ML20245A591
| ML20245A591 | |
| Person / Time | |
|---|---|
| Issue date: | 04/10/1987 |
| From: | Speis T Office of Nuclear Reactor Regulation |
| To: | Novak T Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20236K304 | List:
|
| References | |
| FOIA-87-380, TASK-082, TASK-82, TASK-OR NUDOCS 8704270012 | |
| Download: ML20245A591 (2) | |
Text
.
APR I O 19p 4
MEMDRANDUM FOR: Thomas M. Novak, Acting Director
)
Division of PWR Licensing-A, NRR I
1 Frank J. Miraglia, Director Division of PWR Licensing.B NRR Bill W. Morris, Acting Director Division of Reactor Systems Safety, RES 3
Guy A. Arlotto, Director Division of Engineering Safety, RES Robert M. Bernero, Director Division of BWR Licensing, NRR FROM:
Themis P. Speis, Director Division of Safety Review and Oversight, NRR
SUBJECT:
STATUS OF GENERIC ISSUE 82, "BEYOND DESIGN BASIS ACCIDENTS IN SPENT FUEL POOLS" As discussed in a recent staff meeting, the enclosures are provided for your information. Enclosure 1 is a copy of the recent Board Notification which transmitted a draft report by BNL on spent fuel pool accidents. is a DSR0 outline of our preliminary views of salient portions of the report. is a copy of recent slides used by DSR0 in a presentation to the EDO.
C Our current activity on Generic Issue 82 is described in detail in the attached draft Task Action Plan (Enclosure 4). Updated schedules will be published in the 2nd Quarter GIMCS report soon to be released. Near term activity will include interfacing with BNL to consider all NRR/RES coments received on their draft report (meeting with BNL scheduled for April 21, 1987 to discuss com-ments).
In addition, a follow-up contract has been funded with Lawrence Livermore National Laboratories (LLNL) to provide estimates of the seismic fragility of actual spent fuel pools and to assess the structural capability of spent fuel pools for loads due to a cask drop. We trust that the systematic investigation involved in our overall Generic Issue 82 program is sufficient to allow continuation with minimal schedule impact.
BT ORIGIIIAL SIGI Themis P. Speis, Director Division of Safety Review and Oversight Office of Nuclear Reactor Regulation
Enclosures:
(1 Board Notification
,b (2
Outline of Staff Coments (3
Slides presented to EDO on March 27, 1987
'i (4
GI-82 Draft Task Action Plan o/
cc:
See page D
)FC :D5RO:R51
(:
IB
- DSRO IB
- 05p: D
- 05RO:D LAME :EThrom:jf
< is
- W inners
- BS uron
- TSpeis
) ATE 204/A/87____ _____:0_4/k/87 ;@G/P/@7
- @4M /87
- @G/ /@7
APR 10195:
2-l cc:
V. Stello E. Beckjord H. Denton K. Perkins, BNL DISTRIBUTION Central FITS DSR0 C/F RSIB R/F E. Throm G. Mazetis W. Minners B. Sheron T. Speis s..
March 27, 1987 i
Docket Nos. 50-275 and 50-323 MEP0c4:DUM FOR:
Chairman Zech Commissioner Roberts Commissioner Asselstine Commissioner Fernthal Commissioner Carr FROM:
Thomas M. Novak, Acting Director Division of PWR Licensing-A
SUBJECT:
BOARD NOTIFICATION PEGARDING BNL DRAFT PEPORT ON SPENT FUEL POOL ACCIDENTS (BN 87-05)
In acenrdance with.NFC procedures for Board I;ctification (BN we are providing for your infomation the following enclosed Brookhaven (BNL)) draft rep i
"Beyond Design-Basis Accidents in Spent Fuel Pools (Generic Issue 82),"
Department of Nuclear Energy, Brookhaven National Laboratory, Draft, January 1987, transmitted by letter from K. R. Perkins (BNL) to E. Throm
- V (NRC),. dated February 5, IPB7.
The report addresses the risk of spent fuel sterage at nuclear power plants and ineludes an evaluation of accident initiating events and their probabilities, fuel cladding failure scenarios arising from uncertainties in Zircaloy oxidation reaction rate data, and the potential for releases of radionuclides under varicus cladding failure scenarios. The draft report presents the results of a study per*ormed by BNL for the staff on Generic Issue 82, "Beyrnd Design-Basis Accidents in Spent Fuel' Pools." The draft report has b.tn distributed widely within the staff for a peer review. Like most of the Generic Issues and Dr. resolved Safety Issues currently being studied by the staff, the likelihood of beyond design basis accidents is being examined to assess whether vulner-aH11 ties could exist which might require further consideration of changes in design or operations. Preliminary staff opinion is that substantial portions of the report will need more critical review because sore assumptions appear to be oversimplified. Comrer.ts will be provided to PML and the final report is expected to be issued this sumer.
The contractor states, based on these preliminary results, that the risk from spent fuel pools is comparable to the risk from core melts in general.
The spent fuel pool risk is primarily due to pnstulated fires from self-sustaining i
oxidatien of the Zircaloy cladding of the spent fuel stored in high-density racks if water is lost from the pool. The dominant causes of water loss are due to structural failures of the pool caused by dropping of a spent fuel cask or by an earthquake. However, all 'of the estimates of the factors affecting the risk are preliminary and the report is less an estimate of the risk than identification of the factors that need further study in order to make a reasonably accurate estimate of risk.
5 n
ps L
. Based or a brief review, the staff expects that the final result of the studyThe r F
is likely to provide a much lower estimate of risk. report is rct base f fuel pools, but of similar corerete structures. assumed that fvlure of the pool would resul fficient to fail pool liner without considering whether the strain would be su Based on other seismic risk studies, the fuel pool could recuire a
a ruch stronger, and therefore less probable, earthquake than that required the liner.
Failure to fail the systems necessary to control reactivity and cool a co l
a few plants.
The report also assumes that all of the cesium would be released fron theWhile th eovivalent of three cores if a Zircaloy fire occurred.in a pool has Powever, in this would prirarily only af fect estirrtes of early f atalities.
the analysis of the fraction of cesium released from the spe l
"If the radial and axial temperature distribution were considered, the ca cu-The lated release fraction wculd likely be less, possibly by a factor of two.
Even report also assumes no retention of cesium in the fuel pool building.
if an earthquake destroyed the building, the debris would likely cause much of
~
Other studies of the cesium retention in similar the cesium to be retained. buildings and the large fraction of cesium retained at Ch overall the C
that the report nar ureeresticate reteatica by a fector of ten. rep The draf t report does not pertain directly to currently ongoing licensing efforts '
hearings.
for scent fuel pool expansion amendment requests by utilities, including However, we believe that the subject ray involve substantial public, press Fe also are providing the report, by copy of this Board Notification, Cergressional interest.
k to the Boards and Service Lists for the Diablo Canyon Plant and Vermont Yan ee report.
Ve will irfom you of further developments.
Staticr..
TFmas M, Novak, Acting Director Division of PWR Licensing-A
Enclosure:
As stated cc:
See next page
- SEE PREVIOUS CONCURRENCE OG5 FD33 D:PVR-A
\\
PDf3 PD*3 CVogan*
H5chierling*
SVarga*
RBernero+
3/27/67 3/27/87 3/27/87 3/27/87 3 /f/87 10Sn4 47pp[
SElacP TSovak
'/#/87
%/d/87
~ ~ ~ ~ ~ - - - - - - _
i.
g.
m Di57:.IEUTION Docket File (50-275/323, 271)
!.ccal PDR PDd3 C. Vrean H. Senier n np S. Verca SECY IP)
OPA EDO OGC ACRS H. Thcepson T. Piovak H. Denton l
R. Volltrer S. Elack J. Scinto L. Chandler W. PcGurren Chairren, ASLB-Chairman. ASLAP C. N. Kohl G. L. Edies P. Nerbut P. Pendonca 4
D. Kirsch T. Speis i
G. Mazetis E. Throm.
W. Pinners E. Possi l
Y. Rooney P. Bernero A. Hodadon r
t Consents On ENL Draft Report *Eeyond DEAs in Spent Tuel Fools" Cl-82 Content 1, Seismic Hazard Curves Page 2'
- 2. Spent fuel Fool Tragility Page 3
- 3. Shipping Cask Drop Page 4
4.
Tallure Assumptions Page 5
- 5. Spent fuel Poo! Inventory (equivalent cores)
Page 6
6.
Tission Product Inventory (ORICEN2 Code Use)
Page 7
7.
Fission Product Release From Tuel Matris Page 8
- 8. Fission Product Release To Atmosphere Page 9
O'-
- 9. Consequences (CRAC2 Code Use)
Page 10 o..
- 10. Kisk Reduction Measures Page il r
- 11. Other Issues Identified During Review Page 12
(.
i Page 1
(
Cessents On ENL Draft Report "Beyond DEAs in Spent Puel Fools" CI-22 1, St!!MIC HA"ARD CVPVES k
ENL used Millstone SEP seismic hasards curves for EVR evaluation and generated a *syttnes1:ed" curve for Cinna (PVR evaluation).
MAJCR COMMERTE. (RRAEIDSRO, EBIDPL-A, EE/DPL-E)
A. Millstone-1 SEP data is higher than more recent EPRI or Utility estimates.
Factor of 2 to 10 too high (for site specific estimates).
B. Use et hasards curve in PRA study.
(1) Methodology used by BNL appears acceptable to most of the teviewers, howeter a better justification for weighting scheme is needed. IThis appears to be a conclusion based on site-specific perception of study ) A more appropriate waighting schete (NUREC-1152) may not change results.
(ii) Methodology recensended by REAB/ DERO is to esamine the sensitivity of conclusions to a set of hasard curves.
(iii) Choice of the upper bound cut-off to acceleration has not been g.,
justified, ENL Indicates it is important.
C. The development of the Cinna curve appears to be in error, small.
IdCC110tf1DAT10N D RESOLVE 1ESUES Select a family (3 - mean, upper and lower b'nnd) of seismic hasard curves o
based on RR&B/DSRO comments and establish an uncertainty range for each fragility cesve ' fragility discussions follow). The data needs to be extrapolated beyond the typical 1.Og limit.
Results can be presented in a generic manner, given large uncertainties in both hasards estimates and site-specific estimates, and failure differences between alewated EVR pool and at-grade PVR pool can be estimated by fragility.
The integration should be provided as a table / curve versus g for each set of f
hasard-fragility curves.
ACTICN; BNL Page 2
-_. __._____________j
Consents On BNL Draft Report "Beycnd DBAs in Spent Tuel Fools" Cl-82 2
STENT TVEL f22k TRACILITY BNL used two fragility estimates, which were intended to be rep resentative of the spent fuel pool structure. ho spent fuel pool fragility was icund and an alternative approach was selected.
Cyster Creek Reactor Building A = 0.75 Bg = 0.37 By = 0.38 Zion Ana.Eldg Shear Vall A = 1.1 Eg = 0.12 BU = 0.20 The f r agility curves were weighted and combined to estimate seismic failure.
MAJOR COMMENTS. (ERAB/DSRO, EB/DFL-A, EB/DFL-B)
A. The estimated seismic capacity of spent fuel pools are overly conservative.
The structures selected by BNL may not be appropriate based on a review of other data (NURIC/CR-4334). Median g values range from 0.9 to 8.2g.
'B.
The spent fuel pool liner needs to fail before water is lost.
E v:
C. The weighting factors u;ed need justification.
RECOMMEND ATION 12 RESOLVE ISSUES
(,
A, A. Technical Assistance project (A-0814) with LLNL has been initiated (Apri!
1, If87) to provide fragility estimates for spent fuel pools. LLNL will also be evaluate the spent fuel pool capacity for shipping cask drop accidents. A task to develop a working definition of " failure" is included in the project.
Estimated completion date for project in the Tal! 1987.
E, in the interim, for the BNL final report, it is recommended that a re-review of available fragility data be made to provide an agreeable set of fragility curves. It is proposed that four fragility curves be selected (two for an elevated BVR pool and two for an at-grade PVR pool). A lower bound (fragile, failure at lower g levels) and an upper bound (better estimate of spent fuel pools) fragility curve should be selected. For each pool type, combined with the three hasard curves, sis estimates of failure will be develope 6. The median hasard curwe results should be presented as "best estimate" with upper and lower estimates used to provide a seasure of uncertain both in hasard and site.
NURIC/CR-4293 (January 1986) provides a range of estimated fragility for three foot thick shear walls which could be reviewed to determine if this data is more representative of spent fuel pools.
ACTION ENL and LLNL Page 3
Consents On ENL Draft Report "Beyond DEAs in Spent Puel Pools" Cl-82
- 3. SH I F F IN; QJ] pflE MAJOR COMMENTS: (RRAElDRSO, PS&EE/DFL-A, DRSS/RES, PEICSE8EE/DPL-A)
A. Human error rate not justified, based on unreviewed data and NRC Handbeck, and develeptent of numerical value not present. The BNL developed data is not justifed, not is sensitivity to data shown.
E A-36 resolution not considered by BNL.
C. Eisk does not address fact that there is little cask movement in the industry at the present time and that there are only a few licensed casks available.
RECOMMENDATION IQ RESOLVE lESUES A. LLNL will evaluate spent fuel pool structural capacity for dropped shipping cask accident.
B'. BNL should review NUREC-0612. " Heavy Loads Control at Nuclear Power Flants" and Generic Letter 85-!!, " Completion of Phase 11 of Control of Heavy Loads at Nucelar Power Plants," and include discussions of heavy loads evaluation.
{
C. Development of risk should include discussion of current shipping rates and pctential futurc rates.
D. A review of pool designs could be conducted to determine which pools at are risk to cask drop.
ACTION ENL and LLNL t
F l
Page 4
Comments On ENL Draft Report *Beyond DBAs in Spent Tuel Poc!s" Cl-82
- 4. TAILVFE ASSUMPTIONS MAJOR COMMENTS: (RRAE/DSRO, LDO, RSE&EE/DPL-A)
A. Instantaneous loss of pool is unrealistic.
- 2. Less of configuration / integrity of fuel assemblies (for beyend design basis seismic events) not addressed.
C Events which could damage fuel assemblies, with or without water, are not addressed (energetic missiles, heary load drop, structural failures).
D. Subject of criticality not addressed.
E. Loss of inventory can occur slowly or in a rapid fashion. In case of seismic events or possibly fires, condition might esist that deprive the operator of control or surviellance capabilities and were not considered.
T.. The seismic event, more than any other event, will affect the entire plant g.,,
not only the fuel pool structure. In PRAs, seismically igduced cgre damage
~
OS*
frequencies were generally estimated in the range of 10~
to 10
. The report does not address failures and consequences of either fuel pool or reactor cota support systems at lower seismic levels. In general, these systems appear more eutnerable to seismic failure than the fuel pool structures themselves (lower g
('
values).
BJCQMMENDATION IQ RESOLVE ISSUES A. LLNL will be esamining failure. Objective: Establish a workable definition of spent fuel pool failure. This will invlove examination of the fuel pool support systems (e.g. make-up capacity), loss of structural integrity and damage to spent fuel pool by falling material within the fuel pool storage building (including dropping of a spent fuel shipping cask).
B. Report should clarify that the assumption is that avents can result in less i
of pool integrity where loss of water asceeds make-up capacity and that fuel heatup calculation are based on
- instantaneous
- loss of water, a recognised limitation of the STUEL computer code. Tor events where the water loss is not l
espected to be rapid, possible recovery and or more appropriate human factor I
data should be included in the discussion.
Issues cencerning seismic response of fuel assemblies, damage to fuel assemblies and critica!!ty are not part of the current ENL project. An agreement on their treatment / assessment in the BNL report is needed. A sugesstion is to note issues as areas where additional work may be tequired ACTION. ENL and LLNL Page 5
Conr.er.ts On BNL Draft Report "Beyond DBAs in Spent Tuel Fools" Cl-82
- 5. 5FINT JV.LL f.22,L INVENTORY (IOU! VALENT CORES)
The risk profiles was developed based on actual operating histories and spent
- ..;,,c;
'.iting as of April 1987. In Millstone, the inventory is approximately three equivalent cores.
MAJOR LOMMENTS: (RSIB/DSRO)
A. Mas potentsal inventory been underestimated considering fact that there are facilities now licensed to eventually hold large inventories of spent fuel? A brief teview of NURIC-0200 data (Licensed Operating Reactors Status Sumanary Esport) indicated a more typical full fuel pool could hold the equivalent of seven cores. Hatch (a dual unit pool) is now licensed to hold nearly 10 equavalent coses.
B.
Tuel rod consolidation could result in even larger inventories.
RECOMMENDATIONS IQ, RESOLVE ISSUES Current inventories can be reviewed to determine "present" risk A b;ief look o-at KUREC-0200 indicates that the current inventory ranges from one to two equivalent cores, with dual units somewhat higher.
Tull fuel invent'ories should be estimated. Should be possible with current calculation with added decay times to discharge batches.
C ACTION: E!IL i
Page 4
i Comments On BNL Draft Report "Beyond DBAs in Spent Puel Fools" Cl-82 6, FIS$10N PRODUCT INVENTORY (OR I C EN2 fdpl jlflj, MAJOR COMMENTS; (DRSSIRES, RS!B/DSRO)
A. Appilcation of ORICEN2, as presented in Appendis A of BNL report.
(1) Average specific power assumption as applied to BVR needs discussion
~
of errors which may be introduced.
(11) Statement concerning total calculated radioactivity being the same as in a real fuel batch may needs further evaluation. Radionuclides inventory versus total energy being the same needs to be addressed.
B. Tission product inventory is dependent on burnup and power level. Since plants used by BNL are in genera! low power plants, inventory may be larger for higher (typical new designs) Mw rated plants. Newer fuel cycles may result in higher burnup values impacting the fission pteduct inventory.
i i
RECOMMENDAT1 DES IQ RESOLVE ISSUES O"
BNL should address cessents provided concerning ORICEN! use.
s..
Highest burnup rates used by BNL are 30,000 MVDIMT (BVR) and 44,000 MVDiMT (PVR). EPRI NP-3765 shows recent (post !?80) burnup values are larger Masisus values 44,000 MVDIMT (BVR) and 55,000 MVD/MT (PVR).
l Issue should be addressed and, if possible, assessed in terms of potential larger fission product inventories. Not actually in BNL present scope, may require additional work.
J ACTION, BNL j
\\
Page 7
l l
Consents On ENL Draft Report *Eryond DEAs in Spent Fuel Fools" Cl-t:
l 7.
Fl!!!ON FRODUCT RE1I AEE 11131211 MATR f!
l l
Two areas of concern:
(1) Ascunt of fuel failed by fire.
(2) Amount of fission products released from fuel matris due to heatup fo!!owing fire.
i MA)Q1 CONCERNB: (DRES/RES)
A. The dynamics of the fire should be considered. The 2ircaloy say burn so quickly that the fuel pe!!sts of the cold fuel say not have time to heat up to lin e tamperaturss tequired to completely vaporise the eclatile radionuclides.
- 8. Pefect ventilation versus imperfect ventilation cases appear to need additional wording.
Rtt0MMENDAT10NS IQ REfolVE ISSUES BNL should reconsider the fire propagation and dynamics. Results should Indicate how much fuel wi!! burn as a result of sofficient decay heat and how much wi!! burn given a rubble bed on the pool floor.
A slaplified heat transfer analysis may be needed to see how fast and how hot
(,,
the feel can get for a Zircaloy fire.
ACTION: ENL l
l I
i Fage 8
i Comments On BNL Draft Report "Beyond DBAs in Spent Fuel Fools" C1 82
- 4. TIS $1QB PRODUCT RELEASE 7) ATMOSPHEFE v s 11Mf?1 COMMENTS : (RIB E'ASIEIDSRO)
A. Possible retention of radionuc!! des on structures not considered.
.B.
No distinction made for non-seismic events.
ij~
RECOMMENDATIONS I E RESOLUTION E ISSUES
}.'.{
BNL should consider possible retention on structures-. May be justifiable for t-
[J,
BVR pools inside reactor building.
.+
g.
~AC7JON. BNL.
)
M
/> S *. -
- e. *.,
l4 A
e
'gs 0,
3 n'
(,
g 4-
- g.
'\\
Page 9
- +_
l Contents On ENL Draft Report "Beyond DEAs in Spent Tuel Pools" Cl-82 9.
CONSEQUENCES (CRAfd 1911 Elfd MAJOE CONCERfil. (RRAE/DSRO, DRSS/RES, EDO)
A. CEAC2 data concerning timings, height and energy of release are not provided.
B. Some radionuclides listing in source term tables are not directly accounted for in CRAC.
C. Consequences for Cases 3 and 4 not provided.
D. It is not appropriate to use NUREC-!!50 data to compare with spent fuel peo!
accidents, !!50 is plant specific study.
E. Shocid more than one site population distribution be considered, and is the conditional interdiction area different for different sites (based on popclation)?
RE,COMMMDATIONS JE RESOLUTION DI IS!UES Provide missing information on releases. Address source tera question.
Consider additional CRAC cases to address sensitivities?
ACTION. EHL Page 10
l Cossants On ENL Draft Report "Beyond DEAs in Spent Tuel Fools" CI-82 l
10 1111 EtDUCTION MEASURES:
MAJDK CONCERNE. (PIICSB/DFL-B)
Consideration of risk teduction seasures are su;erficial.
RECOM*END ATIONS E RESOLVE ISSUES The BNL project objective was to. identify possible measures based on potential cost benefit, no valvellspact assessment was requested. Report should be clarified to indicate that BNL has not attempted a detailed valuellspact assessment.
ACTION. ENL s..
s l
Page 31
Ccar.er.t s On ENL Dr a f t Report "Beyond DBAs in Spent Tuel Pccls" CI-82
- 11. CTEIR 15 5UES IDEh""!EI ED DUR ING REVI EV E2 5 "- r e.$
(RRAE/DSRO, DRSSIF.;S, RSIB/DSRO) l A. A;proach is not systematic enough to assure that all aspects of Generic issue 82 are covered.
- 5. Report uses several bounding assusplions and conservative boundary condition in several areas. Report is a reasonable screenine study in that it c:r.::r ra t ive bounds the problem.
C, The data presented can not be judged as being best estimate, point estimate, c;nserretire. Uncertainties are not, in general, addressed. It is not possible to tell if data is generic or plant specific, probably neither.
E1*,QM?:ENDATICNS,IQ RESOLVE ISSUES The stated objective of the report should coincide with the objective as described in the Technical Assistance project work statement. The report is not intended te be an " integrated risk assessment" of spent fuel pools. Given the s,,,,
possibility of Zircaloy fires, ENL was to re-evaluate the likelihood of I
draining a spent fuel pool and assess the risk associated with a drained pool.
The report needs to be modified to provide an assessment of values selected by
{,
ENL which are not supported. For example, VASH-1400 values and NURIC-0933 values are used solely because they exist, without qualification or assesskent.
ENL meeds to qualify these data.
To the estent practical, an assessment of uncertainty is needed. Human error rates need examination.The amount of fuel involved in a release needs more attestlon.
A* TION; ENL Page 12
a g
sen i
t s
y ad l
t e
o o
S e l
d y a
c o
gel s
c dx e P
npu r
eek s
i i
po Zl r
t a
i i
ai o
ns u i
k l
eh v f o Ueq s
e h s e op kh i
u r
r at 3
F P np yt n u r 3l r
a 9 o F a t
u eqe Sh n i
0 o l
o t h t
t a i
b -
s Gp t
n mrh h
a r n i
E a El et s
e e
g ew R e s
hl ou r
o Uu l
p at e S
S au po of d Nf r
r f ot f
get d h u
e e
n yh nt t
n nt n t
O ot ia s
ck g
i i i i
i r s e s
l u
p en ab e Mis t i di r s
d ns gi ua fa Un t
A o
ga abs l
I it ug vo D
f l
d o
r r E s e B
ed s ap e po Ml m t
D e s i
f u
o e
l r
e e
i t oi s e r
r r e
s dp e
er nt r d
eso o
dh a s a r o n
d t
r c
is n m e, gb dy ai i
r o
e zh er o if a
n z a el v a
l y
ue n h dg nt i
e s uid cg t
i e
i i
i ml es r
sf e
t n
ct e oib B
ei r
i em rd mcd i
a i l t
r e a cngu o
pr r d
s 2
i ir er s e
id e pa a
8 epas hl enh e
p nS L a T c Si t uL m i
e sNo G
G1 2
3 B c s
I
m2 w
s loo P
g a
l s
e 0
o e
u 8
t us am l
F odgy t en c
.ve t
)
t ri n emgsy ede s
n g 8. d au t
r s e
- 8. 0 isleq nase
?-
p 0 mn e
e h g S
odf d e gi
,r n
i o c chi
(
l r
e f
t e ct h o n
s a
ear Aet o t
v h
a ebmlu r a c e
i O
ea dg i
i d
s r
g not a r
l uai t sf os an i
A Cah da e
Cef a t
r e
s d
v B
d o
t sdr e rl y
ee et e ui t d
l a
D rP o l
va r cwi yrleue t
e v
aE n om v
zS 0 i
t sl t a
d a
l f i
vrist e doc i
1 i
n H1 go e
es S r oae f
a gny ozapr o
ceo rd i t r t al n
F nl i
y mo l
i r
a e
h e e l
i e clo gh n
u o e
st o r
l st o
ag o cf i
B el c owpori i
si c
l yPFh Si a P oT mt ir 2
MF nn L
a s pi ee 8
e:
e:
f f
e ph uLf uLf mSSt I
sNta sNta o
G s
s BS BS C1 2 I
i
=
~
u d
e e
d s
e u
a t
e y
s i
l s
u n
e r
r o
f a
g lo o
c a
s n
t e
m C
e o
e b
s v
P e
f n
r u
r a
e o
i if
.c 2
le gr e w
e s
r 1
r u
neb o
e
.d t
o l
t F
rl r s r u
c i
uf e eio o
i r r f
t d o w eb b
o o
h e e a
n r
t
,nl pd e d
n%
l e
e h
e e0 p
u e e n. sog p t
d1
,l f r ns a
i S
efuomio c
ct i
mh t t d cu i
ual m d
Ao n
opmb i
s r s ut n
b O
Cf o oiotba ea i
i r
r s
fdmf t
t r
s o
n oe inyt A
oetandob e
Cr s
e d
e o
i B
n a oietcns
.i eh e
t c
r p D
ol t c aoaL r
c uri eNa es e
a itc rdrt f t vo l
d a
ef a neb em r ee r l
r ss y
S f Ca eest nb n
t es a
e s pa mlerCa o o
o ef a
h t
l m
o s oee y
t a
rl eef t
l e
et R oa s e n
e e% %R o sh oe t
l B
s s 2
TleC 0
R0 0 ir a 01 2
1 a e 1
ple 8
e:
f mr uL f
sN a
os I
G sB S
CC t
i
~
I dna es s
u e s
as ca lo egr o
b e
t e a P
el e
br lbog l
a n e
)
p s
t s
u c
t mei l
F e
u o oa r
sf s
c t
r e e
c e ef s
r C
R e
a o
n r "s t
ch t
?
e n
t d
C nat P e
eF et la n
p A
s sS y d n e a
dt f
S y
cleh R
i il f r
ca C
cuo a,o I
c s' et r
(
n o
Avt n s
Ae r n n
ua e
s rP t
i FAe i
n e
s i
e r qn s
e v
a r
,S n
s v
oei y
om ni A
n C
m g
l e
a Ce no 7 or e n
R t
i i s I
B t
e odol A
e-yicC cb nr D
c r
r u
et s o
e eoodrr e d
v3C1 N c
vst d
o e
n e
nteg r
e S een r
r S
P vI a n
P o
l u
l n
t o q
of n o
o t
f f i
n t
P e
o oo y
o n
e s
n s
e d
n sCs e is ole i
o m ms B
s s
u c
o s
i r r i
r u F
rF c
C i
r e e ea 2
a A
at gt c pt p
r 8
e mn e
e e
e r
u mInanb u
l I
I s
op o
s o
G s
CS C
s C1 2
I i
de e
p ka p "s u
iht q
s n h
)
s r
a t
ot r
l l
oP a
o cn e
t o
a r
f
,e e
P dw s
ne a
ar o B
l oP mt e
us r n
u w c.h a
g e
se F
i 4
i s
iv l
e 4,rl c l
t e a aeuL e
D n
R eeuNN u
f Ryf B
d e
f p
a rg f
t t
n eay n
o t
es b
e y
S S
pusad p
e g
5 cad s
b n
n
- eo e n
r r
i E
i e
o g
- 1. b L r
yid f
n s
u A
D 3 mvs y
w i
(
ean le o
B d
pl eo k
d e
oboHc l
l D
oe i
i r
l f
f s r f e n
Dpob i
s t
u ne d
n r
e kt o
s noo
.e d
l n
d i
a ert yb t d o
Crt c
i t
a a o
y s
r nd i
l e
e guoe at ye c
u n cCe rb B
s id e
t n
pa*
i ir e vd s
2 pt 6 y Cg ol i
o3 a d
c u r
i 8
e h
nam eu eo h
S RJ R c I
t G
O 1
2 3
l l'
s l
s o
s k
o e
c l
P u
a L
r deN y
le h L t
L is c
s k u
S n c F
h e a t
Si r
.d r t
Cwa l
m ye e
w n
M y
a t ul A
r oifs e
I e
/e gl nG Tn o nenf nt u p
T S
oh o ri t
P det i
pn a n gt d h s e n
cauu s e f
oo nr g p
i l i
iP t
rb gtohiR s i
O h
.r W
o nt a n
s s
a R
A No o
nB a n iMei W.
it b
t i t
B dc c
dorp s r
a c
e D
aAd d ail ef r r o
e mer ou uos k yt no d
al p onisf t h fd B
st a st e
i s
e i
n gTii t
t nr c n
gc u w,S ngoao i
o ih r n ooo pi m y p at en er l
ut y
f sl t e
ssi a
e nIswP e
s B
a F n RisP b52 y r
F er e e
l r eS o kSi r v
P l
eu i
a s uo o
t 2
ednnu i
id qf c
t r
i 8
r n
dee e
ii l st ao ef AR b R
u nnt oobs I
I t
t e
o u
G FCCOR P1 23 4
3113 ACTION 11AB B E YOND D E S I GN jf_11) A C C I D ENTS J}] SPENT IyJ1 STORACE POOLS CENERIC ISSUE 11 REVISION J TEBRUARY 1987 Lead Organisttion:
Division of Safety Review and Oversight Office of Nuclear Reactor Regulation Task Manager:
Edward D. Throm Reactor Safety Issues Branch Division of Safety Review and Oversight TTS 4?2-4714 Mail Stop 244 Lead Supervisor:
Varren Mirners, Chief Reactor Safety issues Branch Principle Organizations:
DSRO Technical Assistance:
ENL PIN No. A-3786 Study of Beyond Design Basis Accidents In Spent Puel Fools (Ceneric Issue 82)
LLNL PIN No. A-0814 (Proposed)
Effects of Beyond Design Basis Events On Spent Puel Fools t
App!!cability:
A!! Light-Vater Reactors (LVRs)
Frejected Completion Date:
January 1988 e e e NOTICE e e a This Task Action Plan is Being Provided for Reference The Major Vork Requirements Vere Previously Defined by the Division of Systems Integration, NRR Prior to The November 24, !?85 Re-Organization The Purpose of This TAP is to Docuser.t Generic issue 82 And Pr ovide the DSRO Plans To Resolve This issue This Revision Updates the Current Status Based On The BNL Technical Assistance Program Findings As Of February 5, 1987 e e e NOTICE e e e Cl-82/ Task Action Flan Page 1/27 Revisicn 2.0 February 1987
~ _ _ _ - - - _ _
Contents Backcround Historical Background 4
I Safety Significance 4
Current Status 5
Obiectivel Alternative 1 6
Alternative 2 6
6 Alternative 3 Alternative 4 6
Alternative 5 6
Documents Affected 6
Ild D e s c r i p t i on s Task 1: Accident Initiating Events and Probabilities 7
- 'l -
Sub-Task 1.a: Loss of Vater Circulating Capabilities 7
Sub-Task 1.b: Structural Tailure of the Spent Tuel Storage Pool 7
Sub-Task 1.c: Drainage of the Spent fuel Storage Pool 8
Sub-Tast 1.d: Heavy Load Drop 8
Preliminary Evaluation 8
Task 2: Evaluation of Fuel Cladding Tallure 9
Sub-Task 2.a: Uncertainties in Osidation Propagation 9
Sub-Task 2.b: Validation of the STUELIV Computer Program 9
Sub-Task 2.c: Validation of Osidation Rate Equation 9
Preliminary Evaluation 9
Task 3: Radiological Evaluation 12 Sub-Task 3.a: Radioactive Inventories 12 Sub-Task 3.b: Radionuclides Potentia!!y Available for Release 13 Sub-Task 3.c: Estimated Releases for Accident Categories 13 Sub-Task 3.d: Comparisons of Releases with Other Severe Accidents 13 Preliminary Evaluation 14 Cl-82/ Task Action Plan Page 2/27 Revision 2.0 February 1987
4,,
Contents Task 4: Risk Profile 16 Preliminary Evaluation 16 Task 5: Evaluation of Proposed Alternatives 17 l
Task 6: Spent Tuel Fool Tragility Studies 19 Sub-Task 6.a: Seismic Evaluation of Elevated Spent Fuel Fools 19 Sub-Task 6.b: Tragility Analysis of Above Ground Spent Tual Fools 19 Sub-Tast 6.c: Tragility Analysis of Balow-Crade Spent Tuel Fools 19 Sub-Task 6.d: Spent Tuel Fool Vall Structural Capacity 19 Task 7: Valuellapact Evaluation of Proposed Alternatives 20 Resource Estinatti
%s* n.
Technical Assistance 21 DS!
21 21 DSRO
(,'
Schedules Schedules 22 Milestone Schedule 23 Annandia 2 Safety laplication of the Chernobyl-4 Accident 24 Annendir i l
Discussion of Seismic Tallure Probability 25 1
1 1
l l
l Cl-82/ Task Action Plan Page 3/27 Revision 2.0 February 1987 l
liiEGROVND Nistorical Backcround I
)
The risk of beyond design basis accidents in spent fuel storage pools were estmined in VASH-1400.
It was concluded that these risks were orders of magni-
-tude below those involving the reactor core.
The basic reason for this is the slaplicity of the spent fuel storage pool:
(1) the coolant is at. atmospheric j
- pressure, (2) the spent fuel is always subcritical and the heat source is low, (3) there is no piping whleh can drain the pool and (4) there are no antici-pated operational transients that could interrupt cooling or cause criticality.
In recent years, increasing knowledge in the geosciences has led to a better understanding that, although highly unlikely, it is possible for nuclear power l
plants in the Eastern United States (i.e.,
east of the Rocky Mountains) to be subjected to earthquake ground motion greater than for which the plants were l
dasigned.
For this reason, interest has developed in demonstrating that nuclear power plant structures and safety-related systems can safely withstand earth-l quake ground motion larger than their design earthquake ground actions (post-1973 safe-shutdown earthquake, SSE, or pre-1973 design-basis earthquake, DBE).
In addition to the potentially greater earthquake ground action, the reasons for re-examination of spent fuel storage pool accidents are twofold.
- First, spent fuel is being stored instead of reprocessed. This has led to the espan-3,_,
sion of onsite fuel storage by means of high density storage racks, which results in a larger inventory of fission products in the pool, a greater heat load on the pool cooling system, and less distance between adjacent fuel assen-blies.
Second, some laboratory studies have provided evidence of the possibil-an air cooled environment.
ity of fire propagation between assemblies in
(;Together, these two reasons provide the basis for an accident scenario which was not previously considered.
Safety Slenificance A typical spent fuel storage pool with high density storage racks can hold roughly five times the fuel in the core.
- However, since reloads typically discharge one third of the core, such of the spent fuel stored in the pool will have had considerable decay time. This reduces the radioactive inventory some-what. More importantly, after roughly three years of storage, spent fuel can be air cooled.
For
- example, the spent fuel need not be submerged to prevent
- setting, although submersion is still desirable for shielding and to reduce airborne activity.
If the spent fuel storage peo! were to be drained of water, the discharged fuel from the last two refuelings would still be " fresh" enough to melt under' decay j
heat.
The sitcaloy cladding of this fuel could be ignited during heatup. The resulting fire, in a spent fuel storage pool equipped with high density storage racks, alght spread to most or all of the fuel in the pool. The heat of combus-
- tion, in combination with decay heat, would certainly release considerable gap activity from the fuel and would probably drive " borderline aged" fuel into a moltea condition. Moreover, if the fire beccaes osygen-starved (quite probable for a fire located in the bottom of a pit such as the spent fuel storage pool),
the het stretnica would rob esygen from the uranium dioside
- fuel, ferming a Cl-82tTask Action Plan Page 4/27 Revision 2.0 February 1987
BACKCROUND liquid misture of metallic uranius, sitconica, caidised sitconica, and dis-solved uranium dioside. This would cause a release of fission products from the fuel matris quite comparable to that of molten fuel.
In addition, although
- confined, spent fuel storage pools are almost always located outside of the pri ary containment.
- Thus, release to the atmosphere is more 11kely for com-parable accidents involving the reactor core.
Current Status The safety significance of "Beyond Design Basis Accfdents in Spent Fuel Fools" has been designated as a medium priority issue.
(Ref:
Memorandum, H.R. Denton is E.J.
- Mattson, Schedule for Resolving and Completing Generic Issue 82 Beyond Design Basis Accidents in Spent Fuel Pocis," dated December 7, 1983.)
The mafor work effort on Generic Issue 82 has been performed under Technical Assistance FIN No. A-3786, with BNL, " Study of Beyond Design Basis Accidents in Spent fuel Fools (Generic Issue 82)."
Most of the work scope for Tasks 1 through 4 of this Task Action Plan was to haea been completed by BNL by September 1986.
- However, BNL's performance has not been up to expectations and the Draft report concerning Tasks 1 through 4
'[~.".
was not completed until February 5, 1987. This has resulted in a significant I
I schedule slippage.
Additional structual analyses to determine the fragility of typical spen't fuel storage pools was espected to be performed in TY-87, by BNL.
A request for (f
proposal had been forwarded to SNL, on May 7, 1986.
- However, as a result of BNL's poor performance, and apparent inability to meet schedules, it was determined that an alternative source should be found for this additional work.
A Request for Proposal was sent to the Lawrence Livermore National Laboratory on December 29, 1986. (Letter from T.P. Spels, NRC to J.S. Hirahara, DOE).
It I
is estimated that an additional 1110,000 wt!! be required for the LLNL project, and that it will require approximately eight calendar months to complete the 1
structural analyses.
The valuatimpact analysis is not included in the technical assistance statement of work. Based on available resources, the value/ impact analysis will either be added to the technical assistance statement of work or will be performed by the 1
Task Manager, DSRO.
As a result of the Chernobyl-4 accident, and the postulated involvement of the spent fuel pool in the radionuclides releases (see Appendia A),
the IAEA sponsored a " Consultants Meeting on Spent Tuel Storage Safety In At-Reactor Pools" in Vienna, Austria.
December i through 5, 1986.
The IAEA proposal to convene an Advisory Group Meeting on the Safe Spent ruel Storage and Post Accident Spent Fuel Manageneat (in May 1987) was adopted.
Cl-82tTask Action Plan Page 5127 Revision 2.0 February 1987
4 OBJECTIVES I
Objectives No generic solution to this potential problem has yet been identified.
Several
- t? :: esist however.
Alternative 1 Reprocess the spent fuel to reduce the spent fuel storage pool inventory.
Alternative 1 Install spray headers to provide cooling even when the spent fuel storage pool
- 1. drainad and not refloodable. (May require seismic systems.)
Alternative 1 Comp.: taentalize the spent fuel storage pool by installing partior,s (and I
individual coolant supply diffusers for each compartment) to limit the estent of the accident.
Alternative 1 Modify storage racks to improve air circulation, should the spent fuel storage pool drain.
e..-
Alternative i Develop administrative procedures to isolate the recently discharged fuel, to Itait the estent of the accident.
(
Other alternatives may ba defined as work progresses on this issue.
Tor ex-
- ample, though unlikely, delaying removal of the spent fuel from the reactor until the decay heat levels are reduced to a sufficiently low value to pre-clade propagation of the sitcaloy osidation.
The objective of this Task Action Plan is to esamine the need for the above identified alternatives and to determine the value/ impact (or cost effectiveness) of each of these proposed alternatives.
Documents Affecter The following documents may be affected as a result of work performed:
Standard Review Plan 9.1.2
- Spent Fuel Storage.
Standard Review Plan f.1.3
- Spent Fuel Pool Cooling and Cleanup System Standard Review Plan 15.7.4
- Radiological Consequences of Tuel Hand!!ng i
Accidents Regulatory Cuidt 1.25 - Assumptions Used for Evaluat4ng the Potential Radiological Corssequences of A Fuel Handling Accident in the fuel Handling and I
Storage racility for Boiling and Pressurized Vater Reactors Changes to the St andard Technical Specifications af f ecting cool-down times prior to fuel stosage and/or limitations on the location of fresh assemblies in the spent fuel storage pool may be required.
l CI-82/ Task Action Plan Page 6/27 Revision 2.0 February
!?B7
HH1 Task 1 : Accide.nl Initiatine Events ing Probabilities Backcroond The purpose of this task is to determine the potential events which can result in loss of water from the spent fuel storage pool and to estimate the frequency of these events.
The events to be considered include:
estreme esternal phenomena (such as earthquakes,
- tornados, hurricanes, floods and aircraft accidents);
prolonged loss of all cooling capability (pool boils dry);
- massive pool failure from internal accidents (shipping case drop, crane collapse, turbine alssiles);
- rapid draining of the pool due to circulation system failure, operator error, or malicious act; and e."
sabotage such as the addition of reactive chemicals or deliberate damage to the pool or cooling system.
The dominant contributor (s) to loss of water from the spent fuel storage pool e
k, will be identified.
It is espected that the seismic event will be the dominant event.
This task is divided into four sub-tasks, based on the type of event being evaluated.
Es11 Scone Sub-Task 1.3 : Loss 31 Vater Circulating Capabilities Identify and estimate the frequency of events which can result in the loss of the spent fuel storage pool water circulating capabilities, or which can result in loss of the heat sink.
hk -Ta s k 1.JL - s t r u c t u r a l r a i l u r e 21 1hi s c e n t Im.s_1. s t o r a c e Z121 Identify and estimate the frequency of events which can result in severe damage to the spent fuel storage pool, the resultant damage causing a significant, or total, loss of water.
Cl-IIITask Action Plan Fage 7127 Revision 2.0 February 1987
IME.l.
l Sub,,7a31 L.s : D r a i n a nt.gj, J.h.g Scent I.E2l Storace f_gfj, Identify and estimate the frequency of avents which can result in the loss of water from the spent fuel storage pool by draining the inventory.
But-Task M : He s e f 111A D. Lip.
l Identify and estimate the frequency of events which can result in sevare damage to the spent fuel storage pool as a result of re-fukling operations, the resultant issage causing a significant, or total, loss of water.
Preliminati Evaluation The fo!!owing estimated probabilities for the loss of integrity of the spent fuel pool and resultant uncovery of the stored fuel for each event group were provided by BNL on February 5, 1987:
Estimated Probability Per R-Y Accident Mil.' stone ( BVR )
Cinna (PVR)
Loss of Pool Cooling Capability 1.4 10~'
5.7 a 10" (1)
. +.. -
Fool 2.2 10-5 1.6 a 10 ~5 Seismic Structural Tallure of Structural Tallure from Tornado Missiles (1.
10~0 (1.
- 10
10-10 (1.
- 10 -10 Structural Tailure from Aircrash (1.
~
Structural railure from Turbine Missiles 4.0 10
- O (2)
Pneumatic Ses! Tallure (during re-fueling)
- 0 (2) 1.0 10
Structura! Tallure from Shipping Ccsk Drop (4) 3.1 10
3.1 10
-5 i
Total Estimated Probability 5.5 10
4.9 10
~'
(1) Vith credit for a third cooling system. Other PVRs which typically have two of spent fuelcoojingsystemswouldhaveanestimatedfueluncoveryfrequency about 1.0 10 (2)
Typical PVRs may have failure frequency due to turbine missiles on the
,7 order of 4.0 10 but Cinna's pool 16 shielding (by the containment) from the turbine.
(3) BVRs de not have a similar pneumatic seal.
(4) After removal of accumulated inventory resumes, either for reprocessing or to alternative.atorages locations (on site or away from site).
Cl-82/Tast Action Plan Page 8127 Revision 2.0 February 1987
IA1E 1 1L31 1 _ Evaluation 21 I.23.1 Claddine Failure Backcround The results of work performed by Sandia (Ref:
" Spent Fuel Heatup Storage,"
NURECICR-0649, March 1979, and "The Potential for Propagation of a Self-Sus-taining Zirconium Oxidation rollowing Loss of Vater in a Spent Tuel Storage Pool," Pisano, N.A.,
11_11, D r a f t R e p o r t, J anu a r y 19 8 4 ) have suggested that in certain fuel racking configurations (a) a self-sustaining sitconium-air o:Adation reaction can be initiated, and (b) this self-sustaining reaction can propagate from one region of the pool to another. There are large uncertainties associated with the phenomenology of strcaloy esidation and its propagation in spent fuel assemblies.
The results obtained were based on both experimental simulation and computer modeling. The computer program developed is SFUELIV.
This task is divided into three sub-tasks.
I Egil Scoot 3ub-Task 1.1 - Uncertainties in Oeidation Procacation 6 *..
The uncertainties in the sitcaloy esidation propagation calculations under inadequate roca ventilation conditions (most typical of the spent fuel storage pool structures) will be determined using the STUELIV computer program.
A sensitivity study covering hot spent fuel decay power in the 20 to 90 Kw/MTU (j
range will be performed. The maniana decay power a!! owed for adjacent spent fuel required to prevent osidation propagation under both adequate and inade-quate roca ventilation conditions will be determined.
Sub-Task 1.1 - Validation gi lhi SFUEt1V Conouter Procram The STUEL1V computer program w!!! be validated by comparing the calculated results with esisting Sandia National Laboratory small-scale experimental re-sults.
Aub-Iggi 1,3 - Verification 31 0xidation Rate Ecuation The reaction rate equation for the osidation of sitcaloy cladding in air used in the SFUELIV computer program will be revised based upon current state-of-the-art understanding of the associated phenomena and by performing sensitivity s tudies on the sitcalcy-air reaction r ate correlation.
Pfalininary Evaluation A
review of current state-of-the-art data suggests that the reaction rate equation used in the original work by'Sandia is appropriate.
For temperatures in the 800-!!50 't range (1470-2100 r) new data indicates that the Sandia correlation is valid, for esposure periods of less than 30 minutes. For longer periods, the correlation may be non-conservative.
CI-82/ Task Action Plan Page 1/27 Revision 2.0 Tebruary 1987
IA11 1 T;.e STUELIV computer program is a finite difference solution of the transient equation for heating of the fuel rods considering:
- The heat generation rate from the decay heat and oxidation of the cladding, l
- Radiation to adjacent assemblies and pool walls, and
- Convection to buoyancy-driven air flows.
The key assumptions and limitations of SFUELIV are:
i l
i
- The water drains instantaneously from the pool, l
- The geometry of the fuel assemblies and racks remains undistorted,
- Temperature variations across the fuel rods are neglected,
- The air flow patterns are one-dimensional, and i
- The spaces between adjacent holders are assumed to be closed to I
air flow.
l After the water is drained from the spent fuel pool, the rods heat up until the i
bourancy-driven air flow is sufficient to prevent further heatup.
If the decay i
heat level is sufficient to heat the rods to about 700
'C (1650
'F) the oxidation becomes self-sustaining.
o..
I o-BNL has concluded that:
i
- The likelihood of cladding fire initiation is not very sensitive to the osidation rate equation, l
- The osidation rate equation in STUELIV is a reasonable representation
()
of the available data, and
- The likelihood of cladding fire initiation is most sensitive to the i
decay heat level and the storage rack configuration (which controls the estent of natural convection cooling),
i It was also concluded that the esidation propagation to older adjacent assen-blies is likely if the decay heat level of the older adjacent assembly is high enough to heat that assembly to within 100 to 200 "C (200 to 400 'r) of the self-sustaining osiation temperature. The radiation heat transfer from the burning assemblies would be sufficient to raise the temperature of the older adjacent assembly to the self-sustained oxidation limit.
l I
i The following descriptions of spent fuel storage rack configurations are pro-l vided:
1 (1) Blah densit, EMR configuration:
In this configuration, the fuel assemblies t
I are tightly packed with neutron absorber asterial used in the rack structure to l
replace the reduced water moderator for criticality control.
The center-to-l center assembly spacing is 10.25 inches, the open gap between assemblies is 0.7 inches.
This configuration is in use in nearly all PVRs, and is r e 'f e r r e d to as high density storage.
1 Cl-82/ Task Action Plan Page 10/27 Revision 2.0 February 1987
IAH I (2) Cylindrics! PVR confleuration: This configuration is typical of the early 1
rack
- designs, used before at-reactor storage of spent fuel was required.
The i
center-to-esnter assembly spacing is 12.75 inches, in a closed cylind:ical stainless steel rack. The typical cross sectional area of a PVR assembly is 8.4 by I.4 inches. This is referred to as low density storage.
I (3) Cylindrical gyg conficeration:
This configuration is typical of early BVR J
spent fuel storage rack designs.
The center-to-center assembly spacing is 8.5 inches. The typical cross sectional area of a BVR assembly is 5.3 inches. This is ca!!ad low density storage.
(4) D i r e c t lLELL [yg c on f i n o r a t i on : In this configuration the BVR assemblies are stored in 6 inch center-to-center racks, with a 5.3 inch open space between rows. No additional neutron absorber material is required in the rack structure for criticality control.
This is considered to be a high density storage i
i configuration for BVRs.
1 i
Because of limitations in the STUELIV computer program, BNL has limited the BVR l
spent fuel analyses to the low density cylindrical configuration. The SFUELIV computer program does not account for air flow between adjacent
- holders, a
- conservative assumption which was based on the storage rack design.
The self-
,,, f sostaining oxidation analysis is $vvern6d by the BVR channel bos design, the air flow through the assembly.
The STUELIV results are not significantly in-fluenced by the BVR rack design.
t The critical cooling time is defined as then decay time to reduce the internal I
heat generation rate to a low enough value to preclude the cladding temperature l
from esceading the self-sustained oxidation limit under air cooling. Consider-ing a fuel cycle of not less than one year, the critical cooling time can be I
converted to a conditional probability of fire initiation.
1 The estimated likelihood of self-sustaining oxidation for various spent fuel I
rack configurations is provided below, based on the BNL draft report:
1 I
5 peat Puel Rack Inlet Orifice Minimon Decay Critical Cooling Conditional I
Configuration Diameter Power Time Proballity I
(inches)
(Kw/MTU)
(days)
(per year)
I l
I High Density PVR 10 11 360 1.0 t
5 6
700 1.0 i
i 1
Cylindrical PVR 5
90 10
- 0.0 l
3 45 50 0.14 1.5 15 250 0.7 l
j l
Cylindrical BVR 3
30 30 0.08 l
1.5 14 180 0.5 Cl-82fTask Actica Plan Page 11/27 Revision 2.0 February 1987 j
Th1L 1 1111 1 - Radioloalcal Evaluation Backaround
.na a ;. i u. ; y of radionuclides contained in apent fuel assemblies depends on the operating history and the size of the plant.
During refueling ~the freshly discharged fuel contains a large inventory of isotopes with short half-lives in the range of approsisately one to thirty days.
These isotopes decay over the course of a year, until the nest refueling outage.
The cider fuel contains radionuclides which have longer half-lives. The older fuel approaches a decay rate which is, inversely properation to the dec4y time.
For eaaaple,. alter four years, the spent fuel contains approsisately one-forth of the specific activity of one-year old fuel.
During each ref ueling outage approminately one-third of a PVR core and about one-forth of a EVR core are off-loaded to the spent fuel storage pool.
It is noted that releases for an accident involving the core are basically noble gases and halogens, while for a spent fuel storage pool accident the
' releases are primarily alkall metals (cesica).
Therefore, it may not be pos-sible to directly relate the consequences of a spent fuel storage pool acci-d'e n t s to a similar core-damage accident because of the different radionuclides involved.
This task is divided into four sub-tasks to define the spent fuel storage pool
(.
source tern and er,timate the releases during accidents.
In addition, releases from these accidents in the spent fuel storage pool will be compared to re-leases for other severe accidents (for example, core melt sequences).
W111 Segne Sub-Tosk 3,3 - Radionuclides Inventories The ORICEN computer program (Ref:
"The ORNL 1sotope Generation and Population Code," ORNL-4628, March 1963) will be used to determine the radionuclides inven-tory of the spent fuel as a function of decay time.
Separate inventories will be calculated for activation products in the fuel assembly hardware and clad-
- ding, and for the fissions products and actinides sealed in the fuel elements.
The data will be obtained for a reforence BVR and a reference PVR.
Millstone Unit I and R.
E.
G!nna have been selected as the reference plants for this study.
A comparison of the radionuclides inventory at different decay times, up to the iime when the spent fuel storage pool reaches a capacity load, to the equilibrium inventory of a reactor core will be provided.
Cl-82/ Task Actica Flan Page 12/27 Revision 2.0 February 1987
4
.Thith 1 Sub-Task 1 ) : Radiennelides Potentially Available 131 Release The source term for any postulated accident sequence is defined in terms of:
- the amount (curies) of each radionuclides,
- the composition, physical and chemical form of each radionuclides,-
- and, the time of release of the radioactivity to the environment.
The physical and chemical processes that would take place in a dr ined spent fuel storage pool are not well characterized at the present time.
It may therefore be necessary to use engineering judgement to estimate the source term.
The STUELIV computer program does not account for relocation of the reaction products (molten un-esidised cladding, fuel dissolved in molten sir.
- conism, etc).
Also the degree to which esposed UO would osidise to U O and 2
g reduce the release of less volatile fission products has not been studie The estimate of the fraction of each radionuclides release will be determined based on available data and on engineering judgement.
In the later case, the as'sumptions will be provided.
Sub-Task 1.1 - Estimated Releases [gt Accident Catacerles r
The dose equivalent of the release estimates depends on many factors including
(,
the location of the spent fuel storage poc! and equipment operability (for esample, with and without filters in the fuel storage structure).
Cesium, for
- osample, is espected to be released as an aerosol and filters may provide an effective removal mechanism.
If the fuel storage structure cracks or if fans fall to function due a seinic event, the release may be substantial. The estimated release to the environment wi!! be determined for each accident category with a significant frequency of occurrence.
t Sub-Task 1.1 : Comparisons 11 Releases ELLh Other Severe Accidents The radionuclides release for spent fuel storage pool accidents sequences will be compared to VASH-1400.
The VASH-1400 spent fuel storage pool estimates may be non-conservative as a result of more frequent discharge of spent fool then originally assumed.
A comparison to other severe accidents wi!! also be performed, for reference.
Cl-82/ Task Action Flan Page 13/2y Revision 2.0 February 1987
1AhL 1 l'
Profitinary Evaluation The radionuclides inventories for both the EVR and the PVR spent fuel pools were calculated using the ORICEN2 computer program and the actual operating and i
f:- $: : p ' steties for Millstone I and Cinna. For both plants, the noble gases and halogens in the spent fuel inventory are a small fraction of the inventory l
in an equilibrium core at shutdown, except for the freshly discharged fuel. The l
cesium and strontium inventories are more than three times the equilibrium Inventory.
l l
The fission product release fractions were c a l c u l a t,6 d for two limiting cases in which a Zircoley cladding fire is assumed to occur.
In the first case the l
claddin; ccttustien is assumed to propagate throughout the entire pool and the entire inventory is involved, in the second case the inventory is limited to c r.l y thc mest recently discharged fuel batch.
In order for a Zircaloy cladding fire to occur the fuel must be recently l
discharged (between 30 and 150 days in a EVR, and between 30 and 250 days in a PVR).
If the spent fuel in stored in high density racks, then the probability for a Zircaloy fire in a PVR is essentially 1.0.
A..re-evaluation of the cladding fire propagation estimates indicates that there is a substantial likelihood of propagation to other fuel bundles that have been discharged within the last one or two years. Subsequent propagation to low power bundles by thermal radiation is highly unlikely, but with a substantial amount of fuel and cladding debris on the pool floor, the coolability of even I
the low power bundles is uncertain.
(..
Tor a
less severe accident in which the fuel is esposed to air but does not reach temperatures high enough to ignite the Zircalor
- cladding, fuel pin failure could occur resulting in a release of the noble gases and halogens. Two i
cases have been considered by ENL. In the first case the entire pool is assumed to be drained but the decay period is one year since the last discharge and 50%
of the pins are assumed to perforate or rupture.
In the second case it is assumed that only part of the fuel is uncovered 30 days after the last discharge and all it.e rods fall.
The offsite consequences for the Zircaloy fire cases have been calculated with the CRAC2 computer program. The following assumptions were used by ENL:
l
- a generalized site surrounded by a constant population density of 100 persons per square mile; I
- generalized meteorology (a uniform wind rose, average weather c:nd!!! cts);
l
- the population in the affected sones are relocated after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
Cl-82/Tast Action Plan Fage 14/27 Revision 2.0 February 1987
IME 1 There are several unusual characteristics of a spent fuel accident that cause
)
I somewhat surprising results in the radiation esposure calculations.
Specific-
- ally, the radiation esposure is insensitive to fairly large variations in the estimated release. This is due princi;411y to the health physies asumptions within CKAC.
For the long lived isotopes (predominately cesica),
the esposure is dte mainly to espesure after the area is decontaminated and people return to their homes. The CEAC code assumes that decontamination will limit the esposure of each person to 25 rea. Thus, for this type of release the long tera whole body dose is limited by the population in the affected sectors (about 0.1 million people in three of the 16 sectors downwind of the release within a 50 I
mile radius) or about 3,000,000 person-rea.
i A more sensitive indication of the consequences of a spent fuel pool accident is the interdiction area (the area with such a high level of radiation that it is assumed that it connot be decontaminated).
The interdiction area of the worst case spent fuel pool accident is calculated to be 224 square miles, about two orders of magnitude higher than the interdiction area for a severe core accident (1 to 10 s'guare miles).
1 The consequences estimated for the Zircaloy fire cases are summarised below:
Case Description Vhole Body Dose Interdiction Area (person-rea (square miles)
Per Event)
- 1. Total inventory 2.6 : 10' 224 38 days after last discharge 58 mile radial sons
- 2. Total inventory 2.6 10' 215 l
y0 days after last discharge 50 mile radial sone
- 3. Last fuel discharge batch only 2.3 a 10' 44 l
90 days after last discharge 50 mile radial sone
- 4. Total inventory (1) 7.1 10 224 1
l 30 days after last discharge 500 mile radial sone (1)
A sensitivity case.
50 mile radial sone is proposed in NUREG-1150 for consequence calculations.
Cl-82/ Task Action Plan Page 15/27 Revision 2.0 Tebruary 1987
.T1.!d A l
M i - gigi Proffle B a c k e r oRiul Preliminary results for this project have indicated that a few older plants may have s;er.t fuel storage pool designs that are vulnerable to seismic events.
In particular those plants which were designed to the pre-1973 design-basis earth-quake ground motion wi!! be evaluated.
It is noted that very little information esists for the fragility of the spent fuel storage pool when subjected 'o a beyond design basis seismic event.
The risk profile may therefore have large uncertainties for the seismic events.
Ta.k 4 of this project w!!! provide additional data concerning the fragility of
)
the spent fuel storage pool and the risk profile will be updated following j
c cs;,1 s t i on o f the Task 6 effort.
I Vell ScoDe A risk profile will be d^veloped for the more vulnerable spent fuel storage pool
- designs, using a prG 2bilistic analysis.
The risk profile will incicde
- c. consideration of the uncertainties associated with sitcaloy cuidation propa-gation coupled with the estimated likelihood of the loss of the spent fuel storage pool integrity.
The estimated release fractions of radionuclides, dependent on the event which leads to loss of integrity, from Sub-Task 3.c will be used to determine the
' (,
risk profile.
Preliminary Evaluatlen The risk profile for each plant, based on the dominant initiating event proba-bilities and consequences provided above are summarised below:
i Accident Spent Tuel Fool Health Risk Interdiction Risk Initiator Fire Probability (person-rem (square miles per I
(per reactor year)
(per reactor year) per reactor year)
I' FVR Selseic 1.6 10 42 3.6 10"
~
i 10~'
81 7.0 10 '
i
~
FVR Cask Drop 3.1 s FVR Total 4.9 10~
127 1.1 : 10" l
BVR(2) Seismic 1.8 10 4
7.6 a 10 l
~
l 10"*
BVR Cask Drop 2.5 10~'
4 1.1 :
BVR Total 4.4 10~'
12 1.9 10~
(1) FVR-high density rack, full inventory, 1.0 probability of Zircaloy tire.
(2) BVR-cylindrical rack, last discharge only, 0.08 probability of fire.
I cl-82/Tast Action Plan Page 16/27 Revision 2.0 Febroary 1987 i
l
Illi 1 i
1111 1 ; E v a l u a t i o n gi F r e c o s e d A l t e r n a t i v e s Backcround The alternatives proposed for the resolution of Generic Issue B2 have the potential to reduce the consequences of a beyond design basis accident in the spent fuel storage pool.
Additional alternatives, which may reduce the likeli-hood of the loss of the spent fuel storage pool integrity,
- say, although unlikely, be proposed as work progresses on this project.
The potential risk reductitn alternatives identified by BNL are:
- Reduction of the stored radioactive inventory in the pool.
l
- Provide adequate space around newly discharged fuel bundles to assure sufficient air circulation, to prevent cladding fires and propagation.
- If the dcminant risk is due to loss of cooling, provide an additional cooling system.
- If cask drop is dominant, improve procedures and equipment.
3
.The potential accident mitigation alternatives identified by ENL are:
lU:
- Post-accident spray system.
- Environmentally qualified air-filtration system thigh temperature, i
high flow rate system) to reduce the releases.
Ver k 11111 The reference (base case) release of radionuclides from the spent fuel storage pool will be estimated for the more vulnerable designs based on full capacity of the spent fuel storage pool, The amount of sitcaloy osidation will be estimated assuming that current spent fuel storage racking procedures are used.
The reduction in the estimated release of radionuclides from the spent fuel storage pool will be determined for each proposed alternative.
The mean whole-body dose rates will be estimated using applicable esposure data.
For
- example, if the event is seismic then it may be assumed that a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> esposure is applicable since evacuation may be hindered.
(Ref:"In-Plant Considerations for Optimal Off-Site Response to Reactor Accidents," NUREC/CR-2f25, November !?B2.)
Cl-82/ Task Action Plan Page 17/27 Revision 2.0 February
!?B7
IE.E.i lui f.: S D e al. E1Lt1 Eg.91 T r a e i I i t v Stueies Batteround Preliminary results from this project indicate that a few older plants may have spect fuel storage pools that are fragility lnerabletoseismicevents.
vy Very little information esists for the of spent foal storage pools and there is a need to understand the importance of the large ! cads laposed by full or naarly full spent fuel storage pools during seismic events. The additional leading due to full spent fuel storage pools is tapected to be particularly
..wese for pools which are raised relative to the concrete foundation (base-ret). Nest EVRs and FRVs have an olewated configuration. Several FVR spent fuel statage pools are located in separate buildings (the scalliary building or the fset storage building) and have direct contact with the ground without the prefection of a thick reinforced baseast.
For these in-ground spent fuel stor-age pools, failure may occur due to soil-structure interaction during a seismic event.
The BNL estimated probabilit! of a seismic event resulting in the loss of
. integrity of the spent fuel storage poc! is approsisately 2 10~
per reactor
- year, an order of magnitude greater than previously estimated in VASH-1400 or in the Generic Issue prioritisation evaluation study (NUREC-0933).
Appendia E is provided to address the change in the seismic probability.
The BNL study concerning accident initiators has identified an additional I
(
potential scenario which could result in a loss of water from a spent fuel i
L' storage pool. The inadvertent dropping of a spent fuel shipping cask on a vulnerable spent fuelpoolwallasaresulto{humanerrorhasbeenfound to i
have an espected probability of 3!
10~
per reactor year.
Although originally addressed in previous studies (for
- example, VASH-1400),
the i
estimated likelihood of a dropped shipping cask was based on mechanical i
failure of the crane and found to be two orders of magnitude less. It is noted I
that newer plants, meeting the requirements set-forth in Standard Review Flan 9.1.2 (Rev 3), " Spent Tuel Storage," are designed such that the transfer pit is separate. ' rom the remainder of the spent fuel pool, thereby essentia!!r pre-I cluding !
event.
This task has been divided into four sub-tasks to address these concerns.
(1)
The objective of a fragility evaluation is to estimate the peak ground motion escaleration value ior which the seismic response of a given component (i.e.,
structural element or equipment) located at a specific point in the structure exceeds the component capacity resulting in its failure.
l Cl-82/ Task Action Flan Fage 18127 Revision 2.g February
!!87
4 IMI I MAIk 113pe I
Sub-Task fin - 1elsmic Resconse 11 riegated Seent 1111 Pools Based on a seismic evaluation performed by LLL (NUREC/CR-2024), the sensitivity j
^
of the seismic capability of a reactor building with an elevated spent fuel storage pool to the fuel pool 1cading as the pool capacity is approached will be determined. The elevated design is characteristic of both BVRs and PVRs but this designs appears to be more severe for BVRs. Therefore, an older SVR design will be selected to estimate the vulnerability. The structural repsonse for the safe ' shutdown earthquake ( 0.2 g for the reference Millstone Unit 1 BVR) for normal loading and with a f573 spent fuel storage pool will be obtained. These data will be used to assess t!ie importance of the leading assumption.
(
Sub-Task 1,j, Fracilitt Analvsis 1,[, M,1y2 Cround Sornt. hil hil,g, A full three-dimensional computer model of an elevated spent feel storage pool, and associated structures, will be constructed.
This model will be used to calentate the localised ottess and strain caused by the safe shutdown earth-quake' and compared to the entinated structural capability of the design.
The g.,
driving spectra will then be scaled up for down) to obtain the failure thres-hold.
Finally, a fragility curve will be derived for input into a revised risk profile (from Task 41.
SulL-TALI.L G : F r a e l I i t v hlggli al J,110 w-C r a d e S c e n t f.111 falla
{
The work scope defined in Sup-Task 6.b will be repeated for a
below-grade spent fuel storage pool design, to evaluate soil-interaction failure.
t A
f's ! !
three-dimensional computer model of an in-ground spent foal storage
- pool, and associated structures, will be constructed. This model will be used to calculate the localised stress and strain caused by the safe shutdown earth-quake' and coepared to the estimated structural capability of the design.
The driving spectra will then be scaled up (or down) to obtain the failure thres-bold.
Finally, a fragility curve will be derived for input into a revised risk profile (from Task 4).
l l
k9b-1LL.k.Lt : 5Dtnt 1111 f. ski 2111 Etruttural Capacite The purpose of this sub-task is to estimate the structural capacity of typical spant fuel pool walls to within the impact loads associated with a dropped i
spent fuel shipping cask.
l GI-82fTask Action Plan Page 17127 Revision 2.0 February 1787 lE
.T1.E 1 Ta s k
.7. - YJ11JtlltDR1 f v a l u a t i on M P r o c e s e d A l t e r n a t i v e s Backeround
'a Januarr 1983, the NRC published guidelines for performing the regulatory analysis required for a broad range of HRC regulatory actions.
The principle po: pose of the guidelines is "to ensure that the NRC regulatory desisions are based on adequate information concerning the need for and the consequences of a proposed regulatory action and to ensure that cost-effective regulatory
- actions, consistent with providing the necessary protection of the public health and safety and common defense and security, are identified."
The purpose of this task is to perform the required valcellspact assessment for the alternative resolutions proposed for Generic issue 82 "Beyond Design Basis Accidents in Spent Fuel Storage Fools."
NURICICR-3568, "A
Handbook for Valce-lapact Assessment," dated 1983, will be used to perform the required assessment.
s ou o
ELLL Scope for each proposed alternative. The guidelines
{'.
Perform a valuellspact assessment provided in NUREC/CR-3568 are to be used to determine the valce/ impact for each alternative, as identified in Task 5.
The results of the fragility studies performed in Tast 4 wi!! be used to update the Risk Frofile evaluation infor-nation from Task 4.
The valuelimpact data, as well as the results from other tasks will be used by the NRC to prepare a Regulatory Analysis for the resolution of CI 82.
In order to provide additional insight into the requirements of a Regulatory Analysis, HUREC-Il09,
" Regulator! Analysis for the Resolution of Unresolved Safety !ssue A-44, Station Blackout," Draft Report for Comment. January, 1986, will be usei as an illustrative esample.
CI-8217ask Action Flan Page 20127 Revision 2.0 February 1987
RESOURCE ESTIMATES Technical Assistance The major work effort on Generic Issue 82 has been performed under Technical Assistance FIN No. A-3786, with BNL, " Study of Beyond Design Basis Accidents in i
Spent Tvel Fools (Generic Issue 82)." Most of the work scope for Tasks 1
through 4 of this Task Action Plan was to have been completed by the and ci Tiscal Year 1986.
Additional structural analyses to determine the fragility of typical spent fuel storage pools was espected to be performed in FY-87, by BNL.
1 The TY-85 and TY-86 cost for TIN No.
A-3786 was $233,000.
The proposed FY-87 i
resources for T1H No.
A-3786, for work associated with Tast 6 of this TAP, was estimated to be an.dditional $242,000.
The total expenditure for TIN No.
A-3786 was fised at $233,000 because of the poor per f ormanere by ENL is meeting i
scheduled milestones.
Additional structual analyses to determine the fragility of typical spent fuel storage posts was espected to be performed in TY-87, by ENL.
A request for proposal had been forwarded to BNL, on May 7, 1984.
- However, as a result of BNL's poor performance, and apparent inability to meet schedules, it was determined that an alternative source should be found for this additional work.
i
- A. Request for Proposal was sent to the Lawrence Livermore National Laboratory 8 *~~ '
on' December 29, 1986. (Letter from T.F. Soeis, NRC to J.S. Hirahara, DOE).
It I
is estimated that an additional 6110,000 will be required for the LLNL project, and that it will require approximately eight calendar months to complete the I
I structual analyses.
Ril (Old HEA Oreanizationi It is estimated that previous ef; orts associated with Generic Issce 82 in the Division of Systems Integration was 1.0 staff-years.
Based on an estimate of
$100.000 per staff-year, the DSI cost was $100,000.
Divison d Saf ete Review AM Oversicht DSRO has been assigned as the lead organization for the completion of Generic Issue 82.
It it estimated that an addition staff-year effort will be required to resolve this issue. The DSRO cost, based on the $100,000 per year estimate, is therefore $100,000.
The total estimated resources for the resolution of Generic Issue 82, "Beyond Design Basis Accidents in Spent Tuel Storag,s Fools," is 2.0 staff-years (in-I house) and $343,000 in Technical Assistance (6233,000 under BNL A-3786 and i
6110,00 under LLNL A-0814).
The NRC development cost for the resolution of Ceneric Issue 82 is therefore estimated to be 5543,000.
Cl-82fTask Action Plan Page 21/27 Revision 2.0 Febrcary 1987
SCHEDULES Schedules The current status of Generic Issue 82, as reported in the first Quarter TY.87 I
Update to the Generic !ssue Management Control System, has been modified to show completion of the technical effort and development of a valvelimpact I
assessment and regulatory analysis for Generic !ssue in January 1988.
The I
schedule slippage is a direct result of the delays in the Technical Assistance i
program with BNL.
In addition, without the BNL input insufficient information i
was available to determine if additional structural analyses were necessary. It has been determined that the structural analyses are required to determine not i
only the estimated failure probabilities, but also the realistic risk reduction for the proposed alternatives.
t A request for an additional 5110,000, in FY-87 funds, has been developed and I
forwarded for management consideration. These funds are needed for the proposed project at LLNL to perform the necessary structural evaluations associated with i
Tart 6 ef this TAP. Issuance of the work order is espected in March 1987.
A milestone
- schedule, in the format used for the Generic
!ssee Management Control System (ClMCS) is provided below.
's **:.
G CI-52fTask Action Plan Page 22f27 Revision 2.0 February 1987
)
i SCHEDULES i
Milestone Sehedele Generle isste il i
Bevond Desien Raill A c c i d e n t s in S c e n t Egt1 S t o r s e e P o o l s Milestone Original Current Actual D e r. e r i c !ssus Assigned to DS!/NRR 12/07/83 Agreement with RES to study sitcaloy cuidation 01/84 01/23/84 CIMICS Approved by Director DS!
02/84 02/02/54 RES assessment on sitcalor osidation 04/84 05/21/84 RES work issued for review 05/84 05/05/84 FIDS Tor TY-85 T/A 04/84 06/84 06/
/84 Issue RTP 08184 08/27/84 Award T/A contract 09/84 11/
/84 Complete preliminary evaluation 04/85 05/
/85 Evalsate preliminary results of T/A 08/85 09/85 11/
/S$
Espand work scope for risk study 07/85 09/85 11/
/85
!ssue RTP f or fragility studies to ENL 10/85 05/07/86
!ssue RTP for fragility studies to LLNL 12/29/86 l
Authorise fragility studies 03/87 I
ms "-:.
Task 1 - Initiating Events / Frequencies 07/86 08/12186 Task 2 - Zircaloy Osidation Evaluation 08/86 10/86 10/21/86 i
BNL completed Draft Report 10/86 02/05/87 I
Task 3 - Consequences / Radionuclides Inventory 09/86 11/86 02/05/87 1
Task' 4 - Risk Profile (Preliminary) 09/86 11/86 02/05/87 i
Task 5 - Evaluation of Alternatives 09/86 11/86 02/05/87 t
Paar Rsview cf ENL Draf t 03/87 BNL issue final NUREC Report 06/87 Task 4 - Tr agilit y Stuties 05/87 10/87 Task 7 - Value/ Impact Evaluation 06/87 11/87 Draft Regulatory Analysis 07/87 12/87 Evaluats Future Schedule 08/87 01/88 Cl-82/ Task Action Plan Fage 23/27 Revision 2.0 February 1987
APPEND 1y &
The following it. formation was presented in the June 19, 1986 issue of Nucleenics Veek19, Volume 27, Number 25.
SV111 IIFLRJ IAll CHERNOBYL 19)) RE ACTOR AND SPENT JEL COOLING Chernobyl-4 suffered a double less-of-coolant accident (LOCA), according to a Swiss scientist who has studied the accident details for the past month and a half.
Carlos Ospina of the Swiss rederal Institute for Reactor Research (EIR) i t.
Vueren!!ngen says that loss of coolant to reactor fuel channels beginning Apr:1 26 must have been icilowed about a week later by loss of coolant to the spent fuel storage poc!
nearby, containing an estimated 75 metric tons of spent fuel. Only this " hyper-loca" hypothesis, he believes, accounts for daverse fallout patterns -- corresponding to both fresh (It4-day) and spent fuel -- measured over Europe between May I and May 12.
Ospina, who worked on graphite reactors in Britain in the 196gs, is corrently fuel cyclic specialist at ElR. Chernobyl-4 had been having fuel failure problems since startup, some of which were detected by Swadish monitors. Published Soviet data indicates that iodine-131 releases from Chernobyl were about 17 times higher in 1984, and three times higher in 1985, then those from the contemporaneous station at Kursk. His scenario assumes that a local fuel failure, perhaps a week before the accident, went undetected during reactor-sector control at low power and eventually degenerated and set on fire the adjacent graphite. Tuel began to melt under
(,
the high temperatures, pushing pressure inside the vessel to 2Dg bars and breaching leak-tightness. The resultant " wave shock" overpressurised the steam headers and led to a tremendous steam esplosion in the large steam separator located just outside the reactor building.
While the initial damage and hugh releases were going on, "the nest catastrophe was already underway," Ospina said. This was less-of-coolant to the spent fuel storage pool, which was deprived of power and subjected to high ambient temperatures. Ospina believes that the water in the pool rapidly drie6 out in an esothermic reaction and esposed the fuel to air, corroding it and leading to a second wave of releases quite different from the first one, The "deflAltive failure" of the fuel pool happened about a week after the original accident, he said. This can be seen in the fallout differences in Europe during two time periods, he said. In the period May 1-4, fallout included both iodine and cesium, while in the May 7-12 period, it was mostly cesium. It also explains the large differences in s
radioactivity concentrations measured in southern Switzerland and northern Italy, or Bavaria and North Rhine Vestphalla, during the same periods of tame, ne saad. The main lesson he drew from the analysis is that wet spent fuel storage is " risky" and should be done only under containment structures, or for short periods of time.
j
- Ann Maclachlan, Davos, Switzerland.
Cl-82tTask Action Plan Page 24127 Revision 2.0 February 1987
APPENDff 1 Discussion of Seismic Failure Probability For i
Ceneric Issue 82 "Beyond Design Basis Events in Spent Fuel Fools" 1
The purpose of this discussion is to identify the changes in the estimated I
frequency of a seismic related failure of a spent fuel storage pool.
In this 1
discussion the term failure is used to identify a breach in the spent fuel storage pool wall and liner of a large enough area such that available makeup is insufficient to prevent the complete draining of the pool.
VASH-1400 (Reference 1.
Section 5 to Appendis 1) discussed spent feel storage pool failures.
Esternal events were addressed, in particular the earthquake related failure.
In VASH-1400 is was assumed that the likejihoodof a Safe i
~
Shutdown Earthquake (SSE) was on the order of 10~
to 10 per
- year, the
,estinsted frequency of escoeding the SSE peak ground acceleration, i
In reviewing the test of VASH-1400, it is believed that the authors' were I
addressing the likelihood of design basis esternal events causing pool failure, I
\\ '#'
b'y assigning a probability of 0.1 that a single system (designed to the SSE)
I two systems yoc!d fall. The resulting failure I
would fall and a 0.1 to 0.01 that frequency was therefore between 10~
and 10~
which is interpreted as the i
frequency of pool draining and fuel melt (either by a crack in the pool wall i
and liner or by failure of the cooling / makeup systems).
Combined with the pool 1
- failure, a
failure of the building air filtration system would be required to g,
release isotopes other than the noble gasgs.
Thecombgnedfrequencyofrelease I
was therefore taken to be between 4 s 10~
and 2 10~
lt was concluded that esternal events did not contribute significantly to the total probability or I
consequences of pool failure, i
In NUREC-0933 (Reference 2), the prioritisation evaluation was based in part on I
the VASH-1400 presentation.7The starting point for the assessment of seismic fa!!are was the 10~
to 10~
range from VASH-1400 and taken to mean the fre-quency of breaching the pool and liner. It was further assummed that 90% of the i
time the hole was small and within the available makeup capacity, and that 90%
1 of the time the cooling and makeup system was available. Thegenditional fail-t urs pictability was therefore 0.19.
Usingthe high value (10~
),
the seismic induced f ailure was computed as 1.9 10" It appears that there was an inconsistency in the application of the VASH-1400 I
information.
The VASH-1400 upper end value can be taken to mean that 1 in 10 t
SSE events will result in a failure (or 9 in 10 that the hole is small and t
makeup is available).
Similarly, the 0.1 to 0.01 probability in VASH-1400 for i
two systems to fail at the SSE could be taken as the probability of either a
large hole or the complete loss of the cooling and makeup system.
Cl-811 Task Action Plan Page 15127 Revision 2.0 February 1987
APPEND 1I 1 Applying the technique used in the prioritisation evaluation, the probability of failure is one minus the probability of success (P = 1.0
- P,),
to the g
frequency of the esegeding the SST wguld yield similar results to VASH-1400, that is 0.19 times 10~
or 1.9 10~
as an upper bound.
It appears that the p:.e;; t.s. ; A ct evaluation was based on a double accounting of possible failure scenarios.
The result was to underestimate, by an order of magnitude, the failure frequency.
l The Systematic Evaluation Program (SEP) seismic evaluation report, published in i
April 1984 (after NURCE-0933, November 1983) revised that likelihood of an SSE at sites in the Eastern United States-Ten plants were evaluated (NUREC/CR-3756.
Reference 3).
An estimate of the frequency of excesded 0.2 g ground I
acceleration for each of the sites is provided (approximations from data curves):
I
- Frequency of Exceeded 0.2 g -
Plant Upper Bound Lower Bound Trequency Trequency 1
(per year)
(per year) p.y; Braidwood 4.0 a 10 1.0 10"*
l
~
1 Sheron-Harris 2.0 10-3 2.0 10
~
l
-5 River Bend 4.0 10~'
1.5 10
-4 Millstone 2.5 10-3 4.0 10 e
l Limerick 3.0 10~
1.0 10" La Crosse 2.5 a 10~'
4.0 a 10 '
~
Volf Creek 1.0 10
1.5 a 10~
Vatts Bar 4.0 a 10"I 1.0 a 10~I Vogtle 4.0 10~
1.5 10 '
I
~
Maine Yankee 3.0 10 5.0 10 '
-3
~
It can be seen that the 10~
to10'rangeuseginVASH-1400has
~
been revised upwards by a factor of 10 to 40, to 4.0 a 10~
at the upper bound.
It would that the seisgic failure frequency should be revised to 0.19 therefore appegt
~
10 to bring the prioritization evaluation number tames 4.0 a 10~
or 7.6 10~g).
up-to-date (from 1.9 CI-82/ Task Action Pism Page 26/27 Revision 2.0 February 1987
1 e
j APPENDit g The VASH-1400 analysis could survive the simplistic approach esployed salmly i
because of the assumption that the spent fuel pool radionuclides inventory was saa!!
about 1/2 of a core. However, if the revised SEP seisale data were used,
{
the conclusions might have been diftetent since 10~ge t
enternal events contti-bution would be 10 t o 4 0 times higher, or 1.4 per reactor year.
i The current
- study, beingperforsegbyBNL, has concluded that the seistic failure frequency is about 2 a 10~
based on the revised SEP data
!ct the seismic hasards curves.
While this value is based on a tragility analysis, the data used are not for an actual spent fuel pool structure and are still open to question.
The failure acceleration is assumed to be between 0.75 and 1.0 g, as compared to an 0.2 g SSE value.
The question of the !!kelihood of I
the seismic induced failure of the spent pool remains open and additional work fuel is justified to obtain values of spent fuel pool fragility.
It could be argued, based on the abeve more realistic presentation, that this issue shcold be reassigned as a HICH priority issue.
References:
i i
- 1. " Reactor Safety Study, An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants," U.S. Nuclear Regulatory Commission, NURTC-75/014 1400), October !??5.
(VASH.
3 I
- 2. "A Prioritisation of Generic Safety I
Office of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Commissio NUREC-0933, December 1983, pp. 3.82-1 through
-d.
1 3.
" Seismic Hazard Characterization of the Eastern United States: Methodology I
and Interim Results for Ten Sites," Lawrence Livermore National Laboratory, NUREC/CR-3756, April 1984.
i I
Cl-82fTask Action Plan Fage 27/27 Revision 2.0 February 1987
-_o