ML20246P944
| ML20246P944 | |
| Person / Time | |
|---|---|
| Site: | Limerick  |
| Issue date: | 08/04/1989 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20246J241 | List: |
| References | |
| NUDOCS 8909110276 | |
| Download: ML20246P944 (112) | |
Text
{{#Wiki_filter:_ _ . _ _ _ . _ _ - - . . _ _ . f 4 ENCLOSURE ( DRAFT A PRELIMINARY ASSESSMENT OF BWR MARK II CONTAINMENT CHALLENGES,-FAILURE MODES, AND POTENTIAL IMPROVEMENTS IN PERFORMANCE August 4, 1989 J P S_-_--_--___--__
t 4 4 ABSTRACT This report reviews challenges to the integrity of the BWR Mark II containment that could arise from severe accidents. The challenges are organized into two broad groups: those where containment integrity is challenged prior to extensive core damage, and those where core melt occurs first, with containment integrity - not threatened until the time of reactor vessel failure or later. Also reviewed are some potential improvements that have the potential to either preven.t core damage or containment failure, or to mitigate the consequences of such failure by reducing the releasa of fission products, and thus the off-site consequences. For each of these proposed improvements, a qualitative analysis of the impact upon core melt frequency and risk is given. a. m 9 4 ii
e EXECUTIVE
SUMMARY
In SECY-87-297, dated December 8, 1987, the staff presented to the commissLon its program plan to evaluate generic severe accident ~ contalment vulnerabilities in a program entitled the
' Containment Performance Improvement (CPI) program. This effort is predicated on the conclusion that there are generic severe accident challenges to each light water reactor (LWR) containment type that-should be assessed to determine whether additional regulatory guidance or requirements concerning needed containment features is warranted, and to confirm the adequacy of the existing commission policy. The bases for the conclusion that such assessments are nee %d include the relatively large uncertainty in the ability of LWR containments to successfully survive some severe accident challenges, as indicated by draft NUREG-1150. All LWR containment types are to be assessed beginnning with the boiling water reactors (BWR) with Ma;-)t I containments. This effort is integrated closely with the Individual Plant Examination (IPE) program and.is intended to focus on resolving hardware and procedural issues related to generic containment challenges. New regulatory requirements from this program, if any, would be developed consistent with the safety goal and backfit rule and would consitute closure of generic containment performance issues. The present report concerns plants with a boiling water reactor and a Mark II containment design.
This report focuses on identifying the potential chal2.enges to containment integrity that can arise from a severe accident and the potuntial improvements that could reduce the frequency of a severe' accident or mitigate the off-site consequences in the event that a severe accident should occur. The impact of these improvements upon core melt frequency and risk is examined qualitatively.
.As the result of the large phenomenological uncertainties, and the state of flux of the ongoing research efforts, the conclusions about potential improvements contained in this report should be viewed as tentative. The estimated costs for selected improvements were taken from previously published information. s They are not meant to be interpreted as final estimates, since.no
.. cost-benefit analysis was performed for this report. -
Severe accident sequences at the Mark II plants can be grouped into two general categories: one in which containment integrity is challenged prior to core melt (ie core degradation), the other in which core melt precedes any threat to containment integrity. In the first category, which includes loss of long term containment heat removal systems (TW) and anticipated transient without scram (ATWS) sequences, the challenge to containment is from over-pressurization due to inadequate containment heat removal. In uhe second category, which includes station blackout and other transients where reactor scram occurs, the challenge can be from either over-pressurization at or near the time of reactor vessel lii
failure, or from over-pressurization or thermal failure several
)
hours after vessel failure. These later challenges generally occur as a result of core-concrete interactions in the containment but may also be associated with late core-concrete-pool events, especially with the Shoreham and Nine Mile Point 2 pedestal designs.
~
Potential improvements for the first category of containment challenges include containment pressure control, e.g. , venting from the watwell through a hardened vent pipe, and containment pressure control and fission product scrubbing, e.g. , the use of containment sprays with a backup water supply. A hardened vent line would allow excess heat in the containment to be rejected to the environment, while avoiding the concerns associated with venting through existing " soft" HVAC ductwork. However, with the high likelyhood of suppression pool bypass via the in-pedestal drain line failure 3hortly after vessel failure, the vent systems need to be a filtered vent, such as the multi-venturi scrubbing system. Containment sprays could be used to condense steam in the containment, thus delaying over-pressurization failure. Use of a backup water ' supply for the sprays would avoid any problems associated with pumping saturated water from the suppression pool. Use of the suppression pool cleanup (SPCU) system to remove water from the suppression pool would prevent failure of the containment in the.wetvell. The reactor water cleanup (RWCU) has been shown by one utility to be capable of cooling the suppression pool suf-ficiently to eliminate the need to vent containment for the TW sequences. For the second category of containment challenges, proposed improvements include containment pressure control, e.g. , a hardened vent from the watvell, improved means to depressurize the reactor, e.g. , enhancements to the ADS and the SRVs, containment temperature control and fissien product scrubbing, e.g., containment sprays with a backup water supply, enhanced operability of the SPCU for removal of suppression pool water and RWCU for decay heat removal, and external cooling of the drywell head, and mitigation of the fission product release, e.g., use of fire protection sprays to enhance fission product retention in the roactor building. The ' hardened vent line (with or without an external filter) . could be used to mitigate late over-pressurization challenges. Enhancements to the ADS and to the SRVs would lower the probability of high pressure vessel failure, thus lessening the contribution of high pressure sequences (such as TQUX, which is a loss of feedwater transient with the loss of all high pressure injection capability and the inability to depressurize the reactor) to the core melt frequency. Containment sprays have the potential to mitigate both late over-pressurization and late thermal challenges. However, a backup water supply would be needed. In addition, the minimum flow rate that would be required in order for mitigation via the containment sprays to be effective has not yet been identified. External cooling of the drywell head by flooding it with water has iv
e. the potential to mitigate late thermal challenges, as well as to l scrub any fission products released through leakage in the drywell l- head seal. In fact, leakage past the drywell head seal night be i considered to be an alternative metnod of " venting" containment, with scrubbing of the release by allowing the containment to " burp" into the flooded refueling water cavity. Finally, some plants may l have the ability to spray large areas of the resetor building using " l the installed fire protection sprays. In the event of primary containment failure or venting into the reactor building, these j sprays could scrub aerosol fission products from the secondary { containment atmosphere, lessening the off-site consequences of a 1 release. J The table below summarizes the qualitative benefits, as well as any negative aspects, for each of the proposed improvements. e9 9 e i 1 V
e 1, Qualitative Benefits and Drawbacks of Proposed Mark II Containment Ignrovements Proposed Qualitative Potential Improvement Benefits Lrawbacks HOrd-pipo o Mitigates Over- o High cost ($2.9M at watwell vent Pressure (OP) Pilgrim) - i failure for TW, o Not useful for SBo, and some thermal failure ATWS sequences o High likelyhood of pool bypass - release fission products External o Mitigates OP o Very high cost Filtered Cont. failure for TW, for Filtra ($10-50M) Vcnt *, SBO, and most o High cost for MVSS (MVSS,Filtra) ATWS sequences ($5M) o can mitigate thermal failure o All releases scrubbed Enhanced o Loweru o None identified SRV/ ADS (with ' probability of d dicated power vessel failure cupply) at high pressure o Better ATWS - mitigation o Relatively low cost ($0.5M) Cont. spray o Mitigation of o High cost if flow >> oystem with OP failure 500 gpm bnckup water o Potential to o Cost to use fire cupply mitigate syst2m uncertain - thermal failure o Use for cooling o Aerosol ex-vessel debris - scrubbing' constrained by EPG o Independent of spray initiation RER limit Alternate o Maintain pool o Very high cost containment subcooling for ARNR ($183M+)" hcat removal o Prevents TW cystem challenge o Mitigates some ATWS sequences o Low cost for RWCU/SPCU ($100K) vi l 1 *
.o-l 1
Proposed Qualitative. Potential Improvement Benefits- Drawbacks Cooling to DW o Mitigates cLCooling must be head thermal initiated manually failure, before vessel failure especially . during SBO o All releases from DW head seal are scrubbed R0 view of o Decreases o None identified manual valves probability of cgainst GDC in cont. bypass 10 CTR 50, Appendix A Icproved H2 o None for Mark o Decreased -de-inerted control II plants time inhibits RCS (inerted) leakage identification
-Uce of fire o Scrubbing of o Not useful at all protection fission plants because of cprays in RR products in RB limited spray coverage.
o Available fire
, systems may be used for,, vessel injection j or cont. sprays O
vii
~
48 ACRONYMS AC Alternating Current LPCI Low pressure coolant ADS Automatic injection depressurization system LPCS Low pressure core' spray ARHR Alternate residual heat LOCA Loss of coolant accident removal LSCS La Salle County Station ATWS Anticipated transient MSIV Main steam isolation , without scram valve CWR Boiling ~ water reactor NMP-2 Nine Mile Point. Unit 2 LWROG BWR owners group NPSH Net positive suction CRD Control rod drive head CS Core spray- ORNL Oak Ridge National CST Condensate-storage tank Laboratory DBA Design basis accident PCPL Primary containment DC Direct Current pressure limit DCH Direct containment PCS Power conversion system heating <* PRA Probabilistic risk DF Decontamination factor assessment ECCS Emergency core cooling RBSVS Reactor building standby systems ventilation system EOP Emergency operating RCIC Reactor core isolation procedures cooling
-EPG Emergency procedure RERS Reactor enclosure guidelines recirculation system ESF '
Engineered Safety RHR Residual heat removal Feature SBO Station blackout FSAR Final safety analysis SCS Supplemental containment report system. HCTL Heat capacity SGTS Standby gas treatment temperature limit system . HPCI High pressure coolant SLC Standby liquid control injection SNL Sandia National Lab HPCS High pressure core SNPS Shoreham Nuclear Power spray . Station HVAC Heating, ventilation, SPCUt. Suppression pool cleanup and air conditioning SRV Safety relief valve INEL Idaho National SSES Susquehanna Steam Engineering Laboratory Electric Station . IPE Individual Plant TAF Top of active fuel Examination TW Loss of long-term LGS Limerick Generating containment heat removal Station WNP-2 Washington Nuclear Lilco Long Island Lighting Co Project Number 2 viii
-1
- 0 e
CONTENTS ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . ii EXECUTIVE
SUMMARY
. .. . . . . . . . . . . . . . . . . iii ACRONYMS . . .. . . . . . . . .. . . . . . . . . . . viii
- 1. INTRODUCTION . .. . . . . . . . . . . . . . . . . . 1
- 2. DESCRIPTION OF MARK II PLANT FEATURES AND CONTAINMENT DESIGNS . . . . . . . . . . . . . . . 3 2.1 Reactor Design . . . . . . . . . . . . . . . . . . . 3 2.2 Primary Containment Design . . . . . . . . . . . . . 5 2.3 Secondary Containment Design . . . . . . . . . . . . 9
- 3. CONTAINMENT CHALLENGES PRIOR TO CORE MELT . . . . . . 26 3.1 Definition of Sequences . . . . . . . . . . . . . . . 26 3.2 Discussion of Containment Challenges and Failure Modes . . . . . . . . . . . . . . . . . . 29 3.3 Potential Improvements . . . . . . . . . . . . . . . 33 3.3.1 Containment Venting . . . . . . . . . . . . . . . . 33 3.'3.2 Containment Sprays and Backup Water Supply . . . . 36 3.3.3 RHR Heat Exchanger capacity . . . . . . . . . . . . 37
- 4. CONTAINMENT CHALLENGES AFTER CORE MELT .. . . . . . 42 4.1 Definition of Sequences . . . . . . . . . . . . . . . 42 4.2 Definition of Containment Challenges and Failure Modes . . . . . . . . . . . . . . . . . . 43 4.2.1 Challenges at or Near the Time of Vessel Failure . 43 4.2.1.1 Overpressure Challenges . . . . . . . . . . . . 44 4.2.1.2 Rapid Steam Pressurization . . . . . . . . . . . 44 4.2.2 Challenges after Vessel Failure . . . . . . . . ,. . 49 4.2.2.1 Late Overpressure Failure . . . . . . . . . . . 49 4.2.2.2 Late Thermal Failure . . . . . . . . . . . . . . 53 4.2.3 Discussion of Containment Failure Modes . . . . . . 54 4.3 Potential Improvements . . . . . . . . . . . . . . . 56 4.3.1 Mitigating Transients with Scram . . . . . . . ,, . 56 4.3.2 Hydrogen Control . . . . . . . . . . . . . . .'. . 60 4.3.3 Containment Sprays and Backup Water Supply . . . . 61 4.3.4 Containment Venting . . . . . . . . . . . . . . . . 62 ix
O e 4.3.5 Core Debris Control . . . . . . . . . . . . . . . . 63 4.3.6 Enhanced Reactor Building Fission Product Attenuation . . . . . . . . . . . . . . . 66 4.3.7 Enhanced Reactor Depressurization capability . . . 67 4.3.8 External Cooling of the Drywell Head Seal . . . . . 68
- 5. CONTAINMENT BYPASS . . . . . . . . . . . . . . . . . 70
~
5.1 Definition of Challenges . . . . . . . . . . . . . . 70 5.2 Potential Improvements . . . . . . . . . . . . . . . 71 , 6. RISK ANALYSIS OF CONTAINMENT CHALLENGES AND IMPROVEMENTS . . . . . . . . . . . . . . . . . . 72 6.1 Core Melt Frequency . . . . . . . . . . . . . . . . . 72 6.2 Sequence and Failure Mode Risk Significance . . . . . 72 6.3 Summary of Potential Improvements . . . . . . . . . . 75
- 7. REFERENCES . . . . . . . . . . . . . . . . . . . . . 90 APPENDIX A-ROLE OF THE BWR OWNERS GROUP EMERGENCY PROCEDURE GUIDELINES IN SEVERE ACCIDENT MANAGEMENT . . . . . . . . . A-1 O
O O e X
- - - - -- - - - -- _ _ _ _ . . _ _ _ _ _ _ __ ._________.______________________o
d e A PRELIMINARY ASSESSMENT OF BWR MARK II CONTAINMENT CHALLENGES, FAILURE MODES, AND POTENTIAL IMPROVEMENTS IN PERFORMANCE
- 1. INTRODUCTION In SECY-87-297, dated December 8, 1987, the staff presented to the Commission its program plan to evaluate generic severe accident containant vulnerabilities in a program entitled the Containment Performance Improvement (CPI) program. This effort is predicated on the conclusion that there are generic severe accident challenges to each light water reactor (LWR) containment type that should be assessed to determine whether additional regulatory guidance or requirements concerning needed containment features is warranted, and to confirm the adequacy of the existing Commission policy. '.The
- bases for the conclusion that such-assessments are needed include the relatively large uncertainty in the ability of l
LWR containments to successfully survive severe accident challenges, as indicated by draft NUREG-1150'goneThe present report concerns plants with a boiling water reactor (BWR) and a Mark II
. containment design. Previously, the CPI Program has analyzed potential improvements for BWRs with Mark I containments.3 Future and in-progress CPI studies will address BWRs with M6rk III containments, pressurized water reactors (PWRs) with ice condenser ccintainments, and PWRs with large dry containments, both atmospheric and subatmospheric.
The present report focuses on dominant severe accident challenges, as identified by current severe accident research, which can threaten Mark II containment integrity. Potential improvements are evaluated as to their ability to arrest the core melt progression, prevent or delay containment failure during postulated severe accidents, or mitigate the off-site health consequences of a fission product release. Consequently, a risk analysis has to be performed in order - to correlate containment ! challenges, resulting consequences, sequence frequencies, and potential improvement benefits. Potential improvements and , benefits are considered for each containment challenge. J A preliminary qualitative risk analysis is presented te relate severe accident sequence frequencies, containment failure mode probabilities, and the magnitude of the off-site consequences. The risk from operation of a nuclear power plant is the sum over all sequences of the frequency of the accident frequency times the conditional probability of each potential containment release mode for each accident sequence times the mean magnitude of the consequence, given the fission product source term for the particular combination of the release mode and the sequence.' consequently, all factors affecting plant risk should be considered in a program to improve containment performance. 1 J
~
o. The BWR Mark II plants and their important safety design features, along with the differences and similarities between the various plants, are discussed in Section 2. In Sections 3 through 5, the important accident sequences and containment failure modes are presented. Section 3 discusses those sequences, originating from inadequate containment heat removal, that could challenge the containment integrity prior to extensive core damage. These . sequences include anticipated transients without scram (ATWS) and sequences with loss of long-term heat removal (TW), Section 4 discusses sequences where core melt occurs prior to a significant challenge to containment integrity. Station blackouts (SBO), loss of coolant accidents (I4CA), and transients with a loss of' injection are the primary contributors to this sequence class. Section 5 briefly considers primary containment bypass sequences, which result from failure.of low pressure valves and piping that release reactor coolant outside containment. Section 6 qualitatively" analyzes existing risk profiles of BWR Mark II containments and potential improvements. The preliminary risk analyses are based on existing BWR Mark II containment probabilistic risk assessments (PRAs) performed at Limerick Generating Station sand Shoreham Nuclear Power Station,7, ' and on the Individual Plant Examination (IPE) at Susquehanna Steam Electric Station.' Once the source terms have been identified, a MACCS computer analysis will be performed to determine tho' risk consequences. . The risk consequence analysis is expected to be complete by the end of 1989.
%D 4
#9 e
2 a
'e k
9
- 2. DESCRIPTION OF MARK II PLANT FEATURES AND CONTAINMENT DESIGNS A general summary of design information for the U. S. BWR nuclear power plants with Mark II containments is presented in this section. As indicated in Table 2.1, there are 9 Mark II nuclea power plants in the U. S. at six sites.' Eight of the nine '
reactors have been licensed for operation in the 1980's. As seen { in the table, many different architect engineers and construction . firms were used to build the nine plants. Design similarities and differences are presented in Tables 2.2 and 2.3. Section 2.1 discusses and compares general features of the reactors. The discussion is limited to general reactor design characteristics and the ' safety systems for water injection. Similarly, the primary and secondary containment designs are discussed in Sections 2.2 and 2.3, respectively. When available, plant-specific design information is presented. Plant-specific information was taken from the Final otherwise Sa noted.gtg'gysis Reports (FSAR) for six plants unless 2.1 Reactor Design BWR Plants with Mark II containments feature the General Electric BWR/4 and BWR/5 reactor product linss. Table 2.2 summarizes some of the important reactor design information for the reactors in the Mark II plants. Three of the sites, Limerick, Susquehanna, and Shoreham, use BWR/4 reactors while the other sites usie BWR/5 reactors. The thermal. power ratings for the nine plants are very similar except for Shoreham, which has a smaller vessel, a smaller number of fuel bundles, and a lower. thermal power rating than the other Mark II plants. A comparisori of emergency core cooling systems (ECCS) is also included in Table 2.2. The BWR/4 reactors feature turbine-driven high pressure coolant injection (HPCI) with DC controllers, AC-powered low pressure core sprays (LPCS), and an AC-powered low pressure coolant injection (LPCI) system. As noted in Table 2.2, LPCI at Limerick injects into the core shroud rather than into the recirculation lines (typical BWR/4 LPCI injection location). The BWR/S reactors use a different ECCS featuring AC-powered high . pressure core sprays (HPCS) with backup power from a dedicated i diesel generator, AC-powered low pressure core sprays (LPCS), and an AC-powered LPCI system. Unlike the BWR/4 ECCS configuration, the HPCS and LPCI injection is over the core and into the core bypass region, respectively. It is important to note that BWR/5s do not use turbine-driven high pressure ECCS pumps. Other high pressure injection systems common to both reactor models include the condensate /feedwater system, the reactor core isolation cooling (RCIC) system, and the control rod drive (CRD) hydraulic system. The RCIC and CRD systems are not part of the ECCS and have a lower makeup flowrate than the ECCS. However, in postulated high pressure severe accidents, these systems may be 3
o a k
'important sources of high pressure makeup flow. The RCIC makeup flowrates are included in Table 2.2. The turbine-driven RCIC system delivers approximately 10% of the HPCI/HPCS flowrate. A survey of plant-specific CRD flowrates was not made. However, at Limerick, the CRD injection rate during normal operations is a maximum of 63 gpm. However, with optimum manual valve lineup, each CRD pump could deliver more than 100 gpm to the reactor vessel. ,
All the Mark II plants include an automatic depressurization system (ADS) as part of the ECCS to depressurize the vessel and allow low pressure ECCS to inject water. Upon receipt of an ADS initiation signal, the ADS opens a subset of the safety / relief valves (SRVs). Effluent leaving the vessel through the SRVs is piped to spargers near the bottom of the suppression pool. Discharging effluent from the SRVs into the bottom of the suppression pool maximizes the condensation of the steam and the scrubbing of any non-noble gas fission products. The SRVs are grouped into banks of valves that operate in unison to protect the vessel'from over-pressurization. Each SAV bank has a successively increasing pressure setpoint to provide graduated pressure relief with increasing reactor system pressure. Once the vessel is depressurized, two redundant ECC systems of low pressure. injection are available, LPCS and LPCI. As seen in Table 2.2, the LPCI is a high capacity injection source while LPCS has a lower injection capacity. At the BWR/5 plants and Limerick, the licensees have verified with the manufacturers that the LP,CI pumps are capable of pumping saturated water without failure. Conversely, the other BWR/4 LPCI pumps are postulated to fail while pumping saturated liquid. (Alternate, or non-ECCS, injection systems are discussed in other sections of the report related to proposed improvements and plant response during postulated transients.) Reactivity control in the event of an accident is provided by two redundant systems: the CRD system and the standby liquid control (SLC) system. During conditions that call for a reactor scram, the CRD system rapidly inserts the boron carbide ( B,C) control rods into the core. The Alternate Rod Insertion (ARI) ' system provides a backup scram signal should the Reactor Protection System (RPS) signal fail. If the CRD system totally fails to control system reactivity in response to the RPS and ARI signals, several manual operator actions, including initiating SLC, are prescribed by the BWR EPGs (described in detail in Section 3.1) . The SLC system injects a sodium pentaborate solution into the lower plenum just below the core plate (Limerick injects SLC into the core spray sparger) . A more detailed discussion of reactivity control during accidents is provided in Section 3.1. 2.2 Primary Containment Design 4
*~
,j The SWR Mark II containment consists of two regions, the drywell and the wetwell (see Figure 2.1). The wetwell region lies directly below the drywell and is separated from it by the drywell floor. Vertical downcomers with a nominal 2 foot diameter connect the drywell airspace to the suppression pool (WNP-2 empic.';s 84 2-foot downconers and -18 28-inch downconers) . When the drywell airspace is pressurized, gases from the drywell are forced tt-ough these downconers into the suppression pool. Since the downconers -
discharge below the water level of the pool (approximately 8 to 12 j feet), all effluent entering the wetwell must pass through the suppression pool. The benefits of the suppression pool include i (a) scrubbing of the non-noble gas fission products, (b) a source { of water for the ECCS, sink for stenin condensation. and a(c)140,000 For example, a largeft, heat pool is capable of absorbing 100 MW-hr of energy with only a 40*F temperature rise." Table 2(3, summarizes general containment design information for the six Mark II plant situ. The total free volume is approximately the same at Limerick, Susquehanna, and Im Salle (within 304). Shoreham, due to its smaller reactor thermal power Washington Nuclear Project rating, has a smaller containment. Number 2 (WNP-2) has the same pool volume as the larger Mark II plants, but har., a smaller total containment volume. Comparison of the containment free and pool volumes to thermal power rating also illustrates that'the Shoreham containment free volume appears to be sized consistently with the other plants, but the pool volume ratio is slightly lower. The reactor pressure vessel is supperrted by an annular pedestal that extends from the containment basemat through the drywell floor to the vessel. The design of this in-pedestal region varies from plant to plant (see Figures 2.1 through 2. 6) . The Shoreham and Nine Mile Point 2 (NMP-2) containments have downcomers inside the pedestal region (referred to as in-pedestal downcomers) . At La Salle, WNP-2, and Nine Mile Point 2, the in-pedestal region is recessed relative to the drywell floor. WNP-2 has two sumps cast into the in-pedestal floor. The design of the in-pedestal region can have a significant influence on severe accidents with debris discharge onto the drywell floor. The presence of - downconers in the in-pedestal region could allow relocation of a larger fraction of the corium debris to the suppression pool than in plants without in-pedestal downcomers (see Figures 2.7 and 2.8) . This design could eliminate problems associated with core-concrete interactions (CCI) but there could be significant problems with fuel-coolant interactions (FCI). Conversely, a recessed cavity 1 would tend to retain more of the corium in the cavity. La Salle has the largest recessed cavity, with the cavity capable of holding approximately two entire core volumes (see Figure 2.9). This design would probably result in the maximum CCI and the minimum potential for cooling the core ex-vessel. The Limerick and Susquehanne plants have a flat in-pede,stal floor at approximately the same elevation as the ex-pedestal floor. This design would 5
- J
.e I
l result in significant CCI, large uncertainties associated with the timing of suppression pool bypass, but does have the potential for ex-vessel cooling of the corium with containment sprays. However, j all designs have drain lines into the wetwell, except for i Susquehanna, und thereby provides a high likelyhood of suppression f pool bypass within 10 to 20 minutes after reactor vessel failure. The containment cavity designs for the six sites are summarized in - Figurew 2.2 through 2.10. The doc ;omer pipes extend approximately 3 in. to 18 in. (site-specific) above the drywell floor to prevent clogging. (Note: Shoreham's in-pedestal downcomers extend only 0.25 in, above the
- floor of the in-pedestal region.) The ex-pedestal downcomers for all plants have a steel cover to prevent direct impingement damage and localized suppression pool heating during loss of coolant accidents (IOCAs) . The number of downconers varies from site to site as indicated in Table 2.3. Other penetrations in the drywell floor include floor drains (see Figure 2.9), sumps, and SRV penetrations.
Figure 2.10 shows the drywell floor layout and typical floor penetrations at the Susquehanna Steam Electric Station. While the in-pedestal drywell floor at Susquehanna does have a thin steel liner, this liner is not considered to significantly delay intiation of CCI. At Shoreham, an inflatable seal is used to connect the drywell floor to the side of the containment. WNP-2 also uses a slightly unconventional method of sealing the drywell floor to the wall of the primary containment. There, a steel " ring girder!" embedded in the drywell floor, is welded to a steel peripheral seal assembly, which in turn is attached to the liner of the primary containment. Failure of a drywell floor penetration, the drywell floor seal, or the floor itself (by core-concrete attack or from excessive differential pressure across the floor) would allow fission products in the drywell to bypass the suppression pool. The wetwell airspace communicates with the drywell through normally closed vacuum breakers. These vacuum breakers are designed to (a) open when there is a positive wetwell ' airspace-to-drywell pressure difference (typically begin opening between 0.25 and 0.50 psid) , and (b) provide sufficient flow to : maintain the pressure difference below the drywell floor upward , design pressure (see Table 2.3). The mitximum , upward design ) pressure loading for the drywell floor ranges from 5.0 to 10.0 psid. No information is currently available as the dryvell floor's ultimate pressure loading capability nor the relationship of the floor capability relative the penetrations seals capability. The containment could also be damaged by a negative pressure difference between the primary containment and the reactor building. However, the BWR Emergency Procedure Guidelines (EPGs) provide procedural guidance for situations where rapid condensation 6
.. y e
could reduce the primary containment pressure too rapidly.is The primary containment external design pressure ranges from -4.7 to
-10.0 psaig. Only WNP-2, which has three 24-inch vacuum breakers connecting the reactor building to the watwell airspace, has vacuum breakers to equilibrate the pressure between the primary containment and the secondary containment.
The Mark II containments were constructed using three general - methods." Susquehanna 1 and 2, Limerick 1 and 2, and Nine Mile
. Point 2 were constructed with the following features: deformed steel bars with concrete reinforcement, a flat base, a 1/4 to 3/8-inch thick steel internal liner, and a steel done closure cap.
La Salle 1 and 2 were constructed from prestressed concrete with a steel cap, steel liner, and a flat base. WNP-2 was built with steel surrounded by concrete with an ellipsoidal base. The primpry containment design pressure ranges from 45 to 56 psig. Ho* wever, the ultimate containment pressure has been satinated to be significantly higher. For example, analyses have been performed to assess the ultimate Mark II containment failure
. pressures as follows: 133 psig by Ames Laboratory for WNP-2, 140 psig by GE for Limerick, 150 psig by Sargent and (Architect Engineer for La Salle) for a standard Mark II,'pundy and 135 psig by Stone and Webster for Shoreham ' a
. The most-limiting containment pressure response to a set of design basis accidents (DBAs) is included in Table 2.3. Comparison of the-DBA results with the containment design pressure and the decign' differential drywell floor pressure reveals various safety margins for the dif ferent - plants. The calculated response is dependent upon geometric differences, core thermal power rating, and safety equipment differences. However, .it is interesting to ncts that the larger number of downconers at Nine Mile Point 2 reduces the drywell floor loading relative to the other plants.
An important system for removing energy from the containment during an accident is the residual heat removal (RHR) system. The RHR system is considered as part of the ECCS and can be operated in several different modes, such as LPCI (discussed in Section - 2.1). Another mode of operation is the suppression pool cooling mode, in which the RHP syr:em takes suction from the suppression pool, cools the water in one of two redundant heat exchangers, and returns the cooler water back to the pool. The nominal ratings of the heat exchangers are previded in Table 2.3. In general, the referenced conditions were for a peak pool temperature near 212*F. As seen in Table 2.3, the RHR heat exchangers are capable of removing 1.7 to 2.7% of rated core thermal power. The RHR system can also be aligned to take suction from the suppression pool, cool the water in the heat exchangers, and discharge the water through the containment spray system to both the drywell and the wetwell. In most accidents, the peak wetwell 7 l
2 3 4 airspace temperature is expected to be at or below saturation conditions due to the presence of the large pool. Conversely, higher temperatures can be expected in the drywell, especially if _ molten corium is present on the floor. Most plants ' distribute approximately 95% of the spray flow to the drywell and only 5% to the wetwell. While the spray distribution was established to enhance DBA performance (although not required), it could also be beneficial during severe accidents. Drain pumps, sumps, and . downconers in the drywell floor are used to return the water to the suppression pool. Nine Mile Point 2 also has a sloped floor to direct water to the sumps. As noted previously, the BWR/5 (and Limerick) RHR pumps.are capable of pumping saturated water. If the RHR pumps are taking suction from the suppression pool, three conditions could cause the suppression pool to boil as a result- of containment depressurization- and fail BWR/4 RHR pumps: (a) containment venting", (b)" rapid depressurization caused by containment spray operation (sca Section 4.3), and (c) containment failure. Three methods are provided for controlling combustible gas concentration in the primary containment following an accident. First, the primary containment is inerted with nitrogen to maintain the maximum oxygen concentration below 54. (However, the plant Technical Specifications permit de-inerting of the containment for short periods prior to shutdown and during start-up). Second, the 3 long-term radiolytic hydrogen generation following a DBA is maintained below 4 to 5% by use of hydrogen recombiners. -The hydrogen recombiners take suction from the containment, combine the oxygen and hydrogen, and return the affluent back to the containment. As seen in Table 2.3, the low capacity of the hydrogen recombiners is cc.asistent with a slow radiolytic hydrogen generation rate. These systems would be of little use during a severe accident and might actually compound problems by exceeding the -hydrogen recombiner inlet composition design specifications (potential for release path). The recombiners were intended to be manually initiated several hours after a DBA was terminated. Finally, the nitrogen purge system and the standby gas treatment system (SGTS) in the secondary containment dilute, filter, and vent . the gases in the primary containment. Hydrogen control via venting through the nitrogen purge system is only used if the other control measures are unsuccessful. 2.3 Secondary containment Design The primary containment; is surrounded by a secondary containment, which is made up of the reactor building and the refueling bay. The design of the secondary containment is site-specific, with the design being determined by the architect / engineer. However, secondary containments for Mark II plants are generally similar to those for Mark I plants. General 8
8 characteristics of secondary containments are discussed below for completeness. l The secondary containment, in conjunction with the SGTS, serves as the final barrier to prevent or mitigate the release of fission products. As seen in Table 2.4, the Mark II secondary containments are large structures. At multi-unit sites, a single secondary containment is divided into separate zones, which may be - 1' isolaced from one another. The operating pressure of the secondary containments is
-typically -0.25 in. of water gauge to prevent leakage to the environment. The secondary containment atmosphere recirculation and filter systems are redundant and are sized to maintain the system operating pressure with 100% leakage per day. The buildings are generally protected against failure with blowout panels near the top of the refueling bay. However, at Susquehanna, the blowout panels are icicated in the lower and middle levels of the reactor building (Shoreham has no blowout panels). If the building pressure were to exceed the design pressure, the blowout panels would open to prevent gross structural damage. The ultimate failure pressure of the secondary containments is plant-specific.
However, the structures were designed to withstand the maximum anticipated external wind loadings, and not large internal pressure loads. For example, the above ground secondary containment design pressure rating at the Browns Ferry BWR Mark I is only 3 psid. Therefore, the secondary containment might fail as a result of hydrogwt deflagrations (subsonic combustion) in a severe accident.21 , Fission product control in the sincondary containment is achieved using different systems at the various Mark II sites. In the event of an accident, the Standby Gas Treatment System (SGTS) (the Reactor Building Standby Ventilation System (RBSVS) at Shoreham) operates to filter all releases Eo the environment and to maintain a negative pressure inside the secondary containment. The filtered effluent is discharged to the plant stack or to an elevated release point. The SGTS, or the filtered exhaust section of the RBSVS, utilizes two filter trains to remove the non-noble , gas fission products. At Limerick, Susquehanna, and shoreham, high capacity fans recirculate gases within the secondary containment. At Limerick, the Reactor Enclosuraie Re. circulation System (RERS) mixes, filters, and recirculates the affluent to the reactor building and the refueling bay. Conversely, the RBSVS at shoreham simply takes a suction from the reactor building and discharges the affluent to the refueling bay without a significant amount of filtering."# oak Ridge National Laboratory (ORNL) has analyzed the response of BWR systems (i.e., reactor vessel, primary containment, and secondary containment) to postulated severe accidents. Preliminary review by ORNL of the secondary containment performance has led to 9
. ~ $,
e several insights." First, as mentioned above, the designs of BWR secondary containments are. highly plant-specific. Plant-specific parameters affecting system performance include (a) compartmentalization of the secondary containment, (b) the filtering and exhaust capacities of the atmospheric fission product control systems, (c) mixing and filtering by ventilation systems, and (d) the area coverage of the fire protection spray system in the reactor building. Other transient-specific insights affecting - fission product retention in the secondary containment include (a) the primary containment failure location, (b) the availability of AC power to operate the atmospheric fission product control systems, and (c) the likelihood of hydrogen deflagrations capable of damaging the structure. Specific details about secondary l containment performance during accident situations are provided in Sections 3 and 4. 6* .,. i i 6 e de 10
.9
. Table 2.1-United States Listing of Nuclear Power Plants with Mark II containments.*
' Utility / Plant- Architect Encineer Constructor 29Egercial Date Commonwealth - -
Edison La Salle'1. Sargent & Lundy- Utility 10/82 La Salle.2 Sargent & Lundy Utility 6/84 Long Island
- Lighting Company Shoreham Stone & Webster Utility Note b Niagara Mohawk **
Power Corp. Nine Mile Point 2 Stone & Webster Stone & Webster 4/88 Pennsylvania Power & Light Ci. Susquehanna 1 Bechtel Bechtel 6/83 Susquehanna 2 Bechtel Bechtel' 2/85 Philadelph-la Electric Company - Limerick 1 Bechtel Bechtel 2/86 Limerick 2 Bechtel Bechtel Note c Wnshington Public Power Supply System WNP-2 Burns & Roe Bechtel 12/84 Notes:
- a. "World List of Nuclear Power Plants", Nuclear News. -
Vol. 32. No. 2, American Nuclear Society, February, l 1988.
- b. Shoreham received a full power operating license on 4/20/89. However, the future of the plant remains uncertain because of the pending agreement between l
Lilco and N. Y. State, which would result in the i closure of Shoreham.
- c. Limerick 2 is 96% complete and is expected to go into operation in the sewnd half of 1989. J j
11 l l-
g -g Table 2.2 Comparison of BWR Mark II Reactor Design characteristics , Plant !. Limerick Susquehanna La Salle Nine Mile- . L Parameter 1,2 1.2 1,2 Point 2 Shoreham WNP'2 0 BWR Type 4 4 5 5 4 5 O VCssel ID 251 251 251 251 218 251 [in) - o Number of 764 764 764 764 560 764 Fuel Bundles O R2ted 3293- 3293 3293 3300 2436 3293 Power [MWe l O Power 48.71 48.71 50.00 50.00 49.16 49.16 Density [kW/L) o Avg Heat 5.'34 5.34 5.40 5.40 5.41 5.40 Generation [kW/ft) o MSIV 26 30 25 25 25 25 Bypass [4] O HPCS (HPCI) (HPCI) (HPCI)
-GPM 6250 5000 6250 6250 4250 625
-NPSM [ft) .
21 12 12 .18 -
-Design Turbine Turbine AC motor AC motor Turbfett AC Motor
-Injection- Feedvs.ter Feedwater Above core Above core Feedwater Above cor Location sparger- sparger .. sparger
-O LPCS (CS) (CS) (CS)
-GPM 6250
- 2 6250
- 2 6250 .
. 6250 4625
- 2 6250
-NPSH [ft) - 9 1 14 13 -
l
-Design AC Ector AC motor AC motor . AC motor AC motor AC motor
-Injection Above core Above core Above core Above core Above core Above cor Location o LPCI
-GPM 10000
- 4 7500
- 4 7067 *3 7400 *3 9650
- 2 7067
- 3 -
-NPSM [ft) 5 9 6 14 14 -
-Design AC motor AC motor AC motor AC motor AC motor AC motor
-Injection Core Recirc 3 places Core Recirc Core Location shroud lines shroud lines shroud O RCIC
-GPM 600 600 600 600 400 600
-Design Turbine Turbine Turbine Turbine Turbine Turbine
-Injection Feedwater Feedwater Feedwater Feedwater Feedwatar Feedwate Location 12
~
Table 2.3 Comparison of BWR Mark II gimarv Containment Desicn Characteristics Parameter Limerick Susquehanna La salle- Nine Mile Shoreham WNP-2
. O Totpl Vol 536,759 520,294 526,880 650,300 406,812 457,727
[ftJ-o Wetwell 289,100 281,500 297,000 346,800 215,400 256,400 - [ft*)
~ containment 163 158 160 197 k" 167 139' Vol/ Power rgting
-[ft /MWt]
o tetwell 87 79 85.48 90.19 105.09 88.42 77.86
.Vol/ Power **
ng ratp/MWt] [ft o In-ped. to Same 1 in. below Below Below Same Below DW Floor o Dohnconers 257 242 295 363 242 309 2 Area [ft)
-f Ex-ped' 87 82 82 121 88 99
-Height 18 18 18 3-6 6 ?
above fir- [in)
-f In-ped 0 0 0 8 4 0
-Height n/a , n/a n/a ? 0.5 n/a above f1r
[in) o Design ^
~ Press
[Psig]-
-Internal 55 53 45 45 48 45
-External ~5 -5 ~5 -5 -5 -2 .
O DW Floor Design dP [psid)
-Downward 30 28 ?S 25 30 25
-Upward 10 ? 5 l'a 10 6.4 o Max Leak 0.5 0.5 0.5 1.1 0.5 0.5
[4Vol/ Day) 13
p .
-1
[. . Table';f.3 (Continued) Limerick Susquehanna La Salle - Nine Mile Parame,tgI 1.2 1.2 1.2 Point 2 Shoreham WNP 2 O RHR HX's . ._
-Removal 122
- 2' 134 *2 156
- 2 95
- 2 89
- 2 122
- 2
' Rate
[MBTU/hr) .-
-t of Core 2.2 2.4 2.7 1.7 2.1 2.2 Thermal
, Power DBA Peak
Response
o DW [ prig) 46. 44. 40. 40. 46. 34.7 o WW [psig) 34. 29. 31. 34. 34. 27.6 O DW Floor 23..
- 22. 24. 17. 23. 19.
Ioad-psid Combustible cas centrol oC:ntainment 4% 4% 4% 4% 4% 4% Oz (%) oH, Comt iner 132.
- 2 ?.
- 2 135.
- 2 150. *2 60. *2 66
- 2
-flow [vcfm] (Both units)
Desien' Temperature-G DW [*F) - 340 340- 340 .340 340 340 o WW [*F) 220 220 275 212 225 275 o The power rating represents the peak pool temperature response during a DBA LOCA e 14
4 8
; Table 2.4 Comparison of BWR Mark II ggggpdary Containment Desian Characteristics Plant Susquehanns Nine Mile Parameter Limerick 1.2 1.2 la Salle 1.2 Point 2 Shoreham m
- CCcondary 1.8 - RB #11.5 - RB #1 4.5 4.6 1.4 - RB 3.5 Oldy Volume 1.8 - RB.#21.6 - RB #2(Both Units) 0.7 - RF Bay 3
.[10 ft ) 2.2 - RF BaF.7 - RF Bay De31gn 0.25 0.25 0.25 0.25 0.25 0.25 Prcssure
[p3ia)
*~
Op. Pressure -0.25 -0.25 -0.25 -0.25 -0.25. -0.25 [in. of water) Lenk Rate 100 100 100 100 100 100 [tvol/ day) Ficsion RERS + SGTS SGTS SGTS RBSVS SGTS Product. SGTS Ccntrol - Cyctums , acity (.5-8.4) (3 - 10.5) '4 *2 4*2 1.16
- 2 4.5 *-2 o[kft C a p/ m i n ]
- ltSGTS)
- 2(SGTS) (Exhaust) 60
- 2 -
6 *2 45
- 2 (RERS) (SEOASS) (Recirculate) 15
___________1____=______________ . . . . . . .. i
>4
~*
[_
'g ( l 1
ee '
^
1
\ ,
.. I, I ,' a s '
\ l
__ \ ' ;- U* bas { _ - - - - - - - - . , , . . _ ., _ _ _ _ j
* * . . asino'onom as eens "
N-
%. ,;)
7.:
', y
> :sse I -t' I j a e-en
* (
. .g ,,
I-- ~ . ... l. , , . . A's i "r!! M L J. l ,- l [. ~s,d *='s. l h /
..l 4/ / ,a j.,- . . ,. "-
.. mt- ~
.l p + t2 , , . ' . . . . , , .
eage t tis * * .* .. f es an ,;
, e n
9 * - g won swe 8 's =,s s wavvem ss.na w=e
.asa6 s w w ,,,,,,",
* *** est assu I .. . ,,
T,,
. ]t . .
t_ =' . . ' er ..e :
- o...,. .
% 3em. - h. .'
~
l 1 W.j .g
.t I ,,,,,.,
, . m .
I ** U-
' * ' }.
,,4 w \ . %
'"*".'n""s ... gm
,..- ====.=s.6 .
2-
'ema
.. y ,,,Tf,,, g .
... u w e., *
.
- m e sure . i*t
-i, m es --
j ; _ = --
.: l .'as a.sw. arw' ear :e coewe ,., v s . .
'n P
.. g .suser* =% ..
j . _ l , d==. =.*
- m e
;,.'.r -
w w m . .3.o. i b, ,fD l .,i., . .,, actru t
- s. .I, ,,9 g
t ** 5 etese
- *
- 81 '3E **e' I'8 4"* *4
.e
.05e *- i . ~ et 4 l
l ; 4 . est eats .
, i co w -.
.: I i s uve - .f j [5 : : hmr
') '-
N.~Tiilf s 1
, .. i um.
me e, .-
- - - - %-. ~_
,7 ..
e,=ise. ,
. :' !: sw,r g . .
. i.
-4 J
w ;
.t ,
a r
, .. t' .. .
'..I '.he
.:: . . . .:: ..*. p. ,: - .
Figure 2.1 Shoreham Containment 16 L _ - - __ _-__ -.-__ _______-_ ___._____-_ .
- 1 4
m. REACTCR (, REACTCR MES$URE g-2, p si7;-g, , vt5sCL (R.P.v.) EL. 325'-40,, - - ma s a a) ' , i l
, EL. 3i.-.v,- R,* mReu .
STAR 8, TCP or 55W g EL.SS 2 - = q
- 1. ,w e( - N= eu.3ee.e-l 1 I Top or esw" l l l if-a'4" - 'E2 W r
a EL. s4'-iVZ _ l ev .4- , _1 s I g a w , l
, FRhlAR' l l hhhg CONTAINMENT REACTOR a l l l SUP*CRT N ,n g / ,
$xlRT Wg /l a EL. 244'-4Yt" 1 9 # 1
- f_
l - en 1 e- - 7'. 3 /2' i
-e
- 5' t'/ 2" ORYWELL FLCOR 1 # - -BEND Lig "
EL. 232'-0" EL. 240 -7
/
[ RCACTOR PEDESTAL i i SI'-0" - 1 . 4.o" e e 2d-3"-e =e - 4' C" { EL.175'-C', 4 , P MAT Figure 2.2 Nine Mile Point 2 Containment 17
t e
! "'"F. . . ..e i
, Amma, g 8 l -
i l1 - - '
- i. m _ _.
l
~
. - - . fl .
% [ . ._
h 1 - - - , 8 I i gn aiwigge
'.yJ b.4 ( GAArst?
,, O ;,'g -
. L j, i i
*
- 8erew43 3.g.g -
yf . ,
/ ._ I w-' " J .. e. e s wi & t.
A
" " ,) J
.t W s- ., . . . *0t *h*A .A t t I
,[ ,y. StFWC* WEE
)j e , a, ---
- .? -
.. , r. m *'.an a_
. =.ac;~P
~~ 'I +1
; y .s I . ., ,
m eg,ar,,?pt / ' weuo i
"' , N. ' ~ War.-
s. a
. , l s . ..._ . , ; , . .. g-.: .s.4.t . .- .
w
..,_,[ .
. .. ..., u ., g on- i
--l g :w *%
- i
- - --. h l rEAc*ot
ag.De tTA
.' M 'It 0.i. c &
c:::::::::::r ; p , -- , i --
.- - - m
. .. .L.:aj w A ma.
. ) <- '
1:"'.' T . ~
-L _ _. . ,i wF.4 .l .
,. 4 - ('
.ggymp, ) ,
* *: q- g ** - +
.'.*.:a. - -
t. Q p'./ J - - - -;%1.:u - '
' , N I . , 3 ., ,,_ .
,EHs.* 4$ e*y mw- g.
~
1 U-y
; I M-
%s g
ra. . > I s
'l an" t -a m m -
- 'ejwye -
rd v f- v .-
- - l,
~ *' ,ls - r .
*m-
.y ,
- w. y Is v// I y
I i -+' 'a . l , g ...,' g '. I1 31 2 l" l ap 'l
. i 6
lg l
= * = .
Figure 2.3 La Salle Containment 18 t .
s
.e 4
1 I 111 ' "
)
Fuel stomge Refueling platform pool ( Steam cryer . If
} anc separator f; -
6 storage pool i < l A f P Reactor 1'j( <. ,
- pressure ,, -
g Reinforcec vessel (-: l l'l;' s concrete
- l
\: support
\.
Core _ \ -Free-stancing steel crywall _ ASME
,.. vessel
\
.;, g[*' "
> j g .. ,
}l k.
m - u - {;-l- l -
- g. 49 u i 4 W -
?? t 4;.
} .y.
mm
= l!
f , Suppression { nu i i um m ; - - i i mi . Cnamcer j
.w y
- rs
. O ,. - - -
$ 4 . 1a i h W -.;w N';%m. .,e.1:2&
4dbhh . I g 1
%Qu \
i r
- u - ~ y ; g q g 4 g ; m .y g ;; , _ , l z_ _ .
__. , .e.-@kDEEZ5FTM?!%W" , )i 1ttse 1 1 I 1 Figure 2.4 WNP-2 Containment i 19 _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - l
s 4 1 .. y .
% l' n
e. M Y ll _ N
~
.ans N N
Vy;RI__
- ! /
'me Pnmery
" "' N Ngi g
'l p fl ,
L
?" ~*' ~ .
~
g 6.s : . i Y f .q
- i[' II" I li.Im l -"I l Dimeheogm % l-g
~--
. ,, l q Il .l1 4 ,
d *
.- .e ] ,
g-- . N [ . 00.0L f ; I . L l l Figure 2.5 Limerick Containment 20
i-I staavastr
$NIELItus A_ 7,., g Pluts - '
k .c
?b 6 G ^,,1% ? '
O atBOYA8LE OttttLL p.tA0 j
. g* (,. p,,,_=
= =_, , g -
; } 1
[ 7
.. ,1 , { .. %..
]
- ft38tt E h N gtes gigtg
.- o j#.=u E b l e
* . p j > ,
..p.~ m w m &.a EttlPg(NT 4* h. ( ,
' L uP f L 8 TT g
$ s. ,
.,4 g
A REACitt - a- , m . m . 4 ,9 o/ d 'c Ptossi n 7.j .b.
- PronT esta = =I'"N$#2A*J.4@I[^aY,o( %.< r; h . -
.a U 'INI SUPPRESSION mt s h g-s.
r. r.C ?. D.
- w. -%j.4
.b e
-,<rw., q - - ,
e .. U l. .'! Jy"
#g 4 #g: !? E
~ UN.M. war.XtdM w S &.1.m d Figure 2.6 Susquehanna Containment 21 4
_ _ _ . _ . _ _ _ ____________._m__ __.______.____.__._____m__ _ . _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
s
.'Q 1 e 1g ,
l ie j aut*es l I-susu j j' .l
- i i Al h
t
. i = , - I
.* . . era:*: l tl ". /.Yf 5 jt % *.? 'y,,,,, ve:ar
~
l l
, );... yy 1
a . s. * .:r' . .:.s . * . .}
. l
.. c 19 / ~.*J
- c. c..r( .
*, j,; ;
t
.-a '.C,'. ;
, .e. , . , G,
___ ; ,. # :vt**ssata
- e 1 %
- e . ,.
.../.e._. i._ .. , . i
')'.,.7 l a,.at:t ' # .. ,,
u e *v e us i
. r.. .. s e~
I f,..i .i ,
-{$.'e'. \ .3,.,,.,
l ff,4 - ',g 9t ,.7
~w i.. ..
ec.: s .. it i.s i. .:t
- 3. /.c. - . .:: c.:
g, **,$l;
**4 to g ; - . q /*.,1""""1 *
\::t a wrs we ?% 6 d eM E ?. '. EE, U**
y y . --- :*v j as7&M.g M.(.# '""""'""'a.n.,5* CL. .,.
- d. , L,,jl,qin n.n u. ...i. m. . n,.
.,,.... .. .. .o,.
m _a...- ,
.e ;
,,c. u. .
MMM9WhW9M/#//iNV///6, \ N
,I l
.q e
*. ,N aut*:2 sps:sf I 'ay
! i d : . .*. ****
n..
.P *p .? a I\ a i,
/j .v.*
(, s.i !
*I N. -V$.
J
\' ,\ lif; h l d, muunuses *k?- y*
')
1 Figure 2.7 Shoreham In-pedestal Cavity 22 4
.e 4 -
2.2 l N "* Sk #tectCe - t1 4
~
-mann e.m
/. ,
sen sos sese6 * *
/.' , ~
es = irw . - . -s
.fl.' v " ~ . ./
../
- 1. ! -
- l. .,.
* ' '~
i ,
- ~
. o 3 a.'.l
- o'". ':o
. c.......
~
ret
-ir f., <j'c,,';
,e 'o, : : ; ' 3 :. j :. .. .' -
- m sss essee
;* ' ; O .': -, e.. . ':
hI
, 0'? .i'. c'Ci' 4 ,
L l e'. .,,,;,< -
*~ '. .m O . .O. .. . . .
1 Cg?loO o o
.. f '.,O.*.a*;.R.
g..;O. . * *O
. $ .ollm -
.p
!,lcd,,.Y.u.i,'..<,. 'e. ' 0 sw up'
. .................4 1..5.... ....(. . .9. 9.."J '
we a.
,eo,
,,.C,...u '
s.f ./...d..!. o.
. . . a . s m a., , ,.e . .., .
Ov ,. . ar
~
"?B',Ws,'.0\ 'Sh 3.*O? 'c,! P' '. ' i .
',C 's M.'s .W, ~ ~ ,4, o.-'- .'::f
. , .9 #-:::: 0
_. . 3
.'J'.
.*O **, , O.*,: ,;,. *
. 'lO ,.. .
** . . p ,:,a n, ., o
, 3.o'
,.....s.o. .,8.. .....
,. . ** .s
'. o ';. e tb 's .'
O ,*'*' , o.' Ol'l Oj',I : ' '; O ,'. O .;.9 ' : #* n,97,******
. -' . ' ,g *... n anac m esmits .
." o,l<l* ew *l-o ol.;o C',.'.o .
.. s .. .
ol,l.,.0 Ol,.slC O'.0 4
- O'.O . e...
N' /e
= 'l-ol*l o s,'.*
- O ;c. *. .*.--
*..' Y.
s .
. . !I u , ce,,
.e. ...
i a'en Figure 2.8 Nine Mile Point 2 Drywell Floor 23
t se
*t- _ * .
. ;, e..
4, " S & L
- C.
AW
- i 1, t & el . , T.
,l *=.~1"".
, .. ~ g T .,,
I ,. i.r %S . M. .g .J.
. _g ; j. -
.a --
f ., g ... ~ ..
! s. ee a / I ej
,r ..
- M M u!$
I h: s L'"'2" FF* r t ~T.*-*sw a
~
.. +
r--- i es--- y _ h ! :!' 4*
.R ... - r
. r- .9
- g 1l .
*..r- & % ..
- .- [ p m, m ,. tag _ . *'..4.s:.~,.;..
% ' % ' * /_t E =% d '_ s .
3" .w.
. ,, ~.s -
m 7- y s .. ~.s . .
- s. *u a _ w __
m. y . %v. . . . . .:1R sw
- L..-_ .. ns s' .
. . 19:t d.tv *m ;, ..
d
, . . o. p'"* ..
" * * *. . . .t.ll .r=.*.,( ". . . ,', j ,
r . m.-
===
me g j i 4, ,
.===. . . , b l-y r
w
- n. :.
...n*'
~n .*
. . P
* ~'
l t::':LI.
, , '. 1*. ........ ,
f ,f 5 s v_-ns,c 1 i_. .- : s . A
'M.s.
_; .s '
-,. d *'
r ..
- qJ.2,,
, . ,i t:-
.' I/$ ,
- M. .,. k .df.f.. a '
' _ - , t:- me- ' --*
w . .i ._-
; ,,,,,,,j ii* *e. ,
====: -
l y '
)
,. ~. .
, r.
4
, j-4...... .
_. w
, . : * \
es , . J l s . - _ - .' s' i- --- : I
. 9. a. i' s . .
;II i
mg, i .
...E..'?
.+ .. .
- 3. . J. ,_
4 L
, . L., . ..
l
. i Figure 2.9 La Salle Pedestal Support 24 l
l 4
- s. ;
j o' l g.3. AttetF VALVE (M.S. A.T.) RYwtLL WElst pept 863C86&A44 PtPE (TYP) , (Typt U 0 1 O 0 00,0 0 0 O 0 O .O O *O 0 0 0 0 00 . O i O*
- O O OO
. O .O.O '%OOO . . O {g OO O ,
O O OO g i 2702 90' 00 0 0 .in.9..L OO h. l
'c""
.O'O *.o OO O O 00 -
O.O O O O O
. O O. co . oO O
,x O
0 0 0'0 0 0 O y.
. .. 0 0 O O O , g;, .
n , _ ., a . 180* Figure 2.10 Susquehanna Drywell Floor 25
... ]
e j
- 3. CONTAINMENT CHALLENGED PRIOR 'r0 CORE MELT The severe accident sequences leading to containment integrity i challenges prior to core melt are discussed in this section. The primary sequences with this characteristic are anticipated transients without- scram (ATWS) and transients with . a loss of ;
long-term containment heat removal (TW). For the first of these sequences, scram is not successful and the vessel steaming rate to the containment ' exceeds the capacity of the containment energy rencval systems. High pressure conditions in the containment can lead to containment failure and loss of vessel injection. During a , TW transient, the reactor scram is successful, but without containment heat removal, the containment slowly pressurizes. If the transient is not mitigated, the containment either fails from over-pressurization or the high pressure conditions lead, directly or indirectly, to a loss of vessel injection. Section 3.1 discusses plant system response during transients that lead to containment 'o overpressure challenges prior to core melt. A discussion of the containment response to the pressure loading is given in Section 3.2. A qualitative discussion of systems or actions that could potentially mitigate the consequences from these sequences is presented in Section 3.3. 3.1 Definition of Sequences Anticipated Transients Without Scram ATWS is by definition a transient where all 180+ control rods fail tb insert as the result of the failure of both divisions of the reactor protection system and the failure of both alternate rod insertion systems (one system for each division of control rods) and the failure of both standby liquid control systems. The transient is assumed to eliminate the feedwater system and to isolate containment, with the control rods still at their full power rod positions. ATWS sequences can, for convenience, be divided into two general classes, depending upon whether core melt occurs before - containment integr4ty is challenged, or whether containment failure induces core melt. In the first case, the ATWS initiating event leads to a loss of all high pressure injection and a failure to depressurize the vessel. Core melt proceeds with the reactor at high pressure and the containment intact. This sequence is more appropriately grouped with sequences in Section 4 and will not be discussed further here. In the second case, the ATWS initiating event results in a high pressure condition in the containment, which leads to failure of injection systems or to containment failure and induced core melt. Initially, injection is provided by the high pressure systems (HPCI/HPCS, RCIC, and CRD) and energy is transferred to the suppression pool by the SRVs. Due to the high energy addition rate 26
i to the suppression pool- during an ATWS, the pool temperature increases rapidly. The EPGs instruct the operator to depressurize the reactor to prevent exceeding the suppression pool Heat Capt. city. Temperature Limit (HCTL)." However, there may be some operator hesitation to depressurize the reactor because of concerns about flow / power instabilities at low pressure. Initially, the NPCI/HPCS and RCIC systems take suction from the CST. Upon receipt.of a high suppression pool or a low CST level signal, the HPCI/HPCS .(and RCIC -
.at some plants) suction is automatically realigned to the suppression pool (likely in TW and ATWS sequences). Since some of the injection water is used to cool the lube oil in the HPCS/HPCI and RCIC pumps, with high suppression pool pool temperature, the HPCS/HPCI injection systems may fail due to inadequate lube oil cooling (for example, the Peach Botton HPCI design water temperature for lube oil cooling is 140'F) " As the containment continues to pressurize, RCIC would isolate on high exhaust back-pressure. (40.0 psig at Shoreham). The CRD system would continue to inject, and with both pumps running, would provide adequ core cooling in all sequences except the most severe ATWS.gteIf the operator successfully depressurized the reactor when the HCTL was exceeded or after HPCI/HPCS and RCIC failed, several low pressure injection systems would be available. However, as the containment cot.tinued to pressurize, the SRVs could close at high containment pressure. Closure of the SRVs would repressurize the vessel,and leave limited injection capability (for reasons stated previously).
Wrren the SRVs clost is dependent upon SRV design, reactor pressure, containment pressure, and control air (nitrogen) pressure. For example,-two-stage Target Rock SRVs use the reactor vessel-to-drywell pressure differential .and the control air-to-drywell pressure differential instandem to reposition the pilot valve and cause the main disc to open. Increasing the reactor pressure would increase the differential pressure between the. reactor and the drywell and cause the SRVs to reopen. If the reactor pressure exceeded the low pressure shutoff head of the low pressure injection systems during this pressure excursion, injection would be temporarily stopped. Reopening the SRVs would depressurize the reactor and re-establish low pressure injection.ao . However, a reactivity excursion following coolant entering the core could repressurize the reactor and again terminate low pressure injection.37 Although this response might be self-correcting and result in adequate core cooling, there are many uncertainties in the neutronic response, valve operation, and core thermal-hydraulics (potential for density wave oscillations, and system chugging) ., instabilities, In plants with three-stage Target Rock valves, the differential pressures between the reactor vessel and the drywell and the control air and the drywell work in opposition. Consequently, the SRVs may close as the containment pressure approaches the plant air pressure and remain closed until the 27
.-----------,----,-,,------,--,--___L-_------.,-,----.,--- _-,---__a_----_.--._----,n_- - - - - - - - - - - - . - - - - - - -
s a reactor pressurizes to the opening setpoint. This situation could lead to large interruptions in low pressure injection. Further plant-specific complications may include the possibility of the plant air being isolated by the containn decaying with time due to system leakage.yt isolation signal and If adequate vessel injection were provided but the transient were not mitigated, the containment would continue to pressurize to failure. Failure of - the containment or venting to prevent - containment over-pressurization could lead to failure of injection systems as the result of inadequate NPSH, followed by core degradation."8' TIL Imes of lona-term containment heat removal The sequence with successful injection, but with a loss of long-term containment heat removal, is a slowly developing accident. After a transient initiator, the reactor is successfully scrammed-and the main steam isolation valves (MSIVs) are closed. Decay heat is transferred from the reactor to the suppression pool via the SRVs. Since the RER heat exchangers are not available, the suppression pool will heat up slowly. Injection is provided by the - high pressure injection systems until the HCTL is exceeded. At this time, the . operator would be expected to depressurize the vesself and maintain vessel inventory with low pressure systems. The CRD pumps, with manual valve realignment, should be
~
Note: capable of providing adequate core coo minutesafterthereactorisshutdown.g-ng within approximately As in the ATWS, there 30 is a concern that the SRVs could eventually close on high containment pressure, causing low pressure injection to be lost. This would not affect injection by the CRD pumps. Since the containment pressurization rate is much-. slower and the reactor is shutdown, there would be no large power spikes accompanied by rapid primary system pressurization. Continued heating of the containment would occur until (a) the injection systems failed, (b) the containment failed catastrophically. (c) the containment leakage matched the energy addition rate, or (d) the recovery of heat removal . systems.2s since the TW sequence is a slowly developing accident, ample - time is available to attempt to establish some means of heat removal before containment integrity is challenged. There is little or no chance that operators would overlook the need for decay heat removal. Consequently, following a transient with successful scram (event T), loss of containment heat removal (event W) is dominated by equipment unavailability. The functionally redundant means of heat removal via containment venting (see Section 3.3), along with the time available to make repairs, are generally felt to result in TW contributing very little to the overall core melt frequency. Accident sequences involving loss of containment heat removal were not found to be a significant contributor to the total core melt frequency in the Limerick, Shoreham, and Susquehanna PRAs.ssis 28 i
.C' A survey of'a few PRAs indicated that, in those PRAs that used simplistic WASH-1400 models, and which gave little or no credit for operator actions ip raccvering from the accident, a high TW frequency (104- 10 / reactor-year) was calculated. When detailed models were used and operator actions (not containment venting) dered, the calculated TW frequency was relatively low warg (10 consj/
10 reactor-year) . As the result, the Limerick Shoreham . I, and Susquehanna PRAs identified a low frequency for TW.h 3.2 Discussion of containment challenges and failure modes The principal containment challenge from TW and ATWs sequences is containment over-pressurization leading to core melt. As stated in Section 2.2, the ultimate containment failure pressure of the Mark II containment is much greater than the design pressure. For example, the , ultimate static failure pressure for Limeri assessed to be 140 psig versus a design pressure of 55 psig.'gk Thiswas large difference provides significant time for mitigative action. In the absence of any mitigative actions, containment overpressure failure at Limerick during the TW sequence is expected to occur in approximately 30 hours versus,only 40 minutes for the case of a full isolation ATWS sequence.' Further, operator actions can be very effective in changing the base case responser (see Section 3.3).
~
l Containment leakage at high pressure conditions may also af fect-the containment response. The Containment Performance Working Group evaluated the effect. of leakage fr typical containment penetrations during severe accident loads.ga Limerick was selected as the Mark II reference plant. The results from the study are shown in Table 3.1. At 75 psig, the equipment hatch in the suppression pool unseats and causes a small increase in the wetwell leakage area. At 85 psig the drywell head unseats, and by 140 psig the amount of dryvell leakage is substantial. Leakage during steam loading has both positive and negative implications. A simple staam critical flow calculation shows that the leakage at 140 psig (equivalent to approximately 3.5% of rated thermal power) wculd easily exceed decay heat removal requirements. Consequently, catastrophic failure of the containment during a TW sequence without core melt may be unlikely, on the negative side, leakage of high temperature steam into the reactor building would likely terminate repair operations and might cause failure of injection systems. However, since most of the leakage is predicted to occur at the drywell head, the injection pumpr. would not likely be affected. Flooding the dryvell head cavity might minimize the negative effects from drywell head seal leakage (see Section 4.3.8). While it may be possible to operate containment at 140 psig without leakage, it may not be prudent to operate containment at that pressure or to rely on containment integrity during a severe accident at a pressure so close to the anticipated ultimate design pressure. 29 l
o Temperature-induced Isakage is also a concern. However, since the early containment challenge is due only to steam loading (for Section 3 sequences only), the temperature in both the dryvell and the vetvell would be close to saturation (360'T at 140 psig). Significant seal degradation due to a high temperature environment is not expected until much higher temperatures are reached (see Section 4.2.2.2). . Research by the Containment Performance Working Group supports the likelihood of large leaka rates prior to reaching the cor.tainment ultimate pressure." geThere in still a possibility that the containment leakage rates have been over estimated and thus cat.astrophic containment overpressure failure could still occur. Several probable failure locations have been . identified through finite element analyses or simplified analysis methods." The deformation predictions are generally regarded as reliable, asnuming the " containment configuration is accurately described, i.o., known geometry, materials, and loads. However, actual productions of leakage are uncertain. Consequently, tae ultimate strength results should be regarded as predictions of deformation fallure. In failure analysis calculations for WNP-2, the lower circumferential unstiffened cylinder was assessed to be the most i likely failure location." This region enco= parses the top half l of the vetvell cylinder. Since the downcomers extend below this region e-gross f ailure in this region would not necessarily lead to suppression pool bypass. If the downcomers..were intact, failure above the suppression pool water line would simulate venting, except that the release would be uncontrolled and could not be isolated. This type of failure would represent a relatively benign failure mode since the suppression pool would still acrub aerosol fission products. The results from the Limerick failure analysis show that the yield stresses at the middle section of the cylinder and the l vall-to-base . junction exceed the allowable value of 150 psig." other critical sections include the diaphragm slab-to-vall junction . and the removable drywell head-to-wall junction as indicated in Figure 3.1. Failure at the wall-to-base junction would drain the suppression pool into the reactor building. Although the suppression pool function would be lost, fission products would be retained in the containment until the pool drained. Furthermore, l the release peint would be low in the reactor building. However, l site-specific flooding problems after lower wetwell failure and l loss of a water supply could lead to more severe conse quences . Failure at either of the other two critical locations, the dryvell head-to-wall junction and the diaphragm slab-to-.vall junction, would cause ir. mediate suppression pool bypass and tend to increase I the magnitude of tha source term. 30 C____________________-_____ _ J
l' Shoreham l' PRA,"Ina the containment analysis probabilistic analysis performad was done forto the 1988the early couple containment failure modes to the sequence class, containment failure pressure, break size, and break location. - A continuous failure probability density function was used to characterize the containment failure pressure, with failure deemed most probable between 130 and 145 psig. Subsequently, sequence-specific failure mode event trees were constructed to quantify the distribution of - the different failure modes. Slow pressurization, such as in the TW sequence, would be more likely to lead to a small break above i the wetwell water line. Conversely, in an ATWs sequence with rapid i containment pressurization, both small and large breaks were deemed equally probable at three different locationr,. Small breaks led to increased leakage and prevented further pressurization, whereas large breaks depressurized the containment rapidly. In summary, sequences that challenge the containment prior to i core degradat' ion can lead to containment leakage or catastrophic l failure. Two important leakage locations for Mark II containments are the drywell head seal and the wetwell personnel hatch. For sequences with slow containment pressurization, leakage might prevent catastrophic containment failure. However, significant leakage probably would not occur until the containment pressure was significantly higher than the containment design pressure. Temperature-dependent leakage is not considered impcrtant until core degradation occurs (sequences with containment challenges prior to core degradation). In the event of a containment pressurization rate greater than the leakage rate, gross contai'nment failure could' occur. The failurer location, hole size, and containment failure pressure are both plant-specific and transient-specific. Eliminating the drywell head flange failure, l over-pressurization failures are most likely to occur in the wetvell rather than.in the drywell. 31 e e o
% /
s e
~ Table 3.1 Containment Peformance Working Group Isakage Estimates for Limerick
.i n Containment Dryvell Wetwell
' Pressure Leakage Imakage resia) (so in) (se in) l 0 0.003 0.002-55 0.003 0. 0 0?.
75 0.003 0.2" 85 1.0* 0.3 i 140 42.0 0.3 l l l Notes: l a. Wetvell personnel hatch unseats at 75 psig.
- b. DryWell hatch unsents at 85 psig.
l M Sy k l 1 l l 32 {
. a 5
i e L. 1 i 3.3 Potential Improvements ATWS or TW sequences can challenge containment integrity , before core damage begins. In both cases the danger lies in high containment pressure leading to an overpressure containment failure, which in turn fails injection systems and leads to core melt in a breached containment. Mitigation of this challenge focuses on preventing over-pressurization by removing excess heat . from the containment. Several suggestions have arisen as to how this might be accomplished, both in terms of using existing plant equipment and improvements that might be made to increase' the probability of successful mitigation. Each improvement is discussed below. 3.3.1 Containment Venting
~^
Containment venting has been touted as a means of both relieving prertsure inside the primary containment and preserving the integrity of the secondary containment,27m and could reduce the likelihood of core melt from TW sequences. Venting can delay but not prevent core degradation and containment failure for ATWS sequences. Mark II plants currently have the ability to vent via existing piping and ductwork to the reactor building HVAC; Revision 4 to the EPGs provides direction for when this should be done." However, vent paths fro some Mark IIs may be via reactor building ductwork which is not designed to withstand the internal prcssure associated with venting at the EPG Primary Containment Pressure Limit (PCPL) .3#" ForemostamongtheconcernsexpressediEtheabovereferences is the effect of venting on the suppression pool and upon the reactor building environment following failure of the HVAC ductwork or of the neoprene boot that typically connects the vent piping to the ductwork. Since venting would likely lead to saturated conditions in the suppression pool, pumps that take suction from the pool could fail due to loss of NPSH." At plants such as La Salle, where the ECCS pumps are designed to operate with the suppression pool saturated," this is less of a concern than at plants where a positive suction head is required to prevent pump cavitation damage. For these plants, operator realignment of pump suction to another water source prior to venting would eliminate this problem (assuming that another water source is available). References 21, 29, 30, and 59 discuss in detail the calculated effects of venting upon the atmosphere in the reactor building. Based upon this work, Reference 24 assigned a very low probability to survival of reactor building equipment following venting, and a high probability to bypassing the reactor building fission product control systems during core mult. As Reference 21 points i out, not all BWR secondary containments are equally vulnerable to the effects of venting. However, venting at the PCPL via existing asoft" ductwork would be expected to cause some problems at all 33
- __-._.-.._._.._____m _ _ _ _ _ _ . _ _ _ _ _ _ _ _ . . _ _ . . _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . _ _ _ _ _ _ _ . , _ _ . _ _ _ _ , _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
~ .a,a
..- I plants, and at the very least, it would make the reactor building uninhabitable for personnel who might need to enter in order to repair equipment, align alternate injection, etc. The solution to this problem is to install a " hardened" vant pipe, which would be designed to withstand the pressure associated with venting at the PCPL. The vent pipe would lead from the wetwell airspace to an elevated release point above the reactor . building, such as the plant stack. As long as the suppression pool were not bypassed by downcomer failure, drywell floor failure, or some other means, the release would be filtered by scrubbing through the suppression pool. If the suppression pool were saturated, this would result in a decontamination factor (DF) on the order of 10 for fission product particulate and 1 (i.e., no scrubbing) for noble gases. The DF for a cub-saturated suppression
. pool has been estimated to be on the ord 100 for fission product particulate and 1 for noble gases.gr of However, all of the Mark II plant's (except Susquehanna) have between two and four 4 inch drain lines' from the in-pedestal area to a holding tank suspended from the in-pedestal floor in the wetwell air space.
These lines are postulated to fail within 20 minutes for LaSalle. Philadelphia Electric Company has estimated that these drain lines for Limerick would fail within approximately six minutes after reactor vessel failure. Thus, fission products could be relersed to the environment if any unfiltered vent system were to be used. There are several factors that should be considered t operability of the hardened vent system when it is needed.g ensureFirst of all, the vent piping should be sized to accommodate the expected containment pressure rise for each sequence that the system is required to mitigate. The limiting sequence in this respect is a high power ATWs share the reactor may be producing 25-30% of rated steam flow. However, the lower pressurization rate of the TW sequence (reactor power at decay heat levels) may be accommodated at a more moderate cost. Second, the system should be able to be actuated under all conditions, including SBO. Most existing vent systems do not meet this criterion since AC control power required to open the valves is typically not supplied from an uninterruptible source. Also, interlocks would currently have to ., be bypassed with electrical jumpers sinca the isolation valves in the vent lines receive a closure signal during an accident. Therefore, even with power available, venting would require , numerous operator actions. Utilities have indicated that local manual operation of the vent valves is possible; however, unless the vent valves are opened very early in the sequence, radiation levels and the resulting steam environment in the reacto could present an unacceptable hazard to the operators.g Because building of this concern, a proposal has been made to open the vent lines immediately during a 580 or ATWs." While this could be effective in preventing containment over-pressurization, it could also result in an unnecessary release to the environment. This problem could be overcome by installing an uninterruptible power supply for the 34 1
q-e vent valves. Third, the vent valves should be capable of being reclosed to isolate containment under certain conditions, e.g., when suppression pool bypass occurs. Another possible modification is the construction of an external filtered vent system, such as the Supplemental containment System (SCS),, proposed by the Long Island Lighting Company (Lilco) for Shoreham. Briefly, the SCS would be a gravel-filled concrete , structure separate from the reactor building, but connected to the primary containment by a high capacity hardened vent line. - The system would be actuated by operator action or by rupture discs set at the vent pressure, The gravel bed would scrub particulate. The L height of the struct.ure would provide for an elevated release, as I well. Reference 24 analyzed the proposed Shoreham installation and found that reductions in both core melt frequency and risk could be achieved. The DF for using the SCS design could be on the order of 1000 for fission product particulate, as compared to the DF of 10 to 100 for"the suppression pool. However, because of the high cost associated with the SCO its installation at U.S. BWRs is not expected to be cost-beneficilsi. Such a system is currently in use at the Barseback Euclear PoWW Station in southern Sweden. A Multi venturi Scrubber System (i M S) (Asea-Aton design) is being incorporated at the Oskarshamm. Forsmark, and Ringhals reactor facilities. This design uses approximat and does not rely on any AC or DC power.pyThis 80,000 gallons design is expected of water to be less expensive than the filtra design (approximately $5M as compared to $10-$50M for the filtra). 5 summary, venting has the potential to. effectively mitigate the containment pressure rise during TW sequences, and delay core failure and containment failure in ATWS sequences. However, if the existing " soft" vent systems were to be used to vent containment as specified in EPG, Revision 4, the soft vent systems would be likely to' result in core melt (due to loss of injection) and contamination or failure of the secondary containment (direct failure of equipment or preventing equipment repair due radiation considerations). Suppression pool bypass is likely to occur shortly after vessel failure. To overcome these problems, a filtered system capable of withstanding expected venting pressures would be required. To prevent problems with inadequate suction head to ECCS pumps, the EOPs could instruct the operetor to realign the suction of ECCS pumps to another source of water, such as the CST, prior to venting. To allow operation during SBO, the vent valves would need an uninterruptible source of power. And finally, for ease of operation, interlocks could be bypassed with a keylock switch in the control room or by providing ex-panel connections for bypass jacks. I 3.3.2 containment sprays and Backup Water Supply Containment sprays could be utilized to lower containment pressure through the action of steam condensation (the use of 35 l
t a containment sprays for debris cooling will be discussed later). Revision 4 to the Primary Containment Control portion of the EpGs directs the initiation of containment' sprays when suppression chamber pressure reaches the suppression chamber Spray Initiation ' Pressure, as long as suppression pool level is not too high (see EPGs for precise definition of "not too high") and drywell temperature and pressure are within the Drywell Spray Initiation Limit (also specified in the EPGs). Note that containment sprays, . ! under the guidance given in the EPGs, would typically be initiated prior to venting, since the spray initiation pressure is less than the PCPL, at which venting is required. The normal source of water for the containment sprays is the suppression pool, with motive force supplied by the RHR pumps. If the suppression pool is saturated, two problems arise. First, the suction head to the RHR pumps may be insufficient for pump operation without cavitation and damage to the pumps. Second, saturated water would not be very effective in condensing steam inside the containment. An alternate source of water, such as the l fire pumps, could be manually connected to the RHR system. Drawbacks.to using the fire pumps include the manual connection that must be made to align the system, and the limited flow rates and lower discharge head that the fire pumps can produce in comparison with the RHR pumps. Note also that AC power or local manual operation would be required to operate valves, unless the valve operators are DC-powered, which is typically not the case. Another source of water could be from the condensate transfer system to the RHR system. However, the injection capacity of the condensate transfer pumps is limited by the size of the connection between the condensate transfer and RHR systems, which is typically only a few inches in diameter. (Revision 4 to the EPGs does not specify either of these systems as alternative sources of water for containment spray.) The above discussion brings to light several possible improvements that could be made to enhance the availability of containment sprays for mitigating over-pressurization. First of all, if the RHR pumps are capable of taking suction from a source other than the suppression pool, problems with pump cavitation . could be eliminated. This would also have the benefit of providing cooler water to the sprays, thereby allowing better steam condensation in the containment. However, the more rapid depressurization associated with the cooler sprays could increase the differential pressure loading on the drywell floor beyond the ability of the vacuum breakers to relieve this differential pressure. one possiblity is for one or more vacuum breakers to be inoperable. Some plants may already have the capability to align the RHR pumps to the CST. Others, such as shoreham and Susquehanna, have RER pumps that can only take suction from the suppression pool.'# Use of an external source of water would also add mass to the suppression pool, increasing its ability to absorb heat from the reactor. This mass addition cannot be carried out 36
m a re without consideration of the limits which exist on suppression pool level, and also upon drywell water level, in order to prevent structural failure of the wetwell walls or the drywell floor. Another improvement related to the use of fire pumps for containment sprays would be a high capacity connection from the alternate water source to the RHR system. Some plants may already , have such a connection; others may have only a small diameter . spoolpiece or a hose connection, which would. severely limit the flow rate into containment. Such an improvement could be of great benefit (relative to drywell floor load considerations, see below) . 1 l The final identified improvement would be to ensure that power is available to the valves that must be operated. This could be done by. utilizing an uninterruptible power source (a large one), or by using DC-powered motor operators for these valves. If the r~dactor vessel has been depressurized when the backup water supply becomes available, the backup water could be directed into the reactor vessel.- For the TW accident sequence, the backup water supply would only have to remove the decay heat and thus could prevent core degradation or terminate core failure. For the ATWS sequence, the reactor is still producing 25 to 30% of rated steam flow and thus the backup water supply would delay core failure but may not prevent or terminate core degradation. In both $ cases, injecting the water into the reactor will not prevent and may not delay containment failure.
. 3.3.3 IRHR Heat Exchanger Capacity ..
The rate at which heat can be removed from the suppression pool (and thus from containment) when the main condenser is not available is limited by the capacity of the RHR heat exchangers. Table 3.3 shows the design heat removal capacity for RHR heat exchangers at Mark II plants, both as a solute value and as a percentageofratedthermalpower.m.natagug. As can be seen from this table, each of the Mark II plants has the ability to remove decay energy from the containment via the suppression pool cooling mode of RHR. Bear in mind, also, that these numbers are design , values; if flow and heat transfer in the RHR heat exchangers were impeded by corrosi n or biofouling, the values could be significantly lower.g8 Biofouling of RHR heat exchangers could result in 1004 bypass and thus no cooling of the reactor or suppression pool water. Installation of larger capacity heat exchangers is possible but not cost effective, because of both space limitations and radiation concerns (the RHR heat exchangers are typically a " hot spot" and thus an AIARA maintenance concern) . However, to maximize the heat removal capacity of the installed heat exchangers biofouling and corrosion must be controlled. This requires a reliable means of controlling aquatic life, such as Asiatic clams 37
b (freshwater) and aussels (seawater), which can quickly clog the heat exchanger tube sheets. This might require modifications to increase the availability of the chlorination system; periodic cleaning; and heat balance (performance testing) and nondestructive examination of the heat exchangers. Performance trending of vital heat exchangers can be a valuable aid in early detection of a degraded heat exchanger. A recent review of operating experience feedback concerning service water system failures and degradations,-performed by the NRC Offjce for the Analysis and Evaluation of Operational Data (AEOD) , found that ... service water system failures and degradations have significant safety implications." The core melt frequency (for all commercial power reactor types) was found to be in the range of lE-3 to lE-5 per reactor-year, based on the estimates derived from the review. As a result of this review, AEOD made a number of recommendations for improving service water system reliability. These recommendations did not encompass major hardware modifications; they were generally in the same vein as those discussed above. Similarly, Generic Issue 51, " Proposed Requirements for Improved Reliability of Open-Cycle Service Water Systems," made recommendations related to proper chlorination levels, operation and maintenance of the water treatment facilities, and _ inspection and testing of heat exchangers and piping. However, these recommendations were not found to be annariegn y cost-effective. They might be cost-effective for some plants. The probability of non-recovery for an event (due to common-sode failure from biofouling) resulting in an unmitigated severe' accident has not been considered in.. current PRAs. This should be considered as part of the IPEs and, therefore, these recommended improvements will not be considered further here. Another possibility for improving containment heat removal reliability is the., installation of a dedicated heat removal system, such as the Alternate Residual Heat Renoval (ARHR) system described in Reference 32. This system would utilize a high pressure injection pump, a low pressure injection pump, a separate ARNR heat exchanger sized to remove decay heat following reactor scram, a dedicated ARHR service water system, and a dedicated power supply. . Most, if not all, of these components were assumed to be housed in Category I structures outside the reactor building. Since construction costs alone would be quite high ($90 - $500M)", addition of such a system would have to reduce risk substantially in order for it to be cost beneficial. It seems more likely that specific improvements on a smaller, more affordable scale could be found. Furthermore, Unresolved Safety Issue A-45, " Shutdown Decay Heat Removal Requirements," has concluded that an ARHR system would not be cost beneficial; therefore, it will not be considered any further. Another potential method of removing the heat to prevent containment failure from the TW sequence is to use the Reactor 38
~-
e. Water cleanup (RWCU) system. This a,ystem hc.s been shown to be capable of removing the decay heat from the reactor to prevent over-pressurization of containment for the TW event and thereby eliminate the need for venting containment. This has the added advantage of maintaining the suppression pool water below the saturaclon temperature and thereby assuring proper NPSH for the
.RHR pumps. The ability of the RWCU of preventing containment over-pressurization has been verified by PP&L for their susquehanna .
plant. The use of the suppression pool cleanup (SPCU) system has the benefit for station blackout of being capable of removing water from the suppression pool. This is important inorder to permit continous containment spray operation to cool containment, scrub fission product aerosols, and to cool the corium on the drywell floor. Without the ability to remove water from the suppression pool, the EPGs would instruct the operators to limit the use of containment sprays to prevent high water levels in the suppression pool inorder to prevent wetwell failure of containment. To have the SPCU operable under station blackout conditions would require the diesel generator power supply to have sufficient capacity to run the SPCU pump and related valves. For the TW event, power is available. In summary, although the addition of larger RHR heat exchangers or a dedicated heat removal system does have the potential for reducing the challenge to containment from TW and some ATWS sequences, such a modification does not appear to be cost-beneficial. The use' of the RWCU for decay heat removal to preserve containment integrity does appear to be cost-beneficial (probably anly costs related to changing procedures and operator training for TW sequences and upsizing of the diesel generator for the station blackout sequences).
.C 39 I
s:
' -h p ..
Table-3.3 RHR Heat Removal Capacity Rate as a Percentage of the Plant Thermal Power Ratina RHR Heat Removal ' Percentage of Plant Rate (MBTU/hr.)* Thermal Power Limerick 244 2.2 Susquehanna 268 2.4 La Salle 312 2.8 NMP-2 190 1.7 Shoreham 178 2.1
~
~ -
WNP-2 - Notes:
- a. Rated RHR' heat removal rate was obtained from the FSAR for DBA conditions. Typically, the pool temperature was near 212*F. At higher pool temperatures, the RHR heat-exchangers.would be more efficient.
- b. Total' RER heat removal capacity divided by the reactor
- thermal power rating.
e O e 40
, i. .,
'T.
6mPV g-% ]\ AMOVASLE HEAD-WALL A=:: --)(g JUNCTION . f-- ,. I i-1 , I l DRYWELL I 84- l l l l 1 *
\ %~ 4
,2 JL
- ;' \
t
APMRAGM DIAPH RAGM - SLAS- WALL SLAS 4s 33 JUNCTION c
\s CONTA!NMENT' WALL YETWELL WALL
, -.< Mip HEIGHT l
WET WALL % s L l PEDESTAL ~ BASE SLABN s N l WALL SLAB
\ <
g--/gvNCTION
, ., )
, p '-
W o l Figure 3.1 Postulated Mark II Containment Failure Locations
r d
- 4. CONTAINMENT CHALLENGED AFTER CORE MELT The sequences leading to containment challenges ' after core melt are discussed in this section. The primary sequences with this characteristic are transients with scram, station blackout sequences (SBO), and loss-of-coolant accidents (IACAs) . For the ~
first type of sequence, reactor scram is successful, but vessel injection is assumed to be lost. Containment integrity is not challenged until the time of vessel failure or later.a similarly, station blackout sequences are a special case of transients with
- scram, where there is a loss of all AC power. The final class of sequences is the IDCA, which was found to be a very minor.
contributor to core melt frequency in past PRAs (see Section 6). Containment challenges and mitigation opportunities for LOCAs are similar to transients with scram. Consequently,14CAs will not be discussed further. Note that peculiarities of design may- make certain plant's vulnerable to accident initiators that are not a generic concern to all Mark II BWRs. Examples of these are loss of a DC bus, loss of service water, loss of a level instrument reference leg, and flooding in secondary containment. Since these initiators have to be evaluated on a plant-specific basis, they also are not discussed further. The containment challenges and potential improvements are addressed in the discussion related to transients with scram. Section 4.1 discusses the plant response during transients that lead to core melt prior to containment over-pressurization. A discussion of the containment response- to the challenges presented at or near the time of vessel failure is given in Section 4.2.1. Similarly, unique long-term enallenges not discussed in Section 4.2.1 are presented in Section 4.2.2. A qualitative discussion of systems or actions that could potentially mitigate the consequences of these sequences is presented in Section 4.3. 4.1 Definition of sequences Transients With scram Transient-initiated sequences are characterized by reactor scram with a loss of inventory makeup to the reactor vessel. The sequences can be divided into two groups: those in which the reactor vessel remains at high pressure, and those in which the vessel is depressurized enough to allow low pressure injection. The core is expected to melt rapidly in these sequences, leading to vessel failure and a potential early challenge to containment integrity. The containment challenges posed by vessel failure are discussed in Section 4.2. Potential core melt prevention opportunities for this class of accidents are discussed in Section 4.3. 42 _ _ __ ___ _ 1 . . _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ . _ . _ _ _ _ _ _ _ _ _ _ _ __ _ ____
s . 4 Station Blackout Station blackout (SBO) sequences are a subset of the transients. with scram. AC powered vessel - injection systems are presumed not to have failed but are unava.ilable due to a loss of power. station blackout sequences are typically divided into two groups. .If vessel injection is lost soon after the loss of all AC . power, because of common mode failure of the emergency batteries or common mode mechanical failure of HPCI and RCIC, the sequence is referred to as a short-term SBo. If -injection - is initially available, but is lost several hours hence because of battery deplation or by some other means, core damage is delayed and the sequence is referred to as a long term Sbo. In terms of preventing " core damage, the challenge is to maintain vessel injection until AC power can be restored. Note: Based on a review of the LaSalle PRA and other existing PRAs for Mark II facilities, and thus for this analysis,, the short-term SBo sequence is assumed to involve the loss of all E power only, i.e., DC power is assumed to be available until the batteries are depleted. In this case, the loss of high pressure injection is due to a common cause mechanical failure of HPCI and RCIC. 4.2 Definition of Containment Challenges and Failure Modes Challenges to containment integrity that arise following melting of a significant portion of the core can be divided into two time periods: challenges at or near the time of vessel failure (early- challenges) , and challenges several hours after vessel failure, and generally af ter the commencement of core-concrete interactions (late challenges). Because of phenomenological uncertainties, challenges to the containment after core melt are more difficult to analyze than early" overpressure challenges occurring prior to core melt. However, some general remarks can be made, and areas of uncertainty can be identified. 4.2.1 Challenges At or Near the Time of Vessel Failure These challenges can conveniently be divided into two , categories: pressure loadings in excess of the containment ultimate pressure capacity, and rapid ex-vessel steam pressurization (including steam . explosions) . overpressure challenges could result from a number of sources, including hydrogen deflagrations (when the containment is decinerted), direct containment heating, and failure of the containment vapor suppression function. Ex-vessel steam explosions could be the result of molten fuel-coolant interactions (FCI) in either the drywell or the wetwell. Each of these categories of containment challenge is discussed in more detail below. 4.2.1.1 Overpressure Challenges 43
~
a In general, the containment pressure loading (not including that from steam explosions) at the time of vessel failure is likely to be lower than in a Mark I containment due to the larger size of the Mark desi .A review of past PRA calculations'Ag"""*gn. shows that this type of loading has not been found to be a dominant contributor to Mark II containment ' failure. This is due to several factors. Fi as past NURIG-1150 studies for Mark I plants have shown,gt 'd the melt progression in a BWR is likely to be a flow-type melt instead of the slump-type predicted by the MARCH code. Second, these same studies found that a high ADS reliability resulted . in a high likelihood that the vessel will have been depressurized by the time of failure. Both of these factors contribute to making the likelihood of direct containment heating (DCH) low. Third, since the Mark II containment is inerted with nitrogen, there is only a small probability of hydrogen deflagration at or near the time of vessel failure. (This probability is nonzero but very small because Technical Specifications do allow operation for limited periods of time with the containment deinerted.) 4.2.1.2 Rapid Steam Pressurization Fuel-coolant contact after vessel failure results in two containment challenges: steam explosions and rapid steam pressurization. Steam explosions, as the name implies, result from a rapid transfer of energy from the melt to the coolant (time scale 3n the order of microseconds). Containment failure from a steam explosion would be a result of the dynamic pressure loading or of a missile generated by the explosion. Conversely, rapid steam pressurization occur on a longer time scale (on the order of seconds). Here the containment challenge is from static over-pressurization. In-vessel steam explosions are believed to be far less likely than originally predicted in WASH-1400.E"'" However, ex-vessel explosions and steam pressurization due to ex-vessel fuel-coolant interaction remain a concern for the Mark II containment. Considerable uncertainty, as well as some controversy, surrounds this issue, particularly in regard to ' plant-specific Mark II designs that might increase the vulnerability to this challenge. The discussion below presents both sides of this controversy for each of the Mark II design variations. An ex-vessel steam explosion at low pressure consists of four phases.'" First, there is coarse mixing of the molten fuel cnd coolant, with heat transfer at the fuel-coolant interface via film boiling. Next, there must be a trigger (pressure pulse) to bring the fuel and coolant into liquid-liquid contact, resulting in a rapid increase in the rate of heat transfer from the fuel to the coolant. The third phase of the process is explosion propagation, in which the heat transfer rate increases still further as the fuel fragments and steam is generated at high pressure. The final phase 44
)
l 1 _ _ _ _ _ _ _ _ _ _ _ _ - _ _ -- _ - - - - - - - - - - 1
is expansion of the high pressure steam against its surroundings, converting thermal energy into mechanical energy. The theoretical conversion ratio is approximately 30%. Note that, without a sufficient triggering mechanism, the process cannot proceed beyond coarse fuel-coolant mixing. What constitutes a sufficient. trigger is one of the major uncertainties surrounding the issue, and this uncertainty is compounded by a lack of applicable experimental ~ data. As' discussed in Section 2, the Mark II design is not identical for all plants. In particular, the design of the area underneath the vessel (referred to as the in-pedestal region) differs between the Mark II plants. For the purposes of this report, the design of the in-pedestal region has been used to segregate the Mark II plants into three classes. First, there are those plants, represented by Limerick and Susquehanna, that have a level concrete floor underneath the vessel with no downconers connecting the in-pedestal r'egion to the suppression pool. The next group, represented by La Salle and WNP-2, has a rather large dry cavity underneath the vessel. The third group, represented by Shoreham and NMP-2, has downcomers connecting the area underneath the vessel to the suppression pool, along with'some means of directing the molten corium into these downcomers. All plants, except Susquehanna, have two or four drain lines fron the in-pedestal floor to a collection tank inside of the watwell. The failure of these' drains from the time of vessel failure has been estimated to be'within 20 minutes for LaSalle and in approximately 6 minutes for Limeridk (as per Philadelphia Electric Compnay presentation on July 27, 1989 to the NRC). These three classes are discussed below. Flat floor cavity with no in-eadestal downcomers (Limerick & Suseuehanna) For the first group, calculations performed by the Containment Loads Working Group'7 indicate that a wetwell ex-vessel steam explosion, if one occurred, would not threaten the integrity of the vessel pedestal or the outer containment wall. This is because the rate processes associated with debris-spreading and heat transfer ' between the debris and the concrete, along with the limited flow area of the ex-pedestal downcomers, limit the amount of debris available to participate in a vigorous fuel-coolant interaction. Rapid quenching of the debris in the dryvell-cavity is not expected to fail the containment since the resulting steam could be condensed in the suppression pool. The amount of water available to participate in the dryvell cavity steam spike is determined by the accident sequence (i.e., I4CA or spray operation), operation of the sumps and drains, and the height of the downconers above the dryvell floor (18 in. at both Limerick and Susquehanna) . Since 14CAs are calculated to occur with low frequency (see Section 6) and spray operation during core melting is unlikely (see Section 4.3.3), and since the cavity floor is at the same level as the ex-pedestal drywell floor, significant quantities of water are 45
i - not expected to be available for ex-vessel steam explosions in the drywell. Ex-vessel steam spikes in the wetwell of sufficient magnitude to fail containment are also not expected. Laboratory's revicw of the Limerick PRA,{nBrookhavenNational quenching of large quantities of corium in the ex-pedestal wetwell region was not containment Furthermore,
- found to result .in failure.
time-dependent corium pour rates consistent with a flow-type melt (such as those calculated by ORNL with BWRSAR)'# are expected to further reduce the likelihood that large quantities of corium would participate in ex-vessel fuel-coolant interactions. Dean cavity below the drvwell floor (WNP-2 fr La Salle) For t,he second group, which has a large dry cavity underneath the vessel, steam explosions could be a concern if a significant amount of water were to be present in the cavity at the time of vessel failure. As in the previous discussion, the amount or water participating in the drywell cavity steam spike is deteru ned by theaccidentsequence(i.e.LOCAorsprayoperation),operagionof the sumps and drains, and the volume of the cavity (-3072 f t below the level of tha ex-pedestal drywell floor at La Salle) . Since LOCAs are calculated to occur with low frequency (see Sectier 6) and spray operation during core melting is unlikely (see Section 4.3.3), significant quantities of water are not expected to be available for ex-vessel steam explosions in the drywell. At La Salfe (see Figure 2.9), two 4 inch lines drain the in-pedistal cavity. These two lines' pass through the , suppression pool air space and have been assumed to fail within 20 minutes after reactor vesselgailurebytheexpertpanelassociatedwiththeNUREG-1150 If the in-pedestal drain lines became plugged and the effort. cavity were flooded, steam explosions could be a concern. The coolant from the ex-pedestal drywell floor communicates to the in-pedestal cavity via 8 inch lines, which extend 12 inch above the drywell floor. If the ex-pedestal drain lines became plugged, essentially no containment spray water would get to the corium and the probability of cooling the corium would be very low, i.e. the core-concrete interation would continue unabated until the in- - pedestal floor (or walls) failed. Structural failure of the pedestal walls and gross vessel movement capable of shearing containment penetrations are two possible failure modes. On the other hand, there is a possibility that any of the molten fuel that fell into the water-filled cavity would be quenched, without significant fuel-coolant interaction. This is an area of uncertainty, for which experimental data are needed before conclusions can be drawn. As in the previous cavity design, the pressurization due to an in-pedestal drywell steam spike could be absorbed by the suppression pool. Steam explosions and/or steam spikes in the 46
~
1 ei suppression pool at or near the time of vessel failure are precluded by the design of this cavity. Cavities with In-madastal Downconers (shoreham & NMP *H For the two Mark II plants with in-pedestal downconers, there is considerable uncertainty surrounding the outcome of' fuel-coolant. . interactions. Shoreham has installed 3 concreto "corium zing" to direct corium into the four in-pedestal downconers. The Shoreham PRA indicated that 33-47% of the corium would enter the downconers without the corium ring, . and up to 90% with the corium ring." Although NMP-2 does not have a corium ring, its recessed cavity and
-eight in-pedestal downconers should also direct most of the corium to the suppression pool. Fuel-coolant interactions were considered in both the in-pedestal drywell cavity and in the suppression pool at Shoreham. To eliminate concerns about in-pedestal drywell steam explosions, the amount of water that could accumulate in the in-pedestal r'ingion was reduced by lowering the height of the four in-pedestal downconers to 1/2 in. above the drywell floor.
Consequently, a maximum of 600 lba of water at a 1/2 in. depth could participate in a steam explosion. . The in-pedestal downcomer height at NMP-2 is not known. The FSAR for NMP-2 states that all downconers range in height from 3 to 6 in. above the drywell floor. Since there. is no known analytical tools that can truly predict steam explosions, there is a possibility that the Shoreham analysis could be nonconservative. Wetwell steam explosions and rapid pressurization leading to containment failure were assessed in the Shoreham PRA to have a conditional, probability of 4.8E-04 (For comparison, the to 3.5E-04, depending upon the sequence. Limerick PRA and the Reactor Safety Study used values ranging from lE-02 to lE-03. These probabilities Are given that the severe accident has occurred and has proceeded to pass the corium into the suppression pool. Thus, the differences in the probabilities are not considered to be significant, given the total uncertainty in predicting steam explosions.) Four containment failure modes were identified in the Shoreham PRA: (1) gross movement of the vessel due to a drywell cavity steam explosion, (2) direct containment , failure due to a small missile, (3) failure of the outer wetwell wall during a steam explosion in the suppression pool, 'and
. (4) failure of the containment by static over-pressurization. The I original Shoreham PRA containment response calculations were performed with MARCH, non-coherent flow-type pour boundary conditions (not characteristic of MARCH but thought to be likely for BWRs). Steam explosions were assumed to occur with a probability of 0.5. Failure of the containment by static over-pressurization was assessed to be the most likely failure mode.
The qualitative reasons for this conclusion are summarized below:
- 1) There is insufficient fuel-coolant premixing.
- 2) There is an insufficient trigger for initiating an explosion. J 47
e
- 3) There is no strong couping mechanism, e.g. , slug impact, that could. transfer enough kinetic energy to generate a missile capable of penetrating the containment wall.
The thermal-to-mechanical energy conversion efficiency for the steam explos calculations was taken to be it based on experimental data (on . This value is low relative to the maximum ,
' theoretical conversion efficiency (30%) . However, more research (analyses or experimental data) is necessary to make a definitivo conclusion.
Although the steam explosion analyses performed' for the Containment Loads Working Group were based on Limerick and had many uncertainties, the authors point out that a steam explosion in an in-pedestal downconer has a much higher like pedestalthandoesanex-pedestalexplosion.fihoodoffailingthe This is because an in-pedestal e,xplosion would stress the pedestal ring in tension versus a compressive stress during an ex-pedestal explosion. (The authors also cautioned about the conservative and preliminary nature of the calculations.) Other calculations for Limerick show the resulting containment pressurization rate due to rapid debris-quenching is dependent upon the amount of pool vater participating in the quench.s containment failure could result from a steam spike with large quantities of debris and poor circulation between the in-pedestal and ex-pedestal regions _ of the suppression pool . Although the calculations are bounding in nature (large corium flow rates), they do point out the potential for a steam spike challenge for sites with in-pedestal downcomers. The Shoreham PRA apalyses considered the entire pool as participating in the quench There is a possibility that wetvell steam explosions could fail one or more of.the in-pedestal downconers. Because openings exist between the in-pedestal and ex-pedestal regions of the wetwell airspace, downcomer failure would create a path from the drywell directly to the watwell airspace, bypassing the water in the suppression pool. Therefore, any subsequent releases, ~' including those resulting from wetwell venting, would not be mitigated by water scrubbing in the pool. On the other hand, if a significant portion of the debris were successfully quenched, the accident might be successfully terminated, thereby greatly reducing the effect of suppression pool bypass. 4.2.2 Challenges After Vessel Failure (Late Challenges) Should the containment survive these early challenges to its integrity, mitigation of challenges from core-concrete interactions and high internal temperature might be necessary. Although the Mark II containment is considerably larger than the older Mark I design, late overpressure or thermal failure could still occur within a few hours after vessel failure. 48 __ 4
V 4.2.2.1 Late Overpressure Failure Numerous of Mark I and Mark II containments *###stydies "b *7 have indicated that the generation of noncondencible gases from the interaction of the molten core debris with concrete can cause the internal containment pressure to exceed , the ultimate containment pressure capacity. The generic likelihood of core-concrete interactions is uncertain because of the design of the in-pedestal region, just as in the case of ex-vesual steam explosions. However, analyses and experiments have shown that the rate of the core-concrete interaction and the gas species produced are dependent on the chemical makeup of the concrete. cavities with In-madestal Downcomers (Shoreham & NMP-2) The two plants with in-pedestal downconers would appear to be least vulneratie to core-concrete interaction since most of the debris would be directed into the suppression pool. Thi.s direction of the corium into the suppression pool is by means of the in-pedestal floor being sloped to the center and at a slightly lower elevation than the ex-pedestal floor. One plant, Shoreham, has a steel liner on the ficor. This liner is postulated to fail immediately upon-contact with the corium and thereby expose the concrete. If this were the case, then late overpressure challenges would not be a significant concern as long as there were no high pressut.e ejection of debris from the reactor vessel, so that very little debris would be entrained into the drywell outside the pedestal region. This would be the case as long as vessel failure occurred at low pressure. A high pressure failure could result in a more severe early challenge. However, since the debris would be widely dispersed during a high pressure melt ejection, there would be a greater likelihood that the debris could be cooled. The Shoreham PRA# only considered sustained core-concrete interactions to be a threat for the case of high pressure vessel failure. Recovery c" containment sprays prior to the occurrence of either containment failure or , gross leakage was assumed to quench the debris. Failure of the in-pedestal downconers due to contact with debris (or failure of the Shoreham drywell floor seal) would bypass the suppression pool and could increase the vulnerability of the containment to over-pressurization (loss of suppression pool heat sink only for the high pressure vessel failure sequence) . Unmitigated accidents were calculated to lead to over-temperature failure or overpressure failure prior to complete erosion of the concrete. Since low pressure failures were assessed to be more likely than high pressure failures (see Section 6) in the Shoreham PRA, the most common and states resulted in early quenching of the debris in the suppression pool. Late venting or recovery of injection to cool ex-vassel debris was effective in mitigating 49
- _ _ n_n _ ,__
4 . . overpressure challenges following a high pressure vessel failure.7 Thus, sustained core-concrete interactions leading to containment failure were not likely. From this perspective, the in-pedestal
- downcomers are a desirable feature.
peen. cavity below drvwell floor (WNP-2 & Ia Salle)' on the other and of the spectrum, the pedestal design with.a-large dry cavity underneath the vessel would appear to be the most vulnerable to core-concrete interactions. ' For . example,- the la Salle cavity is large enough to hold approximately two entire core volumes below the drywell. floor grade and effectively out of-reach of containment sprays. Eight-inch drain lines provide a path for spray: water to flow-from the ex-pedestal drywell floor.to the in-pedestal . cavity. The confined in-pedestal ' geometry would decrease the likelihood of successfully cooling the debris. Sustained in-pedestal core-concrete interactions' would lead to pressure loading from noncondensible gas generation and to erosion of the cavity floor. At La Salle (see Figure 2.9), failure of the-3.75 ft thick drywell floor by erosion would lead to relocation of debris to a dry in-pedestal chamber in the wetwell. A large reinforced concrete plug located beneath the cavity .would significantly delay challenges to the containment basemat. Conversely, at WNP-2, a pool of water lies below the drywell cavity. Consequently, failure of the drywell floor at WNP-2 could lead to successful quenching of the debris, but with the potential for fuel-coolant interactions. An alternative drywell floor failurs mode might include the opening of a large hole in the cavity floor due to molten debris flowing through t g cavity floor drain lines and ablating the surrounding concrete. These drain lines at La Salle were esti minutes after vessel breach. gated It by is draft reasonable NUREG-1150 to expecttothat fail 20 as long as the corium is fluid, it will tend to drain through the failed drain lines and leave only a small quantity on the cavity floor. This small quantity may be insufficient to continue core-concrete interaction. Wetwell venting to prevent containment over-pressurization ' after suppression pool bypass might actually increase the off-site consequences (an earlier unscrubbed release). Sustained core-concrete interactions at La Salle could lead to containment over-pressurization failure." In WNP-2, the magnitude and rate of the fuel-coolant interactions would determine the challenge to the containment. -Based on observations by the Containment Loads Working Group'7, factors affecting core-concrete interactions include: (a) Type of concrete - Higher drywell temperatures and pressures are encountered with limestone concrete while the deepest vertical penetration is found with basaltic concrete. The concrete composition changes the ablation temperature and the physics during core-concrete 50
-.ma.mi_._ _a_.._ u .-__ _ - _ _ _ ___-._____-__..__m__._________.-.___-.m. _ _ _ _ _ _ . _ _ _ _ . _ . _ _ - - _ _ _ . - - _ _ . _ . _ _ _ - - - _ . _ _ - - _ _ _ _ _ . _ _ _ _ _ _ _ _m.-_ _ _ _ . _.-
l.' l: .' .
'b interactions. Since core-concrete interactions-are an l endothermic process, 'a lower ablation temperature (typical of basaltic concrete) " absorbs" more energy from the melt but-leads to higher erosion rates and. greater dilution of the corium. Conversely, less energy absorbed by the concrete leads to higher radiative and convective heat transfer rates from the debris surface to the ~
I atmosphere. (b) Watersin the concrete - A higher percentage of water in the concrete leads to higher drywell temperatures and L pressures, and ' greater vertical penetration. Water released during core-concrete interactions promotes oxidation of metals in the melt (an exothermic reaction) . (c) Corium temperature - Higher initial corium temperature increases both the containment temperature and the pre'ssure loading, as well as the concrete erosion rate. (d) Steel in the corium - Reducing the steel content of the corium reduces the pressure and temperature loading on the containment, but increases the concrete penetration rate, thereby decreasing the time to structural failure of the drywell floor. . As indicated by recent MARCH calculations for La Salle, drywell floor' failure is assumed to occur when the ablation has reached one foot into the concrete floor, i.e. prior to complete erosioii of the concrete floor.as An assumption used in these calculations was that the floor would fail if the debris were to penetrate a vertical distance equal to one third of the floor's thickness, i . e. , . to a depth of approximately one foot. Floor failure and containment over-pressurization during a high pressure )j short-term station blackout (TBUX) were calculated to occur at 80 and 420 minutes after vessel failure, respectively. As stated J previously, these slump-type MARCH boundary conditions may have limitations associated with melt composition and pour rates. Furthermore, failure of the 4 in, drain lines was not considered in the analysis. . In summary, sustained core-concrete interactions are most likely for the deep cavity design. The confined geometry of this cavity affects debris coolability by limiting the heat transfer surface area. Sustained core-concrete interactions could lead to drywell floor failure and eventual containment failure. The l presence of a pool beneath the WNP-2 cavity versus no pool beneath j the La Salle cavity affects the challenges subsequent to dryvell floor failure. Uncertainties affecting the over-pressurization calculation include (a) phenomenological uncertainties related to the melt boundary conditions (flow-type versus slump-type), (b) structural integrity of the dryvell floor during core-concrete interactions, (c) fuel-coolant interactions after drywell floor J failure (WNP-2) , and (d) debris coolability in the cavity given the ' 51 l mammm-u_-__--ma-,-,-u--_w_--..- a- a - - - - - - - - -
L -. , late recovery of drywell sprays. Failure of the drain lines is anticipated to result in early suppression pool bypass for both plants. At Ia salle, the core-concrete interation would continue on the dry wetwell floor below the in-pedestal cavity while the corium would be cooled by the pool below the cavity at WNP-2. l Flat- floor cavity with no in-madestal downcomers (Limerick fr ~ Susquehanna) The plants with a level in-pedestal floor at the same elevation as the drywell floor appear to lie somewhere between the other two designs as far as vulnerability to core-concrete interactions is concerned. More of the corium would be expected to be retained in the ex-pedestal region of the drywell, with less of the corium quenched in the suppression pool than in the design utilizing in-pedestal downcomers. In addition to the factors affecting sus,tained core-concrete interactions presented in the
. deep cavity discussion, the following observations pertain specifically to the flat floor cavity design:"
('a) Corium . temperature and spread - A higher corium temperature, which causes the corium to spread further, leads to higher drywell temperatures and pressures but lower concrete penetration. (b) Failure of downcomers - Failure of the downcomers, with the failure located in the watwell airspace, would lead
- to suppression . pool bypass and earlier failure of the containment (loss of suppression pool heat sink).
Downcomer failure was not discussed in Referenca 47. Howeve,r, further analysis or experiments appear to be warranted to reduce the uncertainty associated with downcomer integrity. The integrity of the downcomers would have an impact on containmerit pressurization and on the effect!.vaness of any venting mitigation strategy. (c) Corium drained into the suppression pool - A higher percentage of corium falling into the suppression pool ' during the melt relocation reduces the amount of melt available for core-concrete interactions. 4.2.2.2 Late Thermal Failure If molten corium is present in the drywell, direct heating of the drywell atmosphere and internal surfaces could cause drywell temperature to increase well beyond the containment design temperature. The Containment Performance Working Group examined both thermal and pressure loadings on containments during severe accidents. Leakage due to pressure loads is discussed in Table 3.1. However, leakage due to drywell head seal degradation was not considered for the Mark II plant (Limerick) In the Mark I evaluation, leakage through the containment purge and vent 52
4 valves, was determined to be significant. However, for Limerick, the Containment Loads Working Group accepted the manufacturers evaluation. All the isolation valves in Limerick had metal-to-metal contact and had been used successfully in conditions ranging from temperatures in the cryogenic range up to 9 0 0* F . Conversely, four 18 in. butterfly purge valves in the Mark I reference plant did not have a metal-to-metal seal. Based upon ., saal. life curves, a temperature-dependent leakage model was developed. Temperatureinduced leakage (of ethylene propylene seals) amounted to a 14 incha equivalent area at temperatures above 500*F. More recent testing" performed by the NRC has examined seal behavior under a variety of cor,ditions. General conclusions from the study were:
- 1. The thagmal ( 3 00'F) and radiation aging (200 Mrad at 1 Mrad /hr) specified in these tests had a negligible effect on-leakage onset temperature.
- 2. Metal-to-metal contact at the sealing surfaces virtually prevented significant leakage.
- 3. Leakage onset temperature did not appear to be significantly affected by increasing the seal compression from 9% to 17%.
- 4. Iqakage onset temperatures ranged from 626*F to 669'T for the five tests of ethylene propylene rubber 0-rings with a gap
~
between the surfaces.
- 5. Leakage onset temperatures ranged from 486*F to 592'T for the five tests of silicon rubber o-rings with a gap between the sur. faces.
- 6. Post-test visual inspection indicated that all gaskets experienced severe degradation, including those that were tested without a gap between the sealing surfaces.
Reference 26 indicates that the Limerick drywell head was originally designed with a double tongue-and-groove seal. However, there were problems with the groove orientation and the design was changed to double gundrop with silicon rubbar seals. The tests in Reference 48 only tested silicon rubber seals in a double 0-ring configuration. More recent tests (in support of the SNL program in Reference 49) included silicon rubber tests with double tongue-and-groove geometry. Six tests were conducted with gaps of 0.01 in. or with metal-to-metal contact. No leakage was observed in these tests for temperatures up to 700*F. It is not known whether the results from the tongue-and-groove test with silicon rubber are applicable to the double gundrop drywell head seal design at Limerick. 53
For the.Shoreham PRA a temperature-dependent leakage model was used (see Table 4.1) ."g The drywell head is sealed by a double 0-ring. . At temperatures between 500*F and 800*F, radial shear.at the- drywell head anchorage caused slippage but no loss 'of structural integrity or increase in leakage. Seal degradation began at temperatures above 500*F; however, metal-to-metal contact between the drywell head and the flange was maintained. Above . 400*F, increased slippage would lead to leakage. An upper bound analysis of the drywell head concluded that a tension failure of the drywell head would occur at approximately 1200*F with an internal pressure of 60 psig. Temperature-dependent leakage from other isolation valves was not considered in the Shoreham PRA. Thermal leakage can be an important challenge to the integrity of Mark II containments. Several conclusions can be drawn from existing iesearch. First, a survey of existing equipment is needed to determine plant-specific vulnerabilities. Second, seals with metal-to-metal contact are less susceptible to leakage. Third, pressure loading may unseat seals and increase the likelihood'of thermal degradation. Finally, seal leakage may occur at a variety of locations, e.g., purge and vent lines and the drywell head. , However, drywell head seal leakage (without water above the drywell l head) would occur directly to the refueling floor and has the potential to greatly increase the severity of off-site consequences. . With the area above the drywell head flooded, the release would be scrubbed, similar to having been passes through the suppression pool. The elevated temperatures might also ; revaporize volatile fission products that have plated out on 1 surfaces inside the containment. - 1 4.2.3 Discussion of Containment Failure Modes k Containment failures resulting from sequences with challenges occurring after core melt can be the result of damage caused by pressure loads, shock waves, missiles, or thermal loads. As discussed in Sections 4.2.1 and 4.2.2, the likelihood and magnitude of the containment challenges affect both the size and location of the failure. Each of the postulated containment failure modes is , discussed below and is related to plant-specific effects when applicable. Pressure load failures fall into t.. classes: rapid pressurization due to steam spikes or to a rapid pressure loading at vessel failure, and slow pressurization due to sustained core-concrete interactions. As discussed in Section 4.2.1, the pressure loads at the time of vessel failure are nct likely to cause gross containment failure although a rapid steam spike might increase containment leakage area. As discussed in section 4.2.1.2, plants with in-pedestal downconers have the greatest probability for occurrence of rapid steam spikes after vessel failure. In other plant designs, rapid steam spikes might occur following drywell floor failure. Depending on the magnitude i 54 l i
[. -
^. e 3
of the pressurization, steam spikes could lead to either increased I leakage or to containment failure. Based on the discussion in Section 3.2, increased leakage could occur through the drywell head seal and the wetwell personnel hatch, while a gross failure would l be most likely to occur in the wetwell airspace or at the watwell wall-to-basemat juncture.
~
Very rapid steam spikes or steam explosions could fail the containment by a variety of means including dynamic failure of the wetwell or pedestal walls due to shock wave impact, creation of a
" missile"- that penetrates the containment wall, and rapid pressurization of the in-pedestal cavity causing gross movement of the vesse that severs lines penetrating the containment boundary."j In Reference 6, failure because of missile generation or gross vessel movement was assessed to be unlikely. Similarly, the Containment Imads Working Group concluded that the dynamic j loading due to an-ex-pedestal st l i t Limerick was not
'likely to fail the wetwell wall." pas exp os on adynamic failure of the pedestal wall was not considered likely in Reference 6 due to the lack of a mixing trigger. However, the Containment Loads Working Group '
suggested that an in-pedestal steam explosion could challenge the structural- integrity of the pedestal walls. Failure of the pedestal walls would not directly imply containment failure, but it could lead to gross movement of the vessel, which could sever lines penetrating the containment walls. The failure location depends on the location of the steam explosion, upon whether there is gross vessel movement, and upon whether a small missile is createt. Gross vessel movement would most likely lead to failure in the' drywell. Direct failure of the containment structure by shock wave impact, if it were to occur, would most likely occur in the vetwell below the water line (suppression pool shock wave energy is more effectively transported through water than through air). The final failure mode is due to excessive thermal loading. Section 4.2.2 provides a detailed discussion of this failure mode and the likely location of the failure. Thermal failures are only postulated to occur in the drywell, with the primary locations ' being the drywell head seal and the various containment penetrations, including the containment isolation valves and electrical penetration assemblies. Thermal failures have potentially severe consequences since the suppression pool would be bypassed and the release could be high in the reactor building. Howeverm there could be additional fission product retention in the reactor building (see section 4.3.6). 4.3 Potential Improvements Now that the late challenges to Mark II containment integrity have been identified, a discussion of several improvements that have the potential for mitigating these challenges is in order. The potential improvements are discussed below. 55
~s j
4.3.1 Mitigating Transients with scram Mitigation focuses primarily on termination of core degradation, on preventing or delaying containment failure, and on reducing the off-site consequences. The sequences can be divided into two subsets: those in which the vessel remains at high - pressure, and those in which it iTs depressurized to a pressure low enough to allow injection from low pressure systems. High pressure sequences will be examined first. With the renctor at high pressure, i.e. , near normal operating pressure, four systems (excluding SLC) are capable of developing enough discharge head to inject water into the vessel: feedwater, RCIC, HPCI/HPCS, and CRD. For plants that have only turbine-driven feed pumps, use of feedwater injection requires that one or more main steam isolation valves (MSIVs) be open and that the main condenser be ivailable. For plants that have a motor-driven feed pump, feedwater would be available' even with the MSIVs closed, assuming that flow was throttled to the available makeup rate.to the main condenser from the CST. For some transients, feedwater would initially be available and is the preferred source of makeup to the vessel. If ' the transient resulted in isolation of the vessel from the. main condenser, the main steam lines could be reopened under some circumstances. Otherwise, RCIC would become the preferred' makeup source. Initiation of RCIC is automatic at Level 2 (RCIC initiates slightly above Level 2 at Susquehanna) . Should<~RCIC be unavailable, HPCI/HPCS must be relied upon for initial high pressure makeup. The initiation of HPCI/HPCS is also automatic at I4 vel 2. Operation of RCIC and HPCI/HPCS is dependent upon the availability of emergency 125VDC power; HPCS (BWR/5 only) also requires AC power, either from off-site sources or from its dedicated EDG. Service water is also required for area cooling and for cooling of the EDG that is supplying the HPCS pump. Both RCIC and HPCI/HPCS initially inject water from the Condensate Storage Tank (CST). The suction of HPCI/HPCS, and also RCIC in some plants,' automatically transfers to the suppression pool when suppression pool level increases above a set level (or CST level ' decreases below a set level). This suction transfer becomes a concern in the case of SBO or ATWS because the high suppression pool temperature would interfere with lube oil cooling. Each CRD pump injects at a flow rate of 40-70 gpm during normal operation, taking suction from the CST. Following a scram, this flow rate increases to approximately 100 gym. Thus, the CRD pumps are a viable source of high pressure makeup. For cases in which the CRD flow rate is too low to maintain vessel level, the CRD pumps are still of benefit in delaying the onset of core degradation. Operation of the CRD pumps is dependent upon cooling from one of the component cooling water systems (TBCCW or RBCCW). At some plants, the component cooling water system may isolate non-safety-related loads, such as the CRD pumps, upon receipt of 56
f e an accident signal (low vessel level or high drywell pressure) . Without cooling, the CR can be expected to fail within approximately one hour.g pumps The interlock that isolates component cooling water can be bypassed with electrical jumpers; however, guidance for doing so is not typically found in the current EoPs. For plants like La Salle, where RBCCW is not designed as an Engineered Safety Teature (EST), this is not a problem since RBCCW , should remain available to cool the CRD pumps whenever the CRD pumps are available. If no high pressure injection is available for coolant makeup, the vessel must be depressurized to a31ow injection from low pressure systems. Doing this is the province of ADS, with manual depressurization by the operator as a backup should ADS fail. Since the issuance of the TMI Action Plan in NUREG-0737,st the initiation logic for the ADS has been modified at some plants to increase the 1.ikelihood that the reactor will be depressurized when depressurizat~ ion is needed. Essentially, this modification involved either the removal of the coincident high drywell pressure signal from the ADS initiation logic, or the addition of a delayed bypass of the high drywell pressure signal if the low vessel level signals are present. Under this revised ADS logic scheme, the reactor should automatically be depressurized upon receipt of a signal indicating that the reactor water level has fallen to approximately 30 to 36 inches above top of active fuel (TAT) (alcng with a confirmatory low level signal set at 172 inches above TAT), a signal that low pressure ECCS pumps are running, and time-out of the ADS timer relay (typically set at 105 seconds). (The normal water level is 200 to 210 inches above TAT 4) The operator can inhibit ADS (e.g., during an ATWS event) , through the use of an inhibit switch that was added to the system in response to NUREG-0737. The addition of these modifications is expected to significantly lower the ADS failure probability. This would in turn decrease the contribution to core melt frequency from the TQUX sequences, where core cooling is lost because of a failure to depressurize the reactor. Other enhancements to increase the operability of the SRVs ' during severe accidents have also been proposed. These include a dedicated source of DC power to the SRV solenoids, assurance that the SRVs would be capable of being opened by the operator under environmental conditions associated with severe accidents, and improved operator training and Emergency op4 rating Procedures (EOPs). Because of the possibility of concurrent failure of both the AC and DC power systems, the addition of a dedicated DC power supply for the SRV solenoids could have some potential for reducing core melt frequency. The containment vent pressure is Primary Containment Pressure Limit defined in the EPGs.', set Thisatdoes the not approach the containment pressure at which the SRVs' might be prevented from opening by a low differential pressure between the
) used to open the valves.
containment Therefore, thisandventing the instrument set pointair (N,ld shou not be a concern for the 57 ______J___________-____
~
plants that have been evaluated. However, it could become a concern for plants with a higher venting setpoint, i.e., a higher primary containment pressure limit (PCPL). Revision 4totheEPGspiscussesvariousalternativemeansof depressurizing the vessel.' For example, interlocks could be bypassed to allow the MSIVs to be opened. This would allow use of , the turbine bypass valves to reject steam to the main condenser, assuming that the main condenser were available. The use of these alternative methods is indicated if less than the minimum number of SRVs required for emergency depressurization is open, and the differential pressure between the vessel and the suppression chamber is above the minimum pressure required to open an SRV (50 psig is a typical value). Once the vs.ssel has been depressurized, a number of systems can be used for low pressure makeup. These are: condercate pumps, RHR pumps in the LPCI mode, LPCS, condensate transfer pumpe, fire pumps, and service water pumps. Each of these sources is discussed below, along with possible difficulties that might have to be overcome before the source could be utilized.
- 1. Condensate pumps: Use of the condensate pumps may be limited by two basic interrelate.d considerations. First, if the McIVs were closed, condenser vacuum would be required if makeup to the condenser were via a " vacuum drag" line from the CST. The av_,ailable flow rate from the condenst.te pumps would then be limited to this makeup rate since the condenser hotwell inventory is only sufficient for.a few. minutes of operation at full flow. Maintaining condenser vscuum could be dif ficult if auxiliary steam were not available as a motive force for the steam j'et air ejectors. Steam from the auxiliary boiler could be used, but this would of course be dependent upon the availability of the auxiliary boiler. The mechanical air removal pumps could also be used, but these pumps discharge directly to the turbine building exhaust plenum, bypassing the offgas treatment system. Plant-specific design differences in the balance-of-plant may, of course, affect the condensate '
pump availability. For example, La Salle and NMP-2 utilize pu= ped makeup to the hotwell under normal operating conditions. La Salle also has emergency makeup pumps, while the emergency makeup at NMP-2 is via " vacuum drag" from the CST.naa
- 2. RHR pumps in LPCI mode: The RER pumps get a signal to start upon receipt of either a low ves.sel level signal (30 to 36 inches above TAF) or a high drywell pressure signal I (approximately 2 psig). These signals also cause the RHR I system to realign to the LPCI mode; the LPCI injection valves do not open, however, until vessel pressure decreases below a set value. At Susquehanna, failure of this valve interlock was found to be the dominant contributor to failure of low 58 i
pressure ECCS systems. Typical LPCI flow rates are on the order of 10,000 gpm per loop. The operator cannot throttle the LPCI injection flow or realign the RHR system to any nther operating mode during the first few minutes of LPCI operation. However, LPCI flow can be terminated by stopping the RER pumps. This might be an action taken during an ATWS to prevent injection of cold water into a critical reactor. ,
- 3. Low pressure core spray pumps: The LPCS ' pumps generally receive a signal to start at the same time as the RHR pumps.
Either LPCS or LPCI is capable of mitigating a design basis j IDCA. The LPCS pumps are capable of taking suction from the CST at some plants; however, this is not true for all Mark II ' plants. This possibility has not been given credit ir. FRAs that have been reviewed.
- 4. Condensate transfer pumps: The above systems constitute what might be" called the " normal" means of low pressure injection.
Now we come to what are sometimes referred to as " alternate" means of injection. The first of these is the condensate transfer pumps. The interconnection between the condensate transfer system and the RHR and LPCS systems could allow the condensate transfer pumps to be used to inject water into the vessel via the RER or LPCS piping. Two restrictions apply, however. First, the connections are via manual valves in the reactor building; an operator would have to be dispatched to th_e reactor building to open these valves. Under some circumstances, the environment in the reactor building could prohibit doing this. Second, the lines-are rather small (on the order of 4 in. in diameter), thus limiting the injection flow rate. However, this is a source that should be considered when evaluating the overall failure probability of low pressure injection.
- 5. Fire pumps: Plants typically have both motoredriven and diesel-driven fire pumps,.which are used to supply water to the fire mains for fire protection. However, via a hose or spoolpiece connection from the fire main to the service water '
system, they could also bit used to inject water into the reactor vessel or into the containment. The above restrictions on the use of the condansate transfer pumps also apply to the fire pumps. An operator must manually connect. l the fire main to the service water system, and the flow rate I is limited by the size of the hose or spoolpiece. Note that AC power is required, even if the diesel fire pumps are used, unless the Movs connecting the service water system to the RER system can be opened manually. Manual operation of these valves would require operator entry into the reactor building.
- 6. Service water- As a last-ditch effort, plant ECPs direct the operator tc, tine up service water to inject into the vessel from the ultimate heat sink connection to the RER system.
59 w___-_____. - _ _ _ _
i These two systems are isolatea from one another by two Movs, which are operated from keylock switch e in the main control room. The valves could also be opened locally, using a manual handwheel attached to the valve operator. Typical PRAs only give credit to the first three of these systems when evaluating the availability of low pressure injection. . The reason the other systems are not included is given as lack of operator familiarity with using the systems for this purpose. This l is not felt to be a valid reason for excluding them from consideration since operators receive extensive training on potential sources of water to be used in an emergency. This includes both classroom instruction and simulator training. The use of these systems is spelled out in Revision 4 to the EPGs, further reducing the likelihood that operators would overlook them in an emergency. Inclusion of these sources would result in a reduction in ,the contribution from the TQUV sequences. 4.3.2 Hydrogen control Because the Mark II containment is inerted under most operating conditions, hydrogen combustion is considered to be a low probability event. The only significant danger is that the containment could somehow become deinerted, for example, through venting. Revision 4 to the Primary containment control EPG provides adequate safeguards to prevent this from happening. Therefgre, no enhancements to the hydrogen control mechanisms currently in place are felt to be justifiwd. However, hydrogen deflagrations in the rasctor building,.should primary containment fail, or in the vent path, are possible. 4.3.3 Containment Sprays and Backup Water Supply There are three sitigative aspects to the use of containment sprays. First of all, by evaporative and convective cooling, sprays can reduce containment pressure. Second, they can be used to cool debris outside the vessel, limiting core-concrete interaction and drywell heatup. And third, they provide some , I scrubbing of aerosol fission products in the containment ! atmosphere, reducing the consequences of a release. The sources of water that could be used for spraying the containment have already been discussed (see Section 3.3). There are also limitations on the use and effectiveness of containment sprays that have to be considered. First of all, since all pumps that can currently be used to spray the containment can also be used for vessel injection, there is a high probability of containment spray failure given that vessel injection has failed (the presence of core debris outside the vessel implies that vessel injection has failed). Therefore, a realistic discussion of containment sprays is likely to be predicated upon the assumption that the pumps are recovered late in the sequence, or that an 60
([
~)
( x 1 alternate means of spraying the containment is available, which is independent of vessel injection. 1 If the vessel were at high pressure without injection, there I is a possibility that containment sprays could be available during core melt. .The RHR system would require remote manual realignment from the LPCI mode to the containment spray mode. There are .I several implications to be considered when arking this realignment. First, containment sprays might increase the likelihood of an ex-vessel- steam explosion at the time of vessel failure. Conversely, since-the vessel was assumed to fail at high pressure, I the sprays might promote quenching of the debris during the high i pressure melt ejection and would also reduce the fission product inventory available for release to the environment. Second, upon vessel failure, if the RHR system were still aligned in the LPCI mode, water would be injected upon vessel depressurisation and could quench ,4ebris left - in the vessel and provide a source of water to the ex-vessel debris through the opening in the vessel. Steam explosions are thought to be unlikely when water is added after_ debris relocation. Consequently, LPCI flooding after the initial relocation of the core in-vessel cculd quench in-vessel debris and reduce the probability of ex-vessel steam explosions. In . addition, the EPG prohibits use of drywell sprays under certain combinations of drywell pressure and temperature in order to prevent containment failure or deinerting. Therefore, if drywalk sprays were to be used, they would have to be initiated relatively early in the sequence. This might be difficult if repairs had to be made to the RHR system
- to make the sprays available. As discussed in Section 4.3.1, other sources besides RHR can be used to spray the containmeryt. These sources include the service water system, the fire pumps, and the condensate transfer pumps, injecting via the existing RNR piping to the containment spray headers.
Operation of the containment sprays would result in over-pressurizing the wetwell vall and could result in containment failure. To continuously use contain:nent sprays, a means of . removing the water from the suppression pool sust be provided. One method is to use the suppression pool cleanup (SPCU) syster pump to pump the water to the radwaste system. To summarize, the use of containment sprays has the potential to effectively mitigate both late overpressure and thermal failure of the containment, and to arrest or slow down corium activity and to scrub fission products, thereby reducing the off-site consequences. Water must be removed from the suppression pool in order to prevent watwell failure. However, some plant modifications might be needed, depending on the plant-specific design features, the alternate water source to be used, and the method selected to remove the water from the suppression pool, in order to ensure the availability of the alternate water supply. 61
>v ,
x
' Analysis of the Mark I containment indicates that a high volume
" spray" is not necessar and pressure reduction.4toHowever, provide the significant actual containmentbenefits of anything cooling less than a - full spray pattern in the Mark II containment are
~
uncertain. 4.3.4 containment Venting _, Section 3.2.1 discussed some potential benefits and downsides of venting to -- mitigate containment pressurization during ' containment challenges that occur erior to vessel failure. Containment venting might also be of limited usefulness in preventing or delaying containment challenges that occur Ahar i vessel failure. In order to-keep containment pressure below the PCPL, the EPG instructs the operator to vent ". . . irrespective of the off-site radioactivity release." With core debris spread outside the vpssel, following the EPG in this regard could lead to an early and possibly unscrubbed (if the suppression pool were bypassed) release of fission products. If plants . have - the intention cf using their containment vents following vessel failure, procedures will-have to be developed to supplement the guidance given in the EPG, particularly in regard to when the vents must be reclosed to limit the off-site release. A hard pipe vant system with an effluent radiation monitor interlocked to close the vent isolation valves with radiation in the vent line above a predetermined. level would minimize the downsides of venting during severe _ accidents and any pressure relieved would reduce the peak containment pressure, potentially below the containment failure pressure. However, the entire vent system (actuator, seals, and control system) would have to be designed and tested to ensure the availability of the sealing function under the harsh containment environment associated with a severe accident.
. Venting to mitigate late thermal failure or to reduce the containment base pressure nrior to an anticipated containment challenge is not addressed in the EPGs. A strategy of venting pre-emptively, prior to suppression pool bypass, could lead to a slichtiv higher risk because a release of. noble gases could occur ,
during the evacuation. Failure of the downcomers in the wetwell airspace would cause suppression pool bypass and loss of fission product scrubbing, and could require that the vent valves be reclosed in order to limit the off-site consequences, since watwell venting without benefit of fission product scrubbing might increase i the risk. For example, a non-venting strategy (after vessel i failure) with successful termination of the accident would likely l result in lower consequences than a continuous venting strategy. conversely, venting might provide some benefit in mitigating a late thermal challenge to the drywell head seal. In the worst case of downconer failure and thermal-induced leakage from the drywell head seal, venting would reduce the driving force for leakage and would release fission products (potentially at a higher rate) at a lower location in the secondary containment. However, externally cooling 62
1-
*.^
the'drywell head seal (discussed-later in this section) appears to offer a greater benefit. The design of the in-pedestal region also changes the-effectiveness of the venting strategy. Suppression pool bypass might occur shortly after vessel failure for the cavity design with - in-pedestal downcomers. (In this design, most of the corium is , orpected to enter the suppression pool, thereby essentially eliminating CCI and providing scrubbing of most of the release. The total containment pressure is not anticipated to exceed the containment design pressure for this scenario.) Suppression pool bypass in the deep cavity design would likely be owlayed until core-concrete erosion failed the in-pedestal cavity (or the. cavity drain lines at La Salle). The risk of suppression pool bypass for the flat pedestal design stems from a combination of tne previous two challenges. Specifically, spreading of the debris could fail ex-pedestal downconers while in-pedestal erosion might lead to floor failure. . 4.3.5 Core Debris Control The issue of core debris control is centered around previding a coolable' debris bed following vessel failure and, at the same time, limiting -the extent of core-concrete interactions. As discussed earlier, individual variations in the in-pedestal design of the Mark II containments could lead to drastically different and stntes_of the core debris. This section considers uncertainties in the phenomena affecting the potential benefits of various-modifications to enhance the control of core debris outside the reactor vessel. Cavities with In-madestal Downconers (Sherehan fr NMP-2) The in-pedestal design at Shoreham makes use of a "corium ring" to direct the molten core debris into the suppression pool via four in-pedestal downcomers located underneath the vessel (see Figure 2.7). The Shoreham PRA found this configuration to be a very desirable feature, since virtually all debris would be ~ expected to be quenched in the suppression pool. As long as the vessel does not fail at high pressure, there is little concern about ex-vessel debris-cooling; it'is taken care of automatically by suppression pool debris-quenching. For the same reason, core-concrete interactions are not a concern. As discussed earlier, the potential that this design presents for ex-vessel steam explosions and suppression pool bypass have to be considered since inadequate experimental and calculational data exist to exclude, with certainty, these potential adverse effects. A modification to seal up these in-pedestal downconers and cover the in-pedestal floor with a layer of lead (Pb) bricks has been considered. The rationale behind this suggestion is that the corium would float on top of the molten lead since the lead would 63 __-__c_-__________-__- - _ _ _ _
.. . . 1 be the. denser of the two materials. The zolten lead would, in effect, insulate the concrete floor from attack by the corium and would result in a more generic Mark II containment design. Normal refractory materials would tend to produce non-combustibles, since they cannot withstand the postulated high temperatures of. the corium, and the dilution from the refractory materials is less than from using lead bricks. , j The . sealing of the in-pedestal downconers could have an adverse- effect on the vapor suppression capability .of the containment. The design' basis IDCA would have to be reanalyzed for each plant to ensure that the containment design pressure would not be exceeded. For example, at Shoreham, the maximum containment .
pressure during a design basis locA is 46 psig, already very near e the containment design pressure of 48 psig. In addition, the added weight of the lead bricks could be a seismic concern. Seismic analyses would have to be redone to answer this question. To date, no analysis has been performed quantifying the risks and benefits of the in-pedestal downconers, and analyzing whether these downconers really do guarantee ex-vessel debris-cooling, with negligible amounts of core-concrete interaction. Furthermore, no inexpensive material has been identified that would protect the downconers from corium attack. Therefore, this modification is not considered to be cost-effective. Dean cavity below drvve11 floor (WNP-2 & La Salle) The deep cavity design also raises a number of concerns about the fate of the molten core debris after it leaves the vessel. The large dry cavity underneath the pedestal presents the possibility that significant' amounts of debris could become trapped, out of the reach of containment sprays (based on drawings in Reference 11, the cavity is over 9 ft, deep). Should this occur, ablation of the concrete cavity floor would generate large quantities of noncondensible gases. Also, the calculated-assumed failure of the drywell floor in Reference 59 indicates that failure of the cavity floor would likely occur at about the same time that watwell ' venting would take place. If this happened, the vented release would bypass the suppression pool. Another mechanism that could lead to suppression pool bypass even sooner than failure of the cavity floor is failure of the cavity floor drain lines because of concrete attack, which has been estimated in draft NUREG-ll50 to occur 20 minutes after vessel failure." If the debris attack were extensive enough to fail the cavity floor at La Salle, the corium could directly sttack the large dry concrete plug that rests on top of the containment base mat. Beyond this point, uncertainties overwhelm the ability to make realistic predictions of the outcome. The Sandia analysis of the la Salle containment response to severe accidents, which is 64
= ..
a
- currently _ in progress, may provide resolution of some of these issues.
Flat floor cavity with no in-codestal downconers (Limerick ' & Suscuehanna) For the Limerick and Susquehanna designs, which have a flat , floor in the in-pedestal region, the issue of core dobris control is centered around preventing suppression pool- bypass due to failure of the ex-pedestal downconers by debris attack. A proposed modification - to ' enhance containment performance is to install sleeves around the most vulnerable downconer penetrations to prevent ablation. The sleaves would have a skirt to prevent radial ablation and would be anchored to the drywell floor. However, no inexpensive materials have been identified that could withstand the high temperature attack of the corium. Therefore, this proposed modification Aoes not appear to be cost-effective. 4.3.6 Enhanced Reactor Building Fission Product Attenuation As discussed in Reference 21, the secondary containment of a BWR may potentially play a significant role in mitigating severe accidents. Mcwever, since there are large differences in the design _ of individual secondary containments, the treatment of r'tigation ability and severe accident vulnerability can only be done on a plant-specific basis. Some calculational work on the secondarycontainmentresponsetoprimarycontgngentfailureand venting has been performed at ORNL and Sandia The following discussion is based on this work and is only applicable to the plants for which the analysis was done. The first observation is that aerosol fission product deposition in the reactor building can be extensive as long as
- reactor building integrity is maintained. The internal surfaces of the reactor building and the equipment located there provide a ,
large heat sink and area for Mssion product deposition. Increased deposition is desirable sinn it results in a less severe fission product release to the an'aironment due to the smaller aerosol source term. The integrity of the secondary containment could be threatened by over-pressurization as a result of a mass release rate into the secondary containment in excess of the exhaust capacity of the ventilation system and by hydrogen deflagration. I Some reactor building design features may heighten the threat to secondary containment integrity. For example, Shoreham has an I emergency reactor building ventilation system that rapidly recirculates the atmosphere in the reactor building, but filters and exhausts only a small portion of this flow to the environment. This could threaten sscondary containment integrity in three ways. 65 i
---,-_--------__------_---_-_-----.___--------_u.--------_---O--
~
[ .
.=
1 First of all, the small exhaust . rate implies that the reactor building could very easily be overpressurized by containment failure or venting (recall that the vent ductwork is predicted to
~ fail, releasing steam and noncondensible gases, including hydrogen, into the reactor building): the most likely failure location would
-be the refueling bay walls, which are ~ predicted ' to fail at approximately 0.5.pgid (Shoreham does not'have blowout panels in ..
its refueling bay) Second, given that the refueling bay walls have failed, any. fission products released into the lower elevations of the reactor building by failure of - the primary containment would be rapidly transported up to the refueling bay by the mixing-action of the ventilation system. This would tend to lessen reactor building retention of fission products, increasing the severity of the off-site release. Finally, Reference 30 predicts that global hydrogen burns would be very likely to occur in the reactor building following primary containment failure or venting. Again, this is primarily due to the mixing action of the ventilation system, which acts to limit the localized buildup of hydrogen released from the primary containment. These global hydrogen burns are predicted to result in a peak reactor ~ building temperature of 12 00'F. and a peak reactor building differential pressure of 6 psi. Note that high connectivity. between the different elevations of the ' reactor building also aids in promoting global hydrogen burns over less severe compartmentalized burns. The connectivity of the reactor building is another feature that varies widely from plant to plant. tee-use of fire protection sprays to scrub fission products in the reacto some plants.*'g building has been For example, suggested at Browns asMark Ferry (a a possibility I plant),fora large percentage of the reactor - building receives coverage from the proaction (fusible link) fire protection sprays. At other plants, reactor building fire protection is limited to deluge sprays in areas where high concentrations of alsetrical cable runs are located, with only a small percentage of the reactor building protected by general area preaction sprays. In any case, unless the reactor building fire protection sprays are supplied from a independent dedicated water source, they cannot be given much , credit for severe accident mitagation. This is due to the fact that any available fire protect 2on water would likely be used for vessel injection or for spraying the primary containment. Note that multiple-unit sites might be able to cross-connect their fire protection systems to overcome this limitation. Finally, as pointed out in Reference 21, if the release from the primary containment could be directed into the lower elevations of the reactor building, tha likelihood of reactor building fission product retention would be increased. In particular, those failures that result in a release directly to the refueling bay should be avoided since they are the ones most likely to directly bypass the reactor building. From this aspect, the critical 66 1 _ _ _ . _ _ . _ _ . _ _~ i
.J 1
primary containment failure mode becomes high temperature deterioration of the drywell head seal. 4.3.7 Enhanced Reactor Depressurization Capability pr6 posed e:hancements to-the SRVs (see Section 4.3.1) would decrease the probability of core melt at high vessel pressure and . I may reduce the probability of suppression pool bypass due to a ! stuck open vacuum breaker. Ensuring the ability to depressurize the reactor reduces the likelihood of direct containment heating, j which could potentially cause rapid containment over-pressurization at the time of vessel failure, and it eliminates dispersive exit of the core debris from the vessel, which could transport a large fraction of the core to areas of the drywell located outside the f pedestal region. This also reduces the number of vacuum breaker i actuations and thus the overall probability of a vacuum, breaker. 1 failing open... 4.3.8 External Cooling of the Drywell Head Seal A proposed improvement te prevent late thermal failure of the drywell head seal is to flood the reactor cavity (the area above the drywell head and below the missile shield plugs at the refueling ficor), allowing the head seal to be cooled directly. Several means for implementing this modification exist. For exampit, a siphon (to be installed only when needed) could be used, a hard-pipe with manual isolation valves could be installed in one of the spent fuel pool-to-reactor cavity gates, or a firt hose could be employed. If cooling to the head seal were desirad, an operator.would be dispatched to take ::he appropriate action. Care would have to be used to ensure that tae pool would not be drained below the minimum allowable water level as specified in the plant Technical Specifications. Providing a remote manual capability to flood this area provides a potential method of inadvertent draining the spent fuel pool, which is not acceptable under non-severe accident conditions. There are some potential drawbacks to this proposal. First of all, once the accident har progressed to the point where external cooling of the head seal is needed to prevent failure, radiation levels in the reactor building could prohibit opagator entry to open the valves. Therefore, during an accident like SBO, which appears to present the greatest potential for head seal failure, the valves would have to be opened early to ensure . cooling. If power is recovered prior to core melt, which is very likely, the water would have to be drained and the reactor cavity decontaminated. However, the frequency of SB0 is low enough that this concern is probably insignificant, particularly in light of the potential consequences should the initiating event leaa to core melt and vessel failure. The benefits of flooding the drywell head will be estimated in a later report. 67 _ - - - - _ _ _ _ _ _ _ _ __ _ _ - _ _ _ - _ _ _ _ _ _ = _ _ _
e . , I Table 4.1 Containment Performance Working Group Leakage Estimates for Limerick.' Drywell Drywell Temp *erature Leakage a ( F) (in )
< 500. 0.002
> 500. 0.025
> 600. 0.050
> 700. 0.075
> 800. 7.2
> 1100.3 144.0 Notes:
- 1.
References:
- 1. MAAF Analysis to Suecort Shoreham 100% Power PRA, FAI 87-80, Volume 1, 1987.
- 2. Study of the Structural Intecrity of Shoreham Primary Containment Under Accident Conditions, 25746-1520145-B4, Stone & Webster Engineering Corporation, 1988.
2h All leakage is to the reactor building refueling floor
-~
level.
- 3. The drywell head would fail un, der tension conditions with an internal pressure of 60. psig.
68 ,
~
g u.- p I
- 5. CONTAINMENT BYPASS 5 .1 - Definition of Challenges In this mode of containment failure, a release pathway is created that bypasses containment entirely. This could happen in- ~
two general ways. First, there'could be a failure to completely L isolate the containment. For example, isolation valves could fail , to close, manual. valves could be left open following a leak rate 1 test of a containment penetration (unlikely), an equipment hatch. 1 could be left open (unlikely), etc. The likelihood that a sizeable containment penetration could be left open without detection is low l since this would be indicated by excessive nitrogen makeup flow to the primary . containment and could be identified as constantly i increasing containment pressure.. However, a leak rate test connection foz; a primary containment isolation valve could be left open,.with ne indication of this condition in the control room. In an arrangement like the ene shown in Figure 5.1, which is common to many systems with lines' penetrating the primary containment, if the inboard isolation valve should fail to close during an accident, the open leak rate testing line (through the starred ' valves) would provide a path to bypass the containment altogether. Note that this path could be difficult or impossible to isolate once there is a significant release of fission products into
- containment, since the leak rate test valves can only be closed manuall,y. y To reduce the probability of this. mode of containment bypass, the utility could require that all such valves be administratively locked closed when not in use, or the leak test-piping could be capped. Note that this is a requirement of the General Design Criteria of Appendix A to 10 CFR 50; however, experience has shown that utilities may not always rigidly conform to this requirement in the case of leak testing, vent, and drain lines. With bypass through such paths eliminated, the probability of containment bypass due to an open penetration reduces to the probability that two isolation valves in one line fail to close.
This considered to not be a significant concern. The other way in which containment could be bypassed is by the so-called interfacing systems 14CA, also known as an Event V sequence in the terminology of WASH-1400. In this sequence, there is a failure of one or more valves that forn a boundary between the high pressure reactor coolant system and a low pressure system outside containment. Such sequences have been found in past PRAs to be insignificant contributors to the overall core melt frequency and to risk. However, because of the 1987 precursor event at the Biblis-A PWR in West Germany, the Office of Nuclear Reactor Regulation (NRR) has initiated an NRC review program to re-evaluate the contribution of the V scquence to risk at U.S. plants. In addition, a recent report by BNI? estimated the core melt frequency for three U.S. BWRs. The estimate ranged from 1.02E-6/yr for Peach dottom to 8. 81E-6/yr at NMP-2. Therefore, depending on the outcome 69
4
~ ..
of the NRC review program, the improvements identified in Reference 57, and possibly others, may need to be implemented in order to lower the contribution to risk from this sequence. 1 5.2 Potential Improvements j i Becauce these sequences have been generally found to be insignificant contributors to core melt frequency and to risk, there is no improvement that is felt to be cost-beneficial at this
]
time. - 1 i i 0 Drywell {*
- 1 I*
Leak Test Connection Mark 11 Contdnment Figure 5.1 Typical Leak Test connection for Containment Isolation valves 70 l
\ ~. N ., h 6. RISK ANALYSIS OF CONTAINMENT CHALLENGES AND IMPROVEMENTS 6.1 Core Melt Frequency Table 6.1 summarizes the core melt sequence frequencies of the dominant sequences in the Limerick, Susquehanna, and Shoreham , risk studies. The primary challenges to core integrity come from . l transients, SBO, and ATWS. Loss of coolant accidents (LOCAs) are l not large contributors to core melt frequency, nor are TW I sequences. Note that ATWS is a fairly miner contributor for I Limerick, as a result of compliance with the ATWS rule (10 CFR l 50.62). For Shoreham, the contribution from SBO sequences has been i drastically reduced by the addition of 3 more emergency diesel generators (EDGs) (total of 6) and the inclusion of a 20MW on-site blackstart gas turbine generator in the loss of off-site power event trees. The Shoreham contribution from ATWS was reduced by i the proposed installation of the Filtra filtered containment vent system, the u'se of highly enriched boron for SLC (shutdown in 7.5 mins.), and the addition of an ADS inhibit switch for use during an ATWS. Note that transients are a large contributor to core melt for all of the above plants. In particular, the sequences involving a loss of feedwater with high pressure core melt (TQUX) were found to be dominant. This is in part due to the use of 2.4E-3 as the ADS failure probability. This value was based on the' ADS initiation logic in place before Reference 51 (TMI Action Items) required modifications to the logic to make ADS failure less likelym e.g., removal of the coincident high drywell pressure signal-from the set of required initiators. An updated analysis of the ADS failure probability should reduce the frequency of depressurization failure (event X), and thus reduce the contribution from the TQUX sequences. 6.2 Sequence and Failure Mode Risk Significance As a first step in cualitativalv determining the potential impact upon risk of various containment and systems modifications, seteral sensitivity studies that have been performed for Susquehanna, Limerick, and Shoreham were , examined.(s e These studies calculated the effects of a particular design feature (e.g., containment venting), or of a proposed modification (e.g. , enhanced SLC system) upon the frequency of each identified containment failure mode and upon risk, with risk being defined in terms of the number of expected acute and latent health effects in the surrounding population. Note: since Shoreham did not have an approved emergency plan at the time of the sensitivity study, only dose-versus-distance probabilities could be used to measure risk. The first study examined was the IPE performed by Pennsylvania Power and Light (PP&L) for Susquehanna.s Since this was essentially a Level 1 PRA, no detailed analyses of containment 71
response or off-site consequences were performed. However, for each of the five initiator categories used in the study, PP&L proportioned the core damage frequency among four and states, these q being core danace, core 3g11 with vessel failure, core melt with j vessel failure and subsequent loss of containment integrity due to ] venting, and core melt with vessel failure and subsequent overpressure containment failure. The base case core melt . frequency is 1.5E-7/ reactor-year (refer to Table 6-2 for additional information). 1 The first sensitivity study for Susquehanna was to examine the effects of using the " traditional PRA" value of 0.1 for the probability that the operator fails to initiate SLC during an ATWS l when SLC initiation is required by the EOP. The base case values in Table 6-2 were calculated using a value of 0.0 for this probability, i.e., the operator never fails to initiate SLC when i it is called for in the EOP. This changes the ATWS frequency from ' essentially zero to 9E-9/ reactor-year, but does not significantly affect the total core melt fro:guancy. , In their review of the Limerick PRA, BNL calculated the effects on core melt frequency, containnsnt failure mode frequenc and off-site consequences of not complying with the ATWS rule. g-The plant modifications consisted of an ARI system and a three-pump, automatically initia tad ' SLC system.. With these changes, failure of the SLC system becomes dominated by the probabuity that the operator overrides the automatic SLC initiation signal. This modification reduces core melt frequency for ATWS by an order of magnitude but does not significantly affect the total core melt frequency from internal events. An INEL report" provided the results of an analysis of the effects of venting on core melt frequency and risk at the Shoreham Nuclear Power Station. Three sensitivity cases were examined: venting via the proposed Filtra system, venting via the existing HVAC ductwork in the reactor building, and not venting. The study included the effects of venting on systems taking suction from the suppression pool, and the effects of venting on the reactor , building atmosphere and on reactor building retention of fission products. The results are shown in Table 6-3 for the assumption that venting via existing ductwork leads to failure of equipment located in the reactor building with a release that effectively bypasses secondary containment. As can be seen from Table 6-3, containment venting was only found to affect the core melt frequency of TW and ATWS sequences. Venting through Filtra was found to recuce the total core melt frequency by approximately 184. However, with the assumption of reactor building equipment failure following venting, the use of the existing ductwork vent was found to have no benefit in reducing core melt frequency. For ATWS sequences, this was primarily due to the size of the vent lines (6 in. diameter) being too small to 72
f
- 1 adequately relieva containment pressure. In the sequences f involving loss of containment heat removal, the venting-induced j equipment failures in the reactor building were assumed to lead to {
loss of vessel injection and subsequent core damage. In terms of the containment release mode, which can loosely be thought of as being equivalent to the containment failure mode . used in other studies, the most favorable results are again obtained by venting through Filtra: almost 94% of all core damage sequences are recovered (release modes A and D) versus approximately 75% recovery for the existing vent and no-vent cases. In comparison with not venting, use of the existing ductwork vent does reduce the probability of early overpressure containment failure (at or near the time of vessel failure), but there is.an increase in the probability of late thermal failure of the'dryvell head seal resulting in late venting with an uncontrolled release of fission products, with essentially no holdup time (release modes C4 and C2, respectively). The risk of a release at Shoreham could only be assessed in terms of the likelihood of exceeding a specified dose at a given distance from the plant. As Figures 6.1 and 6.2 show, venting through Filtra is effective in reducing risk by approximately a factor of five in comparison with both the no-vent and exiscing vent cases. On the other hand, use of the existing ductwork vent is seen to actually increase the risk in comparison with the no-vent. case by a factor of approximately five. This somewhat surprising result is, as before, related to the small size of the vent lines: in order to mitigate a TW sequence, MAAP calculations . indicate that poth the watwell and drywell vent lines would have to be opened. Opening the drywell line would cause fission products to bypass the suppression pool, resulting in an unscrubbed release. This is compounded by the failure of the reactor building to provide significant fission product retentien. This increase in risk occurs despite the reduction in the probability of early - overpressure containment failure provided by venting. The reason for this is that a fraction of the early releases would be scrubbed through the suppression pool (containment failure in the watwell , airspace), reducing the severity of the release. Also, the increased probability of late thermal failure contributes to higher risk because the failure location is in the drywell head region. This produces a release to the refueling floor, directly bypassing the reactor building. Reference 24 also perfermed a gross estimate of the allowable costs for installing Filtra, based on the perceived reduction in risk (dose). The Shoreham base case risk was estimated at 154 man-rem / reactor-year. Using a value of five as the Filtra risk reduction factor gives a Filtra base case risk of approximately 31 man-rem / reactor-year, p. net benefit of 123 man-rem / reactor-year. Therefore, the upper bound total cost benefit was estimated as 73
e 7 N .e (123 man-rea/ reactor-year) ( $1000/ man-rem averted)(40-year plant life) = $4.9M. The costs of actually backfitting the Filtra system to a U.S. BWR are not known. However, because of the large amount of construction required, and the need to install new lines passing through the primary containment, the allowable cost calculated , above is probably only a small fraction of what the actual cost would be. Therefore, installation of a Filtra-type filtered containment vent system is not likely to be cost beneficial at U.S. Mark II BWRs. 6.3 Summary of Potantial Improvements Numerous han'ean improvements and operator actions that could improve containaset performance during risk-significant accidents were discussed previously in Sections 3, 4, and 5. The discussions of the potential improvements were organized by transient type, i.e., containment challenged prior to vessel failure, containment challenged long after vessel failure, or containment bypass. In general, the improvements were selected to mitigate the containment challenges identified in Sections 3.2 and 4.2. The results' from the qualitative risk analyses are summarized in Table 6-7. . Representative costs from Mark I risk reduction studies are given in Table 6-8. Improvement 1, a hard-pipe vent system, prevents containment failure due to overprem are failure. containment overpressure challenges are most important in sequences where containment integrity is challenged pricir to core degradation (Section 3). In particular, containment venting with continued vessel injection could prevent core degradation in TW sequences and could reduce the containment pressurization rate during ATWS sequences. For sequences with core degradation prior to containment challenge (Section 4), venting would at best mitigate the off-site release and, at worst, could lead to an inadvertent release. Phenomenological uncertainties related to suppression pool bypass (from downcomer failure) suggest that the release is. not assured of being scrubbed following vessel failure (See . Section 4.2 for design-specific discussions) as compared to the Mark I hardened vent where all releases through the vent would be scrubbed. If a reliable method of preventing suppression pool bypass were available, wetwell venting would provide a controlled release and would reduce the driving pressure for other containment failures. Although a filtered vent system would alleviate the problems associated with suppression pool bypass, the high cost of the system would likely exceed the benefit. Similar to containment venting, alternate residual heat removal systems could prevent core degradation in TW sequences. However, the low frequency of TW sequences (see Table 6-1), the high system cost, and the potential for operator actions to 74
K
- i mitigate the' sequence (see section 3.4), would likely keep this systan from being cost-effective.
Enhanced reactor depressurization capability would allow reactor depressurization independent of site DC power and would ensure operability during high temperature conditions. This capability would be particularly important during station blackout - sequences when DC power would be limited or unavailable. A reliable means of depressurization would permit successful operation of. safety grade or alternative low pressure injection systems. The combination of enhanced depressurization capability and the variety of low pressure injection systems would make a reduction in the core melt frequency of high pressure transients possible. No potential drawbacks were identified. An alternate supply of water to vessel injection or containment s, prays could provide both preventive and mitigative benefits. If the reactor were at low pressure, the alternate source of water could be injected into the vessel to prevent extensive core damage. . Conversely, if the vessel were at high pressure or already failed, the water could be sprayed into the containment to mitigate (possibly prevent) the release of fission products. However, due to the low flow rate of postulated alternate pumping systems and uncertainties associated with fuel-coolant interactions, the benefits are uncertain. In addition to cooling the debris, the containment sprays would provide some scrubbing to fission products released from the melt. Although the cost erf this improvement could be low, many operator actions may-be required to utilize the system. As discussed in Section 4.4.5, the' potential benefits of core debris control are uncertain and plant-specific. Additional research and a.nalyses are needed to examine ex-vessel debris behavior for different Mark II cavity geometries._ Due to uncertainties associated with steam explosions, rapid steam spikes, and debris-spreading and coolability, the benefits of debris control improvements are unknown. In general, the costs associated with modifications to the drywell floor, downcomers, and changes in the cavity design are expected to be high. Plugging drain lines that penetrate the drywell floor may offer the greatest benefit. For example, at La Salle, drain lines in the bottom of the cavity are postulated to quickly ablate and fail following attack by molten corium. In a meeting on July 27 between Philadelphia Electric Company (Limerick) and the NRC, the licensee stated that the in-pedestal drains would fail approximately six minutes after vessel failure. Early failure of the cavity floor would result in suppression pool bypass. Flooding the drywell head has the potential for preventing or mitigating leakage through the drywell head seal. As discussed in Section 4.3, high temperature and high pressure conditions may lead to drywell head seal leakage. A pool of water above the drywell 75
1 i
., e. 'e head could provide cooling to the seal and would scrub particulate fission product releases.
Hydrogen control is already provided in Mark II plants by nitrogen inerting of the primary containment. Other than pre-emptive venting in anticipation of containment failure, no significant benefits were identified. . Past studies were identified in Section 4.4.6 that showed the potential bansfits of reactor building sprays. However, use of the diesel fire pumps has been suggested as an alternate source of water for the containment sprays. It is anticipated that direct cooling of corium in the containment would provide more benefit than spraying the reactor building. i 6 h p' 9 9 l i
)
76 i _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ ._ )
~
a Table 6.1 Core Melt Frequencies for BWR Mark-II Plants (Internal Events Oniv) Core Melt SNPS Initiator Ifin SNPS Filtra SSES La Salle 3 Transients 5.4E-6 8.2E-6 2.1E-5 7.0E-10 2.0E-6 . SBO 6.7E-6 5.6E-6 4.0E-7 7.6E-8 2.3E-5 ATWS' 1.1E-6 1.4E-5 1.5E-6 9.3E-9 neg. I4CA neg. 1.0E-6 1.5E-6 6.5E-8 neg. TW 5.5E-7 4.3E-6 5.9E-7 neg. 1.0E-5 Special 2 ,. none 6.1E-6 7.8E-6 neg. neg. Total CM Freq 1.5E-5 4.1E-5 3.3E-5 1.5E-7 3.5E-5 Notes:-
- 1. Includes sequences.in which core melt occurs both before and after containment failure.
- 2. Includes sequences initiated by loss of a DC bus, loss of service water, level instrument line break, and manual shutdown due to high drywell temperature.
- 3. Values for La Salle are estimates from the latest draft of NUREG-1150."3 l
l l 1 l I 77 i 1 - ..
- - - ~ ' --
. T N
4 Table 6.2 SSES Plant Damage State Frequencies (Base Case Values)' CM With RPV Failure and Loss of Primary Containment Intecrity- - Plant Damage Core CM With Wetwell Containment State Damace RPV Failure Vent Failure Transients 7.0E-10 1.9E-9 1.3E-9 1.4E-9 ATWS 9.3E-9 nil nil 3.3E-11 SBO 7.6E-8 3.9E-8 1.6E-8 1.6E-9 LOCA ,. 5.5E-9 4.9E-10 3.0E-11 2.5E-9' LOCA/ATWS 5.9E-8 nil nil nil Total 1.5E-7 4.1E-8 1.7E-8 5.5E-9 Notes: 1. Values'taken from the SSES IPE. 4 e S S e e 78 - _ - _ _ _ _ _ - _ - - = = _ _ _
. .. E ;n ?.1 .
4 Table-6.3 Effects of Ventinc on the Shoreham Core Melt Frecuency Accident CM Frequency CM Frequency CM Frequency
, class Filtra Ventina Existino Vent w/ No Ventinc Transients 3.0E-5' 3.0E-5 3.0E-5 SBO 4.0E-7 4.0E-7 4.0E-7 Ioss of CHR 1.9E-6 9.0E-6 9.0E-6 IDCA 1.3E-6 1.3E-6 1.3E-6 ATWS (CM < CF) 1.8E-8 2.5E-6 2.5E-6 ATWS (CF < CM) 3.5E-7 1.1E-6 1.1E-6 ATWS with 2.1E-6 0.0 -0.0 release via Filtra-Total . 3.6E-5 4.4E-5 4.4E-5 Notes: 1. All core melt frequencies are per reactor-year.
I 79 , u____.-. _ _ _ _ _ _ . _ _ _ _ - _ _ _ . _ _ _ _ _ _ _ _ . _ _ _ _ _ _ . _ _ _ _ _ _ . - - _ . _ _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ _ _ _ . _ _ _ _ __ _ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ______._.___.-._,___a__
(
'l
. l
.- . I Shoreham Release Mode Probabilities
~
Table 6.4 With'Ventina via Filtra
' Release. m.
i Mode Release Cond. . Mode . Prob. Prob. Release Mode 61.9% 2.20E-05 Al Recovered in-vessel, no CF 0.1% 2.19E-08 A2 Recovered in-vessel, vented 0.0% 0.00E+00 A3 Recovered in-vessel, late CF 61.91% 2.21E-05 0.1% 3.31E-08 B1 Early small-slow CF, pool + Rx-B DF 0.0% 3.58E-03 B2 Early small-slow CF, Rx-B DF only 0.0% 2.01E-99 B3 Early small-slow CF, no DF 0.0% 4.44E-09 B4 Early large-slow CF, pool + Rx-B DF 0.0% 0.00E+00 B5 Early large-slow CF,.Rx-B DF only 0.0% 8.74E-10 B6 Early large-slow CF, no DF 0.5% 1.69E-07 B7 Early small-mod CF, pool + Rx-B DF 0.0% 1.27E-08 B8 Early small-mod CF, Rx-B DF only 0.2% 5.95E-08 B9 Early small-mod CF, no DF 0.4% 1.43E-07 B10 Early puff CF, pool + Rx-B DF 0.0%'O.00E+00 Bil Early puff CF, Rx-B DF only 0.1% 3.58E-08 B12 Early puff CF, no DF
- 1. 3 0% -4. 63E-07 5.1% 1.80E-06 C1 Late vent, controlled release 0.1% 3.10E-08 C2 Late vent, uncontrolled release 0.0% 1.17E-10 C3 Late vent with DW fail (OT), RxB DF 0.1% 2.09E-08 C4 Late vent with DW fail (OT), no RxB DF 0.0% 2.38E-12 C5 Late OP CF with pool DF 0.0% 2.49E-12 C6 Late OP CF without pool DF 0.0% 2.68E-13 C7 Late puff CF, with pool DF 0.0% 2.74E-13 C8 Late puff CF, without pool DF 0.0% 0.00E+00 C9 Late OT CF with Rx-B DF 0.0% 1.29E-10 C10 Late OT CF without Rx-B DF .
5.20% 1.85E-06 9.8% 3.48E-06 D1 Recovered ex-vessel, vented 1.3% 4.60E-07 D2 Recovered ex-vessel,, vented NG release 20.5% 7.30E-06 D3 Recovered ex-vessel, no CF 0.0% 8.03E-10 D4 Recovered ox-vessel, late CF 31.56% 1.12E-05 100.0% 3.56E-05 80
4 , Table-6.5 'Shoreham, Release Mode Probabilities With Venting via tha Existina Containment Vent Release-Mode Release Cond. Mode Prob. Prob. Release Mode . 50.2% 2.20E-05 Al ' Recovered in-vessel, no CF 0.0% 0.00E+00 A2 Recovered in-vessel, vented 0.0% 0.00E+00 A3 Recovered in-vessel, late CF 50.214 2.20E-05 2.54 1.08E B1 Early small-slow CF, pool + Rx-B DF 0.0% 3.60E-09 B2 Early small-slow CF, Rx-B DF only 0.1%'5.56E-08 B3 Early small-slow CF, no DF 0.3% 1.32E-97 B4 Early large-slow CF, pool + Rx-B DF 0.04-0.00E+00 B5 Early large-slow CF, Rx-B DF only 0.1% 3.76E-08 B6 Early large-slow CF, no DF 3.6% 1.56E-06 .B7 Early small-mod CF, pool + Rx-B DF 0.0% 1.45E-08 B8 Early small-mod CF, Rx-B DF only 0.9% 4.08E-07 89 Early small-mod CF, no DF , 1.6% 7.20E-07 B10 Early puff CF, pool + Rx-B DF 0.0% 0.00E+00 Bil Early puff CF, Rx-B DF only 0.64'2.55E-07 B12 Early puff CF, no DF 9.72% 4.27E-06 0.0% 0.00E+00 C1 Late vent, controlled release 0.1% 4.29E-08 C2 Late vent, uncontrolled release 0.0%'O.00E+00 C3 Late vent with DW fail (OT), RxB DF 15.0% 6.58E-06
- C4 Late vent with DW fail (OT), no RxB DF 0.0% 1.76E-09 C5 Late OP CF with* pool DF 0.0% 2.57E-09 C6 Late OP CF without pool DF 0.0% 2.18E-10 C7 Late puff CF, with pool DF 0.0% 2.64E-10 C8 Late puff CF, without pool DF 0.0% 4.65E-13 C9 Late OT CF with Rx-B DF 0.0% 8.61E-09 C10 late OT CF without Rx-B DF 15.12% 6.63E-06 .
0.0% 0.00E+00 D1 Recovered ex-vessel, vented 0.3% 1.18E-07 D2 Recovered ex-vessel, vented NG release 24.6% 1.08E-05 D3 Recovered ex-vessel, no CF 0.0% 1.31E-08 D4 Recovered ex-vessel, late CF 24.92% 1.09E-05 100.0% 4.39E-05 81 ( L_______________ ____
~ Table'6.6 shoreham Release Mode Probabilities With No Containment Ventina ,
Release Mode Release Cond. Mode Prob. Prob. Release Mode 50.3% 2.20E-05 Al Recovered in-vessel, no CF 0.0% 0.00E+00 A2 Recovered in-vessel, vented - . 0.0% 0.00E+00 A3 Recovered in-vessel, late CF 50.28% 2.20E-05 9.1% 3.99E-06 B1 Early small-slow CF, pool + Rx-B DF 0.0% 5.00E-09 B2 Early small-slow CF, Rx-B DF only 0.2% 8.36E-0A B3 Early small-slow CF, no DF 1.0% 4.22E-07 B4 Early large-slow CF, pool + Rx-B DF O.0% 7.92E-10 B5 Early large-slow CF, Rx-B DF only 0.2% 7.28E-08 B6 Early large-slow CF, no DF 10.2% 4.47E-06 B7 Early small-mod CF, pool + Rx-B DF 0.3% 1.40E-07 B8 Early small-mod CF, Rx-B DF only 0.7% 3.12E-07 B9 Early small-mod CF, no DF 2.3% 1.01E-06 B10 Early puff CF, pool + Rx-B DF-0.1%'4.89E-08 Bil Early puff CF, Rx-B DF only 0.6% 2.42E-07 B12 Early puff CF, no DF 24.64 % 4.08E-05 0.0% 0.00E+00 C1 Late vent, controlled release 0.0% 0.00E+00 C2 Late vent, uncontrolled release 0.0% 0.00E+00 C3 Late vent with DW fail (OT), RxB DF 0.0% 0.00E+00 C4 Late vent with DW fail (OT), no RxB DF 0.0% 1.81E-09 C5 Late OP CF with pool DF 0.0% 2.65E-09 C6 Late OF CF without pool DF 0.0% 2.25E-10 C7 Late puff CF, with pool DF 0.0% 2.71E-10 C8 Late puff CF, without pool DF 0.0% E.09E-13 C9 Lats OT CF with Rx-B DF 0.1% 2.89E-08 C10 Late OT CF without Rx-B DF . 0.08% 3.38E-08 0.0% 0.00E+00 D1 Recovered ex-vessel, vented 0.0% 0.00E+00 D2 Recovered ex-vessel, vented NG release 24.7% 1.08E-05 D3 Recovered ex-vessel, no CF 0.3% 1.31E-07 D4 Recovered ex-vessel, late CF 24.96% 1.09E-05 100.0% 4.38E-05 82
' ' - ~ ~ " ~~~"" ^ ~ ~ ~ ~ '- -
_ __.____________.__1___________
c Table 6.7 ~ Qualitative assessment of benafits and drawbacks of nrenosed Mark II containment improvements Potential .. Incrovement Potential'Benfits' Potential Drawbacks
- 1. Vent Systems o Prevents overpressure o Filtra - very high cost a.' Filtered failures for MVSS - high cost '
containment transients with scram o Does not prevent thermal vent system o Delays ATWS failure, steam explosions, o Preemptive venting or steam spikes reduces base pressure o can lead to inadvertent prior to core damage releases of noble gases o Reduces hydrogen - secondary containment burning o Assures all releases will be scrubbed o Unaffected by suppression pool bypass
- b. Hard-pipe o Prevents overpressure o Very high likelihood of vent system failures for suppression pool bypass, with transients with scram would increase risk dedicated o Delays ATWS o Moderate high cost power source o Preemptive venting o Does not prevent thermal reduces base pressure failure, steam explosions, 7 prior to core damage or' steam spikes o Reduces hydrogen - o Can lead to inadvertent secondary containment releases burning
- 2. Alternate o Maintains suppression ' o Very high cost for ARHR containment pool subcooled ($180M+)
heat removal e Prevents overpressure
-system challenge form TW o Reduces pressurization rate from ATWS #
o RWCU enhancement los cost ($100K)
- 3. Enhanced o can prevent high press o None identified reactor core melt transients -
depressuriza- o Relatively low cost tion system o Reduces late challenge o Reduce pool bypass by vacuum breaker failure
- 4. Improved H2 o Wetwell venting would o See venting drawbacks control in mitigate / prevent o May lead to primary secondary reactor building burns containment deinerting containment 83
Table 6.7 Continued
- 5. Additional -o Prevent core melt in o Requires reactor at low supply of water low pressure pressure for injection to the reactor transients with scram o Analysis of EPG spray or containment o cooling and scrubbing initiation limit required- -
spray system of ex-vessel debris o Procedures for concurrent o Prevent or mitigate fire, if fire sys used thermal failure o Low flow rate will reduce o Independent of RHR pressure reduction o Relatively low cost, capability if fire sys used o Requires many operator actions & add. piping o Increase potential for steam explosions
- 6. Core Debris o Decreases probability o Benefits uncertain due to control of suppression pool FCI uncertainties
- a. Eliminating bypass o Increases the probability in-pedestal o Decreases the of CCI and ex-vessel downcomers probability of a steam fission product release explosion or rapid o Requires re-analysis of steam spike containment pressure suppression capability
- b. Adding in- o Increases likelihood o Increases the likelihood pedestal of quenching the core of steam explosions / spikes downcomers ex-vessel o Increases the probability o Reduces importance of of suppression pool bypass containment sprays and o May be extremely expensive venting o Require re-analysis of containment pressure suppression capability
- c. Strengthening o Decreases the o Few high temp. materials ex-pedestal probability of available (high cost) downceners suppression pool o Increase CCI probability -
bypass o Does not reduce erosion of the drywell floor
- d. Plug in- o Decreases probability o Requires system mod.
pedestal of suppression pool o can increase cavity steam cavity bypass explosion probability penetrations o Decreases probability o Requires seismic of a steam explosion reanalysis of primary or rapid steam spike containment 84
. .1 I
Table 6.7 Continued
- 7. External o Mitigates or prevents o Must be initiated early in cooling of drywell head seal the accident '
drywell head failure o toes not prevent laakage o Makage would be cif other isolation valves scrubbed by overlaying or penetrations water poci 8.Use of fire o Scrubbing of fission o Limited spray coverage protection products o May provide a greater sprays in the o Hardware already in benefit as an alternate reactor place containment spray or RPV building injection system l 1 "It l .. l l l l 85
.. }
I L Table 6.8 En==rv of cost estimates from erevious studies l l Potential Incrovement cost estimate Reference
- 1. Vent system $0.69M Ref. 60 SEA Report 87-253-07-A:1 (
a.Hard-pipe $2.90M Ref. 31 Estimated from Boston Edison Company . vent system (DPU 88-28 Request No. AG 13-6) and with does not include Tech Spec mods. I dedicated $.19M to Ref. 4 Draft NUREG/CR-4551, includes. l power source $6.1M replacement power costs. Does not
- include hard pipe (only modified power supplies to vent valves) i b. Filtered $14-33M Ref. 4 Draft NUREG/CR-4551, Appendix F, containment Several sources vent system,. $30M Ref. 61NUREG/CP-0095, November 1988, R. O.
(Filtra) Schlueter and R. P. Schmitz, estimate of the Swedish Filtra
- 2. Alternate $61-77M Ref. 4 Draft NUREG/CR-4551, with $845K containment recurring costs heat removal $100K Estimate for enhanced operation of RWCU/SPCU system system
- 3. Enhanced .Co.50M Ref. 60 SEA Report 87-253-07-A:1 Depressuriza- $1.99M Eef. 31 Estimated from Boston. Edison Company tion Capability (DPU 88-28 Request No. AG 13-6) does.
not include' Tech Spec, training, etc.
$3-14M Ref. 4 Draft NUREG/CR-4551
- 4. Enhanced $0.81M Ref. 60 SEA Report 87-253-07-A:1 containment $2.4M Ref. 56 Estimated from Boston Edison Company spray system ,(DPU 88-28 Request No. AG 13-6) does not include Tech' Spec, training, etc.*
- 5. Downcomers n/a Any downconer modification is expected to be ,
- a. Eliminating very expensive. Improvement 4.b is expected in-pedestal to be the most expensive of the three
- b. Adding options. Improvement 4.d has the lowest in-pedestal cost. May only be applicable to La Salle
- c. Strengthening ex-pedestal
- d. Plug in-pedestal penetrations 86
1 l Table 6.8 (continued) Potential Innrovement cost estimate Reference
- 6. Cooling to n/a No cost estimate is available. However, the .
drywell head hardware costs associated with a small pump or siphon are expected to be low. . Operations with the drywell head water seal, I training, technical specifications, etc. should be considered in the system cost. l l I
- 7. Improved see 1 Primary containment is inerted.
hydrogen consequently, hydrogen venting would be ur,ed control to prevent burns in the secondary
~
containment in the event of containment failure. S.Use of fire n/a ~ Existing fire sprays should initiate on high protection reactor building temperature after primary sprays in containment failure. Additional spray , reactor protection costs are not available. ! building , 6 e l J l I l l l b 1 87
1
,e ,l ,-in
- I
' I e
1
! I I -e
}
- =
! ! a i ,
- ~
l l o l I f l
! 'i 1
- f v U -
.t w
4 ~
., 6
, I I /
/ e D -N I
/ ,
^
/ E i . . . . .
o o e e N e n e n N - 0 l (s-30 s e*wl.O sweg g swipeeses se mea me *essw I Figure 6.1 Shorehan Probability of Exceeding a Dose of 5 Rens as a Function of Distance 88
- t. _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ . _ - _ _ _ _ _ _ _ _ _ _
( 4 o I l-I l
! >I l
d
~
C
..- -=
. 7 JIF
/
I i w$ gw l i
, 1 h
-p
/
t I , i
../ ,$ .
' l
/ .
, l f8-"
I
/ *
~
M
~
f rh C -
~
/ .
J
~.'C c N e e e o tw - o (9-3ct so wu.)
swehl Cet tu!** ears go seeA see oose Figure 6.2 Shoraham Probability of Exceeding a Dose of 200 Rams as a Function of Distance 89
M
. e
-7. 0 REFERENCES
- 1. SECY-88-147, U.S. NRC, Integration Plan for Closure of invere Accident Issues, V. Stallo to NRC Commissioners, May 25, 1988.
- 2. NUREG-1150, (Draft), U.S. NRC, Reactor Risk Reference Document,, February, 1987.
- 3. NUREG/CR-5225, Wagner, K. C., et al.,'An Overview of BWR Mark I containment Ventina Risk Innlications, November 1988.
- 4. NUREG/CR-4551 (Draft), Volume 3, Amos, C. N., et al.,
EYaluation of Severe Accident Risks and the Potential for Rick Reduction: Peach Bottom. Unit 2, May 1987.
- 5. NUREG/CR E 3028, Papazoglou, I. A., and Karol, R., Review of the Limerick Generatina Station Probabilistic Risk Assessment,. February 1983.
- 6. Probabilistic Risk Assessment. Shoreham Nuclear Power Station, Science Applications, Inc., Prepared for Long IFland Lighting Company, June 24, 1983.
- 7. E. R. Burns, et al., Shoreham Nuclear Power Station Full Power PRA. PRA Undate: Sucolamental Containment System Implementation, IT/Delian Corporation, Prepared for Long Island Lighting Company, February,198&.
7a Z. T. Mendoza et al., Containment and Phenomenoloeical Event Tree Evaluation At Full Power for the Shoreham Nuclear Power Station, Science Applications International Corporation. Prepared for the Long Island Lighting Company, February, 1988.
- 8. Pennsylvania Power & Light, Susquehanna Steam Electric Station Individual Plant Examination, NPE-86-001, 1986 .
- 9. "World List of Nuclear Power Plants", Nuclear News, 32, No.
2, February 1989.
- 10. Limerick Generating Station Final Safety Analysis Report.
- 11. La Salle County Final Safety Analysis Report.
- 12. Nine Mile Point Nuclear Station Final Safety Analysis Report.
- 13. Susqu,ehanna Steam Electric Station Final Safety Analysis Report.
90
t .
- 14. Shoreham Nuclear Power Station Final Safety Analysis Report.
t
- 15. Washington Nuclear Power Station Unit 2 Final Safety Analysis Report.
- 16. Silver, E. G., " Operating U.S. Power Reactors,", Nuclear Safety.~24, September-October 1983.
' 17. NUREG/CR-4242, Castle, J. N., et al., Egrvey of Licht Water Reactor containment Svstana. Dominant-Failure Modes, and Mitiaation Onnertunities. Final Renort, R&D Associates, Jrnuary 1988.
18.- NEDO-31331, General Electric Company, BWR Owner's Group Emeroency Procedure Guidelines, Revision 4, March 1987.
- 19. NUREG/CR-3653, Greinann, et al. Final Resort. Containment Analysis-Techniaues a State-of-tha-Art Sn==ary, March 1984.
- 20. Harrington, R. M., and Hodge, S. A., " Containment Venting as a Severe Accident Mitigation Technique for BWR Plants With Mark I Containments", Nuclear Enaineerine and Desian 108, 1988.
- 21. NUREG/CP-0095, Greene, S. R., "The Role of BWR Secondary Containments in Severe Accident Mitigation: Issues and -
Insights From Recent Analyses", Proceedings of the Fourth Workshon on Containment Intecrity, November, 1988,
- 22. Annroximate Source Term Methodoloav for Boilina Water Reactors, Fauske & Associates, Inc., FAI/86-1, December 1986. ,
- 23. Wagner, K. C., Analysis of a Mich Pressure ATWS With Very l Low Make-un Flow, DOE /ID-10211, October 1988.
1 ! 24. Galysan, W. J., and D. L. Kelly, Containment Ventinq > l Analysis for the Shoreham Nuclear Power Station, EGGG Idaho, Inc. Informal Repor,t, January 1989.
- 25. NUREG/CR-4700, (Draft), Volume 3, Amos, C. N., et al.,
Containment Event Analysis for Postulated Severe Accidents: Peach Bottom Atomic Power Station. Unit 2, May 1987.
- 26. NURIG-1037, U.S. NRC, Containment Performance Workine Groun, May 1985.
- 27. NUREG/CR-4696, Blackman, H. S., et al., Containment Ventino Analysis for the Peach Bottom Nuclear Power Plant, EGG-2464, December 1986. l
- 28. Leonard, M. T., and Freeman-Kelly, R., A Preliminary 91
l e Evaluation of containment Ventina at La Salle County Station. Unit 2.
- 29. Dingman,_S.-E., BWR Reactor Buildina Environments After containment Failure, SAND-88-1515C, December 1988.
- 30. Greene, S. R., An Anzensment of the Shoreham Nuclear Power .
Station's Secondary containment Mitiaation Canability, June-26, 1987.
- 31. Bird, R. G., Boston Edison,.Information Renardina Pilarim Station Safety Enhancement Proaram, BEco ltr. 87-111 to S.
A. Varga, US NRC, dated July 8, 1987..
- 32. NUREG/CR-4767, Hatch, S. W., et al. Shutdown Decay Heat Removal Analysis of a General Electric BWR/5 Mark I, July 1987. .
- 33. NUREG/CR-4165, Dallman, R. J., et al., Severe Accident Seeuence Analysis Procram - Anticipated Transient Without Scram Simulations for Browns Ferry Nuclear Plant Unit 1, EGG-2374, May 1987.
- 35. .Papazoglou, I. A., " Risk Evaluation of the it ternate-3A .
Modification to the ATWS Prevention / Mitigation System in a BWR-4, Mark-II Power Plant", Erpceedina from the International Meetina en Licht Water Reactor Severe Accident Evaluatien. Auaust 28 to Saotember 1. 1983.
- 36. Anderson, J. G. M., et al., Analysis of Anticipated Transients Without Scram in Severe BWR Accidents, EPRI NP-5562, December 1987.
36a, Chambers, R., " Integrated SCDAP/RELAPS Analysis of a BWR High Pressure Boiloff", Proceedings of the Fourteenth Water Reactor Safety Information Meetina. Gaithersbura Maryland. October 27-31, 1986. i
- 37. NUREG/CR-3470, Harrington, M. E., and Hodge, S. A., ATWS at Browns Ferry Unit 1. Accident Secuence Analysis, July 1984. -
1
- 38. Yang, J. W., Pratt, W. T., " Mitigation _of Internally ]
Initiated Severe Accidents for a BWR Mark-II Power Plant", 1 Proceedings: International Tooical Meetina en Probabilistic I Safety Methods and Anolications, EPRI NP-3912-SR, Volume 2, Special Report, February 1985. 1
- 39. Ludewig, H., et al., "An Assessment of Core Melt Accidents in the Limerick Facility", Proceedina from the International Meetina on Licht Water Reactor Severe Accident Evaluation.
Auaust 28 to Sectember 1. 1983. 92
]
'. e
- 40. Perkins, K. R., et al., " Containment Performance for Core Melt Accidents in BWRs with Mark I and Mark II Containments", Proceedings: International Tonical Meetinc en Probabilistic Safety Methods and Aeolications, EPRI NP-3912-SR, Volume 2, Special Report, February 1985.
- 41. Erdmann, R. et al., " Mitigating Severe Accident Consequences ,
at the Shoreham BWR," Proceedina from the International Meetina en Licht Water Reactor Severe Accident Evaluation. Aucust 28 to Sectember 1, 1983.
- 42. Greene, S. R., et al., Peach Bottom Containment Resoonse C_ calculations for Unmitigated Short-term Station Blackout, ORNL, Letter Report dated February 1, 1988.
- 43. WASH-1400, U.S. NRC, Reactor Safety Stugv--An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants, (NUREG-75/014), October 1975. ,
- 44. NUREG-1116, U.S. NRC, A Review of the Current Understanding and Potential for Containment Failure from In-Vessel Steam Erelosions, June 1986.
- 45. Theofanous, T. G. et al., "An Assessment of Steam-Explosion-Induced Containment Failure, Parts I, II, III, and IV", Nuclear Science and Encineerina. Volume 97.
No. 4, December 1987.
- 46. Tong, L. S., Princieles of Desian Improvement for Licht Water Reactors, Hemisphere Publishing Corporation, 1988.
- 47. NUREG-1079, U.S. NRC, Estimates of Containment Leads from Core Melt Accidents, December 1985.
- 48. NUREG/CR-5096, Brinson, D. A., and Graves, G. H., Evaluati2D of Seals for Mechanical Penetrations of Containment Buildines, August 1988. ~
- 49. NUREG/CR-4944, Bridges, T. L., containment Penetration Elastomer Seal Leak Rate Tests, July 1987.
- 50. Study of the Structural Intecrity of Shoreham Primary Containment Under Severe Accident Conditions, Stone &
Webster Report, prepared for Long Island Lighting Company, 1988.
- 51. NUREG-0737, U.S. NRC, Clarification of TMI Action Plan Requirements, November 1980.
- 52. !1 berg, D., et al., Containment and Phenomenolocical Event v I
Tree Evaluation At Pull Power for the Shoreham Nuclear Power Station, Science Applications International Corporation, 93 1
r, 3 4
<s.
February, 1988.
- 53. Nienczyk, S. J., " Potential Effects of the Fire Protection System Sprays at Browns Ferry on Fission Product Transport",
i Proceedings; International Tonical Meetina on Probabilistic Safety Methods and Anelications, EPRI NP-3912-SR, Volume 2, Special Report, February 1985. .
- 54. NUREG-0460, U.S. Nuclear Regulatory Commission, Anticipated Transients Without Scram for Licht Water Reactors,-March 1980.
- 55. KUREG/CR-4626, Vol. 1, Neitzel, D. A. , et al. , Innrovina the -
ggJiability of Oman-Cvele Water Svatama, September'1986.
- 56. . NUREG/CR-3179, Harrington, R. M. and Ott, L. J.,_The Effect of Small Canacity, Mich Pressure Iniection Svstama on TOUV Seauences at Browns Ferry Unit One.
- 57. NUREG/CR-5124, Chu, T. L., et al., Interfacing Svatan= LotA:
Boilina Water Reactors, February 1989.
- 58. NUREG-1775, Volume 3, U.S. NRC, Operatina Ernerience Feedback Renort - Service Water System Failures and Degradations, November 1988.
- 59. NEREG/CR-4550 (Draft), Wheeler, T. A., et al., Analysis of Core Danace Freauenev: Ernert Judaement Elicitation on Internal Events Issues. Volume 2. Part 1--Ernert Panel, and Part 2--Proiact Staff, December 1988.
- 60. NUREG/CR-5278, Claiborne, E., et al., Cost Analysis for Potential BWR Mark I Containment Improvements, January 1989.
- 61. NUREG/CP-0095, Schluater, R. O., and Schmitz, R. P.,
" Filtered Vented Containments," in Fourth Workshoe On Containment Intecrity, November 1988.
- 62. Letter from D. Kelly (INEL) to J. Ridgely (NRC) dated April 5, 1989 (DLK-03-89).
- 63. Conference call between D. Whitehead (SNL) and S. Greene (ORNL) on February 9, 1989. Reference provided by J. N.
Ridgely, NRC-RES.
- 64. Hodge, S. A., et al., Primary Containment Resnonse Calculations for Unmitigated Short-Tern Station Blackout at Peach Bottom, ORNL Letter Report dated November 28, 1988.
- 65. Conference call between A. Payne (SNL) and P. K. Niyogi (NRC-NRR) on May 1, 1989. Reference provided by J. N.
Ridgely, NRC-RIS. 94
--- -;____-___._-_..__-___-_._--.1-. . - - _ - . . _
l n-
- 66. NUREG-1289, U.S. NRC, Reaulatory and Backfit Analysis:
Unresolved Safety Issue A-45. Shutdown Decav Heat Removal Requirements, Novenbar 1988.
- 67. Letter from D. Kelly (INEL) to J. Ridgely (NRC), dated June 9, 1989, (DLK-06-89).
- 68. - Soderman, E., Mitiaation of sever 3 Accidents, Nuclear Europe, pages 18-19, dated 11-12/11'87.
- 69. Pennsylvania Power and Light Company, The PP&L Aneroach to Risk Manaaement and Risk Assessment, Technical Report Number NPE-89-001, Revision 2, dated January 1989.
- 70. NUREG/CR-5030, Theofanous, T.G., et al., An Assessment of steam-Induced containment Failure, February 1989.
e 4 e 6 es 49 95
~ ~ **
e _
.,- j 4
Appendix A 1 l ROLE OF THE BWR OWNERS GROUP l EMERGENCY PROCEDURE GUIDELINES IN SEVERE ACCIDENT MANAGEMENT 4
)
This appendix discusses the interface between Revision 4 to ' the Emergency Procedure Guidelines (EPGs) and the concept of severe accident management, i.e. , actions taken by the operators and plant staff to prevent or mitigate an accident that threatens the integrity of one or more barriers to the uncontrolled release of fission products to the environment. The disdussion has been placed in an appendix because the information presented, while important to the overall program for responding to severe accidents in all BWR nuclear power plants, is not related to hardware or system aspects of improving the performance of the Mark II containment. Revision 4 to the EPGs ... generic symptomatic emergency procedure guidelines"'gontains to direct the operator response in the areas of reactor pressure vessel (RPV) control, primary and secondary containment control, and radioactive release control. The EPGs are written so as to be svmeton-oriented rather than event-oriented. Thus, the operator is not required to diagnose the cause . of an off-normal condition before taking corrective or mitigative action.
- 4' Revision 4 to the EPGs was structured to provide guidance to the operators in responding to situations that extend beyond the licensing design basis accident. However, preventive and mitigative strategies remain that Revision 4 does not address. In addition, there are questions as to whether Revision 4 provides the optimum strategy for dealing with ATWS events. These two areas of concern are the focus of this appendix.
Strateales not included in Revision 4 One strategy thr.t is potentially useful both for preventing ' containment fr.ilure and-for mitigating the effects of an off-site release is the use of containment sprays. Revision 4 only addresses the use of RHR pumps, with suction from the suppression pool, as a source of containment sprays. Also, the use of containment sprays for cooling debris present on the drywell floor after reactor vessel failure is not addressed. In fact, the Drywell Spray Initiation Limit in the EPG might very well prevent the use of sprays for debris cooling. As discussed in Section 3.3.2, the fire pumps and the condensate transfer pumps are potential sources of containment sprays in the event that the RHR pumps are unavailable. However, with no procedural guidance on the use of these pumps for this purpose, their use can be given little credit in a quantitative risk A-1 l-_________.____.-____.2__-_.________-.-._---.__._..----___._. _ - - _ _ . - - - . . . _ _
4 . A' analysis (ir. PRAs that have been reviewed, no credit has been given to these two. sources). Another concern is the direction given to the operators for venting the primary containment. Revision 4 directs that the primary containment is to be vented to prevent exceeding the PCPL, regardless of the off-site. radioactivity release. If the - suppression pool were not bypassed, releases through a wetwell vent
.line would be scrubbed by the water in the suppression pool.
However, if the' suppression pool were bypassed, e.g., by-downconer failure, the release would not be scrubbed, and the off-site consequences could be significantly more severe. This is an accident management. question that the EPGs currently do not address. Onarator actions durincr an ATWS Preventi'on of core melt during-an ATWS focusesprfmarilyon insertion of negative reactivity to reduce core power Failing this, actions must be taken to prevent containment failure since this would be - like?.y to induce core melt. The ATWS sequence is unique in terus of operator response because it requires that actions be taken that would be inappropriate for other accidents. For example, the ADS must be prevented from depressurizing the reactor since this could result in the rapid injection of large quantities of relatively cold water into a critical core. Furthermore, vessel level might have to be lowered to the' top.cf active- fuel (TAF) to lower reactor power. In addition, should depressurization become necessary after boron-injection has begun, e.g., to nvoid exceeding the suppression pool HCTL, the operator would be faced with a severe challenge in controlling the depressurization and subsequent low pressure injection to prevent the boren from being flushed out of the core. Entry into the ATWS portion of the EPGs is specified whenever a condition exists for which reactor scram is required, and reactor power is above the Average Power Range Monitor low level trip setpoint, or is undetermined. There are other symptors that could . be' indicative of an ATWs, e.g. , multiple open SRVs, but by focusing the operator's attention upon these two, which can be checked very quickly, the^ procedure minimizes the operator response time by not having the operator check too many redundant indications. This is important since any delay in operator response could exacerbate the situation by increasing the heat load on the suppression pool, lessening the time until containment integrity is challenged. Should the entry conditions exist, the procedure then directs the operator to manually initiate ARI, if it has not already been initiated, and to trip the recirculation pumps. If ARI is successful, the ATWS will be terminated by control rod insertion. If it is not successful, tripping the recirculation pumps will cause rapid core voiding and a substantial reduction in power. A-2 l
Y
. e a
Once ' these actions have been accomplished, the procedure then directs the performance of a number of actions concurrently, all sfmed at inserting negative reactivity. The first of these actions is to inhibit ADS and initiate Standby Liquid Control (SLC) to inject boron before reaching the boron injection initiation temperature in the suppression pool - (typically 110 T. ) . A delay in operator response could exacerbate the situat. ion by increasing the heat loading on the suppression pool, lessening the time until containment integrity is challenged. The procedure also outlines several alternative methods of inserting control rods into the core, including a) manual insertion of individual control rods,. bypassing the Rod North Minimiter and Rod Sequence control System as necessary to allow the insertion of high worth rods first, b) initiation of a manual' scram after resetting the RPS and ARI logic to drain the scram discharge volume, and .c) insertion of rods by de-energizing the scram solenoids, venting the scram air header, opening individual scram test switches, or venting the over-piston volumes of the individual hydraulic control units (HCUs). There are a number of positive features of this procedure for First of all, as pointed out above, the procedure ATWS mitigation.~ number of entry conditions, reducing the time has a minimal required for operator diagnosis of the situation. Second, should boron injection via the SLC system fail, the procedure directs the use of- alternate systems (such as RCIC) for injecting boren. Third, actions that can be taken to insert control rods are detailed in the procedure. Defeat of:HPCI/RPCS and RCIC suction transfer to the suppression pool is also required by the procedure. Therefore, failure of high pressure injection due to loss of lube oil cooling is not a concern if the procedure is followed. On the negative side, the operator might weigh the risks versus benefits in his or her mind, and thus might fail to initiate j SLC, or might delay initiation for too long. Typically, a value of 0.1 has been assigned to the probability that the operator fails "to initiate SLC. Due to varia.tions in the training philosophy at . individual plants, this value may be conservative in some cases. For comparison purposes, the Susquehanna IPE effectively claims a value of 0.0.a i Since the SLC system hardware is very reliable, variations in I the above value can have a profound impact upon the outcome of the analysis. At Susquehanna, for example, the frequency of 4 overpressure containment failure due to ATWS increases from 3.3E-11/*fr to 5.1E-6/yr when the probability of operator failure to iniziate SLC is increased from 0.0 to 0 .1. 8 With this one cht.nge, ATWS sequences became the dominant contributor to corpressure containment failure; previously, ATWS had been a l hegligible contributor. A-3
- y e *+
}
4
~ .
A Some plants have taken steps to reduce the time required for shutdown of the reactor following SLc _ initiation, either by increasing the SLc injection flow rate, or equivalently, by increasing the concentration of the sodium pentaborate solution. The not effect of thess changes is that the initiation of SLc can be delayed longer without intolerably adverse effects. The 1988 update to the Shoreham PRA, for example, gives 30 minutes as an - allowable delay time, based on the calculaged time to shutdown of 7.5 minutes .after boron injection begins. Limerick instituted the so-called ATWS-3A modification, in which utilized, with an automatic initiation signal. gree This SLc pumps changes are the probability that the operator fails to initiate SLc to' a probability that the operator will override the automatic initiation signal, a similar concern. As an aside, the SLc system at Limerick injects through the core spray sparger, eliminating concerns about inadequate mixing of boron in the lower plenum of the reactor v,essel. There are also questions about how effecti've the guidance of the EPGs would be in terminating or mitigating an ATWS event. Concerns include 1) the efficacy of lowering the reactor vessel water level to the TAF (a RELAPS calculation was done at INEL for 4 a BWR/4power),gedicted2) that power in the range of 17-20% of rated thermal the operator's ability to control level at the TAF due to inaccuracies in indicated level versus actual level, and
- 3) the potential undesirable effects associated with manual reactor pressure control. For example, reclosing SRVs fcdlowing manual depressurization could generate power and pressure spikes. This is because the change in void fraction with respect to changes in pressure is approximately two orders of magnitude larger at 85 psig than it is at 1035 psig.
A human factors analysis of an ATWS at Browns Fer g Unit 1 found uncertainty in the timing of boron injection. The uncertainty was rooted in ". . .the considerable dif ficulty operators would have in deciding to execute the tarJe and the high level of stress accompanying the decision." A nominal human error i probability of failure to initiate boron injection of 3.69E-2 was . calculated, with upper and lower uncertainty bounds of 2.59E-1 and 1.47E-2, respectively. Because of the above considerations, the human factors analysis concluded that " ...it may be more appropriate to take the worse case scenario and use the upper [ bound) as a more conservative estimate." The effect on the Suaquehanna core melt frequency of increasing the probability of operator failure to initiate SLc from 0.0 to 0.1 is shown in Table A-1. There are also uncertainties associated with the EPG direction to control level at the top of active fuel (TAF) while boron is being injected into the core. First of all, there is uncertainty as to what reactor power would actually be with level at the TAF and pressure near normal operating pressure. A RELAP5 calculation A-4
- - - - _ ___Em._____._____.______.______[*$' "*
~$ 1
~
,1 A
i was dono at INEL for a BWR/4 that predicted power in the range of 17-20% rated ' thermal power, significantly above decay heat power. ,ofHowever, thia calculation is sensitive to the axial power ! profile in the core, which is in turn a function of the inlet coolant enthalpy. The inlet coolant enthalpy would depend upon the amount of feedwater heating that resulted from uncovery of the feedwater spargers as vessel level was lowered. - Another uncertainty in controlling level at TAF stems from inaccuracies in indicated level versus actual level. Since the wide range level instruments are calibrated " hot," they would give reliable indication of level during an ATWS. However, the wide range instruments do not extend to the TAF; with the wide range instruments at the bottom of their indicating range, vessel level would still be several inches above the TAT. The post-accident flooding instrumentation does extend down below the TAF, but these instruments are calibrated " cold." Therefore, they could not generally be relied upon for accurate indication during an ATWS. The uncertainties associated with level control could be overcome if operators were to control injection flow rate, rather than level. Since reactor power is determined by flow rate through the core during an ATWS, a flow rate could be specified that would place reactor power near the decay heat removal capabilities of the
~
RHR system. Flow control has been shown to be an effective means of controlling the reactor power and slowing the containment heatup rate. -Very low flow rates would result in very low reactor power (-44) g would significantly slow the containment pressurization rate. However, on the downside, very low flaw rates may lead to partial core uncovery with core cooling provided by steam flow. Core heatups might lead to significant Zircaloy cladding oxidation and to increased hydrogen production." There could also be undesirable effects associated with operator attempts to control pressure in accordance with the EPGs during an ATWS. First of all, the operator is directed to depressurize the reactor if the suppression pool heat capacity temperature limit (HCTL) is exceeded. As noted in the EPGs this - would be a complete depressurization to below 200 psi.'g If condensate pumps were running, this ceuld result in injection of cold water into the vessel. There is also a concern that LPCI and LPCS could inject very large quantities of water since both are initiated automatically at Level 1. The operator is relied upon to terminate all sources of vessel injection except CRD and SLC. Also, if the RHR system were being used for suppression pool cooling when vessel level dropped below Level 1, it would automatically realign to the LPCI mode. Because of this interlock, continuous heat removal from the suppression pool could prove to be difficult until level is permanently restored above Level 1. Secondly, attempts to manually control pressure after the vessel has been depressurized (e.g., by reclosing SRVs following A-5 y e *% e .
'c9 ,
l manual depressurization) could generate power and pressure spikes. This is because the change in void fraction with respect to changes in pressure is approximately two orders of magnitude larger at 85 l psig than it is at 1035 psig. Conflicts might also arise between different sections of the EPGs- when responding to the ATWS, For example, emergency ~ depressurization would be required if suppression pool temperature could not be maintained below the HCTL. However, rapid depressurization could result in boron being flushed out of the core, along with the need to control high capacity low pressure injection from_ the condensate system, LPCI, and LPCS in order to avoid recriticality and power / pressure spikes. To obviate this concern, PP&L has instituted a special ATWS HCTL of 208'F to ' allow more time for boron to be injected before approaching conditions in the suppression pool that would require emergency depressurization. PP&L also takes exception to the EPG instruction to lower level' to the TAF, because of the possibility that the ADS would depressurize the reactor; PP&L instead advocates allowing HPCI, RCIC, and CRD to inject at full flow (5700 gpa), with level stabilizing below Level 2 but well above the TAT. Doing this would also promote efficient boron mixing in the lower plenum, avoiding the need to carefully restore level after the necessary amount of boron has been injected. An analysis is not given in the Susquehanna IPE to justify the use of this special ATWS HCTL. However, if raising this limit could Je substantiated, then the allowable operator delay in initiating boron injection could be significantly increased. This would in turn lower the probability of operator failure to initiate boron injection in a timely fashion. Consequently, there would be a greater probability of successfully torsinating the ATWS with no core degradation or challenge to containrer.t integrity. e The strategy of allowing high pressure systems to inject at rated flow during an ATWS, rather than lowering level to the TAF, is also predicated upon the use of a higher HCTL for ATWS. This is because the equilibrium reactor power with an injection flow ' rate of 5700 gpm is approximately 28% of rated power, resulting in a rapid heatup of the suppression pool. If injection were not throttled, the NCTL specified in the EPG would quickly be reached. Again, the use of an increass6 HCTL for ATWS would allow additional time for injection of boron, even for the case with unthrottled flow from the high pressure injection systems. In summary, an ATWS sequence might be effectively terminated by taking the actions outlined in the EPGs. Calculations done by ORNL using BWR-LTAS show that either SLC initiation or manual rod insertion alone are sufficient to bring the reactor to a shutdown condition, with no challenge to containment integrity from over-pressurization.3' However, as mentioned above, there are concerns that following the EPGs might not terminate or mitigate the A-6 __________.____________.______._m__ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
.7
*' . l 1
4 . !
- l consequences of an ATWS event. There may be alternative actions that are easier to perform than the actions specified in the EPGs.
For example, if the MSIVs could be opened (by bypassing interlocks in accordance with the EPGs), then the turbine bypass valves could be used to reject heat to the main condenser. Drywell sprays could he used to condense steam in the containment, but this would be temporary measure, unless water could be removed from the ~ suppression pool, since high suppression nool level would eventually require that sprays be terminated. l Table A-1 SSES Plant Damage State Frequencies (Failure to Initiate SLC of 10%) CM With RPV Failure and Loss of Primary containment Intactrity ctire CM With Initiator Damage RPV Failure Wetwell Vent Cent. Failure Transients- 7.OE-10 1.9E-9 1.3E-9 1. 4 E-9 ATWS 8.4E-9 nog nog 5.1E-6 SBO 7.6E-8 3.9E-8 1.6E-8 1.4E-9 IDCA 5.5E-9 4.9E-10 3.0E-11 2.5E-9 IDCA/ATWS ' 5.3E-8 n'eg nog .. 5.9E-9 Total 1.4E-7 4.1E-8 1.7E-8 5.1E-6 O i A-7
.}}