IA-87-640, Summarizes Rept on Regulatory Actions Taken in Various Countries to Enhance LWR Safety. Related Documents Encl
| ML20237B770 | |
| Person / Time | |
|---|---|
| Issue date: | 03/05/1987 |
| From: | Messieres Candace De Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| Shared Package | |
| ML20237B537 | List: |
| References | |
| FOIA-87-640 NUDOCS 8712170139 | |
| Download: ML20237B770 (174) | |
Text
{{#Wiki_filter:_ - _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ . _ _
'p 'dtod , o UNITED STATES
! j , "... (f[,'; NUCLEAR REGULATORY COMMISSION
- e- - '/ c ADVISORY COMMITTEE ON REACTOR SAFEGUARDS o, *+ 8 WASHINGTON, D. C. 20555 s.,
[j March 5,1987 MEMORANDUM FOR: ACRS Members FROM: Dr. Monideep K. De, ACRS Fellow MD
SUBJECT:
Summary for Report on " Regulatory Actions Taken in l Various Countries to Ennance LWR Safety" l l 1 l For your convenience, please find attached a summary of the subject report by the ACRS Fellows . 1 cc: ACRS Technical Staff i 8712170139 871210 PDR FOIA / /1 SOR C I B 7-640 PDR \
et b u p on Ti ( y f t c o a n y h e e g H gn t u ra i o y e l r a ms Y h c ea N ib t e h A M am i e g y D t nn R l t n c y ai . E es i n c d a C ry t a n nr S n d e ut F d e n g d 0 ny v u r eh al d e rc C p d e m a I yp e r E d er Y L t u r l( e T B i S e d l o E U c t n a f mn F P E i r l a i - e A l e si s scmrtd S R pw ef ey eeeusn uo g gt gpt oyo R L dP nt nit nssf sc W A an asp a y L R d y h e h re h aS a ry E ec cmn cec c e E D sn vn lll t n C E ae ei eio el al a e N F eg ra rdc ravawg A rr at a at otd r H ce wn wd2 ws ms ee N nm d o dd+ d nenem E IE rC rAN rI RIf e a a a O . H . H . H . . T S E I u R s T r N o r U r e o O o .ta n r f C f s r o u l e S s s r S a n it C s m eo U ree s c C ey e t t os g e e E ra t a IO ) t e r j pr s wd R n apl r g u n f p y e A g rme eus u n d i o ws s d e V i d e o s ef npe e pz I N D s e e i m. gt d e c o it n e c or m o r n yl t ol i nne n t r o miy ul tk n t r v p f t oem p paia N d ca ps P a en p rv n E r il uy d y r cewi u W eat K E a rok s t e c e i st a s F nr A Cd t oc d u r d n v d uet Ai ge T N n cca e b e r eo pk t e ee pg o h nd y e b n o t nt m nera g S A a l r w oc l or co t nc a n ed or Rt seooo e N O FS e t t p l a el si ef lem sti d eo eit nfi d lpme sv eif f o t I W gnc vB g ve gwaaat n p grok S T M nea eC e nt e nsl sca ti nd n C - avenD D n an D e a un e n u a nae A hi ri o h e t h ccornn eq h eot t 0 crdgre bd cm sa d e a 0 si ciriono cni F 3 et n em ti tf oi n it es O 1 e n uv r a ei ra ri eaca ct mey eb cgn ( rmit o ut ut rmersrcept rreae X aal r dS at dl aorapeet ee a unrd I weosp e wn eU wt pmt jsdf wt n on R dt oa m c3 d o c - d uoeunnyna d ot o T rS cGI oH rC oU rats pI isI s r2C sC A M a H . . . r P . a H . r P . Ha. . . a H . s s o L o t la yt v t u o i o m lk e
- c R b a s al t t rB n a e e e nn d H E l o i V ui c y I Vt c r T a A a C tt d E nS -
e n J a/ r S o B l r e C c O P e v C e w e E S eo s p S e e a e g v v eC o r o r rA a p c n p ef a m m Do M I I A B C D. jt'
.
- L
- i . . , . , .' } . ,' i '. ' - . * , _
l l!
R t e i ae E l c s i z i i s ct H f v e ei T e r pw O t d p SS n p /( e f u d emy
)
m e s n)i )t el idl dd at a ane nncst t ar oai yI nf cl dS ore ere . cer s edl m zu z au dt s dtd vi eis eieog d we d wd ml dS r dSd ee A(P A(ARB g Y n T it E n-F A es S ve c di R ev m W re e L ed t t s E l f y C sie s N N efi A E g l R H D nt e H N E anr D E W h e O S cme nr d e T eiu i S ras f at s i _ E wne d I R d or o rCP M T a N H . . . U O C S U s O r I R o t A a V r _ pe _ N mn I ue pG s _ N ps _ E gl mp _ K ne pum s mpu _ A M is p T O) ge u p m DB ri pS u S aD I R p N G h gPH O Nl nHR S I l c% I T K e 0 igyy W F v n0 C D e e1 rtt A A E r v aii Tt id h cc d F I S re sca a pp e O N( dt e t
- a gnaa a X U mg necc g I
ae av e R er hi%% r T t g cr00 g A se d00 e M s e 1 1 s d rm er aarr r d u weuu u d o dt oo o AF rsFF F a
. . H . . . .
s s o L o l t a v yt o t u m i o e l k R c ib a t s t a al n rB e e e d H nn i l o y E ui c r Vt c a V A I a d T tt e n C nS r S o E a/ e C c J l r v C e B P e e E S D w S eo s p e e e v v a g o o eC a r r rA p p c n m m ef a I I Do M B C D. A
. . . . t i i , . 2 -} ,
SUMMARY
OF ACTIONS TAKEN IN VARIOUS COUNTRIES TO ENHANCE LWR SAFETY I FRANCE
- 1. Added the following hardware to enable plants to cope with a loss of all AC power for up to 3 days.
Steam-Driven electrical generator for reactor coolant pump seal cooling Gas turbine to back up diesel generators l Improved DC Electrical Power System l i A special emergency procedure, H3, was developed to respond to a blackout , event and utilize the extra hardware.
- 2. Decided to install capability for controlled, filtered containment venting during severe accidents to prevent gross containment failure and to decrease releases to the environment.
- 3. Improved the reliability of the ECCS through the following:
Automatic switchover frctn injection to recirculation phase Separation and dedication of ECCS pumps for safety use only Interconnection between the Low Pressure Injection System and the , Containment Spray Syster. '
- 4. Improved the reliability for secondary heat removal through the following.
Added two turbine-driven AFW pumps Connected demineralized water storage tank for gravity feed to the Condensate Storage Tank.
- 5. Located RHR system inside containment to decrease risk from interfacing LOCA or Event V sequences.
- 6. Provide independent plant support systems (e.g. cooling) for safety equipment (e.g. ECCS pumps).
- 7. Developed special Ultimate Emergency Operating Procedures for respcmse to core melt and severe accidents.
I j
11 FEDERAL REPUBLIC 0F GERMANY
- 1. Added diversity and redundancy in the ECCS system through the use of the N+2 concept.
- 2. Increased the reliability, monitorability, and duplicity in the Emergency Power Supply Systems.
- 3. Brursbttel: Installation of a special Emergency Decay Heat Removal System which is immune to earthquakes, external explosions, sabotage, !
and aircraft crashes.
- 4. BBR at Krlich: Installation of a fourfold redundant emergency feedwater system, each train of which has an energency condenser.
- 5. Decided to install capability for controlled, filtered containment venting during severe accidents to prevent gross containment failure and to decrease releases to the environment.
III UNITED KINGDOM (Improvements made to the SNUPPS Design for Sizewell-B)
- 1. Added Stean-Driven Charging Pumps for increased emergency cooling capability.
- 2. Upgraded the isolation between the Reactor Coolant and RHR systens to reduce the probability of a LOCA outside containment.
- 3. Improved ECCS: Four 100% capacity HPIS pumps with larger capcity and lower shutoff head, larger (50%) SI accumulators, and four 100% capacity RHR pumps.
- 4. Added an Emergency Boration System for rapid boration of the RCS in the event of ATWS.
- 5. Added a Backup Reactor Protection System for increased reliability of RPS.
- 6. Added four segregated AFWS Pumps for increased cooling capability and increased equipment redundancy.
- 7. CCW, ESW, Dry Cooling Towers design changed for increased redundancy and capability.
- 8. Added four segregated 100% diesel generators for increased reliability. ;
I
IV SWEDEk
- 1. All containments wil' te protected from overpressure by means of devices for controlled pressure relief and filtered venting. This is to reduce the risk of gross rupture of containments and uncontrolled release during severe accider.ts.
- 2. Improved the Decay Heat Removal capability to reduce the contribution to core melt f requency f rts DHR failure from 50 % to 10 %.
V SWITZERLAND
- 1. Added a second suppression pool to the Muhleburg plant to limit overpressure in the reactor building and to provide a long-tem heat sink in the event of total loss of all normal RHR capability.
- 2. Installed an independent bunkered Special Emergency Heat Renoval System for external floods, icss of control room, major fires, and to provide automatic cold shutdowr, particularly during third party intervention.
- 3. Installed a filtered venting system (C0SA) in the Leibstadt plant.
VI FINLAND
- 1. Installed high-capacity safety valves in BWR containments in case of loss of suppression p:ci and for rapid venting following a pipe break.
- 2. Connected Fire Protectien System to Containment Spray Systen for added redundancy and diversity for containment cooling.
- 3. Fuel assemblies replacec with dummies to reduce embrittlement rate of the pressure vessel.
VII NETHERLAND
- 1. Borse11e: Installed a totally independent alternate decay heat removal system.
VIII BELGIUM
- 1. Installed a Second Level Decay Heat Removal System (SLDHRS) to withstand flooding, aircraf t crashes and external explosions.
IX ITALY
- 1. Installation of a total'y independent Special Emergency and Heat Removal (SEHR) system to protect against sabotage.
_ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ b
X # PAN
- 1. Erhanced and detailed safety inspections are conducted for the reliable and safe operation of a nuclear plant and to prevent the occurrence of ;
problems. '
- 2. Installed seismic scram system for safe shutdown during seismic events.
?
l
e, . ATTACHMENT 1 to Enclosure 2 ; i i
)
1 i s s
, PROPOSED BWR SEVERE ACCIDENT CONTAINMENT REQUIRDENTS l
l l l
; R. M. BERNER0 l=
DECEMBER 22, 1986 s I 1 i e
ATTACHMENT 1 to Enclosure 2 3 6 PROPOSED BWR SEVERE ACCIDENI CONTAINMENT REQUIRDENTS R. M. BERNERO DECEMBER 22,1986 s l I I O esp 8 4
BACKGROUND e 0 MARCH 28, 1979: TMI 2 ACCIDENT 0 OCTOBER 2, 1980 FEDERAL REGISTER INTERIM DEGRADED CORE RULE ADVANCE NOTICE OF PROPOSED RULEMAKING LONG TERM DEGRADED CORE O DECEMBER 1980: IDCOR FOUNDED 0 NRC SEVERE ACCIDENT RESEARCH PROGRAM 0 JANUARY 6, 1982: SECY 82-1 PROPOSED SEVERE ACCIDENT POLICY 0 APRIL 13, 1983: PROPOSED POLICY FOR COMMENT 0 AUGUST 8, 1985: NRC SEVERE ACCIDENT POLICY ~
i
,9-f w {
i SEVERE ACCIDENT POLICY STATEfENT I e OPERATING NUCLEAR POWER PLANTS REQUIRE NO FURTHER ACTION TO DEAL WITH SEVERE ACCIDENT ISSUES UNLESS NEW SAFETY lilFORMATION ARISES TO QUESTION WHETHER ADEQUATE ASSURANCE OF NO UNDUE RISK TO PUBLIC HEALT SAFETY. e IN THE LATTER EVENT, A CAREFUL ASSESSMENT SHALL BE MADE OF TH SEVEhE ACCIDENT VULNERABILITY POSED BY THE ISSUE A THIS VULNERABILITY IS PLANT OR SITE SPECIFIC OR OF GEN IMPORTANCE. e THE NDST COST-EFFECTIVE OPTIONS FOR REDUCING T SHALL tie IDENTIFIED AND A DECISION SHALL BE REACH WITH THE COST-EFFECTIVENESS CRITERIA 0F THE COMn! LACKFIT POLICY AS TO WHICH OPTION OR SET OF OPT - ARE JUSTIFIABLE AND REQUIRED TO BE IMPLEMENTED. e IN THOSE INSTANCES WHERE THE TECHNICAL ISSU REGULATORY REQUIREMENTS, GENERIC RULEMAKING WILL BE THE I IN OTHER CASES, THE ISSUE SHOULD BE DISPOSED OF SOLUTION. THROUGli THE CONVENTIONAL PRACTICE OF ISSUING [ - ORDERS OR GENERIC LETTERS WHERE MODIFICATION THROUGH BACKFIT POLICY, OR THROUGH PLANT-SPECIFIC DECISIO HAKING ALONG THE LINES OF THE INTEGRATED SAFE PROGRAM (ISAP) CONCEPTION. I a
IDCOR/NRC PROCESS t 1 e TKO PARALLEL PROGRAMS TO STUDY SEVERE ACCIDENTS IN REFEREllCE PLANTS ] NRC SEVERE ACCIDENT PROGRAM IDCOR a COMPARE AND RESOLVE TECHNICAL ISSUES e IDCOR PREPARE AND SUBf;IT IPE METHODOLOGY FOR HRC REVIEW , e NRC GENERIC LETTER TO DO IPE WITH GUILELINES & CRITERIA BY APPROVED l'.ETHODOLOGY e CONDUCT IPE
- e IDENTIFY AND EVALUATE OUTLEIRS i
e ORDER FIXES , i i
5 X U. S. BOILING WATER REACTORS s e l e 24 BWR 2/3/4 WITH MARK I CONTAlf#EhT (ALL LICENSED) e 5 BWR 4/5 WITH l%RK II CONTAliFENT (8 LICENSED) e 4 EWR 6 WITH l' ARK !!! LONTAINIENT (4 LICENSED) 8 W 4 k f
n,---,---,--- , - - - - - - , - - - - - - - -
\r. e 3 I
3 33 j 131133
'll!!!
3 J i 11:
,!c r 3 i n, 2-g _ , , i: ::
E!!E 101 Il 1 l- d d 55 XXXIXX X
# 3 I5g 5 2 gj yy 5 ill 11 1s1 11 11 1131111 E
ilms 1 e >-- - - - - > >- >> ----- >- ll l 1,1111.11] 1 1 I 11111 11,ill 1.1.- s' I.h- sst ts s s ss st ss.sst
- t i E!a 232 23 2 2 33 23 IIIIII 3
~
No e s 1
$ !! EEE EE E E na EE E!!!ii E I
-_ -_-_== =
? ! iii ii i i f f iiiiiiii i
; 18 tii it i i it it sit tii i
=
l3 EEE !! ! l !! !E !!!!!i g - j s's Esi $ b l b r 6 a
! Je l ltir gia.,
- g. r gAg ,g ],.1,g=
s
= ] le j [I]I =lE=tl-
- J. ,
I E b ! I IEE .
~E
- n ##
==
S y E iabab h h l - sl1 15$5"ii aa .i:31 a3 2 v
- z vala =
ua= . aa ss G
. ..r . . . . . . . . :. .n ar. =. - u. .: - u . e-in=aa --- .. : : -' ' - - -
14 t n n, wigio, tuseaser esauiamond
/
./~ ~
% m .
^
b .
% .f N*s shw won . . I QT* ' " {
~. n s' ,,m -
i V ., Mne' '~~~,,, , vent pie
, yo n ~
L*=== s
- N
,,Was heesw -
s . A
,t u _ n :. )
l
' N f. , Tow someten w
s. i sheaew h.s l V .
-- p,
~
8.Te someten _. 1. ***l t
\
I- l 6 1 4 t g Figure 4. BWR Mark I containment configuration
.. 4 l
1 l l NUREG-1037 2-24 I 1
I bE l EWR C0riTAli4 MENT ISSUES - lMRK 1 l l e e SMALL VOLUE MORE RAPID OVERPRESSURE ESPECIALLY VULNERABLE TO HYDROGEN BURii e Sl%LL DRYWELL FLOOR LOWER HEAD AREA CLOSE TO DRYWELL WALL POTENTIAL FOR DIRECT DEBRIS ATTACK DIRECT RADIATION AliD CONVECTION HEATING , e LIMITED PASSIVE CAPABILITY BUT OPTIONS FOR ACTIVE RESPONSE 1 e 5-ELEhENT APPROACH { HYDROGEN CONTROL SPRAY IN DRYWELL PRESSURE RELIEF I DEBRIS CONTROL- I PROCEDURES AND TRAINING i
Qp KEY RESULlS FOR BWR CONTAINi',EliTS e REACTOR SAFETY STUDY - PEACH LOTT0f' 90% EARLY RELEASE o IDCOR - PEACH BOTTOM 20% EARLY RELEASE o VERM0i4T YANKEE - 7% EARLY RELEASE e NUREG-1150 e l 1 t i ii
- E R L_ . A Z g y F N q t mi m ET m VN N
,[ = = !5 : i=. ! . .i e El - I'= .-
A_ : . ,4 = ! - :i=.! . .isIl-Ed W ~ A3 NV O CO
. * - .!ili '
Y_ O RB A-EF C .7 .=: 1 .e: _= _=
=_Ci
=~E
- l:.
- l:. % _E :9!
=
A _ E : 8! k - ._- E
- 4 YP
~
mt0 D U ufW 4 r_ . m T. wRM 3 A B . O R 3 - i 1 0 0 0 0 0 O 0 0 0 O 9 8 7 6 5 A 3 2 1
. . 0 0 0 O 0 O 0 0 O
E R U + t
+ +
L ,
. + +
A + + - F ++
. 1
+
T
+m N A E-F I
RIF U VN A NA I EIT - k h Y A A PLLTL D h T OE ER TE N M Ntv YMl LIL IH b Y O RL k C~OA A D EON S H TI b N M Y-0
~
AEE ND E E UB W l A &- R KET bG M L 3 ARE TUR TN R HIN LC hT MI A GAO SN A UFC hE EV a EH C O RTF S HNO hL E TE lL A A T TMG hE O LNN N EII TT W A n MAL h E L TE RND N UW A-E A A A YP RU N EO0 O NCN I ER O A A T I T L A I E t R I D M L iF L D A 1 t I T . - B HT M A T EL - A A B I WM O R +
- P - - I
- ~
- 0 4 8 2
- 3 2 1 1 i
. / t-BWR CONTAINMENT VENTING 0 TO PREVENT OVERPRESSURE FAILURE POSSIBLE RELEASE OF NOSLE GAS ACTIVITY 0 FIRST " APPROVED" IN GENERIC LETTER 83-05 (FEBRUARY 8, .1983)
EPG REV 2 SAFETY EVALUATION SUPPLEMENTAL DEVELOPMENT NEEDED 0 SECOND " APPROVAL" IN LETTER TO BWROG (NOVEMBER 23, 1983) EPG REV. 3 SAFETY EVALUATION STILL AWAITING FUTURE SUBMITTALS ON CRITERIA FOR CONTAINMENT VENTING PRESSURE , 1 REQUEST TO CONS fDER FURTHER TECHNICAL STRATEGIES FOR A DEGRADED COI CONDITION. SEVERE ACCIDENT MANAGEMENT)
, G ALL LARGE BWdS HAVE VENTING PROCEDURES VARIOUS PRESSURES (0.5-2 x DESIGN)
EXISTING DUCTS e
.I
{
~
. l
/3 &
VENT DIAMETER, In.. o a v e , 7 " N * ' g .6 f I I I "I 4 E 5'
- E
. 8
=
1 m l g M
.e _ I e .
a
~ E b
% eq
- - 88
.a 8 8E -
. . zg E.
I 9 m ,, . . .I .5 d 3 g .. ta i n 5 o 8 _tEY g 8 ' Ila i
-B _ 8w 3 11 I
.gC- .
ce > a'R f H$ >3 2 ", I
- "50 .
=
h ga "
'l d8 geo 8 I - c 2 =.
5H R .
$I os- ;5 e i . l . l .
o e o 3 g. . e w w a "" s 6 6 6. d W ' 8 3 .1. 3 W Y l 0 .1, N 3 A .
~
t i, i
- eLp V*
i l?.
** ugu n n\ i*t-
-== su dsuma.m =--
= =% -- -asu nm--. _- ____ - = .r.- _ -
I 1
, - ___--___- --__ m =_- .
_--_ -___-z ==_=___=__-__ m____ _ FIGURE _3__, _ 8== -iEi s= m~==-=_-_= _- = -
= - - -
' *g;gijiiiiiii :M iiiie EXPECTED ifHOLE 400Y RADIATION 00$E (REM) iiG i i iE'"
- f'i< %: k;Oi1 !...
FROM RELEASE OF 1005 N08LE SA$ES 4 g g; .i t i t i e i,i: . . - i gs5. men:tssaaest _
- jf_ RELEASE) (1 HOUR FROM 3412 DECAYED Ntt LWR V$.0!$TANCE AND 5 HOUR $ DURATION Ol
_ __ = - a m_ _ _ _ _ __ - _ - _ - - . -_- = - -_ _ _ ___
--2 2 - -
- b$ -
.,b
- ,_-3_
f ;;- ,.;-
- - -2
_ _ _ _ _ _ _ _ BmD i n...- :: . ... - -
-. _-. . ^ . ' -
- . : : - :n 5,,.!.!i:.j':i
,,,,,.......m.. ........ gT(( ~
;^ ; ii, ii .
u;;;;;;;;;;=
; ; l l H;;ll ; ; .1. Graphs assume one hour boldup and decay prior ;; l l l l l l l;!!
- M06 *to release. Greater delay in ulease can produce (
Iower doses (e.g., as such as a factor of about
> t ' 30 at one mile for 12 hours of inreactor holdup l c6apared to one hour).
*=;_ __ _= - {
j
~' -- -
=
~ _
{ __ il Dose estimates are based upon NCCS computer
, code calculations estag revised (relative to CRAC and CRAC2) meteorological sampling models. . i =
. "** t - u __ _ -
W*m - - - - = i = E _ ! 3. The likelihood of exceeding the M i : 3' : .
- : .i b estimated $5 percentile dose is less {
a; - r-- : w::: i than 65 given release of .3005 of I 3 ; , l.i .' i::::l: ^'
, i noble gases as',specified above.
g yy m 33'
,I .iE 1q .I II 1 IIlI I I 1 .E
>E 1 E I , .
. 1 , . . .
1 , I I I I I I I I I I
,,I l I' s I I , 1 I
.t "1 1' I I I I I I b too - / _1 #m%_, .- m. w; e c . demanummuunqquussuusuungsgu
[ l a f hM!! _- == bM d r b I O N h I3 [RO N b b b 10 METER#$ --
.==-__:: --
g g =- := _ : : :: : -- c- : _- - ::n x __ --
- - _ - _ _ .: = = : --
- - : _ : -- = - _ _ __ _
n:
=
3 1 g I l n 4 tt . E b:;.F4 ;): .c; ELEVAT i as9 % ;,es % #, t . Gn gHEiii:l!!! Liilli! H ii! Ei-iMaili H M;, We i~...: ,'e M E _- . Xiii
*. =:_ :- _ __ : =_= ce--= =5_
l'i : -::-__-_ __: =: : : ::-- :: : :__
=
- - ~'-- : : _ := : : _
r
- 5 : .E E=-: E- 5:.5 II *ri:e -
=. : 5 - s_,1 5 : : E =. m - : : : : - i.E-:
=- - -- : : : : x : _ _ _ : : . ..: _ :: : -
.g e -
= -
m
'_' ::: .:: : : : : :: ::::_n-u-:=:-
; . E: :!:: ::= :: L :: : : : : : =- : : . : :
E, ' : : : : ::: ::T ':! l v : U ;;; ;; ; ;; 1 110 w g , , , g EI IE EmIE I IE 1 EM l II Jl IEEmIEI I ml IE
;;m E
,I 1 EE II IEI I E I IE IEl M 1 I El IE A IN IE E 151 1E I bel IE I M IE; 2 E 1E I A E :T, E I E E EIF IE E EEI IE I II] I E I IEl IE I LJ IE II I IE I E 7 ERIII
- :::::':L ;:::: - 1 g 5 : ir pt 1 53:ces= s*
. m na m stes u
,. > F J p,- t-!!!y - - - _ ====n== -
=-
~----espr GROUND LEVEL RELEASE 10 METERS
- - ==== ..=-
=. = - _ - - _ ___ _
- = = = = --
= ===_ __
= _ =
= = =
=
mk Osi!;TE El llBitBI N
@ g W' j i
^ d ty e t . n::;; ;_ _ ;_ .;_ _
y v _ _ _ = - - ._ g N_EE= _ E i E E i E = = -5 E - ' i;- g = - -_ z . E =- -E - =
. a5M _ _:f ELEVATED : RELEA%} 100 METERS -
!. j i. .- ' ._ .- El -H __ !!5 my_.
, ,__ ___ .__ __. __. .__ 4 4_. _ - __
= .:
gs _
;;sgg!!!
y14 ;;; g p __.s _ _. - 4:: $: ::: :: :: = ::: q:y n g, 0 h - - - N+f4 ~-
+- -
. ;ll : : '- '
DISTANCE FROM RELEASE P0 INT (KILES) 5 ' ' ' 'I I > "I~ L L 'pd n --iii i i i-ii- i i i ii i i i iii i i i ii i i i ii i i i ElIIE l ! Il"
/5'dl CONTAINMENT IMPROVEMENT STRATEGY I l
e i e PREVENT HYDROGEN C0t3UST10N BY INERTING . I 1 e REDUCE DRYWELL SPRAY FLOW RATE ; PERMITS ALTERNATE SUPPLIES TO PRODUCE SPRAY l EXTENDS WATER SUPPLIES e PROVIDE RELIABLE BACKUP SUPPLIES FOR DRYWELL SPRAY PROVIDES SHALLOW POOL OF WATER Oli DRYKELL FLOOR DIRECT SPRAY COOLING OF ANY CORE DEBRIS LEAVING LOWER HEAD AREA SPRAY SCRUBBING OF DRYWELL VOLUVE DIRECT COOLING OF WALLS e WETWELL PRESSURE RELIEF TO STACK l - POOL SCRUBb!NG i - ELEVATED RELEASE e DEBRIS CONFINEMEiff. : , o TRAINED OPERATORS .
}
i
/L < j I
PROF 0SEP REQUIREMENTS I
- 1. HYDROGEN CONTROL PRESENT REQUIREMENTS li'f0 SED BY 10 CFR PART 50.44 AND THE TECHNICAL SPECIFICATIONS SilALL BE ADHERED TO, H0 ADDITIO;ML REQUIREMENTS ARE PROPOSIID. i
- 2. CONTAINMENT SPRAY ALL BWRs WITH MARK 1 CCNTAlHFaiT SHALL PROVIDE AT LEAST TWO BACKUP WATER SUPPLY SYSTEr.S FOR THE C0ilTA!HMENT DRYWELL SPRAY, ONE OF WHICH SHALL BE FUNCTIONAL DURING STATION BLACKOUT.
WATER TO THE SPRAY SYSTEfi FROM THESE BACKUP SUPPLIES SHALL i BE AVAILABLE BY REMOTE 1%WUAL OPERATION OR BY SIMPLE PROCEDURES FOR CONNECTION AND STARTUP WHICH Call BE IMPLEMENTED s DURING A SEVERE ACCIDENT SCENARIO. IN ADDITION, THE SPRAY H0ZZLES SHALL BE ADJUSTED S0 THAT i AN EVENLY DISTRIBUTED SPRAY PATTERN WILL BE DEVELOPED IN THE DRYWELL WHETHER WATER itS SUPPLIEB BY THE PRifARY SOURCE OR EITHER OF THE SACKUP SOURCES. A FLOW RATE ON THE ORDER 0F 1/10 0F THE PRESENT FLOW RATE IS CONSIDERED TYPICAL, l THE LICENSEE SHALL SELECT THE FLOW BASED ON AN ANALYSIS OF l PLANT SPECIFIC PARAMETERS. i l
l' (7 X 1 PROPOSED REQUIREMENTS (CONT'D.) l l e
- 3. PRESSURE RFIIEF THE LICENSEE SHALL SELECT A PRESSURE BETWEEN DESIGN PRESSURE AMD It tit'IS DESIGli PRESSURE AT WHICH TO OPEN AN EXHAUST PATH FR0f4 THE WETWELL VAPOR SPAE TO THE HIGHEST
- VENT POINT (STACK OR PIPE) AVAILABLE. THIS LlHE SHOULD BE CAPABLE OF HANDLING WATER VAPOR FLOW EQUIVALENT TO 1%
DECAY HEAT AT THE VENT PRESSURE SELECTED WITHOUT SIGNIFICANT CHANCE OF RUPTURE BEFORE THE DESIRED RELEASE POINT. THE LINE SHALL BE EQUIPPED WITH ISOLATION' VALVES WHICH CAN BE OPENED AND RECLOSED BY REMOTE FANUAL OPERATION OR BY S!HPLE PROCEDURES WHICH CAN SE IMPLEMENTED DURING i SEVERE ACCIDENT SCENARIOS INCLUDING STATION BLACK 0UT.
- 4. CORE DEBRIS CONTROL THE LICENSEE SHALL EHSURE THAT THE WATER'IN THE SUPPRESSIO POOL IN TE EVENT OF TORUS FAILURE IS HELD WITHIN THE
- : CONFINES OF THE TORUS ROOM AND THE CORNER ROOMS AND CANNOT FLOW OUT T0. OTHER PARTS OF THE PLANT.
- 5. PROCEDURES 4ND TRAINING THE LICENSEE SHALL liiPLEMENT EMERGENCY OPERATING PROCE AND OTHER PROCf.DURES BASED ON ALL SIGillFICANT ELEMENTS APPROPRIATE TO ITS PLAkT OF El1ERGEHCY PROCEDURE GUIDELINES, REVISION 4.
I
//W ,
C0hDIT10NS e QUALITY AND DESIGN STANDARDS SINCE THESE REGUIREfENTS ARE INTENDED TO BE AN OPTIM12ED USE 0F EXIST!HG EGUIPPENT IT IS EXPECTED THAT ADDED EQUIPMENT, OF ITSELF, NEED NOT IEET THE QUALITY OR DESIGN STANDARDS OF SAFETY RELATED EQUIPMENT. NEVERTHELESS, MODIFICATIONS TO OR NEAR EGulPMENT OR SYSTEMS WHICH ARE ALREADY SAFETY RELATED SHALL NOT COMPROMISE THE QUALITY OF SUCH EQUIPMENT OR SYSTERS. IMPLEMENTATION THE EQUIPfENT CHANGES REQUIRED HEREIN SHALL BE INSTALLED DURING THE FIRST REFUELING OUTAGE Wii!CH EEGINS NINE (9) MONTHS AFTER THE EFFECTIVE DATE OF THIS LETTER. THE PROCEDURES AND TRAINING REQUIRED SHALL BE It1PLEMENTED ON A SCHEDULE REVIEWED AND AP GIVEN THE IliPLEMENTATION OF THE GENERIC IMPROVElEHTS BY THE NRC. OF MARK I CONTAINMENTS THERE IS NO HEED FOR AN INDIVIDUAL PLAN l ' THIS DOES NOT REMOVE EVALVATION (IPE) FOR CONTAINMENT PERFORMANCE.
'THE HEED FOR AN IPE WHICH COVERS THE SYSTEM RELIABILITY OR C MELT FREQUENCY PORTION OF THE SEVERE ACCIDENT QUESTION.
I
. 8 (9 l l
l WHY DO THIS 1 1 1 0 " COMPLIANCE" WITH THE SAFETY G0AL? - MOST CONTAINMENTS NOT NEEDED ! i i 0 NEEDED FOR SAFETY? POSSIBLE ARGUMENT RESPONSE TO MARKEY OVESTION 0 JUSTIFIABLE BACKFIT? - YES ( e e . t 4 i r
C0ffilSSION RESPONSE TO A HEARING GUESTION
' JULY 16, 1986 QUESTION IS A 90 PERCENT CHANCE OF FAILURE IN THE EVENT OF A CORE MELTDOWN AN ACCEPTABLE FAILURE RATE 7 ANSWER THE NRC HOLDS THE POSIT 10fl THAT THE LIKEllH00D OF CORE ELT ACCIDENTS IN A14Y PLAliT SHOULD BE VERY LOW AND, IN ADDITION, THAT THERE SHOULD BE SUBSTANTIAL ASSURAt4CE THAT THE C WILL MITIGATE THE CONSEQUENCES OF A CORE MELT SHOULD ONE
' IT IS NOT MERELY IN ORDER TO ENSURE LOW RISK TO THE PUBLIC. A QUESTION OF HAVING LOW RISK BUT OF HAVING AS WELL THE D l IN-DEPTH ASSURANCE OF COGINED PROTECTION BY PREVENTION MITIGATION... {' I
+
l 1 s 0 l i
TABLE 3 COST-BENEFIT ANALYSIS e COST: $0.7-2.2M [ BENEFIT:(1) FCH CCFP CCFP AVERTED AVERTED BEFORE AFTER LOSS /YR LOSS PRES. VALUE BASE
~4 CALCULATION 1x10 /yr 0.5 0.05 $4x105 /yr $3M/$12M 1x10 /yr
-5 0.5 0.05 $4x104 /yr $0.3M/$1.2M LOWER FCM LESS CHANGE
~4 $4x100 /yr IN CONTAIMENT 1x10 /yr 0.5 0.1 $3M/$12M BETTER CONTAIMENT TO START 1x10
~4 0.2 0.05 $2x105 /yr $2W$6M
[ i "0PTIMISTIC"
-5 4.2 0.05 $2x104 /yr $0.2M/$0.6M CALCULATION 1x10
" PESSIMISTIC" CALCULATION 3x10
~4 0.9 0.1 $2x105 /yr $16M/$60M (1) FCM = Frequency of Core Melt CCFP = Conditional Containment Failure Probability
' AVERTED LOSS PRESENT VALUE expressed as A/B where A is the averted loss I per year times 8 (roughly equivalent to discount at 12%/yr rate) and B is the averted loss per year times 30 (no discount). -
f.
S Y METHOD OF ACTION I O RULEMAK(NG? COMMISSION CLOSED SEVERE ACCIDENT RULEMAKING COMMISSION CHOSE DEVELOPMENT OF A SAFETY G0AL F0F GENERAL PERFORMANCE STANDARD TEDIOUS TO START A SERIES OF 50.44 TYPE RULES 0 GENERIC LETTER /0RDER? CLASS SPECIFIC , PUBLISH FOR COMMENT OPTIMUM TREAlMENT OF DETAILS PREFERRED FOR PROMPT EFFECTIVE USE OF RESOURCES i: 16 l-I
! l
T. P,cxe.as BWR OWNERS' GROUP
'2 ~' 2 ~4 SEVERE ACCIDENT CONTAINMENT ISSUES NtMARC APPROVED SEVERE ACCIDENT CONTAINMENT ISSUES APPR0ACH
- 1. QBJECIIVE EVALUATE CONTAINMENT INTEGRITY. IF APPROPRIATE, ASSESS POTENTIAL IMPROVEMENTS TO MINIMlZE OFFSITE RELEASES F0P.
SEVERE ACCIDENT CONDITIONS (BEYOND DBA) WITHIN AN APPROPRIATE COST / BENEFIT 60AL.
- 2. IDENTIFY CHALLENGES TO CONTAllMENT
- 1.) H2 GENERATION 2.) OVERPRESSURE 3.) TEMPERATURE 4.) CORE DEBRIS ATTACK 5.) FISSION PRODUCT CONTROL 6.) HUHAN ACTIONS 7.) DIRECT CONTAfHNENT HEATIII6 r 3. IDENTIFY INITIATORS TO EACH CRALLENGE FROM EXISTING ANALYSES PICK KEY EVENTS / INITIATORS (MOST SEVERE - LESS SEVERE)
SEQUENCE THAT PRODUCES MOST SEVERE CHALLENGE
- 4. ASSESS PLANTS' ABILITIES TO MEET CHALLENGES
- 5. ASSESS PLANT VULNERABILITIES
- 6. PROPOSE ALTERNATIVE 5 TO ADDRESS VULNERABILITIES i 7. EVALUATE ALTERNATIVES
- 8. REACH DECISIONS
$
- OTHER ISSUES TO BE ADDED OG37/12.12
l' 1l\\l' 1ll b - 3 h* - I I I
, W K
. O -
v .%w w.4:.M 7 R A M y
. '. ',>\
l t 9 - W
, I S
.* N G _
I S E D M E T S Y S _ k, g N O 2 , s .' 1 1 I S S I K (D l oO R A E R P P
,\ M U S -
r
.~
E
,~ R
. L!! U S
S E R P C . I _ R T C r r
. E L
e E e L _ O L _ n O A n r P , R _ w NE - E OW o N . c IT SS . ._ E t L,. L S Y ,'t G nr LES auEf nwr rr nPePr t y .=
.s;r
- .n
.i 1 aea EUs i
- I PDWSv . .
- . e r 1 K 2
-==== .
- :r. . R .
12345 D i . e A M G I F x.h.i
.. 2 r
. , '[ : e
=. i 4:
M.
. i
. i-Ill li
f 03 t
!% E E
C
' e aK "
jy . w .a ?l U g(eil j igx
...;gue - --
.::1 .
9' 1 < ,'
? '
- %g Q y
[\.-- V g
. WA II I 'I I
'{..... b *
~
Q - (,, y -
~
g g- ' e e. x gI
-e g hk
-s gB {
ED #
\
~ .
! s, Ii H=
t gs i. I Ia . 1 5 jwII si , 4 1
-
- i. 3 s i
_ e i = p .n,
\ n l w
' ,/ . n, t
2 C .~ s -
- z. o -
\
9i M w -
\
/ l.L - -
g "
;. ., ou
-- g -
i
}a t .
31 x '
- 1b -
1 5 M. .
Y 1 - PRIMARY CoNTAIl#453ff 2 = DRY 1 ELL 3 - wrmu, 4 - svernessIon ecos, s - v wr sys m 1M.9 F.E'J 1L , i e%
%, cr r T4',' .- :- h 3 e; ; ;-lyi I 6 L( ', ; i )Lla L i 5 3 I i r
. . l . ,
. I L ; .
,4 - ,_ <. . ,
hs
- s. , , ,
,s . -
["p' 2 "' s ,,- w[ - o - u ._ f" 3 - p' .
- l - -
l?'
,,0 7 l
\
t f ' .
\\
- , , c
=
k I ab
% ' 2 E ha mey ,
["' ___ _ mma .a#w1mm j SAULINIE 69 BAULINIE 72 FIG. 2.3 KRAFTWERK UNION PRESSURE SUPPRESSIVE SYSTDt8 .,, l
\
.. i
M s . p I p . .:. ,....;. u .;. n
- j { .
l
- h I
r' ~ ,l ,4 9' e _ - . - . . . . $ , c .3 . , . e . . g . . . 7,
. . .; .u d ; I r \' INN um..=l ll'* ;:
~
?
=] y
- n. ,*e
,. . ._ br_Z ..1 f [>;
9 n a.:. ,....,,, .
'd(
ll 1 g
);
p ( . yw , .. . . .y..m Jr.- ' - 6 ' .
.- r
- 4. p - =: ; s r
Q
^
(
/ s M'#i d _ .
~g q; i
j . . - .,.,s... m . m
.l
. _ . . . s ,
k I jf j' j h E
.L_..z.J j ll 4 II JII a i g I
9 1li l t I .,,,,, , [, )
;'; -p-
/ ' S ; l
, s h- ,7 .
p j
- y. - . . - . , ,.
jl
- 1 1 r_ e.y. J ng h a e a
- 3) :<
17 u l l Q . ) %-p I i
- I -- w s.. . d;$r
. o,
~
f
~
.[ !
/
l
** M k
i, ll (i _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ -- Q l
4 Dryvell floor Ares
.\ ,
(
'\ ,
o A
'N wwl Y*** ,s'%.,
f
'd#;
- l .
suy . 9 ,- I 4 'y' r hr Detail su: I SvPAW yW" , Baure 5-14) ,
*s'
)
'. \...-
s ; > s..! PL AN- DRYWul ft. Don I f . ) . Figure 5-13 i
- 167 -
- i. . .
pq Dryvell Shc11 (Detail A-A of Tigure 5-13) i \ . f ) I i \ E-JJ9dd
- I 465 yMb r
i Ob, . 'fbd 4l38$ O" _
- ., n - m.34-f!G35'AY N *.*
..,.,=.A...gg.
*.a. p' ;'s * . g * ,
- I
,? .Ji . -
i i 'p',5 p V, ,-
~M Drywall LIhitR i
A-A l
)
l l Figure 5-14 )
- 168 -
l I l-
9 PROPOSED BWR SEVERE ACCIDENT CONTAINMENT REQUIREMENTS
. --R. M. BERNER0 DECEf'.BER 9, 1986 i
N ( /.
NRC SEVERE ACCIDENT POLICY e AUGUST 5, 1985 e PRESENT REACTORS ARE SAFE EN0 UGH, BUT. . . e SEARCH F0R OUTLIERS e CONSIDER EALANCE OF PREVENTION AND MITIGATION
- SPECIAL CONSIDERATION OF CONTAIMiENT PERFORl',ANCE
I s THE SEARCH FOR OUTLIERS
)
l e SEARCH FOR SIGNIFICANT VULNERABILITY
- FIND OUTLIERS NOT NECESSARILY GU;,iiTIFY INLIERS e INDIVIDUAL PLANT EXAMINATION UNLESS ALREADY DONE IDENTIFY OUTLIERS BACKFIT AS APPROPRIATE 4
e WHERE TECHNICAL ISSUE GOES BEYOND CURRENT REGULATORY REQUIREMENTS GENERIC RULEi.AKING PREFErlRED
- ALSO USE BULLETINS, ORDERS OR GENERIC LETTERS
-4 GDC 16: CRITERION 16 - CONTAliiMENT DESIGN. " -AN ESSENTIALLY LEAK-TIGHT BARRIER AGAINST THE UNC0i4 TROLLED RELEASE OF RADI0 ACTIVITY TO THE ENVIRONMEiiT AND TO ASSURE THAT THE CONTAliii'E.NT DESIGii CONDITIONS IMPORTANT TO SAFETY ARE NOT EXCEEDED FOR AS' LO!iG AS POSTULATED ACCIDEliT C0iiDITIONS REQUIRE." GDC 50:
"--AS REQUIRED BY CRITERION 50 - CONTAINMENT DESIGN BASIS.
SECT 10ii 50.44, ENERGY FROM METAL-WATER AND OTHER CHEMICAL REACT 10iis THAT 11AY RESULT FROM DEGRADATION BUT NOT TOT FAILURE OF El'.ERGEliCY CORE C00Ll!1G FUNCTIONING, (2) THE LIMITED EXPERIENCE AND EXPERIMEiiTAL DATA AVAILABLE FO ACCIDENT PHEN 0liENA AND CONTAINMENT RESPONSES, AND (3) THE CONSERVATISM 0F THE CALCULATIONAL MODEL AND INPUT P
IDCOR/NRC PROCESS e TWO PARALLEL PROGRAMS TO STUDY SEVERE ACCIDENTS IN REFEREilCE PLANTS NRC SEVERE ACCIDEiiT PROGRAM IDCOR e COMPARE AND RESOLVE TECHNICAL ISSUES
. e IDCOR PREPARE AND SUBf;1T IPE METHODOLOGY FOR NRC REVIEW l i
e NRC GENERIC LETTER T0 D0 IPE WITH GUIDELINES a CRITERIA BY APPROVED t'.ETHODOLOGY ( e CONDUCT IPE l l
)
e IDENTIFY AND EVALUATE OUTLEIRS e ORDER FIXES l l
U. S. BOILING WATER REACTORS e 24 BWR 2/3/4 WITH MARK I CONTAINMENT (ALL LI ENSED) e 5 BWR 4/5 WITH FARK II CONTAINMENT (8 LICENSED) o 4 BWR 6 WITH t' ARK III CONTAINf'ENT (4 LICENSED) e I
ss pr- e 3 3 33 3 333{33 3
-le .! ss a ,1 J ,! , ,! ,!
'r ' p', s i i ii i ee iiiiii i lEE~i :
Ins a:s ss s es : : sanssi 2 3 3
.s ,W a ,, ,j,j i a
iv,i ill 11 i!1 11 11 1s1s1111 1 ( I 2-- - -- - _ 2 3- 2>- ---- - >- i in
-n Ig iji l
_u__ _ s _i _, i i i i _e 4o __. _4 _a _ _. _. _ _. _.ij1ji12 _
. E
! eet y r e te ey ee -ee ey eeet e=ee.e=
?
e
'_I B s_ ee.e ee e
- -- - .a.~ 4 c.
U 5 . g- - - EE EE ,,,EEE nl EEE EE E E zzz E - I
. m -__ __ _ _ == ====== =
555 55 5 5 er 55 er k t k k k k 5, vver 5 4
. 2, vre ve 1 1
- 55 III II I I II IIIIIIII I i
m --- -- D - -{ { E 8j~- eIna e r A
- L?jeS',e{=gg 5 s _.
~g=
E. w l11r -- aI
" E.r ~g 'I ~ e A. . M S.
I - ! I = & I r ISit -a t W $_ a $_ r__rs grE_ _ h_ _ = E --
!!!}} iI -8 33 .
13131: sa 65 Il "ists J "J si a i !jj J j- A: I1 3 g : : : a : aa ss aa a l
7 KEY RESULTS FOR BWR C0NTAINFENTS e REACTOR SAFETY STUDY - PEACH 50TT0f. 90% EARLY RELEASE e IDCOR-PEACHBgTTOM20%EARLYRELEASE 1 e- VERMONT YANKEE - 7% EARLY RELEASE e NUREG-1150 - l l l
-- -----_o
i 8 ) O I r,;. - , I
',_,,l~~**
~
A r "I : i f fj: n 9 PA1MAny gygygg y, y sms
+
, . ... .m eimv .,
~&. ,
%,; G ~
m , TYPICAL MARK I CONTAINMENT DESIGN. FIGURE 1
0O0ne o5500Y t n n t oons n44bS0 'e 0N10oo54470ln o os e 1 0 i 0
. l.
I 0 L L) 3 u, n, 0 6 6 2 9, O, 9 t l, h, Y, 5, 0, a n/7 ,-i, .t U2 t O1 , , , ,,,, EVF 999446664 9 i '. Y 5 9 /.o222221 0 / / 7 0662 W 2222201 1 93il oT 1 t l k 1 I 1 1 1 1 1 1 1 1 1 t 1 1
- RU l 1 1 1 1 1 1 1 1 N EI C O
I WA( T 000o056600006ia oOOnOo660 nn i t
- o
' ni 35 0 A 000 o8 63' 0501 6 't t s H n i t s
R I t LE1
)
000,1 i 4220,750'O,9*_i i Ut 1
, , ,,, , ,,, , , , , , , i, 220, , ,
t O EMF 99944711 003 4 6 6 Y. 6 q os Sb/ p 8 1 u4 G1 Wl! 555r 63551 i '- 4 4 ( 4' Hi7/4 bh3 l I YLU 1 1 1 1 1 1 1 1 1 t l 1 1 1 1 1 1 t I i 1 1 t 1 ON ROC l O DV( CC A SE ET ER) I I < N DL G E SI s o RM TSG 000000000oO0 000e na i s t t 00op LN NE' * . . . . . . . . .... . - . . . . . . . . . . OR O / 2666226 2 TI AA N 'l CP 6662262266t 5556656655SMS66 s 66226665S5665 t E L N B GO D I 00055055005hbO0n i 0o550 n A T NC I EWM 0007/ 0 2 2 0 0 / / / O 0. HuhAs / 0si i 22.02 000//763374SSt 665'.//n
* .. H N / / 9 l
I CI I l' i
' c. c 5'. /
tI PALU 0 00hh/ ~
'si 0' l 1 l LRF( 1 1 1 l l
l K t H i A SSSSSS S SH l i U S. S SSSSSS M CCCCCC CC: CCC I C CCCCCC PPPf PP PPPPPPSSPPPI iPSP H . i tLLLLCCLLt l LLCL LLLLLL T / / / / / / PP//////P/ ////// I 1 I 1 I I I 1 I I I I I LLI I I 1 1 I t I W N CCCCCC//CCCf I CC/C CCCCCC I ' PPPPPP LLt Lt LCCLLt P P P.i PPPPP.PI O l' I I l 1 t Ctl t t LLLL I / / / / / / PP////// -1 I / i / i //// A I 1 I 1 I 1 LLI I I I L1 I t i1 1 1
- CCCCC SR CCCCCC//CCCC: C/: ' -
1 1 t PI I 1 PPI CU PPPPPl t
.i:b1 Ct if i l i l l t l i l I I i l l l t l I Ui i I I HII t i
t I l I CCl I I l l I il
-t l it l I Ll / / / / / / P P / / / / /. /.W / l 1
- / / /. /. /. /
C E CCCCCCHll C C C :l t t 4 CL1 t I t t t N I I I I I I / / I I I i i l / t / / I i i t i 1
. ; Ct O.
( CCCCCCCCCCC.C R R R l< l i i l I ( i l l d l e I t l e t i t i l l l
'1 1
l i
< <i i l l t
a t c l I
) .
333661 77026663l o00' f u1 3 1 T 99933022593339l /. 5 3 4 t : Y 2 *.9 9 15 9 "/ CRW 1 I WM 22244355624442'i 4 5 LP( 333222221 322'* 32I 1 1 331 221 44444433444444352"'_44- ' . '. DE ON RRRRRRRRRRRRiRRRRRRRRRRR < RI PL WWWWWWWWWWWWWWWWWWWWW<WWW UBDDDDDDDDil s e 1 8 s Dl L t i t a t et l e 1 a t D 1 i 1 23 n n1 2 . o t s 1 2 e yyy i d 'e l
- n. n. i rrr1 2t l l i
l a . sn E rrr a o e l o et i t s ea M eeekkt23n i ee1 ertt t i iY A FFFccS r r enn - l l Coo' s. nn t t N i i nnA2t I c r a a 1 I t i a t ssswwree a E t.i. s M r e s i C :t T nnnsseddeirhh e. i el t r wwwnnpssnma ccel et ccql d n N t r A ooouuoeeartt t pl nnse al aar t L rrrrrorruea aaoioiyeei t n iu P DBBBDCDDDFl l l l l l l MMNOPPPOuV -
lu 1 BWR CodTAlHMENT ISSUES - PARK 1 e SMALL VOLUME l MORE RAPID OVERPRESSURE 1 ESPECIALLY VULNERABLE TO HYDROGEN BURN { e SHALL DRYWELL FLOOR LOWER HEAD AREA CLOSE TO DRYWELL WALL POTENTIAL FOR DIRECT DEBRIS ATTACK DIRECT RADIATION AND CONVECT 10ii HEATING e LIMITED PASSIVE CAPABILITY BUT OPTIONS FOR ACTIVE RESPO o 5-ELENENT APPROACH
~
HYDROGEN CONTROL l SPRAY IN DRYWELL PRESSURE RELIEF DEBRIS CONTROL PROCEDURES AKD TRAINING l O _ _ ~ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
a i i C_0NTAINMENT IMPROVEMENT STRATEGY e PREVENT HYDROGEN COMBUSTION BY INERTING l e REDUCE DRYWELL SPRAY FLOW RATE PERMITS ALTERNATE SUPPLIES TO PRODUCE SPRAY l EXTENDS WATER SUPPLIES i e PROVIDE RELIABLE BACKUP SUPPLIES FOR DRYWELL SPRAY i PROVIDES SHALLOW POOL OF WATER ON DRYWELL FLOOR DIRECT SPRAY COOLING OF ANY CORE DEBRIS LEAVING LOWER HEAD AREA .- SPRAY SCRUBBING GF DRYWELL VOLUME l DIRECT COOLING 0F WALLS e 'iETWELL PRESSURE RELIEF TO STACK POOL SCRUBblNG ELEVATED RELEASE e DEBRIS CONFINEMEilT e TkAINED OPERATORS
o E r m g ? s*.
-. 0 2 4 4 3 2 o 8 n t
0 i o 1 0 t a 9 g i t i M l 0 t I 0 n 8 e d i c c i ) 0 r A c I 0 eR i 7 wW oB s ' P p f
- fo o 8 0 n I 5 w n o
( I 0 oi9 6 M i t t a c a n u P Y ul a M 0 T I F E v I 5 I 0 C - s"a 5 A at 0 . P t r
=
.j A n e po 0 C m e I
E I 0 er r G R R - 4 N i u O U I q CC e DI S T R S N e l E 0 E z mo I 0 V i rs
~ R P )f t
3 S f t do 8 n cp e e T 2 V u I N3 I 0 0 d o E( p 2 . Mm N 2 ev e r 2 r s I 0 I A0 T1 0 0 0 u g i eR i N I 0 F r O= = 1 i g CLf f
, s i
- _ - . - h
'O T 0- 8 6 4 2 0 (
' 0 0 0 0 E jUm2$o g5>
4 -- m=u m sw.sm-= n - o H{libw- h M- _-=-; -ee:7 il l i : H-e =_ s.. m sse ruinirs+m_ w e - - ___ _ =- -- .
,= FIGURE 2
_E
-:_ _ == === = -- === -
= -== _ _ ---
_= __-_- e
= = id ii ! - E M EXPECTED WOLE B0DY RADIATION DOSE (REMOi';= g iii iw
'l !! 1 I
, y 3 f: # =i m;c:=E md!=ri:' m FROM RELEASE OF 1005 NOBLE GASES u- -
(I HOUR DECAYED AND 5 HOURS DURATION4L~==D: 0F 2 s
=;i i ;=gy3_:3 _ -7
; RELEASE) FROM 3412 Et LWR V5. DISTANCE E i;-!) 'Ll l
=.-:.. .. .
- =E- _
a ". I;;;;llllll 1 - ICTES: i ;;';;ll;;; ' 11 ::::!! il:I )
- 1. Graphs assume one hour holdup and decay prior l l l l l l ' I i i ' ' l l "to release. Greater dela in release can produce lower doses (e.g., as muc as a factor of about i 7~ 30 at one mile for 12 hours of inreactor holdup l g compared to one hour).
=
e ===e i=1 ; _ = _
;.- - 333====__'
y,i si ., i ei 2.' Dose estimates are based upon NCCS computer =._=. g code calculations using revised (relative to 1
'(t y .
CAAC 'and CRAC2) meteorological sampling models.
= b ----
Ei!!!!! 1Ui'
=1
- l E
- 3. The likelihood of exceeding the !~
l6 ! 2- estimated 95 percentile dose is less E ' Ei a i : ii than 55 given release of 1005 of '
'i noble gases as's)ecified above.
- 'i!!ili! iiii:! !!i j
=,
g
- M: Tl;;;;;'lll i I 11 1 II I i I I I I 1 1 i lll I I I 1 1 i i i r 1 i1 1 E 1 1 Mg i ig !EIli IE I I I II i 1 1 I I I I 4 I I I II I II I too i Js' ; s. : . ar 3 m: b umummmmmmuuuuumuuuuggggggg
- 3'
! kb L. Urn W
fi i
. e a! r e - B EST' MAT "5aN'y P a
- v- : M""' 10 nETERS k
--- =- -=1 g
; -- _- gg --
{'
, - ;_- _ -g=
5 3-
!u. I s, g
=
3 M"shetenue b
=- -- ---
!: i itiamma O ESTIMATED 95 PERCENTILE Eu . -
-ELEVATE.D _ = RELEA3R 10_0 MET
+: c- _ _ _ _ -E-
.l.,
;; : ;~ ~
agg :- - q: s
=
= - = ; -:
5- =- o 33j - i E =- ! !
=- w!
ea _ E : - i .E E
. 1:
- i. i
- Jd
-u b I ,- : : .- : :. :: .-
.: Fi
- - :: : R:
m : .i := w: 9
;;; j;:
, t :- :
- :::._ ::, :',2 ':: : ::: : : :
3 . ::- i e w . g
= ; ::=%: ;: =
...m. - ,.1 -.
l l:: i
,imi i 1I, i.i o . .
. > .. i 1 L o
,i 1 im 3 m I IIEm EN 1II Il 'Em 1 I EI 1 1 IEl III I L-- EI I I I III A1 I I I I "
!k
, = - _ t= 2i: ;: -
= -
=u -
-5 l ll GROUND LEVEL RELEASE 10 METERSs:===
*o
*E y "iM i i.
l!! i! -
.. 13 == =_ . !M - i =! ! ! E! illE!!
W a q ' it-
$i -
r 12 %ItS!ti Q! ? I Pl H $ l l l M $ M
'o i
= = - -
1;- 71-
=.
=
ii-
- 1
_ = - s-ii i; = = 3
@ E
_ i; i E; Ei E .
- y
. if ~-
-=
-E --
-- ti
' E: iE Ei E= E i-
'E
'E S ::-*
WE.EYATED *- RELEASE 100 METERS -- .--- .
~
~ '-
E Ei' E E
- r :
E : b ::: M' ::
^ ~ - '
' ' =
g:::: . .--
- ": :: d'
;-: :: :: :::::::::: ::::: l :::: :::
~
- l : : := :: ::: '= Ej m::
we -- : =: :: =, : ::: ::: :: ::: :: ::: ::: :: ::: :: ::
""' . - . - - . -- .-- =_.: - - --- -- --- -- -- --- -- --
I IE E II I E .Il l
- T:E " STANCE FROM RELEASE POIK1' (MILES) -- g; j ji ; ; ;M
" M is ii i ii i ' i iii i i i iii i i i ii i i i ii i i i ni oj in' i in u in n
! i t'4') ' 4 5 ' 6
'I 8 9 W 11 11 15 14
la ' PROF 0 SED REQUIREMENTS
- 1. HYDROGEN CONTROL PRESENT REQUIREMENTS IMPOSED BY 10 CFR PART 50.44 AND THE TECHNICAL SPECIFICATIONS SHALL BE ADHERED TO, NO ADDITIONAL REQUIREMEifTS ARE PROPOSED.
l 2. CONTAINMENT SPRAY ALL BWRs WITH FARK I CONTAlilMENT SHA'LL FROVILE AT LEAST TWO BACKUP WATER SUPPLY SYSTEMS FOR THE C0iiTAINMENT DRYWELL SPRAY, ONE OF '4HICH SHALL BE FUNCTIONAL DURING STATION BLACKOUT. WATER TO THE SPRALSYSTEti FROM THESE BACKUP SUPPLIES SHAL , BE AVAILABLE BY RE!50TE t'ANUAL OPERATION OR BY SIMPLE PROCEDURES FOR CONNECTION AND STARTUP WHICH Call BE IM LURING A SEVERE ACCIDE!iT SCENARIO. IN ADDITION, THE SPRAY 110ZZLES SHALL BE A3 JUSTED SO THAT AN EVENLY DISTRIBUTED SPRAY PATTERN WILL BE DEVELOPE DRYWELL WHETHER WATER IS SUPPLIES BY THE PRifARY SOURCE OR EITHER OF THE BACKUP SOURCES. A FLOW RATE ON THE ORDER OF 1/10 0F THE PRESEllT FLOW RATE IS CONSIDERED TYPICAL, THE LICENSEE SHALL SELECT THE FLOW BASED 'ON AN ANALY PLAWT SPECIFIC PARAMETERS.
so PROPOSED RE UIREMENTS (C0NT'D.)
- 3. PRESSURE RELIEF THE LICENSEE SHALL SELECT A PRESSURE BETWEEN DESIGN PRESSURE AND lb TIMES DESIGN PRESSURE AT kHICH TO OPEN AN EXHAUST PATH FROM THE WETWELL VAPOR SPACE TO THE HIGHEST VENT POINT (STACK OR PIPE) AVAILABLE. THIS LINE SHOULD BE CAPABLE OF H'ANDLING WATER VAPOR FLOW EQUIVALENT TO 1%
DECAY HEAT AT THE VENT PRESSURE SELECTED WITHOUT SIGNIFICANT CHANCE OF RUPTURE BEFORE THE DESIRED RELEASE POINT. THE LINE SHALL BE EGUIPPED WITH ISOLATION VALVES WHICH CAN BE OPENED _ AND RECLOSED BY REMOTE MANUAL OPERAT ~ OR BY SlHPLE PROCEDURES WHICH CAN BE IMPLEMENTED DURING SEVERE ACCIDENT SCENARIOS INCLUDING STATION BLACKOUT.
- 4. CORE DEBRIS CONTROL THE LICENSEE SHALL ERSGRE THAT THE WATER IN THE SUPPRESSIO POOL IN THE EVENT OF TORUS FAILURE IS HELD WITHIN THE CONFINES OF THE TORUS ROOM AND THE CORNER ROOMS AND CANNOT FLOW OUT TO OTHER PARTS OF THE PLANT.
- 5. PROCEDURES AND TRAINING THE LICENSEE SHALL liiPLEMENT EMERGENCY OPERATING PROCE AND OTHER PROCEDURES BASED ON ALL SIGNIFICANT ELEMENTS APPROPRIATE TO ITS PLAhT OF EMERGENCY PROCEDURE GUIDELINES, REVISION 4.
m , 10 \ i CONDITIONS QUALITY AND DESIGN STANDARDS SINCE THESE REQUIREMENTS ARE INTENDED TO BE AN OPTIMIZED U OF EXISTING EQUIPMENT IT IS EXPECTED THAT ADDED EQUIPMENT, 0F ITSELF, NEED NOT MEET THE QUALITY-OR DESIGN STANDARDS OF SAFETY RELATED EQUIPMENT, NEVERTHELESS, MODIFICATIONS TO OR I NEAR EQUIPMENT OR SYSTEMS WHICH ARE ALREADY SAFETY RELATE SHALL NOT COMPROMISE THE GUALITY OF SUCH EQUIPMENT OR SY IMPLEMENTAT10ll l THE EQUIPtENT CHANGES. REQUIRED HEREIN SHALL BE INSTALL THE FIRST REFUELING OUTAGE WHICH BEGINS NINE (9) MONTHS A THE PROCEDURES A'iD TRAINING THE EFFECTIVE DATE OF THIS LETTER. REQUIRED SHALL BE lhPLEMEiiTED ON A SCHEDULE REVIEWED A BY THE NRC. GIVEN THE liiPLEF.ENTATI0il 0F THE GENERIC IMPROVEME OF VM K I CONTAINMENTS THERE IS N0 HEED FOR AN INDIVIDUA THIS DOES NOT REMOVE EVALUATION (IPE) FOR CONTAINMENT PERFORfWICE. THE HEED FOR AN IPE WHICH COVERS THE SYSTEM RELIABILIT MELT FREQUENCY PORTION OF THE SEVERE ACCIDENT QUESTION
ll SEVERE ACCIDENT POLICY STATEtiENT I e OPERATING NUCLEAR POWER PLANTS-REQUIRE NO FURTHER REQULATORY ACTION TO DEAL WITH SEVERE ACCIDENT ISSUES UNLESS SIGNIFICANT NEW SAFETY INFORMATION ARISES TO QUESTION WHETHER THERE IS ADEQUATE ASSURANCE OF NO UNDUE RISK TO PUBLIC HEALTH AND SAFETY. e IN THE LATTER EVENT, A CAREFUL ASSESSMENT SHALL BE PlADE OF THE SEVERE ACCIDENT VULNERABILITY POSED BY THE ISSUE A!!D WHETHER THIS VULNERABILITY IS PLANT OR SITE SPECIFIC OR OF GENERIC IMPORTANCE. l e THE MOST COST-EFFECTIVE OPTIONS FOR REDUCING THIS VULNER SHALL BE IDENTIFIED AND A DECISION SHALL BE REACHED CONSIST WITH THE COST-EFFECTIVENESS CRITERIA 0F THE COMnlSSION'S LACKFIT POLICY AS TO WHICH OPTION OR SET OF OPTIONS (IF ANY ARE JUSTIFIABLE AND REQUIRED TO BE IMPLEliENTED. e IN THOSE INSTANCES WHERE THE TECHNICAL ISSUE GOES BEY REGULATORY REQUIRE?1ENTS, GENERIC RULEMAKING WILL BE THE PREFERRED SOLUTION. IN OTHER CASES, THE ISSUE SHOULD BE DISPOSED OF THROUGH THE CONVENTIONAL PRACTICE OF ISSUING BULLETINS ORDERS OR GENERIC LETTERS WHERE MODIFICATIONS ARE JUST THROUGH BACKFIT POLICY, OR THROUGH PLANT-SPECIFIC DECISION-fMKING ALONG THE LINES OF THE INTEGRATED SAFETY ASSES PROGRAM (ISAP) CONCEPTION. ____._____ .____m_._. _ _ _ - _ _ _ . _ . . _ _
l', .l. , , COMNISSION RESPONSE TO A HEARING OUESTION JULY 16, 1966 _ QUESTION IS A 90 PERCENT CHANCE OF FAILURE IN THE EVENT OF A CORE MELTDOWN AN ACCEPTABLE FAILURE RATE? ANSWER THE NRC HOLDS THE POSITION THAT THE LIKELIHOOD OF CORE MELT ACCIDENTS IN ANY PLANT SHOULD BE VERY LOW AND, IN ADDITION, THAT THERE SHOULD BE SgSTANTIAL ASSURANCE THAT THE CONTAINMENT ~ WILL MITIGATE THE CONSEQUENCES OF A CORE MELT SHOULD ONE O IT IS NOT MERELY IN ORDER TO ENSURE LOW RISK TO THE PUBLIC. A QUESTION OF HAVING LOW RISK BUT OF HAVING AS WELL THE D IN-DEPTH ASSURANCE OF C0l*BINED PROTECTION BY PREVENTION A MITIGATION...
l 1.. , l-l 1 l CURRENT NATIONAL SOURCE TERM POSITIONS AND PRACTICES IN OECD MEMBER COUNTRIES SWEI)EN Question 1: What is the philosophy of source term use and de-velopment in your country? t In Sweden the use of source tems is primarily considered as
, forming part of a complete s,afety analysis of the nuclear power plants in order to understand the consequences in the event of accidents and to evaluate means - in terms of facilities and
([ management - of improving and optimising the safety. Different sets of source tems are used to analyse plant response to various types of accidents and corresponding environmental consequences. Typically it rests with the licensee to propose the various sets of source terms used in the safety analysis report submitted to the authorities for review and approval. Question 2: What is the source of current source term values used. in your country, for source terms related to licen-sing and for emergency planning source terms? Current values of source term to the containment used in licen-sing as a design basis for syste=s and containment performance are based mainly on earlier USNRC regulations. So:ce modifications have been made as a result of increased knowledge of fission product behaviour after TPl. Emergency planning is presently based on a set of source terms derived mainly frcxn WASH-1400, recognizing that the radioactivity
, released from the reactor core, in the event of an accident ,
L. causing damage to the core, would most probably be largely re- { tained in the reactor containment, whereby the release to the l environment would in the majority of cases be minor. I Question 3: According to your country, what are the key source term issues that require further development? f For plants where the possibility of early containment failure modes during a severe accident could not be neglected, the key I issues would be the mode and the timing of possible containment failure.
/
. _ _ _ _ _ _ _ _ _ _ - _ _ _ _ i
6 2 For late containment failure, the estimate of the s.ource term - would mainly depend on a correct understanding of the interaction between fission product behaviour and thermal hydraulic con-ditions. Uncertainties to be resolved relate mainly to factc:-s determining the degree of retention of the fission products witnin primary systems and containments, e. g. with regard to resuspension of aerosols and revaporization phenomena. An important issue at present is to verify the pefor:ance of filters for filtered containment venting, to be i=plemented at all remaining reactors in Sweden corresponding to the FILTRA system already installed at the Barsebaeck plant. - Developments are required particularly in the field of management of accident conditions. ' Question 4: What are the source term research progrs=s (experimental, analytical) in your ccunt.ry. The national severe accident research program in Sweden is partly concerned with fundamentals, studied jointly by the Swedish authorities and the industry, and partly with applied research in consideration of relevant plant specific features. The fundamental research program (FILTRA - RAMA - PJJ.A II - RAMA III) is being conducted in order to provide the necessary ana-lytical tools (computer models etc) for plant specific severe accident analysis as well as in the development c accident management techniques and plant modifications to cope better with severe accidents. The initial stage of this program, FILTRA, completed in 1982, was aimed at studying the feasibility of filtered venting to be applied to the Barsebuck nuclear power plant, located close to Malm5 and Copenhagen in the south of Sweden. As a result, the { BarsebMck filtered venting system is now operational since Novem-ber 1985. The subsequent stages, RAMA, completed in 1985, and RAMA II, completed in 1987, were concerned with the rMning 10 reactors for which provisions aimed at mitigating the consequences of severe accidents, as judged to be needed, are to be implemented by the end of 1988. RAMA provided the first set of analytical tools, based mainly on the IDCOR MAAP code, together with a pre-liminary validation, enabling plant specific studies to be. carried out and reported by the utilities in 1985. FlMA II was aimed at refining the analytical tools and providing further validation, largely based on current international research. The current final phase of the RAMA project, RAMA III, is concerned with comparisons of the MAAP code with corresponding NRC severe accident codes; further studies of critical matters of uncer-mm__________________ - _ _ - -
3 l tainty, particularly these related to melt progression, coolabi-lity and fragmentation of zelt, and certain chemical questions; and evaluation of accident Isanagement procedures. An essential part of the fundamental national research programme has consisted in participation in' a number of international re-search programs such as LOFT-FP-2, SFD, IDCOR, LACE, NKA (the Nordic Nuclear SAafety Research Programme) and MARVIKEN-V-ATT. The fundamental research continues to be focused principally onte validation and plant specific adaptions of the IDCOR MAAP code, adopted as the prime analytical tool for severe accident and accident management analysis of the Swedish nuclear power plants. The applied research, conducted by the utilities, resulted in preliminary recommendations ty the utilities in 1985, in turn forming basis for recommendations by the Swedish Nuclear Power i Inspectorate and the Swedish Radiation Protection Institute to the Government, as to measures to be.taken at each particular nuclear power plant, in addition to the two Barsebuck reactors. The development of protective measures and accident mitigating systems is presently in prcgress, scheduled for coc:plete imple-mentation by end of 1988. As part of this programe a comprehen-sive testing programe is under way for evaluation of the type of filters, FILTRA-MVSS (Multi Venturi Separation System), chosen . l for filtered containment venting. Question 5: How is current scurce term information applied in the licensing area im you country.
' According to policy decisions by the Swedish parliament in 1981 and by the government in 1981 and in February 1986 there should be a strengthening of the prctection against radioactive releases to the environment, should a severe accident involving extensive
~ core damage and/or meltdown cccur in a Swedish reactor. To this end, filtered venting of the Barsebuck plants was required by
{'. 1986, at the latest, and is actually operational since November 1985. Provisions aimed at granting the same level of protection, irrespective of location, are required for the remaining ten Swedish nuclear power plants for implementation no later than by end of 1988. Question 6: How is current scurce term information applied in the design of reactor safety systems (including contain-ment atmosphere control systems) in your country. Question 7: How is current scurce term information applied in reactor accident consequence assessment in your country. esamen
4 Acce,, table environmental consequences have been defined and the corresponding level of protection prescribed in a Governmental decision of February 1986:
- The same basic requirements regarding the vrimi quantity of re' eased radioactive substances shall apply +w all reactors ir espective of site and power
- Lard contamination, which impedes the use c' large areas for a lo:g period ought to be prevented
- Deatcs in acute radiation disease shall rxst occur.
The requirements are considered as fulfilled if the release of fission products, in the event of a severe reactor accident causing damage to the reactor core and possibly a core melt, is
." limited to 0.1% of the inventory of the cesium isotopes 134 and 137 in a reactor core of 1800 MW thermal power, provided that other nuclides of significance in regard of land contactination are retained in the same proportion as cesitz:. Release of noble gases is accepted in case of a severe accident if *2ecessary to protect the containment.The release is assu::ed to be esticated using most recent source term informtion as represented in vali-dated computer codes.
Certain hardly conceivable accident sequences of extremely low probability, although possibly potential of causir4 higher re-leases, are then not to be considered. It is envisaged that the new source term it'emation developed in the research progra=s will also be used in a review of the Swe-dish emergency planning program. Questico 8: Cceplete text describing current ratiocal source term positions and practices in your country.
- In addition to the requirements in regard of prote::tive measures L to be taken,as described above, the Swedish Goverment pre-scribes, in its decision of February 1986, that fellowing condi-tions shall apply for continued operating licecces for the Swe-dish reactors subject to the requirements to be implemented be-fore end of 1988:
- 1. In the first hand core damages shall be ;revented by means of high quality standards as regards daily cpe: ation and main-tenance work. The on-site preparedness for accidents within the nuclear power station shall be organized to take care of al' possibilities to re-establish core eccli.ng, whenever acci-de:taly lost, before extensive core damages ocer.
t 5 4
- 2. .In case of an accidect with core damage there shall be pre- i pared strategies of acting, specifically suited for the plant in question, with the aim to protect the function of the coo-tainment and to reace as soon as possible a stable final state -
with the damaged ccre properly cooled and covered with wa:e .
- 3. It is in partieula :.mportant that the function of the c:t-tainment remains intact during the first ten to fifteen n:u:s after a core accident. The protection against direct -d:r.I_;es to the containment and to the containment parts of partic:J.ar
- importance to the function of. the containment, as a conse-quence of the accidental conditions or by the core melt, shall
(. therefore t a reinforc:ed.
- 4. In order to protec, t.ne containment against overpressure damages at serious accidents and to improve the possibilities (
(;: to reach a stable fical state after the accident, a contrc11ed pressure relief of the containment shall be feasible. The relief devices are tc be designed so as to enable activation - independent of ope at.or actions and independent of the Arc-tioning of other safety systems if the design pressure of the containment would be exceeded to such an extent that unaccept-able leakage or damag'e to the containment might occur. The pressure relief devices shall also be available for activation by operating personne.1 as a means of accident management. They shall furhermore be designed to ensure,. together with other measures for prote: tion of the containment, that the radic-active release to the environment will be properly controlled i as prescribed. ' Furthermore it is stated, that before any protective measures a e actually taken at the nuclear power plants, technical specifi-
/
cations are to be presented to the Swedish Nuclear Power In-
' spectorate for approval. In particular it shall be accounted fc: , l to what extent proposed =easures may have a negative impact cc l the over-all reactor safety.
i
)
l 1
. I m_.___.___._.____..______.
6 -1 cr r et 4 '$ (a s</* *
~
- b4 t G Oo *fdf' / h4db . ~~ k v
The Swedish Program on Prevention and Mitigation of Releases in the Case of Severe Accidents. , Lars Hoegberg Director, Office of Regulation and Research Swedish Nuclear Power Inspectorate (SKI) P.O. Box 27106, 5-10252 Stockholm, Sweden
SUMMARY
j In 1981-82 the Swedish Government and Parliament cecided on the gene-ral guidelines for the Swedish program on prevention and mitigation of radioactive releases in the case of severe accidents. Installations of a filtered venting system (FILTRA) was required as a condition for a jk. continued operating license for the Barsebeck 1 and 2 nuclear power plants af ter September 1st,1986. The FILTRA system at Barsebeck be-came operative on October 31, 1985. For the other ten operating reac-tors in Sweden, measures for the prevention and mitigation of releases in the case of severe accidents should be implemented by 1989.
, Plant-specific severe accident studies have been sunmitted to the inspectorate by the utilities as a basis for decisions on such mea-SurCS.
This paper presents the principal policy decisions taken on the severe accident issue. The current technical status of the Swedish severe cccident program is summarized. I. INTRODUCTION
- The present Swedish program on nuclear reactor safety is in large
(- parts based on the report presented in Decereer 1979 by the Swedish Government Committee mi Nuclear Reactor Safety (1). M On the political level, the general guidelines for the current safety
- program are stated in the 1980/81 energy bill which was passed by the
( Swedish parliament in May,1981, (2). The guidelines were reconfirmed in the 1984/85 energy bill. l (_J i ) l According to these guidelines, the main priority will be to prevent { expecience as well as plant-speci-core damage. Feedback fic probabilistic safetyof analyses operating (PSAs) are therefore important com- { ponents of the Swedish safety program. As a matter of fact, experience !
- has shown that the Swedish utilities, during the process of carrying j through these PSAs, will implement changes in the plants to ensure l l
i N
- - - _ - _ . _ - - I
s 2
+ .
o that the calculated core damage frequency falls below 10-5/ reactor-year. However, it' is clearly stated in the Swedish reactor saf ety program that it must be recognized that accidents involving severe core damage may nevertheless occur, and that measures therefore should be taken to ensure that releases in such accidents are kept low. Furthermore, the mode of operation of the Swedish regulatory system simuld be noted. According to the Swedish legislation, the plant ownero have the full responsibility, not only to maintain but also to develop and improve nuclear safety. The government and regulatory j authorities set general safety objectives, to which the utilities respond by proposing technical solutions to be reviewed and approved by the authorities. Detailed rules and regulations are limited to as ( few areas as possible. This type of technical interplay between the regulatory authorities and the utilities has been used also in the severe accident area. II THE FUNDAMENTAL POLICY DECISIONS IN 1981-82 The fundamental policy guidelines for the Swedish program on severe occidents were given in the 1980/81 energy bill to the Swedish parlia-ment. In that bill, the Swedish Government, through the Minister of 6 ergy , made the following statement (2):
"Despite the fact that existing reactor installations present an extremely small risk for uncontrolled releases of large amounts of radioactive material, which would result in radioactive ground contamination, I am of the opinion that all possibilities should be exploited to further reduce the risks for such releases.
This is particularly the case for plants, such as Barsebeck, which are situated near densely populated regions. Such measures should be taken even if they involve a not insignificant cost for the owners, ss seen in relation to the reduction of the release risk. Tiltered venting of the reactor containments in Barsebeck should be ready to be taken into operation in 1985 at the latest, or at (. the next following revision. It rests with the Swedish Government J to issue further directives in this matter. Filtered venting of the reactor containments in Ringhals, Oskarshamn and forsmark may also come under consideration. It is however important that the j - (_ I experiences from Barsebeck and the technical developments, which I are under way within this field, are taken into account. Direc-C tives considered in other countries using nuclear power, such l as USA, concerning measures to reduce the risks for radioactive fi ground contamination, should be taken into account, if possible, j when specifying the requirements for the later reactors. If conti-nuing research shows that other methods than filtered venting of 1 l l l I -
,- -3 :
. l o
- the reactor containments give comparable reductions Os the risk I for large releases of radioactive material, or if the cresent risk assessment for accidents which result in large releases of radio- '
active material is considerably changed, the safety requirements for the nucleer power plants in Ringhals, Oskarhamn asd forrmark should be adjusted accordingly. The necessary decisiens in this matter should be taken at such a time that the resulting measures are completed by 1989". A couple of points in that statement merit special attention. firstly, it should be noted that special priority is given to the rrevention of ; ground contamination due to the extensive social conseove7:es that may be anticipated in connection with large-scale evacuatic . Secondly it is pointeo out that simple cost-ef festiveness arguments eill not be
~
#P accepted. Thirdly, priority is put on measures at the Strsebeck site
- k. as these two Swedish reactors are the ones situated closes: to large urban areas; large parts of the Danish capital Copenhace, as well as ,
Malmoe (the third largest city in Sweden) being within 20 km from Barsebeck. . f ' (? In a letter, dated October 15, 1981, to Sydsvenska Varm6craf t AB, owner of the Barseoeck nuclear power plants, the Swedis7 Government required installation of a filtered venting system as a condition for a continued operating licence after September 1, 1986. After making reference to the appropriate parts of the Swedish Atomic Energy Act, previous licensing decisions concerning the Barsebeck plants and the statement in the 1980/81 Energy Bill cited above, the Government issued the following directives (3):
"At the power station at Barsebeck, arrangements shall be made for filtered venting of the reactor containments.
The equipment should be so designed that it comes bito operation when the pressure in the reactor containment excee:s the design pressure of 0,5 MPs as given in the current safety specifica-f tions. The pressure at which the equipment is desiyued to come in-f .
' to operation shall not, however, be so great that leakage from the reactor containment can be expected. The equipment small be con-structed in such a way that at least 99,9% of the reactor core in-
~ ventory of each radioactive isotope, excluding nccle gases, will jr be retained in the reactor containment and the filter system in g#
the event of venting through the filter system during a severe reactor accident. These measures shall have been taken, and the equireent shall be operative, by the end of 1985 or at some later date to be deter-mined by the Nuclear Power Inspectorate, however not later than September 1st,1986." W gamump
. I, Furthermore, the Government empowered the Swedish Nuclear Power l
Inspectorate to review and approve the technical solutions submitted by the utilities and to issue such additional guidelines or conditions as it demand necessary for safety reasons. This government decision thus defined the following important safety concepts and objectives in the area of severe accidents:
- The concept of a safety valve function to come into operation only if the containment pressure significantly exceeds the design pressure. In this way any negative impact on safety for events within the design basis will be minimal. :
- REl~ eases of noble gases may be accepted in the case of a severe f accident going beyond design-basis if such a release will limit i
. { releases of other nuclides causing ground contamination. Obvious- I ly, it is an advantage not having to discuss this issue between plant crew, authorities, local and central government in the j middle of a real emergency.
( (
- A goal to limit releases to 0.1% of the radioactive inventory of a
" 1800 MWth core, noble gases excluded, in order that only limited l areas near the plant (a few tens of sq.km) should be affected by i l
land usage restrictions, at the same time providing ample margins to acute radiation illness and deaths in the case of a severe accident. In this context it may also be noted that the present design basis containment leak rate is 0,1%/24h, also assuming the old USNRC source term to the containment atmosphere. On January 15, 1982, in a letter (4) to to ~ owners of the plant, the. ] Swedish Nuclear Power Inspectorate in cooperation with the National Institute of Radiation Protection issued further technical guidelines for a system for filtered venting of the Barsebeck reactor contain-ments. In this letter, the inspectorate writes as follows: l
" Choice of design basis parameters with regard to venting and
(' filter cacacitv. etc. - l
\
As a starting point for the choice of design basis parameters with ) regard to venting and filtering capacity as well as for the gene-ral design of the installation, an analysis should be made of a ( number of typical cases, chosen in such a way that, when conside-('a red together, they can be judged to cover such sequences of events which give the most important contributions to the risk of con-tainment failure due to overpressurization in connection with very severe reactor accidents as derived from a safety analysis of the nuclear reactor plant." h e C_-__._____.__...
5 Furthermore, the systems for filtered venting should be designed in such a way that they come into operation passively in the case of a severe accident and then function passively for about 24 hours. Af ter that period, credit can be taken for such active measures that have been prepared for. Filter removal efficiencies with reoard to various isotooes i for the safety analysis of filter removal efficiencies with regard to radioactive isotopes the following starting points should be used. In the analysis, special weight should be given to those
'isotepes and their chemical compounds which nay cause ground con- -
tamination of importance from a radiological point of view. More-over, the analysis should deal with those isctoces in the core in-ventrey, excluding noble gases, which are cons;0ered important f, from a radiological point of view in current reference literature. - l When describing the removal efficency in the chosen design, the ; e, starting point should be a so called "best estisate" calculation gq/ based on the best scientific and technical cata available and with a reasonable safety margin added, taking into account, inter alia, i uncertainties in the technical data. When considering what to be , l regarded as a reasonable safety margin, it should be taken into account to what extent the substance in question contributes to ground contamination and otherwise to the general risk picture in the case of accidents where filtered venting is triggereo EMd, , hence, releases of noble gases can be expected". III. THE FILTRA PROJECT Based on the recommendations of the Swedish Government Committee on Reactor Safety, a research project, named FILTRA, wes started early in 1980. It was aimed at investigating the possibility of reducing the risk for a large release of radioactivity, in the case of a severe reactor accident. The project was a joint undertaking between the Swedish huelear Power Inspectorate and the three Swedish nuclear utf-(" lities. The project steering group initiated theoretical and experi-mental research. It also made an early tour of existing containments, trying to identify weak spots, should a severe accioent occur. The , latter, practical approach lead to early realization of the design-specific aspects of containment performance in the case of severe
. accidents. .
In a status report from the FILTRA project, published in March 1981 (5), the feasibility and possible performance of a filtered venting system for BWR containments of the Barsebeck type were discussed. This report was used as a technical background for the policy decisions cited above. e W - - - - - _ _ _ . . _ _ - - _ _ _ _ _
l 6
?
To meet the regulatory repairements, the owners of the Barsebeck plant I started their own FILTRA oes:.pn and Instruction project running in parallell with the research :roject. preliminary safety analysis I report for the filtered vent system was submitted by the beginnireg of ! May 1982. The design propose: by Sydkraf t (the owners of the Barseoeck . plant) was approved by the inspectorate and the Radiation Protection Institute (6) in Decemoer 19E2 after review by Swedish and Germv1 con-
. sultants. ,
The FILTRA Research Project was completed by the same time, providirq more detailed theoretical anc experimental data on filtered venting during severe accident seque: ces in a BWR containment of the Barseoeck : type. The Barsebeck FILTRA systen (figure 1) has been described in detail l~ f - elsewhere (7). Its main features are" a 10 000 m3 gravel bed acting both as a filter and a passive heat
-~ sink, being able to concernse all stearf from a broken primary cir-
[ cuit and from residual decay heat for 24h. Actually, the 24h pas-sive heat sink capacity requirement determines the dimensions as t_ much as the required filtering efficiency. venting pipes with a diameter of 600 mm connecting the wetwells cf
- the two containments to tre filter bed via rupture discs set to 0,65 MPa, the nominal cesign pressure being 0,5 MPa. The dimensic+
ning event for the vent c:apacity is a LOCA with failure of the pressure suppression finction through a 0,3 m2 drywell to wetwell ' leakage area. The containments could also be actively vented - through smaller pipes, fitted with valves and connected to the upper drywell. In April 1995, Sydkraft submitted a final safety analysis report on the TILTRA system, on the basis of which the inspectorate granted a permit to make FILTRA operative. On October 31, 1985 the isolation valves were opened, connectirq the Barsebeck containments to the joint { filter chamber via the rupture discs. That meant that the TILTRA project was completed on schedule - the original target date set by j Sydkraft was Nov 1 - and slightly below projected budget; the final j
~
total cost for the two reacto:rs being 130 MSEK. No additional outage time was required, all internal installations being performed during
-{ normal refueling outages. ,
W IV. PROGRAM FOR ISSUE RESaltrTION FOR THE REMAINING TEN SWEDISH REACTORS. According to the Swedish Gove.nment directives cited in section 2, f studies of severe accident issues for the other ten Swedish reactors
~
should be carried out and decisions be taken on such a time schedule - that any resulting release sitigating measures should be completed by 1989. ammmessa
-7 s
To comply with this requirement, the inspectorate, the Radiation Pro-tection Institute and the utilities launched the RAMA research project by the end of 1982. The purpose of the RAMA project was to evaluate results from various international research programs and to produce, verify and validate a program package suitable for plant specific severe accident analysis of the Swedish reactors. With the RAMA pro-ject as a coordinator of Swedish efforts, Swedish participation in a number of other research projects was arranged such as IDCOR, SFD, LACE and Marviken V. The RAMA project was originally planned for completien by early 1985 and a report' was published on schedule. However, the participants con- - cluced that some further computer program development, verification and validation was justified, now being pursued in the RAMA 11 pro-je00- . To comply with the 1989 time schedule the inspectorate furthermore asked the utilities involved to submit plant-specific severe accident
. s analyses using knowledge and models developed in the RAMA project, in-(, cluding proposals for possible mitigation measures to be implemented by 19B? to comply with the general level of safety indicated in the
(_ 1980/81 Energy Bill. The reactors involved are listed in table 1, and the general lay-out of the different containment types shown in figures 2-4.
- These studies (9,10) were duly cubmitted in May 1985 and are presently being reviewed by the inspectorate and the Radiation Protection Insti-tute in order to prepare recommendations for Government decisions in early 1986 on more specific requirements than in the 1980/31 Bill. In parallell, plant specific level 1 PSAs have been completed for most of the reactors as a part of the ongoing Swedish recurrent safety analy-sis program (the so called ASAR program).
Ongoing technical discussions between the two regulatory authorities and the utilities presently focus on inter alla the following issues: (- c (1) Severe accident management strategies: how to ensure a high proba-bility of preserved containment integrity and with a pressure-less i-containment filled with water above core level as a desirable final state in case of a severe core melt accident.
-{ (2) Possibilities to protect weak spots such as penetrations that may k_' be damaged by direct heating in some containment designs in case of a reactor vessel melt-through.
(3) Possibilities to protect PS-type containments in such beyond-design basis events ns failures resulting in substantial' dry-well/ wet-well leakage during a LOCA (essentially preventing core damage as a consequence a '. . o .: of containment damage due to overpressure). [. .1- - c:. . 9 m eassius
-8 (4) Options for controlled containment venting in the medium to long-term time perspective after on-set of core damage and retention ,
efficiency needed, i.e. with respect to remaining uncertainties $ about fission product behaviour. , (5 ) Containment by-pass sequences . Obviously issues (2) - (4) are closely related to the severe accident management strategies under issue (1). Assuming that the Swedish (lavernment will not change the general safe-ty policy on severe accicents issues described in section 2, the cis-cussiens within the inspectorate are focussing on a technical safety strategy containing the following key objectives: ( - protecting the containment ageinst early damage in connecticn with beoynd design basis accidents which may include core melt - enabling, in various ph.ases of a severe accident sequence, the { cnation of margins against a more deteriorated situation, incres-g
'~
sing the risk of damage to the containment. Such margins may , involve: o increased margins to containment overpressurization creating improved protection against pressure spikes that may occur in various phases of the accident or increasing the time margins available to restore lost containment heat removal systems o margins against uncertainties about the behaviour of airb:ene radioactive substances (e.g. aerosol resuspension) in various phases of a severe accident sequence in order to keep con:ein-ment venting a viaole option in accident management f facilitating the achievement as soon as possible af ter a severe accident, of a long tern stable state featuring a depressurized containment with preserved integrity. This technical strategy it Lased on the following type of considera-tions: Even if theoretical calculations indicate a fairly low prcoaci- j lity for large releases, it does not appear prudent to let a contain-ment remain at elevated pressures - perhaps exceeding the design pres- ,
,k sure - for extended times after severe core damage, firstly, diffuse
(~ leakage of radioactivity to auxiliary buildings, etc may create serio- l us problems for rapid accident management actions bringing the plant I to a stable state. Secondly, there are and will probably remain un-certainties about containment failure pressures and failure modes * (e.g. leak before break) as well as about the behaviour of various radioactive substaces during a wide range of possible severe accident conditions. Thirdly, it should be recognized, that a containment with , a damaged core and in a state characterized by elevated or increasing i pressure, will create difficult decision making and public information problems for the external emergency management authorities.
)
1 J
-9 o' '
~0bviously, the Barsebeck FILTRA system meets the objectives of the '
safety strategy outlined above. For the other reactors the utilities ; may propose alternative technical solutions, based on the better understanding of a number of severe accident phenomena that has been achieved in the past five years. V. REFERENCES (1) Safe Nuclear power? Report of the Swedish Government Committee on Nuclear Reactor Safety (5001979:86) ; (2) The Swedi.sh Governm'ent Energy Bill 1980/81:90, Appendix 1 (3) Swedish Government Decree, Octobe'r 15, 1981 (4) Swedish Nuclear Power Inspectorate letter to Sydsvenska Varmekraft AB, Jan 15, 1982. b- (5) Filtered Atmospheric Yenting of LWR containments, Project FILTRA k Progress Report, March 1981. (6) Swedish Nuclear Power Inspectorate. Review Memorandum on the proposed FILTRA Design, Dec 2, 1982.- .
. "(7) A. Persson, T. Andersson, FILTRA: Filtered Plant for Severe Reactor Accidents, Nuclear Europe No 5 (1983) p 22.
(8) Project RAMA Final Report, January 1985.
'~
(9) OKG Power Group: Severe Accident nalysis Status Report for the Oskarshamn 1-3, April 1985. (10) Swedish State Power Board: ( . Project MITRA Report - Severe Accident Analysis Status Report for Ringhals 1-4 and forsmark 1-3, April 1985.
?-
6
. A : rue.- n .. . . .s?
. . c. . " .
. .:r ., ;..
, . .c m
-10 f
3 C - C o
% C o G " na y
2 G C U _a 6" W 8 8 3 *'* & U D g 3 -- v C ya m _u C C w a e # """' U. t 6 . w . . h -z u U~
- C s 3 C 3 s b a O
x o o m a hh u -. 5 a . m .-
.
- C E D :
U 1 = - --
.C h .e. .,
" ~ z & 3 g =c 2
u , e c - O C e k x
"$ue U
v
- E . , , u
. e u h ..e 0 g in g
' f E i
O
.sb is 5 in E
\ ,, m h
- u a E 3 --
h8 i, e .= 0 en
.O. m 03 . g
.* ?i, M .
e *# aD. W a
@
- C e a.
(^ - ==
- a G e , cJ mU mO 6
v e v g
- s =d b .
C E ** % . W4 -e a L " c e M h O C (D W . C In ,,,,
> &* a. e U a @@ G c e CC D< c g #
ge3 8 a a w 3 3 "o E .u . C ~ UJ - a G == w =* ** @
- O
% . h M t O .C =
te O O . . 2 f = m O4 a 4- C
>w c >
h $ c C
- 8 8
a C g s. mh A h C e
.e m (- o
@ m CT D "* q =>
-= c s. O C - C y =>= m CD e CD *a L !
-= N E OO C C **
amU a 9CC .l"n s .C. yE ~ s f hs .e ec C O a a c om % C: O C
.-e C g* O-m 3 o U
Ih m@ h" C U C C h C C C C C C m C Y C CJ c d o N & C e & cCe a (. OM u e
> " " N N n N N c: OC < et
' e a e e D M N n G D w w w im L (n" . $
g
- 2m g }@ **
Li
- b > m
- Oy g C T U .*e o e g g E 2 E == E.
. - - er g9 ,g E 6 D
.- 5M @C y "# E b- 40 W W ars
- U r
m E> ag 3 K j C ha i oo a so e U y
# N N n d e.- v.
s - 4
- ne 4 ao
. .o g g ~ n
. O I 8 rm
# 3 J m e
. e -e
~
e u e u :.: - E # m m e e g b W 4 E m E w mm C p rp !
- X m X C se s.* 2 I e m -. o om ,C, ,C,,
e C C a: g ga g g w . m I
-11 '
s t B. l P t k
,!'*^ . , . . - - .
fr~=t p-j
... ,. i 3 A ', We i _4 '
O
'y
- n. .. J
.s]-
?, 1 .-
.e -
f N, g .; , #l *1 til e
- N" k
) r" i --V r :*j - ; :' s./ :.
- -;, , _ . m t.;-
q- -
-4
-L%--
g^*M I, -'
}9
. w -_ ..
~ g- y 1A r
). .~...A*~*d. ..
- ( ,. -. .
- 41_= .I~I-- ~_ g_.
. t l
Figure 1: The Barsebeck FILTRA filtered venting system with the ' '
- a. 10 000 m3 gravel bed c:r.6nsor/ filter in the middle.
( DRYWELL f HEAD REACTOR X4' - PRES 5U RE
. r -. VESSEL
?G 'h,
., h N -?,;,
*/, !9 a '. .
.s 9 *- N. O g .*>a :Ji. [* .'s e- ;,..]
W7 &; :&
, t -
e . : o =
, f, 1 s
? DRYWELL - r p 3 ,
,I
. W. * *
(~ $ Ia g"fl I c;e . DI APH R FLOOR
~~~~
s ACM .: , ,u 4 n y' lO a i ' 2. f.* ! w'
&}~rmia L
gbN$b.D.. g,,,, PE D ES T A L ( f : 5 Y i' f I l p. p" $UPPRE55 TON PRESSURE 1 ,,. .. g.- '. - VENTS l W ETWELL - ? ' i .., ,,,,3, g , f l n
.e
- 9 .i;
- ...'. m./ .<; {;
, .o l
I l i a E . ;- - i ! iE / - N - 4 CONDENSATION
. f h.--f,* i .}'w POOL 3 - ,. 5 ,
. %.1.*;';; ? '-J. nn a ';/, p#, . s. * ... vy .
Figure 2: Pressure suppression containment for BWR with internal pumps (Ringhals 1) gemaan 9
12
)-
~
l
, Figure 3:
.e ..';..
- Pressure suppression
;* **.. . p *
-*' "*. "$ Containment for BWR mEACYoR ,*
/ g' with. internal pumps e=Es5unt _ ,,
.i _ oRYwEtt (Forsmark 1 & 2) vE55EL i '1 -
o - s ... - - .
>r:n ..
- s. , *;' *
':-- ?eto. * *[
.=.. s.
A **
'.2 '
\
k.3~6- ;,(:.
*; .... Y..'W>MA.
;,, e; , ,,
s Y.:
"' R -
4' -- WETWELL h, I i 6 i.4 .w
- v. p P.
ll.s. 4 F 't LJ {
., .. E1 ,- $
Qf '
. . l _ J h_.'d 4;--- g m a
- . - - . ., t mz._' ._ m.
?. -- .,. f,, .. s
- 74. . -
'y _.-.;
gg
,,, CONDENSAfl0N y __
*1
., 4 .* j ,===:=:: n,'
?
- POOL -
, 1 .;r r* - PEDE5TAL e
(. . .?.Ia w L., 7 ' r'--
, I* . -- ; .i, e * * *?. ; i.E
. . l;. '. .'.W.
.* b-w, .- p. *
- n
- w ..
-l
'y'.. '. l'j.h .:t, ,..' j .,;. *
. * . m CONTAINMENT SPRAY w
c.I - ' " ^ ^ ^ ^ +; Figure 4: 6 V, f;- Large, dry PWR
, containment ..
(Ringhals 3 & 4) - o fe , 1 STEAM C- , M CENERATOR
, (.., 5
- c/.
y ( e l P '
..~ .< > .
f'
. o , , p g .
(: i gg k y@"bb v
- l M
- W "
hf l /f ._ REACTOR i c *.' % , t VE5SEL a
~ ;
... ). * *l(' .
;. , eW.'.an . F@. ,,,: .--% ,
\ REACTOR CAvtTY SASEMENT , j e
1985-12-30 9 . PUBLIC INFORMATION NOTICE Swedish Authorities Reconmend -Improved Protection against Releases in the Case of Severe Reactor Accidents In letters to the Swedish Gevernment, the Swedish Nuclear Power Inspec-torate (SKI) and The Naticnal Radiation Protection Institute (SSI) pro-pose improved protection against releases in the case of a severe acci-dent for the nuclear power lants'in forsmark, Oskarshamn and Ringhals. The proposed requirements a:e based on reviews of plant-specific severe accident studies submitted by the plant owners in. April, 1985. Some main points in the proposec requirements are:
- Improved protection acaimet early and direct damage to the contain- '(
ment in the case of a ccre inelt. Measures proposed by the owners inclu- . de improved protection of particularly exposed penetrations and support , structures, as well as installation of additional equipment enabling rapid flooding of the lower dry-well in the forsmark 1-3 and Oskars-hamn 3 BWR's. f -
- All containments shoulc be provided with devices for pressure relief, protecting the containment against the major number of such beyond-design-basis events that may lead to loss of containment integrity due to overpressure. The relief devices should be so designed that they can function independently of operator action and of other safety systems if the containment desig, ;:rressure is substantially exceeded. In addi-tion, operators should be arble to use the relief devices actively as a part of accident managenen actions. The relief devices should be so designed that they, together with other safety systems (e.g. contain-ment sprays), ensure witn nigh reliability that releases to the envi-ronment are kept below 0.1 percent of radionuclides inventory in a 1800 MW(th) core, noble gases excluded, and with some regard to how various radionuclides contribute to radiation risks, especially ground contamination. Appropriate measuring equipment should be proviced to enable prediction and nonitoring of releases through the relier de-g vices.
t'
- The improved protective measures against releases in the case of a severe accident should be implemented by Sept.1,1989 at the latest. ,
[ The release protection level required for the forsmark, Oskarshans, and
' Ringhals plants is equal to that required for the Barsebeck plant with its filtra filtered vent rystem, which became operational in November, 1985. However, alternative technical solutions to the large Barsebeck filter may be envisagec, according to the SKI-SSI review report. Such alternative technical salt :tions are presently being discussed with the
. utilities concerned and include:
- Large area, unfiltere: and reclosing relief systems to cope with dry-well to wetwell isolation. f ailure in combination with a LOCA i:n BWR pressure suppression c:ntainments, thus preventing containment =amage which, in turn will precably cause failure of core cooling in such sequences.
e
4 y,
- Improvcd containment spray cystems.
- Small area venting systems with retenti:n cevices such as scrubbers or filters to provide long term pressure co~r:1 in the containment.
The proposed requirements are ' based on the general severe accident policy guioelines issued in 1981 by the Swecish Government and Parlia-ment. These guidelines put a high priority en prevention of extensive I ground contamination in order to limit social consequences in the case of a severe reactor accident. This priority is the main basis for the 0.1 percent release limit, which also provices good margins against acute radiation illness and fatalities. 1 In addition to these general policy guidelines, SK1 and 551 inter alla states the following in their review report as a basis for the proposec requirements: (. - The primary safety objective is still to ::revent core damage. This preventive safety level has improved in Swecish reactors in the past five years due to safety-enhancing measures 1.mplemented on the basis of operational experience and plant-specific P .A's. Also, the safety level is better known. Nevertheless, it must be taken into account that seve-(( re accidents can occur.
- There is substantially improved knowledge on severe accident phenome-na, indicating that containments have good espabilities to withstand severe accidents, provided that " weak spots" identified in design-spe-cific studies, are strengthened. Furthermore, design-specific accident management strategies should be prepared, aisned at protecting the con-tainment function, and, in the long term, at reaching a stable plant state with the damaged core cooled in a containment at normal pressure and possibly filled with water above the level of the damaged core. ,
- Even if theoretical calculations indicate a fairly low probability for large releases, it does not appear reasonable to let a containment with a damaged core remain at elevated pressures - perhaps exceeding the design pressure - for extended times. Fi.rstly, diffuse leakage of radioactivity to auxiliary buildings, etc. may create serious problems
'for accident management actions bringing the plant to a stable state.
Secondly, there are, and will probably remain., uncertainties about con-tainment failure pressures and failure mo:es (e.g. leak before break) as well as about the behaviour of various radioactive substances during
- a wide range of possible severe accident conditions. Thirdly, it must be recognized that a containment with a dam; aged core and in a state
{ characterized by elevated or increasing pressure will create difficult decision making and public information problems for the emergency management authorities. A Swedish Government decision on the proposed requirements is expected early in 1986.
/
M/ #2 ars H ec, berg Director, Of fice/ of Regulation and Researen SWEDISH NUCLEAR POWER INSPECTORATE
. . . O W*'* t%r-Our ref I Saedish Nuclear Powerinspectorate STATENS KARNKRAFTINSPEKTON b**'**'" fr Yewe ref.
i 4 l
... E b,[
I 4 1 J N I I a p . -I N et ov.* e .% e.. h /.yAc!/ j{I l ,, ?- r l El I'd
., i ., m .w . . .. . .. - .;..,+. . ,.:.'c.'s.tP. W ..y.t.t,3 .>.w .:. ,;. ,- w . 4 ,'t - - - (-
.... ~-. _.
- l. 'N
.'. o..
l g. .L Q \ h* .b
*^ -
p: 4
- .
- m' i :, s<
-e
. v.$,1 ci f ,- - r - - - - : ' ::.
T. Y.
,I ??
- fj4 s e. 4 . *J1 k m ; e . ... .e > cm ,
' '6
);.% il:..7 8 * :( _: .* *f.
"''_'h yu.A p e;M M D'.*%$
j ; - - ol .a l' .
' './OJ *..
l j
} .
- .,'- !:m s a =, ,.
'/d ?E 7 l:L' A '* :so ; '" ** '=
- l .
, p"_...~.+..,....,
0, l '.ti.. 7 , H+ 8?
.,k f'
l 0 *.* it. A/ 9
;' .C
} v a r i o' g <, .,. g ,. .
ll
- Ye'., lI YA ~
5' % h v . , a ,' s.~. e
- E'. *.s **?e's5.h v. ' '-&'$~?.[ste f -
,t ,
\-
. - - n' d' ' I.. a' l 'gA. "d,,'*
7.* b*b' . p *, .
, g.A ... e ' "v' r I.s;.a.* w .r:6... n ;
\
O 4 W.
- O 9
9 9,g r 3 3 ( dL z Postodress.Poswess Gotuodress. Office Teldon. Telephone Telegrom-Cobie Telen g Box 27106 Sehistedtsgoton 11 08-63 55 60 SWEDATOM, Stockholm 119 61 SWEATOM S
- S-102 52 STOCKHOLM Teleiax 08-61 90 86
O M** w .N ,g b h'* - Y **8' ** tr . veve .,# STATENS Kre*#IAFTINSPEKTION Sato'Sh Nucecr 3 :wwer inspectorate 4 t I l A Lim F' , 4
)
)
AI.'.C.!.d 4 s . , . , . . . . ... r , v,
~ *
( ~ . ; e
. n s . s. . . ' ' ' ' 'L ' l
(
^
- I
@3fm. . . . C~'g .* 6
~- .
e~. p . 4 i
' ~ 'y ,
m a
< l . . s. . e. s: .
. . e, .; - ..
, dA > , G *. .. ... ,
.1 } %1 E
< . , C, * #.
1 P. o *
. s1.
,. g. , ; . , 3 g ;, n'/ ,
h
= I h 5 LJ 4
< . . . .. M , ' g, s
; A,..
i g
- s .. .* **
.. -..* - ..*.*,..w,e -
<S . .w . s *. w ? f .. . ; t, e . ,. . , g , e , .
- 8
- A .* I
* === ,
A
- g. .
N d I l av
.. ~ ,
, Postortress.Postocaress Gatuacross .Off ge Telefon Telephone Telegrorn-Cable Telen Box 27106 Sehtstecspeton 11 08 63 55 60 SWEDATOM. StockW 119 61 SWEATOM S
*S 102 52 STOCKHOLM Teleios I
*' 08 61 90 86
,, FILTR A.Mf 55 Sida 9 (to)
,, . 1986-11-19 l FILTRA - Mvss =
f 6 9 5 .
<=,
i e
~'
-i N;I .
~
_.2.: J DMm i
)
s k% J h a
~~N f(\s sf f r j
~
s N -n fjfA 2, ss PWR D! 11 - a i s\ -s
- s. q 4a $$
BWR , t
.O Y ', ,
1 AS E A-ATO M IMEM. m 5 E a 6 - _ _ _ _ _ _ _ _ _ . _ _ . _ . _ _ _ _ _ _ _ . - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
(- q9 i e 1 I l bNb I l ! s i GesellschaftfUr Reaktorsicherheit (GRS) mbH i_ l b Review of Selected Aspects of the Proposed Vent-Filter System in Barseblick (FILTRA) [ 1-i l- i i-l h ;
. K. Bracht .
w H. Friederichs _ i E. Hicken ; D. Risse J. Rohde Cologne, November 1982 - o i Glockengasse 2 5000 K6ln1 Telefon(0221) 2068-1 Teiex 8861807 grsc
i g i t -1 - I j f. t
, a L
Review of Selected Aspects of the Proposed i Vent-Filter System in Barseb5ck (FILTRA) GegeEgl_Remapks The contract between the SwxEsh Nuclear Power Inspectorate (SKI) and Gesellschaft fur Reaktorsicherheit (GRS) defines the scope of work and work packages as follows:
"The principal purpose of the GRS review would be to check I.
the basic assumptions and models used in the analysis, e.g.
- ; so that no important phenomena have been overlooked or treated in an inadequate way as can be judged against the ,
background of the current GRS research programme on severe accidents. Thus, it is anticipated that the GRS review would not include a detailed review of all calculations, etc. I l l It was agreed, that no detailed investigations and calculations ; were to be performed by GRS within the frame of this work. Within this general scope, the GRS review includes the , following packages: I) Choice of accident sequences for the functional 4 requirements and comments on possible residual risk. h@ II) Modelling of in-vessel thermohydraulic phenomena, such as { coremelt and slump, steam spikes, zirconium reaction, RPV l penetration, etc. l t c ; III) Modelling of ex-vessel thermohydraulic phenomena, such as i I RPV drainage, debris formation and coolability, heating of containment, melt transport to and coolability in pool, etc. 4 IV) Modelling of source term and fission product behaviour in l containment before and after break of rupture disc etc. ,
.)
. t
, f
~
V) Design and performance parameters of the vent-filter {, system such as disc rupturs' pressure, vent capacity, ' I condensation and filtering efficiency, etc. VI) Comments on accident management, instrumentation, etc." I The review was done by GRS under the aspect of realistic assumptions ("best-estimate") for the description of the ' accident sequences and the modelling of separate effects. In addition efforts were given to point out the margin of possible existing uncertainties. The overall aspect of this review was to check on the choise f- of accident sequences, including the functioning capability of the.FILTRA plant for the time period of 24 hours after the ' initial event leading to a core meltdown but without the j possibility of steam explosions. t c
$ 6 J g.
s t. f. I' i l e , 6 I 1 I L__-__--_ _ - _ _ - - - - _ _ - . . _ _ J
3-1 .
< ).
l t
- I. Choise of Accident Secuug 's for the Functional l Requirements and Comments a Possible Residual Risk { _)
i The task of FILTRA should be to prevent serious environmental , , consequences due to chains of events, which involve , overpressure in the reactor containment. Two complete investigations of release have been performed for the FILTP.A-project as reference cases: transient with loss of all AC power (TB-case) and pipe' rupture in the cooling system combined with simultanects f ' J- leaks in the intermediate floor, no emergency core-cooling and { E-containment spray system (ADE-case). [,
).
i~ The TB-case leads to a core meltdown, followed by a pressure , built-up in the containment system, finally ending with the , opening of the rupture disk. Hereby the steam production is caused by the decay heat of the melt in the containment pool, when the water becomes saturated. In the other reference-case (ADE) the rupture disk is opened already during the blowdown, while the core meltdown happens later on. In both cases the meltdown of the core takes place under low I pressure conditions, because in the TB-case it is assumed that
&& the decompression of the reactor vessel is initiated automatically -
by the opening of the pressure relief valves after the water ;L < f ~ J level inside the vessel falls to less than half of the core height. For this the assumption is made, that DC-power is available'for several hours after the loss of all AC-power. :~ That a loss of DC power during the first two hours is unlikely [: o ' to occur is considered as an important guarantee. : 1 l t i mm g J
- t t
~
l' i
,- L -
Since forced blowdown is assumed to occur automatically before the core meltdown starts, then according to /1/ a core meltdown under high pressure conditions followed'by a ' melt-through of a pressurized reactor vessel in Barseb5ck I is unlikely. Attention should be drawn on one point. - - After the decompression of the RPV, when the pressure in the ;. vessel has reached a level less than o.5 MPa/13/ the pressure : L relieve valves will close again. The following core meltdown , i-will cause a new pressure increase within the vessel due to '
- f.~
y temperature rise and steam production. This will lead again : the relieve valves to open and operate under high temperature h l conditions, so that a loss of the functioning capability 1 of the valves can be presumed. i. I There would- be not difference to the reference case TB, when h I some valves will stuck open under such conditions. Furthermore k information from /13/ are, that there exists additional F possibilities to depressurize the system. This should be checked l for considering the accident management. An inportant point for the discussion on residual risk - is the possibility for leakages bypassing FILTRA. The j q (j; effectiveness of the FILTRA-concept to retain fission products F l will be affected by leakages from the containment directiv to ;
> 1 J the reactor building and - as the filtering system of the
[ reactor building is not operable because of' loss of AC-Power- [ to the environment. The normal tightness of 1 Vol. %/d might ! rapidly increase by orders'of magnitude, if penetrations . (airlocks, ventilation ducts) are not properly closed in the
)
case of the accident or if seals in e.g. cable penetrations become untight due to temperatures beyond the design limits (superheated steam, possible local hydrogen-burning) . ? i
- - - - - - - - - - - - - - - - - 9
I y 8
)
i
~
, . l The fission product release via leakages might exceed the release {
through the FILTRA-System. The probability of an untight
~
containment can at least be compared to that of a failure i of the FILTRA-system. The probability of such failure of the ' isolation should be much less than the probability of a rupture of a vent pipe. If a leakage
-, of the containment is detected , one should try to depressurize the containment by opening the bypass valves to FILTRA.
Confirmatory calculations for radiological optimization are i
, recommended.
j
, r On the Tenth Water Reactor Safety Research Information Meeting i (Oct. 15.82) a paper was presented discussing the results of a study about a hypothetically small break loss of coolan: ,
accident outside of containment (Browns Ferry Nuclear Plan:) /16/. The accident studied would be initiated by a break in the ! scram discharge volume piping when it is pressurized to full . reactor vessel pressure as a normal consequence of a reactor scram. If the scram could be reset, the~ scram outlet valves f would close to isolate the scram discharge volume (SDV) and the piping break from the reactor vessel. However, reset is possible only if the conditions that caused the scram have been cleared. t l (y. was assumed in the study that the scram signal remains in , J
, effect over a long period of time. From this investigations the '
(' conclusion could be drawn, that the possibility for an f.. accident sequence leading to a discharge of coolant directly l through the break into the reactor building atmosphere ' thus bypassing the FILTRA-p.' ant would exist. i The possibility for such an accident sequence should be i checked for the Barsebhck plants. I Another point of interest for the discussions on possible residual risk is that of a non inerted containment situation , during the startup and shutdown pe.riod of the plant. 1 l
O o., The most probable opportunity for ~ malf unctioning of switches, valves, pipes, grids, etc..is given during startup and shutdown operations with adjacent transients. During this time span, when-the power is below 65 % the containment is not < necessarily inerted. The. probability for a severe accident with a non inerted containment seems to be at least not much smaller than with inerting. All investigations for the accident sequences assume inerting and no hydrogen burning.Several accident sequences A ~ with hydrogen burning are possible, if the containment is not inerted , , One of the most severe cases would be the loss of AC-power. (TB)
~
Steam and. hydrogen enter the wetwell first via the SRV and
,, during loss of RPV integrity through the vent pipes. Rough o'
5: calculations point out, that if loo kg of hydrogen mix with ' the air in the wetwell, and the burning starts (detonation), then I. the static pressure coulq reach the value of 1,2 MPa with temperature above 2 coo C.The floor between.wetweil and drywell as well as the containment wall and penetrations would probably not withstand these loading conditions. This depends on the functioning capability of , the vacuum breakers. Also the vent pipes can be affected. ! Although So kg of reacting hydrogen would lead to approximately ' o,75 MPa, which would still mean too much pressure on the ficor. Facing such loads it is recommended, to take into account for
, the accident management that the containment will not be
(;, inerted, when an accident occurs. Inerting should be immediately I started in the case of accidents if it is possible. The plant i m should always be inerted unless it is shutdown. i. < From the review , the conclusion could be drawn, that the most
~
probable accident sequences for the layout of the FILTRA-plant , are covered by the choise of the two reference cases, that means , the TB- and ADE-case and that the possibility for a core meltdown under high pressure conditions is unlikely to occur. mm 1
t y s 1 6
.i
'J of importance seem to be the problems of leakages bypassing FILTRA, including the check about the accident sequence for a ,.
break in the scram-system. In addition it would be necessary to think about a non ' inerted containment situation when considering a reduction of the possible residual rich. I '.
*,v r ;
(o , i. I' WS t
.~
e h ammen.
* ;T g 1 II Modelling of In-Vessel Milt Thermohvdraulic Phenomena General Remarks The course of core meltdown accidents has been analysed i for the sequences: - '
Complete loss of AC-power with depressurization of the RPV ("TB-case") LOCA, failure of emergency cooling and loss of AC power for more than 24 h (ADE-case) e Detailed analyses have been performed, especially for the ' i TB-case with depressurization. To a minor extent also .
,t calculations have been carried out for the TB-case without depressurination (only in-vessel melt; no containment atmosphere investigation). !
L In the following, comments are given in chapter II and III on hcw c. sequences have been analysed and on models and assumptions which have been used. Possible weak points in the analyses and their effect on FILTRA cre discussed. g-t (* r - o>
,' l
. ! N Y
i l 1 1 1 Core Meltdown ' i j Cgsg ! I l
~
For core heatup and meltdown analyses (performed by Studsvik t 3 Energiteknik, KWU) the Battelle Columbus code BOIL has beer. , used. The analyses have been supplemented by hand calculations. j. .
- i. 1
! -)
<- Also at GRS, BOIL as a subroutine of the MARCH-code is used ;
for core meltdown analyses. Based on benchmark calculations with the german core meltdown code MELSIM we expect that BOIL estimates the time.for core heatup and core degradation in [ a sufficient way. ;.
Conclusions:
i
+
We expect that BOIL is an appropriate tool to analyse accidents, leading to a complete core meltdcun like the sequences analysed in the FILTRA project (no delayed ' ECC injection when core may be already partially degraded). g of course it is necessary to take care of model simplifica:ica ,
, in BOIL when interpreting the calculated resul+.s (e . g. with
(" respect to H 2 -generation or to the modeling of the water level inside a BWR vessel /10/). f[ 1. J mm
, - lo - ,
i 4- . Zirconium - Water Reaction, H- Generation L ______________________________2_ __________ l The amount of the Zirconium-water reaction with hydrogen generation during core degradation may i influence containment pressure history (H2 partial pressure); ( effect the energy release into the drywell and wetwell; cause violent H 2 -combustion if oxygen is present in the f ; containment atmosphere 0 l influence core heatup because of the exothermal Zr-H,0 .
. 4 '
reaction For the TB-case, the expected probable amount of the oxidized ' fraction of Zircaloy, estimated by using BOIL, is between 8 % (with depressurization) and nearly 25 % (without depressuricatic: , corresponding to hydrogen masses of about 12o kg and 350 kg i respectively (ADE -case: 120 kg H2 ) . A hydrogen mass according
~
25 % Zr-H 2 O reaction is also taken into account for the containten pressure analyses / 4 /. F 1 With respect to the Zircaloy inventory of the Barsebsck core *
'j 3
pg (~s34 . 1o kg) the generation of about 1500 kg H is possible 2 by the Zirconium - water reaction providing complete oxidation. I- .j I-t - Hence,the estimated probable amount of hydrogen during the . l i core meltdown phase is comparatively low with respect to the l existing potential. , G t-O m 9 e sd w_____--_.__._
11 - t One reason for the expected low amount of H is, that in the 2 sequences which have been analysed in detail (TB with depressurization; ADB), the coolant level inside the RPV . is already below the core when core heatup starts. Therefore ; nearly an adiabatic heatup will take place without significant Zr-H2 O reaction. Even in the TB-case without depressurization, where core meltdcwn starts, and when the cociant level is still inside the core, the hydrogen generation is low as mentioned before. t One has to consider ( as pointed out in /5/) that BOIL neglects {, the effect of EWR fuel assembly shrouds and control rods on o the Zircaloy oxidation. Furthermore we have noticed that the ' fraction of oxidized Zircaloy calculated by BOIL depends on different parameters like fusion temperature, meltdown model etc. (typical values for the oxidized Zircaloy fraction we have calculated for a german PWR range between 4o % and 7o %). Additionally one should have in mind that rather little is known about the kinetic of the Zr-H 2O reaction at temperatures higher than 150o *C. Sometimes the relations l- in BOIL by which the reaction can be described (Baker-Just, { Carthcart) are assumed to be conservative in respect { to overestimation of the reaction. On the other hand there is no model in BOIL which can describe a possible failure of k 1 - oxid layers after a certain time of reaction (" break away") , inducing an enhanced oxidation of Zircaloy. This neglect may cause underestimation of H -generation. 2 A higher H2 -release as calculated will enlarge the H2 -partial pressure which influences the time when the design pressure of the rupture disc will be reached. Based on the data in / 4 / where a maximum H2 partial pressure of about o.8 bar is expected in case of 25 % Zr-water reaction, under the (pessimistic)
c
, . [
I'
- i
, a assumption'of loo % Zr-H2 O reaction with an corresponding partial pressure, the time for reaching the rupture disc design e
pressure will be shortened by about two hours. This shortening r is expected to be of less importance in comparison to the overall time of, accident up to the moment of venting. A large H2.relea e (maximum 150o kg) possessing high temperature may cause a temperature rise of the water in the condensation pool in the order of lo-15 'K. Because of the highly subcooled conditions in the pool water this effect seams'be of minor importan. .
, , The effect of a greater release of energy in case of a Zr-H O l
, 2 s, reaction to a. higher degree than expected, may shorten core heatup in the order of some minutes, the overall effect i on the accident time scale is unimportant.
At high power level (2>65 %) of the plant, the Barsebuch containment is inerted by nitrogen. Therefore one is not to be concerned about violent H - combustion in case of accidents 2 initiated during that operation mode. There may be a different situation in case of accidents initiated during low power operation, when the containment is not or insufficiently inerted. This subject is discussed in chapter I with respect to the ( question of residual risks. As discussed there violent H combustic: 2 might have severe consequences with respect to the integrity of ,
, containment structures.
s !
Conclusion:
Even H2 generation during core meltdown to a greater L' extent than expected in the analyses will not influence the course of accident in a significant way (provided H2 -combustion in case of accidents without inerted atmosphere). 1 I C l We
i l 1 99 E9_ E9 EE ggatigg3_ggit_ g lgggin g_19 tg _t hg _ Lggg g_ Bgy;glgg gg ' t Great effort has been undertaken in the FILTRA-project to analyse , the phase of core _ degradation and melt slump into the lower plenum of the RPV. ,,; In general it is believed that the detailed mode of core [ '. j degradation can be covered by the use of two models in BOIL, j describing'different behaviour of the molten core material ;, l (melt pool formation inside the core before slumoino into t, the lower plenum; fmmediate slumping). p t 1 We expect that this procedure is an appropriate way to [' , N (' j estimate the time scale of core degradation and the il mass- and energy transfer from the RPV into the containment in I. the case of accidents with complete core meltdown under [ low system pressure as well as under high system pressure. ; i Based on the results of BOIL and supplementary hand calculations, in /5/ an attempt was made for the TB-case to estimate the most probable kind of core degradation. I It is difficult to decide to what extent the way of core degradation described there may be realistic due to lack ; of information about material behaviour under core meltdown , condition e.g.
'~
We believe that it may be helpful but not necessary to analyse !. k$" in detail the way in which the core looses integrity, by using
~
,, models (as performed), which are expected to cover the {;
e~ scenario in a simplified way. ! i O m- ____.________.m _ _ _ _ _ _ _ _ _ __
l
* - t I
a
+
h Core Material Behaviour inside Lower Plenum of the RPV l Again it is difficult to presume if the core material in the lower , plenum of the RPV will-form a kind of melt pool or a debris bed. We expect (like in the analyses performed) that the formation , of molten material at the bottom of the RPV may be the more , probable way than the formation of a debris bed. Even in the case where initially a debris bed is formed, one should consider that dry out inside the debris bed can lead to a remelt of i the material. . For the filter system the special kind of core material behaticur , is not important (still in-vessel melt). p Reactor Pressure Vessel Failure We agree that a local failure of the RPV has to be considered much more likely than a circumferential failure because of i the great number of penetrations through the RPV bottom, the short failure time for these penetrations in comparisen tc, the calculated f ailure time of the RPV-wall, the po:sible low viscosit'y of core melt in that situation, which enables a rapid discharge of large anounts of melt within a shc_~ ni e (_/ interval, if a local failure occurs,and prevents RPV of ; v further attack by the melt. ' ( The failure of the RPV marks the end of the in-vessel-melt phase ' of a core meltdown accident. >- k l ?
.E l
p s" l III. Modellina of Ex-Vessel-Melt Thermohydraulic Phenomena Elew_ef_ Melt _9ut_9f_the_RPV i As the probable situa' tion for the flow of melt out of the RPV after its local failure, in the safety report / 1/ , a i discharge of melt over a time period up to several hours 1 i is expected. F ' We concede that there may be a possibility for a discharge over such a period, but we are not of the opinion that a generali-
'l zation with respect to " probable mode" is permissible in that j re way, because ' l i
1 i it is shown in /6/ and we agree that the { properties of core melt in case oj local failure of the RPV give the possibility of a discharge of the complete mass of the molten core within a few minutes during that phase of an accident; , the situation inside RPV before its fcilure, e.g. the mass of molten material inside the lower plenum, has to be considered uncertain because of the insufficient information - about the detailed course of the core meltdown and slumping process ( chapter II). b5 In case of rapid melt discharge from RPV one could expect
~
local loading of the concrete floor below the RPV by a jet. ; t j It is more likely that in case of a rapid discharge ; a jet will be hindered by constructions in the space between RPV-bottom and the concrete floor, the jet might spread up over the distance of about 1o a between RPV bottom and concrete floor. m-
- i d
P Both mechanisms will weaken the intensity of a jet. Even if penetration of the concrete floor by the melt may occur, this event will influence only the time when melt comes into contact with the water of the suppression pool.
Conclusions:
1 We do see a potential,for a different discharge mode '
, of melt from RPV than it is expected as the probable mode
~
l - in /1/. In case of deviations from the assumed scenario A for melt discharge from RPV consequences of major importance ' l on the accident and on FILTRA respectively are not expecced. 1* l 1 l 1 1 4 14 [' [ O O erp 1
17 - , d Steam Generation and Steam Release after Melt Flow into the Condensation Pool When melt flow into the condensation pool is expected, the water in the' pool will be subcooled (<v So K) II Thus, the energy of the complete mass of hot core melt can be absorbed by the water without reaching saturated condition, ,
, hence without generating steam. This behaviour can be expected at low melt flow rate into the pool in any case.
If core melt will be discharged in a more rapid mode, one may suppose that steam, generated by quenching the melt, may be condensed incompletely, so that steam can be released ; into the containment (extreme event: " Steam Spike"). For the design of the vent capacity a possible rapid steam release into the containment is an important feature. Therefore great effort has been done in the FILTRA-Project to analyse i this subject /7, 8, 11/. i The height of the water in the condensation pool is abou' ! 6-7 m. From the results of numerous experiments the condensation
,{ could be drawn, that a strong condensation effect for steam generation in much lower water layers will excist. ,
)
. i
! 1
- For instance, experiments for the Bodega-Bay- and the Humboldt-Say i. 1 i
plants and in other PSS-experimental facilities showed, that j large quantitles of steam introduced into water with a small {' submergence of partly less than 1.2 m and with subcooling ; 1 temperatures less than 2o K were still condensed completely /15/.
~
1
- 1) Compare the remarks at the end of this chapter t
l l l b
Even in the extreme and rather unlikely case of an instantanicus melt flow of about 16o . 10 kg into the condensation pool and assuming that the energy of the~ melt is completely used for ' steam generation, it can easily be shown that containment integrity is not threatened without any venting capacity, assu=ing only half of the stored energy in the core melt would be , transmitted to the pool by steam condensation. This mode of melt flow into the pool has to be expected as a theoretical mode with less probability. even if the melt is of low vis=:sity like water, a few minutes are always necessary for a flow through the drain pipe . and the steel door into the pool, r during the ex-vessel-melt phase, melt of high temperature will cool _swn rapidly as a consequence of radiation. By this the viscosity of melt (constituents of the melt are mostly oxide with high resolidification temperatures) wi 1 _ increase, which will hinder melt flow. melt flow over a certain time interval will induce convectio~n inside the pool transporting more subcooled water to the place where steam is generated (which may enlarge condensau :- , 1
Conclusion:
hi A rapid steam generation in the condensation pool is according to us not a threat to containment integrity because ' of the potential for high condensation rates. i I I i 9 l l O Wm m_______ _ _ _ - _ _ . --
s (..
- s. '
39 - ,' i l gggaghs_ggggggning_sghgggligg_gf_thg_ggg1 It is stated in /'/, page 6/, that a subcooling of 84 K will at least exist, while the melt is cooled down in the pool. In fae: the subcooling depends' on the par'tial pressure of steam above : the pool and on the extent of thermal equilibrium between ' pool and a'tmosphere. If the temperature of the atmosphere , and the pool is always the sare,there is a subcooling of.v 30 K at the beginning decreasing to Le9 K when the rupture disc wil'_ ' open. If the temperature of the atmosphere is lower due to the L partial ' pressure of non condensable gases, the pool will start , g l,'* to boil. No subcooling is Icft when pool temperature slightly I exceeds loo C. During the time, when the core-melt drops into ; the pool and is cooled down, the subcooling of the pool will decrease according to our opinion from So Eo 20 K. 4 a
- 5 a
"e 1 u j
,, . - 2o - l Q
Lona Term Behaviour of Core Material in the Condensation Pool and Influence on the Containment Atmosohere l The safety report /1/ summarizes the results of'the behaviour of melt inside the condensation pool as follows i (steam explosion excluded)- 1 falling melt will be fragmented when it passes
~
~
,. through the water in the pool
's the melt will probably be fragmented in pieces of contimetre size
.f.
~ .
these fragments will be resolidified, whan reaching the bottom of the pool. The final conclusions are, that "a porous bed of particles will be formed on the bottom of the pool ..., pool water will be able to penetrate this bed and ensure its cooling". We have some reservations on the general conclusions. As peinted out before, melt release with high as well as with low flow rates seems to be possible. A low flow rate will make the initial ( formation of a debris bed at the bottom of the pool probable. be
. In case of a high f3cw rate an overall fragmentation of the
( melt becomes more questionable. Fauske and Ass. /8/ : -] suppose that the metal grate below the concrete floor , ,] (to disperse the water surface during pool swell) will act like a r
. particle former and disperser, when melt f alls down through i i the drain pipe and through the grate. The formation of particles of 1 cm in diameter is expected by that mechanism. ;
t
)
l mm
T In our opinion this kind of fragmentation is not the essential mechanism, because the grate may f ail. .Furthermore the greater part of molten material may eventually flow through ,. the steel door directl'y into the pool 73/. Even if there will be fragmentation of the melt by any , . mechanism and may be coolable as single fragments, the achievment of the coolability when the particles settle dcwn and form a debris bed can not be ensured. The size of the fragments as well as the height of the debris bed and its power density may vary locally. Coolant must reach every particle f" ' of the debris bed, otherwise dryout will occur. A lot of theoretical and experimental work concerning the dryout heat flux has been done during the last recent years or is underway. In general the experiments are of small scale and performed for idealized conditions, mostly , i for uniform particle size. The few experiments which have been performed for debris beds consisting of particles or ,
't different size, show a strong influence of the particle distribution on dryout heat flux (reference /3.6-6/ in /1/) .
('. Because of the low thermal conductivity of dry particles only a small dry zone is needed to reach temperatures which are :
- sufficient for remelting the particles. Thus, even if there j v' is a coolant covering the particle bed as expected in the [
condensation pool, melting propagation of the debris bed may - 1 1 occur. In that case the melt will be separated by a crust zone - from the coolant while at the bottom side, the melt may attack I the concrete floor of the pool (core melt-concrete interaction). An attack may already happen by hot debris without remelting /14/. I
\
l l
~ 3
... '. ' 1.
- o i h
The onsetting of a core melt-concrete interaction has j consequences on containment loads (as already shown in / 3 / but not discussed in the safety report /1/). l
- Without interaction only saturated steam will be released into the containment during long term phase of the accident With interaction additionally the release of noncondensible
[ gases (.H 2, CO 2, CO) has to be taken into accounc. p ..
.f' - E i
8 Therefore it'is to be expected that the composition of contain ent atmosphere may differ at the time when the rupture disc will burst in the cases with and without concrete erosion: I large fraction of steam, small fractions of nitrogen and - hydrogen, in case of no concrete erosion smaller fraction of steam and nitrt. *n, largerfraction of non condensible gases, especially hydrogen in case i of concrete erosion. l N.* The results in / 3 / concerning containment pressure history i show even like calculations we have performed in the frame ( of different studies to containment pressure rise during I. core meltdown accidents, that a smaller release of vapour and a l higher release of hydrogen'will reduce the pressure increase. p That means that the rupture disc will burst at a later time than expected in the safety report /1 / in case of concrete erosion. m
.__ _M_ _ _ _ _ _ _ _ _ _ _ _ _ _ . . _ _ _ - . _
~ l
, .i
! )
1
~
I::portant structures of the containment could be attacked bv concrete erosion. An erosion or penetration of the side walls of the condensation pool seems to be less probable because of the ' i great distance between the walls and the expected location of a melt release into the' pool (below steel door and drain pipe) . Estimation by a confirmatory study, whether the core melt attack / t-concrete erosolon may threaten the pillars, which support the l , main structures of the containment should be performed. Nevertheless, mass 2y concrete erosion has to be considered as a long term prccess f E i-
Conclusions:
J"% " 1 t- - Concrete erosion leading to a higher release of hydrogen l in connection with a smaller relace of steam than expected i. does not influence the pressure increase and the time for ; venting in a negativ sense in comparison to the results : estimated in the FILTRA project; for long term accident management the possibility of the release of noncondensible gases (H2, CO2, CO) as a consequence of concrete erosion, should be considere': J with respect to the effectiveness of FILTRA, the different [ compositon of containment atmosphere at the moment, {} when the rupture disc bursts should be taken into account (see fig. 3.1-1 /1/ concerning the effect of steam content y on particle penetration); i (.. l-r. p for long term accident management the possibility of an }- attack of the concrete 'or of other structures like the !
, pillars by core material should be considered. l e
f e 9 emuussup a _ - _ . _ . _ . . _ _ _ _ _ _______a
< ~
24 - s Contain10.ent Atmosobere Analysis' Calculations concerning containment pressure and temperature histories are performed mainly for the loss-of-AC-power case > l (TB), documented in /1, 4/. No figures are given for containment atmosphere behaviour in the LOCA-case (ADE) with core melticwn. For seperate effects (steam spike, vent capacity) results are shown in /8/. Containment atmoshere analyses are " intended to have the character ; of best estimate" /4/. On the other hand it is pointed ou:
,r that "due to uncertainties in the calculated models and possible variations in the assumptions regarding accident initiation and accident progression one must ... consider the possibility that the real histories (temperature, pressure) may differ significantly from the results presented" /4/. Results of /3/ concerning containment atmosphere analysis are not taken into account recognizably in the safety report /1/.
If data would have been estimated for the band of uncertainties's of the results they should be published. Y i M4*
25 - o
.f*
IV Modelline of Source Term and Fission Product Behaviour in Containment before and after Break of Ruoture Disc etc. Source term ' The estimation of fission product and inactive material release takes into account recent experiments and theoretical considerations and is believed to describe properly the aerosol source term. As the behaviour of particles in the containment is strongly dependent on the amount of released masses, the
,3 choice of a three times lower mass release is suitable to get an impression of the uncertainties associated with this release.
Aerosolbehaviour in the containment . The aerosol codes EAARM 3 and NAUA which have been used in the FILTRA projcct have been identified by a group of e:<perts as the b'est suitable codes available at the moment for LWRs. NAUA has in addition included a model to calculate particle growth by condensation. Howevor, up to now the thermodynamic codes are not able to give the amount of supersaturated steam for very short time intervalls needed as input to NAUA. That is why it is not possible at the moment to calculate condensation on particles and the differences between the two codes, NAUA and HAARM J
~
p are not that big. u . The input data for the HAARM 3 calculations have been checked. There are no objections. Moreover the influence of the different models and parameters for the retention of fission products in the containment are thought to be of minor importance as regard to the high retention efficiency of the crushed rock bed filter especially in the case of avery early failure of the rupture - disc which represents the most severe load to the FILTRA-cystem. 4 4W
26 -
- t f
V Desien and Performance Parameters of the Vent-Filter System
~
i such as Ruoture Disc Pressure, Vent Capacity, Condensation and Filtering Efficiency ^ p b 92EEEal_Bggaghg The FILTRA system is designed to reduce the consequences to the public in the case of an accident leading to an overpressure b, failure of the containment. The system shall cope with severe ; accidents with core degradation without any active action for about 24 hours. i Comments concerning the Design of the Vent-Filter System VSEE$Ug_gagag[gy The opening pressure of the rupture disc is o.65 MPa, ' its opening diameter 6co mm. This together with the ' resistances of the piping forms an effective diameter of 365 mm or o,1 m 2 . The venting capacity is well large enough to cope with longterm (_ steam production by decay heat and there is some additional (m, capacity to maintain unlikely events like leakages between drywell and wetwell in the case of LOCAs or like some steam
'- escaping the pool during the cooling of the melt in the pool. The deg -
of additional capability for such events depends very much - on the accident sequences. Some calculations were made with COPTn., but no detailed information were given in /1/ and /17/ so thz the results could be checked. The results indicate, that leakages between-drywell and wetwell could reach geometric areas of o.3 m 2 to more. ' The probability for even larger leakages (open doors, several penetrations) is not smaller than that of the mentioned areas. High effort should be put on all openings and penetrations ro mumre a tight separation between drywell and wetwell as soca as the reactor is under pressure. Possible measures are suction of twenseals, position indicators, keys etc.
<- l i=
l 1 . . l
'~
C_o_ n_d_e_n_ _s a_t_ io_n_ _' C_a_n_ a_c_ i t_y_ _o_ _f _ F_ I_L_T_ _RA_ L The FILTRA - building and - function is desbribed in /1/ and , according to the performed experiments (the results of which are yet to be published),., the desired condensation effect is f likely to be obtained. The capacity of FILTRA to condense steam is sufficient /1/
~
- for the steam procution of the first 24 h. This assumes an ,
uniformly distributed heating of all crushed rocks. But the experiments were conducted with extreme 1-dimensional boundary
,. conditions (o,8 m2 x 30 m) where the steam is more or less y forced to meet all crushed rock in the test filter. The real plant has the dimensions 315 m 2 x 35 m and is originally filled -
with nitrogen. Whether the steam front is really flat in this i large area, is a subject to be questioned and no experiments are known for this kind of condensation-processes. If the steam front is not as flat as assumed, or if parts of the whole volume
~
stay filled with nitrogen, there might be the possibility that some steam may escape from the FILTRA - building before 24 h. We recommend to estimate the possible degree of ununiformity of the steam front by simple analyses.
. I
[ There is no knowledge of the longtime behaviour of such a kind {_, of filters. The FILTRA building should be inspected regularly to
- examine possible changes of the crushed rock bed. When the icvel +
f of the crushed rock bed has changed due to settlement, then the : w bed should be regularly refilled up to the ceiling to avoid empry spaces (hydrogen burning). Organic growth in the rock bed might, influence the filter efficicncy. Adequate measures to handle i i that possible problem should be taken into account. l 1 l l i 4 _.__._________Q
, . l l
EBOgyaygg_Cggagity_1g_[1LT33 , The capacity of the sump in the FILTPA building of Soo m' was checked by GRS as keing well sufficient to take the steam r equivalent of the first 24 h (ca. 27c MWh), without considering. the capacity of the pool, which additionally is sufficient for at least the first 7 h (ca. 12o MWh). If the AC-power returns after 24 h, the water can be pumped , to tanks in the auxiliary building. Thus no problem is seen , in the capacity of the FILTRA sump.
,C 1
O Y e 1 e a s e (- ' o enuman.
I 1
----------.s.c Ruoture Di and Valves No detailed information on the construction of the disc and the valves was available up to now.
Our reccamendations are as follows:
- 1. The disc should not be able to damage the following pipes and valves to put them out of operation or make them untight.
- 2. There should be a specified space in th,e piping system to catch the disc, so that it cannot plug the pipes.
-3 ?
- 3. The pressure wave after opening the disc propagates through the piping system. It should be asured that no damage td the system due to pipe whip is possible.
- 4. The valve of the not affected plant should be closed before the rupture disc of the plant were the accident took place, opens (in order to prevent the dice from the pressurewave, missiles etc.).
- 5. The bypass-valves, connected to the drywcll (two in series),
[ should only lead to the FILTRA-system. They should be operable at any time immediately after the beginning of the accident. The minimum calculated cross section of these valves will at least amount to o,o2 m 2 , m s ap
s,
- 3o - , ,
i + l Ccmments Concernina the Filterfhi'Efficiencv of the Vent-Filter System 9E9EEal_gggagks I F A reduction of the consequences is, however, limited by the 1 following aspects: noble gases and methyl iodide are not retained (the contribution of the noble gases to late fatalities can ! be 1 %) (( leckage of the containment and the FILTRA system time of containment failure The influence of the last two aspects.can be understood in the following exampie: In the case of a transient followed by total loss of AC power ; tha rupture disc bursts after 13 hours. Assuming a leakage of the i containment of about 1 %/d (design leckage of the german BWR) F one can calculate a penetration factor during the 13 hours of about 7 . lo- . This has to be compared to the fraction of the ' initial
-3 released mass which is still airborne at that time n (lo till 7 . 1o-5). This example shows that the leaked mass 4
has nearly reached the value of the airborne mass, so that a (- filtration of the aerosols will only slightly reduce the consequences, e (- [ Igdigg_ggggggigg , The critical point in the estimation of iodine retention in the crushed rock bed is the chemical form of the released iodine. In WASH-1400 and in the german risk study ohe has assumed that ' nearly all of the iodine is in elementary form. This is justified because there was no experimental verification at that time and l
31 - ! t e a 1 one had to adopt the most volatile chemical form of iodine. In the meantime, however, there are strong' indications that the released iodine will be present in the form of iodine, which behaves like an aerosol particle and will be retained very l effectively in the crushed rock bed filter. Elemental iodine on the other hand is very reactive. Due'to the large surface area one can assume that elemental iodine has also a great
. retention factor in the crushed rock bed especially if a .
!n%'
scrubbing pool with sodium thiosulphate is used. As methyl . iodine is very difficult to retain in the crushed rock bed, this
/" chemical form becomes very important and can finally determine the overall filtration efficiency of iodine. Recent experiments to determine the volatility of HOI give an indication that ,
i organic iodine compounds will be formed. According to /,12/ the EOI in the water reacts very rapidly with organic substances and i forms volatile organic compounds. The question of organic iodine ; species is very important and much work remains to be done in l this area of research. 8 IDfl92DSE 9f_DY$E9EED_gggygsgigg_gg_ggggggigg It is stated, that when cooling in the reactor containment is resumed, air can be sucked back through the FILTRA plan into th (J containment. In this case hydrogen formed during the accident _ can react with the oxygen in the air in a more or less explosive [, way. This event influences the retention capability of the fil:ar j. system in different ways. [ t
~
Aerosols deposited on the crushed rocks during depressurization 'e can be resuspended by the increased turbulent flow and thus f released to the atmosphere. If one has to assume a shock wave i passing through the FILTRA plant, this would be an even more effective way to resuspend aerosols from the rock surfaces. I . l fa
i e i Iodine adsorbed at the rock surfaces might be desorbed due to , the temperature increase during a hydrogen combustion as adsorption is rather sensitive to temperature changes. The same might be true'for the effect of chemisorption. : Depending on the type of the chemical compund formed between t iodine and the rock material, the rising temperature can lead , to chemical decay and following release to the atmosphere. It is - besides others - for that reason, that similar design concepts studied for American BWRs and PWRs include hydrogen c. c- ,- igniters. - As nothing comparable is provided for the FILTRA plant one has to demonstrate the effect of hydrogen combustion on the filter
- efficiency. This must include the check that design targers are still fulfilled.
i l l E5EEagglatigg_gf_gggggigggial_;gsults The experimental results have been obtained mainly on a test ? rig with a 7.5 m high and o.5 m diameter gravel column. These
/' results form the basis for the verification of codes, while t2 . these codes have been used to calculste the retention facters for the designed filter system.
The first question that arises in the context is: have the ccndi-tions during an accident bden modelled in the experiments in an 4 adequate wap !? Considering the experiments aerosol size and !. density, flow velocity, steam contents, concentration and parricle materials have been investigated as influencing paramters. For aerosol retention these variables may cover 'the actual conditiens l in a sufficient way. For iodine retention, parameters like 1 i r , l -- L ~. l
r---------"---------- , 33 - ;, i
- . I o ,
t radiation and poisoning by chemicals seem to have been neglected. These may be rather important for the formation of volatile
. iodine species in the later period of the accident. Another ,
problem is, that iodine adsorption strongly depends on the j sorption material involved. In case one depends on the experiments made, this means that the same material has to be used in the designed FILTRA plants and the same sorption behaviour ; has to be' checked. There is no comment, how this shall be assured. t-The second question relating to the extrapolation of experimental results concerns the influence of the geometric carameters. That g< means the scaling up of experimental results for the design has to respect the change of the retention behaviour due to changes , in the dimensions of the filter. For the aerosols the results of ' the calculations indicate, that this,has been observed, although ; the rise in retention factors for deeper beds in some cases is ' i hard to explain. But this is only of minor importance for the results. For iodine no che'ck could be made, as the programs used I are not known and no detailled results of experiments which are still in progress as well as of calculations have been presented. It was only stated that elemental iodine is completely retained in the bed. Aerosol retention
/7 The theoretical calculations of the filter efficiency of the I L.
crushed rock bed for aerosols were done with a modified versica of the HAARM 3-code. The basic assumptions in this code are: E homogeneous distribution of particles in each compartment I log normal particle size distribution in each compartment . impation is neglected. In a crushed rock bed, however, there is a concentration gradient and also a changing of the particle size distribution along the bed due to the different deposition processes. In order to account a f 0
. .- - - 34
'I
~ r o
~~ i for this situation the crushed rock bed was divided into '
different sections, which give a better approximation. This approximation is believed to be satisfactory because the l assumption of a homogeneous distribution of particles in a compartment gives conservative values compared to a concentration - I gradient. In the case of a concentration gradient the depletion of particles at the front of the bed is much higher so that , finally more particles will be retained than in the case of a s hemogenious concentration. r Impaction is important at high flow velocitics and for large
. particles. It could play a rdle at the time of the failure
~
of the rupture diaphragm and in the first segments of the crushed rock bed. The calculation with HAARM 3 without impaction will overestimate the penetration and thus will be conservative. The calculated acrosol decontamination factor lies in the range of 1o - lo . Compared to operating experience with sand bed filcar in nuclear installation this range seems somewhat unrealistic.
, This holds although residence times in operating filter units usually are shorter th'an in the FILTRA design. It can be seen, that from some bed depth upward no marked increase in retencion t
,{
( ', can be attained due to leackage and resuspension effects. As .
. those effects have not been considered in the codes used for the design of the FILTRA plant the results overestimate the additional effects of the higher column compared to the experiments as was already mentioned before. Without extended '
analysis it is not possible at the moment to quantify the effect
~.
of bed depth. It is therefore recommended to study in detail the influence of leakage and resuspension on the filter efficiency. _ _ _ _ _ _ _ _ _ - _ _ _ - _ - _ - - - - = - - - - -
l'
~ '
j. o i
' ' ~
' Conclusions to chacter V The design and performance of this passive system meet the '
necessary requirements for the assumed reference cases. There is also the possibility to take this system into operation by g bypassvalves parallel to the rupture disc, if accident sequences , should make this necessary. As it cannot"ce excluded, that hydrogen and air mix inside the building, the building and piping systems are designed against such loads. The check of i the adjacent experimental results was not possible by GRS. L ,
,r . The main recommendations are in short:
1 the tightness between.drywell and wetwell during b enstre normal operation (e . g. door to wetwell, vacuum breakers)
- the rupture disc should not damage any valves er pipes, and should not plug any pipes the piping system should be designed with respect to pips whip problems i the bypass-valves should be operable at any time from
, immediately after the beginning of the accident experiments should demonstrate the equally distributed
, condensation of steam in the large cross section of the '
m FILTRA building hh - regular inspections of the crushed rock bed. ( Three critical points have been identified, which could effee the filter efficiency of the designed crushed rock bed: i The amount of organic compounds of iodine, which will be formed in the course of the accident, is of fundamental importance to the iodine filter ef ficiency. The dif ferent ' formation processes for organic iodine species should be ' studied carefully. G SO
,$ < e l
*. l, >:
n.
* -j l
a a { Y.
- In the case of hydrogen combus'tilon deposited aerosols ;_
could be resuspended and adsorbed iodine might be [ i desorbed. It should be checked that the designed filter efficiencies are still fulfilled after a hydrogen combustion. [
- A higher bed depth in the FILTRA-plant compared to [
c the experiments could led to an overestimation of the retention !' capability of the crushed rock bed. The influence of this 4. effect will be relevant in the cases of very low calculated [ penetration factors. l [ b 6 i
.f',
l
<- r l
h j S k d' l 1 W ,@
,,1.,- .
4
' I' VI. Comments on Accident Management and Instrumentation
i In case of an accident a prepared accident management procedure should start immediate precautions .. to undertake the necessary measurements, to analyse the systems situation and behaviour, to find the. proper diagnosis and .l to initiate operational requirements. For the diagnosis of core melt accident sequence information - on following points are necessary. er i I %. pressure in drywell and wetwell ' temperatures in drywell and wetwell air space temperatures in different elevations in the wetwell pool waterlevel in the wetwell : pressure and temperatures in the RPV top and bottom '
'with wide range instrumentation j
composition of the atmosphere in drywell and wetwell, taken I by sample probes inspection of the drywell and wetwell by optical equipmen-pressure and temperatures in the FILTRA-building
,o composition of the atmosphere in the FILTRA-building f waterlevel in the FILTRA-building i
~ -
, Knowledge about temperatures in the core area of the RPV k- [
during the accident sequence could be useful. t-For the longterm accident management it becomes additionally useful to preplan the - l . } . accessibility of components (e . g. for cooling, venting) and instrumentation under severe accident conditions repairability of instrumentation and components relevant ! to safety aspects. ; Operator training for accident management, analysis and repair of components should be forseen. b
38 - ; r The cabling and instrumentation of the electrical monitoring systems need special efforts for the sure~t'y of the transmitted informatica (see several years of development for the Marviken- and HDR-test- , instrumentation). In addition to these general remarks we like to make recommendations for some special aspects: Prepare a procedure to ensure measurements of radioaccite ,~
, material in the surroundings of the plant in order to make decisions whether people have to be evacuated. The authority who takes the decisions and how the necessary information are j- supplied to him should be fixed before hand. +
R Guarantee the safe opening of SRVs during accidents without core cooling systems operating (e . g. TB) in order'to prevent high pressure in the RPV, in Iase the core starts to melt. I Consider accident management actions for the case of severe basemat or pillar attack in the wetwell. ' The vent valve of the not affected second plant.should be clcsed as soon as possible after an accident was detected. This shou ~d_ prevent the rupture disc of the second plant from possible effects !
- s. of the burst of the disc in the affected plant.
h1 Consider actions in case of open penetrations or of leakages grea: (*s. than 1 %/d from the containment during an accident that the > bypass valves of the rupture disc can be opened immediately in , order.to depressurize the containment via the FILTRA-system. 1 F.- - If AC-power returns, the cooling system should be started so that possibly no air is sucked back into the containment. If possible, refill the containment with nitrogen as steam is condensing.
~ / -
I e
$'inal Conclusions i f
While evaluating the safety report on the FILTRA-plant, we , identified a series of problems which in our opinion can be solved. - i Summarizing our general remarks and conclusions gi/en in the preceeding chapters we finally conclude the following:
~
Accident analyses for the reference cases have been performed in general in an adequate way. In our opinion !- the analyses cover sufficiently the spectrum of important and
- e. expected sequences for the design of the FILTRA-plant.
t The possibility of an accident scenario due to a break in the piping system to the scram discharge volume tank leading to bypass FILTRA should be examined. In case of its existense additional investigations are required. S FILTRA contributes to a reduction of public risk yet to be quantified, especially in case of accidents leading to an early failure of the containment, due to overpressurization as a consequence of steam and gas generation. Alsv in cases
,, where late containment failure due to overpressurization is anticipated (in case of long term steam and gas generation) ky a reduction of risks by FILTRA is to be expected. l i I No essential disadvantages to the power plant caused by i FILTRA have been identified by GRS. Only during a long term phase of an accident, pro,blems may be encountered regarding containment integrity and fission product retention in FILTRA in cases when air is sucked back through the filter into the containment and as a consequence H 2 -burning becomes a possibility. !
w 0
i 1 e d' ~
- Ao -
i c .
+
t The possible scope of the formation of irregular condensation , l profiles in the crushed rock bed of FILTRA should be assessed by simple theoretical calculations, considering local density differences and radial flow behaviour. i i m . i The effectiveness of the FILTRA-plant should be ensured i by the possibility of repeated inspections. As problems might occur especially the possible changes of local density in the crushed rock bed as well as the possible growth s of organic materials. , The implications on the static of the concrete internals by
,.e
> , an attack on some pillars due to melt concrete interaction x~
should be examined. Furthermore for long term management the possibility of a local basemat attack should be considered. In general, reduction of risk by installing FILTRA may be limited due to the following reasons: ! noble gases'and methyl iodide are not retained, leakage out of containment may bypass FILTRA (via the reactor build _ng or turbine hall), , C possible H2 -burning in cases of accidents during start up and shutdown of the plant (non-inerted containment situation) might cause a bypass of FILTRA due to loss of containment { j integrity. , Due to the non inerted containment situation we see some possibilitic g to reduce risk by for example early inertisation (before 65 % power-level is reached) ) running the emergency diesel during start up to ensure pcwer supply to relevant safety systems.
o.1
- I
,e !- -
. " + ,
i REFERENCES'
/1/ FILTRA-Safety Analysis - DRAFT i i 1982-o8-o4 -
. ,. i F l 1
/2/ FILTRA-Design Considerations for Implement 1ng a '
, Vent-Filter System at the Barseb5ck Nuclear Power Plant K. Johansson, L. Nilsson and A. Persson i.
H International Meeting on Thermal Nuclear Reactor , Safety, August 29 - September 2.1982 - Chicago
/3/ Kernschmel: analyse fur den 17o0 MW h -Siedewasser- t reaktor BarsebEck KWU, KfK - Report, Mai 1982
/47 FILTRA BarsebHck Containment Pressure and Temperature Histories During an Accident Initiated by Loss of Power r
Studsvik Technical Report: NR-82/94 Roland Blomquist, 82-02-12
-7
/5/ Core Melt Analycis up to Vessel Failure for a
{g BWR after Loss of AC-Power '
. Studsvik / NR - 82/14o i
,f - -
.G. Vicider and F. Carlson 1982 - o6 - 3o ,
/6/ Assessment of Accident Progression / Regression for Hypothetical Core Damage Scenarios for the BarsebHek Nuclear Power Station ,
Fauske and Associates, Inc. FAI/82-4 January, 1982 l eunusse __aa_---._ _ _ _ _ _ _ .
V w , s (' t'
- l-i l .
4
/7/ An Assessment of the Vent Area Requirements for the Barsebsck Nuclear Power Station Royal Institute of Technology ,
Kurt M. Becker, Stockholm, Dec. 23, 1981 (KTH - NEL - 31 ) ' *
/8/ Vent Area Requirements for a Rapid Quench.
of Core Debris I a 4 Fauske and Associates, Inc. FAI-KB/81~-1 R.E. Henry.and H.K. Fauske J-E r
, /9/ . FILTRA - Transport and Filtering of Aerosols after a BWR - Core Meltdown due to Major LOCA, Failing '
Pressure Suppression System and Failing Emergency Core i Cooling Studsvik/NR - 82/96 Hans Huggblom 1982-o2-12
/1o/ Green, S.R.
- Application of MARCH to BWR P
Severe Accident Analysis presented at l o "h Water Reactor Safety Research Information Meeting { October 15, 1982, Gaithersburg, Maryland, USA (1's /11/ Boiling Heat Transfer from Core Debris and i condensation of Steam in the Wet-Well of the Barsebsck Reactor F. Mayinger 81-12-o2 . .) 5- 1 I i 1
/12/
~
D. Torgerson, private communication, 2o.1o.1982 i 3 F l I } l
/13/ I Private communication, information meeting,
[ j GRS-Cologne, 27/28.1o.1982 ' i O m-m__________________ _ _
l 7 l
, .. y
/14/ W.W.'Tarbell, D.R. Bradley~ '
Attack of Fragmented Core Debris on Concrete with Presence of Water Sandia Nat. Lab., New Mexico presented at l ot ,h Water Reactor Safety Research Information Meeting October 15, 1982, Gaithersburg, Maryland, USA , ( /15/ Tests of a Full Scale 1/48 Segment of the Humboldt Bay Pressure Suppression Containment '
- GEAP 3596. 196o '
/16/ R.M. Harrington, S.A. Hodge Scram Discharge Volume Break Studies Part 1: Accident Sequence Analysis R.P.Wichner, C.F. Weber et al:
Part 2: Fission Product Transport Analyses presented at 1o th Water Reactor Safety Research Information Meeting October 15, 1982, Gaithersbrug, Maryland, USA Filtra - Tryck; reaktor inneslutning vid (
/17/
r6rbrott med samtidigt luckage 6ver mellanbjnlklaget
; ASEA-Atom, A. Johansson, 1982 - o3 - 24 *( -
e
.=
~
l mp
~ . _ . . . _
4 1 4 t UNITfD STATES
; [,,. .,, I'.,
NLC E AR REGULATORY COMMISSION
! $ .'i' } REGION l
.j h-
**,*s.,
' U [#[ 631 P ARK AVENUE
*..e* w.r+0 or PRusSI A, PENNSYLVANIA A 19406 v'.
'l
..g
, . y
. y~ :,'
.. ~
, :; _ t ;. ' :+" ;.
l
~ - : >
s e % s g AM f) A2 d? - .
.l .
b l Q,A, *
.l %
J
- 1 1 l l
S f I
+
l
< ,i I
s *$ - 3 9
- I *
- g,* 4 f'
**H.* '*a' ' -\ s' , '
. ,, , <W','
. ;; , . :. v : , ,
e s s
-*:-.- m.. - - . -. -.
* -,.7,.. ., _
7 - _ _ . ,
- s. . < -
g..,- 4 *
, b,
", h + = ,s e e p
- e 4
4 4 e 8
- e
- g
- b 1
Pbve pant cesign , l 1 m C 9 Coping with station blackout c
'~
By Frigyes Reisch An international meeting is being held in London this month to
. discuss emergency power supplies at nuclear power stations. Of particular concern is the reliable operation of these supplies in
. the event of station blackouts. The occurrence of a statxm blackout failure of the emergency a.c. supplies. (that is the loss of all (both emergenc) During the well publicized grid failure and normal) off-site and on-sate a.c. in Sweden in December 1983, all 24 power sources) is a low probabihty diesels at the three affected nuclear event. The likelihood of a blackout stations operated satisfactorily. occurnng at the same tirne as another Nevertheless, a station blackout is of independent accident is even low er. serious concern because it has a major So far only two station blackouts have impact on the availability of systems for been reported both in the United removing decav heat fr'o m the reactor States - Fort St Vrain in May 1983 and core. There is a risk that all these Susquehanna in July 1964. There were systerm might fait simultaneously. no serious consequences, despite the in its annual safety review for 1983. o Frigscs Rench a muh $kt. the Seedah Sucic,, the IAEA points ou't that if the maxi-mum probability of core melt from
~
Pomer inspectorate. S-lu 52 5 ocLadm. 5. eden o ""T4 NUCLEAR ENGINEERING INTERNATIONAL
t h Fbwer pant cesign , Table 1. Summary of data on totallosses of offshe poser ser US nucker station sites.1948-1943.
- l Humber of sHo-years to the end a 3ecember 1983 = s33.
I g05, armemency of j
- t. Causes of loss of acumirrence Median duretion
{, offshe power Number per sele-year) (hours) I Plant-centred 30 0.056 03 g Grn:t blackout 10 2.019 07
$ 601 S*"'e storm 6 0011 2.6
$ Total 46 C D66 05
$g y s Source Nureg 1032 k \ \,
J 000 N g TatHe 2. Summary of US data on diesei generator retaanery start attempts and f ailures for both 2 " Q" tests and actual demands. No. of units = 4s; No sr roector. years = 194; No. of diesel NNQ generator years = aba. Auto start D1 1C De No. of Faeures per No of auto failures / Duration mours! Cetegory No.of demands tallures amaed start failures demand Fig.1. Loss of off site power frequency , Test 13 665 253 IPS 55 0 004 duration. Loss of offsne power 100 5 105 3 0 03 *' AM emergency comands $39 14 1:25 5 0.009
"' O# Source:Nureg 1032 t' 30- M 196182 g that the organizers did not think to fhi. -including modification of pro-
)3' invite papers from such countries for cenim and administrative controls to j 20 '
, presentation at the meeting. eman:e system rehabihty and avail-
{ 15 Earlier this year the International. ablitC. It also calls for the implemen-1 10- Atomic Energy Agency issued a report 2 tatoc of an emergency a.c. power [ on the safety aspects of station black. 5 a ; svsw a reliabihty programme to main-
- o. m L.i outs which contains reports of desiFn tan a high level of performance U9h} {j{ and operating experience in several thuurnout the plant operating lifetime, pj{g( o L countnes, includmg descriptions of station blackouts that have occurred, ant the training of personnel to mn tror recovery time and in the use
% and also a long hst of conclusions and o' those plant systems that would recommendations. In particular the s enac available during a station Fig. 2. Contribution of djesel generator sute , report recommends that the design and backout.
systems to failure rate. operation of the normal and emergency station blackout were to be bmited to
. Power systems should be reviewed kqukanons. In the United States exist-4 and where practicar weaknesses recti- iry r;pulations do not require explicitly 10 extensne bactfa-tin or ould10" be reautre Ycar'd. tbt ziuelear stations be capable of on and 6sne Fig. 3. As example of a swa with core cooling wrtistandmg a station blackout. but the power supphes fail. the reactor tnps and systems independent fr m the a.c. busbars. Stclear Regulatory Comrnission has the reactor coolant pumps and the snain feedwater pumps cease to operate. .
l
. Alternate on-site systems for renxmng Rehef eve g the heat from the core during a blackout C 5 fil
; can be powered by: -"
- 9 Steam turbines,or 9 Dedicated diesel enerators. F In all light water reactors, the ability X N"^ ~~l to remove decay heat is adequate tor p 1--- oo.--
some short penod of time, but the a abihty to cope with a station blaa.out # I I
, will vary from plant to plant dependmg on the specific (Jesign and capabihe) of j h
x N
~
Nl 7i . b,'"d'[,h the equipment installed, the quahey of x, . := remo. ; the emergency procedures and the [ sw ; abihty of the operators to handle the situation. j ~
'" ~ "b -
The rehabihty of emergenes pcmer supphes is of international concern and h O Suwession poor 3,,%,. the papers hsted for presentation ac the i OECD-CSNI conference in LomJon i J this month (ltrIS Octoberl should wo- @ _ _ _,t ges( vide usefut records of operating ex:ien. - ence'. But since gnd system rehabtbr) is c c,,u,,,, omrmar er menz et 7 v_._ gd hkely to be a particular problem: m 7 l developing enuntries it is dnappouming {
- Ociober 1985 __
a9 j
1-I f her pant cesign i devg .ated station blackout as an "un- nbie a. Poemasse sectors hmmng me abuny to cope wem a stenon bisce e l reso.ec safety issue". An NRC report clso pubhshed earlier this year provides PWR BWR 2/3 SWR 4/$/s e voluable review of tne extensise tech. Acs pump easieakage x x ; nical studies performed by the NRC on Acs w o m keupa-a-this srect met the past four years. It 3,j$,*,*,7l[',"" ( ( ide:16ew the dominant causes of black- DC canery ca paaty & x x x out accidents at nuclear power stations compresseo air tvaive es- :m x x x ir. 5 Und States and sugpts how W 0**"y*M' *'"J"'f"Y,[,,$$,,,,,,,,,,,g
- A *
- the . : q asues could be resobed. oj,",7n'g Weer thn will lead in a nde makmp co c oircontac3 x y v sec=> aoubtful in the present pohtical Containment x ch . eyurrounding nuclear reculation ;0{"$;' Dl3 in W* l nitCd blatCs. , Mary Wilmng x x The NRC draws part.cular attention wWS room) (Hoct RCiC roon :
to :ne degree of on-site a.c. power % , , , sys:em redundancy and the reliabihty of indaidaal systems. to the abihty to suddenly reduced to about 5 per cert cf restore off-site power m less than eight normal to match only the in-hcuse
, hours. and the capacity and perform- _ 1 /gn ..,,, electricity demand. Since the los of macr of the deca 3 heat removal systems.
q sysum
**' off. site power is often the tesur ce
'It estimates the likely frequency of sinam veltage or frequency fluctuations. 6e .
core-damage from a station blackout as """' hems switchoser to house load operatiot n ' typraD3 - of the order of 107 per % ,, ,, . . , _ , difficult and in practice has usually reen rein .c year. To reach this lesel at says no. te ar,,rary S. ,, q unsuccessful. req. ares an ability to cope with a station oumo ua'5
~~
An ahemative is to use gas turb:tes blaciout of at least Tour hours. and perhaps eight to 16 hours. emergency uisu oc'.er on emergerd W, hoo* cw e
@M on or near the site or local hydre I i electric stations when the normal pnd diesel reliabihties of 0.95 per demand or '--~~-~~~~7~4yl supply is lost. This arrangement ha, bener with relatisely low susceptibilities worked successfully.
to cornmon cause failures. Reactor type was not by itself found to be a dominant A.ns neaci NC ' However, the backbone of te emergency a.c. system in a nuccar fae:or m ' determining the likelihood of 7
- power station is provided by the core damage due to a station blackout. l
, esnmentsel standby diesel generators. There are but emphasis is placed on particular sfo.ome + two to four of these depending on the anects of design such as the ability to test pec O desiF n for each reactor. They are. m n:amtaan the integrity of reactor coolant m, ,n, RC" general, sery reliable. but there t.ne pu:::p seals, on batter) capacity and the me tT been cases of simultaneous fashres.
effect of the operating environment on Again most of the data on diesel gener-
'he performance of instrumentation. Fig. 4. Simplified flow diagram cl a Feench ators faihng to start have been gathered 1 ant iccation is also considered of gicat eum with two steam rartiunes. One, as is in the Umted States and a summam is im;antance. normat, drhes the anaar, seedu sier n.iniem given in Table 2 and Fig. 2.
SurprisinFl y, neither report refers to pumps. The addeniumaa. special, sicam. All the reports of nucicar staion the possibihty of the loss of normal gnd turbine drives a teneramer einich supphes the 8883 inJ'Clion Pussys and battery blackouts refer to light water rescurs.
. supphes through sabotage or in a h P-chargus, Many rwns and swas are provided nth g limited war. Yet confidence that a equipment that can be useful st a Glatmn can be safely shut down in such Most of the failures occurred at the station blackout. Both reactor types.for j circumstances is also necessary. power station site and were caused by example. are dependent on d.c. sup- i either equipment ma!fum: tion or human plies for control and instrumentaur Designing for emergencies. Clearly the error, and the loss of power lasted for purposes. Besides the provision of stifi- l bes way of handhng a station blackout only a short period. The United States, cient battery capacity the training or I is to avoid it, This can be achieved by like most indur.triahzed countries, has a operating staff to handle such an esetr- ;
assurm; a sufficiently high reliability of very stable pubhc gnet system, and tuality is a key factor. the normal and emergency power sup- therefore grid failures are rare end There are also certain imporant plies. Normally a nuclear plant is con- recovery is achieved cuickly. Some- differences between runs and nwm m nec:ed to the grid through its switchyard times it was also po,sible to rectify the way blackouts are dealt with. The by two transmission knes i. Inside the quickly the e*fects of weather mduced classicai rw n design has a steam tur me plant there will be two to four redun- disturbances (c.F. when 4 sirgle piece of tirisen auxiliary feed-water pump Se , dan: datnbution trams depending on equipment was hit r> hpntning). But long as there is water in the stonge the specit'ic design. when storms caused widespread de- tanks there is a d.c. supply to :nc Only in the United States hae the struction. an outage cowlJ last sescral instruments and to the d.cja.c. cor er-cau es and frequency of failures m the hours. ters, an'1 the valve actuators and c:'en-grid suppl 3 53 stem to nuclear power Most nuclear pomer stations hase pressed air is also available to the vas es stamns been >>stematically reported equipment that enaNes the main gener- which need it, then the plant can and anahsed The tailures which occur- ator to be switched ne- to house load withstand a station blackout for seseral red sere dnided by the NRCmio three operation uhen the gre connection n hours, in the Umted States .ae cate;ones: plant centred; gnd blackout; lost . In these cirmn stances pow er " classical" r w a design ofiguratior a3 and u cather. tSee Table 1 and Fig.1.1 output from the reaor has to be not been changed. l 1
.-So NUCLEAR ENGINEERING INTERNATICNA.
I _d
~
/. S- ,
s
. 3 r, !
~
me Q g $.g
, ,- ver qy%.-
g
< t h'$$hY{ : :b Y w l T.
For station blackout purposes, owns Dg. 5. The Swedish State Power Board bus nal steam turbine driven auxihary (HP) have bten dwided into two functionally recently invested in this 2.8MWe anMhr feedwater pump and a steam-turbine. different classes: diesel generator to act as a standby malt for driven h.p. core spray pump this diesel e Those that use an isciation condenser coohng system for de:ay neat remosal Q h',
, ,Ily,Q*[ ",'{'
(four per reactor). See Mt. August 1985, pt.. enables the station to withstand a longer blackout than presiously. and do not have an a.c power- Clearly much has and is being done to (Photo: Nohab Verkstads.) independent make up capability improve the abihty of nuclear stations (BWR4 and -3 designs). to w1thstand a station blackout both 9 Those with a reactor core isolation incorporates two steam turbine-drive: through backfittmg and design innova-coohng system and either a steam- auxikary feedwater pumps, two sicarre tions. But is it enough? Smee the TMI turbine driven high presscre coolant turbine driven primary water charF ng i accident, the NRC points out. most injection system or high pressure core- pumps and a 12 hour battery supply nuclear power stations in the Umted spray system with a desrated diesel with two additional small diesel gener- States hase been required to have at (Fig. 3), a,y of which are adequate to ators for battery charging. These to- least one a.c. power independent decay remote deca) heat from tme core and gether enable the station to cope with a heat removal tram Not surprismgly. control water inventon conditions m station blackout for an extended penov therefore. the NRC found that the the reactor vessel (BWR4, -5 and -6 Mr R. D. Bye of the CEGB has aise initial failure of emergency supplies was designs). drawn t.ttention to the desirabihty ct a "small. but not significant" contri-
, Since owns are deugned as natural making it standard practice to start the butor to core melt frequency. Most of circulation rractors, at leas:t at reduced standby generatcrs on reactor tnp (as the dominant failure modes were power levels. the loss of rearctor coolant well as on grid failure) so that the por associated with a failure to sustain recirculation poses no special con- tnp sequence is less likely tc be inter- decay heat removal durmg a station
. sideration. Moreover, reactrvity control rupted if the grid subsequently fails. blackout and the NRC report points out
. during cooldown is adequately main- In Germarry, Belgium and Swit- that "very limited work has been done
,- tained by control rod msertion, an zerland the approach with ewas has at nuclear plants to look at the capa-(, action which would occur automaucally been to provide additional diesel-driven bility and reliability of systems in con- ;
on loss of all a.c. power (Table 3). feedwater pumps placed in a specia' ditions of sustamed loss of a.c. pow er" build ng situated at a distance from the This gap in our knowledge could best Seeking solutions. Outsade the United mam reactor buildings. According to be filled by a concerted miernational States the design ot' *w as has been the IAEA report. the Soviet rwn desigt research effort. It is an encouragmg sign rnodified in several countre to increaw can withstand a station blackout of that the I AEA asked the noted French the abihty to cope wi1 a station several hours without any additiona' safety expert Mr Pierre Tanguy to gne blackout. equipment. a lecture on station blackouts as its To avoid the possibiaty of a pnmary There are owns in the Umted States general awerably m Septernber. O leak occurring as a result of a 1.sck of that are equipped with steam turtime-seal injection to a tr.am circulation dovcn auxiliary feedwater pumps umr %,,, pump the French hase installed m all lar to those used m the nwns described their ew as a specui steam drn en 3 egg,p,,,,,,,,,,,,,,,,,3,,,,,,,,,, above. But some older siwns with isole- og,,ar,ng enpmence retavos to o+s,re e,ecir,c turbme-generator. Thn a uwd to tion condenser > do not hase t hese power sources 16-18 0ctober 198$ Lonoon i charce the battenes and to supply the auuhars feedwater pumps and these ! IAE A. Satery as;>ects of star on cracaours at motor of the hydrostat,c test pump which injects water to the seals of the reactor $ hase been equipped to dea with station blackouts by the ins t a!-
,'",,]'gl'[8"',s C eccoo332. IAE A.
3 Nac. Dai.,ar,on or starson c;acaovt ace,- pnmary circulation pum;s A plant lation of dedicated diesets mdependen: cents at nuevea, po.cr pian ts %,,eg 1032 equ:pped m tnis was can withstand a f rom the a e. buses. %c'ea' Regwawy commiss.on wasnmoton station blackout for 'much longer than the ongmal design tFi; 4 e in Sweden a ow n of more than ter nj,ajs an a p, e5
, p, years old was recently equipped witt m,nt, m po.,, , gyp,, n,,c, %c,,,, gngm.
In Bntam the Sucwed B design such a dedicated diescl. With the orip cermg mrernarvonauanuary 1980
~~ October 1985 -
51
q
,sh e.en _ n a n' y., -
lT, & ; A Y 1
" #a g%.=gMgg_ v . ..s .. ...
i f,"s33 -'
\
's
?a cg__ ._..~s m W ex v .
. u.
a; t."" _ 2 l m...,,y- I
. ,hMtisAdhimhpy
_ e$$hh-y h l $ g":;,
. . ... = .-
, ,1 .I }h[g --N));M h
e e emmum
. . 7a
** Ateeds lad.strial Forwm, his. l
'* j
< 7101 Waconem Avenue gemenen. 9@ 20014-4891 l
- *
- Telephens 30t)6s4 9260 !
j TWK 7104249602 ATO8AC FOR DC
.J
__ F- . m ~. .,. .-. .., h w January 9, 1987 To: Members, Steering Group, Committee on Reactor Licensing and Safety
Subject:
IDCOR Re-evaluation of Technical Bases for Filtered-Venting Enclosed for your information is a copy of the sunnary report of the subject effort. This reassessment was undertaken . following the accident at Chernobyl and the announcements that various European regulatory agencies were moving to embrace filtered venting. The summary reviews past IDCOR work on this issue in light of new information and assess ~es the more recent European concepts. The enclosed paper was provided to Harold Denton in December 1986 by IDCOR. Sincerely, RWH:wdp - Enclosure ec: Frederick Buckman Cordell Reed
- ~ . . . . . . , , ,
-/Y, me
a SIAttARY Fli. TIRED VENTED C0tCA1NMENT EF7ECTIVDIESS ASSESSMENT l The Chernobyl accident and regulatory and political actions in Europe have agais focused international attention on the essirability of filtered vented containments. Although it is clear that a iiltered vent would have had no
. impact on the offacts of the Chernobyl accident, the Germans have recentir followed hamn and France in adopting filtered went requirements f or their plants._ Other foreign countries are re-examining their postuon on tne need
~ana ef festiveness of filtered vents. j The United States nuclear industry severe accident progree - IDCOR - had conducted evaluations of several iiltered vented containment concepts several years before the Chernobyl accident. The IDCOR technical evaluations presented to the NRC in 1984 showed that costs of filtered vents vers not offset by corresponding risk reduction benefits.
Only the very lew probability accident sequences showed Overall, IDCOR substantial concludedsource that tore reduction from using iiltered vents.Further, IDCOR found that for the filtered vents are not cost effective. limited cases where vents were effective (notably the MARK I containment overpressure failure with later over-temperature failures) a venting mechanlas'is already available in existing designs and that capability is just as effective as a separate filtered vent. IDCtX has reassessed its previous technical conclusions concerning filtered vents based on new information developed during the past two years including foreign advances and coveitsents. As part of this work, IDCOR re-evaluated
' the effects of filtered venting f or all important accident sequences for U.S. l reactor designs including very low probability events.
IDQR concludes: l
- 1. Filtered vents are not effective for the category of accident sequences i involving containment by-pass.- Ttsse types of sequences may and of ten do dominate severe accident risk for existing domestic plant designs. l
- 2. Filtered vents do not appreciably reduce the source tems for PWR containment overpressure f silure accident sequeness (the most likely PWR failure mode).
- 3. Filtered vents do not improve the source terms released for the category of accident sequences which are already scrubbed in a suppression pool for DWR. Even the MARK I sequences with containment overpressure f ailure and later temperature f ailures are pool scrubbed until the later failure occurs.
- 4. The costs of a filtered vant cannot be justified on the basis of overall risk reduction.
j
I 4
- 5. Filtered venting cannot be justified for land contamination reduction since installation costs would ascoed averted costs.
- 6. The French filtered went containment system is not effective for most circumstances due to its small sise and low officioney. !
This paper provides the supporting information which land to th=== conclusions of our re-evaluation. The paper is arranged as follows: o Review of the concepts studied by IDCOR ' o Summary of the cost benefit basis of IDCOR's position o Risk reduction potential ' o Cost evaluation based on CRAC-2 analysis o Vent effectiveness by sequence , o Evaluation of filtered vents planned or implemented by foreign countries. e e W b
im addition to these studies TVA embarked upon conceptual designs of various is.Itered vented containments during the Sequoyah licensing process in 1980. These designs were prepared for order of magnitude cost estimates and provide ' a reasonable basis for estimating costs in todays dollars. The criteria applied in these designs are nearly identical with that recently chosen to asialyse the Mtc (Bernero) SWR containment performance issues. Therefore , these estimates should be comparable with the BWR venting concept. MR Task 9.1 - Core Damaae Prevention This task looked at a variety of equipment that could be used to reduce the likelihood of core damage. Containment venting is one me.ans that could be emed to mitigate certain sequences and thus avert economic loss and public risk. Containment venting in combination with a makeup water source are capable of preventing containment f ailure in loss of containment heat removal cases. The reduction in core damage due to venting / water makeup was determined to be small. No other contribution to core damage was identified.^ I:DCOR Task 19.1 , Task 19 evaluated two concepts that goalify as a form o'f containment wanting. Concept 1 is a filtered vent through a suppression pool (see Figure ! 13 Concept 2 is a compartment vent in which the challenged containment is relieved into another large volume. No radioactivity would escape in this concept if contairusent isilure is averted. (See Figure 2) The following functional requirements were considered for evaluating the c:encepts . o Reduce probability of containment overpressure . o Reauce radioactive material releases
~
o Maximise the evacuation time available to the public o Minimise adverse effects on dMisign basis. l Dose reductions considered adequate in Task 19.1 would be to reduce WASH-140 release categorias 2 and 3 to below category 5. A factor of 100 in todine ! and particulate would bring about the desired reduction. Sandia studies performed around 1981 found that additional ef ficiency would have very little additional impact on risk. It is interesting to note that the barseback filter has a DF ef about 1000. The Task 19.1 study did not consider whether land contamination or loss of land was significant. This study found that a venting system utilizing the BWR suppression pool l would be 25 to 30 times more cost effective than a filtered vent containmer.t. I however, even with the low cost of the non-filtered vent the cost benefit analysis conducted by Sandia found this vent to be marginally beneficial for BWR. The filtered went was definitely not cost beneficial ior PWRs.
*
- e ,,_
n _ e.'p,..._m_._.y
-l,-
- e - L :..- . _.. _.
,- "f
_ s.
"~
= ,
y . . =
=
nc-
. = .
=
- 5. , _ -e_. .
- r c .
P._ s e .
.E s
=. ... m. ;, .
G-
. = ..
p,
..e=._,__G...._.,
- .~.
..I mTN.___ . "..*""" N
. - ^ .
..e.-
=
=1 R -
E E .- V . "'s.. _-===- ~ T ., A nT. _-
.-.=.m,C
=
m W _
=
. . C.-
_ _. I,_ , - 6. _~.
. a y
. R . ._*_,e, -.."t.
, c t g ., _~H -n. .. E- ".
..P.,% ,., . ,
9, w R g. . f=..n,_,_.Sy..,.. P 3 P_ _ _ .
....... O *.. .
r..- .
.. S .. =._ M . p .~"
T.** .
.=..
A . . . _~- :_. .
- J-= _ . D . ,*.".. -- - .%
- " =.=E e *. _ =
T ,, -
.m - ..
.R E . e f"_._*,-
- m. . .- .*_
...T
._ L
.g
._~_
I i -
- m. F. _ -w ._.
_...E.._." A A
- 1. - .
. 5:-. .. ..e.
- A A a A. O.
- V T._~.
I N.__ - e;. U _~ - ..
. 5 =. . L.N ...*.
, w .;. . . . . i . . . -
. . . ..== :. .
i ,
.,.~>.f_ ..
;;; .._ _.. . . . -.._. M
.. . - a
- , _= .
.- - 5,
.: - . .- .~". .
m _. 3 .. m
e~.2C___.,_ n_kd.h. _w
,n T -.
~
=
._T I
f.N'
- C~V r .
U e. I_ r ue . E- _ n ~ ,_. _ . : ~~.~ i...~_- i -
- ,~ , -
c
-- = ,._.'s .==sP-
.r G
- - s ~:
E , .: L9 L,__:. V N r,- A a .'. 2' V _ T _ .~
=
f
" =- ~^ .N e.
w .- '
~
. . E .=.,~ .s
.N- .wV.,
t ._ _T- . T. E GG ~ MN N - - NIT D I I - L I
.N..."
E
.Q
= TE
_AN NV U B = M_.. . .-
.. C - - .
_.O.T... 3._r :- 9
. R ._ _.
6 __ - P _ T"
- ~
E - V m, -
- . L. M. .-
7 L 9i , .. O A V =- .
.~d ~
i- . .. .. _. .C_e
..=r...~._
._._:,-_h,,.
i m_ . _ i r
._G _ _. ,. ,. 6~
i ~.. . i i i =
- . _T..:._, I 1 :. = m . ,/ v;. .
v . . U7.=. i.
= : .N .
r =.%. -[~.
..w~."=
t
. n .-
_ _ = =. a: 'Z ,- ..
.. _ . . %=m
.~
e ., ~ L.
.~'3=_
u t
e Task 21.1 - Risk Reduction Potential IDCOR Task 21.1 was required to evaluate a broader spectrum of risks than Task 21.1 evaluated the worth of a number of mitigative or Task 9 or 19. preventive devices to study their effect on core damage prevention, Filtered and non-filtere containment f ailure, and doses to the population. wants were studied for the four reference plants. The impact of the non-filtered want was found to be insignificant on both core damage prevention and consequences (f actor of two reduction) for all plants. Filtered vented containment was found to have varying degrees of risk reduction varying from a f actor of 2.6 reduction for Bowever, Peach Bottom to a f actor Grand Gulf of 10 reduction for Grand Gulf in latent cancers. already has a very low risk of latent sancer risk and the filtered vent would avert only a small fraction of a man-res per year. Peach Bottom releases Thus it ' would be reduced by 16 man-res/yr for the committed risk profile. would be worth spending only $640,000 on a filtered went to remain cost beneficial. The drywell high temperature failure is a major asstanption that j lessens the impact of the filtered vent in these aftalyses. The high , temperature drywell fmilure is postulated to occur even in the event of ' containment venting and subsequent suppression pool scrubbing of fission products is lost. The amount of additional dose that could be averted, Therefore f however, is about 50% more than the reported estimate. approximately $1,000,000 is the most that should be spent on a filtered vent for Peach lotton based on the estimated risk to the public. IDCOR's revised source term analysis has produced lower estimates of the source ters and thus On the thehand other estimates the NRCforestimates cost benefit should be higher are significantly reduced accordingly. "
'and would raise this value. f.conomic costs were not considered in the conclusions of Task 21.1 2 Task 18.1 - Evaluate Atmosoberic and Liould Pathway Dose
{ This task included an analysis of offsita economic costs in evaluating the impact of severe accidents. The analysis is based on a large number of economic factors (see Tables 18 & 19) used in the original reactor safety study and included in the CRAC2 code. The dollar estimates are based on 1980 dollars and presumedly represent actual cost of property, and expenses. Costs were calculated for each accident sequence and compared to the cost of loss of the plant; however, no. risk based conclusions were drawn from the costs. Of fsite costs are shown for each reference plant by sequence in tables 20 through 23 of the Task 13.1 report (Tables 20 and 22 included here as examples). The highest mean of f site cost shown are for the Peach Bottom TW sequence at slightly more than one billion dollars. The TC non vent case was The highest identified as the next highest cost at 770 million dollars. Sequoyah and the Grand Gulf estimate for Zion is 130 million dollars. nuclear plaats have mean offsite costs far below 100 million. If one assumes a seven percent annual cost escalation f actor, these values should rise by
- about 50 percent to be reflected in 1986 dellars.
1 i Considering the probability of the Peach Bottom accident, TV (2 x 10*7 /hr) an estimated dollar risk, similarly calculated to the dose risk to people, is r only $300/ year. If SST1 source terms were used, this number would be and i l increased by a factor of 5 ($1500/ year)d /y ear,f the frequency a dollar risk cost ofof an event producing SST1 source term were 1 a 10
$75,000/ year or $3,000,000 over the life of the plant would result. Thus, it does not appear the filtered vented containment would be a cost ef f ectiveOther measure even in the extreme case if the conventional appresch is used.
measures of risk may be needed to assass land contamination and abandorusent. The following material is taken from Task 18.1. It identifies fission product levels that would require land restriction measures beyond temporary evacuation. 4 9 9 9 9 e# 4 6
-e ms
l Table 18 NON SITE-SPECIFIC ECONOMIC INPITI DATA Value Coenents Parameter (1980 dollars)
$499 per acro From CRAC2 Manual Decontamination cost for .
farm areas (for DF of 20)
$3349 per person From CRAC2 Manual i Decontamination cost.19r residential, business and public areas (for DF of 20) 0.2 WASB-1400, Compensation rate per Appendix VI, Pava year for residential, 12.4.2.1 busige.ss and public area '
Value of residential, $31,527 per person From CRAC2 Manual business and public areas 44344 per person From CRAC2 Manual Relocation cost
$135 per person From CRAC2 Manual Annual cost of milk consumption Annual cost of constanption $685 per person From CRAC2 Manual of non-diary products
$165 per person Value suggested In Evacuation cost letter to CRAC2 i users from Sandla National 1. abs.
dated 6/24/82 Data change 1.82/06/24
Table 19 SITE SPECIFIC DATA Seeding Earvesting Farm Land Dairy Prod. Annual Value of State Month Month Traction Traction Sales Farmland . 1980 $ 1980_J,
.Per acre Per acre 4 5 9 0.077 0.182 250 485 1 MAINE 9 0.097 0.444 ISO 802 2 N.B. 5 657 9 0.283 0.791 177 3 VT 5 '
5 9 0.123 0.283 372 1366 4 MASS 5 9 0.081 0.220 476 2133 5 R.I. 5 9 0.140 0.313 500 2158 6 CONN 9 0.315 0.579 188 642 7 N.Y. 5 2222 8 N.J. 5 9 0.197 0.162 376 9 0.307 0.413 239 669 9 PA 5 5 9 0.618 0.153 183 1516 10 ORIO. 9 0.728 0.067 206 1498 { 11 IND 5 1786 j 0.795 0.041 213 12 ILL 5 9 955 5 9 0.285 0.238 197 l 13 MICH l 14 WIS 5 9 0.520 0.598 1% 807 9 0.563 0.185 160 854 15 MlNN 5 5 9 0.944 0.050 242 1458 , 16 20WA 0.724 0.079 111 674 17 MO 5 9 f 9 0.922 0.047 45 306 18 N.D. 5 66 257 l 19 S.D. 5 9 0.922 0.074 9 0.967 0.027 99 470 e 20 NEBR 5 5 9 0.915 0.034 92 437 1 21 KANS 4 lo 0.A71 0.046 506 1725 l 22 DEL 4 10 0.414 0.227 273 1799 23 MD 4 10 0.371 0.171 126 864 24 VA a.4 472 25 W.VA 4 10 0.270 0.203 ; 10 0.368 0.056 261 819 I 26 N.C. 4 635 27 S.C. 4 10 0.327 0.063 1&B l 28 GA 4 10 0.417 0.058 1H 609 4 10 0.368 0.077 233 930 l 29'FLA l i enunum
e Table 19 (Continued) SITE SPECIFIC DATA Seeding Harvestint Yaru Land Dairy Prod. Annual Value of Month Month Fraction Fraction Sales Farmland
} ttate 1980 $ 1980 $
Per sere Per sere 4 10 0.557 0.117 141 792 30 KY 669 4 10 0.507 0.140 118 31 TENN 144 515 32 ALA 4 10 0.400 0.041 4 10 0.475 0.047 135 520 33 MISS 34 ARE 4 10 0.494 0.030 158 . 691 10 0.332 0.087 137 763 35 LA 68 442 36 OKLA 4 10 0.782 0.051 37 TEXAS 4 10 0.811 0.053 54 3 54 9 0.658 0.026 20 186 38- MONTANA 5 0.294 0.114 93 485 39 IDAHO 5 9 9 0.560 0.024 15 119 40 WYOMING 5 10 0.570 0.039 69 332 41 COLORADO 4 10 0.600 0.056 21 100 42 N. MEXICO 4 0.556 0.069 36 134 43 ARif,0NA 4 10 0.236 0.215 36 265 44 UTAE 10 0.127 0.117 19 104 45 NEVADA 4 10 9 0.369 0.138 132 586 46 WASH 5 9 0.300- 0.093 68 330 47 OREGON 5 0.318 0.119 316' 936 48 CALIF 4 10 0 0 0 0 49 NOVA SCO 5 9 0 0 0 0 50 QUEBEC 5 9 0 0 0 51 ONTARIO 5 9 0 0 0 0 0 52 BU A CAL 5 9 0 0 0 53 SONORA 5 9-- 0 0 0 0 54 CEINUANU 5 9 0 sammeum
, , e e - e*=.
6.5.6 . Areas Contaminated by Deoesited Fission Products Among the outputs produced by CRAC2 are CCDFs ior areas requiring . decontamination and areas from which people might be excluded for long times, In each case because deposited fission products are predicted to emit gamma rays and deliver radiation dose at an unacceptable rate to individuals residing in those areas. Figure 13 shows the conditional mean for these two areas as a function of the cesium release fraction. The reader is referred to the CRAC2 model description (Q) for a precise definition of these areas. For most of the range of source term magnitudes, a f actor of 10 reduction in the cesium release f raction leads to a f actor of 10 reduction in the area, as shown by lines AB and CD. It is apparent, however, that the current IDCOR source terms do not alisinate the need to decontaminate er relocate in the event of a sevege accident (cesium release fractions would have to be reduced to 10~' and 10~ respectively for this to be possible). 6.5.7 offsite Costs Figure 14 shows the predicted conditional mean of fsite costs as a function of the cesium release frection. AB is a line exhibiting direct proportionality between the cost and the release magnitude. There are significant departures from this line, particularly for the smaller source terms, where the costs become almost independent of release ma3nitude, see particularly the Grand Gulf results. This 1.s because at these low levels, the only cost incurred is that for evacuating the EPZ. That is, for source terms with cesium release frections that are less than about 10~3 to 10~', the costs incurred are equal to the number of people in the EPZ times the cost of evacuating an individual ($165 per person in the present work). Thus, the ratio betwgen , the Grand Gulf AZ cost and the Zion V cost at a release fraction of 10** is ~
'almost exactly equal to the ratio of population within the EPZ.
For higher release fractions, the slope of the relationship between costs and source term magnitude becomes much closer to that of the line exhibiting direct proportionality, AB. This is because, at higher cesium release fractions, the enst of decontamination dominates off-site costs. As can be , j l seen from Figure 13, the area requiring decontamination is almost directly l { proportional to source term magnitude. l I i l , 6.6 UNCERTAINTIES l l l' A thorough discussian of uncertainties lies outside the scope of IDCOR Subtask 18.1. This is partly because only limited sensitivity analyses have However, been done on the source term magnitudes and other characteristics. some relevant comments can be made. First, the conclusions about there being l very few early f atalities er early injuries seem to be quita robust with respect to uncertainties in source ters characteristics. For example, most of the IDCOR source terms would have to be increased by an order of magnitude or more to bring them into the hatched area on Figure 9, i.e. average 1, Cs and .Te release fractions.
r jI
- {
- ~ l I a I T f 8, *. 1 D at g W r S. o o
- s eST-1 n
O P VN 'Tr no oi I O C e i t
} # T- - -
t c l O S u 3 aa nr I T F 1 O "' P e-t w T e 01 iT n s ac
) - t e
# Y
- I C e ' na N
- I' oc W I C 1 M '
cl ee T S P
- T S
P E M S 2 C t t F w V >. ( , F Dp gs
,O
,I
_ V 0 9 L M O w. T-Me i nC rf
, A T S Z 2 P w c, 'n, uqn io
, R T2 r i -
P r Z i A
- t 0
- S
.O
- i t
M 1 e. e eo V -i N
, m..e f, T
- t l
c a e Rh mc t OC
. t I
w t o I rGO t S OI e an Ct i T 3 o,- Ms F n eu e se A rr
. s, I
F s, T rH ev a V s t L ,Y f 2
- p52 s s n C - M -- C as ea T S T-
-S _
3 o0 3
' M s
- P G o I O s 5 0 n n P S
- a,, t i
C3 a gh I 1 l o ai nt
- c C.
v 2 C I D oc ii Z ( 7 R
~
3, - td 'w T e C T G C ir de G G g v V I C - V[ . nt on CI
+
G - T W O 03 s " .
~
s e 2 g 0 3 1 dv T 1
- V e e t e
e s M G U r ew r O G O Dg u g o e C h t r M oI O t N t T i r e O Og G ht s-s . oO er ene .e. I T A G e s e C O V $ e s d de e o d G v E Z
- I ~
r e r nt 0 t e e e e t Z A 1 m m e.e D-G d d t e et mm, e s t c i t i m G' ed rno no . v 5 e e c c t t e e A In n DDz I - 8 O 4g3OOg D 6
- 0
- ~ 1 1
- 1 -
3 7 0 ' 0 1 0 1 1 -
- h
- I=5 5 $0 1 gTI- r hjI.I814l .
I i
' . 1!
,lfIl l
. 8 4
W am. 4 c E d O
. k
- a O > u.
- , b N 3 '
u O g @
= l e.=
e"1 y 2 :* Es o E w -
*- ~
m u
- a. _
~
~
g# E t C e s - IO
-- hl 8
- o-
&E= }
E z J . 5 -s a E . 8: o.z . Wu ky%n I D m a e l
~
b Wee '21 0 3 b $$, l on - U w N w l C
- c c
> > t 5 @.
p. I o - e 8
-o 4 u 0
!3 j
a y s o y-e e w W a [ h
.. 5 5 5 C
j o e oa i e i i l o SS
- 6-23 l
ETTECTIVENF55 OF VDTTING BY SEQUENCE As moted in the summary of the task 21.1 evaluation, containment venting does not mitigate every sequence. Several important risk sequences woeld not be mitigated at all, although they are generally low in core melt frequency. The following tables identify the dominant sequences that j result in fission product release and summarise accident timing and source terms. For Peach Botton, the SgE sequence is a low contributor to core melt frequency and similar to station blackout. It will not be discussed J further. Table 3 provides the fraction of core melt frequancy whers ! containment venting may be beneficial. l Peach Bottee
'1M Sequence: Venting in this sequence can prevent the loss of containment function and the assumed loss of core in3ection systems. In the current .
analyses these do not occur before 34 hours. Venting in this case would substantially delsy f ailure, but cooling must be re-established. If cooling is not re-established, the pool will eventually deplete until the pool is bypassed. If core injections remain functional'and water is ; injected from axternal sources, then cooling is established and venting I will be successful. TC Sequence: Venting is only partially successful in eliminating releases of this sequence because of the assumed f ailure of the drywell wall. Bowever, releases are much less if venting is performed. If the drywell wall does not f ail, then releases will be only noble gases. The use of sprays with venting may prevent this containment f ailure mode f rom occurring. , f TQVW-Sequence: Benefit from venting could be obtained in this sequesce, if the uncertainty surrounding marly drywell f ailure can be averted. 1
)
I Crand Gulf l T23QW: Similar to Peach Bottom except overpressure creates path through suppression pool. , l T23C: The K1VS sequence with containment f ailure 1 saves Pathwsy through { suppression pool intact. AE: long tern overpressure failure of containment lesves suppressi n pool intact. T 1QUV This sequence is equivalent of station blackout. Containment f ailure leaves suppression pool intact. Other: There is no V sequence for Grand Gulf. Containment bypass is not as significant as TW. Randos f ailure of a drywell penetration would render a containment vent inef fective unless an external filter were also installed to supplement the filtering action of the suppression pool. j
--- _.-___m___
1
. j
\
TABLE 1 ACCIDENT SEQUENCE DESCR.IPTION 1 Accident Plant Sequence Description Peach Botton 'N Transient with failure of decay beat removal j TC Transient with f ailure to scraa TQW Station blackout Grand Gulf T230W Transient with failure of decay heat removal T23C Transient with f ailure to scram , T1QW loss of of f. site power and .f ailure of injection AE Large 1DCA with f ailure of injection Zion TMLB Station blackout V interf acing system LDCA bypassing containment Sequoyah V Interfacing system LOCA bypassing containment S2H7(DB) LOCA with failure of recirculation and sprays, (drains blocked) S2HF(DO) thCA with failure of removal and sprays (drains
. open)
T IMLB Station blackout Containment Failure Modes 4 - Containment isolation failure f - Overpressure due to steam generation NCF - No containment iailure m*
g; 1
'. ~ '
))) )))) ))) ))))))
o 233 5555 235 453435 M - - - - - - - - - - - - - - - -
- ((( (((( ((( ((((((
u 511 1111 111 111211 R < (( ( < d e ))) )))) ))) )))))) s 555 5555 555 555535 a a - - - - - - - - - - - - - - - - e B- ((( (((( ((( (((((( 11t1l 1 l 311 1111 111 e r < < <<<< < < <<<( ( < R S S y ))) )))) ))) )))))) T r b 332 4455 425 455435 N o S- - - - - - - - - - - - - - - - - (((((( A t ((( (((( ((( L n e 811 2813 241 118241 F e T < < v - E n C i N E e b ))) )))) ))) )))))) R r R 132 4455 534 453344 E o - - - - - - - - - - - - - - - - - (((((( F C s ((( (((( ((( E C 233 3817 582 143274 , R f o R
- O n ))) )))) ))) ))))))
F o 132 4455 534 553344 i r - - - - - - - - - - - - - - - 1 3 t B ((( (((( ((( (((((( U t c - 233 3811 382 663264 I a 1 <> E r T F E r C )))) ))) )))))) R K ))) U - 000 0000 000 000o00 O e ((( (((( ((( (((((( S X 111 1111 111 111111 D E T g C n E ie) L E ns rih i r 780 6276 623 691n19 S aT( 1 1 4 54 2 3 a 1 . W 2-1 1 n eo si) E. 1 at r 000 0000 267 796o27 B eah 853 8511 21 1 t2 A l r( T eu I RD f e os) ' 1 ar 098 6387 843 81 61 39 - eeh s 4 1 5 54 2 2 31 4 1 0 1 al( l e TR s i a n s o esp d t f t l - - - a e o u h ))) e 3, a BOOS- p r
/n t e c B h
t nS e-G d
$(6
~
- V
- 8(5 y
o DDD ((( n. B ) nu aq c a SVW n a WC3U 33- Q n o BB u q FFFt L HN1 nM 1 x l e e WC} r 2 2E g e 222ig ( PS P 1T1 t jt TTAT i Z Mmv S 555TTv *
~
I I , i l!
ee ))) K K ) f u 8 8 f q 667 - 3 - - ae nr UF ((( 343 3 7 7 )M)7 ( 3 ( 1 ( 3 t n mre i n r oV o ef i T d 8 t ee 6 2 i esr 0 s cae - - o ret 1 p 5 ull 33 s oei eo 00 e 0 SRF 0 0.Y N d t c u d o) E .
' rr pe S h t g s A nn E in oe L nWi id E i l s n R no s o eoo i c F viC f O it t as e Y cni e rc t s oi C eir ua j N fbb a n E f me ss ss b e U EoD eeee l mi Q C YYYY ne (
E i or 3R t r F ce E. N t uv BI d o A e TN Rl O o I mr rt T t n en C To U e c D V yyyy e E rrrr cn R d aaaa ri e ua E r rrrr V e oooo pppp ot S n 5 t ss I T l smmmm eeeee oa i s eoo eeoe C i Nm YNN YYNYNN E F YTTTT F F e E d d e e m m u u s s : s s a a g g n n i i t t 5 0 n n y - ~ ec e e 0 0 cn v v ) 1
)
1 ne eu ))))) 1 K )) 0) K ))) ) )))) 0) uq 766571 7617 6876 7 687 1 qe - - - - - - 9 - - - - 3 - - - - - - - er (((((( (((( ((( ( (((((( SF 2343343 2711 7 631 6 6531 69 _ e . c 9 n e n o
/
t n * [i -aJS _ u t 6 f , q e t o f
/0V eD l u T *'y "7 h ))Je B t T G E'G a BOO Gn-)-qT S - -
/ - y DDD -
e) V - - o (( ( BB t h - d N TV1( NE u FFFLL l n c n t C U ". BL1B. i o - - [ a 3 3E a O_ q 1 i11 i i ' r
Sumsmary: Containment venting not important for controlling fission
. product s.
Is1SL".! TMLB Station Blackout with overpressure f ailure at 23 hosra. Releases are estimated to be very lov MX). Filtered containment venting would be applicable in this sequence. TMLB with impaired containment: . Containment integrity is assumed violated .; at T=0. Vent inef fective. Fission product release 1 to 41. V sequence: Contaissent is bypassed. Vent is ineffective. Release may be large if the break is not under water. Summary: For Zion, a filtered containment vent is ineffective for the highest source term events. . Secuoyah , 52HF: Containment is norms 11y challenged on overpressure. Vent would protect f rom overpressure failure, but source term reductions would be small due ta large deposition that occurs in the ice condesser. TgMLB: Station Bischoot resulting in an overpressure of containment could be protected by a filtered vent; however, the predicted releases are very low thus the source term value of the filtered vent is minimal. The i filtered vent could also be rendered inef fective if hydrogen collected in j suf ficient quantities and ignited to overpressurite containment in absence ) of functioning ignitors. , TgMLB with impaired containment: Tiltered Vent would be inef fective; however, fission product release is low.
- O V sequence: This sequemce bypasses containment, thus a filtered vent would not be ef fective. -
l l l l l I I l l
. i l
Filtered Vents Planned or Implemented by Foreirn Countries Sweden-Filtra Preiset l In the early 80's, a filtered vented containment concept was designed for l the Barseback BWR plant in Sweden. Thefiltersygtemconsistsofa cylindrical reinf orced concrete vessel 353,000 ft in volume, 65.6 ft in diameter, and 131.2 ft in height. The vessel is filled with 1-inch site crushed quartsite gravel. The gravel bed serves three functions: o a heat sink for condensation of sten vented from,the containment o an expansion volume for reduction of the gas pressure in the containment o a filtering medium for removal of the aerosols and iodine in the . steam and gases form the containment The FILTRA system has one vent pipe from each of the Barseback containments (see Figure A.19). The vent pipes are connected to the wetvell gas volume in each containment. Outside cach containment the vent pipes contain a rupture disc about 2 feet in diameter followed by an isolation valve. The isolation valve is normally in the open position The vent pipes to the filter vessel have an equivalent area of 1.1 f t.y gravel condenser, Bowever,inordertoincreasetheresidencetimeintpThegravegbed the vent line has a flow restriction of about .11 f t. is said to be capable of condensing and accommodating about 500 m of condansate. To prevent any hydrogen burn during the initial containment venting phase to the filter, the entire filter system, including vent pipes and gravel iLiter, is inerted with nitrogen during standby conditions. Experiments were performed that showed that hydrogen burns within the f11ter system could be tolerated. The filter is not designed for large seismic events. The FILTRA system also has a vent path leading from the drywell. Although the priuary path is through the wetwell, this alternative path adds additional flexibility to control some sequences that may lead to by pass of the suppression pool. - Many large and small experiments were performed to evaluate the capability of the FILTRA system to provide the three functions previously listed. The results of experiments related to steam condensing and radionuclides retention provided data in support of analytical models that permitted e: extrapolation of the results to a full scale system. It was shown that less than 10 4 of the mass of radioactive particulate will penetrate the gravel. Iodine will be completely removed except for the organic Iodine which is less than 1% of the total. e
FILTRA will have little effect on some events comprising the residual risks. In this category there are accidents that could be caused mainly by asternal events such as earthquakes, but also some internal ones, e.g. lack of containment isolation, reactor tank rupture, I turbine missiles. At low probabilities the internal events also I include the core melt events that could result from pipe breaks with l excessive laakage between drywell and vetvell and some of the { transients without reactor shutdown. ) 1 Evaluation: A major conclusion from this data is that the gravel filter is very l effective for fission product removal and for steam condensation. However, the FILTRA study did not show that a substantial risk reduction j was obtained. Furthermore, the Swedish have stated that the major ! objective of the system it to reduce releases of those radioactive {; substances which could cause long tern land contamination. Cost was estimated in 1982 at 20 million dollars, i i
+
O ee 1 l
~
{
TABLE FILTRA Proiect System Characteristic Parameter Decontaminstien factor 1000 Risk Reduction Fassivity Yes Manual / Auto Initistion Auto (rupture disk) ,,, i Y1ow Rates 2-3 cm/sec Vant Pressure 94 psia Vant Site 1.1 ft. Reliability high Cost Estimate $20 million
~
volume of Gravel Bed . 353,000 ft.3 Dimensiens of Cylindrie.a1 65.6 ft. dia. x 131.2 ft. - Filter Vessel F S Oe e GBWAmp m . 6
-___ _ _ _o 4
1 l0 1 i
^
,_ Off pas line , - Grevel !
0,01 m2 sondene
~"
4~" O o' { 10 000 m3 _ . [- Rupture dise Vent line , 550 kPa f' 0,1 as 2 @
/___g -- , _ ~--3.
D . , i
. .e
- - v. -
- wy -
s i l
\
].. 7 ShuteN esive I
g 1 Tigure A.39 FILTRA Vent filter plant arrangement Q 3,) 'l I g a,
i French Filtered Vent for the 900 MW PWR Desian The French ff.Itering system is based on an inexpensive sandbed filter placed between the contairument and the stacks. The gases being released ar; monitored by the stack radioactivity control system before release. Specifications for the design of the venting system are as follows: o maximum gas flow rates: 3.5 kg/s; o typical gas compositions air (331), CO2 (33%), CO ($1) i system (28%); o gas density: 4 kg/m3 ; e inlet gas temperature: 1400C; , o inlet pressure (before decompression in the pipes): 5 RAR (70 PSI); . o pressure drop through the f11ter: 0.01 KPas o pressure at f11ter outlet: nearly atmospheric l o maximum airborne aerosol concentration: 0.1 g/m3 total I amount: 5 kg; l o aerosol mean diameter: 1 microns and l o filtration efficiency: 10 (objective). / The filtering system has about 40 M 2 in surf ace area and 0.8 M ef sand height. It appears to be insuf ficiently sized to constitute a heat sink
- large enough for the bulk of the steam passing through to be condensed.
It may be adequate for situations that ,may occur very late in m: accident sequence. l The sandbed iilter has been specified according to the results of the
" PIETAS-filtration" arperimental study, in which the best combination of l iiltration af ficiency/ pressure drop parameters was established. i The design utilizes existing penetrations, rupture discs and ionistion l valve and passes into the sand he16 housing. Control of the device is accomplished through manual operation of isolation valves. Details of the j tioncept are uriavailable, and the ' device may only be in the conceptual y stage at present. (See Figure F.3). There is apparently no provision l made for the possibility of a combustible gas ignition within the piping )
or filter system. Unless carefully designed for this possibility the sand l bed could be dispersed causing an unfiltered release if the disruption is not detected. The filter design ef ficiency of 10 is not satisf actory to ; prevent doses to the public or land contamination from occurring in all i situations; however, there will be a substantial reduction in the extent of contamination. Both land area involved and the degree of contamination
-. i a
4 r_ - Outer ~ containment 2 5 ina.r ._ _ containment
- m m '
~
Steam generators s t .mmm. L. v v I
,9
- vy Reactor- '
]
,f Quench tank -- ( ) i , instrument-
_( l OM /_ cables
- . O_. -
O
)
Figure F.I . rrench PWR Containment RCV/k 11/W30-3495
- Revised 3/11/8e e
PROCEDUR5' U2 i
~
( l SPI ou SPU lU2 Y v
\' V'
- [ Pas d'activith't ' #1 ,
(cheminee f - 1r I M . . n.s t,,,g<,,,f,,,,,,
- Scrutation - Local contami
)d ; , },g aL p* FermerLocalisation du defaut - A'"%
vanne{s) en ddfaut' F vg,%[ ou diancher la fuite r ou confiner le local .. M. l 1 r 1 r- ,4 v.< 1 , -
- F Pas d'activits T o hulsards BAN, BKjj L .+d- 4 N O J,.. h is R l* Reinje:16n des effluents dans le 4
~ 3 r 1r f
; g,.q(Tr&s haute activild) l enceinte J g, "1 ,. ,
y 1 r tam / Fg,,s 3 . 1 r (p$, ?!.J of I-Sulvi de P O
- Confirmation isolement des U1 traversies exterieures aux BAN, BK, BAS pour 1303 l
. .c
. I, t Favar 'F.2-
_ _ _ - - - - - - - - - - - - ' - - - ' - - - ~ - -
t p F g ,uer F. 3 Proo6 dure US
. Ll=
WWE ! E se. v Caract6rirsves
#- de la r6f6ranoe 4
dJ Oaz & Rtrer :
- 3,5 kg/e ,
140' C 6 bars P6netration 0.1 grem a6ro d s ,, , existante i . 900 MWe 900 MWe 1 2 ....
... .. .. ... . .,. . .. , .. ...o.... , ..
,, _i
.............. . m...
...w .....,
....................................... ....w. ..... . ........... ... . . .
- v. u,.
iu..
.. .. .................. m u,uuunu t- ._
iuio>uu
.................m...,................................
. ................... 144............. ....,
......., i4...< .,
....M........................... ...
.de sde
. . .. .. ,m
_.x
. . . .s . . . . . _
- . z.- . ..............
.. .... . . ' , UChe e Sab>-
.. .... u F. -,
e IR'O
.u.............m_mmmmmm m ...
m,o. o..,
.....+
vm ..... (d6tente +1 bar) Calsson de Rtration 2 , Ap ,c 100 rnbar surface 40 m - l l efficacit6 > 10 pour a6rosols
'vttesse du gaz 10 cm/s n
- s. .
h ==== rant undersizine af *ha f 41*-- fino f t3) and will be reduced. _of nrovinian for drainate or coolin. == ban the desian the monarant 1=*h highly vulnerable to steam me===tae af ff== inn eroducts free the filter ! and releasing fission products to the environmentuor an alternative . possibility is that the trapped fission products will generate sufficient heat within the filter to cause them to be released. The deficiencies in the design could be counter balanced by added instrumentation and controls over the venting process. The operators by carefully monitoring the venting process could intermittently operate the vent to prevent disruptive levels of combustibles, control the beat lead on the filter, or stop release if the filter becomes saturated. , The cost of the French filter system has been stated to be less than
$1,000,000. If esisting penetrations are used along with small lines and the simplistic design and construction indicated in the concept, then the cost may remain below $1,000,000. This concept does not appear to be .. ,
adequate for reliability.or sufficient reduction in fission product i release to assure no land contamination or to substantially reduce public exposure.. Thus, the filter in our opinion is not an adeouate engineering ' _fix for ventina but a relatively inexpensive political solution. It appears to be vulnerable to criticise f roe competent interveners who have access to the design information. This opinion may change when more information becomes available concerning the design. German Filtered Vent Concept There is very little information available concerning the German concept. No design parameters have been reported. The concept, however, appears to
. be similar to the Swedish and French concepts (See Figurie G.1). The i concept utilises a rupture disc, isolation valves, line drainage capability, passage through a filter and release of gases through a stack. The efficiency of the filter appears to be about 100 based on the .
. releases shown in Figures G.2 & G.4. If true, this systes, is likely to be much more expensive than the l'rench. No detailed evaluation can be made at this time, but it is apparent from the figure that the German System will also be vulnerable to containment bypass or impaired sequences. l i i l 1
I SB RR HAG
~
- N
)
k, b :%: T g x bl A
.. ~
{ _ .' CONTAINMENT ANNULUS AUXIL1ARY
... BUILDING
.M ZUR DRUCKENTLASTUNG DES SB
@ SYSTEMSKlZZE DESIGN SCHEME.FOR CONTAINMENT VENTING F u **E f,i
10 0 3d 4d 5d s 1 V'. YL C 1 elem. Jod J /8' 3 ag. Jod g aeros. Jod (CsJ)
!u
= a E
4 4" ) g . .{ 19 a ( y 3 E /- , y
'n E,
! 5 10-6 95 '
# ~
( ' 104. ' 72 96 120 h - Zeit nach S16rfalleintritt = TIME ArTER INCEPTION OF Acc! DENT ! g SPALTPRODUKTFREISETZUNG BEI !
!E DRUCKENTLASTUNG, HD PFAD FISE10N PRODUCT RELEASE IN THE CASE OF CONTAllNENT VENTIE (HIGH PRESSURE CASE) l pi at C. L o
10-10 ,0-12
l .
~ .. :- l I
2d 4d , 6d 10:FAILURE PRESSURE bar,. versagensdruck ( 8 ' w DESIGN PRESSURE Druckentlastung
.~
!'6* I'9
- 95d S Ck E /'
g 5 .4 /l
- , [
2 . 0
. 0 -
20 40 s 60 C Zeit = 104 .I Tmn - l DRUCKENTLASTUNG DES M' SICHERHEITSBEH LTERS NACH EINEM
$ KERNSCHMELZUNFALL, HD-PFAD . _
PRESSURE TIME HISTORY FOR LATE CONTAINMENT VENTINr. (HIGH PRESSURE CASE) Figuer G.3 h
~.
109 .
- f .
, 10-(
r 8 5 g . f 4 A Kr Xe ~~ E 10-2 t elen Jod g
. 3 org. Jod .
g ,E 4 aeros. Jod (CsJ) E - E E.10',3 ,
! f -
r y-
<'0-4l 1
l 0 l- 10 'O 8 16 24 32 h 48 l Zeit nach St5ffalleiritritt = { TIME AFTER INCEPTION OF ACCIDENT j i g SPALTPRODUKTFREISETZUNG bel VERSAGEN
. DES LOFTUNGSABSCHLUSSES, HD-PFAD FISSION PRODUCT RELEASE FOR FMLURE OF CONTAINMElli ISOLATION (HIGH PRESSURE CASE)
Figuar G f ____m_.___ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _
Land Aba=4a-=nt Around Chernobyl
~
Releases from Chernobyl conti= =d for several days after the event began. Contamination was predominantly to the northwest and north. Wind , conditions at low and high sititudes coupled with the hot plume release , condition resultad in substantial transfer of material far from the site as well as high deposition in several areas to the north of the plant out .. to 30 kilometers. Readings out to 10 kilometers were as high as 1 R/hr on April 27 and 500 Mr/hr on April 28. At 30 kilometers, readings were , approximately 1 Mr/hr. No deposition beyond 30 kilometers approached levels that would suggest interdiction. , l Some of the area in the northernly sections out to 10 kil'ometers from the plant have been exposed to sufficient fission product deposition that temporary abandonment of the land has been required. If 1/3 of the land within 10 kilometers of the plant and to the north is taken as abandoned, then the land area lost would he about 20 square miles or 12,800 acres. .. Unless an unusual concentrattua of expensive housing, laboratories, or ' businesses are involved, the economic cost of the land abandonment would
~
remain within previous estimataa for U.S. Plants. For example, assuming the average cost of land, buildings, etc., is $10,000 'per acre (This is , J such greater than the cost of farmland) the estimated cost is
$123,000,000.
Susanary of Filtered Vented Containment considerations and conclusions Filtered venting cannot be justified based on cost benefit considerations. Given the uncertainty in the cost estimates (most notably civil suit awards) and taking a conservative estimate to core damage frequency, some BWR plants may show a marginal cost benefit for vented containments through the existing suppression pools, although vented l containment alone may be insufficient to substantially reduce releases of - j; the volatile and particulate fission products because of the assumed
- thermal isilure of the drywell.
~
Although land interdiction is not justifled on a cost benefit basis, a benefit may be gained with a highly ef ficient filtered went or vented , containment (in the BWR Mark I er Mark II case) by reducing the fission .
)
product releases to essentially noble gases for a portion of the important accident sequences. A larger increase could be achieved by taking ' measures that would reduce the importance of impaired containment and l bypass sequences. The BWR Mark III plants would not benefit free l containment vent unless the vent system also includes an external filter - l because the Mark III design is protected in all cases except Drywell bypass (the potential for an interf acing system 1DCA, V-Sequence, is too , l low to be meaningful). A system similar to the Swedish design or as considered by IDCOR having a reduction f actor of 100 or more is satisf actory to assure that essentially noble gases would be released. For many BWRs, the availability of means to cool the core debris that reaches containment may also be necessary to provide the assurance.
e
,. 's ' \
The French filter design does not appear adequate to provide assurance that land contamination would not occur. Bowever, it may be adequate for a limited set of conditions. Much more information is needed about W French severe accident response of containment and the design of the filtered vent to conclusively rule out the design or resolve concerns i about its capability. The German design is very asuch in the concept stage I in so f ar as publicly available information describes their system. Bowever, it appears to be a conventional desige. Another factor that may be important for a decision on filtered vented containment concerns recovery from a severe accident event. Use of a vent to control pressure prevents the unpredictable failure of . containment and allows the releases to be stopped so that recovery teams are not concerned with continual cloud doses or with plugging the caetainment breach. I
! 'One negative factor in a filtered vent decision is that the use of a vent during an accident will result in releases of noble gases. For most PWR - , accident calculations this represents about one third of the dose , calculated for individuals. Filtered vents may be viewed very negatively by the public as a result. ,
A significant reduction in land interdiction frequency may be possible with the filtered vent but the area of land in geestion does not appear to be very large. Conditional doses to individuals may be substantially reduced in some cases but overall the reduction may be less than a factor of 3. The cost of a reliable filtered vent for a FWR is estimated to f all k" within the range of 10 to 35 millions. Inexpensive systems may be possible while limiting the conditions that can be handled. BWR systems asy . I utilise the existing suppression pools but assare that the filtering effect is not bypassed. . j In summary, filtered vented containments do not appear justifiable solely ' on an economic basis or from the standpoint of a r educed public risk. ee e A e m e sa.. m ee . im - _ _ _ _ _ _ _ _ _ 9
O. Berg et al.: Early fault detection and diagnosis C Carl Hanser Verlag.!vi6nchen 1987 O. Berg, R.-E. Grini and M. Yokobayashi Early fault detection and diagnosis for nuclear power plants i process simulator, called NORS, and a control room section . l Fault detection based on a number ofreference modeh is dem- equipped with workstations for all anain surveillance and con-i onstrated. Thu approach is characterued by thepossibility ofde- trol functions where both softwats and hardware can be rent.
.i tectingfaults before a traditionalalarm system d triggered. even ranged easily to create a wide spectrum of realistic experi-f in dynamic situations. Further, by a proper decomposition mental conditions. The aim othe system development work
! scheme and use ofavailableprocess measurements, theproblem in Halden is to design and to build computer systems working area can be conpned to thefaultyprocessparts. A diagnosis sys- in this environment that can assist and support the operator in tem using knowledge engineering teck uques is described.1)pi- his various tasks and through shis improve the total perfor-calfauhs are classsped and described by rules involving alarm mance and safety of complex plaat operation. One such sys-patterns and variations ofimpo, tant parameters. By structuring tem, the core surveillance system SCORPIO, was implement-thefault hypotheses in a hierarchy. the search space is limited, ed at the Ringhals Nuclear Power Plant in Sweden (1). An which is importantfor real time diagnosis. Introduction ofcer- integrated disturbance handling system for use at nuclear taintyfactors umprove thepasbility and robustness ofdiagnosis plants is under development. This paper describes the activi.
by exploringparallelproblems even when some data are mining. ties of fault detection and diagnosis which are carried out A new display proposal shouldfacilisate the operator interface within the scope of disturbance handling. and the integration offault detection and diagnosh tasks in db-turbance handling. 2 Detection 2.1 General Fruhreitige Fehlerfindung and Fehlerdiagnose in Kernkraftwer-kea. Methoden der Fehlerfindung, due sich auf mehrere Refe- Detection of faults in large indestrial process control systems ren:modelle stGt:en, werden erlautert. Garakteristisch fDr sie is an important part of the total surveillance functions due to ist die M6glichkeit. Fehler auch in dynamischen Situationen the potential large consequences if they develop into acci. auf:ufirden. bevor das konventionene Alarmsystem ausgel6st dents or emergency situations. An ently warning can reduce wurde. Die Problem:one kann durch ein diferenziertes Zerle- the risk of severe accidents, case the diagnostic task by giving gungsschema unter Benut:ung verfirgbarer Pro:epgebpen auf better localisation capabilities, and \ cave more time for coun-diefehlerhaften Bereiche eingegren:t werden. Em Diagnosesy- teractions. If it is possible to isolase the problem area before I sitm, das Knowledge-Engineering Verfahren bennt:t, wird be-
} schrieben. Typbche Fehler werden durch Regeln, die Alarmaus-15semuster und Enderungen wichtiger Parameter einbe:ichen, L
klassifi:iert und beschrieben. Der Suchbereich wird durch Struk-turierung der Fehlerh>pothesen in einer Hierarchie eingegren:t, wa,sJIer eine Echtzeit Fehlerdiagnose wicksig ist. Die Einflihrung t"I~ W
.i _
y g von Sicherheitsfaktoren verbessert die Flaibilitat und St6runan-
. fdlligkeit der Diagnose mittels Untersuchung paraueler Pro. -
! blemstellungen. auch wenn einige Datenfehlen. Ein vorgeschia.
genes neugestahetes Display soll die Operatorschnittstelle
& Qf. Q g %9 entlasten und die Integration von Fehlerfmdung und Fehlerdia- l gnose bei Storungen erleichtern. t e tau 5 l 1 Introduction ,0 l The surveillance and control of large industrial processes i t comprise a number of tasks and functions which have to be * * '
shared between human beings and computer systems. The de- carms I carms gree of computerization and which functions should be left , 1 { for operators in control rooms have been widely discussed in : the past and will probably be a continuous item for discus-
' sions also in the future. However, it is a clear trend in design- waa
, ing new control rooms to replace com entional panels by cath- lig.1. Space-time domam categoruanon of alarm systems and Early l ode ray tubes (CRTs) from which operators survev the process fau/r Detection. EED. Afault is armhred at a certam time, and it is d-d^ em j and exercise the necessary actions for its control. The Halden N "7,jfp',^',, $ *"[,,7 M'* ',^', "%P f ,f ,,
rea; Project has an experimental control room laboratory which prove the spatial resolution and gne warumgs before traditional alarm l contains an advanced large scale pressurized water reactor limits are reached i 90 Kerntechnik 50 (1987) No.2
--_ A
K. Geyer et el.: Accident manzgement matic activation or during the initiating phase of ue ac- e fewer redundancies than conservatively assumed m the li-cident. censing procedure are required, and o To establish long-term residual heat removal. e considerable time delays are acceptable for a.:vadon of Ifit is not possiole to fulfil all these objectives, stage 2:n.ay the safety measures. nevertheless be considered under control provided thu at Table 2 contains characteristic periods of time pnor so core nec_st the residual heat removal path is ensured. meltdown following a loss-of coolant accident (low-pressure An example for prevention accident management is 4eed- path) and core meltdown following emergency power opera-nater supply by mobile resources. tion (high-pressure path) for KWU pressurized waar reaaors o At the third stageit is assumed that core degradatiax has (5,6). tdvanced so far that the remaining reactor core geome:ry is The delay time until the start of core degradanon is also no longer coolable. De progress of core meltdowr md serv important with respect to the possibility of r:amaal ac-subsequent reactor pressure vessel failure cann:t Ne tions inside the reactor building. In the light of experience stopped. Then, the only objeai,6 h tv wnfine the rch pined from fuel assemblies in pressurized water rum, the radioactivity inside the containment. lt is assumed that the radioactivity contained in the reactor coolant is nepigible. In wide range accident monitonng instrumentation it sull the event of coolant escaping from the reactor coohru system available to provide information on the conditiom ma$e before the initiation of core degradation, no serious em a mi- j the cor.tainment, e.g., pressure and temperature. nation of the compartments concerned has to be capected, j ne objective of accident management at the third stage Rus, the accessibility of the compartments concerned, partic-
]
of core degradation is to mitigate core-melt consequmm ularly those of the annulus,is not noticeably impared; there- { espedally: fore, manual actions can still be taken. I o To actuate containment isolation unless this has aheady l been effected automatically or during the mrtaang 4 Conclusions j phase of the accident. This measure is identical to ftat at l stage 2. Research done in recent years, especially risk analysis and ! o To prevent serious consequences o r the combusten of core-melt studies, has shown that the total core-mdt frequen- { hydrogen. cy of modern KWU pressurized water reactors is in the rance o To prevent containment failure in the long term die so of10-6 per year. Severe consequences to the emirocment are overpressure. prevented by the containment with a failure probability less As regards environmental impact,it makes no diffeence than 10-2 per demand. His high reliability of the coctain-whether accident management measures are effective wth re- ment isolation was confirmed already in phase A of the Ger-spect to prevention of core melting (core cooling) or to mtip- man Risk Study [3). Taking into account accident m.anage-non of core-melt consequences (containment isolation; The ment measures as discussed in this paper, the resadual risk is reason for this is apparent from Phase A of the Germaz Rsl negligible even without consideration of offsite emergewy re-Study [3]. Realistic evaluation reveals that, from the risk rand- sponse actions. With respect to offsite emergency response point, release categories 5/6 (core meltdown, contamme plans,it is sufficient to take into account core melt evenes with esolation ensured) and release category 7 (core cooling effec- isolated containment. tive containment isolation failed) are to be considered brEtfy equisalent. His equivalence is particularly important for such (Received on February 27,1987) anitining events which may potentially cause loss for emr=- ment function. The authors of this contribution 3J Boundarycondaions Dipl.-lng. Karlhein: Geyer. Dr. Otto Gremm. Dr. Unch Krug-mann. Dr. Harald Roth Seefrid. Kraftwerk Union AG, Ham-The limits imposed under the licensing procedure are deter- merbacherstraBe 12 + 14, D-8520 Erlangen. unined with a considerable safety margin to hazard Emin Therefore, the implementation of effective accident manage. References snent measures requires a realistic assessment of 1 Ncif. H.; Grycr. K.H.; MGller. R.;Orundsatze der BewdI=nr von o the efficiency of operational and safety systems: Storfallen im Betriebshandbuch von KWU-Druckr o the efficiency of alternative measures to be applied uben Kerntechnik 50 (1987) 101 all these systems fail. 2 Gcver. K.H.; Reince. H.: Computer aided incident rneprm: Use I o the periods of time available untilloss of coolable g:ome- gj9%pNe displays for pressuriud water reactors. KecuerMk 50 try; 3 Deutsche Risikostudie Kernkraftwerke. Verlag TtW Rheanland. s the accessibility of compartments for the on-site p:rfcr- Koln 1979 mance of manual actions (in particular in the annuus of 4 #derner. H.: Fachgesprsch der Gesellschaft f0r Reakt:rsacfberheit, Koin 12.-13.11.1986 PWR plants)- 5 Heuser. f.: Fachgesprsch der Gesellschaft f0r Reakscruc:herheit, From research done in the recent years, especially withm Kein.12.-13.11.1986 ~ phase B of the German Risk Study [4),it is apparent that n ee- 6 Hassmann. K.: Hosemann 1. P.: Consequences of Deg aded Core der to prevent core degradation Accidents. Nuclear Engineering and Design 80(1984i.T Kerntechnik 50(1987) No.2 89
l K. Geyer et al.: Accident raanagement TsNe 1. Safety objectives. RtJared plant parameters and mmurrum rr- Table 2, Core meltdown. Charactenstic tsmes for KWU perssaced gereneentsfor K WUprenurued water reactors m<ater reactors i S4er) Obyecuves Related Plant Parameters and Time af*er start of acciden Minimum Requirements i Low pressure high presure
- Neutron Flux < min core meltdown core mehmem
- Neutron Flux Periode not posiuve
{
, h b
- Control Rods inserted E N 1 ; Water level reaches core upper > 0'7 > 2a
- Boron Concentration > min iedge Core Coolmg { Start of core meltdown > 1.1 > 2.1
- RPVCoolant Inventory > min Start of reactor pressure vessel ,3 , 3 ~.
e Pnmary Coolant Inventory . Borated Water Tanks, Level < min ld'U#
- core and reactor outlet tempera-
~
, tures: 9
" < max 1, falling below of the safety objectives cannot at all or only partsally he ful-
~ 'I e Pnmary Heat Transport T Sec Sat + AT when semod- j filled. This will essentially depend on the degree of ctre de-
" Ymax en es ual Heat I struction. Three stages have to be regarded (see Fig.2):
Removal System in operauon e
- Thefrst Stage is characterized by the fact that damurr .
the fuel assemblies is still negligible but that the safeb.# lain Steam Pressure of all SG < jective oriented emergency procedures of the opr rmg 4 e Secondary Heat Sink
- Main Steam Pressure of 3 (N-1) manual are not effective. The minimum requirements for
- SG falling < max 2 (depends on e the safety objectives are the same as under design-basis condenser availability) '
conditions (see Table 1). The information required ty the
- SG levels of E (N 1) SGs > min shift personnelis assured by the accident proofinstnmen-and < max tation provided for design-basis accidents.
S " " I ~ e At the second stage the core is assumed to be partialy de-e Feedwater Supply fo co# ntrol ng graded but the remaining core geometry is still su5o,ent
- Feedwater > min before cooling i for residual heat removal. It can no longer be assiimni that the instrumentation inside the reactor pressure vesselis op-
- Reactor Coolant Pressure: erable; however, information on the other plant campo-w l e Reactor Coolant "1 3 p ke[cadary Hear,
" < max 2. when Residual Heat r nents and systems remains available as at stage 1.
The objective of accident rnanagement at the fm' t and Pressure Limitation Removal System in normal op i second stage of core degradation is to prevent or sig core
. erauon , melting and to ensure integrity of the reactor sesselespe-
< max 3, when Flood / Sump ; cially-operanon e To activate any kind of injection into the reactcr pres-berrniv Control sure vessel. Injection must be performed at a minmum l t imnnoc of Activity
- vent stack activity < min rate to prevent the expected production of hycrogen i Release to the
- total waste water activity < min from becoming excessively high at low injection noes.
Emwonment
- SG tube leak in 51 SG e T) limit the quantity of hydrogen formed inside the con-
- Cont. pressure < desir" Pressure tainment if necessary.
- ContAemperature < desi n temp. e To actuate containment isolation as a precautionary Comamment integrity
*Q"y5I,ntb"on <p te,,i requ measure unless this has already been achieved by auto-
- prevention l mitigotron -
1st stage l safety oDjectwes l n rn des y T;e safety objettree-e e"et I pac:edves of **e cw 7 I maiunt are not et'r*wt { g i N safetyf ulfooyethres dled / ,, NN ~ l 7,3 saf ety o0;ett'ves I stop core melt ms de l reactor pressare vessel g N safety objectives fulfilled / /l l 3rd stage
,,, I safet, oo;etens l
keep radicottmty artside l containment l l \ setety oD ettivesJ , j } l \ f ulf dled / res as l I procedures for longterm venhng' of f sde ear 7e tv { fig.2. Stages and safety objectives of accident tongte,. .t, conor.n n.com32st,on ,,soonse ees y namagement l e3 Kerntechnik 50(19F No.2 L - __-_-_
1 K. Geyer et al.: Accident mmemm i I i q@te' (onditaff 3r Del.;r 005 I &ncee N occident can'rst by avsmats atten snehetsen Of labitsiert guld reatter protector sysree j )* h theth 54f ely 80 ft*19ss bit et oriented diagnosis Ond fun (f.oning si sa'e*) syst ems l sof tts 00pettnts %tfi.et l Fel 8e l _ J. Indenhf y attraent by , at(ident decision tree
** e ,
8(Esdent igenhfit c hen poss;ble yes no l u '
,1 ,
d . J. l f evenboriented safe'y at te(twe-erentet ; frocedures arstedures { jll take the plant inte o longterm safe condit on take its planf wrz a longtern saf e tenahen i O dk ljl safety objectives fulfilled safety selettnes fa%eS{ A l yes l *o , res l M [ i l 'l 1 l j; !I - i ! l OttioW identihtrer I g
** 5 5* ' I h, I, proceoures l ., j, j Fes l m Q {j! take the plant init c l ,l Longterm safe (Mdition l
. 0: b
! ,! safetr cbjectnes fulfdled l I yes no l
e s;
! offsite emergeacy j response actions ,
FWL Emerg procedure gaadehme fa , necessary i KWUpre.tsume waterreaaors t i plement each other, and that t ansitions from one to the other e The performance of the intended manual actions is ren- f are possible at any time, dered difficult became of unfavorable ambient conditions. ' e The intended manual actions might cause increased release 3 Accident management of radioactive materials. . 3.1 Cseria Due to the importarre of the events, these accident man. agement measures do not require consideration of certain it must be possible during aD esent sequenet.. npecially on boundary conditions whch have to be observed under design escalation into beyond-design-basis ewats, to i,tilize all safety basis conditions, such as the strict process and physical sepa-reserves of the plant. To a cenain extent this is done already ration of the individual r:dundancies. in the safety objective oriented procedures. Accid 9nt manage-ment extends them by measu*cs which are characterized by 3.1 Safety objectives , one of the following features: o Configurations and operabonal modes of systems are cho- For emergency procedtres tsee Section 2), the relevant safety sen which are permitted by the sprem design, but which parameters were arrangd hierarchically and assigned to a were not originally envisand for pLmt operation (includ- manageable number o'sa ety r objectives (see Table 1). Of ing manipulation of proteane interioeks). course, the definition of:hese safety objectives is also valid for o Manipulation of safety systems or safety instrumentation beyond design-basis events It is necessary, however, to place and control systems is necessary (e.g, deactivation of the different minimum requnments on the safety parameters un-intended function of the reactor pfutection system). der accident managemest conditions. It is possible that some Kerntechnik 50(1987) No.2 87 l _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ l
~ - K.Geyer et al.: Accident management C Carl Hanser Verlag. MQnchen 1937 K. Geyer, O. Gremm, U. Krugmann and H. Roth-Seefrid Accident management measures for KWE nuclear power plants dure Guideline illustrated in Hg.1. His guideline indicates j for beyond-design-basis ever.ts, the operating personnel can the sequence of actions to be taken by the shift personnel in ;
make use of the existing plant equipment in order so prevent es- the course of abnormal events. calation ofsuch events into uncontrolled condmons or to limit As a first step, the operator is instructed to find out the sta-the consequences ofsuch events. De proceduresfor such manu- tus of the plant. By briefly checking the safety objectives (see al actions are called accident management procedures. Dey below) and the proper functioning of the safety systems he have to be consistent with the emergencyprocedures ofthe oper- can make sure that the plant has already been brought into a ating manualfor upset conditions and design basis acddents. safe condition as a result of the automatically initiated ac-Accident management measures funher reduce ne extremely tions. He then has to make a diagnosis of the event and the low risk of Kraftwerk Union (KWU) nuclearpowerplants to a longterm measures to be taken. ne diagnostic aids, such as
. negligible extent, even without consideration ofogsite emergen- the accident decision tree of the operating saanual[1] and the l cy response actions. CRT displays of the Process Information System PRINS [2]
assist him in performing this task-Accident-Management-Mallnahmen fur KWU-Kernkraftwer. De extensive automation designed into the plant for the Le. ETir Ereignhabidufejenseits der Auslegung kann das Be- initiation of accident control measures ensates that, even un-triebspersonal die vorhandene Anlagentechnik nucen, um durch der accident conditions, the operators have enough time to in. Handeingnfe die Ausweitung solcher Ereignbabuufe su unbe- form themselves about accident sequence and plant status, to herrschten Zustunden :u verhindern oder die Ausurkungen sol- considet the possibilities for long-term accident control, and cher Ereignisabidufe :u begrenzen. Die AnwetsmagenJDr diese to take appropriate measures. Handeingnfe, Accident Management-Pro:eduren genannt, massen konsistent sein mit den Anweisungen des Betriebshand. 2.2 Eventorientedprocedures buchsfDr Ereignhse bei anomalem Betrieb und bes Auslegungs-
, sidefdflen. Die Accident Management Mapnahmen reduzieren Normally, the special procedures defined for individual das ohnehin schon extrem geringe Rhiko von Kernkraftwerken events in the operating manual can be identiRed with help of der Kraftwerk Union (KWU) auf einen rernachidssigbaren the accident decision tree. In addition to mastructions on the Wert, selbst ohne BerGcksichtigung von anlagenerternen Not- measures to be taken, these procedures indude instructions fallschutzmg#nahmen. for continuous observation of the safety ob.jectives. He ope-rating personnel is provided with explicit ieructions for han-
. dling standard situations, which are onenard towards selec.
1 Introduction tion of the optimum method of controlling the plant. In the event of deviations from standard situations, e.g., he safety systems of Kraftwerk Union (KWU) auc!cas power failure of further components, the specification of both the plants make beyond-design events with severe consequences objectives of the individual actions and of the reasons for to the environment very improbable. However, even ifit is as- them allow the shift supemsor to foDow the event oriented sumed that all safety systems fail, there still rernains ampic procedures as long as the safety objectives are fulfilled. time for manual actions of the operating personnel e to prevent melting of the core or 23 Safety-objective orientedprocedures e to limit core-melt consequences to the emironment. His has been shown by extensive research done in recent In case an accident cannot clearly be associated with one of years. the event oriented procedures desenbed in tne operating man. De procedures for such manual actions of the operating ual, the Emergency Procedure Guideline instructs the shift I personnel, in the following called accident unanagement personnel to follow the safetrobjective oriented procedures; procedures, have to be consistent with the emergency proce- i.e., if one or more of the plant parameters assigned to the i dures of the sperating manual for upset conditions and de- safety objectives (see Table 1) run outside their limits, they { sign-basis accidents. For KWU nuclear power pl. ants,the con- have to be brought back into the safe range by taking safety- i sistency is achieved by the overall engineering responsibility objective oriented measures. Accident control by taking safe- { of KWU and its long-term experience with turnkey supply. It ty-objective oriented measures also ensures that the plant is 1 should be noted that in this paper the term accident manage- brought into the long term safe condition. f ment does not include offsite emergency response planning. It is apparent at the stage of observation of the safety objec- I tives during the status oriented diagnostic phase of accident 2 Emergency procedures for upset conditions and control that one or more of the safety objectives are not ful-design-basis accidents filled, the operator is instructed by the Emergency Procedure 2.1 General Guideline to adopt the safety-objectin oraented procedures directly. In KWU plants, the emergency procedures for upset condi- It is evident from the above that event oriented and safety-tions and design-basis accidents follow the Emergency Proce- objective oriented procedures do not rule out but rather com-86 Kerntechnik 50(1987) No.2
W. Braun: Safety must not depend on humat rehability C Carl Hanser Ye .At. Monchen 1937 W. Braun Safety must not depend on human reliability ... s De following nine papers are con- huma factors influencing reliability during its constner- on site. He cerned with various aspects of the inter- and safety of nuclear installations: the answer to this quemica is quality assur-face between man and machine in nu. entre safety, including design, fabrica- ance. clear installations. This is a very com- tior, construction and operation, must At least for wate cooled reactors in , plex field, and no more than a few not depend on human reliability. the Western wodd,ine safety related en-aspects can be mentioned in this intro- Oeviously, maloperation by " press- gineering knowledge h sell established duction. ing the wrong button" has to be, and and so epenly exchmg:ed that hitherto It can be observed that the majority defrmdy is, excluded by automatic in- undetected basic enors and not yet real-of accidents and incidents in the nuclear tedocks wherever such faults may be ized gaps of knowledge may be ex-field as well as in any other technology safey-eelevant. But also the actuations cluded. %e layom and the design of has not been caused by breakdown of whrt are immediately necessary to safety relevant equpenent, systems or l hardware but by human failures. Both brirg an unscheduled transient under procedures of a nudear power plant are the Three Mile Island accident in 1979 control and to mitigate its consequences not only documested and counter-and the Chernobyl disaster in 1986 were must not be left to the control room per- checked internally but fully analysed or rnainly caused by human malbehaviour sornet Though normally such tran- even recalculated be an independent in-and maloperation. In less significant nu- sietzs develop not quickly but in mi- spector organization or the licensing au-l cle r incidents, maloperation played a nuus, the personnel need some time to thority. The same holds for specifica-role as well. Transients could have been ana$yse the situation, identify the acci- tions. manufacturing and the test proce-stabilized at an earlier time had the per- dem, and decide upon suitable proce- dures for safety-re1xed pressure-loaded sonnel reacted with more promptness duns. Only then intelligent measures hardware. The whcie pc'ocess of manu-and intelligence. With regard to the fi- can be taken. As a rule, therefore, no facturing is observed cruically and spec-nanciallosses caused alone by unsche- safay-relevant manual actuation must imens are tested 3b ndependent quality dulid outages of large nuclear power be accessary in German nuclear power control groups and professional inspec.
,, plants, not to speak about the possible plaats for at least the first 30 minutes. tors. Especially the tesming of all welds consequences of a radioactivity release GeneruBy, in order to exclude human and the evaluation a(pressure tests are i tecident, the problem of minimizing ma3 operation and in order to guarantee carried out by up to three inspectors maloperation and its consequences is of the fuEinment of the fundamental safety working incwpendently. [
vital importance to nuclear power. The ob.iecdves, the safety systems have to be By this network of qmality asurance ' sceptical public, easily enough to under- actsated automatically. In addition, to and quality controlhurman failures and stand, ask for " error tolerant" technolo- prori& for intelligent actions of the per- negligences during desgn and construc-gies and " forgiving" reactors. sonnel the presentation of the process tion can be broughtu a minimum num-
, The following papers deal with infortcation should be further opti- ber and to very snan extents. It has aspects of human factors during normal mind s.o as to allow for quick identifica- been made sure thz the unavoidable re-and abnormal operation. It is shown tiot of the accident and for understand- mainder of faults is far from being able how by improved training, by more er- ing of the transient, its history, and its to cause a dangeroui failure.
gonomical operational manuals and progreu Of course, automatization of We can certain13 neicher accomplish procedures, by more intelligent process safery functions and modern informa- a flawless technolorv as should we as-information, by diagnostic systems, and tior systems help to improve also nor- sume an error free ope:rator..However, by many other refinements, the failure ma' pumt operation and assist in in- esery responsible ergineer should make rate of the operating personnel can be creasic:t its reliability. his product as failtre-elerant as possi-further reduced and the mitigation of Considering not only the safe opera- ble and optimize tie interface with its transients and accidents can be im- tiot of a nuclear plant but also its many user. We are convnced that recent nu-proved. However, safe operation and ef- othe s.afety related aspects, the question clear power plants n the Western world fective mitigation ofincidents under the arius Y and how human error is ex- are very close to thse goals.
- influence of not absolutely failure-proof clut.ed m the design of the plant, during human beings is not the only aspect of ma:ufacture of its components, and Kerntechnik 50(1987) No.2 85
Autorenfortdruck aus der Zeitschrift ,'*-l
- ~~
/ A ~' "
9- ~' ATOMKERN- -
~
i T ENERGIE J__ _ A__\__ Independent Journal on Energy Systems and Radiation l 1987 Carl Hanser Verlag M0nchen l l l
**=- l Alle Rechte. auch die des Nachdruc,ks der photomechanischen Wiedergabe d:cses Autorenfortdrucks und der Ubersetzung. behalt sich der Verlag vor
f
q,, UNITED STATES NUCLEAR REGULATORY COMMIS$10N
.[
.g
.,.. g j WASHINGTON, D, C. 20666 k .... / DEC 9 1986 MEMORANDUM FOR: Robert M. Bernero, NRR Richard W. Starostecki, IE i Richard E. Cunningham, WSS Denwood F. Ross, RES Clemens J. Heltemes, Jr., AE00 Joseph Scinto, OGC l
FROM: James H. Sniezek, Chairman Comittee to Review Generic Requirements
SUBJECT:
CRGR MEETING NO. 104 The Comittee to Review Generic Requirements (CRGR) will meet on Monday, J December 22, 9 a.m.-1 p.m. in Room 6507 MNBB. The agenda is 'as follows: 1 9 a.m. - 1 p.m.,- R. Bernero (NRR), will present for CRGR review, a pro-posed Generic Letter concerning improvements to BWR Mark I Containments. NRR will provide a copy of the proposal to each member and eight copies to Walt Schwink l by December 12, 1986. (Category 2 item.) If a CRGR member cannot attend the meeting, it is his responsibility to assure that an alternate, who is approved by the CRGR Chaiman, attends the meeting. Persons making presentations to the CRGR'are responsible for (1) assuring that the information required for CRGR review is provided to the Comittee (CRGR Charter - IV.8), (?) coordinating and presenting views of other offices, (3) as appropriate, assuring that other offices are represented during the presenta-tion, and (4) assuring that agenda modifications are coordinated with the CRGR contact (Walt Schwink, x28639) and others involved with the presentation. Division Directors or higher management should attend Metings addressing ' agenda items under their purview. In accordance with the ED0's March 29, 1984 memorandum to the Comission con-cerning " Forwarding of CRGR Documents to the Public Document Room (POR)," the proposal which contains predecisional infonnation, will not be released to the 1 PDR until the NRC has considered (in a public forum) or decided the matter addressed by the information. 9 - wdh ames H. Sniezek, Cha rman omittee to Peview Generic Requirements cc: See next page , h -f(4 9 A ft R g - -
m
.y DEC 9 1986 k
7 cc: SECY . Commission (5) V. Stello, Jr. Office Directors Regional Administrators
. W. Parler Distribution:
JSniezek JRoe RFraley. TPehm ROGR Staff DEOROGR cf JZwetzig BZalcman MLesar FHebdon WLittle RErickson Central File PDR(NRG/CRGR) JClifford JZerbe PPabideau JMurray W0lmstead JNorberg 4
\
l
- ROGR l
- POGR
- DEDROGR : :
7.....:.. -
- WS
- ...h...mk ,
- J er e j
- JSni ...:............:............::.__.........::...........
- 1?/ 9 /86 :
12/q/86 :12/'7 /86 : : : - 0FFICIAL RECORD COPY
..~ . ? /% umino : Tarts NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20005 \, DEC 8 1986 FEPORANDUP FOR: Robert M. Bernero, NRR Richard W. Starostecki, IE Richard E. Cunningham, NMSS Denwood F. Ross, RES Clemens J. Heltemes, Jr., AEOD Joseph Scinto 0GC FP0M: James H. Sniezek, Chairman Committee to Review Generic Requirements
SUBJECT:
CRGR MEETING N0.103 l 1 The Comittee to Review Generic Requirements (CRGR) will meet on Monday, December 15, 1-5 p.m. in Room 6110 MNBB. The agenda is as follows: 1 l 1:00 - 2:00 p.m. - G. Arlotto (RES) will present for CRGR review the proposed Regulatory Guide entitled, "EE 404-4, Draft 2. Qualification of Connection Assemblies for Nuclear Power Plants." (Category 2 item.) CPGR deferred review of the proposal at Meeting No.107 until Peeting No.103. A copy of the proposal was enclosed with the announcement for CRGR Peeting No.102. ! 2:00 - 4:00 p.m. - J. Murphy (RES) will present for CRGP review the I proposed draft NUREG 1150, Reactor Risk Reference Document (Main Report and Appendices). A copy of Dr. Ross' memorandum concerning this matter is enclosed. (Please note: RES has forwarded a copy of NUREG 1150 directly to each member. Icategory ? item.) 4:00 - 5:00 p.m. - M. Jamgochian (RES) will brief the CRGR concerning the evolving 10 CFR Part 50, Appendix E revisions. (Category? item.) If a CPGR member cannot attend the meeting, it is his responsibility to assure that an alternate, who is approved by the CRGR Chairman, attends the meeting. Persons making presentations to the CRGP are responsible for (1) assuring that the information required for CRGR review is provided to the Comittee (CRGR. Charter - IV.B), (2) coordinating and presenting views of other offices, (3) as appropriate, assuring that other offices are represented during the presenta-tion, and (4) assuring that agenda modifications are coordinated with the CRGR contact (Walt Schwink, x?8639) and others involved with the presentation. Division Directors or higher management should attend meetings addressing agenda items under their purview. ,
\
emw _ 2P, Y
4
, ~ , .
DEC 8 586
-p.
In accordance with the ED0's March 29, 1984 memorandum to the Commission con. cerning " Forwarding of CRGR Documents to the Public Document Room (PDR)," the enclosure and NUREG 1150 (Draft), which contain predecisional information, will not be released to the PDR until the NRC has considered (in a public forum) or decided the matter addressed by the information. Origina1 signed by James H. Sciezek James H. Snierek, Chairman Committee to Review Generic Requirements
~
Enclosure:
D. Ross memo dated 11/7/86 cc: SECY Commission (5) V. Stello, Jr. Office Directors Regional Administrators W. Parler J. Murphy M. Jamgochian G. Arlotto Distribution: JSniezek JRoe RFraley TRehm j POGR Staff DEDROGR cf JZwetzig BZaleman !' MLesar FHebdon WLittle RErickson Central File PDR(NRG/CRGR) JC11fford JZerbe PRabideau JMurray ) W01mstead JNorberg DFC :ROGR :POGR :0EDROGR :
. . . . . : . . 4 4. . . . . . : . . . . ,.} . . . . : . . . . . . . . . . . : . . . . . . . . . . . . : . . . . . . . . . . . . : . . . . . _ _ . . . . . : . . . . . . . - - - -
- JZ :JSnierek : : : :
NAME : W5ptyfnk e/ l DA TE : 12 /86 / h.
- 17.. 8. /86
. . . . :1?/ : . . . / .P,6. . . . . . . : . . . . . . . . . . . . : . . . . . . . . . . . . : .
i 1 _ _ _ _ _ _ _ - - _ _ _ ___. _ _ 3
[ g e( o NUCLEAR REG A WM WASHING TON, p.c. 20S55 g November 5, 1986 orren or na COMhKS$4oM88 ntMORANDUM FOR: ictor S gjye 0 ector for Operations k FROM: Thomas M. Roberts
SUBJECT:
COMMISSION BRTEFING ON GE/ NARK I CON At the Commission Briefing on the GE 1986, one slide was prese/Ma rk I Containment Program on hc f m the Commission's Severe Accident Policy Statem "In those beyond instances current where the technical l rulemaking added). will be the In oHer ca ses , preferred solutionregula (emphasis tory re ( of through e the conventional practice of issuingthe isfue s ; Bulletins and modifications areOrders justifiedor throughGeneric Letters where beckfit { i policy, or of the Integrated Safety Assessmentthrough conception." e lines plantI Program (ISAP) At that )
\
the preferred approach to resolving its conctime containments. This rulemaking as I aske \' ensuing discussion. ques tion was not specifically answerederns with the Mark I Therefore, I would again in the the s taf f's ra tionale for not pursuing rulemaki like to ask, "What is 1 ( solution which clearly (ingoaccordance beyond currentwith Commission ngPolicy) as the preferred to add ! affect a class of plants?" ress issues regulatory requirements and which cc: Chairman Zech i l Commissioner Assels tine ' Commissioner Bernthal Commissioner Carr SECY OGC H. Denton R. Bernero k}}