ML19261A717
| ML19261A717 | |
| Person / Time | |
|---|---|
| Site: | Atlantic Nuclear Power Plant  |
| Issue date: | 04/16/1975 |
| From: | Shapar H NRC OFFICE OF THE EXECUTIVE LEGAL DIRECTOR (OELD) |
| To: | Case E Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML19261A713 | List: |
| References | |
| NUDOCS 7902070019 | |
| Download: ML19261A717 (30) | |
Text
-
April 15, 1975 Note to Edson Case, Acting Director, Office of Nuclear Reactor Regulation
SUBJECT:
" CLASS 9" ACCIDENTS Attached is a legal memo regarding the need to discuss " Class 9" accidents in safety evaluations and envirar. mental impact state-ments.
The cemo forms the basis for cur recommendation that alte in the " Report on Consideration of Core Meltdo.sn" for the FNP applicaticn be adopted.
However, wa could not agree to circula-tion of a DES (which would be lecally deficient) pending the ccm-plation of an environmental assessmsnt of the core melt event.
As the memo indicates, consideration of " Class 9" accidents would be appropriate in both the safety and environmental reviews en which experience has been developed", and th The ccnsideration should be given by the staff in this case to use of a supplanental siting methodology in addition to the criteria in 10 CFR 5100.ll(a) that would. include consideraticn of " Clas accident consequences.
The meco also contains a ganeral discussion of " Class 9" accidents as treated in environmental impact statements for pcwer reactors.
I would like to discuss this matter with you further.
Some additional comments on the FNP Recort are written in the margin of the Report forwarded to us for ca.T=ent, which is returned herewith.
Mcward X. Shapar Executive Legal Director At achments:
1.
Le;al memo 2.
FNP Report o1.01
.s.
s REPORT O!T CO:ISIOE2ATIO!! CF' CO?.E !C-CT.i AS A DESIG'T 3 ASIS RECCIRCT,7 AND ALE'! ATE COURSES OF ACTIO!! TO CONCLUDE RI7ID OF Ei?.V?LICATIC:i-jb ',h h 1 GC?tWtt 6 ICD
/Md
(
b J
}LC~
4' dLLC4,titot C jyf-/JA 4P 4
/
<Affaced otka t4pa4
^
( 9 d N L c i $ trr A u._ 4, I w h Ucuf4 dh d4x) pas Ifarch 24, 1975 M
TA3LE OF CONTI'?TS SECI' ION PACI NO.
I.
STATIMIiT OF I"rII PR03LN 1
A.
Staff Use of ?robabilitf Assess =ents 1
3.
Recent Considerations S
C.
The P:cblen 7
II.
?ROPCSID ?oSITIO::S 7
A.
Consideration of Core !!altdown in Sa.fatf Ivaluatiens 7
3.
Alternative Courses of Action to Conclude Review of Mi? Application 11
~,
,, =
I.
STATD.E;T OF THE ??.OBLEM A.
Staff Use of Probability Xssessnents Radiological safety assessnents perforced by the Cc-4ssion staff purs.:a.ut to its responsibility in the licensing of power reactor facilities have to date consistently required that the design of tha facility be such that no credible si=gle event or sequence of events be capable of resulting in unaccept:ble
[ cc: sequences to de public. Al dou;h our reg.lations do n:t
- n.. /,
d2fi:2 de ter= unac:2ptab12 : ns:quen:2s, vm have for s:fety N '.)jd D
,3+[
a:sessnent purposes generally interpreted thes to be desas gC
,j S Q:. 3 eL-substant4=117 in excess of the guideline values specified in h
. :.e
/ 0'# V 10 CFR 100. Similarly, the tern credible is not defined in P.
's.\\ N*q.
our regulations. Ecuever,'our past licensing actions have indicated that the likelihood for an accident resulting in consaquences substanHnfly in excess of the guideline values specified in 10 CFR 100 should be in the order of one cha:ce in one nillion per year per reactor. In s"-ary, our practice has been to require that a reactor, plant be designed so that the probability of having,an accident occur that would result in consequences
'IC.DM substantially in excess of the guideline doses specified in V
-6 10 CFR 100 veuld be no greater than about.10 per year per reactor.
e eum -
.e w
ene e e we e eew = e- -w e
.i,.
_2_
Since the consaquances of a gr:s3 c:re sci dcwn in a pcuer rea : tor plant will, for_ all designs approved to date, clearly result in calculated doses far in excess of the 10 C7R 100 guide-line values, it is clear that in the past we i=plicitly concluded that the probability for a gross core =eltdown in each approved reactor pl' ant was,less than 10' per year. The staff has in the past nade esti=ates of probabilities for specific events and cenbinations of events and, on the basis of the results of those quantitative esci=ates, deter =ined whether or not the event or cr=bination of events need be censidered as a design basis for the reactor plant. ' he folleving are ens =ples of instances wherein the staff has =ade quantitative prebability assessnents to deternine whether or not a postulated event need be considered in the design of the plant.
1.
Reactor Pressure Vessel Failure The staff concluded on the basis of a generic evaluation that the probability for failure of a reactor pressure vessel
~
is less than 10 per year per vessel and, therefore, that such an event need not be considered as a design basis for
~
a light water reactor power plant. This evaluation is descrhed in WASH-1318, Analysis of Pressure vessel ' Statistics
~~~
fron Fossil-Fueled Power Plant Service and Assessnent of Reactor Vessel Reliability in Nuclear ?over ?lant Servi:e (dated "ay 1974).
^
2.
Aircraft I=sact The probability for a plane to i= pact a nuclear plant is quantitatively assessed in each case where the location of an airport or of aircraf t landing and/or takeoff patterns are such that the probability for an aircraf t c 2ch at the site is of concerr.. Quantitative assess =ents of at:-
i t
craf t i= pact probabilities have been =ade for nu=erous applications, including those for the following facill:1es:
(a) Three M.ile Island (b)
Zion (c) Shorehan (d) Saabrook j
(e) Douglas ?c. int i
The calculated probability in each case was used to determine whether or not and to what extent aircraf t i= pact needed to be considered ~in t,he design of de facility. These evaluations are doc =ented in the records of these applications.
3.
Erolosiens For sites *+.ere ship =ents of explosive =aterinis create a potential safety pro' ale =, the staff has assessed the probability for such explosions and dete=ined the extent to whi.ch they needed to be considered in the design bases for the specific plants involved. Specific applications for which this has been done include 3r=svick and 3yron-3:aidwcod. Doc =entation for these evaluations is pro tided in the applicable record.
~
- l..i 4
?.aactor Scram The p::babilitf for reac:or scras upcn deand has been assessed by the staff on a quantitative basis. This assess-t
[
=ent is described in WASH-1270, Anticipated Transients Without Scras for Water-Cooled ?cver Reactors (dated Septe=ber 1973), and forns the basis for cur current positien on ATWS and the need to i= prove the reliabilitf of the reac er ser:s syst:= in future pl:nts.
In addition to the above, a caraful search of the records vill provide other c:ses where the s:2if has cade qu:sti:2:1 re probability assess =ents which se: red as the basis for licensing action. In general, however, whether or not an event or ce=bination of events needed to be censidered in the safe:f assass=ent of a reac:or plant was dete=ined en the basis of a qualitative probability esti= ate.
If the esti=ated value was clearly less
~
than about 10, it was neglected; if it was significantly greater
~
than about 10, design changes were required either to reduce the probability to an acceptably Icw value or to al:er the plant so that it would be able to withstand the event in an acceptable manner; that is, lead to consequences within the dose guidelines established in 10 CFR 100. The entire structure of our safety evaluations has frem the on-se: of the regulatory pr: gras been based on es:4--tes of the likaliheed for the occurrence of specific
g
. t e
i.
'0.11e several of thes e assess:c.t: hate been quantitative events.
.sst have b een ade en the ba. sis of u:dec ranted qualitaci e group judgnents. In any event, for each casa ve evaluate, we as a staff convince ourselves that the probability for experiencing an accident in the nuclear pcver plant that would result in dose ay consequences si:;nificantly in excess of 10 CFR 100 is so Icw
..p
(<J. ~
that it need not be consid' red as a design basis. In part, since e
[
a gross cars =altdcun accident vould lead to consequences well W.[. W I,
above the guideline li=its, we also consistently cenclude that y
the probability for a core neltd:wn in a licensed facility is
/
D e.,\\v 4 g.eh 'g,(/, g a.L -
)
~
less than 10 per year.
t
'U A'1 3.
Recent Considerations
. i.9.a The occurrence of two events during the past year or so qd has, as a result of their c==bined inpace, led the staff to
- ~ -
k(\\d b
rea-=- 4ne its position 'rfth regard to c:nsideration of f&
{jy core neltdown. The first event was the proposed u,se of nuclear M -g/
power plants installed on floating platforms located miles t
, L.,
10 g.E offshore at ocean sites. On the basis of preli=inary consideration, we vere convinced that the probability for a core =eltdown accident for a floating nuclear, plant (EIP) would differ little from that
. for a land-based unit'.
Cn.this basis, ve. have consistently refrained frot fornally requiring the Ei? applicant to consider core reledevn as a design basu event or to investigate the consequences of a =eltdown.
g, e og e.
e e 6 6M 6
- O N 6 **
l 6_
i The Advisory Cc~e 1::ee en F.ea :or Saf egu:rds (Cc 1::ta or AC'C) hcs he n 1 volved in tha rr.-1:v of the T.? since 1:s ini:ial proposal. Pron the onset, the ACRS has voiced its concern that core celtdown should be considered for the F27 concept. Without any support fro = the staff, the Cc==1::ae has encoura~,ed the applicants involved in the PJP proposal to investigate the consequences of a core =eltdown at an offshers ocean site and to consider the poten:ial for =itigating those co:: sequences by use of special design provisicas such as a core catcher. The Cc_ it:ee has =aintainad a censis:sacly 3::::g posi len with
- spa:t :o :h'a need :o ::nsider the cora saltdsva si:uaci:n for a FNP.
The second event was the issuance of the draf: version for public ce==ent of WASH-1400, An Assess =ent of accilant Risks in U. S. Cc==ercial Nuclear Power Plants (dated August 1974).
This draf t report indicates that the average probability for a core =el:down is 60 ti=es greater than the 10- per year value discussed previously. The draft report was reviewed by a Regulatory staff Task Force during the latter part of 1974.
One of the conclusions of the Task Force was that the core =elt-down probability predicted by the Stu'dy Group that issued 'the WASE-1400 docu=ent was too high by a factor of at least ten.
4 e
O w-e se
=
e 6
ee ee og e -
e.>_e eh
e**=
eme ee-
.,, e
7-C.
S e b l em The pr:bla= faced by the staff can be sap--'
=4
'- o the follo. ring ita=s:
l.
Is the evidence provided in WASH-1400, in conjunction with
/
h the concerns expressed by the ACES vith respect to the 7J' P
,%C 4-9
(*jt'("4ge'P. \\
g?,
concept, sufficient to support a staff conclusion that s d
' ') c? 4,f MTP core ahltdown consequences should ba
\\
h h,* h. h.
evaluated in detail?
w
' g,,.r,,s, y 2.
c'.J,(. ' '
4 If the answer to 1 above is positive, then should s#-U sr
[ y Tg, p
avaluations be nade for land-based nuclear plants?
\\p ",* II.
V
?'0?oSID PCSITIC'TS g
Saverni alternative courses of action are being eensidered by the staff in order to bring our ongoing review of the Ei? proposal to a conclusion. These are described below.
Consideratien of Core keltdern in Safetv Evaluatie-s A.
',;hile no special in-depth investigations have been conducted by the Office.,of helear Reactor Regulatian with respect to probabilities for accidents leading to core neltdcun, we do have a reasonable knculedge of the WASE-1400 draf t report and of the reasoning behind the core =eltdown concerns of the ACRS.. Further, we have a good understanding of, the basic design differences,
bar:.reen a land-based and a floating nuclear plant and can reasonably assess bov these design differences and siting differencas can, e
.e ee-w
.m
- =*
-e - e mse==
=.....
3-I l
in a general way, affact the probabilir7 for core =eltde.n and g
the consequences of such a =eltde.n.
On the basis of this l
. y,U.T, ju k=culedge, we have reached the folleving conclusiens with respect i
i.'#
to the questice as to whether sufficient evidence is nov available p~.
s.)1 ak-d f
to justify detailed evaluation of core neltdevn consequences.
v S *4. $,[.
1.
It is likely t'.st rigorcus evaluatien of all pertinent Y}, s).)
- j,3 v.*
infor=ation, including, the WASE-1400 draft report and the a
uf fW
- J,g;$,d co=ents nade by the Regulatory staff Task Force on the i
i.w s# h' draf t report, vould lead to a significant reduction in the
. 7*.~g\\[t,
[
. r.).
- P ^..
co-- - 'tdcun pr:babilities given in the "ASE-liOO drait i\\
\\
". t,':* 'g#. l s
,g..
report.
1 h
V a' gr t j
V'p^ s
(,'*
2.
On the basis of the WASE-1400 analyses, there are reasonably o
clear-cut design and operating and =aintenance procedural
.c cha=ges that can be =ade for nuclear plants to further reduce the probability f' r core =eltdown. These include o
el1=i=ating the potential for. single equip =ent failures that could lead to system failures, #-f-ing dependence upon operators to perfor= emergency functions, and careful scheduling of naintenance.
3.
The probability for core =eltdevn in an approved floating nuclear plant will not differ significantly f, rom that for a land-based plant.
6
e
~9-4.
Le acuta or i=ediate radiological consequences to indiriduals i
of a core eltdov:2 in a floating nuclear pl:n: located at an ocaan site so=e siles fron land vill not diffar g t.tly from those for a land-based plant. In the event of a core 2
=eltdova in an ice cendanser centsi=ent, such as is pr: posed i
I for the floating nuclear plant, the =ost probable = ode of I
initial conni-act failure for either a land-based pl..n:
1 or a floating nuclear plant win be due to overpressurization ra*5-- ** an =elt-thr: ugh of the contai=an: base. It is i
1.ikely that the at=cspheric releasas will control the extent of the acute :: saquances. The doses to piopia fro = such releases vin, on the svarage, be less for a floating nuclear plant since it vill be located sc=e siles frec any populated land areas. The acute consequences to individuals fron the relaase of the =citen core to the ground under the land-based plant and to the ocean under tha floating nuclear plant vill be less than those associated with the at=ospheric releases.
Prald-4"ary a=alyses by the floating nuclear plant applicant have indicated that for that concept, releases to the saa vould not result in any fatalities, while the at=ospheric
. releases could lem'd to several fatalities. W ile we have not as yet reviewed those enTyses in any depth, qualitative assessnents of the potential acute consequences for either e
- ~. _ _..
g, f -
l the la:d-based or fleati=g nuclear plant indicate that post-accident c:atrol =sacures.i'l limit the c:uta ec=-
sequences from the water or grou=d release pathways to a secondary order of cincaru co= pared with the acute con-sequences frc:2 the at=ospheric release pathway.
5, The ic=ger-term, or environ = ental, consequences of a gross core meltdown in a land-based plant compared with the consequences of the saca event in a floating nuclear plant are likely to diff2r to a greater extent. Even for these c:nsequences, h:v: var, tha diff arezca =ay :ot be as vide s ci;ht initi:11-y ha su ri:24.
0:.a of tha 2:s: si;-.ifi-3 cant differences between the floating nuclear plant and
=ost land-based plants vould be the loss of control of radioactive =aterial that selted through the plant. For cost land-based plants, the release of radioactivity to the ground water could be controlled to a significast extent by the use of per'i=eter wells and flow barriers. For the ocean-sited floating plant, dispersion of radioactive material to the ocean could be #-adiate and would not appear to be a= enable to practical control. A similar
- }.
problem.could,however,.existforsejaland-based plants where the =olten core could possibly enter a near-surface aquifer before preventive =easures could be taken,
ag 11 I
e t
and cente inate the water used for a vide and populated
.rea.
- ddition,1s=d-based plants located = ear the ocean share, a lake shore, or river bank are likaly to l
conta=inate these usters is due e'-= via the diffusion I
of long--lived radiocuclides. It is difficult to see how i
the az:ast of conta=1:ation of water bodies could be as great for a land-based plant, but it should be recogni:ed that the water bodies could serve as sources of dri:@.isg I
vater or for food irrigatica purposes.
Another ebvious e:vir:n ental differe=ce results frc= the fact that dispersica to the occas vill conta=i=ata large
.I e
areas of " interna.tional vaters" and result in potential food. chain doses to residents of countries othar than cur ovu that vould be, difficult to esti= ate with any degree of accuracy at the present ti=e.
6.
The practical desi:;n of core catchers or other, devices which veuld =1tigate the course of events of a grossly
=elted racetor core is beycnd present-day technology.
3.
Alternative Courses of Action to Conclude Review of Ei? Amelication The following alternative courses of action are available to the. staff to genclude' its. review of the E.T appficatica:
a e*
e enum.====
e.
e e
=..
ww em
i i.
i Alt e: na -ive 1
'hintain the position that the probability for gross core i
l neltdown is lov enough that it need not be a design basis event for any light water reactor plant. On this principle take the following steps:
I 1.
Issue our SER vithin a few months holding to our position on the low probability' for gross core =eltdown and indicating lf in the SER that this is a basis for our favorable findings.
i p'..(f.
J 96 w/k Acte =pt to convince the ACRS, on the basis of qualitative
- 5
(\\,y.2 6
arguments, that the probability for and consequen:es of a
. c3-y V. C',i M
- ss cora neltdcun for an occar.-sited nuclear pc ser plant
\\
', \\
do not diffe. greatly fro:2 those for land-based plants.
,T
+
2.
c' (p Proceed with our present plans with respect to environ = ental
.s s 7
i= pact state =ents; i.e.,
to issue the E;? generic DES with 0@, ).W g\\>.g little change in our position on class 9 accidents realizing u
v,-
I the potential for DES recirculation.*
e 3.
Have the Office of Nuclear Regulatory Research e:ctend the I
y WrF 400 study to floating nuclear plants on a generic
. s. s,
.4 i
9(.f~
I he,(; #_thout any ties, schedular or otherwise, to the
\\
i 1
processisy of the.FNP license applications presently before
)
NRC. Current plans call for extension of WASE-1400 to the iLetter dated 2/21/75, A. Gia: '.usso to E. Case, " Accident Analysis for Generic DES - Offshere ?ouer Systens (C?S)
l s
m 13 _
l l
L !?32 ETGF., and F:;? (nuclear povered merchant ships should be included in the F'i? assessment for co=pleteness). We esti= ate that this effort vill be ec=plete in two to three years.
Alternative 2 t
Proceed as in Alternati-e 1 except that va vould conclude
%'f 303
- c
~
N g.ug in our SER that the =anufacturing license should be restricted 8). M
-u
,v 1 l
to a limited nu=ber of ocean-sited units (say two to four).
t
- l..:
% [-
Indicate to OPS that its application could be = ore readily
\\f extended to a larger nu=ber of units if the additional units
\\
l vere located at protected, non-ocean sites, such as a sealed i
lagoon.
In such locaticas, the consequences of a gress core
=eltdown could qualitatively be sh..wn to be not significantly different from those for a land-based plant.
& >?ative 3
~
roceed as in Alternative 1 except that while we would issue 6
y '.
o u:: SER vithin a few =enths holding to our positica on the icv 0
\\
,3 (;,
probabi31ry for gross core =eltdown, we veuld indicate that r4 'j V
because of the persistent ACES conce as about the consequences M
,n@
. L of a gross core =eltdown, we are requiring the applicant to per-g..\\ sp r\\,
y additional analyses of the acute consequences to confirm Qg y
that their =agnitude vill not differ greatly fres those calculated for land-based plants of s1=ilar design. The applicant would be requested to. provide these ::alyses to the staff and ACIS during e me
-w e=-sea se e ee eiem _
e e=
e-m e e e m** * **
- N
- N m - * *-
- 14 _
I the course of the review. Tae st.aff would review and cc=ent on these analyses to the AC7.S, but they vculd not be used as a basic I-for the staff rec:=endatics in the SIZ re;2rding is uence of the canufacturing licanse. This vocid probably require an extension of six to nine =caths in our schedule.
Alternative 4 Expand the scope of our review to include the consideration l
of the probability and the consequences (both acute and le=g-4 i
l ters) of a a:ss core celtd:wn. ?:cceed in the follo -ing -'mer:
i 1.
Issue our SE?,vi di: a few =ench: but vit.'.h:ld en: fi:-l conclusiens pending'the c pleti:n :f ::: :::e::=2nt cf the gross core meltdown event. Tais will per=it the applicant, V
C' the staff, the Coast Guard, and the ACRS to resolve any outstandi=g issues on the basic design while the degraded accident concerns are'being addressed on a different ti=e scale.
2.
'Jithin the Office of Nuclear Reactor Re;ulation, develop' a program of six to ni=e =enths' duration to i=vestigste and sy.
ce= para core celtdown for the floating nuclear plant located at a fixed ocean site and a land-based plant located at a few typical sites.- Specif?c areas to be co= pared are (a) core.
=elt probability, (b) radiological (acute) consequences, and (c) environ = ental (long t.e:2) consequences.
RL (1:.'?.-1)
- e
...-e.,m
-e-.-
=====e
=
= --
une e-
-e-e
15 1
would assu=e responsibility for ite= (a), TF. for ite= (b),
and ZL (I?) for ites (c).
ne applica:t v:uld be requesced
~~
to provide additional infor=ation in all thesa areas. Tais 6
i 8
program would aim toward issuance of a supple =ent to the SE2, late this year.
3.
Eave the Office of Nuclear Regulatory Research undertake a short-tern evaluation *of core celt for floating nuclear plants. The progra= should be li=ited to consideration of a floating nuclear plant at as ocean location and sheuld consist of an investigation of about six =enths' duration to bound the consequences of ocean dispersien of the radi.~-
t M
activity released fro = a nelted reactor core deposited q
within the proposed breakuater for the Atlantic Generating Plant. This vould provide needed validation for our short-ter= assesscent described is (2) above and for the concurrent investigation that is to be conducted by the applicants.
It is our present understanding that such a study would require about six =cuths to perfor= but could not be co=pleted until about a year to a year and a half from now because of other priority work assign =ents and =an-power T'~4tations.
a e
O e
S FW
-em e 4
mem+ 6 ee ee eN e N& g a 6e g ee h
=N e
es e eeee me w
ee.
l m
~
)
1s _
l 4.
Have the Office of Nuclear Rayf.atory Research e:etend JASE-1!.00 g--
to esvar ficating nucicar plants as specified in Alter :.acives 1, 2, and 3 above.
1' 5.
Issua our ganaric DES as soon as practical but withhold our final conclusions pending the cc=pletion of an environ-4,t 3
C-?@(.).V./,-
cental assesscent of the core meltdown event. This vill
- gut, Q
g't.
per=it the identification and possible resolutien of any 3
0 {Ypid proble=s that develop vith respect to our basic assess =ent i
l
(
,;(,
while the degraded accident assess =ent is being per'o=ed
.#w e&. q on a longer cine scale. The schedule ceuld be directed
,w 6@ '1e> '
.g, tcvard 1: stance of an anended DES, if needed, by about l
'.,.;50~
b
- F' the end of the year; that is, in concert with the planned l
o
', j., 0f '&
- sb,
.\\-
SER Supple =ent addressing the probabilities of cors =elt-
... t.p-
'9
!.... P,,
devn events and the safety related consequences of such
.' i'f-
=eltdcuns.
cra-h 6.
The DES for the manufacturing li x::: vas issued in.Tuly of'last year. The co==ents are to be incorporated into the FES for the generic reviev which would be expected
.k#u to be issued se=e four conths after the DES for the generic l'
review; on this basis, the FES sight be expected to. be issued is the. fall of this year.unless an a=enduent is needed in which case it would be is the spring of next year.
.eem eg e.as e eme g
- "e
s T
17 -
nis alternative vould pe:.it the basic SIR and DES (generic) to be icsued withis a ::sth c:
.co, the first SIR Iu;.ple=sst to be issued near the end of the year is concert with an a= ended DES (generic), if needed, both of which would address the degraded 8
accident event in appropriate perspective. The fina.l. SER Supple-i (following a final ACRS =seting and favorable letter) and
=ent (c'v the FES (generic) =ight be' expected to issue in the spring of
.S f
1976 should as a= ended FES be required. The hearing could u
co==ence, and s==eti=e during the hearing, the results of the short-cars investigatics by the office of Nuclear Regulater7 Rasearch should be available. The results of the 1:ng-te =
review could be issued before issuance of the nanufacturing license is needed (no sooner than 1978) and should be issued in ti=e for our review of the final design.
Alternative 5 Eave the Office of Nuclear Regulatory Eesearch undartake a co=prehensive, detailed, long-term study and comparison of the probabilities and consequences of core nelt in land-based
($
and offshore float 1=g nuclear plants. nis would involve as in-depth study of both acute and long-tern consequences. of core.
= cit for offshore plants, and also an e= tension of
'n' ASH-1400 to investigate the long-term, environ = ental core celt consequences e
e w*6 ego am.
e e
w y p.
..O a g g w e musuD6 me e
e, e
eg.
.s I
- IS -
for land-based plants.
~ is affar: ould require three to fise a
ye2rs to co ple.e z.nd would inJolte both thac:etica'
-peri-d
= ental effort to achieve the desired result. Issuance of a
=anufacturing license would be delayed until the study is co=plete.
Alternative 6 i
t Inform the RTP, applicant that N?.C cascot, in the next several years, issue a =anufacturing license for offshore sited floating
- h' i
pl:sts. This decision would be based on the need for =uch : ore b~, *
.t 9 infor=ation about the consequences of core =e'.t than is prescstly
\\y".I"3[Cd'$ '
a a11able end the fact dat ths :::esca f inf:=ntica ess::- be
' ~}'* ~
htsined withcut a :sf or, in-depth etudy of the ;r:ble= 0.::h as is described in Alter =ative 5).
I Point out to the' applicant that if his application is li=ited to plants located at protected, non-ocean sites, such e
. '.f v
.A.
.s cs a sealed lagoon, the probability and consequences of core n#g t*,
=elt could easily be shewn to be ce= parable to land-based plants.
- 0. q-. V If the ETF applicatica were so 11=ited, the :TRC revie i could
,1 h
' proceed without the =ajor delay described above.
L t.
I
.t
/
Gg.p=
GW 0
,% (,%= c e
e e
s S
S w 4
.*g6 4 N O+,
g6 h 6h
- M'*'
N MN
s, m
UNITED STATES NULLEAR REGULATORY CC?.r/ISSION W ASHirJcToN. o.
C.
20555
- .:A? 3 L 1975 L2e 7. Gessi&., Inecutive Director for 0;erations CORE MILT CONSIDERATIO!!S AND ALTZ?l:ATIVE PROCICPIS 703. PRCCISSING TEE FLOAIING NUCLEA3 PLANT APPLICATIONS The enclosed report discusses a major policy issue that has received considerable staff attention during the past year. The report also presants alternative courses of action that =ight be taken with respect to conpletion of the licensing process for the floating nuclear plant.
The Director of the Division of Reactor Licensing and the Acting v/
Director of the Division of Technical Review both recon =end adootion
.[Kg' f
of Alternative 1.
While I believe I understand the bases for tb.is '
reco=nendation, I do not think this Alternative is viable and instead
.J g
"j.i.
believe Alternative 4 should be followed.
y..?.,f.)3-e P.
I have requested co==ents on the enclosed report and proposed alter-native courses of actions fron H. K. Shapar, S. E. Hanauer, and E. J. C. Kouts.
In addition, I have requested that Dr. Kouts provide esti=ated dates for the start and co=pletion of the work -bat we believe could be done by his Office under each of the alternative courses of action. I have asked that the co=nents be provided to
=a by April 7,1975.
In addition, RL and TR have suggested that the issue concerning the degree to which core =eltdown analyses should be perfor=ed in individual
.l-cases be considered on a generic' basis outside the context of any
'ff *l,L specific case, so that the Con =ission itself can be involved in the g
- d. ;.l j. f decision without concern about ex parte linitations. Howard Shapar
. O.p. ?
d.
has been askad to consider this approach.
g(9gf..
.Q I would like to discuss this =atter with you af ter receipt of the
,.f requested infor=ation from Dr. Kouts, Dr. Hanauer, and Mr. Shapar.
, p '(
I will let you kncv when such a neecing wo tid be appropriate.
/.'%
~
Edson G. Case, Acting Director
. ' Office of Nuclear Reactor Regtilation
Enclosure:
As stated
@Cv r.hc O'O c
1
=*
2,,:.. /
t m
_ w,,
z=-
.,,,,.4; 1 %
J5 %w 4
'?
.196 7'
a s,
=%
Us tTE D STATYS NCCt.I A R M CU LATO RY CO?M.*:ssiCN W MHIN G TON. D.
O.
20555 4 ril 13, 1975 Edson G. Case, Deputy Director, Offich of Nuclear Reactor Regulation CORE MELT CONSIDERATIONS FOR FLOATIt!G MUCLEAR POWER p!. ANTS While I agree with you that Alternative 1 is untenable, I do not think that the issue you raise can be considered solely within the four corners of the floating plant question. True, we mght do as the lawyers suggest and try to treat fl:ating plants specially because of the lack of experience, but tHs does not seem to me to be a correct position.
Moreover, I do not believe that your proposed Alternative 4 really solves the problem.
The licensability of the floating plant depends, it seems to me, on two decisions:
(1) Whether the plant - reactor, platform, moorings, break-water, etc. - are adequately designed, and (2) whether it is alright to put this sort of machine out in the water near the coast.
If only design basis accidents are considered, and if such problems as breakwaters and ship collisions are properly resolved, there is no difference in concept between land reactor ~ safety and floating reactor safety. But this is not by any means the first time that such questions have affected reactor licensing. Indeed, if only design basis accidents and the words of Part 100 were considered, we could allow plants to be built in Burlington for sure, and probably at Edgar and Ravenswood.
Yet all these three sites were rejected because they are too close to large populations. One of the rationales for this was the unlikelihood of successful evacuation of large dense populations. However, behind this was another consideration in the backs of everybody's minds that very large accidents are possible and that their possibility, even though they are
'!ery improbable, dictates keeping reactors out of highly populated areas.
I believe that the ACRS is correct in asking the FNP applicant to consider accidents worse than the design basis. I believe that NRC should also develo basis)p an adequate appre i tion whether bad accidents (outside the design would be " catastrophic" in the FNP. Now, unfortunately I don't have a good definition for " catastrophic" or a good definition how low the probability should be before I am willing to accept a " catastrophe".
Ideally, one would have at least a comparative Ras.::ussen-like study.
In the real world this is some time off.
fs#c+h c
l 4
=
x 2
%s.at*>
. Y:ur Altarnative a is an atterpt to rea-h same decisions paMing the completion of a g::d safety study for the.:.'iP.
I don't sse hcw it can be made viable. Having opened the question about core melt consequences, you can't proceed without an answar in my opinion. Tne work you propose to get done and the papers you propose to circulate are in: mplete, and in a certain way trivial, without an adequate consideration of the core melt probl em. I much prefer your Alternative 5 with some rearranged priorities and maybe some intensive work by the applicant. That's what I think Alterr.ative 4 would end up looking like anyway.
Tnere are some other important questions about the floating plant, such as outside power reliability, storms and breakwaters, ship collisions, etc.
I think these should be pursued without waiting for the core melt question.
Tnere are lots of plant design details for which a suitable resolution ~is
~
sure to be available. I question whether doing a lot of work cn these at the present time is really worthwhile.
Attached are a draft talking paper on the ?,asmussen Study and how it affects Class 9 accidants and also some detailed co=ents on the materials transmitted to me on F. arch 31st.
AN
- -c-3 37.xq
- ::. ' %.. e..
m._
~.::j'.._~ :-in ;
p ShphenH.Hanauer
-Technical Advisor Enclosures
- 1. Oft Talking paper re Ras =ussen Study
- 2. Detailed Corm:ents
~
C'2. AFT T."LKII:3 PAPER PPZLI:lI:' RY C01 S!:ERATIC:IS It!.'JHAT MIGHT 3E a
APPROPRIATE TO DO ABOUT THE PASMUSSEN STUDY RESULTS A.
The problen being discussed here. is the apparent discrepancy between the safety study and the Regulatory safety goals.
1.
The safety study shcws a core melt probability of 6 x 10-5 Although AEC Regulatory has cct=ented that this figure seems too high and has pointed out some of the conservatisms, it dcas not seaa to me very likely that the final study report will have a significantly different number.'
- 2.. The current safety design basis and the current rules 1-d regulations'
._____ _ governing reactor safety are not expressed in_guantitac; e probabilistic
~
-:=--:..__
~~-
. terms." Rather, the concept of " credible" is used to differentiate
'between. postulated accident sequences to be included in the safety design basis and others, not so included because their probabilities are
~
so low that they are " incredible". It can be seen that sc=e notion of
. probability must be employed even in this formulation in order to distinguish '" credible" frca " incredible". Mcwever, little or no guidance is given in our rules as to what constitutes " credibility"..
- 3.. The growth and increasing respectability of quantitative probabilistic methods has impelled industryand intervenors, as well as the staff, to consider whether and to what extent probability numbers should be considered in safety evaluations. Soth on its cwn hook (in the generation of the ATAS report, 'JASH 1270) and in response to e
. cententions and questions from inte: tenors (preceedings at Zica and Monticalic), the staff has stated in vari:us ways some opinions as to what constitutas " credible" and as to what might be suitable quantitative safety goals.
i
.i In the ATWS rsport thl' staff suggests as a safety goal that events a.
with consequences rore severe than 10 CFR 100 guidelines should '
have a frequency no higher than 10-3 per year for the USA. Today, with 50 reactor running, that goal translates to 2 x'10-5 per reactor year. In A.D. 2000, with 500 reactors
- running, the goal would be 2 x 10-6 per reactor year.
b.
In testimony at Zion, th'e staff suggested that the pro' ability of c
a LCCA with failure of ECCS and substantial breach of contair:nent
~
~
.should be less. than,10-6 per reactor year with that facility.
c.
In a draft Standard Reviaw Plan 2.2.3, an acceptance criterion is proposed for potential accidents caused by hazardous operations off-site such as transportation of explosives in railroads, canals, etc. Such events are to be included in the plant design basics if their probabilities, evaluated realistically at the 50% confidence level, are calculated to be above the range 10-7 _ 10-6, 4.
It seems clear enough that the safety study result, that core meltdown has a frequency of 6 x 10-5, is semehcw in apparent disagreement with the various safety goals suggested recently by the staff in the range 10-6 for severe consequences.
l
- 3.
Oces a discrepancy actually exist?
,1.
It is by no means clear thac th2 core meltdown fraquiacy is as high as 6 x 10-5 per reactor year as given by the study, Tne study itself' recognizes substantial conservatisms in this a.
i esti= ate. AEC Regulafory, in its cor=ents on the draft tudy, ~
pointed out that this number seemed high, perhaps by a factor of 10 or more..However, it seems unlikely that the number will be changed significantly in the final study or that any reasonable a=unt of effort will produce a better number. 'Je are therefore going to have to use-this one.
b.
An important result of the study is that the consequances of core
. celtdown are highly variable and that the cost likely consequence 1.;: 2:.~ : -(,,. _ tis very mild indeed., By contrast, the present Regulatory approach implies that the consequences of 1.e.," incredible" events (not in the design basis) can be and are likely to be extremely severe.
In fact, the two approaches to safety evaluation are very different and the apparent "discrepane,y" in the frequency of gross core failure cannot and must not be separated from the equally important discrepancy in consequences of such fafTures. For the single distinction in 10 CFR 100 between credible accidents whose conservative calculation consequences cust be within guidelines values and events outside the design basis (Class 9 accidents) whose consequences are not considered, the study substitutes the-I
- cceplate spectru.m of accident consequancas with the fraquancy distributian function to go with it:
C.
'Jhat can ve do about the " disc. epancy"?
1.
We can explain it away as in paragraph 3.1.b. above. This is really not very useful by itself,-since the discrepancy in numbers will renain to confusa the public and delight.the intervenors. Ofrect and correct explanations tend to sound defensive and unconvincing because of the widely diffehnt premises of the two approaches and the apparent ig'noring of the less 'avorable approach. It really isn't le favorable for public risk but appears to be so as far as core meltdown is concerned.
2.
The "10-6 goal" can be changed. NP,C could announce that in view of the
. ;.- ~. insight gained _ in_ the safety study and the apparent satisfacto.ry.
=...
state of public risk shown in the study, the safety goal is now to be 5 x 10-5 per reactor year for severe events. This wculd have the effect of per.nitting the design basis of, for example, floods and plane crashes, to be relaxed substantially. The resulting effect on real public safety would not be neg!ible. A new " discrepancy" would then appear since the relaxation would lead to a new value for core meltdown frequency which I predict would not be consistent with the new safety goal.
It is noteworthy that industry reaction to the safety study seems to be in this direction.
3.
Improvements in safety could be required to bring the core meltdown fraquency dcwn to the order of 10-6 per reactor year by AD 2000. This
- scunds great and apparently i. proves public safety, but makes no real sense if the risks shown by the safety study are in fact acceptable. Resources spent in decreasing risks already acceptable are resources wastes. We don't need sugh a pointless increase in the cost of energy.
4.
Change the safety evaluation basis to be more nearly.like, or equivalent to, the risk, evaluation basis. Tnis is a fonnidable technical and policy task.
a.
It cannot be acccmplished satisfactorily by waving a magic vand and decreeing that the new safety design basis would be a quantitative probabilistic risk evaluation on every nuclear pcwer plant. We do not have, at this time, the technology to
^ (;
..,3.;acccmplish _ t'his. task, nor the resources, nor, the data'. The_.
,respense would be a mad scramble to hire all the fault tree experts in the country with the result a massive and mostly.
heaningless numbers game. The most important drawback is the lack of real meaning to the numbers thus assembled. This does not mean that the safety study is meaningless (for judging whether nuclear power is unsafe) but just_ that it is not now valid for making case-by-case licensing decisions.
b.
Some more modest change must be found that preserves the good features of the present systen. We know it has good features because the level of risk that has resulted seems satisfactory.
We knew it needs improvecent because of the insights gained in the safety study, and we knew that the new basis must be
i
. ccnsistent with availabla technoi:;y and a reasonable amount of resources.
D.
Hcw can we get started with the r.ecessary reforms?
1.
Obviously, the final versicn of the study must be issued. This will 3
he definitive in the sense that it would take a lot of time and resources to do any better. One hopes that the ccnservatism of the results will be more carefully appraised, especial!y as it related to core meltdown probability.
~
2.
The ATWS question will have to be reexamined at least to some extent.
A chorus of industry ccmplaints states that the safety study results, as given in the August 74 draft, negate WASH 1270 and the staff's earlier avaluation.
'a'e have already initiated (::py attached) a
,-t-;p.a. scoping. staff study of these apparent disagreements withcut waiting for.
the final study draft, which will in turn no doubt have to be stud.ied some mor'e for the same purposes.
3.
There is no obvious analogous discrepancy in.the area of risk due to external phencmena. The study relied.on the adequacy of the regulations -
this area. Sane further study work would seem to be indicated and was recommended in the AEC Regulatory review. We shall have to see the final study ta find out whether there is a problem here or not.
4.
In my opinion, the correct approach is the difficult one of recensidering the present safety evaluation basis in the light of both the methods and t results of the safety study. Scme possibilities for this are suggested in paragraph C.4.b. above. Scme care detailad ideas w:uld be the subjecc ofafutUrepaper.
t 5.
In ths meantime, peadir.g such a stu:y and, if a :a;ted, such ref:m, the sensible and technically correct and res;cnsibia thing to da ses.:s to is to ackn:wiedge the apparent " discrepancy" of paragraph A.4.
I and to explain it on the basis of differences in approach as discussed
~
in paragraph B.l.b.
It might-b'e well if a position paper were prepared on this subject.
l i
4 t
a
- e e
me 4 esp
. e ee e
as w.
.g ee p.m.
.,,m,
_,e
- - * * ** **. ~ ',*
- ",,,,,, ' " ' ~
[
O e ((.
[*
e "Q*
O G
G e
O
- = :::.:r 91 a.91,6 ;J ww..1;e l J Fa;e 1 - The frequen:y discussed here follcus 'fe:r the assun: tion that about 1,000 react:rs are in operaticn. S me distinction s?.cuid be made bat..aan :: day with 50 reactors runr.in; and A.J. 2003 with 500 to 1,C00 reactors running.
~
Page 2 - One of the lessens of the Rasmussen study is that cost core
=eltdcwns do not have drastic consequences.
This puts " catastrophes" further down on the probability scale because of their unlikelihcod even if the core celts. The discussion here does not make this distinction and therefore fails to laarn an important lesson from the Safety study.
Page 4 - The summary in the last half of this page is core sicplistic than the facts. The caterial in thd last three lines, and on top of page 5, neglects entirely the sort of considerations that have led to the rejection of proposals to build reactors at Suriington, Ravenswood and Edgar. Thus, while each word of this discussicn is true, it f ails to take an important factor into account.
Reactors must be shown to be safe in consideration of all possible occurrences. The spectrum of such cccurrences ranges frca the trivial to the catastrophic. For. occurrences within the design basis the consequences must be shown to ce tolerable by conservative evaluations.
-.ce postulated events outside the design basis the prcbability must be shcwn to be low and _ acceptable and qualitatively depend on whether the probability is low enough in relation to the consequences. The value of the Safety study is just in that it makes these relationships axplicit and quantitative.
Page 5 - We have indeed implied that core celtdcwn frequency is icwer than 10-0 per reactor year. This problem is discussed in detail in a separate attachment.
Page 8 - Item 3 has to be demonstrated.
It has not been so demonstrated yet in a way that is convincing to ce.
Page 9 - The basis for Item 4 is not evident to me. Are there evaluations or calculations that show this in any quantitative way? In particular, the discussion of differences in failure mode between ice condenser containment and the contain=ents studied in Wash 1400 should be substituted or deleted.
Similarly, Item 5 seems to contain several unacknowledged conjectures.
Page 12 - I agree with you that A' ternative 1 is untenable.
Even if I thought it was true, I doubt if an acceptable defense of it could be sold to the board, the courts or the public. I do act believe it is true.
Page 13 - I don't see how we could rec:rmend building a small number of plants with this kind of an open questien regarding their safety.
8 Alternative 3 ignores the problem unac:eptably and,: reposes that :ne staff be an estrich.
.s i
2-
? age 14 - Altarnative 4 - I have cc=ented on this at length in :.y covering
- E.o.
Fage 17 - Alternative 5.
I think this is the only course and propose that it 5a done e.xpediticusly instead of sicwly as prcposed.
Page 18 - Altarnative 6 is what happens if ycu don't do Alternative 5 with considerable speed. I don't know whether lagoons are a viable alternative for a few years.
.. ~ ~ ' - *
- o S
e I
_