ML20116L407
| ML20116L407 | |
| Person / Time | |
|---|---|
| Site: | 05200001 |
| Issue date: | 11/03/1992 |
| From: | Buchholz C GENERAL ELECTRIC CO. |
| To: | Dangelo T, Palla B, Poslusny C NRC |
| References | |
| CEB92-54, NUDOCS 9211180294 | |
| Download: ML20116L407 (20) | |
Text
{{#Wiki_filter:" ~ bYOCl l ~ .u .p. Advanced RenatorProntam ^ San Jose, Califomis Phot;r (408) 3251785 Fax M08) 9251193 CEB92-54 Soy- // 3ct Tue, Nov 3,1992 f To: Chet Poslusny, NRC Bob Palla, NRC Tony D'Angelo From: Carol E. Buchholz
Subject:
Response to Questions from September 29 - October 1 Meeting The following issues were raised in the dncussions we had during the subject meetings. I have not yet been able to addrer.s all of the issues that were raised but I J wanted to get this information out to you. I hope it is useful in resolving any i questions you may have as you write the SER. l I Issue 1: Confirir the type of concrete used in the pedestal, j Resp ~n 1: l The pedestal of the ABWR is deSned as the sidewalls of the lower drywell. This structure supports the vessel and the wetwell/ upper drywell diaphragm floor. The type of concrete to be used in the pedestal is not specified. Basaltic concrete is [ required for the floor of the lower drywell. Basaltic concrete was used for the lower dgwellin determining the response of the containment to core concrete attack. This type of concrete is often used in the United States. The other type of concrete which is frequently used is limestone. common sand. Basaltic concrete is more rapidly eroded during core concrete interaction than i: limestone-common sand concrete. Therefore, one would ex)xct that iflimestonocommon sand concrete were used in the ABWR pedestal (i.e. th: side walls), the sideward e ;sion rate would be slower than that considered in the uncertainty analysis for core concrete interaction which was develeped in CEB. 17 133 (h I i 9211180294 921103
- l PDR ADOCK 05200001 o
A PDR l bi- . nou 4aem i g3_ m u.n m y, j
w X. Therefore, the times estimated in that analysis for the time at which pedestal integnty coeld be threatened are expected to be conservative if non basaltic concrete is used in the pedestal. The other key impact of the type of concrete is the production of non<ondensible gas. Limestone-common sand concrete produces more non condensible gas than does basalde concrete. IIowever, this will not have a significant impact on the analysis presented in CEIF92 X because the surface area of the sidewall will be only ten to fifteen Furthermore, percent of the floor area if core concrete attack should occur. the shape of the debris pool will be pancake-like. The gas generated at the side wall will not be able to reach into the debris pool and cause more rapid metal water reaction in the debris pool Rather, it will bypass the debris. Therefore, there will be little impact of the gas generation on the rate of attack due to any enhanced metal water reaction. In summary, the type of concrete to be used in the pedestal side wallis not specifled. If non basaltic concrete is used in the pedestal the rate of sideward ab'ation may be somewhat reduced as compared to the analysis presented in CElk 92 X. The rate of non-condensible gas generation may be slightly higher. However, because of the relative areas of the sidewall and the floor the impact will be small. The conclusions of the ancertainty analysis will not be affected by a different choice of concrete. Issue 2t Confirm mass of material used in core concrete interaction calculations. Response 2: Plant Design Data: 1 Mass UO2 = 171600 kg Mass Zr T550 kg (cans and clad) = Parammrs computed / input in MAAP: Mass UO2 - 171600 kg Mass Zr = 72575 kg The total UO2+Zr mass is closer to 244,000 kg in the MAAP analyses than the C 2+2r+ Steel mass as discussed in the meeting. The amount of":el that MAAP adds from the RPVis approximately 20,000 kg which represents the mass of the lower core plate. Due to the prompt penetration failure, there is not sufLcient time to melt the lower plenum steel (i.e. CRD housings). In addition, a significant amount ofiron is added as a result of the rebar in the concrete. The results of the FMX1P calculation indicate approximately 2000 kg/hr ofiron added to the melt. CEB92-41-2 November 3,1992 " T" " t t ^ y~m '"ru re
i' Immie St 4 Determine the time of Zr depletion for the bounding sequence for the core concrete interaction sequence. Resporwe 3: - For case FMX1P, the Zr is completely depleted by about 20,000 seconds into the event. The plots provided show that the onset of CO production is coincident with the depletion of Zr. During the early steges o CCI, the following exothermic jl r reactions are depleting the Zr mass: Zr + 2H2O > ZrO2 + 2H2 Zr + CO2 -> ZrO2 + C Therefore, the total hydrogen gas generation is not equivalent to the gas which would be generated from a 100% metal water reaction even though the Zr is depleted. Issue 4: Provide additionaljustification for the use of 100 kW/m2 as a limitiag value for the upward heat transfer limit. Response 4: The heat transfer between the water and the debris can be !imited by: a) Conduction within the debris, b) Critical heat flux, or c) Film boiling. The last is of concern if the debris surface temperature remains so hot that the water cannot wet the surface, i.e. if an insulating blanket of steam forms. Film boiling has been observed in well controlled laboratory environments using aolis).ed surfaces. However, it has been observed that the smallest of surface maerfections or contaminants would quickly result in a transition to nucleate bouling. It seems highly unlikely that the irregular surface of the debris would be able to maintain itselfin film boiling. Therefore, film boiling is not a credible limit to upward heat transfer. Critical heat flux is sufficiently high that it would not impose a practical limit on debris coolability. Therefore, a lower limit on the upward heat flux may be obtained by consideration of the conduction limit. The biggest unknown is whether the debris remains in an intact slab-like configuration, an intact configuration with irregularities which increase the heat transfer area and act as fins, or if the debris CEB9241-3 November 3,1992 neu usam tn m_r m..,,,,, i
develops cracks which allow water to ingress. The presence of cracks would increase the heat Oux, Therefore, let us consider the worst situation (intact s'ab). The temperature distribution in steady state, assuming a homogeneous debris mixture, is given by: B'T h 2 + g* = 0 (1) dx where: k = thermal conductivity (3.5 W/mK), g* = volumetric heat generation. It is sufficient for our purposes to consider the case of 1% decay power For a total debris 1. ass of about 244,000 kg, this implies and average initial volumetric heat generation rate: q" = 1,3 m In a one-dimensional flat geometry, integrating Equation (1) twice yields: - q"x' T= + Qx+ Q (2) 2k Ifwe 1) assume nucleate boiling is maintained at the surface, 2) conservatively assume that the bottom of the debris in contact with concrete is adiabatic, 3) assume molten debris is at uniform temperature, and 4) impose tl.e condition that the debris not ablate concrete, we have as boundary conditions: Q=0 Q2 = 1550K 7(6,)= 450K 33 j where: 6s. = debris thickness. CEB92-414 November 3,1992 non 4cem t :n 1_n_g y e
r Substituting into Equation (2), we have for the limiting debris thickness for coolability: 6, = 0.08m 3 This means that if we are in nucleate boiling at the surface, we can just remove decay heat purely by conduction through the debris slab at a thickness of 8 cm. The surface heat flux is: 9" = g'"6 = 1003fW / m' 3 The heat flux which would result from critical heat flux would be substantiall l higher than this value. Thus, one could view this as the lowest possible upward heat transfer given the boundary conditions. A higher temperature at the bottom of the crust or heat transfer into the crust would both increase the debris-to-water heat
- transfer, i
This rather low heat transfer would be increased if the surface was of non uniform l thickness (fin effects) or especially if the surface cracked sufficiently to allow water to ingress. 1 Issue 5: Provide a discussion of the susceptability of the RIPS to failure as a result of contact with molten core material during a severe accident. Response 5: See ABWR SSAR Figure 5.4-2 for a pictoral description of the location of the RIPS in the RPV. Figure 5.4-1 shows more RIP detail. Since the core melt progression is expected to contain the corium inside the core shroud, debris would not approach the RIP impellers or RPV RIP nozzles which are located outside the shroud. However, if the shroud is perforated by the corium, the corium might then enter the top of RIP impellers and possibly enter the stretch tube / shaft annulus. This is extremely unlikely since this annulus thickness decreases in the downward direction to 1.5mm (The variance between the 215mm diameter RIP shatt and the 218mm inside diameter of the stationary stretch tube.). Any molten material relocating through the RIP would quickly freeze or flow through the pump ather than flowing along the pump shaft. In the event that the corium did flow down the stretch tube / shaft annulus, the motor housing to RPV nozzle weld might fail allowing the RIP / motor to drop. I ABWR SSAR Figure 1.2-3b shows the two RIP vertical restraints which connect the bottom of each RIP motor housing to the RPV bottom head. These restraints i prevent the RIP / motor from dropping out of the RPV in case the motor housing weld falls for any reason. Therefore, m the exceedingly unlikely ever.1 of RIP failure, 1 wouId be small.the nump will not fall from the vessel, and the penetration through the vesse 1 CEB92 415 November 3,1992 L mn msm m t i-o s em ru ra
Neverthelea, the corium is expected to! freeze and, consequently, not flow down the annulus into the motor housing. Therefore, the RPV RlP nozzle motor housing Reactor Coolant Pressure Boundary would not be breached. Issue 6: Indicate the impact of suppression pooi bypass on the probability of recovery for the RHR system. Response 6: i It was noted that late RHR recovery is c'onsidered prior to pool bypass in the CET. Hence, the effects of pool bypass on the probability of RHR recovery are not explicitly considered m the model. The probabilities oflate RHR (non-) recovery under various conditions are summarized in Table 1 (see attached).;Th: 2 probabilities assess whether RHR is recovered prior to operation of COPS or overpressure structural fr.ilure of the containment. The probability of RHR recovery will vary for difTerenct accident subclasses, for sequences with the core damage progression terminated ia-vessel, and for sequences with active injection!to the lower drywell after RPV failure because of differences in the availability of AC power, sequence timing and other factors. Minimum RHR Recovery Probability with Pool Bvpass MAAP analysis indicates that the time dvailable for late recovery of RHR prior to COPS actuation would be 5 hours for sequences with pool bypass. Using the standard exponential non recovery formula for systems (Section 19D.5.4): p,, j-T/M) where: P, = Probability of failure to recover system (for one division), T = Time MITR = Mean Time to Repair (19 hours) For the three redundant divisions of RHR and considering potential common cause failures and limitations on the availability of operators, the probability of failure to recover 1 of 3 RHR divisions'was estimated using: P,;vy - n 5Pg Hence, with T = 5 hours and MTIR = 19 hours: i P,=0.4 s CEB92414 l November 3,1992 m" + _f
1 ( This non recovery probability is considered a bounding value which will be applied for sequences in all accident cluses under all conditions for a conservatin estimate of the impact of pool bypass and RHR recovery on COPS operation and t containment structural failure. Note that late RHR recovery is only an issue for those accident classes with RHR unavailable at core damaCe initiation.- Accident cluses IA,,,0 and Illa _0 have RHR anllable and need not be considered. . Furthermore, the impact of pool bypass on RHR recovery is not an issue for accident clus 11 since these sequences are transients where steam discharge occurs directly to the suppression pool. The corrected probability of RllR non recovery can be approximated from the previously calculated probability (which negheted the effect of pool byyr by adding an additional term to account for sequences which would others. c have had successful recovery. The probability of these sequences is multiplied o, the probability of bypass and the probability of recovery for sequences with pool bypass. Noting that the total pool bypass probability (for large and small bypass events) is 0.022 and assuming afl ool bypus events would decrease the recovery time to 5 hours, the adjusted RH non recovery probability is estimated as P,..., a pt.m + (1 - P,..) x 0.022 x 0.4 where: P,.m the existing non. recovery probabilities shown in the CETs and in Table 1, P,.m_n = the modified non-tecovery probabilities, 0.022 = the probabiliti) af pool bypus, 0.4 = the non recovety probability for sequences with bypass. The modified RHR non-recovery probabilities are shown on Table 1 for comparison with the existing values. ! Seouences with In vessel Core Damn e Mitication For trr.nsient sequences with core melt arrest in vessel steam discharges which occur will be directed into the suppression pool and the existence of a pool bypus pathway (open vacuum breaker) does not impact containment heatup and the time availabfe to recover RHR. Only for LOCA sequences tenninated in-vessel in accident cluses !!!A_1 and IllD would pool bypus potentially impact containment pressurization rates and the RHR recovery probability. However, the probability of class Illa _1 sequences with core melt arrest in. vessel is 3.87E - 12 x 0.05 = 1.9E -13 which is negligible. The probability of clus IIID sequences with core melt a.rrest in-vesselis: 2.lE - 10 x 0.9 = 1.89E - 10 CEB92-417 November 3,1992 ClassIB1 Existing Modified Results Rimb i CEB92-11-8 November 3,1992
,1 l No Cont Leak !2.11610 2.0SL10 RD Open ! 4.43D11 4.66L11 DW Head Fall 8.94L18 9.40L13 Class ID Existing Modified Results Results 2 No Cont Leak i 5.07L11 5.01L11 RD Open l.82L11 1.88E-11 S.67E 13 S.79L13 DW Head Fail 1 Conclusions The release category frequencies were modified to account for the changes in IB1 and ID sequences. The overall frequency of COPS operation is 2.08E8 and that for late drywell head structural failure is 5.25E-10. The impact of pool bypass on the late RHR recovery probability is to increase the frequency of COPS operation by 4.8512 (an increase of 0.02%) and of drywell head failure by 7.7E-14 (an increase of.02%). Thus, consideration of pool bypass in,the calculation of RHR recovery has no impact on risk, haue 7: Indicate the number of alloy blends which will be used in the passive flooder system. Response 7: 1 The passive flooder contains a fusible: plug which is a blend of several different metallic alloys. A blend is required to obtain a plug melting tecnperature of approximately 260C. A single blend will be used. No need exist.s for mtiltiple blends b:cause sample valves will be tested before installation and during periodic refueling outages to ensure that the desired melting temperature is achieved. The sentence in CEB92 X y hich erroneously referred to multiple blends will be I removed. Issue 8: Confirm that there are no openings in the support skirt which could allow water to flow from the upper drywell to the lower drywell. Response 8: There are no penetrations in the ABWR RPV support skirt. CEB92-41-9 November 3,1992 1 nw muu m . w.m wn w m
Iamri i Discuss the use of the firewater saray to provi6 flooding of the lower drywell. In particular, explain the potential for the operator to use the firewater in d:ywell spray mode as opposed to vessel injection mode. Response 9: The primary goal of the ac independent water addition system (firewater addition system) is to provide cooling water to the vessel in the event that all other means of vessel injection are inoperable. If the vessel has failed, water added to the vessel will drain into the lower drywell to allow fdr the quenching of the debris. If not t required to assure adequate core cooling, the ac-independent water addition system l may be operated in diywell spray mode to proside primary containment control. Drywell sprays can be used to reduce high drywell temperature, reduce high ] suppression chamber pressure, and provide steam inerting in the event of high containment hydrogen concentration. The only time the ac independent water addition system will be used for primary containment protection when adequate I cort cooling is not assured is when containment failure is likely to occur without i himediate corrective action. i f This operational philosophy of the ac. independent water additional system is i l discussed in the Emergency Procedure Guidelines (EPGs) contained in the ABWR SSAR Appendix 18A. The text in the EPGs will be modified slightly to ensure that the ac-independent water system is not removed from the service of assuring adequate core cooling unless containinent failure is hkely without immediate corrective action. Issue 10: I Providejustification for the probability of power recovery in the interval from 8 to 10 hours in the IB-2 sequences. -{ Response 10: The conditional probability of failing to recover offsite power in the two-hour interval from 8 to 10 hours is obtained from the EPRI KAG 9(7/88 draft) Table D2-2 (see attached). A value of 0.6 is obtained by dividing the non recovery probabilities for 8 and 10 hours. Issue lit Provide additionalinformation to indicate that the use of the fuewater addition system represents the most probable mechanism for the flooding of the lower drywell. CEB92-4110 November 3,1992 i h CM 4CPJC5!!97 1g r, -g-; 93:97 pg pgg
L Response 11: The following text will be added to the discussion of the pusive flooder system sequences in the Appendix 19E of the SSAR. Similar discussions will be added to the summary discussions in Chapter 19. i L The passive flooder system is tiesigned to cause the lower dqwell to be flooded when there is no water aver 16,g core debris in the lower dawell. If there is no overlying water pool the fusible material in the valve will h<:at up, and melt the fusible plug. If there is water overlying the debris pool, thr: lower dnwell will not heat up sufIlciently to cause the passive flooder to oven. Exammation of the Containment Event Trees (Subsection 19D.5.11) shows that the firewater addition system is expected to operate in most of tic accident sequences. Therefore, the passive !)ooder is dot needed in the majority of accidents. Rather, the lower def..ch flooder h viewd as a passive backup system which floods the lower drywell,in order;to keep the temperature in the dnwelllow, and in order to allow quenching of the core dehns. Issue 12: l i Confinn that the teflon disk is designed to slide out of the pLsive flooder rather than melting If necessary, consider the potential for the teflon disk to become stuck in the exit of the flooder line. Response 12: The teflon disk resides between the stainless steel disk and the fusible plug in the flooder valve. Its purpose is to insulath the fusible plug from the relathely cold suppression pool water. Ifinsulation was not provided, melting of the plug might not be uniform and operation of the flooder valve might be impaired. The disk will not melt or stick in the valve because teilon has a softening temperature of approximately 400 OC and a maximum continuous operating temperature of 288 oC both of which are above the plug melting temperature of 260 OC. Furthermore, teflon has hl h chemical resistance add will not adhere to the stainless steel plug S nor the fusible plug. Issue 13: i Provide information about the pressurization of the containment in the event of SRV discharge line failure. Note this was DCH Question 2 of the Staffs earlier letter: Providejustification that the reactor depressurization system is highly reliable during seismic events, and will assure a very low absolute frequency of high pressure reactor vessel failures 16 seismic events. This should include -l discussion of: (1) the impact of SRV discharge pipe failures on the ability to i depressurize (indicated to be a concern in draft section 19E.2.3.3.4), and (2) 1 quantitative estimates of the availability of wetwell sprays in these events. l 1 l CEB92-41-11 November 3,1992 4 FFCM 406M511M 11-O p n 03:g7 yg pig
I f TABLE 1 Existing Modified RHR Non-RHR Non-Core Melt Active Recovery Recovery AccMent Arrest in Injection Probability Probability Subclass M Iiwer DW .( Pr. 1 LPr.m.nl IA_l, IIIA_1 CM ARREST l 0.05 0.% NO ARREST IMJECT 0,01 0.02 NO ARREST NO INJECT 0.1 0.11 IBl_0 CM ARREST 0.0 0.01 NO ARFEST INJECT 0.01 0.02 NO.r REST NO INJECT 0.1 0.11 a IB2 0 CM ARREST 0.05 0.06 NO ARREST INJECT 0.1 0.11 NO ARREST NO INJECT 0.1 0.11 IB$_0 CM ARREST i 0.0 0.01 NO ARREST INJECT 0.05 0.06 NO ARREST NO INJECT 0.1 0.11 ID,IIID CM ARREST 0.2 0.21 NO ARREST INJECT Of 0.21 NO ARREST NO INJECT 0.2 0.21 i i i i l CEB9241-13 November 3,1992 F E M 4069251193 11 03:07 FM Pl? -- -.. - C S - 9 2
~. Response 15: In the unlikely event failure of the discharge 'alping docs occur, the ability of the reactor to depressurize will not be affetted, The ADS system would automatically initiate w'.nen the water levelin the vessel reacned Level 1. There are no directions to the o9erator which would direct the' ADS to be inhibited. Rather, failure of the dischargh piping would lead to a LOCA type condition except that the break ;ould be downstream of the SRV. i The issue addressed in part (2) of this' question will be discussed in a later submittal. Issue W Investigate the difference between the consequence results calculated by GE and BNL Response 14: Hal Careway at GE and Art Tingle at BNL have been in contact. In their discussions it was determined that the discrepancy in the consequence results ap acars to be the result of differences in the MACCS and CRAC codes. BNL ran a supp emental CRAC calculation which essentially duplicated the results of GE's evaluation, BNL is now investigating the differences between their MACCS and CRAC results. GE will suppert further dialog as required to resolve this issue, Issue 15; Provide additional information to support the analysis of the pedestal integrity if concrete ablation has occured. Response 15: Discussions were held between Gary Ehlert and Gutam Bagchi. In that meeting it was agreed that Gutam would look further at the GE submittals to date and indicate to GE if further information was required. GE will support further information and dialog as required to support this issue. i i i 1 November 8,1992 CEB92 41-12 4 I 93 11-0?-90 03:0? PM .F11
NOV, 3 '92 1: 06?!! GEIAbEAIAgd{ 18783 ?,13/20 4IJGGESTl!D ALWil REUAEILITY D ATA: INITIAL DRAIT (75.118) l I p \\ Table D2 2 CUMULATIVE NDN.RECDVERY PROBABILITl!S
- Probability of not Probability of not Time (hr) recovering power Time (hr) recovering power 0.5 0.48 13 7.3 x 10 3 1
0.36 t4 5.3 x 10 3 2 0.22 15 3.9 x 10 3 3 J.14 16 2.8 x 10'3 4 0.11 17 2.1 x 10 3 5 0.079 18 1.5 x 10 3 ~9 1,1 x 10'3 6 0.064 1 7 0.050 20 8.0 x 10 4 8 0.036 21 S.9 x 10** Q 0.021 22 4.3 x 10 4 4 10 0.021 23 3.1 x 10'd 11 0.012 24 2.3 x 10-4 12 0.010 Data es'imated from Ilcensee event reports, and repoded in NUREG'CR 13635 for valves. NUREG/CR 12056!for pumps, and NUREG/CR 13627 for d:esel generators: 1 Additional data compiled for diesel generators and reported in NUREG/CR-2989;8 l The data fer diesel ge nerators tiported in NSAC/108;8 The data base compiled for the Accident Sequence Evaluation Program.10 which is based largely on data from the; Reactor Safety Study;11 The data provided for the Nenheast Utilities system, as reported in the draft version pf the Al.WR PRA Key Assumptions and Groundrules Document:12 Military data for non nuclear installations reported in NPRD 2;13 and The data for some electrical cobponents and Instrumentation reported in IEEE. 500. l In addition, raw data were extracted from available sources for several speelf',c plants. These sources included the following: I The plant specific experience summarized in the Oconee PRA.3 e The data raported Ior indian Poidt Units 2 and'3 in the Indian Point PSS,13 The av. dance for Zion in the Ziob PSS,16 1 014 " ~
_
19EL DETERMINISTIC EVALUATIONS l .1 INTRODUCTION ) j .2 DETERMINISTIC ANALYSIS OFl PLANT PERFORMANCE 2.1 METHODS AND ASSUMPTIONS [ eWkp'tlon 2.1.1 2.1.1.1 MAAP3.0B 2.1.1.2 ABWR Modifications i 2.1.2 ABWR Configuration Basis '
- 4.2.1 ABWR Configuration Assumptions 2.1.2.2 Station Blackout Perfonnance 2.1.2.2.1 Sumnary 2.1.2.2.2 CoreCooling 2.1.2.2.3 Prirnary Containment Vessel (PCV) Integrity 2.1.2.2.4 Operator Actions 2.1.2.15 Recovery Following Restoranon of AC Power 1 Assumptio!.
Conclusions 2.1.2.2.6 2.1.3 Phenomenolo ns 2.1.3.1 Steam Ex osions 2.1.3.2 Degree o Metal-Water Reaction 2,1.3.3 Supples:lon Pool Bypass due to Additional Failures 2.1.3.4 Effect of RHR Heat Exchanger Failure in a Seismic Event 2.1.3.5 Radiation Heating of the Equipment Tunnel 2.1.3.6 Basemat Penetrauon i 2.1.3.7 Hydrogen Burning and ENplosions 2.1.4 Definition of Base Case Ar. sum ptions 2,1.4.1 Core Melt Progression and Hydrogen Generation 2.1.4.2 In-Vessel Recovery I 2.1.4.3 System Recovery After Vessel Failure and Normal Comainment Leakage 2.1.4.4 Early DrywellHeadFailure uences of Suplnession 'Jool Drain 2.1.4.5 Cons 1 2.1.4.6 Stuck nVacuum 3ttaker l 2.1.4.7 Containment Structural Failure Pressure l 2.1.4.8 Overpressure Relief Rupture Disk 1 2.2 - ACCIDENT SECUENCES ~ (Subminedin CEB92 39 on June 30) 2.2.1 LCLP ( 2.2.2 LCHP l 2.2.3 SBRC i l 2.2.4 LHRC 2.2.5 LBLC 2.2.6 NSCL l 2.2.7 NSCH 2.2.8 NSRC 2.2.9 Summary l 19 Outlina l Page1 Nevernber 3.1992 FRCM 4009251193
- t g3 g,
- c 3. p g p, j 4
2.3 _ JUSTIFICATION OF MNOMENOIAGICAL ASSUMPTIONS 2.3.1 Steam Explosions i ^ 2.3.1.1 (as originally samitted) ion Proc'ess The Steam Explos 2.3.1.2 Pmvlous Studies 2.3.1.3 Theoretical Considere.tlons A l 2.3.1.4100% pplication to ABWR Metal Water Reaction l 2.3.2 (as originally samitted) l 2.3.3 Suppression Pool Bypass Paths (Don Knecht's work} 2.3.3.1 Introduction 2.3.3.2 Identification and Description of Suppression Pool Bypass Pathways 2.3.3.3 Evaluation of Bypass Probability 2.3.3.4 Suppassion Pool Bypass Resulting from Extemal Event Analysis 2.3.4 Effect of RHR Heat Exc! sanger Failure in a Seirmic Event (as oririnally samitted) l 2.3.4.1 RHR Equipment Room Flooding 2.3.4.2 Dynamic LoadsInducclby Chugging 2.3.4.3 RHR Equi xnent Room Structural Integrity 2.3.5 Potential for F ashing During Venting (samittedin CEB9212 on April 2) 2.3.5.1 Critical Time Constants for Blowdown Response 2.3.5.2 Pool Swell 2.3.5.2.1 Pool Swell due to Suppression Pool Flashing 2.3.5.2.2 Pool $welldue to Flow Frorn Drywell 2.3.5.2.3 Steam Source ! 2.3.5.2.4 Application to ABWR 2.3.5.3 Carryover due to Entrainment 2.4 ADDITIONAL SEOUENCE ANALYSES INAME CH4NGEl (as oririnally samitted. duplicates with :tection 19E.2.6 will be remod: 2.4.1 Core Melt Progression and Hydrogen Generation 2.4.2 In. vessel Recovery 2.4.3 System Recovery after Vessel Failure and Normal Containment Leakage 2.4.4 Early Drywell Head Fallure; 2.4.5 Suppression Pool Drain 2.4.6 Stuck Open Vacuum Breaker 2.4.7 Containment Structural Fallure Pressure 2.4.8 Effect of Overpressure Relief Rupture Disk 2.4.9 Effect of Debris Coolability in the Lower Plenum 2.5 IDENTIFICATION AND SCREFNING OF PHENOMENOLOGICAL ISSUES (new. Samittedin CEB9212 starting p,10. Note title change) 2.5.1 Review of NUREG/4551 Grand Gulf and Peach Bottom Analysis 2.5.1.1 Grand Gulf 2.5.1.2 Peach Bottom i 2.5.1.3 Application of NUREO/CR-4551 Results to ABWR 2.5.2 Review of NUREG 1335 2.5.3 Review of Recommended Sensitivity Analyses for an Individual Plant Examination using MAAP 3.0B (EPRI), 2.5.4 Review of ALWR Requirements Document 2.5.5 Summary and Conclusions l I 19 oudms Pags2 November s.1992 l I I C 'I +23325II93 1 _ f ; 7, ; g3.g, p y, y g.. j C v c-
26 SENSITIVITY ANALYSIS AND SCOPINO STUDIES FOR PHENOMENOIDQICAL ' ISSITES l l 2.6.1 (new. S&mittedin CEB92 X) ion and Ilydrogen Generation Core Melt Progress F!s(new Samittedin CEB92 X}sion Product Release from Core t 2.6.2 (new. Samittedin CEB92 36) 2.6.3 Cal Re evaporation (new SamittedCEB92 36) 2.6.4 Time of Vessel Failure (new. S& mined CEB92 36) 2.6.5 Recriticality During In. Vessel Recovery (new SamittedCEB92 X) lity l Potendalfor Recridea 2.6.5.1 2.6.5.2 Implicadons of Recrideality 2.6.5.3 Conclusions 2.6.6 Debris Entrainment and Direct Containment Heating (new Sumittedin CEB92 36) 2.6.7 Fuel Coolant Interactions (Put Summry Here - Details in At'tachment 19EB) 2.6.8 Core Concrete Interaction and Debris Coolability (new Submittedin CEB92 36) l 2.6.9 Fission Product Release Location Fis(new suminedin CEB92 36)sion Product Release Flow; Area i 2.6.10 (new-Submittedin CEB92 X) 2.6.11 Suppression Pool Bypass l transferred to 19EE) \\ Ill(gh Temperature Fallure of Drywell 2.6.12 new. Samittedin CEB92 X) l Su(ppression Pool Decontamination Factor 2.6.13 (new-Samittedin CEB92.X) 2 DETAn Fn PHENOMENOLOGICAL UNCERTAINTY STUDIFS 2.7.1 Direct Containment Heatingj (Put summary here. details in Attachment 19EA) 2.7.2 Debris Coolability l (Put summary here - details in Att'achment 19EC) 2.7.3 Suppression Pool Bypass (Putsummaryhere detailsin Attachment 19EE) 2.8 - SEVERE ACCIDENT DESIGN FEATURE CONSIDERATIONS 2.8.1 Containment Overpressure Protection System (submittedin CEB92 X) 2.8.1.1 Pressure Setpoint Determ.nadon i 2.8.1.2 Variability in Rupture Disk Setpoint 2.8.1.3 Sizing of Rupture Disk ! 2.8.L4 Comparison of ABWR Performance with and without COPS 2.8.1.5 Supp:ession Pool Bypass; 2.8.1.6-Summuy 2.8.2 Lower Drywell Flooder O(submittedin CEB92 X Doug revising primarilyfor clarity)ium Shield) rium Protection for Lower Drywell Sump [Cor 2.8.3 (Put summary here details in Attachment 19ED) 19 omiina [ Page 3 November 3.1992 l FEOM 4009:51103 ig g3_gg. g) g7 pg pyg
29 REFERENCES 3, CONSEQUENCE ANALYSIS 3.1 SITE ASSUhWr10NS 3.1.1 ' Meteorology i 3.1.2 Population 3.1.3 Evacuation i 31 CR AC INPLTT DATA - 3.2.1 Input Which Differs From Standard CRAC Assumptions 3.2.2 Input to CRAC from Performance Analysis 3.3 COMPARISON OF RESULTS TO GOAIS 3.3.1 Goals f 3.3.2 Results 3.4 REFERENFFF. l EA. DIRECT CONTAINMENT HEATING (sulmtined in CEB92 39 revised on July 15)) i 1
SUMMARY
DESCRIPTION 2 DESCRIPTION OF EVENT TRFF ANALYSIS 2.1 Event Headings 2.1.1 Containment Pressure Pdor to RPV Failure (COhTPRES) 2.1.2 RPV Pressure at RPV Failure (RVPRES) 2.1.3 Mode of RPV Failure (MODRVFAIL) 2.1.4 Fraction of Core Inventory Molten in Lower RPV Head (RVCORMASS) 2.1.5 High Pressure Melt Ejecuon (HPhE) 2.1.6 Fraction of Entrained Debris Fragmented and Transported to the Upper Drywell (FRAO) i 2.1.7 Peak Containment Pressure Followin g RPV Failure 2.1.8 DrywellHeadFalls Following Vesse Failure i 3 DETERMINISTIC MODEL FOR JCH 3.1 Debris Dispersalin the ABY/R 3.1.1 Velocity Required to Transport Debris Particles 3.1.2 Argonne Experiments on Debris Dispersal 3.1,2.1 Experiment on'Zlon Configuration 3.1.2.2 Experirnents on Grand Gulf Configuration 3.1.2.3 Application to OE ABWR Configuration 3.2 Pressurization due to DCH I 3.3 Calculation of Vent Clearink Time 3.4 -- Calculation of Dispersal Time Constant 3.5 Application of DCH Model to ABWR 3.6 Sensitivity to Various DCH Parameters 3.6.1 Base Case 3.6.2 DispersalTime Constant 3.6.3 Debds Temperature 3.6.4 Nodalization i 4 19 outlins { Pip 4 November 3.1992 I F M :40 m I 93 gg_gg_gg g3 g7 pg. p;7
3.6.5-Zr OxidadoriExaVessell 3.6.61 InidalDrywell steam Pracdon : 3.6.7 11 en Combustion-3.6.8 : Vent caring., 14 -. CONTAINMENT ULTIMATE ST ENGTH AND UNPRRTAINTY - 4.1 - Ultimate Strength. __ l 4.2. Uncertainty in the Failurt Pressure f 5
SUMMARY
OF RRRULTS 5.1 Quantification of Decomposition Event Trees : 5.2 Impact on Containment Fallure Probability 5.2.1 Sensiti of Containment Failure Probability to Assumptions-5.3 Impact on teDose 6 CONCLUSIONS 7-REFFRENCFS l ,I EB. FCI i (submuudin CEB92-X) 1 1 INTRODUCTION l 1 APPIlCA111 ITY OF EXPERIMENTS 2.1 Fuel CoolantInteraction Tests - 2.2 Experiments With a Stratified System 2.3 BETA V 6.1 2.4 High Pressure Melt Ejection Experiments 3 EXPLOSIVE STEAM GENERATION 3.1 Phenomenology i 3.2 Bounding Analysis MaximumImpulse Pressure l; IMPULSE M 4 - 4.1 -- 4.2 Impulse Duration -1 4.3 -Pedestal Capability 1 .4.3.1 Elastic-Plude Caladatio'n .4.3.2 - Comparison to NUREG 1150 Orand Gulf Pedestal ' 4.4 l Capabliity of the ABWR to Withstand Pressure Impulse a -5 ' WATER MISRT11S' L 5.1 -Maximum Lise Het ht. 5.2 L Available Rise He t-5.3 Capability of AB to Withstand WaterMisalles 6 CONTAINMENT OVERPRRASUPI7ATION 1 xl eam Generati Rates - --, 6.2.1. Water Added to Debris j.
- 6.2.1.1
' WaterInventory from Lower Plenum 6.2,1,2 Passive Flooder Flow 1 19 Ousitne Page$ kvanber 3,1992 .i FROM 4089051193- _gg.0?-S2 2 03ic7'PMt Pie _ '..........s..........
63.1.3 ECCS and Firewster Flow 6.2.2 Corium Pou? from Vessel into 1% existing Pool of Water 6.2.2.1 Probability of Pnettooded 14wer Drywell 6.2.2.2 Steam Generation Rate for Pre flooded lowei Drywell ) 6.2.3 Explosive Steam Generation Rates 6.2.4 Maximum Steam Generation 6.3 Containment Pressurization i 6.3.1 Drywell Connecting Vent Flow 63.2 ' Vent Clearing 6.3.3 Horizontal Vent Kow 6.4 Summary of Overpressurization Limits i EC CORE CONCRLTEINTERACTION AND DEBRIS COOLABILITY (submittedin CEB92 X) i ED. CORIUM SHIELD (su'xnittelis, CEB92-47 on August 7) f 1 ISSUE 2 PROPOSED DESIGN 3 SUCCESS CRITERIA FOR PRO SED DESIGN 4 ANALYSIS OF SHIFI D FRFF71NG ABILITY 4.1 Assumptions I 4.2 Initial Freezing of Molten Dibris in Channel 4.3 Required Channel Length to Insure Freezing 5 LONG-TERM ABII TTY OF DEBRIS TO REMAIN SOI.TD 5.1 Upper Shield Wall (Above Lower Drywell Floor) j 5.2 Lower Shield Wall (Below Lower Drywell Floor) 1 6 EXAMPLE CALCULATION 7 DETATI FD DESIGN ISSUES 8 REFERENCFS EE SUPPRESSION POOL BYPASS (combination of submittab in CEB92-36 and CEB92 X) I 1 INTRODUCTION I l 2' DESCRIPTION OF DECOMPOSITION EVENT TRER ANALYSIS 2.1 Vacuum Breaker (V/B) Stuck Open 2.2 VacuumBreakerLeaks(VB LEAK) 2.3 Ativsols Plun Leakage PathILEAK PLUG) 2.4 Suppression 7001 Bypass (POOL _BP) DETERMINISTIC ANALYSIS f. 3 l 19Oudins Page6 November 3,1992 i rFou scegut;9, 9E 03 M N P19
a 3.1 Method 3.2 Results 3.2.1 Late Suppression Pool Bypass with no Plugging 3.2.2 Pwexisung Suppression Pool Bypass with no Plugging Late Supxession Pool Bypass with Pluggin3'1ugging 3.2.3 Pre-exirt.ng Sup1xession Pool Bypass with 3.2.4 - 3.2.5 Suppression Pool Bypass with Drywell Spray 3.3 Conclusion of Deterministic Analysis 4
SUMMARY
OF RFSULTS Impact of Release Fractions } Quantification of DET 4,1 4.2 4.3 Impact on Time to Rupture Disk Opening 5 CONCLUSIONS l i 6 REFERENCES i i i l i i I 1 l l l f I 19 Outline P86: 7 November 3.1992 .l 'L i non 4cesas t : r,3 .,}}