ML18101B287

Technical Evaluation Rept of Plant IPE Exam Back-End Submittal, Final Rept

ML18101B287
Person / Time
Site:	Salem
Issue date:	08/31/1995
From:	ENERGY RESEARCH GROUP, INC.
To:	NRC, SCIENTECH, INC.
Shared Package
ML18101B285	List: ML18101B284 ML18101B286 ML18101B287 ML18101B288
References
CON-NRC-04-91-068, CON-NRC-4-91-68 ERI-NRC-95-101, NUDOCS 9603260229
Download: ML18101B287 (50)

Text

\\

ERl/NRC 95-101 TECHNICAL EVALUATION REPORT OF THE SALEM INDIVIDUAL PLANT EXAMINATION BACK-END SUBMITTAL

~.

9603260229-960321

?DR ADOCK 05000272 V

PDR Final Report August 1995 Energy Research, Inc.

P.O. Box2034 Rockville, Maryland 20847 Prepared for:

SCIENTECH, Inc.

Rockville, Maryland Under Contract NRC-04-91-068 With the United States Nuclear Regulatory Commission Washington, D.C. 20555 ENCLOSURE 3

\\

ERI/NRC 95-101 TECHNICAL EVALUATION REPORT OF THE SALEM INDIVIDUAL PLANT EXAMINATION BACK-END SUBMITTAL FINAL REPORT August 1995 R. Vijaykumar and M. Khatib-Rahbar Energy Research, Inc.

P.O Box 2034 Rockville; Maryland 20852 Prepared for:

SCIENTECH, Inc.

11821 Parklawn Drive Rockville, Maryland 20852 Under Contract NRC-04-91-068 With the U.S. Nuclear Regulatory Commission Washington, D.C. 20555

1----

E.

EXECUTIVE

SUMMARY

This Technical Evaluation Report (TER) documents the findings from a review of the back-end portion of the Individual Plant Examination (IPE) of the Salem Generating Station.

E.1 Plant Characteristics The Salem Generating Station (SGS) consists of two four-loop, 3411 MW(t) Pressurized Water Reactor (PWR) units of Westinghouse design, each of which is housed in a large dry containment. The two Salem units are very similar to the Zion plant, except that the rated power is about 5 % larger, whereas their containment free volume is about 7 % smaller.

E.2 Licensee's IPE Process The methodology employed in the Salem IPE submittal for the back-end evaluation is clearly described, and the IPE is logical and consistent with GL 88-20. The front-end analyses used the fault-tree linking methodology, and the event trees were quantified using the SETS computer code. Twenty-one event trees, and 140 supporting fault trees, were developed. Of the 818 core damage sequences from the 21 event trees, 152 sequences survived the 10-10 per reactor year truncation limit. The calculated Core Damage Frequency (CDP), including both internal events and flooding, was 5.2 x 10-5 per reactor year for SGS Unit 1, and 5.5 x 10-5 per reactor year for SGS Unit 2. The outcomes of the front-end analyses are grouped into Plant Damage States (PDSs). Systemic sequence screening criteria (i.e., all systemic sequences that have a CDP greater than 10-7 per reactor year, all bypass sequences that have a CDP greater than 10-s per reactor year, and all sequences that contribute to more than 95 % of the CDP and the total frequency of containment failure) were applied to front-end results to ascertain no important sequences were left out during binning of core damage sequences into PDSs. 18 PDSs were found to remain after applying the above criteria, and they include 51 core damage sequences.

In addition, similar PDSs were further grouped together which resulted in 9 condensed KPDSs for use in Containment Event Tree (CET) analyses.

Probabilistic quantification of severe accident progression involved the development of a CET with 30 top event questions. The CET was quantified using the RISKMAN code developed by PLG, Inc. The information found in NUREG/CR-4551 (for the Zion plant) was extensively used in the quantification of split fractions for the various CET nodes. The results of the CET analyses lead to an extensive number of end-states which are subsequently binned into a very limited number of release categories. The MAAP code was used to simulate the containment response and to calculate (to a more limited extent) the radionuclide release fractions. A very simplified methodology was used to assign source terms to release categories.

The SGS IPE was a cooperative utility-consultant effort, with most of the work being performed by PSE&G staff. PSE&G provided overall coordination of the SGS IPE through its Reliability

& Assessment group, provided engineers to support the study, performed portions of the PRA tasks in-house, and reviewed the overall results.

However, for some portions of the Salem IPE Back-End Review ii ERI/NRC 95-101

examination, outside contractor assistance was used. Contractor assistance on the SGS IPE is summarized as follows. EI International, Inc. provided technical guidance for the original Level 1 PRA. PLG, Inc. provided technical guidance to update the Level 1 PRA, and also provided guidance for the back-end analyses.

ABB Impell Corporation performed the SGS-specific containment capacity analyses.

PLG also provided assistance for human factors analyses.

Utility Resources Associates developed a reactor core model for the SGS MAAP file. Gabor, Kenton & Associates, Inc. provided technical assistance for modelling unique SGS characteristics with the MAAP code. Reliability and Performance Associates performed a review of the IPE submittal. A number of plant and containment walkdowns were performed, and containment and plant drawings were reviewed.

E.3 Back-End Analysis The submittal reports a Core Damage Frequency (CDF) of 5.2 x 10-5 per reactor year for SGS Unit 1 due to internal events and internal flooding initiators. Station Blackout (SBO) sequences

( 41 3 ), Loss of Coolant Accident (LOCA) sequences (15 3 ), transients (25 3 ), and flooding

• sequences ( 15 3) were the leading contributors to core damage. The corresponding CDF for Unit 2 was calculated to be 5.5 x 10-5 per reactor year, consisting of transients (36 3 ), SBO sequences (313), LOCA sequences (163), and flooding sequences (133). The two SGS units are nearly identical in the primary system, steam supply system, functionality, and maintenance, operating and emergency procedures. The differences in the calculated value of CDFs for the two units is attributed to the differences in the plant-specific data that was used to obtain the initiating event frequencies for transient events. An additional flooding event was also identified for Unit 2.

PDSs were used to bin the end-states of the front-end analyses, and served as entry points to containment analyses. The PDS attributes include:

Reactor pressure at the time of vessel breach (high, low)

Containment isolation system status (isolated, unisolated)

Containment cooling status (fans and/or sprays)

ECCS status (vessel injection and/or recirculation)

Probabilistic quantification of severe accident progression is performed using the event tree methodology.

The Salem CET has 30 top event questions.

The CET considers accident progression starting at the time of core damage and ending at the time of containment failure.

Each PDS was analyzed using the same event tree. Most of the important phenomena of interest to PWR severe accident phenomenology, including hot leg or steam generator tube rupture due to natural circulation; failure of PORVs to reclose following a protracted cycling period; in-vessel core coolability; hydrogen generation and combustion; in-vessel steam explosions; high pressure melt ejection; direct contairunent heating; molten core-concrete interactions; containment failure in the early and late phases of the accident; hydrogen combustion in the late phase of the accident; core debris coolability; and basemat melt-through are treated in the CET.

Split fractions to account for phenomenological uncertainties were obtained after a review of the Salem IPE Back-End Review iii ERI/NRC 95-101

NUREG-1150 analyses for the Zion plant. Special features of the Salem containment design, such as the design of the cavity and the instrument tunnel, the possibility of direct liner attack due to corium dispersion, and Salem-specific containment failure modes, such as basemat flexure, were considered in the CET analyses.

The end points of the CET analyses were binned into radiological release categories. The release categories were classified in accordance to the time of release, size of the break, and break location. Information on radiological releases and frequencies was provid~ for these three broad categories:

Early (containment failure) and large releases Early (containment failure) and small releases Late containment failure Severe accident progression in the Salem plant was analyzed using version 3.0B, Revision 17.02 of the MAAP/PWR code. Results from the containment analyses indicate that, upon core damage, the conditional probability of radiological releases (including containment bypass events) for SGS Unit 1 is 0.44. The corresponding result for SGS Unit 2 is 0.53.

Table E. l summarizes the results of the SGS CET analyses leading to containment failure likelihOOd.

Table E.l Containment Failure as a Percentage of Total CDF Containment Failure Mode Salem Salem

• Unit 1 Unit 2 Early Failure 4.8 5.1 Late Failure 36.7 44.0 Bypass (V) 0.9 0.9 Bypass (SGTR) 0.5 0.2 Isolation Failure 1.2 3.0 Intact 55.8 46.8 To a large extent, the licensee used the information presented in the NUREG-1150 analyses for the Zion plant. However, the conditional probability of early containment failure at the Salem Units 1 and 2 (due to.overpressurization and ex-mode failure) is 4.8 % and 5.1 %, respectively, and is approximately one order of magnitude larger than that calculated for Zion.

The conditional probability of late containment failure at Salem Units 1 and 2 (due to overpressurization and combustion of non-condensable gases generated by MCCI) is 37 % and 44 3, respectively, and is larger than that calculated for Zion (24 3 ). These differences are mainly attributable to the differences in the CDF profiles between the Salem and Zion plarits.

The CDF of SGS Units 1 and 2 include a larger contribution from high pressure sequences initiated by internal flooding or by loss of service water or ventilation (SWENF).

These Salem IPE Back-End Review iv ERI/NRC 95-101

sequences contribute to 39.1 % of the CDF for SGS Unit 1, and 47.3% of the CDF for SGS Qnit

2. These sequences are binned into high pressu~e PDSs C3B and C3D, and do not have ECCS available. Recovery of ECCS is not credited. These high pressure sequences have a high conditional probability of early containment failure (10 % for flood sequences, and 5 % for the SWENF sequences).

Together, these sequences lead to the calculation of a conditional probability of nearly 5 % for early containment failure at the Salem plant, and a calculation of conditional probabilities of 37% and 44% for late containment failure for Salem Units 1 and 2, respectively.

The licensee's process for the evaluation of containment failure probabilities and failure modes is consistent with the intent of Generic Letter 88-20, Appendix I. The dominant contributors to containment failure are consistent with the insights obtained from the NUREG-1150 analyses for the Zion plant. The licensee has considered the failure of the containment isolation system and containment bypass.

Failure of electrical and mechanical penetrations at elevated temperatures was considered and ruled out.

A number of sensitivity analyses have been performed. These two important sensitivity analyses (for SGS Unit 1) are described below:

0 0

Relay and switchgear room eyewash line rupture and flooding.

Because of the fact that internal flooding in the relay room and switchgear room led to core damage sequences with large CDFs, it was assumed that the eyewash basins were replaced with smaller volume designs to eliminate flooding. The total CDF was found to decrease by 14 %, and the frequency of large, early releases was reduced by 22 %.

Subsequently, in the cover letter that accompanies the IPE submittal, the licensee states that "IPE reviews and field walkdowns have identified existing floor drains previously unaccounted for in the IPE analyses", and that the floor drains have reduced the-internal flooding event frequency by a factor of two.

Procedural Change for ISLOCAs.

A weakness was found to exist for certain ISLOCAs. Leakage through RHR discharge check valves in the Salem plant is diverted to the Pressure Relief Tank (PRT), arid when the PRT disk bursts, a small LOCA is indicated. However, there are no procedures to indicate ISLOCA for this scenario. A procedural change for operators to transfer to the ISLOCA procedure was considered, and the impact was modelled. The overall CDF was found to reduce by 1 %, and the early, large release frequency was found to reduce by 4%.

E.4 Containment Performance Improvements Generic Letter 88-20, Supplement Numbers 1 and 3 [7-8] identified specific Containment Performance Improvements (CPls) to reduce the vulnerability of containments to severe accident Salem IPE Back-End Review v

ERUNRC 95-101

challenges. For PWRs with large dry containments, it is requested that the licensee evaluate the IPE results for containment and equipment vulnerabilities to hydrogen combustion (local and global), and determine the need for procedural and/or hardware improvements. The submittal has analyzed the vulnerability of the Salem containment to the potential for hydrogen pocketing.

A containment walkdown was performed to identify potential locations for hydrogen accumulation. It was postulated that the areas above the RCP seals and the pressure relief tank blowout disks had the highest potential for accumulation. A region, about eight inches deep, below the ceiling of the refuelling deck and above a plug that is located in the refuelling floor (overhead of the RCPs), was identified as an "inverted box" where hydrogen can accumulate.

However, the amount of hydrogen that can accumulate and bum is not expected to cause a significant pressure rise in the containment. In the region overhead of the PRT, the containment is open, and there are metallic gratings in the ceiling which permit the hydrogen to escape to the upper containment.

In summary, no locations where hydrogen could accumulate, bum (deflagration or detonation), and lead to containment or equipment failure, could be identified.

E.5 Vulnerabilities and Plant Improvements No definition of "vulnerability" as pertaining to containment analyses can be found in the IPE submittal. However, the licensee alludes to two "vulnerabilities" which can be ameliorated by modifications. The licensee has identified two modifications, one a hardware modification, and the other, a procedural modification, based on the results of the IPE analyses, and these include:

(1)

A design change to replace eyewash stations in the relay room and the switchgear room by a limited water supply station.

(2)

A procedural change to revise the ISLOCA procedures when leakage from RHR check valves outside the containment direct leakage to the PRT.

As noted in Section E.3, the licensee subsequently determined that the recommended design change to replace eyewash stations is no longer valid due to the identification of existing floor drains. The licensee has also stated that PSE&G has submitted the procedural modification mentioned above to the Westinghouse Owners Group (WOG) for appropriate changes in the Emergency Operating Procedures (EPGs). The Westinghouse Owner's Group responded that a check of auxiliary building radiation was needed to address the possibility of a LOCA outside the containment, and the Salem EOPs were appropriately modified to require verification of radiation detection in the auxiliary building and the pressure relief tank.

E.6 Observations The important points of the technical evaluation of the Salem IPE back-end analysis are summarized below:

The Back-End portion of the IPE supplies a substantial amount of information with regards to the subject areas identified in Generic Letter 88-20, and NUREG-1335. For Salem IPE Back-End Review Vl ERI/NRC 95-101

the most part, the separate models used in the Salem IPE Back-End analysis,are technically sound. Extensive use is made of the NUREG-1150 studies for the Zion nuclear power plant.

The licensee has addressed all phenomena of importance to severe accident phenomenology in PWRs.

The licensee has also identified an additional mode of containment failure, i.e., liner melt-through due to direct contact with dispersed debris, owing to the design of the cavity and instrument tunnel at the Salem plant.

The analyses have led to one procedural modification.

The containment structural analysis for SGS is well executed and generally superior to such analyses that have been developed for other IPEs. The treatment of temperature distributions and their effects on capacity to resist pressure failure is also well done.

The submittal includes a detailed evaluation of equipment survivability under severe accident conditions.

The submittal has addressed the recommendations of the CPI program, requested as part of GL 88-20, Supplements 1 and 2.

Salem IPE Back-End Review vii ERI/NRC 95-101

TABLE OF CONTENTS

1.

INTRODUCTION....................................... 1

2.

1.1 Review Process 1

1.2 Containment Analysis................................. 2 CONTRACTOR REVIEW FINDINGS

.......................... 5 2.1 Licensee's IPE Process................................ 5 2.1.1 Completeness and Methodology......................

5 2.1.2 Multi-Unit Effects and As-Built/ As-Operated Status..........

5 2.1.3 Licensee Participation and Peer Review of IPE.............

6 2.2 Containment Analysis................................. 6 2.2.1 Front End/Back End Dependencies....................

6 2.2.2 Containment Event Tree Development.................. 11 2.2.3 Containment Failure Modes and Timing................. 19 2.2.4 Containment Isolation Failure....................... 22 2.2.5 System/Human Response.......................... 22 2.2.6 Radionuclide Release Categories and Characterization......... 22 2.3 Quantitative Assessment of Accident Progression and Containment Behavior......................................... 25 2.3.1 Severe Accident Progression........................ 25 2.3.2 Dominant Contributors to Containment Failure............. 26 2.3.3 Characterization of Containment Performance............. 28 2.3.4 Impact on Equipment Behavior...................... 30 2.4 Reducing the Probability of Core Damage or Fission Product Release... 31 2.4.1 Definition of Vulnerability......................... 32 2.4.2 Plant Modifications.............................. 32 2.5 Responses to the Recommendations of the CPI Program........... 33

3.

OVERALL EVALUATION AND CONCLUSIONS.................. 34

4.

REFERENCES......................................... 36 APPENDIX A IPE EVALUATION AND DATA

SUMMARY

SHEET....... 37 Salem IPE Back-End Review viii ERI/NRC 95-101

LIST OF TABLES Table 1

• Summary of Key Plant and Containment Design Features for the Salem Plant........................................... 3 Table 2 Comparison of Containment Capacities......................

• 3 Table 3 SGS Unit 1 PDS Matrix............................... 9 Table 4 SGS Unit 2 PDS Matrix............................... 10 Table 5 Containment Failure as a Percentage of Total CDF: Comparison with Other PRA Studies.......................................... 27 Salem IPE Back-End Review ix ERI/NRC 95-101

AFW ATWS CDF CET CHR CPI OCH ECCS EOP EPRI ESF EVSE GKA HCLPF GL IPE ISLOCA IVSE KPDS LOCA MAAP MCCI MOV MSIV NRC PDS PORV PRA PRT PSE&G RC RCS RCP RHR RPV RWST SBO SGTR SGS SRV TER VB NOMENCLATURE Auxiliary FeedWater Anticipated Transient Without Scram

... Core Damage Frequency Containment Event Tree Containment Heat Rejection Containment Performance Improvement Direct Containment Heating Emergency Core Cooling Systems Emergency Operating Procedure Electric Power Research Institute Engineered Safety Features Ex-Vessel Steam Explosion Gabor, Kenton, and Associates, Inc.

High-Confidence of Low-Probability of Failure Generic Letter Individual Plant Examination Interfacing Systems Loss of Coolant Accident In-Vessel Steam Explosion Condensed Plant Damage State Loss of Coolant Accident Modular Accident Analysis Program Molten Core Concrete Interactions Motor Operated-Valve

• Main Steam Isolation Valve Nuclear Regulatory Commission Plant Damage State Pilot-Operated Relief Valves Probabilistic Risk Assessment Pressure Relief Tanlc Public Services Electric and Gas Company Release Category Reactor Coolant System Reactor Coolant Pump Residual Heat Rejection Reactor Pressure Vessel Reactor Water Storage Tanlc Station Blackout Steam Generator Tube Rupture Salem Generating Station Safety Relief Valve Technical Evaluation Report Vessel Breach Salem IPE Back-End Review x

ERI/NRC 95-101

1.

INTRODUCTION This Technical Evaluation Report (TER) documents the results of the review of the Salem Generating Station (SGS) Individual Plant Examination (IPE) Back-End submittal [1,2]. This TER complies with the requirements for IPE back-end reviews of the U.S. Nuclear Regulatory Commission (NRC) in its contractor task orders, and adopts the NRC review objectives, which include the following:

To determine if the IPE submittal essentially provides the level of detail requested in the Submittal Guidance Document, NUREG-1335, To assess the strengths and the weaknesses of the IPE submittal, To provide a preliminary list of questions based on this limited review, and To complete the IPE Evaluation Data Summary Sheet.

The remainder of Section 1 of this report describes the technical evaluation process employed in this review, and presents a summary of the important characteristics of the Salem Generating Station related to containment behavior and post-core-damage severe accident progression, as derived from the IPE. Section 2 summarizes the review technical findings, and briefly describes the submittal scope as it pertains to the work requirements.

• Each portion of Section 2 corresponds to a specific work requirement as outlined in the NRC contractor task order. A summary of the overall IPE evaluation and review conclusions are summarized in Section 3.

Section 4 contains a list of cited references.

Appendix A to this report contains the IPE evaluation data summary sheets.

1.1 Review Process The technical review process for back-end analysis consists of a complete examination of Sections 1, 2, and 4 through 7 of the IPE submittal. In this examination, key findings are noted; inputs, methods, and results are reviewed; and any issues or concerns pertaining to the submittal are identified. The primary intent of the review is to ascertain whether or not, and to what extent, the back-end IPE submittal satisfies the major intent of Generic Letter (GL) 88-20 [3]

and achieves the four IPE sub-objectives. A draft TER based on the back-end portion of the submittal was submitted to the NRC in January 1995. A list of questions and requests for additional information was developed to help resolve issues and concerns noted in the examination process, and was forwarded to the licensee.

The li~ensee responses [9] were reviewed. The final TER is based on the information contained in the IPE submittal [l] and the licensee responses to the NRC Requests for Additional Information (RAls) [9].

Salem IPE Back-End Review 1

ERI/NRC 95-101

1.2 Containment Analysis A detailed description of the Salem containment and plant data are provided in Section 4.1 of the submittal.

Figure 4.2 (of the submittal) illustrates some of the design features of the containment.

The Salem generating station consists of two four-loop, 3411 MW(t), Pressurized Water Reactor (PWR) units of Westinghouse design.

The containment building is a reinforced concrete cylinder with a steel liner.

Table 1 provides a summary comparison of the key plant and containment design features of the Salem, Zion and Surry plants. As can be seen from this table, the Salem plant is very similar to the Zion plant, except that the rated power is about 5 % larger, whereas the containment free volume is about 7% small. The following plant-specific features are expected to have a bearing on severe accident progression ~nd mitigation in the Salem plant:

The cavity in the Salem plant is connected to the lower containment (inside the crane wall) by an inclined instrument tunnel. The room under the seal table (called the incore instrument room) is dissimilar from the Zion design in one aspect; the flow from this room is to the annular compartment of the containment, through an opening in the crane wall 3 ft (0.91 m) wide and 8 ft (2.4 m) high. Thus, core debris dispersed by High Pressure Melt Ejection (HPME) after vessel breach, directly enters the annular compartment of the containment.

There exists a potential for direct attack of the containment liner by melt ejected from the vessel.

However, since the annular compartment floor is in the lowest elevation of the containment, it is expected to be flooded with several feet of water if RWST is injected into the containment. Even if RWST is not injected into the containment, there is expected to be several inches of water on the containment floor due to condensation on the containment walls. Whether liner melt-through (and associated leakage or failure of concrete containment walls) will occur, depends on structural and thermal-hydraulic considerations.

The cavity floor area in Salem is about 585 ff, which is about 24 % larger than the floor area in Zion. The increased floor area can assist in the coolability of core debris on the cavity floor. Water can -enter the reactor cavity by overflow of the injected water that accumulates oµ the annular compartment floor through the opening on the crane wall, over the curb in the incore instrument room, and through the instrument tunnel. If the entire RWST inventory is injected into the containment, water will fill the cavity, and surround the lower third of the RPV.

The RCS water volume-to-power and containment free volume-to-power ratios impact containment pressurization due to blowdown. Table 1 shows that these ratios are smaller for the Salem plant by about 10 % as compared with the Zion plant, however, the containment pressurization by blowdown is expected to be similar in these plants. The Salem IPE Back-End Review 2

ERI/NRC 95-101

Table 1 Summary of Key Plant and Containment Design Features for the Salem Plant j~re Salem Surry Zion Power Level, MW(t) 3,411 2,441 3,236 Volume of RCS Water, m3 357 283 368 Free Volume of Containment, m3 75,039 46,440 81,000 Mass of Fuel, kg 101,787 79,652 98,250 Mass of Zircaloy, kg 20,684 16,466 20,230 RCS Water Volume/Power, m3/MW(t) 0.10 0.12 0.11 Containment Volume/Power, m3/MW(t) 22 19 25 Zr Mass/Containment Volume, kg/m3 0.27 0.35 0.25 Fuel Mass/Containment Volume, kg/m3 1.4 1.7 1.2 Maximum Quantity of H2 Generated by Zirconium 900 720 890 Oxidation, kg Maximum H2 Concentration, 10-3 moles/m3.

5 9

6 Table 2 Comparison of Containment *capacities I

Feature I

Salem I

Surry I

Zion I

Design Pressure 47 psig (4.2 bars) 45 psig ( 4 bars) 47 psig (4.2 bars)

Failure Pressure 112 psig (8. 7 bars) 126 psig (9.8 bars) 134 psig (10.2 bars)

Concrete Type Limestone Basaltic Limestone ratios of fuel and zirconium masses to the containment free volume impact containment pressurization due to Direct Containment Heating (DCH) after vessel breach. These ratios are about 10 % larger for the Salem plant as compared to the Zion plant.

Hydrogen generation and maximum hydrogen concentration are expected to be similar between the Zion and Salem plants. Finally, the t}rpe of concrete in both the plants is limestone, which is expected to lead to the generation of significant quantities of noncondensible gases by core-concrete interactions in the late phase of severe accidents..

Salem IPE Back-End Review 3

ERVNRC 95-101

In several aspects, severe accident progression and containment response in the Salem plant is expected to be very similar to the Zion plant. No differences could be identified between the two Salem units regarding the plant and containment design.

Therefore, use of the Zion NUREG-1150 study as a reference for the Salem !PE, appears to be reasonable.

Salem !PE Back-End Review 4

ERl/NRC 95-101

2.

CONTRACTOR REVIEW FINDINGS The present review compared the Salem IPE submittal to the intent of the Generic Letter (GL) 88-20, according to the guidance provided in NUREG-1335. The responses of the licensee were also reviewed. The findings of the present review are reported in this section, and follow the structure of Task Order Subtask 1.

2.1 Licensee's IPE Process 2.1. l Completeness and Methodology The IPE submittal contains a substantial amount of information in accordance with the recommendations of GL 88-20 and NUREG-1335. The methodology employed in the Salem IPE submittal for the back-end evaluation is clearly described, and the IPE is consistent with GL 88-

20. The front-end analyses used the fault-tree linking methodology, and the event trees were quantified using the SETS computer code. Twenty-one event trees, and 140 supporting fault trees, were developed.

Of the 818 core damage sequences from the 21 event trees, 152 sequences survived the 10-10 per reactor year truncation limit. The calculated Core Damage Frequency (CDF), including both internal events and flooding, was 5.2 x 10-5 per reactor year for SGS Unit 1, and 5.5 x 10-5 per reactor year for SGS Unit 2. The outcom~s of the front-end analyses are grouped into Plant Damage States (PDSs). Systemic sequence screening criteria (i.e., all systemic sequences that have a CDF greater than 10-1 per reactor year, all bypass sequences that have a CDF greater than 10-s per reactor year, and all sequences that contribute to more than 95 % of the CDF and the total frequency of containment failure) were applied to front-end results to ascertain no important sequences were left out during binning of core damage sequences into PDSs. 18 PDSs were found to remain after applying the above criteria, and they include 51 core damage sequences. In addition, similar PDSs were further grouped together which resulted in 9 condensed "KPDSs" for use in Containment Event Tree (CET) analyses.

Probabilistic quantification of severe accident progression involved the development of a CET with 30 top event questions. The CET was quantified using the RISKMAN code developed by PLG, Inc. The information found in NUREG/CR-4551 (for the Zion plant) was extensively used in the quantification of split fractions for the various CET nodes. The results of the CET analyses lead to an extensive number of end-states which are subsequently binned into a very limited number of release categories. The MAAP code was used to simulate the containment response and to calculate (to a more limited extent) the radionuclide release fractions. A very simplified methodology was used to assign source terms to relea&e categories.

2.1.2 Multi-Unit Effects and As-Built/ As-Operated Status It appears that all relevant Salem plant and containment systems are modelled as a part of the IPE. A numb~r of plant and containment walkdowns were performed, and containment and plant drawings were reviewed. To supplement the review, videotapes of the SGS Units 1 and Salem IPE Back-End Review 5

ERl/NRC 95-101

2 were taken, which were then reviewed. The walkdown focussed on the following items:

configuration of reactor cavity; instrument tunnel; and flow paths from the instrument tunnel (and the cavity) to the other regions of the containment; areas where hydrogen could potentially form pockets in the containment; location and configuration of the containment sump; and a comparison of SGS Units 1 and 2, so that the containment analyses of the two units can be performed in a similar manner. As far as the back-end analyses are concerned, no differences between the two units were identified.

2.1.3 Licensee Participation and Peer Review of IPE The SGS IPE effort was a cooperative utility-consultant effort, with most of the work being performed by PSE&G staff. PSE&G provided overall coordination of the SGS IPE through its PRA group, provided engineers (including senior reactor operators) to support the study, performed portions of the PRA tasks in-house, and reviewed the overall results. However, for portions of the examination that the licensee determined to be beyond the expertise of PSE&G personnel, outside contractor assistance was used.*

Contractor assistance on the SGS IPE is summarized as follows. EI International, inc. provided technical guidance for the original Level 1 PRA. PLG, Inc. provided technical guidance to update the Level 1 PRA, and also provided guidance for the back-end analyses. ABB Impell Corporation performed the SGS-specific containment capacity analyses. PLG also provided assistance for human factors analyses. Utility Resources Associates developed a reactor core model for the SGS MAAP file. Gabor, Kenton & Associates, Inc. provided technical assistance for modelling unique SGS characteristics with the MAAP code.

An independent review team composed of four PSE&G staff members and one outside consultant from Reliability and Performance Associates, was formed to review the IPE results.

The detailed review comments and responses are provided in Section 5 of the submittal.

Few comments pertain to the back-end analyses. The review team felt that the in-core instrument room would be subject to severe loading after vessel breach (since there are only small openings leading out of this room), and a structural analysis of this room must be performed. However, the licensee felt that this was not necessary, since they felt that the DCH analyses performed in the IPE submittal were conservative. The review team concluded that "the analytical process,

• methodology, scope and results (of the IPE) to be standard, reasonable, and expected".

2.2 Containment Analysis This section provides a review of PDS binning, CET analyses, release category definitions, severe accident analyses, and the containment structural analyses in the submittal.

2.2.1 Front End/Back End Dependencies The entry points to the containment event trees are the plant damage states. PDSs are groupings of Level 1 core melt sequences, based on similarities in accident progression and availability of Salem IPE Back-End Review 6

ERI/NRC 95-101

containment safeguards. In the Salem IPE submittal, PDSs are a combination of five separate binning characteristics as described below: *

1.

RCS Pressure: RCS pressure at.the onset of core damage can affect melt progression in-vessel, and can impact the degree of melt ejection after vessel breach. In addition, RCS pressure is also an indirect indicator of the core damage initiating event. Two RCS pressure levels have been considered in the Salem IPE:

(i)

High pressure ( > 600 psig), and (ii)

Low pressure ( < 600 psig).

For high pressure PDSs, (1) accumulators are not expected to inject; (2) PORVs may or may not cycle, and can possibly be stuck open; (3) High Pressure Melt Ejection (HPME)

• is considered aft~r vessel breach.

2.

Containment Isolation and Bypass Status: The following four situations are considered:

(i)

Containment isolated, and not bypassed; (ii)

Containment not isolated; or failed prior to core damage; (iii)

Small containment bypass, i.e., an unisolated SGTR; and (iv)

Large containment bypass, i.e., an interfacing systems LOCA.

3.

Containment spray operation: This parameter addresses the availability and operation of containment sprays. Two modes of operation are possible:

(i)

Injection from RWST, and (ii)

Recirculation from the containment sumps.

4.

Containment fan cooler operation: SGS has five safety grade fan coolers that will start automatically in an accident.

5.

RWST Injection Status: Injection of RWST into the containment can lead to flooding of the containment floor and the cavity. The RWST can be injected into the containment if either (a) the sprays operate in the injection mode, or (b) RHR system operates in the injection mode (as ECCS). Approximately 65 3 *of the inventory of the RWST has to be injected before the cavity can be flooded.

The PDS definitions are provided in Table 4.3.1 (page 4.3-8) of the submittal. Each PDS is represented by a three letter combination, of which the first character has no significance. The Salem IPE Back-End Review 7

ERI/NRC 95-101

second character is a number between 1 and 7; numbers 1 through 3 represent high pressure PDSs. Numbers 4 through 6 signify low pressure PDSs, and number 7 represents ISLOCAs.

For the high pressure PDSs, 1 represents sequences that have ECCS injection only, 2 represents sequences that have injection and recirculation, and 3 represents sequences that do not have ECCS systems functional.

Similarly, numbers 4 through 6 represent the breakup of the low pressure PDSs into various ECCS functional states. The third character in the PDS definition is a letter between A and H. Letters A through D represent an isolated containment, and letters E through H represent an unisolated containment. Letter A represents PDSs with functional fan coolers and sprays, letter B represents functional fan coolers only, letter C represents functional sprays only, and letter D represents PDSs with no containment cooling. Steam generator tube rupture initiating events with an unisolated steam generator, are denoted by a prime (') after the three character designator. Thus, there are a total of 97 PDSs possible from the above definition of PDSs. 18 PDSs were found to remain after applying the truncation criteria listed in Section 2.1.1.2 of this TER, and they include 51 core damage sequences. In addition, similar PDSs were further grouped together which resulted in 9 so-called KPDSs for use in Containment Event Tree (CET) analyses.

The PDS definitions appear to be adequate. However, some important characteristics that are important for CET analyses, which have been considered by other PRAs, are not considered for PDS definition in the Salem IPE submittal.

Examples of such characteristics include the following: AC power status, DC power status, and the status of auxiliary feedwater system.

The definition of PDSs, together with a listing of their corresponding CDFs, are provided in Tables 3 and 4 for SGS Units 1 and 2, respectively. For the Salem Unit 1, the PDS group K3D is the dominant contributor to CDF (34.5 3 of the CDF). It is a high pressure PDS composed of transient, LOSP, and flooding-initiated sequences.

No RWST injection occurs and all containment safeguards are assumed to be failed. The next dominant contributor is K6D damage state (21.2 3 of the CDF), which is similar to K3D, except that the RCS is at low pressure at the time of core damage. A majority of sequences in this PDS have RCP seal LOCA prior to core damage, leading to low RCS pressure.

The third leading contributor to core damage is the K3A damage state (19.83 of the CDF).

RWST is injected, vessel pressure is high, and all containment safeguards are functional in this PDS. The dominant initiators are station blackout, feedwater line break, and A TWS. The next leading contributor is the K4A damage state (18.5 % of the CDF), and it is similar to damage state K3A, except that the RCS is at low pressure at core damage. The fifth leading contributor is the K6A damage state (3 3 of the CDF), where the RWST is injected into the containment, vessel pressure is low, fan coolers operate, and sprays are available. The next four dominant damage states are K3H (1.2 % of the CDF), K7 (1 % of the CDF), KlB' (0.5 3 of the CDF),

and K3G (0.273 of the CDF). They include an assortment of unisolated containment and bypass sequences as summarized below:

o 1.5 3 of the CDF is composed of sequences that have containment isolation failure.

• Salem IPE Back-End Review 8

ERI/NRC 95-101

~ --z

ti n

\\0 VI I....

0....

Table 3 SGS Unit 1 PDS Matrix CONT. ISOL STATUS ISOLATED NOT ISOLATED RCS PRESSURE AT TIME OF CORE CONT. S.G.'s STATUS FC+CS FC cs NONE FC+

FC cs NONE UN CO VERY ONLY ONLY cs ONLY ONLY ECCS STATUS A

B c

D E

F G

H INJECTION ONLY,

K1B 1.50%1 HIGH INJECTION AND 2

RECIRCULATION NONE 3

K3A K3D K3G K3H 119.8%)

134.5%1

(.27%)

11.2%)

INJECTION ONLY 4

K4A 118.5%1 LOW INJECTION AND 5

RECIRCULATION NONE*

6 K6A K6D 13.0%1 (21.2%1 7

K7 (0.94%1 INTERFACING SYSTEM LOCA (RHRSI Abbreviations:

SG's = Safeguards FC = Containment Fan Coolers CS = Containment Sprav llniectionl

Table 4 SGS Unit 2 PDS Matrix CONT. ISOL STATUS ISOLATED NOT ISOLATED RCS PRESSURE AT TIME OF CORE CONT. S.G.'s STATUS FC+CS FC cs NONE FC..o.

FC cs NONE UNCOVERY ONLY ONLY cs ONLY ONLY ECCS STATUS A

B c

D E

F G

H INJECTION ONLY 1

K1B

(.25%1 HIGH INJECTION AND 2

RECIRCULATION 0

NONE 3

K3A K3D K3G K3H (17.3%1 (31.7%1

(.26%1 (2.9%1 INJECTION ONLY 4

K4A (13.9%1 LOW INJECTION AND 5

RECIRCULATION NONE 6

K6A K6D (4.6%1 (28.1%1 7

K7 (0.94%)

INTERFACING SYSTEM LOCA IRHRSI Abbreviations:

SG's = Safeguards

• FC = Containment Fan Coolers CS = Containment Spray (Injection)

o 0.9 % of the CDF is composed of sequences associated with ISLOCA.

o 0.5 % of the CDF is associated with sequences with unisolated SGTR initiating events.

The previous discussion was focussed on the Salem Unit 1. The CDF for Salem Unit 2 is 5.5 x 10-5 per reactor year, and the breakup of the CDF among the PDSs is qualitatively similar.

However, there are numerical differences*in the values of the frequencies for the various PDSs, and they are reported in Table 4.6-8 of the IPE submittal. The breakup of the Salem Unit 2 CDF is described below. The K3D damage state is once again the dominant contributor to CDF (31. 7 % of the CD F) The next dominant contributor is the K6D damage state (28. 1 % of the CDF), which is similar to K3D, except that the RCS is at low pressure at the tiine of core damage. The third leading contributor to core damage is the K3A damage state (17. 3 % of the CDF). The next leading contributor is the K4A damage state (13.9 % of the CDF), and is similar to damage state K3A, except that the RCS is at low pressure at core damage. The fifth leading contributor is the K6A damage state (4.6 % of the CDF), where the RWST is injected into the containment, vessel pressure is low, fan coolers operate, and sprays are available. The next four dominant PDSs are K3H (3 % of the CDF), K7 (0.9 % of the CDF), KlB' (0~25 % of the CDF), and K3G (0.26 % of the CDF) damage states.

They include an assortment of unisolated containment and bypass sequences, and summarized below:

o 3.2 % of the CDF is composed of sequences that have containment isolation failure.

o 0.9% of the CDF is composed of sequences associated with ISLOCA.

0 0.3 % of the CDF is associated with sequences with unisolated SGTR initiating events.

2.2.2 Containment Event Tree Development Probabilistic quantification of severe accident progression is performed using an event tree methodology. In the Salem IPE submittal, severe accident progression was modelled by one CET composed of 30 top event questions. The CET considers accident progression from the time of core damage until 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after accident initiation. The back-end event tree is depicted in Figure 4.5-1 of the submittal. The development of the event tree is discussed in Section 4.5 and the quantification of the event tree is discussed in Section 4.8 of the submittal.

Containment failures occurring prior to or within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after vessel breach are categorized as early failures, whereas those occurring between 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after vessel breach to 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after event initiation are categorized as late failures. The 4-hour time interval has been selected to qualitatively address natural deposition mechanisms (e.g., aerosol settling and volatile fission product plateout), which would significantly reduce the in-containment airborne source term and reduce the offsite release associated with a late failure.

The CET was evaluated using the RISKMAN computer code. The discussion of the top event questions of the SGS CET follows.

Salem IPE Back-End Review 11 ERI/NRC 95-101

Top Event 0 - Plant Damage State (IE): This top event identifies the PDS for which the CET will be evaluated.

Top Event 1 - Containment Not Bypassed Prior to Core Damage (BY): The second top event considers containment bypass prior to core damage. The PDS designator automatically provides the split fraction for this CET question. Only V-sequences are considered in this node, as SGTRs are considered at a later node. Success of Top Event BY implies that there is no bypass of the containment at the time of core damage, whereas failure implies containment bypass.

Additional CET event questions related to the bypass node include the size of the bypass (Top Event LB), possible recovery actions before significant core damage (Top Event IR), and whether the bypass is scrubbed by water (Top Event Z2).

Top Event 2 - Bypass Prior to Core Damage Is Not Large (LB): For bypass sequences, this top event question determines the size of containment bypass.

Top Event 3 - No Significant Reactor Coolant Pump Seal Leakage (SL): If conditions that can lead to a RCP seal LOCA exist (no seal injection and no thermal barrier cooling), top event question 3 determines the degree of primary coolant leakage out of the seal. The concern for RCP seal leakage is relevant for slow station blackout events.

Top Event 4 - Operator Maintains Controlled Steam Generator Cooling after Battery Depletion in slow Station Blackout Events (CS): This is the only operator action explicitly included in the CET. It considers operator actions in long term station blackout scenarios to maintain the turbine-driven AFW and steam generator depressurization when the station batteries are discharged.

If operator action is successful, secondary cooling is available after battery depletion, providing additional time for AC (and DC) power recovery.

A value of 0.1 is assigned for this event, based on discussions with operations personnel.

A detailed HEP calculation was not performed for this action [9].

Top Event 5 - In-Vessel Recovery of Damaged Core (IR). This top event question addresses the possibility of core coolability inside the vessel, thereby preventing vessel breach.

In-vessel core coolability can be established in several ways. The most important path to recovery is by recovery of onsite electric power before coolable core geometry is lost, allowing ECCS injection to arrest core damage.

In addition, operators can successfully activate an alternate means of core cooling after core damage but before coolable core geometry is lost.

Once the core relocates to the lower plenum, it may be difficult to prevent vessel failure. In the submittal, recovery after a substantial amount of core debris has relocated into the bottom of the reactor vessel is not considered to be likely. The potential for the prevention of vessel breach by ex-vessel cooling of the lower head is not credited if! the submittal.

In addition, the likelihood of successful operator action to arrest core damage during the time period between the commencement of core uncovery to the time of significant core relocation is considered to be relatively small for the following reason: For the important transient and small LOCA sequences, the time interval between accident initiation to the beginning of core damage is Salem IPE Back-End Review 12 ERI/NRC 95-101

relatively long due to the large steam generator inventory. If recovery is possible, it is more likely to occur before core uncovery. The progression of core damage is fairly rapid, in particular, due (o exothermic metal-water reaction, allowing only a short time for recovery of a degraded core.

Top Event 6 - Pressurizer PORVs and.Safety Valves Properly Reseat after Cycling Open (SV):

This top event question addresses the potential for a pressurizer PORV or safety valve to be stuck open prior to vessel breach due to passage of high-temperature steam, and possibly, hydrogen and volatile gases. Such failures can occur in high pressure non-LOCA sequences with auxiliary feedwater unavailable. A stuck-open PORV or a SRV valve will depressurize the RCS, reducing the possibility and extent of HPME. Given that a PORV or SRV is stuck open, there is uncertainty whether the valve fails fully open or partially open.

Deliberate depressurization of the RCS by the plant operators after the onset of core uncovery is also addressed. Such operator actions are directed by the SGS emergency procedures. However, stuck open relief valves and intentional depressurization would convert otherwise high pressure core melt sequences into lower pressure core melt sequences, thereby reducing the propensity for ISGTR, induced hot leg creep rupture, and HPME-induced DCH. A split fraction value of 0.5 is assigned for accident sequences at system pressure, and 0.0 for accident sequences at lower pressures, based on the information presented in NUREG/CR-4551 [9].

Top Event 7 - No Induced Steam Generator Tube Rupture (IS): The Induced SGTR issue is addressed by this tope event. Note that this event addresses consequential steam generator tube failure due to (1) high accident temperatures and pressures, and (2) steam generator tube rupture as an initiating event. However, if vessel breach occurs shortly after ISGTR, and the steam generator safety valves properly reseat, the offsite release will be less significant. This possibility is further addressed in Top Event 17 (Z2).

This event is dependent on the following *factors:

The level of water in the steam generafor secondary side (SGTR requires that the secondary side be dry).

Natural convection flows from the core and reactor vessel to the steam generator tubes should be sufficient to heat the tubes to high temperatures. (This requires high primary side pressure).

Creep rupture of the hot steam generator tube(s) occurs.

The steam generator tube(s) fails before any other breach in the primary system, such as the hot leg piping, the surge line, or the reactor vessel bottom head.

Top Event 8 - No Induced Breakage of RCS Hot Leg or Surge Line (IB): The possibility of depressurizing the RCS by natural convection-induced heating and creep rupture of the RCS pressure boundary (other than induced steam generator tube ruptures) prior to vessel failure is Salem IPE Back-End Review 13 ERI/NRC 95-101

considered by this top event question.

This issue is addressed in a manner similar to the NUREG-1150 analyses for Zion plant, with a value of 0. 722 assigned for RCS pressure of 2500 psia (setpoint pressure), and a value of 0.357 being assigned for hot leg rupture at a pressure of 2000 psia.

For lower pressures, hot leg rupture is assumed improbable. Similarly, the NUREG-1150 values are used for induced SGTR.

Top Event 9 - High RCS Pressure at the Time of Vessel Breach (HP): This and the next CET top event evaluate the RCS pressure immediately before vessel breach. The RCS pressure at the time of vessel breach is dependent on the plant damage state scenario and on the success of preceding Top Events SL, SY, IS, and IB (Top Events 3,6,7 and 8).

The SGS CET analyses categorize the RCS pressure at the time of vessel breach into three categories, as follows:

1.

2.

3.

Low Medium High

< 200 psia 200 to 2,000 psia 2,000 to 2,500 psia Scenarios with no feedwater or ECCS injection will be at the system setpoint. Slow station blackout sequences with turbine-driven feedwater available will be at intermediate or high pressure depending on RCP seal leakage and on whether the steam generators have been and remain depressurized, and medium or large LOCAs will be at low pressure.

The only point that should be noted here is that the range covered as medium pressure is very broad.

However, the vessel pressure is evaluated for specific accident scenarios in the CET.

Top Event JO Medium RCS Pressure at the Time of Vessel Breach (MP): Success of this top event implies medium RCS pressure (200 to 2,000 psi), whereas failure implies low RCS pressure (less than 200 psi).

Top Event 11 No Containment Isolation or Structural Failures Prior to Vessel Breach (Cl): This event considers the possibility that the containment is unisolated, or is failed prior to vessel breach.

In the time frame prior to vessel breach, the challenges to containment structural integrity considered in the submittal include:

1.

Hydrogen combustion, and

2.

RCS blowdown prior to reactor vessel lower head breach.

A special case applies to the situation when in-vessel recovery is successful. In this case, release of some gaseous fission products and hydrogen into the containment is possible and this top event questions whether containment isolation is successful for source term determination.

Salem IPE Back-End Review 14 ERI/NRC 95-101

The treatment of hydrogen combustion in the CET is described subsequently. Using an adiabatic pressure rise model, containment pressure rise was calculated, for varying steam fractions, burn completeness, and varying initial conditions. It was noted that pressure rise due to hydrogen combustion prior to vessel breach (if it did occur) was not a threat to the containment.

However, a special case where hot leg rupture prior to vessel breach was found to be capable of releasing a substantial amount of hydrogen to the containment. If sprays are activated at the time of hot leg rupture, containment failure due to hydrogen combustion is possible.

A probabilistic distribution of in-vessel zirconium oxidation and hydrogen generation was developed following the NUREG-1150 results for Zion.

The corresponding containment pressure loads were calculated using the adiabatic combustion model, and convoluted with the containment capacity curve.

A probability of 0.010 was calculated for the conditional probability of containment *failure. It was also noted that local concentrations that can cause a Deflagration to Detonation Transition (DDT) to occur, were possible. Although the submittal does not evaluate the conditional probability of containment failure for this scenario, nevertheless, a value of 0.02 is assigned for containment failure induced by a detonation.

Top Event 12 - Containment Failure Prior to Vessel Breach Is Small (Ll): This top event question is asked only if containment failure occurs prior to vessel breach and addresses the size of the containment breach. The success path for this top event implies a benign (small) breach in the containment, whereas the failure path addresses a large isolation failure or catastrophic*

breach of the containment.

Top Event 13 - No Containment Failure due to In-Vessel Steam Explosion or Rocket Failure Modes (AL). This top event question addresses the probability of an in-vessel steam explosion of sufficient energy to fail the reactor vessel and generate a missile, which, in turn, fails the containment; i.e., the so called a-mode failure in WASH-1400. While the probability of such events is low, in-vessel steam explosions are more likely when the RCS pressure is relatively low. This top event also includes the probability for RPV rocketing-induced containment failure, which is addressed in the Zion NUREG-1150 APET but is considered by the licensee to be unlikely in plant designs like SGS or Zion with large and relatively open reactor cavity designs.

The conditional probabilities assigned by the licensee for the IVSE-induced containment failure are 8.0 x 10-3 for low (RCS) pressure scenarios, and 8.0 x 104 for high pressure scenarios.

Top Event 14 - No High Pressure Melt Ejection (ME):

This top event question addresses whether the mode of vessel breach results in a rapid and energetic pressurized ejection of core debris from the vessel. The quantification of the split fractions for this top event is strongly dependent on the scenario-specific RCS pressure at the time of vessel breach as well as on the uncertainties associated with the mode of vessel failure and the amount of debris exj,elled from the vessel. Whether a high pressure melt ejection (HPME) has occurred influences the amount of debris dispersion from the reactor cavity and the magnitude of any direct containment heating effects. The submittal uses a value of 0.81 for high pressure scenarios, and 0.62 for medium pressure scenarios. In contrast, the NUREG-1150 analyses for the Zion plant use values of 0. 79 and 0.6 for set-point pressures, and high (and medium) pressures, respectively.

Salem IPE Back-End Review 15 ERI/NRC 95-101

Top Event 15 - No Containment Failure at Vessel Breach (C2): The integrity of the containment can be challenged by events tbat occur at vessel breacb. The pressure spike resulting from the combined effects of OCH is superimposed on the pre-vessel breach containment pressure. The OCH effects include containment pressurization from the blowdown of primary system gases at vessel failure, possible steam generation if water is on the reactor cavity floor, heat transfer between core debris and the containment atmosphere, exothermic energy generated during blowdown and debris dispersal by the oxidation of the metallic corium components, and by hydrogen burns.

The key uncertainties that are associated with the OCH process are assumed to be the following:

The fraction of corium that is initially ejected from the vessel.

The fragmentation of the ejected corium and the subsequent amount that is transported to the large free volume of the containment above the operating deck.

• The amount of metal oxidation and subsequent hydrogen generation as the hot corium debris and superheated steam are ejected out of the vessel.

The amount of hydrogen burned in containment.

The rate and quantity of heat transfer to the containment gases and the relative timing of the various pressurization phenomena.

The method of evaluation of containment pressurization and failure at vessel breach is quite detailed, and discussed in Section 4.8.2.3 of the submittal. The containment pressure prior to vessel breach is obtained from MAAP calculations for each scenario. The containment pressure rise at the time of vessel breach was provided in NUREG-1150 as a function of the fraction of core ejected, size of the vessel breach, and whether the cavity was wet or dry. The containment pressurization load uncertainty distributions obtained from NUREG-1150 analyses were used after scaling them to account for the differences in containment free volume. The distribution of containment pressure loads were convoluted with the containment fragility, to arrive at the conditional failure probabilities for the Salem containment. The containment loads assigned (for the evaluation of containment failure probabilities) implicitly include the effect of ex-vessel fuel-coolant interactions [9]..

Top Event 16 - No La.rge Containment Failure at Vessel Breach (L2): Given that containment failure occurs at vessel breach, this top event question determines whether the containment failure is large or small. The method of determination of the failure size is adopted from the Zion NUREG-1150 study.

Top Event 17 - La.rge Containment Failure is Well Below Grade (Z2): Large, independent containment failure modes can occur in the containment dome (a direct release into the environment) or in the containment basemat in the Salem plant.

This top event question differentiates between the two failure modes.

Salem IPE Back-End Review 16 ERI/NRC 95-101

Top Event 18-Containment Sprays Operate after Vessel Breach (S2): If containment sprays are operating before vessel breach (as denoted by PDS attributes), this top event determines whether the sprays would continue to operate after vessel breach. The principal failure mode considered here is the possibility of debris generated at vessel breach plugging up the containment sump.

A conditional probability of 0.001 is assigned for sump plugging and spray failure. It is also assumed that a large containment failure (either in the dome or the basemat) will fail containment spray because a dome failure is assumed to fail the spray spargers, and a basemat failure is assumed to result in pressurized water ejection from the containment sump to the environment.

Top Event 19 - Core Debris Cooled (DC): If the RWST was fully injected prior to vessel breach, the water elevation would be such that it will overflow the curb surrounding the instrument tunnel opening into the incore instrument room into the cavity floor. In this case, the entire reactor cavity would be filled with water, and the lower third of the reactor vessel would be submerged. On one hand, the water would inhibit dispersal simply due to debris quenching as well as viscous and inertia effects. On the other hand,* steam production as the corium thermally interacts with the water would provide an additional mechanism for enhanced dispersal.

This top event questions whether the released corium can be sufficiently cooled to inhibit concrete ablation. If the RWST is injected, debris cooling will be very likely since the corium spread area is relatively large; the reactor cavity, the annular and lower compartment floors will remain flooded. If the reactor cavity is dry and the pre-breach vessel pressure is high, it is possible that a portion of the core debris will be dispersed into the annular compartment and spread over a sufficient area such that either residual RCS water or upward natural convection and radiation cooling of the debris is sufficient to maintain the corium-concrete interface temperature below the concrete melting temperature to prevent concrete ablation. The portion of the core that was not molten af the time of vessel breach will subsequently melt and drain down to the reactor cavity and potentially ablate the concrete floor. This top event addresses these possibilities. If water is injected and cavity is flooded prior to vessel breach, it appears that a high conditional probability is assigned for deb.ris bed coolability. Conversely, for a scenario with a dry cavity it appears that a low probability (0.01) is assigned for core debris coolability.

Top Event 20 - No Containment Liner Melt-Through due to Localized Corium Thenn.al Attack (IM): The SGS reactor cavity and instrument tunnel configuration can result in corium sweepout into the annular compartment. If sufficient corium sweepout occurs in cases where the RWST has not been injected, localized liner melt-through is possible. The floor elevation of the annular compartment is about 2 ft below that of the lower compartment, so any condensation of the initial RCS. and accumulator water would tend to collect on this floor. The containment is heavily reinforced in this area, and even if liner melt-through were to occur, no significant release would be expected until gross concrete cracking were to develop. Failure of this event implies gross concrete cracking and radionuclide release into the soil, which is approximately 20 ft below ground level.

Salem IPE Back-End Review 17 ERI/NRC 95-101

9.,

A simple conduction heat transfer model was used to evaluate the heat transfer to liner from the core debris. Details of the model are presented in Section 4.2.2 of the submittal. The licensee concluded that, in the presence of water (injected from RWST), the temperature of concrete rebars had to increase to 800°F, before cracking can occur. This value appears to be optimistic.

However, if the hot corium boils off all the water, subsequent heatup is possible, and it appears that failure will occur approximately 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after dryout. Hence, it was concluded that liner melt-through is a late mode of containment failure. The conditional probabilities assigned for liner melt-through are the following:

Case 1:

5 x 10-2 for the case only RCS water is present on the containment floor, Case 2:

1 x 10-2 for the case where sprays come on after vessel breach, and Case 3:

0.0 for the case where RWST water is injected prior to HPME.

Top Event 21 - Fan Coolers Operable after Vessel Breach (F2): Given that fan coolers were in operation prior to core damage, this top event addresses the possibility of failure of fan coolers due to environmental effects resulting from vessel breach. These effects include the pressure and temperature transient in the containment, the steam environment, possible hydrogen burns, and airborne aerosols that could plug the fan cooler system. Failure of this top event indicates that the fan coolers will not function after vessel breach. It is assumed that the fan coolers will

• fail due to aerosols if debris cooling is not available.

Top Event 22 - No Hydrogen Burn within 4 Hours of Vessel Breach (HE): Depending on the reactor vessel pressure at the time of vessel breach, whether the containment floor is flooded, and whether containment spray is operating, substantial airborne aerosols can remain in containment for some time after vessel breach. Normal aerosol agglomeration and settling processes, as well as volatile fission product condensation on cooler surfaces, will decrease the in-containment airborne radiological source term over time. However, there could be significant radiological releases if containment failure occurs while there is high airborne activity. Because of this concern, containment failures that did not occur at the time of vessel breach, but did occur within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after vessel breach, are treated as early failures in this study. Hydrogen combustion is the only likely mechanism for causing containment failure within this time frame.

Top Event 23 - No Containment Failure due to Early Hydrogen Burn (C3). This event questions whether containment failure occurs due to combustion within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after vessel breach.

Top Event 24 - No Large Containment Failure from Ea.rly Hydrogen Burn (L3). This top event is similar to Top Event L2 (event 16) and is asked only if containment fails due to hydrogen burns after VB.

Top Event 25 - Large Hydrogen Burn-Induced Containment Failure Is Below Grade (Z3). This top event is similar to Top Event Z2 (event 17) and addresses the probability of large, subsoil containment failure modes when hydrogen burn fails containment after vessel breach.

Salem IPE Back-End Review 18 ERI/NRC 95-101

• Top Event 26 - Long-Term Containment Heat Removal (HR): For sequences in which the core debris is cooled (see Top Event DC), maintaining long-term containment integrity will require containment heat removal by either the containment recirculation spray system (rejecting heat to the RHR heat exchanger) or by the fan coolers. Containment spray and fan cooler failures at vessel breach are addressed in Top Events S2 and F2, respectively. This event is included to address possible long-term failures or recovery actions. The negative impacts of late CHR recovery (possible combustion) were assessed using this node, but it was noted that a series of small pressure spikes due to combustion (insufficient to threaten the containment) were calculated were found to occur in the late phase of the severe accidents [9].

Top Event 27 - No L<Jte Containment Failure (C4): This top event question is used to evaluate the probability of a late containment failure.

"Late" is defined as within 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after event initiation. Failures beyond 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> are not considered in the IPE. This node is used to treat containment failure due to overpressurization in the late phase of a severe accident, including the effects of combustion of non:-condensable gases generated by MCCI [9].

Top Event 28-Long-Term Containment Failure is Small (IA): This top event question addresses the size of the containment failure following any long-term overpressurization of the containment.

Top Event 29 L<Jrge, L<Jte Containment Failure Mode is Well Below Grade (Z4): This top event question differentiates between above-grade and below-grade failure when Top Event IA fails.

Top Event 30 No Basemat Penetration (Bl): Long-term basemat integrity is always considered to be maintained for sequences in which the debris is cooled. This top event question is used to determine whether basemat melt-through is possible within the 48-hour time frame.

In summary, the CET analyses are very detailed and treat all phenomena of importance to the severe accident phenomenology in PWRs. Details of the quantification of the CET basic events are provided in Section 4.8 of the submittal.

2.2.3 Containment Failure Modes and Timing A probabilistic evaluation of the pressure capacity of the SGS U nit-1 containment structure was performed. Failure of the containment structure due to transient (i.e., quasi-static) and long-term (i.e., static) over-pressure loads, for various temperature conditions, was considered. The containment failure pressures were investigated for temperatures in the range of 300 °P to 800 °P.

Several modes of failure of the containment were evaluated; of these, ten were judged to be sufficiently significant (in terms of potential quantitative impacts) to be included in determining overall containment fragility and failure-mode likelihood. Piping (mechanical) penetrations and electrical penetrations are examples of failure modes that were screened-out from further evaluation, based on judgment. The ten most important failure modes are listed as follows:

Salem IPE Back-End Review 19 ERI/NRC 95-101

1.

Dome Meridional Membrane Stress

2.

Basemat Flexure

3.

Liner Tear (Local Strain Concentration)

4.

Cylindrical Shell Hoop Membrane Stress

5.

Cylindrical Shell Meridional Membrane Stress

6.

Dome Hoop Membrane Stress

7.

Basemat Shear

8.

Wall-Basemat Junction Shear

9.

Equipment Hatch

10.

Personnel Airlock (Inner)

Median pressure capacity, logarithmic standard deviation due to strength variability ({38),

logarithmic standard deviation due to modeling uncertainty ({3M), and high-confidence of low-probability of failure capacity (HCLPF capacity) are presented, for each failure mode and for three temperature conditions, in Table 4.4-1 of the IPE submittal. The four most-dominant failure modes are: first, dome meridional membrane stress failure; next, basemat failure; third, linear tear; and fourth, personnel airlock bulkhead failure. Figures 4.4-1 through 4.4-5 of the IPE submittal provide plots of cumulative failure probabilities for these modes, at temperatures of 300°F, 425°F,.550°F, 675°F, and 800°F. Figure 4.4-6 of the submittal shows plots of total cumulative failure probability for these various temperatures.

All failure modes were categorized according to the following two types of expected characteristics of failure/leakage: (1) small, controlled leakage; or (2) large, uncontrolled break.

These failure characteristics clearly have significant impacts on the expected rate and magnitude of radioactive releases.

Of the ten significant failure modes, only linear tear (Mode 3) is expected to be characterized by a small, controlled leak; all other failures are associated with the more-catastrophic large, uncontrolled break. The linear-tear failure mode has an estimated median area of 7.2 in2*

To evaluate the overall containment fragility (i.e., due to all modes combined), the IPE developed a correlation matrix of failure mode variabilities. Correlations were assumed to be either perfectly correlated or perfectly independent, relative to modeling and strength variabilities. Thus, two failure modes may be perfectly independent; perfectly dependent in strength variability and perfectly independent in modeling variability; perfectly independent in strength variability and perfectly dependent in modeling variability; or perfectly dependent in Salem IPE Back-End Review 20 ERI/NRC 95-101

both strength and modeling variabilities. Table 4.4-3 of the IPE presents the correlation matrix.

This matrix was used in statistically combining the* various failure modes.

The manner in which failure modes were combined in the containment failure analysis depended on the type of load condition. If the pressure load builds up over a long period of time (i.e.,

slow pressurization occurring over several hours), the temperature distributions and likely succession of failures (and expected fraction of the various types of failures at varying temperatures and pressures) were determined.

Under this case of lm1ding, the median containment pressure capacities were calculated as: 112 psig at 300°F, 107 psig at 550°F, and 102 psig at 800°F. In the case of transient pressure increase (i.e., rapid pressurization over a few seconds or tens of second), Monte Carlo simulation among the fragilities for the dominant failure modes was conducted for given pressure pulses.

This simulation was conducted to determine the fraction of containment rupture associated with large breaks above the ground surface (i.e., dome rupture) versus the fraction associated with large breaks below the ground surface (i.e., basemat flexural rupture). The occurrence of liner team preceding either of these two types of uncontrolled failures was assumed to not avert uncontrolled rupture, because the pressure increase would be too rapid.

The IPE states that Unit-2 of SGS is sufficiently similar to Unit-1 to enable the application of Unit-1 containment fragility results to Unit-2.

The containment failure analysis for SGS is well executed and generally superior to such analyses that have been developed for other IPEs. The probabilistic format implemented is well conceived. The treatment of temperature distributions and their effects on capacity to resist pressure failure is also well done.

The IPE states that plant-specific analysis of dynamic pressure effects was beyond the scope of the study.

Although, it is recognized that many IPEs have not undertaken plant-specific assessment of dynamic load resistance, such loading conditions are *important, they have a significant effect on structural capability, and they are not impractical to implement. It is not clear that the failure analysis considered effects of local, edge-load stresses at points of discontinuity in the containment shell (e.g., springline, knuckles, juncture with base, etc.) Local stresses often have important effects on failure capacity. Despite these observations, the containment failure assessment for SGS is a significant strength of the IPE.

The effect of elevated temperatures upon containment penetrations was considered in the IPE

[9]. A review of electrical penetrations was performed as a part of the containment capacity analyses. The electrical penetrations in the SGS containment have closure plates on each end.

The seal materials lose their geometric integrity at temperatures greater than 300°F, and leakage through the seal materials is an issue. However, tests performed in Sandia National Laboratories indicate that the seal materials contained within penetrations that have closure plates (similar to those in Salem) do not extrude and leak at elevated temperatures. Hence, the licensee concluded that electrical penetrations were not controlling failure modes at elevated temperatures.

Salem IPE Back-End Review 21 ERI/NRC 95-101

2.2.4 Containment Isolation Failure An analysis of the containment isolation system is presented in Section 3.2.1.21 of the submittal.

A listing of all the fluid systems piping that penetrate the containment, and the valves that are required to isolate the containment, are provided in Tables 3.2.1-1 and 3.2.1-2 of the submittal.

The overall conditional probability of isolation failure is 1.5 % for SGS Unit 1, and 3.2 % for SGS Unit 2. The licensee attributed the differences in results to the differences in initiating event frequencies between the two units. It is noted that the frequency of containment isolation failure is large in the SGS submittal, in comparison to other IPE submittals.

Containment bypass was also analyzed as a part of the front-end analyses. Interfacing systems LOCAs, steam generator tube ruptures and thermally-induced steam generator tube ruptures were all considered. The results of the ISLOCA analyses were documented in a separate report.

After a screening process, two potential ISLOCA paths were identified.

The paths were associated with the high pressure RCS water entering the RHR system via the suction line from

• the hot leg or the discharge line from the four cold legs. A CDF of 5.6 x 10-7 per reactor year (corresponding to a conditional probability of 0.93 of the total internal events CDF) was calculated for the ISLOCA sequences.

2.2.5 System/Human Response Two operator actions are credited in the SGS CET. The first action is credited in top event 4, "Operator Maintains Controlled Steam Generator Cooling after Battery Depletion in slow Station Blackout Events (CS)". This action was discussed earlier, and is the only operator action explicitly included in the licensee discussion. It considers operator actions in long term station blackout scenarios to maintain the turbine-driven AFW and depressurization of steam generators when the station batteries are depleted. If operator action is successful, secondary cooling is available after battery depletion, providing additional time for AC (and DC) power recovery.

A.value of 0.1 was assigned for this event after discussion with plant personnel. A second action considered in the submittal involves operator action to depressurize using PORVs in high pressure sequences. A split fraction value of 0.5 is assigned for accident sequences at system pressure, and 0.0 for accident sequences at lower pressures, based on the information presented in NUREG/CR-4551 [9].

2.2.6 Radionuclide Release Categories and Characterization The endpoints of the CETs represent the outcomes of possible in.,.containment accident progression sequences.

These endpoints are binned into a number of release categories, characterized by similarities in accident progression and source term characteristics. Associated with each source term category is a release frequency and a release magnitude. The release category definition in the Zion NUREG-1150 analyses uses eleven different attributes. The Salem IPE submittal notes that it would be impossible within the scope of the IPE submittal, to define and evaluate the magnitude of the source term of similar complexity. Hence, in the Salem IPE submittal, even though use is made of the NUREG-1150 results, the scope, the Salem IPE Back-End Review 22 ERl/NRC 95-101

definition and evaluation of the source term bins is considerably more simplified.

The licensee definition of the source term release categories encompasses all but five of the characteristics used in the NUREG-1150 definitions. The source term bins were classified based on the time of release and size of containment failure, as follows:.

Time of Release Containment failure is categorized as "early" if failure occurs up to and within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after vessel breach. "Late" failure is defined as occurring beyond 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after vessel breach. If failure does not occur at all, then the associated releases are placed in the "intact" category.

Leakage Size (and Mode of Release)

Small holes in the containment (for instance, those occurring due to liner tear) are termed "leaks", and the other failure modes are referred to as containment "breaks". The three possible locations are in the dome, in the basemat, and liner tear at the penetrations. A Salem-specific containment failure location is the basemat-bending failure mode, where releases would have to traverse about 80 ft. of soil before they can reach a water table.

Soil can provide significant decontamination, but the degree of decontamination depends on the soil properties, and water table elevation. Based on these factors, two modes of release, namely, direct (from all release locations other than the basemat), and subsoil (due to basemat-bending failure mode), are defined.

There is a discussion of the effect of sprays and cavity flooding in Section 4.9.1.1 of the submittal.

However, these effects are not used in the definition of release categories. In addition, four ISLOCA bins and two SGTR bins are considered. The classification of releases for the ISLOCA bins are based on the size of the ISLOCA and whether or not the releases in the auxiliary building are scrubbed by fire sprays. In the IPE submittal, it is concluded that the break in the auxiliary building cannot be submerged by water owing to the location of the valves. Two SGTR release categories, one with a stuck-open relief valve, and another with a cycling relief valve, were defined in the submittal.

The licensee evaluated the radiological release fractions (to the environment) for ten CET end-states that comprise 99.1 % of the CDF for SGS Unit 1, and 99.2 % of the CDF for SGS Unit

2. For the SGS Unit 1, the frequency of large early releases is estimated to be 3.95 x Io-6 per reactor year (6. 7 % of the CDF), and has significant contributions from the following modes of containment failure:

Large, early, direct releases to environment from the containment dome (69 % )

Large, early, subsoil releases (27 % )

Salem IPE Back-End Review 23 ERI/NRC 95-101

SGTR with stuck open SRVs (93)

Large ISLOCA, unscrubbed (3 3)

Large ISLOCA, scrubbed with firewater sprays in the auxiliary building (1 3)

For all of the above release categories, except the SGTR sequence, the source term code ZISOR,

• provided in Reference [5], was used to evaluate the source term. For the SGTR sequence, release fractions were evaluated using the MAAP code.

For the large, direct release bin, approximately 100 3 of the noble gases, 19 % of cesium, 16 3 of the iodine, and 3.2 % of the tellurium are reported. For the large, early releases to the subsoil, the releases provided are, 100 3 for noble gases, 2.4 % for iodine, 2 % for cesium, and

• 4 % for tellurium. A soil decontamination factor of 8 appears to have been used.

Releases for the SGTR sequence (stuck-open SRV) were obtained from MAAP simulations. The release fractions are calculated to be 90 % for noble gases, 35 % for cesium iodide, and 0.05 %

for tellurium. The release fractions calculated by the MAAP code for cesium iodide are slightly larger than similar calculations performed using the MELCOR and STCP codes, but the release fraction of tellurium reported in the submittal is considerably smaller. The small magnitude of the tellurium release fraction is not explained in the submittal.

The calculated releases f~r the bypass, sequences are presented below:

Release Fractions for Unscrubbed. Large Bypass Releases Noble gases Iodine Cesium Tellurium 100%

16%

16%

2.2%

Release Fractions for Scrubbed. Large Bypass Releases Noble gases Iodine Cesium Tellurium 100%

3.2%

3.2%

4.3%

These results are identical to those reported for Zion in Reference [5].

The release fractions for the Release Category II (small, early containment failure end-states, corresponding to approximately 0.1 % of the CDF in the SGS IPE submittal) are listed below

[9].

Salem IPE Back-End Review 24 ERI/NRC 95-101

Release Fractions for Release Category II Noble gases

'99 %

Csl 6.1%

Tellurium

1. 7 %

Source terms were also calculated using the MAAP code for Release Category III (late containment failure). The release fractions calculated were less than 1 % for all radionuclides, except the noble gases.

In summary, the licensee has performed a limited number of source term calculations, but they do represent the spectrum of radiological releases expected from severe accidents in a PWR with a large dry containment.

Generic Letter 88-20 states that: "any functional sequence that has a core damage frequency greater than or equal to 10-6 per reactor year and that leads to containment failure which can result in a radioactive release magnitude greater than or equal to BWR-3 or PWR-4 release categories of WASH-1400, and any bypass sequence that has a core damage frequency greater than or equal to 10-7 per reactor year" should be reported by the IPEs. Although the licensee chose not to address this request, the submittal provides enough information to identify such sequences. The reportable sequences are the following:

SGTR Sequence S4IsgAasPo1 Y ci (CD F of 1.1 x 10-7 per reactor year)

SGTR Sequence S4IsgHa5Pol (CDF of 1.1 x 10-1 per reactor year)

ISLOCA Sequence V-Seq-Discharge (CDF of 5.1 x 10-7 per reactor year) 2.3 Quantitative Assessment of Accident Progression and Containment Behavior 2.3.1 Severe Accident Progression Analyses were performed to predict the progression of severe accident sequences that were determined to be dominant contributors to core damage. Severe accident progression in the Salem plant was analyzed using version 3.0B, Revision 17.02 of the MAAP/PWR code. Details of the MAAP parameter input file and the MAAP simulation results are included in Section 4.7 of the submittal. A number of adjustments to a typical PWR input model had to be made to account for some peculiar features in the SGS. Examples include: the modelling of flow from the lower to annular compartment instead of the upper compartment, modelling of station blackout sequences including actions to account for secondary side depressurization in case of a long-term station blackout, modelling of floor areas in annular and lower compartments in addition to the cavity floor for core concrete interactions.

A number of scenarios were analyzed, including the following:

Salem IPE Back-Erid Review 25 ERl/NRC 95-101

( 1)

Long term station blackout with steam-driven auxiliary feed water cooling available for 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

(2)

Short term station blackout scenario.

(3)

Transient with loss of feedwater.

(4)

Long term station blackout with seal LOCA.

(5)

Long term station blackout with AC power recovery.

(6)

Long term station blackout with stuck-open PORVs.

(7)

Large LOCA with failure of ECCS.

(8)

SGTR.

For each of these scenarios, a number of variations and sensitivity calculations were performed.

Results for key thermal hydraulic parameters such as containment and RCS pressures, in-vessel and ex-vessel hydrogen generation, fraction of radionuclide inventory released or retained, etc.,

are provided in Section 4.7 of the submittal.

However, the number of sensitivity analyses performed is somewhat limited, and many of the sensitivity analyses recommended in the EPRI document [3] were not performed.

2.3.2 Dominant Contributors to Containment Failure Table 5 of this review shows a comparison of the conditional probabilities of the various containment failure modes of the Salem IPE submittal (for Units 1 and 2) with the Zion and Surry (NUREG-1150) results [2]. All comparisons are made for internal initiating events only.

The Salem core damage frequency for internal events is comparable to the CDF calculated by NUREG-1150 for Zion [5]. It was already noted that the Salem plant and containment are similar to the Zion plant. The conditional probability of early containment failure in the Salem Units 1 and 2 (due to overpressurization and a-mode failure) is 4.8% and 5.1 %, respectively, and is approximately one order of magnitude larger than that calculated for Zion. The reasons

• for the difference in the calculated probabilities of early containment failure appear to be the following:

1.

As shown in Table 1 of this review, the ratios of the mass of fuel and zirconium to the containment free volume, are 10 % larger in Salem as compared with Zion. The ratio of the mass of fuel and zirconium in the plant to the containment free volume provides an indication of OCH-induced containment loads.

Salem IPE Back-End Review 26 ERI/NRC 95-101

2.

The median failure pressure of the containment (112 psig) is approximately 16 psig less than the failure pressure of the Zion plant.

3.

The CDP of SGS Units 1 and 2 include a larger contribution from high pressure sequences initiated by internal flooding or loss of service water or ventilation (SWENF).

These sequences contribute to 39.1 % of the CDP for SGS Unit 1, and 47.3 % of the CDF for SGS Unit 2. These sequences are binned into high pressure PDSs C3B and C3D, which do not have ECCS available. Recovery of ECCS is not credited. These high pressure sequences have a high conditional probability of early containment failure (10 %

for flood sequences, and 5 % for the SWENF sequences). Together, these sequences lead to the calculation of a conditional probability of nearly 5 % for early containment failure at the Salem plant.

Table 5 Containment Failure as a Percentage of Total CDF: Comparison with Other PRA Studies

++

Containment Salem Salem Failure Mode Unit 1 Unit 2 Early Failure 4.8 5.1 Late Failure 36.7 44;0 Bypass (V) 1.0 0.9 Bypass (SGTR) 0.6 0.3 Isolation 1.2 3.0 Failure Intact 55.8 46.8 Core Damage Frequency, yr-1 5.2 x 10-5 5.5 x 10-5 Included as a part of Early Containment Failure After Charging Pump Modification Surry Zion+

NUREG-1150 NUREG-1150 0.7 0.5 5.9 24.0 7.6 0.2 4.6 0.3 NA++

1.0 81.2 73.0

4. lxIQ-5 6.2xIQ-5 In addition, the licensee has also calculated a higher conditional probability of late containment failure.

The use of li~estone for concrete is the main reason for the calculation of higher probabilities of late containment failure in Salem (and Zion), as compared with Surry. The lower containment capacities at higher temperatures (102 -107 psig for the Salem plant), is an explanation for the calculation of higher probabilities of late containment failure in Salem as compared with Zion. Other minor reasons include: identification of liner melt-through, and identification of basemat flexure and subsoil releases, as containment failure modes in the Salem Salem IPE Back-End Review 27 ERI/NRC 95-101

IPE submittal.

However, the main reason for the calculation of the higher conditional probability of late containment failure is the difference in the CDF profile between the Salem and Zion plants. As discussed previously, SGS has a large contribution from internal flood and SWENF-initiated high pressure sequences.

These high pressure sequences have a. high conditional probability of late containment failure (90 % for flood sequences and 95 % for the SWENF sequences). Together, these sequences lead to a calculation of conditional probabilities of 37% and 44% for late containment failure for Salem Units 1 and 2, respectively.

Since the dominant containment end-states are intact containment or late containment failure (with small releases), the radiological releases are dominated by SGTRs and interfacing systems LOCA sequences. Bypass sequences (SGTR and V-sequences) contribute to 1.6 % and 1.2 % of the CDP in Salem Units 1 and 2, respectively. The contribution of these bypass sequences to releases is larger than that calculated for the NUREG-1150 analyses for Zion In summary, it appears that the methods and assumptions used for containment analyses in the Salem submittal and the Zion NUREG-1150 report are, for the most part, similar.

2.3.3 Characterization of Containment Performance The progression of various. accident scenarios were modelled and quantified using an event tree and quantified using the RISKMAN computer code. The results of the containment analyses for the important release modes, are discussed below.

Release Category 1: Large. Early Containment Failure and Large Bypass Sequences (Unit 1)

This release category has a frequency of 3.95 *x 10-6 per reactor year, and accounts for 6. 7 % of the total CDF.

It includes direct releases due to structural failure (60 % ), releases through subsoil (27 % ), releases due to bypass (6 % ), and releases due to SGTR (9 % ). From an offsite risk point of view, this release category encompasses the most severe accident scenarios. The biggest contributors, by initiating event to the large, early, releases are floods ( 40 % ), isolation failure (18%), fast station blackout (10%), SGTR (7%), loss of feedwater (6%), and V-sequences (6 % ).

Release Category 2: Small. &rly Containment Failure and Small Bypass Sequences (Unit 1)

This release category accounts for less than 1 % of the total CDP in SGS Unit 1.

Release Category 1: Large. &rly Containment Failure and Large Bypass Sequences (Unit 2)

This release category has a frequency of 5.27 x 10-6 per reactor year, and accounts for 8.3 % of the total CDP. The most important difference in the results between SGS Units 1 and 2 is due to the differences in calculated containment isolation frequency. The conditional probability of failure to isolate containment in SGS Unit 2 (3.2 %) is larger than that calculated for SGS Unit 1 (1.5 %).

Salem IPE Back-End Review 28 ERI/NRC 95-101

Release Categozy 2: Small. Early Containment Failure and Small Bypass Sequences (Unit 2)

Once again, the small, early containment failure release category accounts for less than 1 % of the total CDF in SGS Unit 2.

Sensitivity analyses were performed for the following uncertain parameters for SGS Unit 1:

o Relay and switchgear room eyewash line rupture and flooding.

0 0

Because of the fact that internal flooding in the relay room and switchgear room led to core damage sequences with large CDP, it was assumed that the eyewash basins were replaced with smaller volume designs to eliminate flooding. The total CDP was found to decrease by 14 %, and the frequency of large, early releases was reduced by 22 %.

Hence, it was shown that a design change could provide a significant reduction in CDP and frequency of early releases.

Human Error Probabilities (HEPs) for controlling auxiliary feedwater flow following battery discharge in long term station blackout scenarios.

In a long term station blackout scenario in the Salem plant, operators are instructed by EOPs to depressurize the steam generators and maintain turbine-driven cooling. A HEP of 0.1 was assigned for this operator action in the base case CET analyses. After discussions with plant personnel, it was decided to increase the HEP for this action to 0.5 and 1.0, respectively, to evaluate the impact of operator actions upon releases, If no operator actions are assumed, the frequency of intact containment or no vessel breach was found to decrease, whereas the frequency of early, large releases was found to increase. On the other hand, increasing the credit for operator action, increased the probability of no vessel breach, and increased the frequency of early, large releases. In summary, the impact of operator actions in controlling AFW was found to be of secondary importance to containment failure and radiological releases.

External cooling for reactor vessel lower head if cavity is flood No credit is taken in the base case for lower vessel head cooling by cavity water. In the sensitivity study, full credit was taken for external vessel cooling. The frequency of no vessel breach was found to increase by 490 3, and the overall conditional probability of no containment failure increased from 55 to 67%. However, the frequency of early releases was reduced only by 1.5 %. The licensee concluded that the external vessel cooling had only a second order impact on "offsite risk".

o Evaluation to assess improvements associated with improved RCP seal materials.

It was shown that the new 0-rings used in Westinghouse pump seals led to small improvements in CDP and release frequency.

Salem IPE Back-End Review 29 ERI/NRC 95-101

0 Hardware and procedure change to allow operator to open PORVs in high pressure scenarios.

Salem EOPs direct the operators to open both PORVs when the core exit thermocouples read 1200°F. Full credit was given for this operator action. Frequency of large, early releases was found to decrease by 50 %. The licensee considers this to be a potential accident management issue.

o

. Procedural Change for ISLOCAs.

A procedural weakness was found to exist in the Salem EOPs for certain ISLOCAs.

Leakage through RHR discharge check valves in the Salem plant is diverted to the PRT inside the containment. When the PRT disk bursts, a small LOCA condition is indicated to the operator, and hence the operator transfers to the appropriate procedures.

However, there are no procedures to indicate to the operator that this scenario is an ISLOCA. A procedural change for operators to transfer to the ISLOCA procedure was considered, and the impact was modelled. The overall CDF was found to reduce by 1 %,

and the early, _large release frequency was found to reduce by 4 %.

An additional sensitivity study was performed for Unit 2 after noting that the release frequency of early, large releases was larger for Unit 2. The principal contributor was the containment isolation frequency. The frequency of containment isolation was reduced by a factor of 10, which lead to a 33 % reduction of the frequency of large, early releases.

2.3.4 Impact on Equipment Behavior

• Equipment survivability in a severe accident environment in the Salem plant is discussed in Section 4.1.4 of the submittal.

The submittal considers the survivability of the following equipment: sprays, fan coolers, Service Water System (SWS) to cool the fan, the Auxiliary Feedwater (AFW) system, the ECCS, and the instrumentation and control systems.

A

. comparison of the equipment environmental test data with the MAAP-calculated containment temperature and pressure profiles, was made in Section 4.1.4. In cases where data is not available, assumptions are made regarding equipment survivability.

Sprays:

The containment spray pumps are located in the auxiliary building,and are not challenged by severe accident conditions. A number of MOVs and transmitters are located in the containment penetration areas. However, these penetration areas were not found to be subject to severe accident conditions, since they were physically separated from the containment. In addition, it was noted that the spray headers located in the upper perimeter of the containment could fail following containment failure, if the failure location was at the dome of the containment.

Salem IPE Back-End Review 30 ERI/NRC 95-101

Fans:*

SWS:

ECCS:

The fan cooler units were examined, by considering the performance of the solenoid valves, switches, fan cooler motors, hydrogen recombiners, and junction boxes. Of the MAAP analyses performed in which the fan coolers were called on to operate, the highest temperature reached was 340°F for a few seconds in one scenario and 334° F for a few minutes in two other scenarios.

After comparing with the equipment qualification ranges of temperature and pressure, it was noted that the severe accident environments cannot threaten the various components that form part of the fan cooler systems.

However, there is a concern that the aerosols released after vessel breach can lead to plugging of the fan cooler filters, and it is conservatively assumed that the fan cooler units fail following vessel breach.

The service water system is used to cool the fan coolers. However, all the pumps and components that comprise the SWS (except for five solenoid valves) are located outside the containment. After reviewing the equipment qualification requirements of these valves, it was concluded that the valves will not fail under severe accident environmental conditions.

The ECCS pumps are located in the auxiliary building. A number of switches, transmitters, solenoid valves, and Limitorque MOVs were identified as a part of the CVCS, RHR and safety injection system, that were found to be located inside the containment. Equipment qualification conditions were reviewed for each component, and it was concluded that all the ECCS components can survive the severe accident environmental conditions. However, no credit was given for ECCS operation after vessel breach in the IPE submittal.

All the components of the auxiliary feedwater system were found to be outside the Salem containment. In addition, a review of instrumentation used to monitor RCS pressure, core exit temperature, pressurizer water level, containment sump water level, containment radiation level,.

containment hydrogen concentration, containment temperature, and secondary side water level, indicated that all of the instrumentation systems can function during a severe accident.

The potential for hydrogen accumulation, deflagration, and detonation, is addressed in Section 4.1.5 of the IPE submittal.

A containment walkdown was performed to identify potential locations for hydrogen accumulation. It was postulated that the areas above the RCP seals and the pressure relief tank blowout disks had the highest potential for accumulation. A region, about eight inches deep, below the ceiling of the refuelling deck and above a plug that is located in the refuelling floor (overhead of the RCPs), was identified as an "inverted box" where hydrogen can accumulate. However, the amount of hydrogen that can accumulate and burn is not expected to cause a significant pressure rise in the containment. In the region overhead of the PRT, the containment is open, and there are metallic gratings in the ceiling above the PRT which permit the hydrogen to escape to the upper containment.

Salem IPE Back-End Review 31 ERI/NRC 95-101

2.4 Reducing the Probability of Core Damage or Fission Product Release 2.4.1 Definition of Vulnerability No definition of "vulnerability" as pertaining to containment analyses can be found in the IPE submittal. However, the licensee alludes to two "vulnerabilities" which can be ameliorated by modifications. The modifications are discussed in the next section.

2.4.2 Plant Modifications The licensee has identified two modifications based on the results of the IPE analyses.

A design change to replace eyewash stations in the relay room and the switchgear room by a limited water supply station.

A procedural change to revise the ISLOCA procedures when leakage from RHR check valves outside the containment direct leakage to the PRT.

Subsequently, in the cover letter that accompanies the IPE submittal, the licensee states that "IPE reviews and field walkdowns have identified existing floor drains previously unaccounted for in the IPE analyses", and that the floor drains have reduced the internal flooding event frequency by a factor of two. Hence, the licensee concludes that the !PE-recommended design change is no longer valid.

PSE&G has also stated that it submitted the above-mentioned procedural modification to the Westinghouse Owners Group (WOG) for appropriate changes in the Emergency Operating Procedures (EPGs). The Westinghouse Owner's Group responded that a check,0f auxiliary building radiation was needed to address the possibility of a LOCA outside the containment, and the Salem EOPs were appropriately modified to require verification of radiation detection in the auxiliary building and the pressure relief tank.

The licensee has identified modifications based primarily on the basis of the release frequency of large, early releases. Operator action to depressurize the RCS is shown to lead to substantial reduction of large, early releases. However, the licensee noted that a majority of high pressure accident sequences have inoperable PORVs due to battery depletion [9]. To realize the benefit calculated by the sensitivity studies, substantial hardware modifications (a new and unique remote power source for the PORVs) would be required. The licensee stated that it is not planning such developmental efforts.

Cavity flooding is shown to increase the probability of intact containment and vessel, but the licensee is not planning any procedural modifications based on this insight. Cavity flooding to levels sufficient to submerge the RPV prior to vessel failure is indicated to be feasible only for small and medium LOCA sequences with recirculation failure. However, the change in overall containment failure probability was found to be negligible, since the contribution of small and Salem IPE Back-End Review 32 ERI/NRC 95-101

medium break LOCA sequences with recirculation failure to the CDP was rather small. Hence, the licensee concluded that plant and -procedural modifications are not warranted.

2.5 Responses to the Recommendations of the CPI Program Generic Letter 88-20, Supplement Numbers 1 and 3 [7-8] identified specific Containment Performance Improvements (CPis) to reduce the vulnerability of containments to severe accident challenges. One of the recommendations of the CPI program pertaining to PWRs with large dry containments was that the utility should evaluate the IPE results for containment and equipment vulnerabilities to hydrogen combustion (local and global), and point out any need for procedural and/or hardware improvements. The submittal does not directly address the recommendations of the CPI program, but has analyzed the vulnerability of the Salem containment to pocketing of hydrogen. The potential for hydrogen accumulation, deflagration and detonation, is addressed in Section 4.1.5 of the IPE submittal. A containment walkdown was performed to identify potential locations for hydrogen accumulation. It was postulated that the areas above the RCP seals and the pressure relief tank blowout disks had the highest potential for accumulation. A region, about eight inches deep, below the ceiling of the refuelling deck and above a plug that is located in the refuelling floor (overhead of the RCPs), was identified as an "inverted box" where hydrogen can accumulate. However, the amount of hydrogen that can accumulate and burn is not expected to cause a significant pressure rise in the containment. In the region overhead of the PRT,. the containment is open, and there are metallic gratings in the ceiling which permit the hydrogen to escape to the upper containment. Local detonations were not addressed in detail.

In summary, no locations could be identified where hydrogen can accumulate in sufficient quantities to burn, and lead to containment or equipment failure.

Salem IPE Back-End Review 33 ERI/NRC 95-101

3.

OVERALL EVALUATION AND CONCLUSIONS The back-end portion of the Salem IPE submittal provides a subs~ntial amount of information in regard to the subject areas identified in Generic Letter 88-20 and NUREG-1335. This submittal uses a large event tree methodology to perform the back-end analyses.

The quantification process is based on plant-specific MAAP calculations, and the information contained in the Zion NUREG-1150 study.

The important points of the submittal-only technical evaluation of the Salem IPE back-end analysis are summarized below:

The Back-End portion of the IPE supplies a substantial amount of information with regards to the subject areas identified in Generic Letter 88-20, and NUREG-1335. For the most part, the separate models used in the Salem IPE Back-End analysis are technically sound. Extensive use is made of the NUREG-1150 studies for the Zion nuclear power plant.

The licensee has addressed all phenomena of importance to severe accident phenomenology in PWRs.

The licensee has also identified an additional mode of containment failure, i.e., liner melt-through due to direct contact with dispersed debris, owing to the design of the cavity and instrument tunnel in the Salem plant.

The submittal has addressed the recommendations of the CPI program (GL 88-20, Supplements 1 and 2).

The analyses have led to the identification of one procedural improvement.

The containment structural analysis for SGS is well executed and generally superior to such analyses that have been developed for other -IPEs. The treatment of temperature distributions and their effects on capacity to resist pressure failure is also well done.

The submittal includes a detailed evaluation of equipment survivability under severe accident conditions.

The conditional probability of early containment failure at the Salem Units 1 and 2 (due to overpressurization and a-mode failure) is 4. 8 % and 5.1 %, respectively, and is approximately one order of magnitude larger than that calculated for Zion.

The conditional probability of late containment failure at Salem Units 1 and 2 (due to overpressurization and combustion of non-condensable gases generated by MCCI) is 37 %

and 44 %, respectively, and is larger than that calculated for Zion (24 % ).

These differences are mainly attributable to the differences in the CDF profiles between the Salem and Zion plants. The CDF of SGS Units 1 and 2 include a larger contribution from high pressure sequences initiated by internal flooding or by loss of service water or ventilation (SWENF). These sequences contribute to 39.1 % of the CDF for SGS Unit Salem IPE Back-End Review 34 ERI/NRC 95-101

1, and 47.3% of the CDF for SGS Unit 2.

These sequences are binned into high pressure PDSs C3B and C3D, and do not have ECCS available. Recovery of ECCS is not credited. These high pressure sequences have a high conditional probability of early containment failure (10 % for flood sequences, and 5 % for the SWENF sequences).

Together, these sequences lead to the calculation of a conditional probability of nearly 5 % for early containment failure at the S3.Iem plant, and a calculation of conditional probabilities of 37% and 44% for late containment failure for Salem Units 1 and 2,

.,respectively.

In conclusion, the Back-End portion of the IPE supplies a substantial amount of information with regards to the subject areas identified in Generic Letter 88-20, and NUREG-1335. Licensee personnel were involved in the back-end analyses, which were performed with the help of outside contractors (PLG, Inc. and ABB Impell Corporation). The event tree developed by the licensee for containment analyses includes all the phenomena of importance to severe accident progression in PWRs.

The licensee's process for the evaluation of containment failure probabilities and failure modes is consistent with the intent of Generic Letter 88-20, Appendix I. The dominant contributors to containment failure are consistent with the insights obtained from the NUREG-1150 analyses for the Zion plant. The licensee has considered the failure of containment isolation system and containment bypass. Failure of electrical and mechanical penetrations at elevated temperatures were considered and ruled out.

The submittal has addressed the recommendations of the CPI program, requested as part of the GL 88-20, Supplements 1 and 2. The IPE methodology is capable of identifying "vulnerabilities", and the analyses have led to the identification of two plant improvements (related to two identified vulnerabilities).

Salem IPE Back-End Review 35 ERI/NRC 95-101

4.

REFERENCES

1.

"Salem Generating Station Individual Plant Examination, " prepared by Public Services Electric and Gas Company, (July 1993).

2.

"Severe Accident Risk: An Assessment of Five U.S. Nuclear Power Plants," NUREG-1150, (1990).

3.

Kenton, M. K., and Gabor, J. R., "Recommended Sensitivity Analyses for an Individual Plant Examination Using MAAP 3.0B," EPRI report (1988).

4.

R. i. Breeding, et al., "Evaluation_ of Severe Accident Risks: Surry Unit 1 ", U. S.

Nuclear Regulatory Commission, NUREG-4551, Vol. 3, Part 1 (October 1990).

5.

C. W. Park, et al., "Evaluation of Severe Accident Risks: Zion Unit 1 ", U. S. Nuclear Regulatory Commission, NUREG-4551, Vol. 7, Part 1 (March, 1993).

6.

F. T. Harper, et al., "Eva! uation of Severe Accident Risks: Quantification of Major Input Parameters," NUREG/CR-4551, Vol. 2, Rev 1, Part 4 (June, 1992).

7.

NRC Letter to All Licensees Holding Operating Licenses and Construction Permits for Nuclear Power Reactor Facilities, "Initiation of the Individual Plant Examination for Severe Accident Vulnerabilities - 10 CFR §50.54(f)," Generic Letter 88-20, Supplement No. 1, dated August 29, 1989.

8.

NRC Letter to All Licensees Holding Operating Licenses and Construction Permits for Nuclear Power Reactor Facilities, "Completion of Containment Performance Improvement Program and Forwarding of Insights for Use in the Individual Plant Examination for Severe Accident Vulnerabilities - Generic Letter No. 88-20-Supplement No. 3 - 10 CFR §50.54(f)," Generic Letter 88-20, Supplement No. 3, dated July 6, 1990.

9.

"PSE&G Response to the NRC Request for Additional Information on the Salem Generating Station Individual Plant Examination (IPE) Submittal," Public Service Electric

& Gas Company (1995).

Salem IPE Back-End Review 36 ERI/NRC 95-101

APPENDIX A IPE EVALUATION AND DATA

SUMMARY

SHEET PWR Back-End Facts Plant Name Salem Generating Station, Units 1 and 2.

Containment Type Large, dry containment.

Unique Containment Features The cavity in the Salem plant is connected to the lower containment (inside the crane wall) by an inclined instrument tunnel. The flow from the room under the seal table (called the incore instrument room) to the annular compartment of the containment is through an opening in the crane wall 3 ft (0.91 m) wide and 8 ft (2.4 m) high. Hence, there is a potential for liner attack by dispersed debris.

Unique Vessel Features None found.

Number of Plant Damage States 18 (condensed to 9 for CET analyses).

Containment Failure Pressure Dry 112 psig (median).

Additional Radionuclide Transport and Retention Structures

_ No credit to auxiliary building structures.

Conditional Probability That The Containment Is Not Isolated 1.5 % for Unit 1.

3.2 % for Unit 2.

Salem IPE Back-End Review 37 ERI/NRC 95-101

hnportant Insights Including Unique Safety Features Five safety grade containment fan coolers.

RHR valves that have larger capacity than other plants of Westinghouse design.

hnplemented Plant improvements One plant modification (See Section 2.4.2 of this review) is under consideration.

C-Matrix SALEM UNIT 1 "Plant Damage Frequency Percentage of Conditional Conditional State" (Per Reactor Total CDF Probability of Late Probability of Early Year)

Containment Failure Containment Failure SGTRIE 3.0 x 10-1 0.5 0.0 1.0 ISOFAL 7.2 x 10-1 1.2 0.0 1.0 VSEQ 5.6 x 10-7 1.0 0.0 0.36 LFWNPO 2.1 x lQ-6 3.6

. 4.5 x 104 0.12 FLOOD 1.5 X 10-5 26.2 0.9 0.1 FSB046 5.8 x 10-6 9.8 0.073 0.079 SWVENF 7.7 x 10-6 13.0 0.95 0.05 SMLLRF 7.1 x 10-6 12.1 0.0 0.0083 ATXLOC 1.6 x lQ-6 2.7 0.0 0.008 LOCAIF 2.2 x 10-6 3.8 0.0 0.008 SSBOR6 7.9 x 10-6 13.4 0.0056 0.0022 SBOPSO 3.6 x 10-6 6.0 2.3 x 10-2 3.3 x 104 SSBORF 3.9 x lQ-6 6.7 0.0 0.0 Total of All 5.9 X *10-5 100 0.37 0.075 Categories Salem IPE Back-End Review 38 ERI/NRC 95-101

---

~-------

11

~

• I SALEM UNIT 2 I

"Plant Damage Frequency Percentage of Conditional Conditional State" (Per Reactor Total CDF Probability of Late Probability of Early Year)

Containment Failure Containment Failure SGTRIE 1.5 x 10-1 0.2 0.0 1.0 ISOFAL 1.9 x 10-6 3.0 0.0 1.0 VSEQ 5.6 x 10-1 0.9 0.0 0.36 LFWNPO 3.1 x lQ-6 5.0 4.5 x 104 0.12 FLOOD 1.5 x 10-s 24.5 0.9 0.1 FSB046 3.0 x 10-6 4.8 0.073 0.079 SWVENF 1.4 x 10-s 22.8 0.95 0.05 SMLLRF 5.2 x 10-6 8.3 0.0 0.0083 ATXLOC 1.5 x 10-6 2.5 0.0 0.008 LOCAIF 4.1 x 10-6 6.6 0.0 0.008 SSBOR6 6.9 x 10-6 11.0 0.0056 0.0022 SBOPSO 3.1 x lQ-6 4.9 2.3 x 10-2 3.3 x 104 SSBORF 3.5 x 10-6 5.5 0.0 0.0 Total of All 5.9 X 10-s 100 0.37 0.075 Categories Salem IPE Back-End Review 39 ERI/NRC 95-101