ML20127A464
Text
{{#Wiki_filter:_- 3 s ~,s. [4 JAN 2 7 f983 NEMORANDUM FOR: Brian W. Sheron, Chief, Reactor Systems Branch. OSI THRU: James F. Meyer, Reactor Systems Branch OSI FROM: W. Brad Hardia, Reactor Systems Branch, OSI
SUBJECT:
SumARY OF MEETINGS WITN ORNL AND NRR CONTRACTORS TO DISCUSS LIMRICE PRA, SPECIFICALLY, puennerurunafCAL, ACCIDENT PROGRESSION, AND CONTAI M ENT LOADING AREAS t On Movember 17, 1982, NRR and RES and with ORNL and BNL to discuss current ORNL ac area of severe accidents for BWRs, in particular with regard to the current PRA assessment of Limerick. OmL presented a brief summary of their activi-ties including their work on the MARCN code improvements being performed under contract to RES. At the conclusion of the meeting, it was agreed that ORNL would perform a review of the backend portion of the BNL draft report on the Limerick PRA (NUREG/CR-3028, October 1982) and that a followup meet-ing would be held to discuss ORNL's comments. It was also agreed that ORNL would perform independent calculations of radionuclide releases for Limerick, to be completed in February, and that ORML staff would brief SNL staff on the MARCH-2 code after its completion. It was noted that the intent of this re-view of the SNL work was to aid the NRR staff in evaluating "best-estimate" accident scenarios for the staff effort on mitigation features for Limerick. Enclosures 1 and 2 include a list of meeting atteMees and a copy of a meet-ing handout provided by Steve Hodge of ORNL. On December 16, 1982, a followup meeting was held 114 8ethesda to discuss ORNL's comments on the draft 955 report. In addition to their review com-nents. ORNL staff made a series of presentations on the MARCH code improve-ment program and related work in the severe accident phenomenology area. Enclosures 3 and 4 include a list of meeting attendees and copies of hand-outs provided by ORNL. From these meetings, it is concluded that there is a substantial resource of core melt accident progression skills at 0WL that could be of important use to MRR in the future. It is strongly rae===a= dad that NRR staff main-tain contact with ORNL, through the appropriate RES program managers, and consider ORNL's further participation in future NRR BWR severe accident effbrts. qrg Wnalsep . Brad Hardin, (S30 2 l 7 0 l21 Rea tor Systems Branch, DSI
Enclosures:
As stated RSB:DSI RSB:DSI RSB:DSI RSB:DSI RSB:DSI BHardin:cs JMeyer WHodges NLauben 1/ /83 1/ /83 1/ /83 1/ /83
\\ JAN 2 71983 Brian W. Sheron cc(w/anc.): E. Che111ah T. Pratt (8E1)fB.%) I. Papezoglow W. Kastenberg I,UCLA)
- 7. Theofanowt IAntae)
I. Catton (CCLA) D. Stenson (A*4) cc (w/o anc.): R. Mattson T. Spels A. Thadant F. Cofflan M. Mattit-R. Martin S.Hodge(0?&) W. Butler DISTRIBUTION Docket File RSB Rdg. RSB Plant / Subject BHardin JMeyer WHodges NLauben Hardin Rdg. I e i I t 5
-. g f *%q'o,% ./ \\ UNITED STATES NUCLEAR REGULATORY COfdMISSION { j wasumaros, p. c. aom \\,*****/ jg 27 %@. MEMORANDUM FOR[ Brian W. Sheron, Chief, Resctor Systems Branch, DSI + THRU: k F. Meyer, Reactor Systems Branch, OSI FROM: . Brad Hardin, Reactor Systems Branch OSI
SUBJECT:
SUMARY OF MEETINGS WITH ORNL AND NRR CONTRACTORS TO DISCUSS LIMERICK PRA, SPECIFICALLY, PHENOMEN0 LOGICAL 4 ACCIDENT PROGRESSION, AND CONTAINMENT LOADING AREAS 'On November 17, 1982, a meeting was held in Bethesda with staff members from -NRR and RES and with ORNL and BNL to discuss current ORNL activities in the area of severe accidents for BWRs, in particular with regard to the current _ mj PRA assessment of Limerick. ORNL presented a brief summary of their actici- . ties. including their work on the MARCH code improvements being performed under contract to RES. At the conclusion of the meeting, it was agreed that ORNL would perform a review of the backend portion of the BNL draft report on the Limerick PRA (NUREG/CR-3028, October 1982) and that a followup meet-ing would be held to discuss ORNL's comments. It was also agreed that ORNL staff would brief NRC and BNL staff on the March-2.b code after its completion. -It was noted that the intent of this review of the BNL work was to aid the NRR staff in evaluating "best-estimate" accident scenarios for the staff effort on mitigation features for Limerick. Enclosures 1 and 2 include a list of meeting attendees and a copy of a meeting handout provided by Steve Hodge of ORNL. On December 16, 1981, a followup meeting was held in Bethesda to discuss ORNL's comments on the craft BNL report. In addition to their review com-ments, ORNL staff made a series of presentations on the MARCH code improve-ment program and related work in the severe accident phenomenology area. Enclosures.3 and 4 include a list of meeting attendees. and copies of hand-outs provided by ORNL. From these meetings, it is concluded that there is a substantial resource of core melt accident progression skills at ORNL that could be of important use to NRR in the future. It is strongly recommended that NRR staff main-tain contact with CPJil, through the appropriate RES program managers, and consider ORNL's furthar participation in future NRR BWR severe accident efforts. Esod b x W. Brad Hardin - Reactor Systems Branch, DSI enclosures: Ar. stated J' cc: See Next Page c ,,n-,-,-- r - nn -~-n-.,~.---,.-.---,,-,---.n-.-n -- ~,-,--,- - --.-,--, - ,.n-
Brian W. Sheron. JAN 2 71983 cc (w/ enc.): E. Chellfah T. Pratt (BNL) I. Papazoglou (BNL) W. Kastenberg ((UCLA) T. Theofanous Purdue) "I. Catton (UCLA) D.Swanson(ASA) cc (w/o enc.): R. Mattson T. Speis A. Thadani F. Coffman M. Mattia R. Martin S. Hodge (ORNL) W. Butler m. e e l i e l l I l p (__ s h
t
- c. ~
ENCLOSURE 1 MEETING ATTENDEES - LIMERICK PRA, NOVEMBER 17, 1982 R. Curtis RES B. Agrawal RES J. Meyer NRR B. Hardin NRR S. Hodge ORNL - T. Pratt BNL e. 4 0 i l i i y .m,-..,- ---v-r
h h t\\C 21QT S. A. Hodge ORNL - SASA Nov. 15, 1982 BOILING WATER REACTOR PRAs Initiation / Plant Type Completion Sponsor Oyster Creek BWR 2 MK I 1977/ Unpublished Utility Millstone 1 BWR 3 MK I 1980/1982 NRC (IREP) Peach Bpetom 2* BWR 4 Mk I 1972/1975 NRC (RSS) Browns Ferry 1* BWR 4 MK I 1980/1982 NRC (IREP) Su?) }#4t. 1980/1983 Utility y ,g Limerick * ,ilWR 4 MK II 1980/1982 Utility ) .r 1 ~ s.4 Shoreham*hYM BWR 4 MK II 1981/1982 Utility i 3 '[J [r ' v., - Susquehanna BWR 4 MK II 1980/1982 Utility Ia Salle BWR S MK II 1980/1982 Utility Grand Gulf BWR 6 MK III 1976/1981 NRC (RSSMAP) GESSAR* BWR 6 MK III - /1982 GE Held by ORNL/SASA. gee-e- -e e smyeame e. ese-ps see ,geg***=* + *=
g '. S. A. Hodge ORNL - SASA Nov. 15, 1982 ACCIDENT SEQUENCE IDENTIFICATION CODES A Large Break LOCA $1, 52 Small and Intermediate Break LOCA TQUV Scram with Loss of injection to Reactor Vessel W Scram with Loss of Ability to Cool the Pressure Suppression Pool ATWS Failure of Control Rod Insertion Loss of DHR TV Sequence including Cases with Loss of Offsite Power and Stuck-Open Relief Valve. l 1 L i l 1 -. n. ;;;...
+ h 8 - s e -1 6 l;' J* ,ji
- j T
\\\\ f Gnerrec i- )' M'coid1,7,$$ IREP - BROWNS FERRY,Q, p WASH 1400 t_ (3X10-5/ REACTOR-YEAR) (2X10-4/ REACTOR-YEAR) Ord dc( -O a P % ~. ... TL E Y ~ 2Vm-S ~ j CORE MELT FREQUENCIES FOR BWR 4 MK I PLANTS e I
4, LIME RICX MEAN - 1 A a 10-5/ REACTOR YEAR THER OTHER TOtJV TOUX T OUX p SJ^* 3 2 EWNM ATW5 E T EVENTS y WASH-1400 MEAN - 3 x 10-5/ REACTOR YEAR sv i TQUV l Tw ATWs I y49 t C6 COMPARISON OF LIMERICK AND WASH-1400 RESULTS l .~-.n.7-,-g,,..,...- -~
'.et S. A. H dga ORNL - SASA LIMERICK DOMINANT SEQUENCES hri @cd (Frequency > 1 x 10~) s 1. T The no-coolant injection acci-E dent sequence with ' fast bollof f Tq V and core uncovery while the con-F tainment is intact (Class 1). -5 > U l.136 x 10 TE T E M d 4 2. T' V 10RV and no coolant injection -7 U 8.5 x 10 accident with fast boiloff and X T C' core uncovery while the con-g tainment is intact (Class 1), but with a high-temp PSP which has an impact on fission product transport. 3 T Stuck-open relief valves cause T overheating of the PSP. -There T -7 is-failure of-heat removal. A F [) PW(P) 5 5 x 10 Class ll accident (containmen t TEI fails first). I 1i 4. T ATWS to 30% power and stuck-T C PU 4 2 x 10-7 pen relle valves. No HPCI or 2 M T FW injection. A Class ill event. F, (ATWS + bolloff and melt with containment intact). 2' 5 T MSIV closure or LOSP, ATVS to p -7 30% p war, n HPCI or FW iniec-3> C UU 2.6 x 10 T N R tion. A Class til event. e i 2 6. T MSIV closure or LOSP, ATWS to p ) -7 30% power, overpressure of con-T,3 1 9 x 10 M 12 tainment. (A Class IV event, assuming that the HPCI high ex-haust pressure trip is at 150 psig, higher than the contain-ment failure pressure).
s P -8
- 16. TfW l.6 x 10 LOSP with containment heat removal unavailable. A class 11 accident r
V -8 17 S qu 1.3 x 10 Medium LOCA, FW and high pressure i x systems unavailable, failure to depressurize or low pressure ECCS unavailable. A class I accident -8
- 18. T C 4W 1.2 x 10 ATVS with RPT, no HPCI or FW, no pg 12 RHR.
This is a class ill accident 9 Generic accident sequence classes: Core Class melt At core melt, containment Ty e i i Fast intact, Icw pressure TQUV, SBLOCA + TQUV 11 Slow Failed Loss of DHT Ill Fast intact, high pressure ATWS Large LOCA TV - TQUV IV Fast Failed ATWS, Loss of CHR
i S. A. Hodga ORNL -- SASA Nov. 15, 1982 DESCRIPTION OF DOMINANT SEQUENCES (Limerick) 1. Each initiating event for this group of sequences provides that the power conversion system (PCS) and the feedwater/ condensate system are unavailable: TE= Turbine trip with loss of offsite power (LOSP). Tq= MSIV closure / loss of FV/ loss, of condenser and the p FW and condensate systems remain unavailable. T Q =. Turbine trip with PCS and conds/FW unavailable. T Tq= Manual shutdown with PCS and conds/FW unavailable. g During the sequence, the high pressure injection systems are not available. HPCl/RCIC unavailable (4.9 x 10~3) U = t and the low pressure ECCS systems do not perform their function because of Mechanical failure of ADS valves or ' failure V = of the LPCI and core spray pumps, or failure of timely manual ADS actuation by the X = operator. 2. The initiating event is an inadvertently opened relief valve at 100% power, which preheats the pressure suppression pool before the scram. T C' represents the case where the operator does not g scram the reactor before the pressure suppression pool tempera-ture requires prompt RHR operation. The complex U is the same as in (1) above. j - X i l l l .; ; :;c; ; :; ;- - : v - -~ - - ~
4 2 3 These sequences represent shutdown, stuck-open relief valves, and failure of heat removal to the environment by either the PCS or the RHR system. 4. ATWS following either turbine trip with bypass' or MSIV closure events. The RPT is effective, so the reactor is brought to 30% power. Multiple SRVs fall to close and there is no high pressure - Injection (except RCIC). 5 ATWS with RPT to reduce power to 30%, no HPCI (U), no RCIC (U } ' R MSIVs closed. 6. ATWS with RPT to reduce power to 30%, SLCS operation, but no con-tainment heat removal. HPCI continues to run, keeping the core covered. Containment failure occurs before core melt. i 7 Scram with loss of PCS, FW, and condensate. Failure of containment ~ ' " heat removal, including the case with stuck-open SRVs. 8. MSIV closure or turbine trip with bypass and mechanical failure of RPS. The rectre pumps trip, but no SLCS pumps operate. '9 10RV, scram, HPCI or RCIC available but no RHR pumps available for containment heat removal and no PCS available. l 10. 10RV, RPS mech falls, SLC successful, but no HPCI or FW available. Rapid bolloff leads to class ill event. l l r l 11. Large LOCA, scram, LPCI and core spray, successful long-term recirculation of water, but no containment heat removal (and PCs not available -- see p. 3-36). i I [ 12. LOSP, SRVs fall to reclose, HPCl/RCIC not available, failure of ADS or LPECCS not available. l
- :.g,*
7 - -
- y." *
, Lt:r ~, + -" r - -- ~~ ~
3 13 Turbine trip with bypass and subsequent MSIV closure (see p. 3-72) or Initial MSlV closure. Mechanical failure of RPS. Recire pump trip. Multiple SRVs fall to reclose. HPCl available, SLCS operates, failure of containment heat removal. ' Result is Class ill 80% (HPCI lost before containment failure due to high back pressure trip) and class IV 20% (HPCI continues to run). 14. MSIV closure / loss of FW/ loss of condenser, SRVs fall to reclose, FW/conds remain unavailable, HPCl/RCIC not available, ADS not actuated or LPECCS not available. Similar to (I), except have 50RVs.
- 15. SBLOCA, scram, high pressure systems not available, no depressuri-zation. Break size is <0.04 ft2 (IIquid).
- 16. LOSP, HPCI or RCIC available, but PCS and RHR/RHR$W unavailable.
2 9
- 17. Medium LOCA 0.004 < break <0.1 f t IIquid so that RV is not de-
,0.0'16 < break <0.08 f t2,g,,, pressurized. FW is assumed unavailable (see p. 3-58) and the reactor scrams. HPCl'is unavailable. There is a failure to de-pressurize via ADS or a failure of the low-pressure systems.
- 18. MSly closure, failure of RPS mechanical, RPT. SLCS operates.
HPCI and FW unavailable. RCIC operates until high exhaust back pressure trip. No containment heat removal. Slight possibility of a class. lV accident if RCIC continues to operate. 6 . 3. ,e ~
S. A. Hodge ORNL - SASA Nov. 15, 1982 SHOREHAM PRA Frequencies of Core Vulnerable Conditions
- Frequency of Core Vulnerable Accident Class (per Reactor Year)
Loss of Coolant Makeup 4%i 2.7 x 105 Loss of. Containment Heat Removal 1.1 x 105 . ATWS W/0 Poison injection 6.1 x 106 LOCA 3.6 x 10~7 LOCA Outside Containment 2.0 x 108 Total 4.4 x 105 From Draft - Preliminary Copy. This PRA has not been released. e 4 p
- 3 w9-g *.
%4
- '~q P
9'
- b
~.,,_ .. ~ - -. -,.. -..
GESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMATION-Class III t Table 3.4-1 ASSESSED FREQUENCY OF CORE DAMAGE PER PLANT YEAR Frequency Representative Cauce of Loss (Event Class Initiating Event of Core Cooling der Year) g Transient w/ loss of Initiated Event 4.6x10-6 core cooling followed by injec-tion systems failure I SB or IB m 1x10 d SB/IB w/ loss of core cooling -10 I LB LOCA w/ loss of 7x10 g core cooling I -0 III ATWS w/SLC w/ loss 9x10 of core cooling 4 V Ex-containment LOCA 2x10 ~ w/ loss of core cooling s.. II Mas of heat removal Mss of containment 2x10 g following no Scram integrity followed by injection systems failure II h as of heat removal 3x10-10 g following a Drywell LOCA ~8 II Loss of heat removal 2x10 T following a transient ~0 IV ATWS w/o SLC but w/ 5x10 core cooling -10 VI Containment LOCA w/ 5x10 failure of vacuum breakers or containment spray 4.7x10~6 6 W 15.D.3-67/15.D.3-68 ..,. ~.. ; r;,.. - ~,
- S. A. Hodge ORNL - SASA Nov. 15, 1982 There are ten plants with MK 11 containment in the U.S.
These plants are distributed over seven sites: Shorehavn 820 MWe Brookhaven, NY BWR 4 Nine Mlle Point 2 1080 MWe Scriba, NY BWR 5 Susquenanna is2 1050 MWe Berwick, PA BWR 4 Limerick is2 1055 MWe Pottstown, PA BWR 4 Zimmer 1 810 MWe Moscow, OH BWR 5 La Salle 1s2 1078 MWe Seneca, IL BWR 5 - WNP-2 1100 MWe Richland, WA BWR S If Limerick is eliminated - Not Shoreham or Zimmer because they are atypical "small BWRs". e ~ Not La Salle because Commonwealth Edison not cooperative. e Probably not VNP-2 because too far away. Leaves Nine Mlle Point 2 (BWR 5) or Susquehanna (BWR 4). e w.
ENCLOSURE 3 MEETING ATTENDEES - DECEMBER 16, 1982 B. Agrawal RES J. Meyer NRR B. Hardin NRR E Chellfah DST-S. Hodge ORNL T. Kress ORNL S. Greene ORNL L. Ott ORNL i S. Nienczyk - - 'ORNL T. Pratt BNL 4 I. Catton dCLA l 4 e
6 (Jib 6U.E ' W /S COMMENTS ON COMPLETENESS OF BNL DRAFT NUREG/CR-3028 S. A. Hodge SASA Project _ Manager Oak Ridge National Laboratory Phillips Building l December 16, 1982 N l l
4 s ACTUATION OF BOTH DRYWELL SPRAY NETWORKS MIGHT CAUSE DRYWELL FAILURE IF PSP TEMPERATURE <105'F LIMERICK FSAR SECTION 6;2.1.1.4 = PRIMARY CONTAINMENT DESIGNED FOR NEGATIVE 5 PSIG " ACTUATION OF BOTH DRYWELL SPRAY NETWORKS IS ADMINISTRATIVELY PRO-HIBITED WHENEVER THE SUPFRESSION POOL TEMPERATURE IS BELOW-105'F." .\\ jl \\, i V, W ,r 3.\\ n >e &}f f:6' kv a\\G e ,e -.-*.e-. r-e s
CONSIDERATION OF DRYWELL COOLER OPERATION IS IMPORTANT TO BWR ACCIDENT ANALYSIS f Maintain drywell temperature below ~/ design limit (protects SRy solenoids)h' dO T s.y :.s. :> Lowers containment pressure, Drywell coolers trip on core spray initiation signal and chilled water is lost on drywell isolation - leads to ADS permissive.
- ~
t.~ 9 ' [
A i OPERATION OF CRD HYDRAULIC SYSTEM SHOULD BE CONSIDERED IN PRA INJECTION SOURCE IS CST, NOT AFFECTED BY STATUS OF PRESSURE SUPPRESSION POOL INJECTION INCREASES FROM 60 GPM to 170 + GPM WHEN SCRAM IS'IN EFFECT 170 GPM SUPPLIES ALL NON-LOCA, POST-1 SHUTDOWN INJECTION NEEDS SIX HOURS AFTER SCRAM CST VOLUME (.190,000 GAL) SUFFICIENT FOR 13.5 BOURS AND. TANK CAN BE RE-FILLED. .__...w. r r
... a. ..... :..i.. N e _ M_ A, e.g / 3 S 2*$$.5 .. ;; i.. g 4: 5{!f'g'f.f{ .: A :' E _m ) .g m! 4-J 5 rv -- - - - i g l ~ 8 1 l L_v l }j _ il f l i 14 g j i, I gil o ..r. 1 l all I 3 p i 2 l 1$ L 8 ly g = 3 ] +. I{ ilajj Mu K 6b g m L y ~,8. E 5 l n o o I y g i ] l.- ~ $5f 5 s, l I I I E 1 J I l h L r_ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _, j d .gg a b4 b2 e g1 'tSciis sI izp =2l y I s. n $%hdd$ .] a# i X I i *l Zt is 8U 3 l L c o.1_.. E i ~
g w ...q, i
- z...
- ' * =.
7,' - C##E"E ** BOLATION VALVE ) ERAM VALVE PILOT AIR -O M ' BOLATION VALVE - '~ - is0LATION VALVC wtTtt0RAWAL RISER E3MAUST WATER RtSER r30t.ATIOt VALVE DRIVE WATER rifer 1 ? IRILATION VALW SCRAA8 PtLOT VALVE A33088LT SESENT st:$ER k;.kg .7,,, ;A,,n anowmo u T R riser f l k. / 0 JUNCTION SCX ,p / M j WOLATION VALVE SCRAM 0:3CMARGE riser ,g W [ ggning TROUcM A33csSLY g. OUTLET SCRAM VALVE INTERCCNMCCTING _ BELET 3CRAas VALVE y CA.= r." ? r. } l.. - CINECTf0NAL CONTROLVALVE h !f
- WITHORAWI i
OutECTIONAL TROL VALM 'N I M \\ OtRECTTCNAt. CCitTROL VAkyt r;$ LINSERTI = ' ' " .j. s [t OutECTIONAL CO*tTROL vat.vt . 'Y anTn0 RAW An0 sETTi.13 annorPvAt.n WATER ACCUMULATOR ORAIN 6* - SCRAM ACCUMUuTOR N2 CTun0ER p MRAM WATER ACCUMtJLATUR y h "' " "# ACCUMULATORN2 PRE!!URE d INotCATOR ~ . FRh88E o = y CARTRIOct VALVE l ,ACCBAJLATOR N; CNA4QMG .* s .<p/ s ACCuuULATcR 7 INSTRuwtNTATICM Assgassky Control Rod Drive Hydraulic Control Unit (Component assemt:fy) g g g. O ..,,,,,.py.y,,, 4 - ~
.t 'l SRV RELIABILITY DATA IS OF QUESTIONABLE VALUE. Three-stage target rock valves have well-defined cause of in-advertent opening and failure to close. Limerick has two.-stage target rock. Operating experience with this design has been much better. SORV probability used in PRA probably conservative. (ilF5 9a 3~N 3R \\.d1t )
i 0280 ASUTMENT -* r REMOTE AIR I , -{. ACTUATOR .C y ') ( BONNET ((NM ((' f .t BELLOWS 3 h \\ 00~ LEAKAGE N\\ .g, 7 3 PRELOAD V/ / /,t [ 5 FACER ESTO p PtL E , -]\\ -hw w c :_ ADJUSTMENT SECOND STAGE S _s PftELOAD SPftlNG i SPRING g SECOND STAGE m DISC (CLOSED) N \\ PILOT STEM f /Y / / / / / R ~ MAIN VALVE \\F -A 1 PftELOAD SPftlNG 7 k YOKE PORTION Q OF MLOT VALVE hh' pt MLOTVA VE \\ DISC (CLOSED) \\ MAIN VALVE p MLOT SENSING PORT MSTON ORIFICE #\\\\ \\ MAIN VALVE \\ OsSc(CLOSED 6 mN - INLET \\\\ \\ \\ HIGH PRESSURE FLuto OUTLET V j Figure 2.6-4 Three Stage Target Rock Safety / Relief Valve 2.6-39 v g, ... 7, .a_-. -,--,,,-,,,,-,,,,,.,..,n.,-..-
n l n. 1 g OfAPHRAG*A TYPE i PNEUMATIC ACTUATOR r. e. J t k ( - AIR INLET j e C 'c t W \\ u PtLOT SECTION k' AIN PIST!QN 11 ii : MAIN DISCS [ llI ( l I I INLET l M Ci3 CHARGE l 4 91880 UMERICK GENERATING STATION UNIT 31 AND 2 FINAL SAFETY ANALYSIS REPORT TWO-STAGE, PILOT ACT1JATED, SAFETY RELIEF VALVE (VERTICAL DISCHARGE) FIGURE 5.24 l
- .*,,ey*,
.g, ,7 g .,.,,,.,..--7.. ,,_..,,.n__ n, -,. -
S 9 REMOTE-MANUAL OPERABILITY OF SRVs IS LOST DURING THE LAST STAGE.0F A TW SEQUENCE AND REACTOR VESSEL REPRESSURIZES. 25-osidifferentialrequiredbetweencontroi a o &d D' * \\ ' gas pressure and drywell pressure. b 'S Mt.N6 l yosw.4 t. j.. a d.S pdt !- @ Necessary differential lost at 80 psia in Browns Ferry co.itainment. .~...s...
- s.,
LARGE POOL STEAMING RATES IN W SEQUENCE REQUIRE LARGE CONTAINMENT-LEAKAGE IF OVERPRESSURIZATION FAILURE OF CONTAINMENT IS TO BE AVERTED 3-in.2 vent only delayed failure from 35 to 39 hours at Browns Ferry ~ Orifice-to-atmosphere of 12-in.2 would be required to preclude pressure in-crease above 130 psia. LINEri CC- -ProdDE<0:\\ LMd c.N O W 'W S'T / y ~ :.\\. ; C 't '.,.n;,: L i : S-6 ... ~ _ - +.
IN TW SEQUENCES, FOUR POSSIBILITIES IF CONTAINMENT PERMITTED TO FAIL BY OVERPRESSURIZATION: 1) No LOCA and reactor vessel a b s'g > injection remains available. \\ot}f* 2) LOCA with reactor vessel in-Jection available. 3) No LOCA - reactor vessel in-Jection lost. 4)* LOCA and reactor vessel in-Jection lost. .gc .e 0 ~ )...O' 'M ~j@ d -&- ',. 3.FI A 9p ec g b(p 'b*,'.,g, 'o. - 4 .c & g' '. .t G S Ai g 0'.) (tlh Q C'NE' =*
BROWNS FERRY SASA STUDIES INDICATE THAT AFTER REACTOR VESSEL FAILURE, CONTAINMENT WOULD FAIL BY 6 its isc O' p9' / y s .a OVERTEMPERATURE. ~ o 43 \\$'E '^ Limerick drywell design temperature is 340*F (wetwell is 220*F). MARCH predicts much higher temperatures after corium enters drywell, at pres- ,, s. jN, sures below design pressure (55 psig). , ', gs g, g; ~ \\- 10 3 . >3 v 1 Q x .jJ ' d,., t c \\ k s.O 9 y e no e s
LEAKING MSIVs ARE OF CONCERN FOR FISSION PRODUCT TRANSPORT. At Browns Ferry, measured 875.0 vice 11.5 SCFH allowed at end of each of last two operating cycles. Pathway through drains into main condenser, from there.into turbine building through,. inactive turbine gland seals. Limerick has MSIV leakage control system: will it work? H *'T'*J{ y > O j,o c.. y I.;,; x c :,:_v.(. ; :.,, Limerick also has auxiliary steam = supply to turbine gland seals: is it always available? i l -s L
LIMERICK ACCIDENT GAS TREATMENT SYSTEM DIFFERS FROM THAT AT BROWNS FERRY BFNP has large SBGTS -(25,000 cfm) exhausting through HEPA filters and charcoal to atmosphere. Limerick has small SBGTS (6,000 cfm) and large RERS (120,000 cfm). Operability of Limerick systems ~~ under accident conditions should be investigated filter plugging - effect of high humidity on Charcoal l .... r r r
- T
REACTOR BUILDING FIRE PROTECTION SPRAYS PLAY IMPOR, TANT ROLE IN FISSION PRODUCT TRANSPORT AT BROWNS FERRY ~ Sprinkler heads on at 225'F in reactor building. 0.3 GPM per square foot. Not a safety system. c> ' ",, gsyy C. Cfl [ M' [,. I ** ', L p;t: c M '. * / - C n's h.i 'E h,,.g. 3.e ( ( G, r.'s W Q E '$. H s u-
- ' j d I" k I
.. g.., -
- \\.\\ e u F...
~.~ -~ 3,s f M ig *,1.l d A T i. \\,. I t .***wre+, en e* _m ..,--,,,-p,, wr*'vv=r
- -=M
( .g MANY IMPROVEMENTS IN LGS NSSS DESIGN OVER BFNP, BUT SOME DISAPPOINTMENTS: Irreversible shift of RCIC pump suction on high suppression pool level. OpicGer Eai&. i,'; cA 4. Vs. Ji\\ch M cnior so:uH & P<t4k si,lg.pf (MS, lu 2 oil.';2i.13 Q) Smaller condensate storage tank ~ Ns). f 1(2h 58{4*S9 (200,000 gals). Vw conae, wuj RHR pumps connot take suction on the condensate storage tank. ,.s
i . ~. S. A. Hodge ORNL Dec. 16, 1982 Other ORNL Com'ments Concerning NUREG/CR-3028 1. P. xxiii. "...the siting model used by us... ." Why did BNL use a different' siting model? 2. P. 1-2. "reviewd". (spelling)
- 3. - P. 2-5, 1st sentencei "As noted.'n Subsection 2.1.2...." This is contained in Subsection 2.1.2.
4. P. 2-5, 2nd sentence: ORNL has verified that RCIC can supply adequate makeup to the Browns Ferry reactor vessels with one stuck-open relief valve. This should also be true at Limerick since both are of the BWR-4 design and Limerick has 14 SRVs com-pared to 13 for each of the Browns Ferry units. S. P. 2-15. Feedwater runback should be included in the list of frontline systems for reactor suberiticality given in Table 2.3. 6. P. 2-16. The RCIC in ection flows given i$t Table 2.4 should be 600 gym for both LGS and RSS-BWR. The additional flow is used for lube oil cooling and does not enter the reactor vessel. Also on this table, the LPCS flow for LGS should be 6250 gym. 7. P. 3-1. It is recommended that the first sentence be rewritten' for clarity. 8. P. 3-3. The word " pressure" in the second bullet under item b) is not understood. 9. P. 3-4. It is recommended that "to reclose" be inserted after "1he success and the failure of the SRVs...". t 10. P. 3-7. The statement " Emergency core cooling functionability has also been removed from the. event tree since LGS identified no physical basis for this event." is not understood. 11. P. 3-10.< In Section 3.4, the comma should be removed after " LGS-PRA". Also, mention of the fact that the Limerick PRA does not consider the impact of pumped flow on lube oil temperature in the HPCI and RCIC systems might be made in this section. l
~ e 4 l 2 -l 12. P. 3-11. " Energy" should be " Emergency" for the ESW in the list of' l systems analyzed by fault. trees. Also, it would seem that the l plant air system availability must have been factored into the fault tree analysis for each system that it supports. - 3 13. P. 3-12. In regard to the' discussion in the first paragr&ph, only. a very small part of the design capacity of the ECCS systems is required in non-LOCA situations. Also, fires should be included in the list of items not included in the analysis. I 14. P. 3-14. Under ADS, 'the mechanism of a common mode failure of all-ADS valve solenoids due to contaminated gas supply is not under-stood. 15. P. 3-16.. In Section 3.5.2, ites a), failure of remote-manual operability should be specified. Otherwise, the reader might think failure to shut is meant. Also, in Section 3.6: - " decision"_ is misspelled. 16.- P. 3-17.,In the first paragraph, clarity would be served if the reader were' reminded that operator action is required because high drywell pressure is required for. automatic ADS actuation. 17. P. 3-19. "section" should be " suction" two places in the last mentence of the first paragraph. Also, something is wrong in the second sentence under'% 3_" where "and those based" is used twice. 18,. P. 3-21. In the 7th line, " Table 3.1" should be " Table 3.3" and in the 9th line, " Table 3.2" should be " Table 3.4". Also, it is recommended that the next to last sentence in the first paragraph end "... prior to core damage if suction is taken on the pressure suppression pool". 19. P. 3-22. Should be "... included in the LGS-PRA." in the 6th line under " Component functional dependences". Also should be " Component-human interaction dependences were included...". L 20. P. 3-33. Should be " Figure 3.11" in two places in right-hand column of table.
- 21. Should be " Figure 3.13" in two places in right-hand column of table.
Also " fig." should be completed under " Mitigation" at top c of table.< 22. P. 3-37. Should be " Figure 3.15" in two places in right-hand I column of table. Also " Fig." should be completed under "Mitiga-tion" at top of table. 23. P. 3-39. " Figure" should be completed in two places in right-hand column of table. Also " Fig." should be completed under "Mitiga-tion" at top of table. d = -
N*. \\ 24. P. 3-42. " Transfer from Figure" should be completed in left-hand column heading.
- 25. - P. 44
'Ihe second item under HPCI is not understood. Which auto-transfer action is meant? Neither should occur within 1/2 hour. Also, item No. 5 is the same as item No. 10 on Table 3.1. tlnder RCIC, the same comments pertain to items No. 3 and 8. In rsgard to item 3 under "FW/Cond," is the auxiliary steam supply to be always available at Limerick ? The FSAR doesn't say so. If not, it would take much longer than 1/2 hour to start up the aux-iliary boiler. " System" is misspelled in item 4 under "FW/Cond". 26. P. 3-45. Which alternate suction valves are meant in item 1 under "LPCI"? Also, the "1/2 hour" for this item belongs in the right-hand column. 27. P. 4-1. In paragraph 4.1.1, " Table 7.1" should be " Table 4.1". Also, this table shows five, not. four, groups under " Transients", f In paragraph 4.1.2, recommend " sixteen" instead of "the sixteen". There are more than sixteen operating BWRs. 28. P. 4-2. In the first line, " data" should be inserted between " prior" and "was". and "truhine" should be' " turbine".In the 6th paragraph, "these" should be "those 29. P. 4-3. Why did BNL choose to assess the frequency of the loss of offaite power on the nuclear experience alone? 30. P. 4-5. In the first line of Section 4.2.1, "PRA" should be in-sorted between " LGS" and "to". Also, "Susquehanna", "representa-tive", and " future" are misspelled in the 4th paragraph of this section. 31. P. 4-7. In the first paragraph, "their" should be "these". With regard to the last paragraph, it should be noted that BWR procedures do require partial manual depressurization after each reactor trip. The reactor vessel pressure is manually cycled between about 900 and 1100 psig to avoid automatic relief valve actuation. 32. P. 4-8. You state that the drywell pressure permissive signal is not satisfied during transients. GE maintains that the loss of drywell coolers will lead to a high drywell pressure signal and our Browns Ferry analyses confirm this as well as our Limerick simula-tor session. You may want to say that there is no direct and reliable way to obtain the drywell pressure permissive during transients. Qv * *:gy,* :q h ys ? 'jg*~*.*****
- '1
---e-y r.- y m
Y 5 45. P. 7-67. De results of ORNL SASA analyses show that the "CRD Injection Water" has a major effect on *IW accident sequences. 46. P. 7-76. It seems that a " bottom line" summary statement is needed to close Section 7.3.1.2. Also, a reference is desirable for " Conversations with a turbine manufacturer..." in Section 7.3.2.1. Again, the fact that the CRD hydraulic system can supply all needed injection after six hours should be mentioned in this section. 47. P. 7-95. He second sentence in the last paragraph needs fixing.
- 48. *P.'7-100.
" Meeting" should be* inserted after " September 3" in the last paragraph. 49. P. 8-4 Table 8.9, referenced in the next-to-last paragraph, is missing from the report. In general, the tables are presented out-of-order. 50. P. 8-12. " Release" is misspelled in the heading of Table 8.5. 6 I i i '89* 6 ** ' [1"7-I
- ##**[*
- 9
f c- ;. S. R. Greene-ORNL Dec. 16, 1982 Comments on Section 7.2 General h 1. Containment sprays do not appear to have been modeled in any se-quence. f 2._ Were Standby Gas Treatment Systems modeled? How? 3. MARCH has no models which allow a molten core to fall into the suppression pool. How was this modeled? h 4. In all sequences, when the core melts through the floor, compart-ments 1 and 2 should be opened to each other with a flow area equivalent to the size of the floor failure hole. This is not possible with MARCH, but due to the relatively small AP between the compartments this should probably have little thermodynamic impact. However, the impact on fission product transport could be signifi-cant since this would open a direct flow path between the.wetwell (WW) airs opening. pace and the drywell (DW) even without the vacuum breakers 5. For cases in which vessel was depressurized, hoV was this modeled? Existing MARCH SRV models may yield conservative (longer) times for depressurization. 6. CRD flows were not considered in any sequence. k 7. Possibility of condensate and RHR service water injection into vessel wasn't considered in TW sequences. c.a.:.2 ; g-Ci '.. c. w ;. ?. p 8. It is not clear that failure of the DW floor would resu'lt in con-a, tainment failure. [ 9. BNL apparently let initial MARCH oxidation layer default to 10-15, Based on ORNL work, this results in a significant speedup of the entire event sequence, with especially strong impact on time to initial melting. 10. Drywell coolers aren't modeled, h
- 11. Term " saturation" is used very loosely and can lead to misunder-standings. As used by BNL, saturation a 212*F regardless of actual containment pressure.
If this defination is not used by BNL, then there are errors since PSP flashing would follow containment fail-ure anytime T,,y >212*F. p ,........,s... m . s..,.
^ S. R. Gr ens Dec. 16, 1982 2 %-12. For all cases where containment fails by overpressure, BNL assumes failure occurs in DW wall when p = 155 psia. Iowa State report (NUREG/CR-2442) says failure would occur in WW wall at p = 133 Psia. WP;f, sd t,W C 2//tG. o';, [*a-A n c.f 7 lo c I G b \\ y . 13. Six things can happen when containment fails: 4.c.,' g ;N. 1. No LOCA Full injection capability 2. No LOCA - Some injection capability 3. No LOCA - No injection capability ~ 4. LOCA Full injection capability 5. LOCA - Some injection capability - 6. LOCA - No injection capability n o. possibility that containment failure phenomena could induce a LOCA was not considered. Specific 1. P. 7-21, paragraph 4. ORNL station blackout (TQUX) analysis shows that RPV and SRV piping-to-DW atmosphere heat transfer can add significantly to DW pressuri:ation rate. This can't be modeled in MARQi 1.1. Breat*to SRVs.'. Should keep in mind that the TQUV is modeled in a very conservative fashion since any injection would help considerably. 1 2. P. 7-23, Item 9. Basis for 70-em DW failure criteria? Item 10.. How was the normal 1/2% volume leakage / day modeled - if at all? Item 11. Unlike actual SRVs, MARCH SRVs work against DW pressure, so that SRV flow. depends on DW pressure. His can increase the time required to depressurize the RPV above the actual time that would be required. p Item 12. BNL apparently let initial oxidation layer default to This results in a rapid oxidation burst when the core uncovers.10-15 Based on ORNL work, this results in a very significant decrease in the time to initial melting as well as a speedup of all the sequence events. Item 13. Did BNL increase number of fuel pins to keep total Zr surface area correct? ae
- + ef *; e
,g; - - - - " * ' ~ * ' " " ~ -
t S. R. Gr:eno Dec. 16, 1982 s 3 .f 3. Pages 7-26 and 7-27. Figures 7..! a.?d 7.S. R ese two figures show i that MARCH predicts -that the suppression pool 6 160*F is in contact with the wetwell gas space (at a temperature of 3501 T 1 700*F) .l for s2 hours. His is an unrealistic artifact of the MARCH PSP model which only evaporates enough water from the pool in each timestep to keep the wetwell atmosphere saturated at the pool temperature. Actual evaporation rates would be significantly higher, resulting in lower wetwell atmospheric temperatures. Since MARCH transfers mass between the wetwell and drywell atmospheres, this would result ~ conservative dw pressure and temperatures. Compared to the actual situation, it is probable ser that MARCH predicts higher WW atmospheric temperatures - due to reduced ,f evaporation, and lower pool temperatures due to: (1) DW atmosphere not entering via downcomers, and (2) well mixed approximation. P. 7-26. g (/ Why no drywell failures on temp?, \\h '4. P. 7-28, paragraph 1. Additional uncertainties in DW temperature: (1) Heat transfer from RPV and*SRV piping (+) (2) Radiation energy lost from top of melt (+) 'j (3) Suppression pool evaporation (-) , s.1.,'.7.h (4) Drywell coolers (-) S. P. 7-29. General comments on TW (LDHR) Sequence: 1. LGS HPCI enters the core via core spray loop 8 piping; RCIC flow is split between the feedwater line and core spray nozzle; ~i LPC1 injects directly into shroud region; and CS is core spray. Therefore, almost all injection is either spray or top ,e b.(,L down flooding (LPCI). MARCH does not model either of these y phenomena, thus adding great uncertainty to Zr-H O reaction, H evaluation, etc. 2 2 2. Head curves should be used for LPCI, CS. 3. CRD, condensate, and RHR service water pump injections not considered. 4 Decay heat model can have large impact in this sequence. I 5. Since WLP, WLP, WLP " core melt ends" times are all N. equal, i.t isn't clear why the head failure times differ as they do. s.. .~
e ~
- s..
'S. R. Greeno Dec. 16, 1982 4 6. P. 7-20, paragraph 2: See TQUV general comment No. 1. 600000.lba looks a little high 81080 psi (%10%) but OK. f.WouldRCICswitchsuctiontoPSP?-onbot ow CS,T and h gh SP? $ MARCH doesn't model ECC turbine extraction steam (HPCI and RCIC). Pool might heat up faster and RPV depressurize more quickly than -predicted. h Would condensate pumps trip? If not, why aren't they considered?
- 7..
P. 7-30. Paragraph I (table). Is " core melt ends" equivalent to MARCH " slump" time? Paragraph 2, line 1. Intent isn't clear. k 8. P. 7-33, paragraph 1. It isn't clear why WLP hd failun h I should be different than W LP, since only difference is pool DCF. Since no melting occurs prior to containment failure, changing DCF should not change containment failure time. Since injection isn't lost until containment fails, core melt time should not change and head failure time should not change unless containment pressure is significantly difforent. V ,., 3 Changing DCF can change containment pressure after melting begins (.this is also after containment fails). Since MARCH SRV models work against containment pressure, this could change primary system pressure. 'It is difficult to determine the exact impact of all of this since much would depend upon the containment depressuri:ation rate (pad therefore failure hole size), core slump time, etc. 9. P. 7-33, paragraph 2. Differences between WLP and TWLP event timings must be due to changes in containment pressures, which, in turn, result in changes in primary system pressure. In actual case, containment back pressure would not have a signi-ficant impact on SRV flow and primary system pressure, so one would not see such differences. 10. P. 7-33, paragraph 3. May not be able to lift SRVs in some TW aaquences, so that WHP may not be a minor concern. a. b
S. R. Gre;n3 Dec. 16, 1982 5 g 11. ATWS - General Comments. Page 7-36 : 1. MARCH SRV flow model limitations could be significant in these sequences. Models employ minimum of critical or orifice flow. 2. MARCH's lack of spatial core kinetics adds great uncertainty here. 12. P. 7-40, paragraph 3. 15% pool flash looks OK if pool is saturated at N135 psia or so. [ 13. P. 7-44, paragraph 3. How does the 10% core get into the pool 9 head failure? Since they say they assume the associated oxidation release enters the DW via the downcomers, it appears that they assume that 10% of the core " spills" into the downcomers. Since the downcomers stick up %2 ft above the DW floor, this can only happen if the core piles up above the downconer lip, or melts the lip and cover, or both. If the core melts the downconer lip, the downconers might fall into the pool, opening the W air space to the DW. The other possibility is that the 10% core melts the DW drains and falls into the pool. This would also open the W air space to the DW. In any event, the analysis does not appear to acknowledge that the 10% core dump event could open a flow path between the W air space and the DW. In a realistic core melt visuali:ation, it is not clear that 10% of the core could fall through the downcomers. 14 P. 7-44, paragraph 4 %15% of pool would flash 8 containment fail-ura in TQUX1 if pool were saturated at %135 psia but it isn't; so why include this in the FPT7 % If the pool did flash in TQUX, a lot of the flashing flow would I enter tus DW through the opening in the DW floor that caused the containment failure. Even if no flashing occurred, the vacuum hreakers would not have to open to allow W-DW flow, due to the hole in the DW floor. k 15. P. 7-46, paragraph 3. Ames 1.ab, Iowa State Study (NUREG/CR-2442) says containment failure would occur at midpoint of W wall, not in DW, Since the containment is assumed to fail due to melt-through of the DW floor, fission products would flow from the W to the DW regard-less of vacuum breaker operation.
- 2.,
I-S. R. Grxn3 Dec. 16, 1982 6 MARCH intercompartment flows are unreliable since code transfers sufficient mass to equalize pressures. Minimum pressure differen-tials due to downcomer static head, vacuum breakers, etc., are not modeled. Comments on Section 7e3 f k. P. 7-89, paragraph 2. Vacuum breakers would not have to open. PSP steam would vent through the hole the core fell through. 6 he 0 0 4=p + ; *e e. .. -. e a,a, t
g % ala.a+ Miscellaneous Comments on the LCS PRA pp. 1-v. The comparison of the Limerick CCDF's to the WASH-1400 CCDF's is misleading because the Limerick CCDF's include differences not only due to design and site differences but also due to ' methodology and data base differences. However, the executive summary implies that the comparison figures reflect merely-the differences in the risk for the different plant designs and sites (Limerick at the Limerick site versus Peach Bottom at the Generic RSS site 9.
- p. 1-2 It is noted that the LCS PRA was to address the known problems in the RSS methodology and that among those problems were the probable underestimations of the effects of fires, earthquakes, and human interference.
However, these potencia11y important problems are later stated to be "outside the scope" of the LCS FRA.
- p. 1-20 It le stated that the " liquid, pathways are not considered.
~ Calculations with the INTER code... preclude more than a 4-ft penetration of concrete. The Mark II containment has a concrete diaphraga and water source directly below the reactor." This result must be based on a very specific set of assumptions and calculations. Yet these assumptions and calculations are neither referenced nor described. At best, the implied rationalization used to dismiss the liluid path-l ways seems suspect.. Only if the wetwell is assumed to be full of water and/or the corium is assumed to spread across the i entire wetwell floor does the dismissal of the liquid pathways seen possible. However, both these assumptions can poten- [ tially be violated in certain accident scenarios. Thus it seems as though the liquid pathways have been dismissed for inappropriate reasons. t[,. L:, * * *~~,
- 7 *,
- =~*r.,
1
_ _ _ ~ e d$ l i
- p. 3-10 "Since the Limerick analysis is performed with much greater I
definition in accident sequence consequence evaluation, there 3 is little or no justification for "scoothing" in the Limerick analysis. Instead, each coupled set of accident class and containment failure mode is calculated explicitly with CRAC..." This does not appear to be correct. For example, see p. 3-140s. 1
- p. 3-116 "For LOCA events, a value of 10-2 was used... For non-LOCA events, a value of 10-4 was used..." These values are stated to be taken from the RSS rebaselining work of NRC, Baetelle, and Sandia.- However, in the rebaselining work, a probability of 10-2 was assumed for accidents with the coolant system at low pressure while a probability of 10-4 was assumed for accidents at high pressure. These do not necessarily
} correspond to LOCA's and h n-LOCA's, respectively.
- p. 3-122 "It should be noted that WASH-1400 used (only) five BWR release categories... For the Limerick analysis, there are i
seven distinct containment failure modes considered, and four classes of accident sequences. This leads to potential 1 ' 7 I 28 separate ex plant consequence calculations, compared with the five performed in the WASH-1400." This is misleading. In WASH-1400, a large number of accidents with different con-i tainment failure modes and various types of initiators were considered initially but these were reduced down to four types of meltdown release categories for the consequence calcula-tions. Similar to WASH-1400, only five different types of releases were considered for the LCS consequence calculations-- not 28. l i I see,, s 9 ee.,* ap o.geoo ,e.-- e e - 4 p gm .,g- .,eme ...,_.__,_.,_me,,-,,_,,.m.m,,,,,-,,_ . -,. ~., _,,. _.,_,-,,%.-,,,.-_r.,,-.-e._.,-,,-,..-.-__cm---,-.------,y_-
o _- +- . pp. 3-123, on p. 3-135, it is stated that the Class IV release fractions ~l 3,140 are based on some unspecified NRC/Battelle calculations. In contrast, on p. 3-123 it is stated that some Class IV release fractions are taken from the RSS, that some of those release fractiona are similar to some unspecified NRC release frac-tions and that some are extrapolated from release fractions for other classes. (Note that the NRC/Battelle calculations, the NRC calculations, and the method of extrapolation are not described or referenced anywhere in the PRA.) Furthermore, inspection of p. 3-140s indicates that neither of the above descriptions of the derivations of the Class IV release frac-tions is correct. Such problems make it very difficult to technically review the fission product release fraction estimations.
- p. 3-140a It is not clear how more'than 8% of the ruthenium in the core is released from the containment in certain accidents if only 8% total is released from the core materials during the acci-dent. (It is our understanding from reading BNL's report that oxidation releases in the LCS PRA are assumed to enter the-suppression pool directly and therefore could not contribute to the fractions above 0.08.
In addition, there would be little likelihood of that type of oxidation release if the wetwell failed below the waterline prior to meltdown. How-ever, for Cg-Y", the release fraction given for Ru is 0.12.)
- p. 3-140a The source of the release fractions used to represent the OIRE category is not obvious. All of those fractions do not appear to be taken from the RSS as is stated elsewhere.
- p. 3-140a The release fractions given for C -Y, C -Y, and C -Y in 1
2 3 Chapter 3 are listed in Appendix 3 (p. D-33) as being for Cg-y', C -Y', and C -Y', respectively. (The notation appears 2 3 to be different in Chapter 3 and in Appendix D.) i I ( . ~. .:m.~.......,.......,... c.
t t <- p. 3-140a The release fractions chosen for each category are not con-servative. That is, upper bound release fractions were not adopted as they were in the RSS. In addition, the rationales behind the choices of the release fractions for the various, i categories are not presented. pp. 3-164/ The discussion of uncertainties underestimates the potential 196 impact of several major sources of uncertainty, for examgle, the effects of retention in the coolant system, the effects of different modes and timings of containment failure, and the effects of different initial release rates from the core sacerials. pp. 3-164/ The consideration of uncertainties concentrates on1h on those 0' factors which significantly affect the early fatality' CCDF, ~ which is just one measurs'of the risk. It ignores those many factors which can significantly affect other measures of the risk as well as those factors which can significantly affect the consequences of individual accidents. Chapter 4 Were the WASH-1400 results recalculated using the same version of the CRAC code that was used for the LCS PRA? If so, the errors in CRAC at the time of WASH-1400 do not appear to have been corrected and exist in the version of CRAC used for the LCS PRA. (For example, latent breast cancers were incorrectly estimated in WASH-1400 due to a programming error. As a result of that error, the latent fatality CCDF's given in WASH-1400 were noticeably underestimated.)
- p. 4-10 If " smoothing" was eliminated for the LGS PRA, shouldn't the LGS PRA results be compared to unsmoothed RSS or rebaselined-RSS results instead of to smoothed RSS results?
? ,..y ,s- ..,4 o., ..m. a e., .........c. ,y
. pp. 4-12, The so-called " design differences" reflect more than just 4-18 4-21 design differences. For example, the RSS used composite " upper bound" release fractions for each release category whereas the LGS PRA used release fractions for a single sequence in each category. As another example, the version of CRAC (COMO) used in the RSS underestimated latent canc but the version of CRAC used for the LGS PRA should not ha 4
- p. 5-5 "The use of accident categories in WASH-1400 required lumping accident sequences having major differences in potential con-sequences -into the same category, for consequence evaluation.
For LCS evaluation, the use of categories was eliminated and each unique sequence type was evaluated separately." This is a gross overescatement of the LGS PRA evaluation.
- p. 5-5
" WASH-1400 used a concepE of smoothing of probabilities between categories, to account for miscategorization and other uncer-tainties. This procedure was unnecessary in the LCS evalua-l tion; because of better definition of accident sequence consequence evaluation." This is another ov.cstatement of the LGS PRA evaluation. Appendix D The assumptions used to consider holdup in the reactor vessel and in the rest of the coolant system are not given. Inasmuch as there is currently no standard way to treat holdup and i retention in the coolant system, this needs to be discussed.
- p. D-28 SAI-REACT was used to verify the CORRAL results for some sequences 4 However, when the SAI-REACT results came out much higher than the CORRAL results, they were said to be in error.
If SAI-REACT is in error for some results, it may be in error for all results. 4 w = =p; ow,:- e-... ye. -~.s ,.. _ _.. _ _.. - - _. -. _ -.,., _.. - ~ ~ - _.,._
Overall Rolease fractions '.+cre not estimated for any LOCA's. Overall The descriptions of the fission product transport calculations in the LCS PRA are generally inscrutable. It is not possible, even with great effort, to ascertain exactly how at least certain portions of the fission product release calculations were performed. Overall Which of the three RSS models was used to estimate latent cancers in the Limerick PRA7 It does not appear to be stated. Overall ~ The consideration of only four accident sequences (one for each class) is questionable inasmuch as fission product releases to the environment are very scenario dependent. -It is not obvious that the entire range of possibilitiesjhas beenadequatelyconsidere). Overall No sensitivity ' analyses are presented for either I;he ove'rall study or the various component models and assumptions. The uncertainty analysis given appears to be entirely subjective. 9 9 _ zy s ;:;',,q; #, .~ ~ ~?"
.y [# f
- Q [ P.s C1,
,,9 4 Y Miscellaneous Comments on the BNL Review of the LGS PRA
- p. 7-44 The BNL modeling of the oxidation releases for Class I acci-dents seems to require that the corium fall into the pool via the downcomers because the entire oxidation release is assumed to enter the dryvell via the downcomers. Yet the effect of that assumption on the releases of fission products from the containment is not considered. That assumption would seem to enhance escape for early drywell failure but to diminish the likelihood of escape for early wetwell failure.
4 i
- p. 7-45 CRACIN = 7
- p. 7-51 The description of the creatment of holdup in the coolant system that was used in the PRA is missing. (As is noted by BNL, this description was not provided in the PRA.)
~
- p. 7-81' To estimate the fraction of fission products retained in the RPV, the fraction of hydrogen escaping from the RPV prior to RPV failtire was used. Such an approximation will tend to underestimate the escape of the more volatile species and to-overestimate the escape of the less. volatile species.
In considering holdup and/or retention of fission products in the reactor vessel and the rest of the coolant system, the following must be considered: 1. Different fission products are released at different times into the RPV. Therefore, their behavior in the coolant system must be followed on an element by element basis. 2. If the retention is considered, then the potential for eventual removal, e.g., by revolatilization or resuspension, must also be taken into account. 3. Retention is critically dependent upon the details of both the timing of the releases of the fission products from the core materials and the steam-hydrogen velocities-4 " ". ? L'~ \\ ',y y f?,*[. },;.'x q'. + L y Q* \\ ~
9 g* f rates through the coolant system. Thus a species released during a particular portion of an accident can have a much different probability of escape from the coolant system than one released at another time. As a result, fractions for escape from the coolant system tend to be highly species-and scenario-dependent. However, only. four very limited types of accidents, none of them LOCA's, have been considered. It should be noted that currently, there is'no standard method for considering holdup and/or retention in the coolant system and that direct application of MARCH and CORRAL to such con-siderations is inappropriate.
- p. 7-92 The very large oxidation release accompanying the postulated dropping of the entire me,1t into the pool may have been ignored. It is not. clear what fraction of the core was assumed to participate in the oxidation release.
Overall Would BNL's version of the CRAC code give the same results as the RSS if the RSS input were used? If so, then the LGS PRA results are suspect because the version of CRAC'used for the PRA and that used byDENL generally give such different results. One would not expect that the results of using the BNL version (or any other version) would be exactly the same those in the RSS due to the correction of programming as errors in CRAC shortly after the completion of tha RSS. However, they should be close. (This should be demonstrated for the BNL version.). Overall It would be helpful if BNL would generate CCDF's for acute and latent fatalities using their updated probability and release fraction estimates, their version of the CRAC code, and com-pare the generated CCDF's to thoses given in the LCS PRA and to those obtained using BNL's version of the CRAC code with the LGS PRA probability and release fraction input.
n
- <;. nyy g :,u.,3 3 y,q m :; ~; r ~.-
w: -4 J ~ A PRA SHOULD 1.' PROVIDE AN ESTIMATE OF THE RISK RELATIVE TO THAT ~ .0F-OT!!ER. PLANTS 2. PROVIDE AN ESTIMATE OF THE RELATIVE IMPORTANCE OF .VARIOUS TYPES OF ACCIDENTS RESULT 1 REQUIRES THAT THE COMPARED PRA'S BE DONE WITH ~ THE SAME OR EQUIVALENT ASSUMPTIONS. RESULT 2 REQUIRES THAT ALL SEQUENCES BE CONSIDERED ON A BALANCED FOOTING. THEREFORE WE WILL DIVIDE THE P0TENTIAL PROBLEMS INTO SEVERAL0 VERI.APPINGCLASSES: ASSUMPTIONS THAT DIFFER FROM THOSE OF THE RSS 1 ASSUMPTIONS THAT POTENTIALLY MISWEIGHT SEQUENCES II. ti1 ASSUMPTIONS WHICH ARE INCORRECT ASSUMPTIONS WHICH ARE INCORRECT BUT A. CONSERVATIVE ASSUMPTIONS WHICH ARE INCORRECT AND NOT B. CONSERVATIVE WE WILL CONCENTRATE ON ONLY THOSE PROBLEMS WHICH AFFECT L THE MAGNITUDES OF THE ACCIDENT SOURCE TERMS. 7.,..- ; ; r-g. v -. ,r , ;g
A _2_ J g - e e i g-a G e 9 l TIMING OF' RELEASES FROM CORE MATERIALS 1 G'AP 2 MELTDOWN 3 0XIDATION i ~ 4 VAPORIZATION. i. e e t i e S + J e 4 O 4[*.
- 8. %
6.
? ITIMING0FRELEASESFROMCOREMATERIALE POTENTIAL EFFECTS SPECIES RELEASED DURING MELTDOWN CAN BE RET COOLANT SYSTEM, ESPECIALLY SPECIES RELEASED DURING PERIODS OF LOW STEAM VELOCITIES SPECIES RELEASED DURING MELTDOWN CAN BE SCR SUPPRESSION POOL FOR ACCIDENTS INVOLVING RELATIVELY SLOW HE CORE MATERIALS, SOMEWHAT LARGER FRACTIONS OF THE LES VOLATILE SPECIES MAY BE RELEASED BEFORE RPV M .THROUGH FOR ACCIDENTS INVOLVING UNEVEN HEATING 0F THE CORE, A VARIETY OF SPECIES WILL BE RELEASED AT ANY O f )* $f, * ' * *' ,----m.
TIMING OF RELEASES-FROM CORE MATERIALS RSS: TIMING OF EACH TYPE OF RELEASE IS THE SAME FOR ALL SPECIES PRA: RSS PkESCRIPTION BNL: RSS PRESCRIPTION (EXCEPT OXIDATION) SOA: TIMING OF EACH TYPE OF RELEASE IS DEPENDENT ON BOTH THE VOL'ATILITY OF THE SPECIES AND THE THERMAL-HYDRAllLICS OF THE SCENARIO a S e 2 e 4
o e I i Gap Relene / Steen Emplemen tb = 025 t 1 OJO 1 025 Men Release Vaporiaaten Reissee S-z..s Ill1111111 Inni,,,fi i i i i 'i f 8 9 9 e 9 s a a Tune, Hews FIGURE VII J-8 Typical Sequence of Spike Fission Product Releases for' Postulated Accidents VII-203 4 e e 8 O , _ ~...
TIMING OF RELEASES FROM CORE MATERIALS TIMING CONSERVATIVE MAGHITllDES CONSERVATIVE .RSS, SCENARIO-SOMETIMES SCENARIO- 'SOMETIMES PRA INDEPENDENT INDEPENDENT BNL (RSS PRESCRIPTION) (RSS PRESCRIPTION) SOA SCENARIO- " REALISTIC" SCENARIO- " REALISTIC" DEPENDENT -DEPENDENT m - e e O 4 ,--...,._p,,..., .7
Y ~ TIMING OF RELEASES FROM CORE MATERIALS POTENTIAL IMPACTS OF SOA ON CONSEQUENCES AND RISK 1. FOR ACCIDENTS WITH EARLY CONTAINMENT FAILURE, THE RSS PRESCRIPTION FOR THE TIMING OF RELEASES FROM 'THE CORE MATERIALS IS NOT CONSERVATIVE. 2 FOR ACCIDENTS WITH LATE CONTAINMENT FAILURE, THE RSS PRESCRIPTION FOR THE TIMING OF THOSE RELEAS IS OFTEN ADE00 ATE. 3 FOR MANY ACCIDENTS, THE RSS PRESCRIPTION FOR THE MAGNITUDES ~0F THOSE RELEASES IS NOT CONSERVAT G 5 W E p *. Ob 94 % 9%
- Gus
- eeme h e*+ e -
E ("
HOLDUP AND RETENTION IN THE COOLANT SYSTEM ~ PROCESSES 1 MASS - (STEAM-HYDR 0 GEN) TRANSPORT - 2 CONDENSATION 3 ADSORPTION 4 . CHEMISORPTION 5 PLATE 00T 6 AEROSOL PROCESSES (AGGLOMERATION, DEPOSITION,...) 7 SCRUBBING 8 I ... -...-....-x;._,....,.-...-.....m.
HOLDilP AND RETENTION IN THE-COOLANT SYSTEM POTENTIAL EFFECTS HOLDUP: NOT ALL OF MELTDOWN RELEASE IS SCRUBBED BY SUPPRESION POOL AND S0 SOME 0F IT CAN ESCAPE DIRECTLY TO DRYWELL AFTER RPV MELT-THROUGH RETENTION: NOT ALL OF MELTDOWN ~ RELEASE ESCAPES FROM. COOLANT SYSTEM FOR TRANSIENT-INITIATED ACCIDENTS, THE HOLDUPS MAY BE CONSIDERABLE (UP TO 1 HOUR OR MORE). THEREFORE, THERE IS A POTENTIAL FOR SUBSTANTIAL AEROSOL DEPOSITION IN THE COOLANT SYSTEM. FOR LOCA'S, THE HOLDUPS WOULD OFTEN BE RELATIVELY SMALL -(ON THE-ORDER OF SECONDS OR M.INUTES). THEREFORE SUBSTANTIALAEROSOLDEPOSITIONISNOTASLIKELYkS IT IS FOR SOME TRANSIENT-INITIATED ACCIDENTS. e e ....,g.....,,....
HOLDUP AND RETENTION IN THE COOLANT SYSTEM TREATMENTS RSS:' FEW-DELAYS CONSIDERED; NO PERMANENT RETENTION IN COOLANT SYSTEM PRA: HOLDUP AND RETENTION ASSUMPTIONS NOT DESCRIBED BUT APPARENTLY SOME NON-TRIVIAL ONES WERE USED BNL: FISSION PRODUCT DELAYS ASSUMED TO BE A FUNCTION 1 OF FRACTION OF. HYDROGEN RELEASED; NO PERMANENT ' RETENTION IN COOLANT SYSTEM SOA: RATE OF RELEASE FROM C00lANT SYSTEM IS SCENARIO-DEPENDENT -- IT DEPENDS O'N TIMING 0F STEAM-HYDROG FLOWS VERSUS TIMING OF INITIAL RELEASES FROM CORE MATERIALS; RETENTION IS DEPENDENT ON BOTH COOLANT SYSTEM STEAM RESIDENCE TIMES AND AEROSOL CONCENTRATIONS G I - ~ -
HOLDUP AND RETENTION IN THE COOLANT SYSTEM HOLDUP RETENTION TREATMENT . CONSERVATIVE TREATMENT CONSERVATIVE RSS. NONE* NOT NONE YES NECESSARILYt -PRA ? ? BNL SOME-NOT NONE YES NECESSARILYt i SOA SOME " REALISTIC" -- SOME " REALISTIC"
- FOR MOST ACCIDENT SEQUENCES.
t NOT NECESSARILY CONSERVATIVE IF TIME-0F CONTA OCCURS BEFORE AND/0R CLOSE TO TIME OF MELTDOWN; NOT NECESSARILY CONSERVATIVE SINCE DISREGARDING HOL RESULT IN OVERESTIMATION OF SCRUBBING BY THE SU -POOL. l f , q, 7 r 7;,..,....
~ HOLDUP AND RETENTION IN COOLANT SYSTEM POTENTIAL IMPACTS OF SOA ON CONSEQUE 1 FOR TRANSIENT-INITIATED ACCIDENTS FAILURE, CONSIDERATION OF HOLDUP IN THE COOL SYSTEM CAN RESULT IN LARGER RELEAS MENT BECAUSE NOT ALL OF THE. MELTDOW ASSUMED TO BE SCRUBBED IN THE SUPPR 2 FOR'SOME TRANSIENT-INITIATED ACCIDENTS, CON AEROSOL DEPOSITION MAY OCCUR IN THE FOR SUCH ACCIDENTS, LARGE REDUCTIONS IN TH OF-MOST AEROSOLS PREDICTED TO BE R IF S0, LOCA'S (CURRENTLY NOT TREATED IN THE P BECOME RELATIVELY MORE IMPORTANT. ALSD THE OVERALL1 RISK MAY BE LOWER THAN PREVIOUSLY ESTIM e + Q e I i e i g '# a-O*
HIGHLIGHTS 1 THE RSS (AND THEREFORE THE LGS PRA) ESTIMATES OF ~ RELEASES FROM THE CORE MATERIALS ARE NOT.NECESSARILY CONSERVATIVE. IN PARTICULAR, THE POSTULATED TIMINGS OF SOME OF THE RELEASES ARE OFTEN NOT CONSERVATIVE. 2 THE RSS (AND THEREFORE THE LGS PRA) ESTIMATES OF HOLDUP AND RETENTION IN THE-COOLANT SYSTEM ARE NOT NECESSARILY CONSERVATIVE. 3 INADEQUATE TREATMENT OF RETENTION IN THE COOLANT CAN RESULT IN THE IDENTIFICATION OF THE WRONG ACCIDENTfEQUENCESAS.BEINGRISK-DOMINANT. O I i e O J 9 , a $ k'W g._* N ' 'e l{*,, ; f;) f' $!, ^s 'h *, '. 5, ( ~ l ~
1 ( ORNL MARCH BWR IMPROVEMENT EFFORTS by Sherrell R. Greene Oak Ridge National Laboratory 4 -presented to NRC Staff December 16, 1982 e 4 e _ m a ~<<. - s- ,-.o. ,- _. 9," t r,
ORNL IS PARTICIPATING IN DEVELOPMENT OF IMPROVED BWR SEVERE ACCIDENT ASSESSMENT METHODOLOGIES 5 bY,3-lan ' =" c, r MARCH 1.lB 2AU ': h4M MARCH 2.0 ,,. t g,.3 :.\\ :9"'* y MARCH 2.1 '2 -' ( a. (* * \\,.), s G.s ~ MARCH 2.B C '.. 'W'
- T:7L'
.% v, ;;,.* p _~:. .,~~;r
v .. F. s .f $ b.'3 t',k;:~' \\ a, 9 f MARCH 1.1B IS ORNL MODIFIED MARCH 1.1 ~ Incorporates unique capabilities containment inerting improved RPV level and mass
- \\,; ' ;.-
(- ADS model corrected containment vP,. e,.;.. '.0 ~ modeling errors - ANS 5.1 - 1979 decay heat .g-Utilized in ORNL SASA studies gM '.i
- c. -
'S / h .s ..f A~ -ye.s+o' e
- dep
.C q.W " L I l w -w -~ y w-
e ._o .a h h. a N n ) ORNL-OWG 82-6489 ETO NL MOST MARCH BWR MODELING DEFICIENCIES RELATED TO UNIQUE BWR INTERNAL STRUCTURES AND CONTAINMENT DESIGN w llllllMR 5
- c..
ll rJ n + ( } 7 v ..t l I-g
I e 4 c ORNL-WS-22666 ETD cansion. "' THIRTY-SIX PROBLEMS ASSOCIATED WITH MARCH 1.1 APPLICATION TO BWRs HAVE BEEN IDENTIFIED
- UNREALISTIC BWR CORE COLLAPSE MODEL e
HEAD FAILURE VIA CRD TUBE PENETRATIONS e NOT CONSIDERED' SHROUD AND CONTROL RODS NOT MODELED e CORE SPRAY SYSTEMS NOT MODELED o 'NUREG/CR-2872, APP. 8 f O t -*** 1 %~,~ 'n - r.. f- ~ 4 * *;
g i ORNL-WS-22667. ETD ^ UNION CARalDE NL MARCH 1.1 BWR MODELING CONCERNS (CONTINUED) PRIMARY WATER MASS AND LEVEL CALCULATIONS INCORRECT e SEPARATE BREAK AND SRV FLOW PATHS NOT ALLOWED e ECCS TURBINE EXTRACTION STEAM NOT MODELED PUMP DRIVEN ECCS FLOWS TERMINATE ON RPV HEAD FAILURE .I .x }. LO 7,f[ / { O'
- ' Y
. p d M d 2,r,.. s c ,7..n, ct + 6.v:j,< c en.y ;4 R( r pc:.b 1 e,7. y,. y., g...,.. - -. *.. ,.,-.,,,.y,.
CRNL-WS-22670 ETD 3 umon canssos NL MARCH 2.0 WILL FEATURE SIGNIFICANT 1 IMPROVEMENTS IN THE FOLLOWING t AREAS
- DECAY HEAT IN-VESSEL THERMODYNAMICS
- BREAK FLOW MODELING
~ STEAM J' ENERATOR HEAT TRANSFER
- LOWER RPV HEAD MELT PHENOMENA
- H AND CO BURNING 2
I e e - * : maw;+ vry y:xw .;.c -v. - ?p',.7c
- ,~--~s'~. ~ ~ ~ - - ':-
-c
MARCH 2.0 - STATUS .FT 77 version w/o restructured subroutines released October 1982 FT 77 version with restructured subroutines release December 21, 1982 e Ji@ g 4t . g6 ' ,/ k e; '.. 3..,\\ ,o '~ li s, \\
- O c,Mg
,.4 cyc-{ ,N J. --- >. ~, g yy y,.p. ? +, ,~
o ONE PURPOSE OF MARCH 2.1 IS ADDITIONAL IMPROVEMENT OF BWR SIMULATION CAPABILITIES Distributed primary model Mark II containmen't model Detailed BWR core melt model - p Improved melt models A & B CORCON. integration \\.8, <7 ..s >gs fY.g. E_ (s:. \\~;g a -......-.7-- ,7,....-
MARCH 2.1 - STATUS ORNL-simplified BWR core model - complete Y gsrs9 i RPI detailed BWR core model - complete U g Model transfer to BCL 21-82 BCL to initiate 2.1 work - ? ta>i> N.16
- g/
, 1. l.1 MARCH 2.1 release - ? \\i y ~' ~ v > ',.f- \\ l
t 'I e MARCH 2.0 & 2.1 WILL NOT ADDRESS SOME OF MOST SIGNIFICANT BWR MODELING DEFICIENCIES / Core spray 4 Lower plenum melt progression Localized RPV failure (CRD melt-through) .i ? Suppression pool Primarywatermbssandlevel ^ ECCS turbine extraction steam 4 I l S . [4 [ **', [ (..' ,,',#W'. p '#.* Y #.I
- REMAINING MARCH MODELING LIMITATIONS CAN SIGNIFICANTLY INFLUENCE
~ Timing and mode of vessel and containment failure Fission product distribution in reactor primar7 and containment Effectiveness of in-vessel and ex-vessel ESFs Impact of 'non-ESF' system opera-tion f, .~.,2k' p ,y \\'6M pg.dT .f ~' g c, @ \\C 9 6 ,,Qh m) /. 3 l>, O j,>qc.>.+ x n Q "', _ e. #~ ca ry.:
- i
'i CgcM e A , _J ' ',3 3900sg(s'g. I, (.>* ' ,.. q.,.,..;.y,.7... .s-
? ~ O ORNL TO DEVELOP MARCH 2.B TO ADDRESS MOST SIGNIFICANT BWR SAFETY ISSUES ORNL simplified core melt model RPI detailed core melt model ~ RPI lower plenum progression and head failure models Suppression pool model Core spray model i f r I .,__u_.,.,,..,.,__..
i 'l..*.' MARCH 2.B DEVELOPMENT SCHEDULE
- MARCH 2.0 Reception 12-21-C' MARCH 2.0 Operational 02-01-83
. ^ MARCH 2.B/0 Operational 04-15-83 PWR Coding Eliminated LJO Core Model Containment Inerting RPV Level ECCS Control ADS SRV MARCH 2.B/l Operational 07-15-83 PSP (Basic + Turbine steam + RHR) RPI Core RPI Core Spray MARCH 2.B/2 Operational 03-15-83 MK II Containment ' Assumes SRG 3/4 Jan.-April, 1 May-LJO 3/4 Jan.-Feb., 1 Mar-e A W. &w' o> Sc6 'gh", 9 g\\;+'gf, 6 9 o e . - -. ~
g ORNL SIMPLIFIED BWR CORE CHANNEL BOX / CONTROL BLADE.- HEATUP MODELS by-L. J. Ott ORNL_. Presented to NRC Staff December 16, 1982 ' <. * ;w y'T'n '
e PRESENTATION OUTLINE BWR core structure review / 0RNL BWR core model review Importance of initial Zr0 thickness to severe accident event timing Impact of ORNL model.on severe accident event timing a Scram discharge volume (SDV) pipe break, NUREG/CR-2672 o Loss of decay heat removal (LDHR) NUREG/CR-2973 Summary . r., a.v:. a :: :. n c c~.
- .~s ;
- , < - :~-
~ BWR CORE HAS UNIQUE STRUCTURAL, DESIGNS Fuel assembly Core support Control blades '"*e f*,.: * ^,, ;., FJ * *,;
- a * * 'S
- l
~ BWR FUEL. ASSEMBLY IS ENCLOSED BY A ZIRC-4 FUEL CHANNEL OD A bh WL r ~ N /p i \\ ~& ) ~ 11o9 l y...g CL *A n S ' W sin 18.i . 7,= :. e ,a FUEL 800- .%I[CI i l 1 a i WWEWI. FUEL ~ Cnuno l-Lotta Tit % = llrew.q nost nics l "J.:: I i g\\ AO mflra s s
- c. k'
,k N pg wp g'Wr c 9 1 ~ .. ~... -. ~,,, y.r,. .-...,.r,
i e e e g6 = G BWR FUEL ASSEMBLY IS SUPPORTED BY THE LOWER REACTOR INTERNALS m <h g g% / I s ( O_ ~ I ( \\ / .O-,d, g 1 q_I pE y@ 6 \\.... _ 4
- J a
y.- H E K _.._ y ....c. l onwican W u w e.o g m m ~ - "'' u" % ar ser i e t =
- r
- <.e<~.w
..:.,.s
4 BWR CONTROL BLADES ARE LOCATED BETWEEN FUEL ASSEMBLIES IN THE CORE f . as . h>[*l= - l! OOOOOOOdj [000,00000OQQ0000 ~ 0000.0000 OOOO~OOOO O__Q Q O.O O O O 00000000 00600000
=
q+ 00000000 20000000 ""'" N 9 00000000 O0000000 I 00000000 00000000 l', ,' j
- 00000000,
,00000000 i
- t
,j, qg .s n rr" l, ~eeeeeeee' ~ ~ o 6 l?e eeeeeeee eeeeeeee c.-.n~ iune...n + gggggggg 1 M.=. > 'W eeeeeeGe eeeeeees meCm 6-ria eeeeeeee
- = g t
o- ,eeeeeeee sun.C 50CRtf I 0'C f ( vp ..._y.. h
1. 'N b a ORNL HAS DEVELOPED A SIMPLIFIED BWR CHANNEL BOX / CONTROL BLADE HEATUP MODEL The model will be used to analyze the heatup, oxidation, and melting of BWR channel boxes and control blades. The model will be used in conjunction with the existing MARCH fuel rod heatup and melt models. I k
CURRENT MARCH FUEL-ROD HEATUP MODEL (VERSIONS 1.1 AND 2.0) APPR0XIMATE A PWR CORE s /- j' f-e w- / _b95> 'N y/l / stram Pd sim. % MARCH generalized fuel-rod-node heat balance: (Ean Ill.A.15, NUREG/CR-1711) ST* 20C x v + t,,y lov O.R
- 9.%*
- 9 melt ~ 9:ad ~W
~I2 g 3, R" D R c {,$ s ,f t _.r .---.~ -
l .0RNL's MODEL MORE CLOSELY APPR0XIMATES A BWR CORE O' 'f 'X' 5 f.9 /l_ '*f}f- <U bk x r-x g {n<W /
- s l/
%=2. Ja ourm zax E g '
- P^
- O w
f,* T
- a#M ht 'g.,
e STATUS OF ORNL MODEL Development is complete Incorporation in MARCH Version 1.1 is complete Two ORNL MARCH runs have been repeated ...i SDVA cco' O** .s ' '.' ' ' _;;,'i c \\ '.;u c LDHR . p. i ' ' \\>,
- , *,.w;p s '
.,. y
- p-
~ .sz_>q, ,.m..
IMPACT OF INITIAL Zr0 THICKNESS SPECIFIED BY MARCH USER CAN CRITICALLY AFFECT SEVERE ACCIDENT EVENT TIMING Rapid. fuel temperature rise and accelerated core response Greater metal / water reaction rate and increased H2 generation J The importance of the initial Zr0 thickness can be demonstrated by LDHR MARCH runs where XZr0 = 0 and XZr0 # 0. 4 .s.
- \\
,.N2 4 THE LDHR MARCH RUN WITH ZZr0 = 0 gj-(O.- EXHIBITED AN ACCELERATED CORE RESPONSE,'d.p' ' A. \\%Yg' (,\\h Y[tv, L X8 5' L sdl' ', g'. c-s / /> XZr0 = 0 XZr0 #0 '/ .S ,.y> Core uncovery 2263 2263 M ,/ Start melt 2323 2387 Core slump 2432 2477 End of BOIL 2435 2480 Bottom head fail 2527 ,2574 End HOTDROP 2547 2594 ').; 's ~ %..y (All times in minutes) ? ':- fe6 -0 c I \\' g 8 Y d* ', 9 & p 'h c\\w#4 'f<f.@ e ~ a t 0 2 s . (.. c s .v .V 4 f,, 'y'J ' gr gd' N
- - * * < ?,, ;c * ? ?.
- ?, p,.g
-~7
1 LDHR MARCH RUN WITH XZr0 = 0 HAD RAPID FUEL TEMPERATURE RISE t-LDER ACCIDENT: u CORE X00.EQ.0 l-1- E E kh hi-1 I 16 11-E E I E 1 3 i. I " nu "s - ut E
- BFN8 LDER ACCIDENT: u CORE XCO.NE.0 h!
I-1- dl 8 8 b 11- \\P I o#
- i. I p gi
/ ,t. c c i = l i. i. l 5-3 ~ ~ na"s vt."n: e.a v e, y,- e,,..
LDHR MARCH RUN WITH XZr0 = 0 HAD A GREATER H2 GENERATION RATE (EARLIER AND PROLONGED) BFNP LDHR ACCIDENT: u CORI X00.ZQ.0 El I. a 1 I 3-E' E i ll- [I-l1-li- __ _ l l l l l El, E OW c m:;gh - BFNP LDER ACCIDENT: u CORI X00.NT0 L.fI] gl-gi-E i g!- gi-j!- ji- ! I giq - ~. " " m"e - m:"n.:n
LDHR MARCH RUN WITH XZr0 = 0 SHOWS GREATER REACTION OF CORE CLADDING BFNP LDHR ACCIDENT; u CORE X00.EQ.0 ? N it it ' l&l& E s, ~ s. 3 s " nu"s - ut.E ~ ~ LDER ACCIDENT: u CORE X00.NE.0 l-l- ll-ll-azA 11 h l1- / I l- / l. I J nus - vem
- 7: a] {+f y g*, *, d;'.
COMPARISON OF SDVA (SCR' AM DISCHARGE VOLUME PIPE BREAK ACCIDENT)* ' MARCH PREDICTIONS VIA VERSION 1.1 CORE MODEL AND ORNL EXTENDED CORE MODEL 'NUREG/CR-2672
- r,. : ;'. ' *?.m:4.y ru:. cL~::, cyp;. <;r; * ;,'~
~ THE MAJOR EVENTS IN THE SDVA SEQUENCE OCCUR MUCH EARLIER WITH THE ORNL CORE 1.1 CORE ORNL CORE. CORE UNC0VERY 443 443 START MELT 561 503 CORE SLUMP 661 551 END OF BOIL 661 551 END HOTDROP 681 591 4 ( - (,'ck +$* f.wg g (All times in minutes, X
- 0)
Zr0 g
- qa
.y (q GV. e f '_ y ) p * 's-W b C, ?yj, _,y Y cf sp v ?)> 0- - - -: - :,. u. ye:.- a :,u
REACTOR PRESSURE, WATER LEVEL, AND WATER TEMPERATURE IN THE SDVA SEQUENCE BEHAVE SIMILARLY FOR BOTH CORE MODELS. w,By 3DVA ACCIDENT: u CORE XCO.NE.0 -g. g l.li
- l. It i
i nw y 2 I ~ ~ s.w.a ~ ~ BFNP SDVA ACCIDENT: u COP.E XC03E.0 BFNP SDVA ACCIDENT. u CORE XCO.NT0 I l! I I 11 5. si-s!- e .1 l [- If Il-l i n 1 E-I- i-3 l I s 1,. ii. s=~ ~ i l'; j'~ g-g eg. ei. j 2 i E. 5;
- l. !
l ~ ~ d. uu;% ~ ~ 4-m% ~ ~ ~ ~ ..~.-....,.x.33-.___
g ORNL CORE MODEL SHOWS A MUCH FASTER FUEL TEMPERATURE RISE FOR THE SDVA SEQUENCE SDVA ACCIDENT: UO CORE XCO.NE.0 l-l- / I I ss m ll I i gi-gi-i s 3 3 h i. 3 ~ ~ d. gg.& " B g, y SDVA ACCIDENT: u CORE XCONE.0 i-I- h / i 8 %I MI li-li L I. I It gi-5 5 3 3 l l i-i- $b 3 i ~ d. utd " l -r--- -,,---,e,,--,.---- -+n e ---,,----,n r - --- -- --, - - - - - - - - -w
THE METAL / WATER REACTION HAS A GREATER ENERGY INPUT WITH THE ORNL CORE MODEL IN THE SDVA SEQUENCE BTNP SDVA ACCIDENT: UO CORE X00.NT.0 BFNP SDR ACCIDM: UO CORE X00.NT.0
- 11. )-
.li Di s 1 ~ 3 l2 i 5 a ~ l= i-8, h 3 3 2 3. I:. 2 i i i. Is. t
- L A
c p t 's= ,_ g. BTNP SDR ACCIDENT: u CORE X00.NT.0 BTNP SDR ACCIDENT: u CORE X00.NT.0 13 Ti 1: s:- s t I g:. 5 3 g g& a i a I i, jr. L 3 5 I 3 3. i3
- 2., :.
i i i $ s. ht g' :.. :. N 5 a 3. 2, 3. I m - m.wrz: ttus - u. tzs w d@+ 9J.,-> ),' #? f arn ;p q'o
THE DEGREE OF METAL OXIDATION WITH THE ORNL CORE MODEL FOR THE SDVA SEQUENCE IS *AN ORDER OF MAGNITUDE GREATER O se i. ll-ll-11 ll ur l1l1 "a g, 3 /m .r..m - 9 uD s s ll-ll- ~ 11-II ll11 I I I ay
- 1. 1
.r..a ~ ~ i 6 ofs s -,.,, - - - -. -,. - - -,, -. -.., - - - - -., -.,.., -. ~, - - - -, _,,. - -. -, -,
~ COMPARISON OF LDHR (LOSS OF DECAY HEAT REMOVAL)* MARCH PREDICTIONS VIA VERSION 1.1 CORE MODEL AND ORNL EXTENDED CORE MODEL 4 .NUREG/CR-2973 4 d 1
THE MAJOR EVENTS IN THE LDHR SEQUENCE OCCUR SLIGHTLY LATER WITH THE ORNL CORE 1.1 Core ORNL Core Core uncovery 2263 2267 Start melt 2387 2402 Core slump 2477 2486 End of B0ll 2480 2486 Bottom head fail 2574 2594 (All times in minutes, X
- 0)
M Zr0 Jg g p t\\ cO#'?rDY u.Sh-a 9 y4-Y ,bt
- i me -
g ,,.-g=+, t _.. _ _, _ ~. _.,, _,,,,
THE CORE PARAMETERS (PRESSURE, WATER LEVEL, AND WATER TEMPERATURE) IN THE LDHR SEQUENCE ARE BASICALLY THE SAME FOR EITHER CORE MODEL 3FNP LDHR ACCIDENT: u CORE xCo.NEO 15 i l' [I (~ l>lt
- l. It
( I i,. t i. I
- r. gsk -
m ar7 wa Accrocm u coaz xco.Nto ,sry wa Accrem u core xco.Nto I T l1 l1 sl-si -It i li. 11 t a a I I. s; s ll. 5 g"i "i-g g!1 3-Ij.13 8. 8:
- 1. i 1
n - mm n - mm ,t
- , g, r -.3.p. 9,
.,c g j : - -', $l4 -,- ?" +
LDHR REACTOR FUEL TEMPERATURE RISE IS SIMILAR FOR BOTH CORE MODELS LDER ACCIDENT: UO CORE XOOSE.0 i l-a s li-It I i s gi. gl. i i 1-I- I, s " " m"s. ucEEs 1 q' LDER ACCIDr.N. T. L1 CORE XCO.NI.0 S 1 I i l - l< a l IiN I!- c / I gi gi- / 5 e i 3 3 / 11 1- / l .ilj 2 .ut - u:.v5 r
METAL / WATER REACTION SHOWS PEAK ENERGY BURSTS HIGHER FOR THE ORNL MODEL BUT THE ACCUMULATIVE M/W ENERGY ' INPUT IS LOWER BTNP LDER ACCIDENT: UO CORE X00E0 BFNP LDER ACCIDENT: UO CORE X00E0 ), )i it *i e L 3 L 1, g,.. s i, 3 N 2. t
- 3., 3.
i i N s. i.:.
- :. yz N
s l. BTNP LDER ACCIDENT: U CORE X00E0 BFNP LDER ACCIDENT: u CORE X00E0 ) 13 lt g3 gt gi. g 2 t 1,. 1, i, =. / =. = j 2 f t. 2 Ia i i-i !,o: i 8,. 5 ;. l , =. I g i ) I ma -wr. :n .:a". un " m w
METAL OXIDATION WITH THE ORNL MODEL IS SLIGHTLY LESS en ACCIDCm UO CORE XCCEO l-I- ll-ll-0.311-y al ll-A045 Me*WA l. I ,_, a en AccrDem u coaE xcono 11 !!- azes. U / SIe.- !!. / I I El1 e-l-I i il l-1] 1 SWE - WI.WHl3 . c.. - - s. * - g.,.p., eg
] ~
SUMMARY
ORNL has developed a simplified BWR channel box / control blade heatup model. The model has been incorporated in version 1.1 of MARCH. The initial Zr0 thickness specified by MARCH user can significantly impact severe accident event timing rapid fuel temperature rise and accelerated core response greater M/W reaction rate and increased H 2 generation The ORNL model predicts that the-major events in the SDVA sequence occur much earlier than with the 1.1 core model M/W reaction energy input with the ORNL model>> than with the 1.1 model The ORNL model predicts that the major events in the LDHR sequence occur slichtly later than with the 1.1 core model. 4 .,:, -_., c_.= .. c
3 BFNP 'LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 a % g. l ^Q ^9 4d 4 g-v. 6 23 M 2: M D =o D b-a A C: L 2 m 2 g-M$e > H- >3 m ;;;- i m p,. c: 2: 2 No x. a g. m.l,- ~ Cg o 2000.0 2100.3 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES i A . --:v...... -.., w_71...s.....,_;,.,..,...
4 BFNP LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 ~ 8 = 125-cd-d d 2 2e w wg Ea,- b g- ~ e m M M t-E- IS IS a R-a n-n w, u l'
- h. :- h 8
G 29-Ll o-c mR ms l L g_ l 8 L I j. 2000.0 2100.o 2200.o asoo.o 2400.o asco.o a TIME - MINUTES Si as ' '* r.q **
- r s '
. ;'9+ ~,y r.se; **,
BFNP LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 8 3 - g n-W / r.2 = s s $a $5 g _- @ H-g ra k2 bo < j-d I c . v s ,a I33 raa >g >= 3.E-3 l-i i 2 W r.a. - r,a b b i I % C C a g. raa rna - E* i .i .1 a A 3 8 5 8= t 8 [ 2000.0 2$0.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES .I e --.w ,e.- e - ~ - - - - - - -
4 BFNP LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 8 i l E' m o h-e c 4 v v i 1 6 de 2 2 M M E-E- M M 2: 2: i O O .O O 5 o 2 3d i D E' 2 2 x x 4 2 2 C i g. g_ l-4 ,,I 8 { 2000.0 21b0.0 2200.0 23b0.0 24b0.0 2500.0 e TIME - MINUTES E 2~ 5 ~ -,, m -v
BFNP LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 a o el sN-N N O m E 6 E o m g_ m g-da da m m .- i 4 m m 3 4 zo m I-o a-oS m-Es. Cn. M M 4 4 5 mg mo E' ca.E-oc R- -H c e a w w M ca a a h o r 2,- z g- ~ 04 d c4 l l 8 8 x x o o ( DE D b\\ h 0 .I +8 -C o. o { 2000.o abo.o 22bo.o sabo.o 24bo.o 2s00.o e TIME - MINUTES E 3 ? E 4C
- 'vn'*
ty y v-
BFNP LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 3a g_. ** 3 9e M v h-U M O M C E-- N-l Q:: s = x o 2_ E. f
- x. -
o M 2 M M 2 Cir.1 h-8 g 2 = = 2 De De Q d. o O 2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES r a:
BFNP LDHR ACCIDENT: 1.1 CORE XOO.EQ.0 3-E e e c-s- c c M M D E-D E-M M m m c-o 49 a o-aq 4 i O O z'- 2 \\ O .O. 3 D 3-U 2-4 4 l x ra. 3 I 8-8- ,I
- i e
o ca o j-2000.o moo.o asco.o asoo.o 24o0.o asco.o TIME - MINUTES ! l Ii l l I 5. ..#*l
- '=*("*a',*
g, .e. e
BFNP LDHR ACCIDENT: 1.1 CORE XOO.NE.0 %. a-l e ^C 8 4 ci ^g- -v. - b k [ [ v Gr2 CM M 3 y DC b @_ g. M M 2 r,a 2 g 'M C S E. M s_ us >= C/2 l-2: M y <2C N" h-O e l 2000.0 3 00.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES
- T.?,
? ,.,+ = =. +. y ~ ' '
BFNP LDHR ACCIDENT: 1.1 CORE XOO.NE.0 8 ~ r 8 c: mg_ -g_ a: a:" y'l-
- ,E-
~ = x M M E S $ g. $ e_ 5 0.. 0 > g- >g E-E- $ $~ o R 53 ma as a, g,
- a. g.
l_ !n 2000.0 2NO.0 2I00.0 2$00.0 2400.0 2500.0 TIME - MINUTES 4 em.
4 BFNP LDHR ACCIDENT: 1.1 CORE XOO.NE.0 = 5-C-3 mo o g Q:: M M N N 3 A gg Mo g d-h $- d M M >o i - @ y-4 h-z me mi 2iO v I a a M g >g >= 15-Lj]- 5-1; m-g i.4 - g S Y a$ O h. m $. 1 M 3 2 *- g-l- .s ~ 1 a l i O $ O's [ 2000,0 2b0.0 2200.0 23k)0.0 2400.0 2500.0 TIME - MINUTES y I -i ,j- -~e w w
BFN,P LDHR ACCIDENT 1.1 CORE XOO.NE.0 4 3 g a. 0 9 o -M b v .v d c.: o 2E 2 E= g-M m b M M Ct: 2: O O O O .e o ME 2* 4 WW D I~ .m.. a 3 2 ' A o . g-g_ I C-2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES d 9 f = = -.. .r..,-e, g.,.g 4
BFNP LDHR ACCIDENT: 1.1 CORE XOO.NE.0 0 N e 2 N N O m E d O m mW >N-m m A mi x x 2 . 2 0 O Q::
- M Gs.
ra. m w \\ E<a E<- s 2 3 %= .i rar=- m E-1 0" C W W ra2 .M -2 a 'z-i f-1 w w I 8 8 a x x o a g [J Z [ 2000.0 21b0.0 22b0.0 2300.0 2400.0 2500.0 TIME - MINUTES t $i .yy
BFNP LDHR ACCIDENT: 1.1 CORE XOO.NE.0 t3 da v, g-i i I n m a nn-W M- . in v $5 I ~ f 0m Cz3 * >e z 16- @ 5-M m e p 2 4 5 U O ~<4 o g m.m-D e- .c m e. m 3 o I m -@! e=-E:lo n- ~ \\ x i c4 l .g I e o =_ e 6 ] 2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 6 i i e TIME - MINUTES Q S i d . :;..,e -.=..,.
FN LDHR ACCIDENT: 1.1 CORE XOO.NE.0 3 d d o o o o <= <= M M ~ m M A A a a O Qo i 2 2 o o <,-.<d s ma e Gs. Dz. I f' - l_ ~ l_ e e l I i i o_ o ( 2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES l. 4 , :. ;.~ n::'.u.. -a.r.'.< :.v_.;;;;,_ c, k ~,. - z-', ~-~r:
e BFNP LDHR ACCIDENT: 1.1 CORE XOO.NE.0 h,,, a ha p W m 2 M2 .o. g - oe!$- g-f E = R e T E- $ 8-gi 4s N. c::' C 1 (2 g j_ ' -S-A g-1 C I g M M z Z .x x $ 2-r2 8-1 3, c E-' q a 1
- s
[ e 8o O*- [ 2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES E 2 i i
BFNP LDHR ACCIDENT: LJO CORE XOO.NE.0 li! 8 $~ E t: 20- C$- v v 6 C me ma u g" u g~ E=, bg e e M M E-E- IS I'g ~ I g 0' g a~ 0 8
- t m s_
s e_ = 28 x 81 l C C h b mR ms l " E- $~ l 3 I g-n f 2000.0 2100.0 z60.o ambo.o 24bo.o 2s00.o TIME - MINUTES
BFNP LDHR ACCIDENT: LJO CORE XOO.NE.0 k. g= gl .\\ s s 1 d g l~ hr M "8 M M k g-bo < lil-s i O v a ,.2 l N se j 5 $, $e $ a i g!- e!- 2 I m ] j 1 85 8= t a-8 [- 2000.o adno asno zibo.o abo.o asoo.o 2 TIME - MINUTES 1 1.
BFNP LDHR ACCIDENT: LJO CORE XOO.NE.0 3 a g-g m e g. g. 2 e w v de da 2e 2 as g-raa Ea 8 Dr3 M c: c: 8 8 1 a,t me e
- 2. !-
2
- a N-
- s 2
-x x l 2 2 C [ k-I ~. I d. 8 -[ 2000.o 260.o 2200.0 mso.o 24o' 0.o 2s00.o TIME - MINUTES i ,: 5-3 -, ~
. ~ : ~ BFNP LDHR ACCIDENT: LJO CORE XOO.NE.0 ".S @"S $ I 2 ma-E- E D 2 i v e z S-E- m h M l 6 h i- $ 5- $ !- 8 i m e E $3 l N S ! 8 2 i 22' 0.0 2000.0 2100.0 0 0 24 .0 2500.0 TIME - MINUTES i . ?, ' i -
BFNP LDHR ACCIDENT: LJO CORE XOO.NE.0 18.e- ** 3 yew v v i m E 0 ~3 3-U- U M 4 8 o T a- $ 2-8i 2 j s $_ $ / 5 A R-f 1 x o 5 M l r4 $ E-M 3-3 1 e J 4 8 g a 3 A $1 3
- a De o
i c o-ge a .j 2000.0 21b0.0 22b0.0 23b0.0 2400.0 2500.0 TIME - MINUTES 3 1 ,r.> r:.:: l
1 4 FN LDHR ACCIDENT: LJO CORE XOO.NE.0 d 3 g g. C i o w5 h g. <a M Dr3 i M M O o a a O Oo z 2: C o bh b h.- gd gd k k [ d d i l l l 2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 TIME - MINUTES - <-- *.. g, re,7. m. + s .+,, - - ,, s yr --.-,.m-%*,-r,,7-- .,_,,+-c -p.-- ,-.----y.--
s / BFNP SDVA ACCIDENT: 1.1 CORE XOO.NE.0 s $~ 8 8 12 5-C5-v v A i mg 2g w* h-w $- E E-2 m w w E*- IH - IS IS a C-a g-o w' M h .h >R 8 M5-m e-c: S mm i O C h h m8 me i l a g-A g-i 8 l S_ s l j 440.0 48tLO 532.0 578.0 824.0 670.0 TIME - MINUTES h I i -- i 4'.1 'e-g e-e p. r--e e-..e r,
) w J BFNP SDVA ACCIDENT: 1.1 CORE XOO.NE.0 C .M =' AC AN 4 Aa. g- ~ mE v V M W. M O W. U c *o Ei-x 2m 2 g Me 38-M g-m Cf3 3 4 m"2 e M h-k. L C 2. 3 440.0 486.0 532.0 578.0 624.0 670.0 TIME - MINUTES 6 e .,9...., ..+_.y.
( BFNP SDVA ACCIDENT: 1.1 CORE XOO.NE.0 s a g-8 .g M N 3 d \\ M M >.5-Ea g M g 8 l N "l-vn. o a-a M M f A g,2 "3 .-J M 3". N M W i!- g :- e A c M M b d 8] 85 440.0 486.0 532.0 578.0 824.0 670.0 TIME - MINUTES . - ~ *..
BFNP SDVA ACCIDENT: 1.1. CORE XOO.NE.0 W O E_ n n w r v v O 2, 2 Y g. g w ~ w m' a: t x .o 1 O O O O 2 22 c.. D *.. 2 2' x x 2 2 O g. d M. .C 440.0 4$6.0 5$2.0 Sb0 6d4.0 670.0 i TIME - MINUTES 4 I
- /.~.; w.7 F ?...:y,.c.*.*
- u?. :.- -
,/
- e BFNP SDVA ACCIDENT: 1.1 CORE XOO.NE.0
%a cg 7 I "' 7 a N O$ 'o e$- E !$- N v D b 00 p M* o e Z N' gg # ~ M M g Z o M ..p g O O n9
- c ac-g e-
= o M C.Q C4 -4 O Cu % e-e Ze I n-OC N O O d. d. 440.0 486.0 532.0 578.0 624.0 670.0 TIME - MINUTES ~ -:,iJy -lg. S Ts t,t;- ;,,*p*
- 7, 37,% 2, :: w 1..:
.$fS-*-
l BFNP SDVA ACCIDENT: 1.1 CORE XOO.NE.0 I! E ! bd g v CO
- E-b E-4 E
g a E-m E-iY h. i! = v 8
- u.,
w v-v- M g [ h N-M N-Ea E-* I D .= x l Do Uo i O e_ d l O ,i - 440.0 4$6.0 5$2.0 Si8.0 6d4.0 670.0 i i TIME - MINUTES l i i e e P 1 ..,..y..y,., .. _... ;..-, 1 :,..., l..
BFF SDVA ACCIDENT: 1.1 CORE XOO.NE.0 I d d l-l- OR t-B R- <e C4 M M M A o j o ao x 2: 2 l-3 i a $- U U ma s 5 i l-1- 1 d d I i i g .i o-. o [ 44aa 46s.o sko sko ad4.o e70.o TIME - MINUTES 2 t
BFNP SDVA ACCIDENT: LJO CORE XOO.NE.0 g 8 3-8 0 68-h.a- ^ s e EE 3% E- $ - E-h-e = h Ws 5e a k-a S-M' raa
- B B
ra2 E-M o: H mm D D Q".$ 04 C: L L 3 8 M-a MO 4d4.0 4$8.0 52.0 5$8.0 560.0 l TIME - MINUTES
- s:
g
Vs Y BFNP SDVA ACCIDENT: LJO CORE XOO.NE.0 D.H- < g-b M Q: . M 3 a _ f~ ~ 2: g x .g 2a 2 g Me > H-M g-M- m Af 2 4No Ol$ "' 0: h~ 6 w L b_ E 440.0 4$4.0 4$8.0 52.0 5$6.0 560.0 TIME - MINUTES .s ~ l
BFNP SDVA ACCIDENT: LJO CORE XOO.NE.0 3 8 g-s m 6 o e 2 m u N N g 5- @de d R-M M o-o k h-k k-i IO v i a a W m 5 b > g-M k- >o e g h s 2 m W m b } W< e g C 3-02- ) 2 2 R a a a a o g -o, 8e g 8 ( 440.0 4d4.0 4d8.0 512.0 536.0 560.0 TIME - MINUTES I a i .:.=..n n,,, ~
BFNP SDVA ACCIDENT: LJO CORE XOO.NE.0 l-a s I g_ ^ ^ w cr v v J da da 2N 2 ral 3 - cr:I g, p ca ca m e o C q .O O e 23. M e_ i' DS DS 2-X i x x 2 2 e o g_ 9_ e
- s o
N. o 440.0 484.o 4ae.o staa sse.o sea.o t i TIME - MINUTES l i !-ii Im
- e...,.me g. -
..m..,. e --.,,..e .,-,--,.,--,--.--.,,-.-._,,,,.-,..,._,n-n,,,,,-....,--,-,,--..a,,
~- 1 BFNP SDVA ACCIDENT: LJO CORE XOd.NE.0 I b9 b-73-7 d i nM mh-2: 7-B v 0x 3 h 5-G R-M g - 2; h O i -g g 5 U O <o o i M d- $ d-1 i M o C\\2 j Z i e h M ~ N a-T S-3 m 5 1 8. R { 440.0 4d4.0 4$&O 552.0 5$6.0 560.0 TIME - MINUTES i .,...s.:..-
-9' BFNP SDVA ACCIDENT: LJO CORE XOO.NE.0 3,L, N h2, g. E M v E { 0-G E = 8
- a !-
T E-
- E A
N 2 m C E g_ x-o M 2 2 M g 2 M [ $ 2-g 2-E i: 1 a C d p 8_ ! 8 O j MO N0 do b0 20 MO TIME - MINUTES . j. I I i I 4 4 + e ww-
v,y o
-,-,-4 ,---m-we--e-- ,.gw,- e -- -e ,+- - - -.-- --,--, -
F' s. SFN? SDVA ACCIDENT: LJO CORE XOO.NE.0 6 o E E t-4- d d c c w w U-U-4 4 w w M M a o a a i O Qo .2 2 o o. i I 4 w Uk U h-zo Ma [ is. ra. h_ h_ l d d I ( d 4 -[ 440.0 464.0 4$8.0 5$2.0 5$6.0 560.0 TIME - MINUTES I i i _}}