ML20064L644
| ML20064L644 | |
| Person / Time | |
|---|---|
| Site: | Limerick  |
| Issue date: | 09/03/1982 |
| From: | Aldrich D SANDIA NATIONAL LABORATORIES |
| To: | Okrent D Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-CT-1497, NUDOCS 8207180083 | |
| Download: ML20064L644 (5) | |
Text
w...
._m.-_. _
+M
- e7-Jy'97 Sandia National Laboratories ##
'\\ f Albuquerque, New Mexico 87185
. n.
.a September 3, 1982
[.
...g i'
- s
,1 Dr. David Okrent 5532 Boelter Hall
.y University of Califcarnia, Los Angeles e
Los Angeles, CA 90024 e+
hf
Dear Dave:
Enclosed is a summary of the very limited review I have performed on the Site Consequence Analysis of the Limerick PRA.
I apologize for it being late.
Please f
let me know if you have any questions or desire addi-5 tional information.
Sincerely, e bdric,k
' -f D. C. Aldrich, Sucervisor i[f Safety & Environmental Studies
-b Division 9415 1
'g[sh; DCA:9415:mjc 4
m Enclosure kE Copy to:
w/enci
,p l,N k. J. Michael Griesmeyer Advisory Committee on Reactor Safeguards U. S. Nuclear Regulatory Commission Washington, D.,C.
20555 E
f;
- JJ A
i l -
A
+
B207180083
t
.= :. ::,.::=6 w..m.-- --. - :--.
6- /Y9'Y is p
Review of the Limerick PRA:
Site Consequence Analysis
~ ;
D. C. Aldrich, Supervisor j
Safety and Environmental Studies, Division 9415 i'^-
Sandia National Laboratories
((
Albuquerque, NM 87185 5&. ' "
1 In my limited review of the site consequence analysis portion of the Limerick PRA, I've attempted to focus only on those aspects i],
of the modeling or evaluation that could significantly impact
- g predicted results or major conclusions.
I have the following
'l comments and observations.
. [L The study made use of the WASH-1400 consequence model, CRAC, d
1.
with site meteorology and population data specific to the 74 Some slight modifications were made to the
?
Limerick Site.
CRAC code, or to input assumptions:
. The vertical dispersion parameter, az, was altered to reflect higher ground-surface roughness than assumed in f
d' WASH-1400 k3 was altered The horizontal dispersion parameter, ay, lease duration V
. to more properly account for accident re U
V:
M:
. Plume rise assumptions were slightly modified 3
. An average factor of 0.57 was included to account for y
reduced inhalation exposures for persons in buildings with b
ltk windows and doors closed
. Shielding factors for persons in buildingc were
'l significantly lower than assumed in WASF-1400 (lower 4
implies better shielding) for two reasons.
First, the h
Limerick analysis assumed that, in areas beyond the 25 mile evacuation distance, people would actively pursue
'g sheltering in basements whenever they were available in k
h homes or larger buildings.
Basements provide a great deal J
of protection against external gamma radiation from j
deposited radionuclides.
WASH-1400 did not assume active 7
sheltering beyond the evacuated zone and assigned average shielding factors for persons continuing their ' normal 7
activities" (i.e., a mix of being indoors and outdoors, with no utilization of basements).
Second, shielding factors were further slightly reduced to account for the large number of homes in the Northeast that are brick and 4
,.i:
hate basements.
f
. None of the above modifications is likely to have a major impact on predicted results.
The assumption of active i
sheltering in areas beyond 25 miles won't have a E
significant impact on early fatality predictions, but C
could reduce predicted latent effects by a factor of 2 or
)
i i:
...o....-
fu b*
p I would expect the cumulative impact of all of the so.
modifications to be a small (- factor of 2) reduction in all 3
predicted health impacts.
2.
Over the past few yearo, a number of improvements and correct.sns have been made to the CRAC code, resulting in the release of CRAC2 from Sandia in 1981.
Few of these 47" i
improvem;nts and corrections were included in the code used Jf1 for the Limerick analyais.
To partially evaluate what Lips.
influence this might nr.e on predicted results, I've r
attached some figures generated as part of the recent International comparison Study of Reactor Accident 7,
- A Consequence Models.*
The figures compare CCDFs (and means)
- q(
predicted by CRAC, CRAC2, CRACIT and NUCRAC, for a
- f, moderately large release at a hypothetical site.
Note that
' I,$
Figures 8-12 assume no immediate evacuation of persons near
.Te the plant, while Figure 13, CRAC, CRAC2 and CRACIT all agree relatively well.
The wide range indicated in Figure 13 most likely reflects the variety of assumptions made, although differences in modeling also contribute.
3.
To provide a direct comparison with WASH-1400 results, the
-!(
Limerick Study made use of the WASH-1400 evacuation model and assumptions.
These assumptions were based in part on a simple statistical evaluation of a small amount of evacuation data gathered by the EPA.
The appropriateness of the assumed times (no delay, 1.7 mph " effective" radial speed) and distance (25 miles) are questionable for use at a specific site.
(Note that choosing "better" estimates for, s
delay time and speed is not easy!
A 10 mile evacuation distance assumption would have been more realistic in light
]{l of current emergency planning requirements.)
As I've
- q discussed in my previous reviews for Zion and Indian Point,
!b predicted early effects are very sensitive to assumed l k9 evacuation parameters,'particularly to the delay time before
!Ph public movement.
4.
According to Revision 4 (6/82) of the Limerick Study, core fission product inventories were calculated using the RADCICINDER code system.
The resulting assumed cesium
, h'i inventories are significantly smaller than in WASH-1400 I
, 7Q find this a bit puzzling.
Using an updated version of
- ky ORIGEN, we have performed some inventory calculations for a large BWR that predict cesium inventories ~40% higher than
- t-f-
those in WASH-1400.
Assumed cesium isotope inventories can have a very large impact on predicted latent cancer fatalities and areas of interdicted land.
Therefore, the discrepancy in inventories should be resolved.
/.
..y
't
- Aldrich, et al., " International Problem for Consequence Modeling:
Results," presented at International ANS/ ENS Topical ![T Meeting on PRA, September 20-24, 1981, Port Chester, NY
'i.
' ~
^
na'er.
......w_
by....
s.-
k 5.
Treatment of uncertainties in the consequence analysis was rather weak, particularly if compared to those for the tion and Indian Point Studies.
Additional calculations showing sensitivity to emergency response and other " key" assumptions would have been beneficial.
~t ;gii,,
t f
6.
The iguestion raised in Mike Griesmeyer's letter of May 1R, hf'?
- 1982,
'Are the proposed interdiction and decontamination 4F ^:
approaches practical for the actual areas involved?" - is a I
good one.
However, I don't have a good answer.
Essentially
,4 all consequence codes and analyses performed in this country
':{
have made the same assumptions in this area.
Work is g
ongoing at Sandia to better estimate the extent to which M
interdiction and decontamination would be viable over the
~%
large areas currently predicted for severe accident
- g categories.
iti' r
n N..
.e ib..
lW
'l C
id b
I t
O Y
. 1
!,nn.'9 -
..........-...~............. ~. ~... - - -- :- +--..;
+ - -
v-
)......
L-e f
-.s The results presented on this page are for comparison purposes only, and are not necessarily representative Note:
4:
of potential reactor accidents.
.s., ( *T-A
' 7 15 T
8,
- 3. -
K i
H 4
H.
4 l
7 5
s gW1 C8'C" "i
E W1 tanc 2
-F mT**
(W I
Canc2 l r.p l sa :
Eh
' p.
h eag3 tanc 6m canc lEM Cw11
( '-h:*.
canc saa I
canc2 Rio
' v?
l cant 2 1m 3
CRAlli N CRAC2
- ?d a
Caecit 3Q acAt 720 gg m
, e.-4
'it essi 30' 30 30 30 30 rg/
'g
'g "g
"g fj eas h,
4,
, x
- N
- w. a.v. a Fi
- 1. Passaa 6. tru, Isiivo Pwast'*
Finias 9. Passa 6, m h :Su %.ti.
Q5 1
8 llh
'i 1
tsat av Canc2 28,
4
/
CaA011 360 g
&, ' f -
t
- / _
acw 1
.anat gu'
/
'?
E 681
~
Caa'81
/
cd
, M
_j
!V CRAC11 Sat,
"I pc g,
! N
- h'*..
3 Lwn 153 5
ta;.t2 m "I
M*C h
arx au y
nam 10' 10 30 10 30' 30' 30 30 30 30*
essi We katas, 2 e
Cancte harms, I
.ty F8sa 13. Puesda L BA2, hastic feraatie.
Fasiss XI. Passan 6, Ir21, hittic Peraation
. {i +
9#
^
9 m
a, 8;
CaA;2 isvaciatse, ne.so.;.as.
s es mar, se esELTsaanG te.36 songsf i
e-
- L se I
I Cancli C un CM1
~
CaAC E
crvactaram, as sanas.
no saw. a.: mas Catc2 g"
c,L n _...,e e,a...
~
g "1 a
/
c= +
3 asasw..s.
M
. $. P 1
3 (ant 2 NJ)
J cant n
cantit mas i
5 canc2 see ty acanc 7x,
}
can:lt e 37 man 30' 30 30 10 30' 30' 30 10 30 30 sam he kat s. x g, g,,,,,,
Fasima 12. Passaa 6, SPFJ, arsrstsc Pocastsa.
F86+8s 13. Ptesaa 7. KI, atarstic Pernataan i
_lo_
. _.._