using wind tunnel data to estimat {4 minimum dilution in plumes released from building vents and short stacks'. The model j development does not make or depend on a Gaussian assumption. When reformulated ,t 14 4. ? . a
estimate modelis concentrations assuming squal effluent and ambient air de { \, 4 WQ) '
" [.K+DDf
( - (16) i where F, is the vent Sow, D, is an inhial dilution, and DD is the dow 4 dilution, which is a function of the rato between the exhaust is peed, exh vel D. = 1.0 + 7.0( / i (17) j where W,is the exhaust velocity. The downwind dilution is given by 1-DD = 0.25/UX( )* . ' . (1e) 1 where is the stretched-string distance. {! Maximum concentrations were predicted for each of the 265 conce i bu!! ding surface data using the Wilson Chul model to evalusto mod t j vicinity of full scale structures. The results of the calculations are compa $ observed concentrations in Figure 8. The Wilson Chul model consiste
@bserved concentrations, as expected. No observed concentration
[ s were higher than the ! model predictions. Therefont, h is reasonable to conclude that the mo upper bound to maximum concentrations in the atmosphere as well as in Wilson l.amb Model . l The Wilson Lamb model was developed using Sold data rather than . j The form of the Wilson Lamb model when reformulate I is identical to the form of the Wilson Chui model, Lt., Equation (16). j the models comes in specification of the initial and downwind dilution. In j . model inhial dilution is given by Briggs (1975) relationship ' 1 j 1 D,1.o+13.o@) .
, (is) i i
i 15 t_.__ __ . _
- . ., S. -
- d. La i : ,
)
!, 1E41 _= -_ _.gg _
. , i . :. ; , i. n, , , i i., - - - - "
1E+00 r#':EkMMTH
.w-_
-_y _ , , _ _
i ' ' ' ' '" 1E41 -
-m m y .- n' . whuy' y w
_ .__- m =: .: - ! D *** !* M 'UA N T R 1E42 - ; ' m' '=' #u.
- I*' ?' II" ^ ' ' ' ' '
w: g i 1 4 7 g. -A. - wa.essi gigwa.aain r i _ t 6 *et ' 4 6 g seg 8, D *e g -,,_ g ,,, , gg.gi f, e s ese j ._
--- p ., - - - -
m,=; : ~ = .; .
;. . 9 .- m
$ . s. / gg a 4 l 4 6 a ei 6 ,aa,ea4 4 e e a a ei. f. 4 ea4 see a . 6 aa+66 _m., - _ _ _ _ , r._ - __y ., _ _ __ e j ""
, _. f ,
I I lIOel l l 6 0 $Il8 [! 0 $ l 3004 l I 8 9 01II
- 0 $'43
- n. _ ,m =_ - - _ ; .u f ,,
I 4 1E 06 --,'a' p' ' " a w' 'a' m'" c ~' ' ' '"~' n" - ---' a' +' w' mm' ' ' a'm' 'u'm' ' i jg7 8
. iet i i ii: lli t i ! i lli i i '
i i' ' j 1E 07 1E46 1E45 1E44 1E43 1E42
- Observed X/Q i
i l Figure 8. Comparison of Wilson-Chul model concentration predictions with
' observed values.
a The downwind dilution is given by DD = Ms (20) where $ is related to the o, the standard deviation of the wind direction fluctuations (in radians) by p c.oss + o.17 . . (21) If o,is not available, the ensemble mean value p = 0.089 may be used. Note that in the Wilson Lamb model the maximum concentration is not a function of building area. . Maximum concentrations predicted by the Wilson Lamb model are compared with the measured concentrations in the building surface data set in Figure 9. The relationship totween the Wilson Lamb model predictions and the measured concentrations is similar to that between the Wilson-Chul model predictions and the measured concentrations. The model generally cverpredicts the concentrations and provides a reas' enable estimate of the maximum concentrations. ~ '
'(
16 _S- -m - --y-, - - . - . - - - .-,,7
! , .. s . . g. . 8
?-
. ,., 1E*01 I y - A -
-- y _ ___ _ __ __
=- _
__g _._.__ gg ) ( 1E+00 dM" .- hhhdYb
.- F-% C -
.d M ydhid 1'E.et ,,,,?
'['.i'4 4 '
I, 31. ; ; ,,; , , ,
. . /.
~ww
--,a i
k1E42 i,'N. ". , n f . . ".. J .!2 L. 2 *-
+ +
- u. .
jg,, g
= ca rya _ggy
- ~
ggg g . k ' j .as .2 1E43 1 2' -
*'9t/M-M$h_y h 1
i - 2-4 t 1E44 MnuNuusiN '! , ~ NNN
- 1E45 = - , - . , ..n , , ,
...e f .ino
+ + '"
- w. g- -
7# -- ~ _ 'i'i, 1
=
1E46 9
, i;n.
j 1E47 > , , i;in. i i i oi- i i i i ,' '
- 1E47 1E.06
< 1E45 1E44 1E43 1E42 t observeexc ! Figure 9. j Comparison of Wilson Lamb model concentration predictions with i observed values. i Comnarlsen of the Revised Modelwith the Attamatives i Figures 8 and 9 establish that the two minimum dilution models provide reas I ~ estimates of maximum concentrations on and adjacent to buildings in the vicinit point Neither modelis a Gaussian plume model. Figure 10 compares concentration i predicted by the revised model with the set of concentrations used to evaluate th i dilution model. The pattom of revised model predictions is not much different than t pattom for either of the other models. The revised model underpredicts eight c r in comparison wtth no underpredictions by the Wlison Chui model and one u { by the Wilson Lamb model. However, the revised model abould underpr {' than the minimum dilution models because the revised modelis designed to predi { average maximum value in a plume rather than the absolute maximum., j , Figure 11 compares the predictions of the three models directly. The revised m i most likely to underpredlet concentrations. Yet, only about 3% of the concentrations a i . . j . 4 j - t ' 17 i '
-e - --g- - . . . -- -
. . ,***).'
, 1E+01 ,,,,,, _._,
. m_m , ._ , w m
- - .m._ . ,y. m % 3 1E400 " ' "
. - - ' =w' '
m m .m. m' w", ,, w a - --- n = n mu"a.
=- mum _. _ r.:=r :
_a ( 1E41 me.' . e c :. . ' " '
.'s s
--~--t'-._ .
ry v-m-g. -. - 'u . r- m'u _ p' P _, _ ~ , . % j. ;g R 1E42 ' ' ' ' * " ' ' ; '#
- ?.l.~~ E
"' ' -'"'# ^" '# ^^' '
I ' ~[ - f
- Y M ' W M M..*' @ "*- j .t 3a& 57 f M' M .
--n".
1E43 ' ' '*^ . .
~' " 'g ' o$ -'.;e:en.
*^"i'""'
m- -
.- 7g ,- - -_ a -
"" ' 'tu
1E44 ' ' '***" g
' _ __- --1 ' ' _ ' ' " ' ' '" "
-. ' _' 't.a 1E45 ' ' -
i F===.<" ' '" _m ,'_'_'.m'" . _ _ ' - .'_'.'n"
... .'_ , ' m' 'u' g' "
1E46 - e i: 4tui f, i i ii ni y _. i $ . 6 i,i 1 i i s iim i i 4,,,; 4. t. _.myr 1E 07
' i ' ' ' ' * ' ' ;i'
.! 1E47 i l ' ' ' ' 1
' 1E46 1E45 ! 1E44 1E-03 1 1E42
- Observed X/Q Figure 10.
Comparison of revised model concentration estimates with observ values in the building surface data set. - 1E+05 p.
== . . .=== :=. === . .
. - - _ _ 2.y .= -
. .m.a q
, 1E+04 3' R'
mm.- n+
==
A S_ f.r:..
..=
x 1E+03 ,r
=.=
y mg .. i d 1E+02 a
= Wesor4hul
!:i.
b =!=i= Wdson Lamb
. .- u -
..a= =: ==d= ~
1E+01 i ? _: .
!!r m .me.
-{[
f g :u1= =:== J
- ii :ui ' ' ' !::u:n:::: :::g:::
k , dM . . . . .j ..
=m.... -
y a_ Revised Model
-- ~
.w
.,.,,.u ,,
1E41 ,
..m. .. . . , ,
.c,_....m_.
= y .m.y .y a, .m...m.
gg. . . . . ., 1E42 ' 0% 20% 40%
~
S0% 80% 100 % Cumulative Frequency Figure 11. Comparison of cumulative frequency distributions for the ratios of
- predicted to observed concentrations for the Wilson Ch'ul, Wilson 4ania and revised models based for the building surface data set. '
18 G e m* , _ .
i j' -
)
) ...- i . I ( , i underpredicted by the revised model, and only 1.5% are underpredicted factor of 2. With the exception of four concentrations underpredicted of four, cumulative frequency distributions for the revised model and are nearly identical. NEAR-FIELD DONCENTRATION EsiMATE5
- l The fourth recommendation of the peer group was to evaluate the .
i an appropriate subset of the experimental data. Figures 10 and 11 prov ! indication that the revised modelis useful for estimating concentrati 1 j building surfaces, even though the Gaussian model may not be stricu purpose. As a final check, concentrations in the building surface data were concentrations measured near the release point in the ground-level release ! resulting data, consisting of 402 concentration measurements were comp predictions as a function of distance from the release point. l Figures 12,13 and 14 show i l ratios of the predicted to observed concentrations for the Wilson Chul, W j ! revised model, respectively, as functions of normalized distance. In t level release data, shown by the near4 eld markers in these figures, ! is the downwind distance divided by the square root of the building a { . building surface data set, the normalized distance is the stretche j the square root of the building area. The ratios shown in the figures indi models are conservative near the release point because they tend to overesti i { concentrations. ! j . ! i s - - 4
- k. .
i
- 19 '
i . 4, 4 J
'i .
i
' )
i ' ' ' 1 1
, 1E+05 ' !
- i i
. Neer Nid . Side Surfaces
! ' ' (- 1E+04 .
. U_ - -
* * *
- t 1E +03 ' * :' ^- * ;
' .; 4 }
i a **
- l ' 7 e, , . .
- ^ s .* * . i i 1E+02 I #^a*,b b'* ^ I. -
I
- i' .e a "'
4-g '8 4 ,, .**> 1 ,.g
- d:j = *% 1 * ,
e ej 1E+01 ^ ^ ^ b l' i
^ ^ ^*.' '** o. . = - r A 2's.q :
as .
- "a 4 - *l p *
a q ** . g I 4 1E+00
* . , p** -
*l .
! e . l 1E41 . e 1 i - i 1E-02 I ! 0 1 1 2 3 4
- 5 4
Normalized Distance j Figure 12. - 1 Ratios of mreciicted to observed concentrations for the W model b a function of normalized distance. e 1E+05 i , t i I
'. ' 6 . .
1
. Near mid . Bldg Surfaces 3 1E+04 1 .,
l ,.j h 1E+03
. [a,l "
j
, - -.T **
E [!l **fj u. )J,*i . i * , i 1 . ' i ** 1E+02 . ,, . . *i a*j. l .. g .' * .
,,t1.s.t.d 4,
! .l ,
*' ; !. )' gjf' # I .
SB 1E+01
.I".-
" * *.,s . 8 a" .5* .' t;*' t *;
~
1E+00 "
* ^^l. .
l~**
- l. i I
1,. 1E-01 n I i I 1E 02 e 1 0 1 2 3 4 5 Normalized Distance ' Figure 13. Ratios of predicted to observed concentrations for the Wilson Lamb model as a function of nortnalized distance. i
'O 9
9 4
,
- s .,'
- g ,
i~, :
- 3 1E+05 . i i
, . e Nest Field
- Sids Surfaces
^
1E+04 -- s-( - l s > *.% 1E+03 F ' i
,.a i'r{
i .!-
,E ,,
,e ,,
.~, . .
i i
~
lols::nt;$n,*"%. - , . . l, 14 0 w.a . .
- i j,E.W,,. i5 $Ts., i,, t !. ; n -,. [
n: 1E-01 1E-02 0 't 2 3 4 5 Normalized Distance ) Figure 14. Ratios of predicted to observed concentrations for the revised model as a function of normalized distance. t The final comparison between models is a direct comparison of the cumulative w quency distributions in Figure 15. The distributions of ratios from the Wilson Lamb and revised model are almost identical. This indicates that these two models whould give similar i results if usat in control room habitability assessments. The Wilson-Chui model overpredicts more of the concentrations than otti,er of the other two models. One difference in models ) 'e ; rught be responsible for the larger number of overpredictions by the Wilson-Chui model is that the Wilson-Chul model does not have a means of accounting for~ increased dispersion at low wind speeds due to mesnder. Both the Wilson-Lamb model and the revised model account for enhanced dispersion at low wind speeds. -
- CONCLUSIONS Evaluation of building wake dispersion models Loginning in the mid Ig80s has shown that models currently recommended in NRC guidance to licensees tend to signNicantly overpredict concentrations during low wind speed conditions. As a resutt, the procedures tr.ed in evaluation of control room habitability and the conse'quences of design basis
- 1 21 -
i s Al
* / QW[
y,....
-~~~~-~~~~^~"~~~~O i, . .
.. ,. )
j
- ..i NM ga_ s.. m..
a=is. :.emp .ge mim _. -..q 7. y
'. %,, ,c............_. J gmmm.y.. . , .
_.x . . . . . . . . .
% m.a . . , J
. 333 ;= .
.g 7
- ,f.=
1E+03 pmu aq.. m g.=m .
.euux E
, 7.., a, ay g,., j 1E+c2 4 ......L.L..........
!D Wdson Ciud '"~~ '
7;- wn.m ' O , ' 1E+01 .
!-- ~ -
- m. p m.
rs
.._ . ,#u w y w. . ._.. . .._m._..- m..se _. y
- 1E+0o isep n'5r~ .
- .e.e m=.
..--= :g=;_q... .g.we e <
g a .
.s,_ .,
, 1E01 .. g ==._mm mm :me me .._ ,e I 1E02 I
- 0% 20% 40% so%
so% 100 % i Cumulative Frequency 1 j Figure 15. t Cumulative frequency distributions of the ratios between predicted and I j observed concentrations for the Wilson Chui, Wilson 4amb, and revised l
- t. model for all data near the relem point.
)
i i i s accidents were felt to be overly conservative. A new model was developed in 1990 for use predicting concentrations near buildings that did not overpredict concentrations at low wind speed. The 1990 model has recently undergone peer review. This istter report describes the i
- disposition of the primt'y recommendations of the peer review panel. Those l recommendations were
!. 1) The turbulence increment generated by buildings should be assumed to proportion l . to the wind speed in accordance with accepted thsory and physical reasoning. I 2) j The effects of meander during low wind speed conditions should be treated ex { in the model, but the treatment should be separate from tha treatment of building wakes. 1 { 3) ' j An approach to determi.)Mg concentrations other than otraight-line Gaussian models i thould be considered when releases are from a building and receptors are on or near the building.
- 4) Appropriale subsets of the evallable data should be used to e' valuate the model after the suggested changes have been made.
. . g 22 e
- . - - -.- - _ . . - - . - . - _ . - . . . . . ~ . - . - . ~ . - - . -
- - .y. -c. .-. ,
.. 4
) -
q
- i,* .
j ( j in response to the recommendations, the 1990 model was revised to exp i enhanced dispersion in the vicinity of buildings as a combination of the effects of i high wind speed phenomena. The low wind speed component of'the enhanc the revised model decreases with increasing wind speed. In contrast, the hig l { component increases with increasing wind speed. Turbulence data colle buildings has been used to model the increase of turbulence in wakes that is enhanced dispersion at high wind speeds. Diffusion data collected in aximrime ', vesclors indicate that the revised model is a significant improvement over the ' l vnodels recommended in existing NRC guidance to licensees. l Two attomative, non-Gaussian models devdoped to estimate minimum dilution (maximum concentratinnt) in plumes from building stacks and vents were identifi tested using a different set of data from experiments at three of the reactors. Th appear to predict en upper bound for concentrations in the immediate vicinity of the re point. Concentrations predicted by the revised model for the same data also tend t
~ higher than the measured values. The differences between the predictions of th dilution models and those of the revised model fall within the range of differences be expected given the intended bias of the minimum dilution rnodels. Further the revised model with the minimum dilution modelindicates that conservative near the release point and become less conservative as the distance from release point increases. Cumulative frequency distribution of the ratios of predicted observed concentrations for the revised modelis nearly identical to the distribution Wilsorkt.amb rWdmum dilution model.
The revisto model incorporates the changes recommended by the peer re and concentration predictions near release points that are comparable to the concen predicted by minimum dilution models. Therefore, the revised modelis considered to be appropriate for use in estimating concentrations for control room habitability assessmen The revised model is also considered to be appropriate fur use in estimating con in the near field for use in evaluating the consequences of design basis aocidents. e U
l , ey. '
- . .: .. s.
) .c
- . SEFERENCES
( . i
- krqpect Analyses. American MeteorologicalnaSocie
) AVnospheric Environment 268(4):513 517. Briggs, G. A. A . l i Reactor Complex on Effluent Concentration.' Nu . GPUSC.1972.' ' Atmospheric Diffusion Experimentse swith SF, T and Nuclear Station under Low Wmd 50288, General Public Utilities Service Corporation.Speed inversion . Conditions ce o. i i i
.. Turbulence and DMus/on.10012041, Commission, Idaho Falls, Idaho.
er c Idaho i Martin D. O. and Tikvart J. A. (1968 A general ! the effects on air quality of one or m) ore source. atmospheric dispersion mode APCA, St. Paul, Minnesota,18p. presented at the 61st Annual Meeting of the j , Murphy, K. G. and K. M. Campe. 1974.
- Nuclear power system plant con i
- design for meeting General Design Criterion 19.' in Proceedings o i
i Conference, Washington, D.C. San Francisco, Califomia, CONF 740807,aning , U.S. Atom i Model Cube.' Atmospheric Environment 16():207 222.O 1 ' NUREG/CR 5055, U.S. Nuclear Regulatory Comm .. Ramsdell, J. V.1990. "Diffus@ in Building Wakes tor Ground. Environment.24B(3):377388. . Start, G. E., J. F. Cato, C. R. Dickson, N. R. Ricks,"G. R. Ackerm 1978. RanchoCommission, Nuclear Regulatory Seco Building Wake Washington, D.C. EWects on Atmospheric DMus - Otart, G. E., N. F. Mukari, J. F. Sagendorf, J. H. Cato, and C. R. D Sullding Wake E#ects on Atmospher/c D/#us!on NOAA Technica Air Resources Laboratory, Silpr Springs, Maryland. - Tadmor, J. nd Gur Y. (1969) Analytical expressions for vertical an sostficients la atmospheric $ffusion. Atmospheric Environment 22(): 688 68 24 e h
i Wf .r
- b. e i ( -l
)
4 .
- Thulliier, R. H. and R. M. Mancso,1.M. Building Effects on Ecluent Dispersion from Roof '
[ Vents pt Nuclear PowerPurnte. @RI NP 1380, Electric Power Research Institute, Palo Mto, Califomia. Thuillier, R. H.1982. ' Dispersion Chvectiristics in the Lee of Complex Structures.' .lournal
- of the Nr Polludon Control Association 12
- 526 532. .
Thulilier, R. H. 1988. . , i Wilson, D. J. and E. H. Chul. 1994. "Innuende of Bunding Size on Rooftop Dispersion of Exhaust Gas.' Atmospheric ErMronment 28B(): .
- 4 Wdson, D. J. and B. J. Lamb.1994. ' Dispersion of Exhaust Gases from Roof Level Stacks j ,
and Vents on a Laboratory Building.' .@nospheric Environment 28B(): . s' . 4 } t 4 i I . 4 i i
. _ _ _ . ~~
a
* ,[
O CONSIDERATIONS ON INPlfTS TO ARCON96 FOR WESTINGHOUSE AP600 DESIGN BASIS ACCIDENT ASSESSMENTS HUREG/CR-6331, Rev.1, " Atmospheric Relative Concentrations in Building ' Wakes," documents the ARCON96 computer code developed for the U.S. Nuclear Regulatory Commission for potential use in the assessment of atmospheric dispersion related to control roam habitability design basis accident (DBA) reviews. Since the ARCON96 code has features that may be appropriate for other uses, certain considerations should be made when applying the code to Westinghouse AP600 DBA assessments. The basic diffusion model implemented in the ARCON96 code is a steady-state, straight-line Gaussian model that assumes the release rate is constant for the entire period of release and may, therefore, not be appropriate for application to other atmospheric dispersion scenarios. Therefore, the following considerations should be made when using the ARCON96 methodology for control room habitability design basis assessments for the AP600. The staff will perform independent assessments for applications citing this or other methods to analyze control room habitability DBAs. The meteorological data used by the applicant should be submitted with the application. Number of Met Files: Five years or more of data representative of hypothetical AP600 sites and long-term conditions (e.g., 30 years) at these sites should be used. If high quality data of this duration are not available, ARCON96 can perform an adequate assessment for DBA reviews with as little as one representative year of hourly data. However, use of one year of data may require additional analyses to demonstrate that the single year of data is adequately representative. To assure a timely review, the hourly data should be provided on 3) inch diskettes in the format specified by Appendix A to Standard Review Plan 2.3.3, "Onsite Meteorological Measurement Programs." The SRP format is that formerly recommended for providing hourly data. Specification of the use of 3\ inch diskettes is a new recommendation. If the data are in a compressed format, the capability to decompress the data should be provided with the data. In addition, joint wind speed, wind direction, and atmospheric stability frequency distributions of the data should be provided. A description and diagram of the meteorological tystem(s) used to collect the data used in the AP600 assessment should be provided, including a layout of the meteorological system (s) with respect to topographic features, vegetation, buildings, and other structures. The heights of and distances between the various features and meteorological tower should be shown. A description of instrument heights, the period of measurement, temporal and spacial representativeness, and quality assurance measures should aho be provided. Distance to Recentor: The obimum horizontal straight-line distance from the postulated source to the.reteptor should be used. If the release height and receptor height are different, ARCON96 will calculate a slant range value. ' Enclosure 4 A
d
- i. ef s
i, l 1 i Adjustments may be made for minimum distance around or over buildings or other ! structures, but supporting rationale must be provided to assure that the distance calculated is the shortest distance possible. For area sources, consideration may be given to use of a weighted geometric mean of the closest
- and furthest points from the receptor. For any case, the receptor should be outside of the area source.
l Release Tvoe: Releases should be evaluated as either entirely ground level or elevated (i.e., not vent / mixed mode) based on the same criteria as specified 4 in Regulatory Guide 1.145. Calculations performed for elevated releases should reflect the potential for recirculation of radiological material (i.e., non-Gaussian steady-state behavior). This recognizes that, for tall stack releases, most of the effluent may initially pass above the control room intake, but, as a result of stagnation or actual trajectory, the material may become available for intake to the control room at some later time during the accident period. Stack Flow: To justify use of the stack flow feature, Westinghouse should demonstrate (e.g., via the AP600 Technical Specifications) that the flow can be maintained for the duration of the period over which it is applied (e.g., the entire postulated accident). Default Values: Other than the initial diffusion coefficients for area source calculations, the default v. lues would not normally be changed. Justification should be provided for departures from the default values, as well as rationale for adopting the default values.
.}}