ML11263A027

E-mail with Attachment from T. Ghosh, RES, to L. Perkins, NRR Et Al, on Fyi Salem SAMA Transmittal

ML11263A027
Person / Time
Site:	Salem, Hope Creek
Issue date:	09/14/2010
From:	Tina Ghosh NRC/RES/DSA/SPB
To:	Leslie Perkins, Bo Pham License Renewal Projects Branch 2
References
FOIA/PA-2011-0113, TAC ME1833, TAC ME1835
Download: ML11263A027 (53)

Text

/'I Perkins, Leslie From: Ghosh, Tina Sent: Tuesday, September 14, 2010 9:56 AM To: Perkins, Leslie; Pham, Bo

Subject:

RE: FYI - Salem SAMA transmittal Attachments: Salem SAMA SE trasmittal memo ML1025107541.docx Please see attachment.

From: Pham, Bo Sent: Tuesday, September 14, 2010 9:55 AM To: Ghosh, Tina; Eccleston, Charles Cc: Perkins, Leslie

Subject:

Re: FYI - Salem SAMA transmittal Thanks, Tina. Please send itto me and Leslie Perkins. She's finishing off this DSEIS.

Thanks.

Sent from NRC blackberry Bo Pham From: Ghosh, Tina To: Eccleston, Charles Cc- Pham, Bo Sent: Tue Sep 14 09:45:33 2010

Subject:

FYI - Salem SAMA transmittal Hi Charles, FYI, I have attached the concurred-on Salem SAMA appendix transmittal memo coming your way soon.

I will send you the joint Chapter 5 input with the Hope Creek SAMA appendix.

Best, Tina 1

MEMORANDUM TO: Bo Pham, Chief Projects Branch 1 Division of License Renewal Office of Nuclear Reactor Regulation FROM: Donnie Harrison, Chief Probabilistic Risk Assessment Licensing Branch Division of Risk Assessment Office of Nuclear Reactor Regulation

SUBJECT:

EVALUATION OF SEVERE ACCIDENT MITIGATION ALTERNATIVES FOR SALEM GENERATING STATION UNITS 1 AND 2 (TAC NOS. ME1833 AND ME1835)

The Probabilistic Risk Assessment Licensing Branch (APLA) has completed the enclosed evaluation of severe accident mitigation alternatives (SAMAs) for Salem Generating Station.

This evaluation is based on the analysis of SAMAs contained in PSEG Nuclear, LLC's license renewal application for SGS, and information provided in responses to requests for additional information (RAls).

In the environmental report, PSEG Nuclear, LLC identified 17 SAMAs to be potentially cost-beneficial. None of these SAMAs relate to managing the effects of aging. Therefore, they need not be implemented as part of license renewal. APLA's review of the analysis found the methods used and the implementation of those methods to be sound.

Enclosure:

As stated CONTACT: Tina Ghosh, NRR/DRA (301) 415-2426

ML102510754 NRR-1 06 OFFICE NRR/DRA/APLA NRR/DRAIAPLA NAME TGhosh DHarrison DATE 09/ /10 09/ /10 Appendix F U.S. Nuclear Regulatory Commission Staff Evaluation of Severe Accident Mitigation Alternatives for Salem Nuclear Generating Station Units 1 and 2 in Support of License Renewal Application Review F.1 Introduction PSEG Nuclear, LLC, (PSEG) submitted an assessment of severe accident mitigation alternatives (SAMAs) for the Salem Nuclear Generating Station (SGS) as part of the environmental report (ER) (PSEG 2009). This assessment was based on the most recent Salem probabilistic risk assessment (PRA) available at that time, a plant-specific offsite consequence analysis performed using the MELCOR Accident Consequence Code System 2 (MACCS2) computer code, and insights from the Salem individual plant examination (IPE)

(PSEG 1993) and individual plant examination of external events (IPEEE) (PSEG 1996). In identifying and evaluating potential SAMAs, PSEG considered SAMAs that addressed the major contributors to core damage frequency (CDF) and release frequency at SGS, as well as SAMA candidates for other operating plants that have submitted license renewal applications. PSEG initially identified 27 potential SAMAs. This list was reduced to 25 unique SAMA candidates by eliminating SAMAs that are not applicable to Salem due to design differences, have already been implemented at SGS, would achieve the same risk reduction results that had already been achieved at SGS by other means, or have excessive implementation cost. PSEG assessed the costs and benefits associated with each of the potential SAMAs and concluded in the ER that several of the candidate SAMAs evaluated are potentially cost-beneficial.

Based on a review of the SAMA assessment, the U.S. Nuclear Regulatory Commission (NRC) staff issued a request for additional information (RAI) to PSEG by letter dated April 12, 2010 (NRC 201 Oa) and, based on a review of the RAI responses, a request for RAI response clarification by teleconference dated July 29, 2010 (NRC 2010b). Key questions concerned:

discussing internal and external review comments on the PRA model, including the impact of the Pressurized Water Reactor (PWR) Owner's Group PRA peer review comments on the SAMA analysis results; clarifying the development bases and assumptions for the Level 2 PRA model; additional details on the quality and implementation status of the SGS fire risk model; the SAMA screening process and additional potential SAMAs not previously considered; and further information on the costs and benefits of several specific candidate SAMAs. PSEG submitted additional information by a letters dated May 24, 2010 (PSEG 201 Oa) and August 18, 2010 (PSEG 2010b). In the responses, PSEG provided: a listing of open gaps and "key findings" from the 2008 PRA peer review and an assessment of their impact on the SAMA analysis; clarification of Level 2 PRA modeling details and assumptions; further details on the SGS fire PRA model; analyses of additional SAMAs; and additional information regarding several specific SAMAs. The licensee's responses addressed the NRC staff's concerns.

F-1

An assessment of SAMAs for SGS is presented below.

F.2 Estimate of Risk for Salem PSEG's estimates of offsite risk at SGS are summarized in Section F.2.1. The summary is followed by the NRC staff's review of PSEG's risk estimates in Section F.2.2.

F.2.1 PSEG's Risk Estimates Two distinct analyses are combined to form the basis for the risk estimates used in the SAMA analysis: (1) the SGS Level 1 and 2 PRA model, which is an updated version of the IPE (PSEG 1993), and (2) a supplemental analysis of offsite consequences and economic impacts (essentially a Level 3 PRA model) developed specifically for the SAMA analysis. The SAMA analysis is based on the most recent SGS Level 1 and Level 2 PRA model available at the time of the ER, referred to as the Salem PRA (Revision 4.1, September 2008 model of record (MOR)). The scope of this Salem PRA does not include external events.

The SGS CDF is approximately 4.8 x 10.5 per year for internal events as determined from quantification of the Level 1 PRA model at a truncation of 1 x 10-11 per year. When determined from the sum of the containment event tree (CET) sequences, or Level 2 PSA model, the release frequency (from all release categories, which consist of intact containment, late release, and early release) is approximately 5.0 x 10-5 per year, also at a truncation of 1 x 10-11 per year.

The latter value was used as the baseline CDF in the SAMA evaluations (PSEG 2009). The CDF is based on the risk assessment for internally initiated events, which includes internal flooding. PSEG did not explicitly include the contribution from external events within the SGS risk estimates; however, it did account for the potential risk reduction benefits associated with external events by multiplying the estimated benefits for internal events by a factor of 2. This is discussed further in Sections F.2.2 and F.6.2.

The breakdown of CDF by initiating event is provided in Table F-1. As shown in this table, events initiated by loss of control area ventilation, loss of offsite power, and loss of service water are the dominant contributors to the CDF. PSEG identified that Station Blackout (SBO) contributes 8 x 10-6 per year, or 17 percent, to the total internal events CDF (PSEG 2010a).

Table F-1. SGS Core Damage Frequency for Internal Events CDF 1  % Contribution Initiating Event (per year) to CDF2 Loss of Control Area Ventilation 1.8 x 10-5 37 Loss of Off-site Power (LOOP) 8.1 x 10-6 17 Loss of Service Water 6.6 x 10-6 14 Internal Floods 4.5 x 10-6 9 Transients 6 4.0 x 10- 8 Steam Generator Tube Rupture (SGTR) 2.7 x 10-6 6 F-2

Loss of Component Cooling Water (CCW) 1.0 x 10-6 2 Anticipated Transient Without Scram (ATWS) 7.4 x 10-7 2 Loss of 125V DC Bus A 6.9 x 10-7 1 Others (less than 1 percent each) 3 1.8 x 10-6 4 Total CDF (internal events) 4.8 x 10-6 100

'Calculated from Fussel-Vesely risk reduction worth (RRW) provided in response to NRC staff RAI 1.e (PSEG 2010).

2 Based on Internal Events CDF contribution and total Internal Events CDF.

3 CDF value derived as the difference between the total Internal Events CDF and the sum of the individual internal events CDFs calculated from RRW.

The Level 2 Salem PRA model that forms the basis for the SAMA evaluation is essentially a complete revision of the original IPE Level 2 model and conforms to current industry guidance.

The Level 2 model utilizes a single CET containing both phenomenological and systemic events. The Level 1 core damage sequences are binned into accident classes which provide the interface between the Level 1 and Level 2 CET analysis. The CET is linked directly to the Level 1 event trees and CET nodes are evaluated using supporting fault trees and logic rules.

The result of the Level 2 PRA is a set of 11 release or source term categories, with their respective frequency and release characteristics. The results of this analysis for SGS are provided in Table E.3-6 of ER Appendix E (PSEG 2009). The categories were defined based on the timing of the release, the initiating event, whether feedwater is available, and the containment failure mode. The frequency of each release category was obtained by summing the frequency of the individual accident progression CET endpoints binned into the release category. Source terms were developed for each of the 11 release categories using the results of Modular Accident Analysis Program (MAAP Version 4.0.6) computer code calculations (PSEG 2010a).

The offsite consequences and economic impact analyses use the MACCS2 code to determine the offsite risk impacts on the surrounding environment and public. Inputs for these analyses include plant-specific and site-specific input values for core radionuclide inventory, source term and release characteristics, site meteorological data, projected population distribution (within a 50-mile radius) for the year 2040, emergency response evacuation modeling, and economic data. The core radionuclide inventory corresponds to the end-of-cycle values for SGS operating at 3632 MWt, which is five percent above the current licensed power level of 3,459 MWt. The magnitude of the onsite impacts (in terms of clean-up and decontamination costs and occupational dose) is based on information provided in NUREG/BR-0184 (NRC 1997a).

In the ER, PSEG estimated the dose to the population within 80-kilometers (50-miles) of the SGS site to be approximately 0.78 person-Sievert (Sv) (78 person-roentgen equivalent man (rem)) per year. The breakdown of the total population dose by containment release mode is summarized in Table F-2. Containment bypass events (such as SGTR-initiated large early release frequency (LERF) accidents) and late containment failures without feedwater dominate the population dose risk at SGS.

F-3

Table F-2. Breakdown of Population Dose by Containment Release Mode Population Dose Percent Containment Release Mode (Person-Rem1 Per Year) Contribution 2 Containment over-pressure (late) 42.9 55 Steam generator rupture 31.9 41 Containment isolation failure 2.3 3 Containment intact 0.2 <1 Interfacing system LOCA 0.6 <1 Catastrophic isolation failure 0.4 <1 Basemat melt-through (late) negligible negligible Total 3 78.2 100 2

'One person-rem = 0.01 person-Sv Derived from Table E.3-7 of the ER 3Column totals may be different due to round off.

F.2.2 Review of PSEG's Risk Estimates PSEG's determination of offsite risk at the SGS is based on the following three major elements of analysis:

• the Level 1 and 2 risk models that form the bases for the 1993 IPE submittal (PSEG 1993), and the external event analyses of the 1996 IPEEE submittal (PSEG 1996),

• the major modifications to the IPE model that have been incorporated in the SGS PRA, including a complete revision of the Level 2 risk model, and

• the MACCS2 analyses performed to translate fission product source terms and release frequencies from the Level 2 PRA model into offsite consequence measures.

Each of these analyses was reviewed to determine the acceptability of the SGS's risk estimates for the SAMA analysis, as summarized below.

The NRC staffs review of the SGS IPE is described in an NRC report dated March 21, 1996 (NRC 1996). Based on a review of the original IPE submittal, responses to RAIs, and a revised IPE submittal, the NRC staff concluded that the IPE submittal met the intent of GL 88-20 (NRC 1988); that is, the licensee's IPE process is capable of identifying the most likely severe accidents and severe accident vulnerabilities. Although no vulnerabilities were identified in the IPE, three improvements to plant and procedures were identified. Two of the improvements were revising SGS procedures related to interfacing systems loss of coolant accidents (ISLOCA) and the third was to install an isolation valve in the demineralized water line to be F-4

used to prevent flooding in the relay and switchgear rooms. All of these improvements are stated to have been implemented (PSEG 2009).

There have been eight revisions to the IPE model since the 1993 IPE submittal. A listing of the major changes made to the SGS PRA since the original IPE submittal was provided in the ER (PSEG 2009) and in response to an RAI (PSEG 2010a) and is summarized in Table F-3. A comparison of the internal events CDF between the 1993 IPE and the current PRA model indicates an increase of about 25 percent in the total CDF (from 6.4 x 10-5 per year to 4.8 x 10-5 per year).

Table F-3. SGS PRA Historical Summary PRA CDF1 Version Summary of Changes from Prior Model2 (per year) 1993 IPE Submittal 6.4 x 10-5 Model 1.0 - Updated plant and common cause data 5.1 x 10.5 8/1996 Model 2.0 - Enhanced the service water system and reactor coolant pump (RCP) seal 5.2 x 10s 8/1998 models

- Added anticipated transients without trip (ATWT) mitigation system actuation circuitry (AMSAC) and valves for containment isolation system

- Eliminated switchgear ventilation as a support system

- Added ISLOCA logic Model 3.0 - Incorporated resolution of 2001 Westinghouse Owner's Group (WOG) PRA 5.2 x 10-5 6/2002 certification comments

- Added switchgear ventilation as a support system

- Addressed HRA dependency issues, updated common-cause calculations, and adjusted initiating event fault tree logic

- Modified how recovery actions were credited Model 3.1 - Revised system models for charging pumps, emergency diesel generator (EDG), 4.1 x 10-5 7/2003 and AMSAC

- Revised models for feedwater line break and steam-line break initiators

- Added human actions to close the service water turbine header isolation valve(s)

Model 3.2 - Enhanced the internal flooding and offsite power recovery models 2.5 x 10-5 3/2005 - Revised models for the switchyard and service water crosstie between units

- Revised common cause failure data

- Adjusted the auxiliary feedwater (AFW) pump failure rate Model 3.2a 3 - Removed recovery from loss of switchgear ventilation and for loss of primary 6.2 x 10.5 3/2006 coolant system (PCS) when the initiator causes loss of PCS

- Removed credit for 1) cross-tying the Unit 2 positive displacement pump (PDP) with Unit 1, 2) cross-tying DC power supplies to power-operated relief valves (PORVs), 3) cross-tying power to diesel fuel oil transfer pumps, and 4) repair of failed EDGs

- Updated the split fraction for a seal LOCA after loss of cooling

- Reduced credit for 1) use of the gas turbine generator in several sequences, 2)

F-5

use of a condensate pump for steam generator makeup, 3) an action to preserve service water availability, and 3) switching from the volume control tank (VCT) to the refueling water storage tank (RWST)

- Removed unavailability of both trains of residual heat removal (RHR)

- Revised operator actions for maintaining AFW suction source

- Changed the loss of DC power initiator

- Revised numerous human error probabilities

- Added new failure mode for component cooling system (CCS)

- Revised modeling of stuck open PORV for SBO and very small LOCA (VSLOCA) sequences

- Revised model to require recovery following loss of CCW and failure to swap charging suction to the RWST

- Changed split fractions in service water logic Model 4.03 - Completely revised and updated the human reliability analysis (HRA) 4.5 x 10.5 3/2008 - Updated failure and common-cause data

- Updated model to better reflect post small LOCA operator actions

- Updated model for loss of control area ventilation (CAV) initiator

- Corrected model to have EDG C fail when EDGs A and B or their associated fuel oil transfer pumps fail

- Updated the service water system and reactor coolant pump (RCP) seal system models

- Reduced credit for use of GTG during grid-related LOOPs

- Updated modeling of DC dependencies Model 4.1 - Completely revised the SGS internal flooding analysis 4.8 x 10.5 9/2008 - Updated model for charging pump upon failure to operate minimum flow valves

- Refined the HRA analyses for SGTR events 1

The IPE, Model 1.0, and Model 2.0 SGS PRAs were performed for both Units 1 and 2; the CDF values shown for these PRA versions are for the SGS unit having the highest internal events and internal flooding CDFs. Starting with Model 3.0, the SGS PRA was performed for Unit 1 only.

2 Summarized from information provided in the ER and a response a NRC staff RAI (PSEG 2010).

3 The internal flooding contribution is not included in the reported CDF.

The CDF values from the 1993 IPE (6.4 x 10-5 per year for Unit 1 and 6.0 x 10-5 per year for Unit

2) are in the middle range of the CDF values reported in the IPEs for Westinghouse four-loop plants, Figure 11.6 of NUREG-1560 shows that the IPE-based total internal events CDF for Westinghouse four-loop plants ranges from 2 x 10.6 per year to 2 x 10-4 per year, with an average CDF for the group of 6 x 10-5 per year (NRC 1997b). It is recognized that other plants have updated the values for CDF subsequent to the IPE submittals to reflect modeling and hardware changes. The current internal events CDF results for SGS (4.8 x 10- per year) are comparable to that for other plants of similar vintage and characteristics.

PSEG explained in the ER that the Salem PRA model is representative of Unit 1, that differences in system configuration and success criteria between Units 1 and 2 are minimal, and that plant-specific data are averaged between the two units. In response to an NRC staff RAI (PSEG 201 Oa), PSEG further clarified that there are currently no differences between Units 1 and 2 that are believed to be important from a risk perspective. The specific design differences F-6

are 1) the recirculation switchover on unit 1 is strictly manual whereas on Unit 2 it is semi-automatic and 2) one component cooling heat exchanger on Unit 1 is of a different design than its counterpart on Unit 2. PSEG also stated that future plant modifications that make the risk profile significantly different between the two units will be addressed by the PRA maintenance and update process. The NRC staff concurs that the design differences between Units 1 and 2 are not likely to impact the results of the SAMA evaluation and that use of Revision 4.1 of the Salem PRA model to represent Unit 2 is reasonable.

The NRC staff considered the peer reviews performed for the SGS PRA, and the potential impact of the review findings on the SAMA evaluation. In the ER (PSEG 2009) and in response to an NRC staff RAI (PSEG 2010a), PSEG described two industry peer reviews of the SGS PRA. The first, conducted by the Westinghouse Owners Group in February 2002, reviewed PRA Model Revision 3.2a. The second, conducted by the PWR Owners Group in November 2008, reviewed PRA Model Revision 4.1.

PSEG stated in the ER that all Level A and B (extremely important and important, respectively) facts and observations (F&Os) from the Westinghouse Owners Group peer review have been addressed (PSEG 2009).

The 2008 peer review of Model Revision 4.1 was performed using the Nuclear Energy Institute peer review process (NEI 2007) and the ASME PRA Standard (ASME 2005) as endorsed by the NRC in Regulatory Guide 1.200, Rev. 1 (NRC 2007). The final report for this peer review had not been completed when the SAMA analysis was performed. In response to an NRC staff RAI, PSEG provided a listing and discussion of eight "key" findings from the 2008 PWR Owners Group peer review (PSEG 2010a). A finding is an observation that is necessary to address to ensure 1) the technical adequacy of the PRA, 2) the capability/robustness of the PRA update process, and 3) the process for evaluating the necessary capability of the PRA technical elements (NEI 2007). Four of the findings were determined to have no impact on the SAMA analysis because it was either a documentation issue (one finding), the current treatment in the PRA model was determined to be conservative (one finding), the finding was determined to be in conflict with other requirements in the PRA standard which were met by the PRA (one finding), or no change to the model was determined to be necessary based on additional analysis (one finding). The other four findings were determined to have a non-significant impact on the SAMA analysis for the following reasons:

• Component availability did not include a contribution from surveillance testing. PSEG explained that component availability is based on Mitigating Systems Performance Index (MSPI) and Maintenance Rule data, which is believed to be accurate, and that any changes in failure rates resulting from a comparison of this data with expected unavailability due to test procedures and maintenance is expected to be non-significant.

" Events that occurred at conditions other than at-power operation or which resulted in controlled shutdown were not considered. PSEG explained that identification of initiating events did include a review of events other than at-power operations and that events occurring during shutdowns and non-power conditions which could have occurred at power were not excluded from the review.

F-7

" The SBO success paths following offsite power recovery do not address recovery and operation of required safety systems. PSEG explained that the likelihood of LOOP, followed by SBO, followed by successful recovery of offsite power, and then followed by multiple equipment failures preventing long-term safe shutdown is very small and that, therefore, the current treatment of SBO is sufficient for the SAMA analysis.

" Omission of failure modes for the EDGs due to the use of only MSPI data and not all plant-specific data. PSEG explained that component availability is based on MSPI and Maintenance Rule data, which is believed to be reliable, and that any changes in failure rates resulting from a validation with other plant-specific data is expected to be non-significant.

In response to another NRC staff RAI, PSEG provided a listing and discussion of the resolution of the 72 supporting requirements (SRs) that did not meet Capability Category II or higher and that remain open in SGS PRA MOR Revision 4.3 (PSEG 2010b). Capability Category II is described as follows (ASME 2005): 1) the scope and level of detail has resolution and specificity sufficient to identify the relative importance of significant contributors at the component level including human actions, as necessary, 2) plant-specific data/models used for significant contributors, and 3) departures from realism will have small impact on the conclusions and risk insights as supported by good practices. PSEG evaluated each of the 72 SRs for impact on the SAMA evaluation and concluded the following:

" PSEG determined that 63 SRs were documentation issues and have no impact on the SAMA analysis.

" Three issues were determined to have no impact on the SAMA analysis because: 1) the finding is principally a documentation issue and the one event cited by the peer reviewer as being mis-classified was determined by PSEG to be appropriately classified (SR IE-A3), 2) PSEG determined that they made appropriate approximations for certain component/failure models where data were lacking (SR SY-A21), and 3) the finding has to do with a conservative modeling issue that does not impact the SAMA analysis (SR IE-C3).

• Six issues were determined to have minimal impact on the SAMA analysis because: 1) the referenced event is bounded by the current PRA model (SR IE-Al), 2) the issue relates to how initiating events are grouped (SRs IE-B3 and AS-A5), 3) the issue impacts only one specific human failure event (HFE) (SR SY-A16), or 4) the un-modeled pre-initiator human errors are viewed as having a low risk contribution (SRs HR-C3 and SY-B16).

PSEG further states that, overall, resolution of the SRs will have a minimal impact on the SAMA evaluation and is well within the uncertainty analysis discussed in Section F.6.2, and that all of the identified SRs that did not meet Capability Category II or higher will be reviewed for consideration during the next periodic update of the PRA model.

F-8

The NRC staff considers PSEG's disposition of the peer review findings to be reasonable and that final resolution of the findings is not likely to impact the results of the SAMA analysis.

PSEG also stated that there have not been any further reviews of the SGS internal events PRA since the 2008 peer review of PRA Model Revision 4.1.

The NRC staff asked PSEG to identify any changes to the plant, including physical and procedural modifications, since Revision 4.1 of the Salem PRA model that could have a significant impact on the results of the SAMA analysis (NRC 2010). In response to the RAI (PSEG 2010a), PSEG explained that one design change and one procedural change have been made since PRA Model Revision 4.1 that have the potential to significantly change the PRA results. The design change was to allow use of two small non-engineered safety feature (ESF) diesel generators to provide power for control and operation of switchyard breakers and to provide a backup source of power to station battery chargers. The procedure change included new procedural steps to provide forced flow of large quantities of outside air to areas supplied by the control area ventilation system. These plant changes resulted in a reduction in the SGS CDF. While the CDF for the updated SGS PRA model, designated as model of record Revision 4.3, was not provided in the RAI response, PSEG did provide the updated SGS release frequency of 2.2 x 10-5 per year, which is more than a 50 percent reduction from the 5.0 x 105 per year used in the SAMA analysis. The impact of this change on the SAMA analysis is discussed in Sections F.3.2 and F.6.2.

In the ER, PSEG explains that, in addition to peer reviews, other measures to ensure, validate, and maintain the quality of the SGS PRA include a formal qualification program for PRA staff, use of procedural guidance to perform PRA tasks, and a program to control PRA models and software. PSEG concludes that based on this quality control process, use of PRA Model Revision 4.1 for the SAMA evaluation was deemed appropriate.

Given that the PSEG internal events PRA model has been peer-reviewed and the peer review findings were judged to have minimal impact on the results of the SAMA analysis, and that PSEG has satisfactorily addressed NRC staff questions regarding the PRA, the NRC staff concludes that the internal events Level 1 PRA model is of sufficient quality to support the SAMA evaluation.

As indicated above, the current SGS PRA does not include external events. In the absence of such an analysis, PSEG used the SGS IPEEE to identify the highest risk accident sequences and the potential means of reducing the risk posed by those sequences, as discussed below and in Section F.3.2.

F-9

The SGS IPEEE was submitted in November 1995 (PSEG 1996), in response to Supplement 4 of Generic Letter 88-20 (NRC 1991a). The submittal included a seismic PRA, a fire PRA, and a screening analysis for other external events. While no fundamental weaknesses or vulnerabilities to severe accident risk in regard to the external events were identified, several potential enhancements were identified as discussed below. In a letter dated May 21, 1999, (NRC 1999) NRC staff concluded that the submittal met the intent of Supplement 4 to Generic Letter 88-20, and that the licensee's IPEEE process is capable of identifying the most likely severe accidents and severe accident vulnerabilities.

The SGS IPEEE seismic analysis utilized a seismic PRA following NRC guidance (NRC 1991a).

The seismic PRA included: a seismic hazard analysis, a seismic fragility assessment, a seismic systems analysis, and quantification of seismic CDF.

The seismic hazard analysis estimated the annual frequency of exceeding different levels of ground motion. Seismic CDFs were determined for both the EPRI (EPRI 1989) and the Laurence Livermore National Laboratory (LLNL) (NRC 1994) hazard assessments. The seismic fragility assessment utilized the walkdown and screening procedures in EPRI's seismic margin assessment methodology (EPRI 1991). Fragility calculations were made for about 100 components and, using a screening criteria of median peak ground acceleration (pga) of 1.5 g which corresponds to a 0.5 pga high confidence low probability of failure (HCLPF) capacity, a total of 27 components remained after screening. The seismic systems analysis defined the potential seismic induced structure and equipment failure scenarios that could occur after a seismic event and lead to core damage. The SGS IPE event tree and fault tree models were used as the starting point for the seismic analysis but an explicit seismic event tree (SET) was used to delineate the potential successes and failures that could occur due to a seismic event.

Quantification of the seismic models consisted of considering the seismic hazard curve with the appropriate structural and equipment seismic fragility curves to obtain the frequency of the seismic damage state. The conditional probability of core damage given each seismic damage state was then obtained from the IPE models with appropriate changes to reflect the seismic damage state. The CDF was then given by the product of the seismic damage state probability and the conditional core damage probability.

F-10

The seismic CDF resulting from the SGS IPEEE was calculated to be 9.5 x 10.6 per year using the LLNL seismic hazard curve and 4.7 x 10-6 per year using the EPRI seismic hazard curve.

Both utilized the IPE internal events PRA, with a CDF of 6.4 x 10-5 per year for quantification of non-seismic failures. While the IPEEE indicated that the EPRI results were believed to be more realistic PSEG assumed a seismic CDF of 9.5 x 10-6 per year based on the LLNL seismic hazard curve in the development of the external events multiplier for purposes of the SAMA evaluation (PSEG 2009). In the ER, PSEG provided a listing and description of the top seven seismic core damage contributors. The dominant seismic core damage contributors for the LLNL seismic hazard curve, representing about 95 percent of the seismic CDF, are listed in Table F-4. The largest contributors to seismic CDF are seismic-induced LOOP caused by failure of the switchyard ceramic insulators combined with random failure of the EDGs and seismic-induced LOOP and failure of battery trains A and B caused by failure of the masonry block walls around the batteries. The NRC staff agrees that the seismic CDF of 9.5 x 10.6 per year is reasonable for the SAMA analysis.

Table F-4. Dominant Contributors to the Seismic CDF

%Contribution Sequence CDF (per to Seismic ID Seismic Sequence Description year) CDF 17 OP: Seismically-Induced LOOP 2.9 x 10-6 31 caused by failure of the switchyard ceramic insulators 33 OP-DAB: Seismically-Induced LOOP 2.0 x 10-6 21 and failure of battery trains A and B 31 OP-SW: Seismically-Induced LOOP 1.3 x 10-6 14 and failure of the service water system 35 OP-IC: Seismically-Induced LOOP and 1.2 x 10-6 13 failure of instrumentation and control capability and equipment in the main control room 34 OP-DAB-DG: Same as 33 OP-DAB 7.7 x 10-7 8 and failure of battery train C 17F OP-FW: Same as 17 OP and failure of 5.4 x 10-7 6 containment fan coolers 21F OP-FW-FC: Same as 17F OP-FW and 2.9 x 10-7 3 failure of auxiliary feed water (AFW)

The SGS IPEEE did not identify any vulnerabilities due to seismic events but did identify three improvements to reduce seismic risk. These improvements are 1) procedural change to ensure long term alternate ventilation for the Auxiliary Building, 2) replacement of identified low ruggedness relays with higher seismic capacity relays, and 3) reinforcement of an 8-foot masonry wall in the 4kV switchgear room. PSEG clarified in response to an NRC staff RAI that the first two improvements have been implemented (PSEG 2010a). The third improvement is F-11

discussed further in Section F.3.2.

The SGS IPEEE fire analysis employed EPRI's fire-induced vulnerability evaluation (FIVE) methodology (EPRI 1993) followed by a PRA quantification of the unscreened compartments.

The fire evaluation was performed on the basis of fire areas which are plant locations completely enclosed by 2-hour rated fire barriers and meeting the FIVE fire barrier criterion related to preventing propagation. Stage 1 consisted of qualitative screening of all plant fire areas to determine whether a fire could cause a plant shutdown or trip, or lead to loss of safe shutdown equipment. Stage 1 also consisted of quantitative screening performed by estimating whether an area's associated fire frequency in combination with the conditional core damage probability given by the loss of functions potentially impacted by the fire was less than the 1 x 10-6 per year. Based on qualitative and quantitative screening all but 38 fire areas were screened out. Stage 2 was to evaluate the remaining fire areas by modeling fire growth and propagation to determine the fire damage state for each fire area. Stage 3 was an evaluation of Sandia Fire Risk Scoping Study issues (NRC 1989) using the tailored walkdown approach provided in the FIVE methodology. Containment performance was also examined to evaluate the performance of containment systems and equipment following core damage resulting from a fire. The final stage was assessment of the functional effects on the plant for each fire damage state by developing explicit fire event trees to probabilistically assess unscreened areas.

Probabilistic credit was given for automatic and manual fire suppression systems. Final quantification utilized FIVE fire data and refined conditional core damage probabilities (CCDPs) from the IPE internal events PRA. The resulting fire induced CDF was calculated to be 2.3 x 10 per year.

In the ER, PSEG provided a listing and description of the top ten fire core damage contributors.

The dominant fire core damage contributors, representing about 99 percent of the fire CDF, are listed in Table F-5. The largest contributors to fire CDF are fires in the 460V Switchgear Rooms, Relay Room, and Control Rooms.

Subsequent to the IPEEE, SGS replaced the CO 2 suppression systems with water sprinkler systems in the 460V Switchgear Rooms, 4160V Switchgears Rooms, and Lower Electrical Penetration Area. In addition, the results of cable wrap tests suggested that the cable wrap would not perform as expected in some areas of the plant and, subsequent to the IPEEE, was removed and replaced. Because of the suppression system changes made to the three areas identified, PSEG did not consider the IPEEE results for these areas valid. PSEG reassessed the fire CDF for these areas using PRA insights from an interim SGS fire model. If the interim SGS fire model showed a higher CDF for any of these three areas, the higher CDF was used for the SAMA analysis. This was the case for the 460V Switchgear Rooms and the Lower Electrical Penetration Area. The fire CDF from the interim SGS fire model for these two fire areas are provided in Table F-5. These insights increased the total fire CDF to 3.8 x 10-5 per year, which was used in the SAMA analysis.

The NRC staff asked PSEG to provide additional information about the interim SGS fire model and, specifically, why it was not used for the SAMA analysis beyond the three areas discussed (NRC 2010a). In response to the RAI, PSEG explained that after the completion of the IPEEE, there was an effort made to develop a fire PRA. This resulted in a partially complete "interim F-12

SGS fire model." However, the interim SGS fire model was never integrated into the internal events PRA model of record (which at the time was Revision 3) and was essentially abandoned because of the forthcoming NUREG/CR-6850 fire PRA development guidance that would render the SGS fire modeling methodology obsolete.

Table F-5. Important Fire Areas and Their Contribution to Fire CDF CDF1 %Contribution Fire Area Description (per year) to Fire CDF 460V Switchgear Rooms 1.3 x 10-i 34 Relay Room 7.2 x 10-6 19 Control Rooms, Peripheral Room, and 7.0 x 10' 18 Ventilation Rooms 4160V Switchgear Room 3.4 x 10- 9 Lower Electrical Penetration Area 3.2 x 10- 8 Upper Electrical and Piping Penetration Areas 1.3 x 10 - 3 Reactor Plant Auxiliary Equipment Area (84B) 1.1 x 10- 3 Turbine and Service Buildings 6.4 x 10-' 2 Service Water Intake 4.2 x 10- 7 1 Reactor Plant Auxiliary Equipment Area (1OOC) 2.9 x 10-7 1

'CDF reported for the 460V Switchgear Rooms and 4160V Switchgear Rooms is from the interim SGS fire model. All other CDFs are from the IPEEE.

The SGS IPEEE did not identify any vulnerabilities due to fire events but did identify two improvements to reduce fire risk. These improvements are 1) procedural change to enhance cooling in the switchgear and control areas in the event of a fire and 2) procedural change for the control of transient combustibles in the turbine building. PSEG clarified in response to an NRC staff RAI that the two suggested improvements have been implemented (PSEG 2010a).

As discussed previously, PSEG identified in the ER that SGS has replaced CO 2 fire suppression systems with water sprinkler systems in three areas of the plant since the IPEEE and that cable wrap has been removed and replaced in several areas of the plant since the IPEEE. The NRC staff asked PSEG if any other fire-related improvements have been made since the IPEEE (NRC 2010a). In response to the RAI, PSEG indicated that the following improvements had been made since the IPEEE: 1) the ventilation system and strategy for maintaining viable working conditions was revised for the 4160 Switchgear Room and the Upper Electrical and Piping Penetration Areas and 2) the maintenance shop was eliminated in the Turbine and Service Buildings in order to reduce the initiating event frequency of fires that would damage the cables for the emergency 4kV buses.

In the ER, PSEG states that an effective comparison between the internal events PRA results F-1 3

and the fire analysis results is not possible because neither the plant response model or the fire modeling methodology used in the IPEEE is up-to-date. PSEG also identified areas where fire CDF quantification may introduce different levels of uncertainty than expected in the internal events PRA and identified a number of conservatisms in the IPEEE fire analysis, including:

• A revised NRC fire events database indicates a trend toward lower frequency and less severe fires than assumed in the SGS IPEEE.

• Bounding fire modeling assumptions are used for many fire scenarios. For example, all equipment in a cabinet is damaged for any fire within a cabinet, regardless of whether it is suppressed. Other examples are provided in the ER.

• Because of a lack of industry experience with regard to crew performance during the types of fires modeled in the fire PRA, the characterization of crew actions in the fire PRA is generally conservative.

PSEG's conclusion is that while there are both conservative and potentially non-conservative factors included in the IPEEE fire model, the IPEEE is judged to have more conservative bias than the internal events model.

Although the arguments regarding the conservatisms in the fire analysis are presented in the ER, PSEG used the modified IPEEE fire CDF of 3.8 x 10. 5 per year in the SAMA analysis rather than some reduced value. Considering the above discussion, the conservatisms in the IPEEE fire analysis as currently understood, and the response to the NRC staff RAIs, the NRC staff concludes that the fire CDF of 3.8 x 10-5 per year is reasonable for the SAMA analysis.

The SGS IPEEE analysis of high winds, floods, and other external (HFO) events followed the progressive screening method defined in NUREG-1407 (NRC 1991b). While SGS is not considered a 1975 Standard Review Plan (SRP) plant, aspects of its licensing basis do conform to the 1975 SRP criteria because SGS is co-located with Hope Creek Generating Station (HCGS), which does meet the 1975 SRP criteria (PSEG 1996). For those events that are based on the location of the site, and not plant-specific features, the 1975 SRP criteria was used for the HFO screening analysis. Progressively more quantitatively based methods were employed for those events that could not be shown to conform to the 1975 SRP criteria. The IPEEE concluded that all HFO events either complied with the 1975 SRP criteria or that their predicted CDF was below the IPEEE screening criteria (i.e. < 1 x 10.6 per year). For the SAMA analysis, PSEG assumed a CDF contribution of 1 X 10,6 per year for each of high winds, external floods, transportation and nearby facilities, detritus, and chemical releases for a total HFO CDF contribution of 5 x 10-6 per year (PSEG 2009).

Although the SGS IPEEE did not identify any vulnerabilities due to HFO events, three improvements to reduce risk were identified. These improvements are 1) modify the circulating water intake structure to protect against detritus (blockage), 2) make improvements to protect against water ingress pathways for external flooding events, and 3) improve the hold downs for hydrogen tanks to protect against tornados. PSEG clarified in response to an NRC staff RAI that the first two suggested improvements have been implemented (PSEG 201 Oa). The third F-14

improvement is discussed further in Section F.3.2.

The NRC staff asked about the status and potential impact on the SAMA analysis of a liquefied natural gas (LNG) terminal planned for Logan Township, New Jersey, upstream on the Delaware River from the SGS site (NRC 2010a). In response to the RAI, PSEG discussed the current status of the LNG terminal as well as the regulatory controls for LNG marine traffic and LNG ship design and the safety record of LNG shipping (PSEG 2010a). The LNG terminal remains in the planning stage and no construction has begun. Further, the state of Delaware has denied applications for several required environmental permits and approvals. PSEG concluded that based on the regulatory process and controls for assuring the safety and security of LNG ships, the safety record of LNG ships, and the uncertainty of the planned terminal, consideration of potential SAMAs associated with the possible future terminal is not warranted. The NRC staff agrees with this conclusion.

Based on the aforementioned results, the external events CDF is approximately equal to the internal events CDF (based on a seismic CDF of 9.5 x 10.6 per year, a fire CDF of 3.8 x 10-5 per year, an HFO CDF of 5.0 x 10.6 per year, and an internal events CDF of 5.0 x 10.' per year used in the SAMA analysis). Accordingly, the NRC staff concurred with SGS's conclusion that the total CDF (from internal and external events) would be approximately 2 times the internal events CDF. In the SAMA analysis submitted in the ER, PSEG doubled the benefit that was derived from the internal events model to account for the combined contribution from internal and external events. The NRC staff agrees with the licensee's overall conclusion concerning the multiplier used to represent the impact of external events and concludes that the licensee's use of a multiplier of 2 to account for external events is reasonable for the purposes of the SAMA evaluation. This is discussed further in Section F.6.2.

The NRC staff reviewed the general process used by PSEG to translate the results of the Level 1 PRA into containment releases, as well as the results of the Level 2 analysis, as described in the ER and in response to NRC staff RAIs (PSEG 2010a). The current Level 2 model is essentially a complete revision of the IPE Level 2 model. In response to an NRC staff RAI, PSEG stated that the IPE Level 2 model was abandoned, with the exception of LERF, with Revision 3 of the SGS PRA model and that the Level 2 model was recreated incorporating current industry guidance as part of the transition from Revision 3 to Revision 4 of the PRA model (PSEG 2010a).

The current SGS Level 2 model utilizes a single CET containing both phenomenological and systemic events. The Level 1 core damage sequences are grouped into core damage accident classes, or plant damage states (PDSs), with similar characteristics. The PDSs are defined based on the following attributes: (1) RCS pressure (high or low), (2) containment isolation status, (3) containment bypass status, (4) containment bypass via an unisolated SGTR, (5) containment bypass via an unisolated, large ISLOCA, (6) containment spray operation mode, (7) containment fan cooler operation, and (8) RWST injection. All of the sequences in an accident class are then input to the CET by linking the level 1 event tree sequences with the level 2 CET. The CET is analyzed by the linking of fault trees that represent each CET node.

Whenever possible the fault trees utilized in the Level 1 analysis are utilized in the CET to propagate dependencies. In response to an NRC staff RAI, PSEG states that the Level 1 and F-1 5

Level 2 models are integrated in that the Level 1 sequences are directly passed to the Level 2 model in the software through the Level 1 sequence fault trees (PSEG 2010a). Twenty-three distinct CET end states or sequences result.

Section E.2.2.3 of the ER describes each of the top events of the CET and states that branch point probabilities for each top event are based on previous SGS Level 2 analyses, recent accident progression research, and similar analyses for other nuclear plants. The NRC staff requested that PSEG describe how the branch point probabilities were developed specifically for top events RCS Depressurization and Containment Heat Removal (NRC 2010a). In response to the RAI, PSEG clarified that top event RCS Depressurization consists of the combination of an existing human action from the HRA and the fault tree for PORV operation (PSEG 2010a). The Containment Heat Removal top event is determined by specific Level 2 system models for containment fan cooler units (CFCUs) and containment spray (CS), either of which can be used for containment heat removal at SGS.

Each CET end state represents a radionuclide release to the environment and is assigned to a release category based on timing of release, the initiating event, whether feedwater is available, and the containment failure mode. Three general release categories are defined: intact containment, late release, and early release. These are further divided into eleven detailed release categories based on the above attributes, as defined in Section E.2.2.6 of the ER.

The frequency of each release category was obtained by summing the frequency of the contributing CET end states. The release characteristics for each release category were developed by using the results of Modular Accident Analysis Program (MAAP Version 4.0.6) computer code calculations (PSEG 201 Oa). Representative MAAP cases for each release category were chosen to either represent the most likely initiators in the release category (intact containment and late release categories) or to conservatively bound the consequences of the release (early release categories). The NRC questioned why PSEG did not also use representative cases that bound the consequences for the late release categories (NRC 201 Oa).

In response to the RAI, PSEG stated that, because the late release categories take more time to evolve than the early release categories, the late release categories are less affected by the initial accident conditions and so result in more uniform consequences than the early release categories (PSEG 2010a). Since the accident sequences assigned to the late release categories yielded similar consequences, PSEG selected representative MAAP cases that represented the most likely initiators within those release categories. The release categories, their frequencies, and release characteristics are presented in Tables E.3-5 and E.3-6 of Appendix E to the ER (PSEG 2009).

The total Level 2 release frequency is of 5.0 x 10-5 per year, which is about 4 percent higher than the internal events CDF of 4.8 x 10-5 per year. The ER states that this difference is due to truncation of low probability sequences and inclusion of non-minimal Level 1 sequences. The NRC staff considers that use of the release frequency rather than the Level 1 CDF will have a negligible impact on the results of the SAMA evaluation because the external event multiplier and uncertainty multiplier used in the SAMA analysis (discussed in Section F.6.2) have a much greater impact on the SAMA evaluation results than the small error arising from the model quantification approach.

F-16

The revised SGS Level 2 PRA model was included in the 2008 PWR Owner's Group peer review discussed above. While none of the eight key findings had to do with the Level 2 analysis, eight LERF analysis SRs did not meet Capability Category II or higher and remain open in SGS PRA MOR Revision 4.3 (PSEG 2010b). PSEG determined that all eight of these findings were documentation issues that did not impact the SAMA analysis.

Based on the NRC staff's review of the Level 2 methodology, that PSEG has adequately addressed NRC staff RAIs, and that the Level 2 model was reviewed in more detail as part of the 2008 PWR Owners Group peer review and there were no findings that impacted the SAMA analysis, the NRC staff concludes that the Level 2 PRA provides an acceptable basis for evaluating the benefits associated with various SAMAs.

The NRC staff reviewed the process used by PSEG to extend the containment performance (Level 2) portion of the PRA to an assessment of offsite consequences (essentially a Level 3 PRA). This included consideration of the source terms used to characterize fission product releases for the applicable containment release categories and the major input assumptions used in the offsite consequence analyses. The MACCS2 code was utilized to estimate offsite consequences. Plant-specific input to the code includes the source terms for each source term category and the reactor core radionuclide inventory (both discussed above), site-specific meteorological data, projected population distribution within an 80-kilometer (50-mile) radius for the year 2040, emergency evacuation modeling, and economic data. This information is provided in Section E.3 of Appendix E to the ER (PSEG 2009).

PSEG used the MACCS2 code and a core inventory from a plant specific calculation at end of cycle to determine the offsite consequences of activity release. In response to an NRC staff RAI, PSEG stated that the MACCS2 analysis was based on the core inventory used in the February 2006 NRC-approved Alternate Source Term for SGS (PSEG 2010a). As indicated in the ER, the reactor core radionuclide inventory used in the consequence analysis was based on a thermal power of 3632 MWt, which is 5 percent higher than the current licensed thermal power of 3459 MWt for SGS. In response to an NRC staff RAI, PSEG stated that the higher thermal power was used to provide margin for a future power uprate (PSEG 2010a).

All releases were modeled as being from the top of the reactor containment building and at low thermal content (ambient). Sensitivity studies were performed on these assumptions and indicated little or no change in population dose or offsite economic cost. Assuming a ground level release decreased dose risk and cost risk by 8 percent and 7 percent, respectively.

Assuming a buoyant plume decreased dose risk and cost risk by 1 percent or less. Based on the information provided, the staff concludes that the release parameters utilized are acceptable for the purposes of the SAMA evaluation.

PSEG used site-specific meteorological data for the 2004 calendar year as input to the MACCS2 code. The development of the meteorological data is discussed in Section E.3.7 of Appendix E to the ER. The data were collected from onsite and local meteorological monitoring systems. Sensitivity analyses using MACCS2 and the meteorological data for the years 2005 through 2007 show that use of data for the year 2004 results in the largest dose and economic cost risk. Missing meteorological data was filled by (in order of preference): using data from the F-1 7

backup met pole instruments (10-meter), using corresponding data from another level of the main met tower, interpolation (if the data gap was less than 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />), or using data from the same hour and a nearby day (substitution technique). The 10-meter wind speed and direction were combined with precipitation and atmospheric stability (derived from the vertical temperature gradient) to create the hourly data file for use by MACCS2. The NRC staff notes that previous SAMA analyses results have shown little sensitivity to year-to-year differences in meteorological data and concludes that the use of the 2004 meteorological data in the SAMA analysis is reasonable.

The population distribution the licensee used as input to the MACCS2 analysis was estimated for the year 2040 using year 1990 and year 2000 census data as accessed by SECPOP2000 (NRC 2003) as a starting point. In response to an NRC staff RAI, PSEG stated that the transient population was included in the 10-mile EPZ, and in the population projection (PSEG 2010a). A ten year population growth rate was estimated using the year 1990 to year 2000 SECPOP2000 data and applied to obtain the distribution in 2040. The baseline population was determined for each of 160 sectors, consisting of sixteen directions for each of ten concentric distance rings to a radius of 50 miles surrounding the site. The SECPOP2000 census data from 1990 and 2000 were used to determine a ten year population growth factor for each of the concentric rings. The population growth was averaged over each ring and applied uniformly to all sectors within each ring. The NRC staff requested PSEG provide an assessment of the impact on the SAMA analysis if a wind-direction weighted population estimate for each sector were used (NRC 2010a). In response to the RAI, PSEG stated that the impacts associated with angular population growth rates on population dose risk and offsite economic cost risk are minimal and bounded by the 30 percent population sensitivity case (PSEG 2010a). This is based on the relatively even wind distribution profile surrounding the site, the tendency for lateral dispersion between sectors, and the use of mean values in the analysis. A sensitivity study was performed for the population growth at year 2040. A 30 percent increase in population resulted in a 30 percent increase in dose risk and a 29 percent increase in cost risk.

In response to an NRC staff RAI, PSEG stated that the radial growth rates used in the MACCS2 analysis provides a more conservative population growth estimate than using 'whole county' data for averaging (PSEG 2010a). PSEG also identified that the population sensitivity case of 30 percent growth was approximately equivalent to adding 6.8 percent to the 10-year growth rate. The NRC staff considers the methods and assumptions for estimating population reasonable and acceptable for purposes of the SAMA evaluation.

The emergency evacuation model was modeled as a single evacuation zone extending out 16 kilometers (10 miles) from the plant (the emergency planning zone - EPZ). PSEG assumed that 95 percent of the population would evacuate. This assumption is conservative relative to the NUREG-1 150 study (NRC 1990), which assumed evacuation of 99.5 percent of the population within the emergency planning zone. The evacuated population was assumed to move at an average radial speed of approximately 2.8 meters per second (6.3 miles per hour) with a delayed start time of 65 minutes after declaration of a general emergency (KLD 2004). A general emergency declaration was assumed to occur at the onset of core damage. The evacuation speed is a time-weighted average value accounting for season, day of week, time of day, and weather conditions. It is noted that the longest evacuation time presented in the study F-1 8

(i.e., full 10 mile EPZ, winter snow conditions, 9 9 th percentile evacuation) is 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> (from the issuance of the advisory to evacuate). Sensitivity studies on these assumptions indicate that there is minor impact to the population dose or offsite economic cost by the assumed variations.

The sensitivity study reduced the evacuation speed by 50 percent to 1.4 m/s. This change resulted in a 4 percent increase in population dose risk and no change in offsite economic cost risk. The NRC staff concludes that the evacuation assumptions and analysis are reasonable and acceptable for the purposes of the SAMA evaluation.

Site specific agriculture and economic parameters were developed manually using data in the 2002 National Census of Agriculture (USDA 2004) and from the Bureau of Economic Analysis (BEA 2008) for each of the 23 counties surrounding SGS, to a distance of 50 miles. Therefore, recently discovered problems in SECPOP2000 do not impact the SGS analysis. The values used for each of the 160 sectors were the data from each of the surrounding counties multiplied by the fraction of that county's area that lies within that sector. Region-wide wealth data (i.e.,

farm wealth and non-farm wealth) were based on county-weighted averages for the region within 50-miles of the site using data in the 2002 National Census of Agriculture (USDA 2004) and the Bureau of Economic Analysis (BEA 2008). Food ingestion was modeled using the new MACCS2 ingestion pathway model COMIDA2 (NRC 1998). For SGS, less than one percent of the total population dose risk is due to food ingestion.

In addition, generic economic data that is applied to the region as a whole were revised from the MACCS2 sample problem input in order to account for cost escalation since 1986, the year that input was first specified. A factor of 1.96, representing cost escalation from 1986 to April 2008 was applied to parameters describing cost of evacuating and relocating people, land decontamination, and property condemnation.

The NRC staff concludes that the methodology used by PSEG to estimate the offsite consequences for SGS provides an acceptable basis from which to proceed with an assessment of risk reduction potential for candidate SAMAs. Accordingly, the NRC staff based its assessment of offsite risk on the CDF and offsite doses reported by PSEG.

F.3 Potential Plant Improvements The process for identifying potential plant improvements, an evaluation of that process, and the improvements evaluated in detail by PSEG are discussed in this section.

F.3.1 Process for Identifying Potential Plant Improvements PSEG's process for identifying potential plant improvements (SAMAs) consisted of the following elements:

" Review of the most significant basic events from the current, plant-specific PRA and insights from the SGS PRA group,

• Review of potential plant improvements identified in, and original results of, the SGS IPE and IPEEE, F-1 9

• Review of SAMA candidates identified for license renewal applications for six other U.S.

nuclear sites, and

• Review of generic SAMA candidates from NEI 05-01 (NEI 2005) to identify SAMAs that might address areas of concern identified in the SGS PRA.

Based on this process, an initial set of 27 candidate SAMAs, referred to as Phase I SAMAs, was identified. In Phase I of the evaluation, PSEG performed a qualitative screening of the initial list of SAMAs and eliminated SAMAs from further consideration using the following criteria:

• The SAMA is not applicable to SGS due to design differences

• The SAMA has already been implemented at SGS,

• The SAMA would achieve results that have already been achieved at SGS by other means, or

• The SAMA has estimated implementation costs that would exceed the dollar value associated with completely eliminating all severe accident risk at SGS.

Based on this screening, two SAMAs were eliminated leaving 25 for further evaluation. The results of the Phase I screening analysis is given in Table E.5-3 of Appendix E to the ER. The remaining SAMAs, referred to as Phase II SAMAs, are listed in Table E.6-1 of Appendix E to the ER. In Phase II, a detailed evaluation was performed for each of the 25 remaining SAMA candidates, as discussed in Sections F.4 and F.6 below. To account for the potential impact of external events, the estimated benefits based on internal events were multiplied by a factor of 2, as previously discussed.

F.3.2 Review of PSEG's Process PSEG's efforts to identify potential SAMAs focused primarily on areas associated with internal initiating events, but also included explicit consideration of potential SAMAs for important fire and seismic initiated core damage sequences. The initial list of SAMAs generally addressed the accident sequences considered to be important to CDF from risk reduction worth (RRW) perspectives at SGS, and included selected SAMAs from prior SAMA analyses for other plants.

PSEG provided a tabular listing of the Level 1 PRA basic events sorted according to their RRW (PSEG 2009). SAMAs impacting these basic events would have the greatest potential for reducing risk. PSEG used a RRW cutoff of 1.01, which corresponds to about a one percent change in CDF given 100-percent reliability of the SAMA. This equates to a benefit of approximately $164,000 (after the benefits have been multiplied by a factor of 2 to account for external events). PSEG also provided and reviewed the Level 2 PRA basic events, down to a RRW of 1.01, for the release categories contributing over 94 percent of the population dose-risk.

F-20

The Level 2 basic events for the remainder of the release categories were not included in the review so as to prevent high frequency-low consequence events from biasing the importance listing. All of the basic events on the Level 1 and 2 importance lists were addressed by one or more of the SAMAs (PSEG 2009). As a result of the review of the Level 1 and Level 2 basic events, 19 SAMAs were identified.

The NRC staff requested PSEG to extend the review of the Level 1 and 2 basic events down to a RRW threshold of 1.003, which equates to a benefit of approximately $50,000, the assumed cost of a procedural change at SGS (NRC 201 Oa). In response to the RAI, PSEG provided revised Level 1 and Level 2 importance lists using SGS PRA model of record Revision 4.3, which was discussed in Section F.2.2, and extended the review of the basic events down to an RRW of 1.006, which equates to a benefit of about $47,000 using PRA Revision 4.3. The review identified the following three additional SAMAs associated with new basic events added to the importance lists (PSEG 2010a):

• SAMA 30 - Automatic Start of Diesel-Powered Air Compressor 0 SAMA 31 - Fully Automate Swapover to Sump Recirculation

• SAMA 32 - Enhance Flood Detection for 100-foot Auxiliary Building and Enhance Procedural Guidance for Responding to Internal Floods A Phase II detailed evaluation was performed for each of these additional SAMAs, which is discussed in Section F.6.2.

The NRC staff asked PSEG to clarify the appropriateness of determining importance factors, and SAMAs, for initiators that are identified as flag events having an assigned probability of 1.0 (NRC 201 Oa). PSEG explained in response to the RAI that fault trees were developed for several loss of support system initiating events (PSEG 201 Oa). Those events that lead to the loss of a support system and are responsible for causing the modeled initiating event were identified as flag events. These events are representative of that initiating event's contribution to CDF and were therefore considered appropriate by PSEG for risk ranking. PSEG further clarified that events whose failure leads to the occurrence of the modeled initiating event will also be listed in the importance list ranking and that the flag probability was therefore set to 1.0 to determine the appropriate CDF contribution of the cutsets. The RRW calculated for these flag events therefore correctly measures the risk significance of the initiating event modeled in this manner.

The NRC staff also asked PSEG to clarify the significance of determining importance factors, and SAMAs, for two split fraction events identified in the importance listing: "RCS-SLOCA-SPLIT" and "MFI-UNAVAILABLE" (NRC 2010a). PSEG explained in response to the RAI that the first event, "RCS-SLOCA-SPLIT," is a flag event that indicates those cutsets in which an RCP seal LOCA has occurred and that the second event, "MFI-UNAVAILABLE," is the conditional probability that the main feedwater system is unavailable given that a reactor trip signal has been generated, irrespective of whether an ATWS condition exists (PSEG 201 Oa).

Because the first event is a flag event, it was assigned a probability of 1.0. SAMA 6, "Enhance F-21

Flood Detection for 84' Auxiliary Building and Enhance Procedural Guidance for Responding to Service Water Flooding," was identified because isolating a service water rupture early could help prevent the conditions that can lead to an RCP seal LOCA. The second event was assigned a conditional probability of 0.3. SAMA 14, "Expand ATWS Mitigation System Actuation Circuitry (AMSAC) Function to Include Backup Breaker Trip on Reactor Protection System (RPS) Failure," was identified to use the AMSAC system to provide a redundant trip signal to help mitigate ATWS events. In over 60 percent of the scenarios in which MFI-UNAVAILABLE is a contributor, AMSAC maintenance is also a contributor. By mitigating ATWS events, SAMA 14 also mitigates scenarios having this combination of events.

PSEG reviewed the cost-beneficial Phase II SAMAs from prior SAMA analyses for five Westinghouse PWR and one General Electric BWR sites. PSEG's review determined that all of the Phase II SAMAs reviewed were either already represented by a SAMA identified from the Level 1 and 2 importance list reviews, are already addressed by other means, have low potential for risk reduction at SGS, or were not applicable to the SGS design. This review resulted in no additional SAMAs being identified.

The NRC staff noted that PSEG's review of these other analyses appeared to have overlooked additional cost-beneficial SAMAs identified during the staff's review of these same SAMA analyses and requested PSEG provide an assessment any additional cost-beneficial SAMAs identified during these reviews for applicability to SGS (NRC 2010a). In response to the RAI, PSEG reviewed the cost-beneficial SAMAs identified in the NRC-issued NUREG-1437 reports for each of the six nuclear sites and concluded the cost-beneficial SAMA either 1) was already identified and evaluated in the ER, 2) was already implemented at SGS, or 3) would not reduce SGS risk (PSEG 2010a). No additional SAMAs were identified from this review.

PSEG considered the potential plant improvements described in the IPE in the identification of plant-specific candidate SAMAs for internal events. Review of the IPE lead to no additional SAMA candidates since the three improvements identified in the IPE have already been implemented at SGS (PSEG 2009).

As a sensitivity case to SAMA 5, PSEG identified and evaluated SAMA 5A, "Install Portable Diesel Generators to Charge Station Battery and Circulating Water Batteries." This SAMA only addresses cases in which RCP seals remain intact, which occurs in a majority of the SBO scenarios. PSEG performed a Phase II evaluation of SAMA 5A, which is in addition to the Phase II evaluations performed for the 25 SAMAs discussed above that were not screened during the Phase I evaluation.

Based on this information, the NRC staff concludes that the set of SAMAs evaluated in the ER, together with those identified in response to NRC staff RAIs, addresses the major contributors to internal event CDF.

Although the IPEEE did not identify any fundamental vulnerabilities or weaknesses related to external events, the ER identified three improvements related to external events (PSEG 2009).

The NRC staff noted that the IPEEE safety evaluation report (NRC 1999) identified five total improvements related to external events and requested PSEG review these improvements for F-22

potentially additional SAMAs (NRC 2010a). In response to the RAI, PSEG reviewed the five suggested improvements and reassessed the three improvements originally evaluated in the ER (PSEG 2010a). As a result of this review, two improvements related to fire events, three improvements related to seismic events, and three improvements related to HFO events were identified. The two suggested fire-related improvements have been implemented, two of the seismic-related improvements have been implemented, and two of the HFO-related improvements have been implemented. The remaining two improvements that have not been implemented are as follows:

" Seismic-related improvement - reinforcement of an 8-foot masonry wall in the 4kV switchgear room. PSEG described the results of an evaluation that determined there was no interaction between the wall and the switchgear bus during a seismic event and subsequent implementation of a corrective action to revise the associated calculation to clarify the lack of interaction. Based on this, PSEG concluded that reinforcement of the masonry wall was not necessary and no SAMA is suggested (PSEG 2010a).

• HFO-related improvement - improve hold downs for the hydrogen tanks to protect against tornados. In response to the RAI, PSEG performed a walk down of the hydrogen racks and determined that the IPEEE suggested improvements to the Unit 2 racks to make the design consistent with the Unit 1 racks was not implemented as indicated in the ER. PSEG further noted that the IPEEE states that these hydrogen tanks "will not have any significant impact on safety structures." Based on this, PSEG concluded that, while the suggested change was prudent, it would not reduce plant risk and no SAMA is suggested.

In the ER PSEG also identified three post IPEEE site changes to determine if they could impact the IPEEE results and possibly lead to a SAMA. From this review, one plant change to replace CO 2 fire suppression with water sprinkler systems was determined to have an impact on fire CDF, which was discussed in Section F.2.2. No additional SAMAs were identified from this review.

In a further effort to identify external event SAMAs, PSEG reviewed the top 10 fire areas contributing to fire CDF based on the results of the IPEEE and interim SGS fire PRA models.

These areas are all of the SGS fire areas having a maximum benefit equal to or greater than approximately $50,000, which is the approximate value of implementing a procedure change at a single unit at SGS. The maximum benefit for a fire area is the dollar value associated with completely eliminating the fire risk in that fire area, which is discussed in Section F.6.2. SAMAs having an implementation cost of less than that of a procedure change, or $50,000, are unlikely.

As a result of this review, PSEG identified five Phase I SAMAs to reduce fire risk. The SAMAs identified included both procedural and hardware alternatives (PSEG 2009). The NRC staff concludes that the opportunity for fire-related SAMAs has been adequately explored and that it is unlikely that there are additional potentially cost-beneficial, fire-related SAMA candidates.

For seismic events, PSEG reviewed the top seven seismic sequences contributing to seismic CDF based on the results of the IPEEE seismic PRA model. These areas are all of the SGS seismic sequences having a benefit equal to or greater than approximately $50,000, which is F-23

the approximate value of implementing a procedure change at a single unit at SGS. The maximum benefit for a seismic sequence is the dollar value associated with completely eliminating the seismic risk for that sequence, which is discussed in Section F.6.2. SAMAs having an implementation cost of less than that of a procedure change, or $50,000, are unlikely.

As a result of this review, PSEG identified three additional Phase I SAMAs to reduce seismic risk (PSEG 2009). The NRC staff concludes that the opportunity for seismic-related SAMAs has been adequately explored and that it is unlikely that there are additional potentially cost-beneficial, seismic-related SAMA candidates.

As stated earlier, other external hazards (high winds, external floods, transportation and nearby facility accidents, release of on-site chemicals, and detritus) are below the IPEEE threshold screening frequency, or met the 1975 SRP design criteria, and are not expected to represent vulnerabilities. Nevertheless, PSEG reviewed the IPEEE results and subsequent plant changes for each of these external hazards and determined that either 1) the maximum benefit from eliminating all associated risk was less than approximately $50,000, which is the approximate value of implementing a procedure change at a single unit at SGS, or 2) only hardware enhancements that would significantly exceed the maximum value of any potential risk reduction were available. As a result of this review, PSEG identified no additional Phase I SAMAs to reduce HFO risk (PSEG 2009). The NRC staff concludes that the licensee's rationale for eliminating other external hazards enhancements from further consideration is reasonable.

The NRC staff noted that, while the generic SAMA list from NEI 05-01 (NEI 2005) was stated to have been used in the identification of SAMAs for SGS, it was not specifically reviewed to identify SAMAs that might be applicable to SGS but rather was used to identify SAMAs that might address areas of concern identified in the SGS PRA (NRC 2010a). The NRC staff asked PSEG to provide further information to justify that this approach produced a comprehensive set of SAMAs for consideration. In response to the RAI, PSEG explained that, based on the early SAMA reviews, both the industry and NRC came to realize that a review of the generic SAMA list was of limited benefit because they were consistently found to not be cost-beneficial and that the real benefit was considered to be in the development of SAMAs generated based on plant specific risk insights from the PRA models (PSEG 2010a).

Furthermore, while the generic list does include potential plant improvements for plants having a similar design to SGS, plant designs are sufficiently different that the specific plant improvements identified in the generic list are generally not directly applicable to SGS, and require alteration to specifically address the SGS design and risk contributors or otherwise would be screened as not applicable to the SGS design. For these reasons, PSEG concludes that the real value of the NEI 05-01 generic SAMA list is as an idea source to generate SAMAs that address important contributors to SGS risk. The NRC staff accepts PSEG's conclusion.

The NRC staff questioned PSEG about potentially lower cost alternatives to some of the SAMAs evaluated (NRC 2010a), including:

0 Operating the AFW AFl 1/21 valves closed.

F-24

• Install improved fire barriers in the 460V switchgear rooms to provide separation, between the three power divisions.

• Install improved fire barriers to provide separation between the AFW pumps.

In response to the RAIs, PSEG addressed the suggested lower cost alternatives and determined that they were either not feasible or were not cost-beneficial (PSEG 201 Oa). This is discussed further in Section F.6.2.

The NRC staff notes that the set of SAMAs submitted is not all-inclusive, since additional, possibly even less expensive, design alternatives can always be postulated. However, the NRC staff concludes that the benefits of any additional modifications are unlikely to exceed the benefits of the modifications evaluated and that the alternative improvements would not likely cost less than the least expensive alternatives evaluated, when the subsidiary costs associated with maintenance, procedures, and training are considered.

The NRC staff concludes that PSEG used a systematic and comprehensive process for identifying potential plant improvements for SGS, and that the set of potential plant improvements identified by PSEG is reasonably comprehensive and, therefore, acceptable.

This search included reviewing insights from the plant-specific risk studies, and reviewing plant improvements considered in previous SAMA analyses. While explicit treatment of external events in the SAMA identification process was limited, it is recognized that the prior implementation of plant modifications for fire and seismic risks and the absence of external event vulnerabilities reasonably justifies examining primarily the internal events risk results for this purpose.

F.4 Risk Reduction Potential of Plant Improvements PSEG evaluated the risk-reduction potential of the 25 remaining SAMAs and one sensitivity case SAMA that were applicable to SGS. The SAMA evaluations were performed using realistic assumptions with some conservatism. On balance, such calculations overestimate the benefit and are conservative.

PSEG used model re-quantification to determine the potential benefits. The CDF, population dose reductions, and offsite economic cost reductions were estimated using the SGS PRA model. The changes made to the model to quantify the impact of SAMAs are detailed in Section E.6 of Appendix E to the ER (PSEG 2009). Table F-6 lists the assumptions considered to estimate the risk reduction for each of the evaluated SAMAs, the estimated risk reduction in terms of percent reduction in CDF and population dose, and the estimated total benefit (present value) of the averted risk. The estimated benefits reported in Table F-6 reflect the combined benefit in both internal and external events. The determination of the benefits for the various SAMAs is further discussed in Section F.6.

F-25

The NRC staff questioned the assumptions used in evaluating the benefit or risk reduction estimate of SAMA 24, "provide procedural guidance to cross-tie Salem 1 and 2 service water systems" (NRC 2010a). The ER assumed this SAMA did not benefit from a reduction in fire risk yet indicates that this SAMA was identified based on a review of the SGS IPEEE fire PRA model results. In response to an NRC staff RAI, PSEG clarified that this SAMA was actually identified from the review of the internal events importance list, that the procedural guidance suggested in this SAMA to perform the inter-unit service water cross-tie is already in place for fire events and that, therefore, implementation of this SAMA would have no additional benefits in fire events (PSEG 2010a). Based on this, PSEG concluded that this SAMA has been appropriately evaluated.

The NRC staff noted that the total of the risk reduction results calculated by summing the individual results for each release category for SAMAs 2, 4, 5A, 18, and 19 was different than the summary results that were used in the SAMA evaluation (NRC 2010a). In response to the RAI, PSEG explained that the release category results provided in the ER for these SAMAs were incorrect, due to typographical errors, and the correct results were provided (PSEG 2010a). PSEG further explained that the SAMA evaluation reported in the ER used the correct release category results and therefore no re-evaluation of the SAMAs was necessary. The NRC staff accepts PSEG's explanation.

For SAMAs that specifically addressed fire events (i.e., SAMA 21, "Seal the Category II and III Cabinets in the Relay Room," SAMA 22, "Install Fire Barriers between the 1CCl, 1CC2, and 1CC3 Consoles in the Control Room Enclosure (CRE)," and SAMA 23, "Install Fire Barriers and Cable Wrap to Maintain Divisional Separation in the 4160V AC Switchgear Room."), the reduction in fire CDF and population dose was not directly calculated (in Table F-5 this is noted as "Not Estimated"). For these SAMAs, an estimate of the impact was made based on general assumptions regarding: the approximate contribution to total risk from external events relative to that from internal events; the fraction of the external event risk attributable to fire events; the fraction of the fire risk affected by the SAMA (based on information from the IPEEE and interim SGS Fire Model results); and the assumption that SAMAs 21 and 22 completely eliminate the fire risk affected by the SAMA and that SAMA 23 eliminates 95 percent of the fire risk affected by the SAMA. Specifically, it is assumed that the contribution to risk from external events is approximately equal to that from internal events, and that internal fires contribute 72 percent of this external events risk. The fire areas impacted by the SAMA are identified and the portion of the total fire risk contributed by each of these fire areas determined. For SAMAs 21 and 22, the benefit or averted cost risk from reducing the fire risk affected by the SAMA is then calculated by multiplying the ratio of the fire risk affected by the SAMA to the internal events CDF by the total present dollar value equivalent associated with completely eliminating severe accidents from internal events at SGS. For SAMA 23, the benefit or averted cost risk from reducing the fire risk affected by the SAMA is then calculated by multiplying the ratio of 95 percent of the fire risk affected by the SAMA to the internal events CDF by the total present dollar value equivalent associated with completely eliminating severe accidents from internal events at SGS. These SAMAs were assumed to have no additional benefits in internal events.

In addition to those SAMAs that only addressed fire events, PSEG evaluated the additional F-26

benefits from reducing fire risk for the following SAMAs that also had internal events benefits:

SAMA 1, "Enhance Procedures and Provide Additional Equipment to Respond to Loss of Control Area Ventilation," SAMA 8, "Install High Pressure Pump Powered with Portable Diesel Generator and Long-term Suction Source to Supply the AFW Header," and SAMA 20, "Fire Protection System to Provide Make-up to RCS and Steam Generators." The benefit or averted cost risk from reducing the fire risk affected by these SAMAs was calculated similar to the method described above with the exception that the fire risk affected by each of these SAMAs were assumed to be reduced based on the same failure probability as was assumed for internal events (i.e., 2.OE-02 for SAMA 1, 1.0E-02 for SAMA 8, and 1.OE-01 for SAMA 20). In other words, SAMA 1 was assumed to eliminate 98 percent, SAMA 8 was assumed to eliminate 99 percent, and SAMA 20 was assumed to eliminate 90 percent of the fire risk affected by these SAMAs. The benefit or averted cost risk from reducing the fire risk affected by SAMA 1 is then calculated by multiplying the ratio of 98 percent of the fire risk affected by the SAMA to the internal events CDF by the total present dollar value equivalent associated with completely eliminating severe accidents from internal events at SGS. The benefit from reducing fire risk was calculated similarly for SAMAs 8 and 20. For SAMAs 1 and 8, PSEG added the calculated benefit from reducing fire risk to the benefit from internal events, which was doubled to account for all external events, to obtain the total benefit from internal and external events. This is discussed further in Section F.6.2.

PSEG also evaluated the additional benefits from reducing seismic risk for the following SAMAs that also had internal events benefits: SAMA 5, "Enhance Procedures and Provide Additional Equipment to Respond to Loss of Control Area Ventilation," SAMA 5A, "Install Portable Diesel Generators to Charge Station Battery and Circulating Water Batteries," SAMA 20, "Fire Protection System to Provide Make-up to RCS and Steam Generators," and SAMA 27, "In addition to the Equipment Installed for SAMA 5, Install Permanently Piped Seismically Qualified Connections to Alternate AFW Water Sources." For these SAMAs, an estimate of the seismic impact was made based on general assumptions regarding: the approximate contribution to total risk from external events relative to that from internal events; the fraction of the external event risk attributable to seismic events; the fraction of the seismic risk affected by the SAMA (based on information from the IPEEE); and the assumption that these SAMAs would reduce the contribution to the seismic CDF from the impacted seismic sequences by 90 percent.

Specifically, it is assumed that the contribution to risk from external events is approximately equal to thatfrom internal events, and that seismic events contribute 18 percent of this external events risk. The seismic sequences impacted by the SAMA are identified and the portion of the total seismic risk contributed by each of these seismic sequences determined. The benefit or averted cost risk from reducing the seismic risk affected by the SAMA is then calculated by multiplying the ratio of 90 percent of the seismic risk affected by the SAMA to the internal events CDF by the total present dollar value equivalent associated with completely eliminating severe accidents from internal events at SGS. For SAMAs 5, 5A, and 27, PSEG added the calculated benefit from reducing seismic risk to the benefit from internal events, which was doubled to account for all external events, to obtain the total benefit from internal and external events. This is discussed further in Section F.6.2.

For SAMA 20, PSEG multiplied the benefit from internal events by a factor of 1.1 to account for F-27

other (non-fire/non-seismic) events and added this to the benefits or averted cost risk from reducing fire risk and seismic risk to obtain the total benefit from internal and external events.

This is discussed further in Section F.6.2.

The NRC staff has reviewed PSEG's bases for calculating the risk reduction for the various plant improvements and concludes, with the above clarifications, that the rationale and assumptions for estimating risk reduction are reasonable and generally conservative (i.e., the estimated risk reduction is higher than what would actually be realized). Accordingly, the NRC staff based its estimates of averted risk for the various SAMAs on PSEG's risk reduction estimates.

F-28

Table F-6. SAMA Cost/Benefit Screening Analysis for SGS(a)

% Risk Reduction Total Benefit ($)

CDF Population Baseline Baseline With Cost ($)

CF Dose (internal + Uncertainty(o)

SAMA Assumptions Dose External) Uncetaintye_

I - Enhance Procedures and Provide Modify fault tree to include a new 34 30 4.8M 12M 475K Additional Equipment to Respond to HEP event, having a failure Loss of Control Area Ventilation probability of 2.OE-02, representing failure of the operator to open doors and align fans. In addition, reduce the fire CDF contribution from fires in Fire Area I FA-EP-1 00G/I F1 -PP-1 OOH assuming the same failure probability.

2 - Re-configure SGS 3 to Provide a SGS 3 (gas turbine) credited for 10 10 1.6M 4.0M 875K More Expedient Backup AC Power weather-related and switchyard Source for SGS 1 and 2 LOOPs.

3 - Install Limited EDG Cross-Tie Modify fault tree to include a new 16 15 2.4M 6.OM 4.2M Capability Between SGS I and 2 basic event, having a failure probability of 5.OE-02, representing failure to cross-tie.

4 - Install Fuel Oil Transfer Pump on Modify fault tree to include a new 16 15 2.4M 6.OM 585K "C" EDG & Provide Procedural basic event, having a failure Guidance for Using "C" EDG to Power probability of 1.OE-02, representing Selected "A" and "B" Loads failure of all three fuel oil transfer pumps. Also modify fault tree to cross-tie Train A, B, and C engineered safety feature (ESF) buses.

F-29

Table F-6. SAMA Cost/Benefit Screening Analysis for SGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

C.DF Popuato (internal + Uncertainty(e)

SAMA Assumptions Dose External) Uncertaintye_

5 - Install Portable Diesel Generators Modify fault tree to include a new 16 11 3.1 M 7.6M 3.3M to Charge Station Battery and basic event, having a failure Circulating Water Batteries and probability of 1.OE-01, representing Replace PDP with Air-Cooled Pump hardware and operator failure of new charging pump. Also, as provided in response to an NRC staff RAI, likelihood of offsite power nonrecovery changed to 1.OE-02 from 2.4E-01 for grid and from 1.OE-01 for site/switchyard-related causes and to 3.OE-02 from 2.4E-01 for weather-related causes.

5A~bl - Install Portable Diesel As provided in response to an NRC 10 10 2.4M 6.OM"d" 770K Generators to Charge Station Battery staff RAI, likelihood of offsite power and Circulating Water Batteries nonrecovery changed to 1.OE-02 from 2.4E-01 for grid and from 1.OE-01 for sitelswitchyard-related causes and to 3.0E-02 from 2.4E-01 for weather-related causes.

6 - Enhance Flood Detection for 84' The failure probabilities of existing 6 1 300K 750K 250K Auxiliary Building and Enhance operator actions to detect and isolate Procedural Guidance for Responding floods successfully were multiplied to Service Water Flooding by a factor of 0.1.

7 - Install "B" Train Auxiliary Modify fault tree to include a new 7 1 410K 1.OM 470K Feedwater Storage Tank (AFWST) basic event, having a failure Makeup Including Alternate Water probability of 1.OE-03, representing Source failure of the alternate water source.

8 - Install High Pressure Pump Modify fault tree to include a new 15 6 1.6M 4.1M 2.5M Powered with Portable Diesel basic event, having a failure F-30

Table F-6. SAMA Cost/Benefit Screening Analysis for SGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Dose (internal + Uncertainty(e)

SAMA Assumptions External)

Generator and Long-term Suction probability of 1.OE-02, representing Source to Supply the AFW Header failure of the new pump. In addition, reduce the fire CDF contribution from fires in Fire Areas 12FA-SB-100/I FA-TGA-88 and I FA-AB-84B assuming the same failure probability.

9 - Connect Hope Creek Cooling Reduce failure probabilities for all 13 11 1.7M 4.3M i.2M Tower Basin to SGS Service Water service water fouling events by a System as Alternate Service Water factor of 10.

Supply 10 - Provide Procedural Guidance for The probability that operators would 1 <1 110K 280K 100K Faster Cooldown Loss of RCP Seal fail to reduce reactor coolant system Cooling (RCS) pressure was reduced to 0.1 from 1.0.

11 - Modify Plant Procedures to Make The probability that operators would 13 12 2.OM 5.OM 100K use of Other Unit's PDP for RCP Seal fail to respond shortllong-term seal Cooling injection demand was reduced to 0.1 from 1.0.

12 - Improve Flood Barriers Outside of Reduce likelihood that the 3 3 550K 1.4M 475K 2201440VAC Switchgear Rooms drains would fail to remove the volume of water assumed in the flooding analysis from 1.OE-01 to 1.OE-03.

13 - Install Primary Side Isolation Valves Reduce likelihood of a SGTR in each 6 30 5.2M 13M 18M on the Steam Generators steam generator from 1.75E-03 to 1.75E-05.

F-31

Table F-6. SAMA CostlBenefit Screening Analysis for SGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Dose (internal + Uncertainty(e)

SAMA Assumptions External) 14 - Expand AMSAC Function to Modify fault tree to AND the current 19 <1 530K 1.3M 485K Include Backup Breaker Trip on event for electrical RPS trip failure Reactor Protection System (RPS) with the top gate for AMSAC.

Failure 15 - Automate RCP Seal Injection Reduce likelihood of failure to isolate 1 <1 42K 69K 210K Realignment upon Loss of Component letdown and realign suction source to Cooling Water (CCW) the refueling water storage tank (RWST) from 1.OE-02 to 1.0E-03.

16 - Install Additional Train of Switchgear Reduce likelihood of operator failure to 1 1 180K 450K 2.5M Room Cooling open doors and establish alternate switchgear room cooling from 5.90E-03 to 5.90E-05.

17- Enhance Procedures and Provide As provided in response to an NRC 3 3 510K 1.3M 200K Additional Equipment to Respond to staff RAI, reduce likelihood of failure Loss of EDG Control Room Ventilation of EDG control room HVAC fans from 4.80E-03 to 4.8E-04 for two fans and 2.3E-06 for the third fan.

18 - Redundant Service Water (SW) Reduce failure probability for the <1 <1 140K 350K 635K Turbine Header Isolation Valve operator action to close the SW turbine header valves from 2.20E-02 to 1.0E-03.

19 - Install Spray Shields on Residual Reduce initiating event frequency for 1 0 34K 84K 350K Heat Removal (RHR) Pumps the 45' elevation Auxiliary Building spray scenario from 7.60E-04 to 7.60E-06.

F-32

Table F-6. SAMA Cost/Benefit Screening Analysis for SGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

SAMA Assumptions Dose (internal + Uncertainty(e)

__________________ Assumptions____ External) ______ ____

20 - Fire Protection System to Provide Modify fault tree to include two new 21 7 5.1M 12.7M 13M Make-up to RCS and Steam Generators basic events, having failure probabilities (SGs) of 1.0E-02 and 1.OE-01, representing failure of the new AFW pump and independently-powered charging pump, respectively. In addition, reduce the fire CDF contribution from fires in Fire Areas 1FA-AB-84A, 1 FA-EP-78C, 1 FA-AB-64A, 1 FA-AB-84B, and 12FA-SB-100/1 FA-TGA-88 assuming the same failure probability of 1.OE-01.

21 - Seal the Category 11and III Cabinets Eliminate the fire CDF contribution from NOT ESTIMATED 870K 2.2M 3.2M in the Relay Room fire damage state 1 RE2.

22 - Install Fire Barriers between the Eliminate the fire CDF contribution from NOT ESTIMATED 330K 830K 1.6M 1 CCl, 1CC2, and 1CC3 Consoles in the Fire Damage State CR16.

CRE 23 - Install Fire Barriers and Cable Wrap Reduce the fire CDF contribution from NOT ESTIMATED 300K 750K 975K to Maintain Divisional Separation in the transient combustible fires in Fire Area 4160V AC Switchgear Room 1 FA-AB-64A, 4160 Switchgear Room, by 95 percent.

24 - Provide Procedural Guidance to Modify fault tree to prevent a 9 4 700K 1.8M 175K Cross-tie SGS I and 2 Service Water complete loss of service water event Systems for events which can affect service water supply to one unit only.

F-33

Table F-6. SAMA Cost/Benefit Screening Analysis for SGS(a)

% Risk Reduction Total Benefit ($)

CDF Population Baseline Baseline With Cost ($)

CF Dose (Internal + Uncertainty(e)

SAMA Assumptions Dose External) 27 - In addition to the Equipment Modify fault tree to include a new 16 11 3.1 M 7.7M 4.2M Installed for SAMA 5, Install basic event, having a failure Permanently Piped Seismically probability of 1.OE-01, representing Qualified Connections to Alternate hardware and operator failure of new AFW Water Sources charging pump. Also, as provided in response to an NRC staff RAI, likelihood of offsite power nonrecovery changed to 1.OE-02 from 2.4E-01 for grid and from I.OE-01 for sitelswitchyard-related causes and to 3.OE-02 from 2.4E-01 for weather-related causes.

30c) - Automatic Start of Diesel-Powered The failure probability for the operator 1 <1 40K 83K 1OOK Air Compressor action to start the diesel-powered air compressor was reduced by a factor of 100 to 6.3E-04 from 6.3E-02.

31(c)- Fully Automate Swapover to Sump The failure probability for the operator 1 <1 27K 56K 100K Recirculation action to swapover to sump recirculation was reduced by a factor of 100 to 5.3E-05 from 5.3E-03.

32(c) - Enhance Flood Detection for 100- The failure probability for the operator 1 <1 50K 1OOK 250K foot Auxiliary Building and Enhance action to isolate the flood source was Procedural Guidance for Responding to reduced by a factor of 100 to 1.0E-03 Internal Floods from 1.OE-01.

F-34

Table F-6. SAMA Cost/Benefit Screening Analysis for SGS(a)

SAMA I Assumptions .... External) .

(a) SAMAs in bold are potentially cost-beneficial.

(b) SAMA 5A added as a sensitivity case to SAMA 5 to provide a comprehensive, long term mitigation strategy for SBO scenarios.

(c) SAMAs 30, 31, and 32 were identified and evaluated in response to an NRC staff RAI (PSEG 201 Oa). The RAI response stated that the percent risk reduction was developed using SGS PRA Model Version 4.3 and that the implementation costs for SAMAs 30 and 31 are expected to be significantly greater than the $1 00K assumed in the SAMA evaluation.

(d) Value estimated by NRC staff using information provided in the ER.

(e) Using a factor of 2.5.

F-35

F.5 Cost Impacts of Candidate Plant Improvements PSEG estimated the costs of implementing the 25 candidate SAMAs through the development of site-specific cost estimates. The cost estimates conservatively did not include the cost of replacement power during extended outages required to implement the modifications (PSEG 2009).

The NRC staff reviewed the bases for the applicant's cost estimates (presented in Table E.5-3 of Attachment E to the ER). For certain improvements, the NRC staff also compared the cost estimates to estimates developed elsewhere for similar improvements, including estimates developed as part of other licensees' analyses of SAMAs for operating reactors.

The ER stated that plant personnel developed SGS-specific costs to implement each of the SAMAs. The NRC staff requested more information on the process PSEG used to develop the SAMA cost estimates (NRC 2010a). PSEG responded to the RAI by explaining that the cost estimates were developed in a series of meetings involving personnel responsible for development of the SAMA analysis and the two PSEG license renewal site leads who are engineering managers each having over 25 years of plant experience, including project management, operations, plant engineering, design engineering, procedure support, simulators, and training (PSEG 2010a). During these meetings, each SAMA was validated against the plant configuration, a budget-level estimate of its implementation cost was developed, and, in some instances, lower cost approaches that would achieve the same objective were developed.

The SAMA implementation costs were then reviewed by the Design Engineering Manager for both technical and cost perspectives and revised accordingly. PSEG further explained that seven general cost categories were used in development of the budget-level cost estimates:

engineering, material, installation, licensing, critical path impact, simulator modification, and procedures and training. For costs that could be shared between the two SGS units, the total estimated cost was evenly divided between the two units to develop a per unit cost. Based on the use of personnel having significant nuclear plant engineering and operating experience, the NRC staff considers the process PSEG used to develop budget-level cost estimates reasonable.

In response to an RAI requesting a more detailed description of the changes associated with SAMAs 3, 5, 8, 13, 20, and 23, PSEG provided additional information detailing the analysis and plant modifications included in the cost estimate of each improvement (PSEG 201 Oa). The staff reviewed the costs and found them to be reasonable, and generally consistent with estimates provided in support of other plants' analyses.

The NRC staff also noted that the ER reported an implementation cost for SAMA 3, "Install Limited EDG Cross-Tie Capability Between SGS 1 and 2," of $4.175M in Section E.6.3 and

$525K in Section E.5-3 and requested clarification on which was the correct value (NRC 2010a). SEG responded that $4.175K was the correct value and stated that this value was used in the SAMA evaluation (PSEG 201 Oa).

F-36

The NRC staff requested PSEG provide justification for the differences in the cost estimates for SAMA 1, "Enhance Procedures and Provide Additional Equipment to Respond to Loss of Control Area Ventilation," having a cost of $475K, and SAMA 17, "Enhance Procedures and Provide Additional Equipment to Respond to Loss of Emergency Diesel Generator (EDG)

Control Room Ventilation," having a cost of $200K, which are similar in that each involves opening doors to provide ventilation and using portable fans to enhance natural circulation (NRC 2010a). In response to the RAI, PSEG stated that SAMA 1 has a higher cost because it is a more complicated modification involving three rooms having differing requirements while SAMA 17 involves four rooms that are basically identical (PSEG 2010a). The NRC staff considers the basis for the differences in cost estimates reasonable.

The NRC staff noted that SAMA 21, "Seal the Category II and III Cabinets in the Relay Room,"

and SAMA 22, "Install Fire Barriers between the 1CC1, 1CC2, and 1CC3 Consoles in the CRE,"

are similar in that each involves installing fire barriers to prevent the propagation of a fire between cabinets and requested an explanation for why the estimated cost of $3.23M for SAMA 21 to modify 48 cabinets is similar to the estimated cost of $1.6M for SAMA 22 to modify just three consoles (NRC 2010a). PSEG responded that the cost per console ($400K) in SAMA 22, is much higher than the cost per cabinet ($35K - $70K) in SAMA 21 because making the modifications to the Control Room consoles is more complicated than making the modifications to the Relay Room cabinets (PSEG 2010a). Specifically, SAMA 22 requires making ventilation modifications due to the significant heat loads in addition to adding fire barrier materials. The NRC staff considers the basis for the differences in cost estimates reasonable.

The NRC asked PSEG to justify the estimated cost of $1 00K for SAMA 10, "Provide Procedural Guidance for Faster Cooldown Loss of RCP Seal Cooling," and SAMA 11, "Modify Plant Procedures to Make use of Other Unit's Positive Displacement Pump (PDP) for RCP Seal Cooling," in light of the statement made in the ER that the minimum expected implementation cost is assumed to be a procedure change at $50K at $100K for the site (NRC 2010a). In response to the RAI, PSEG explained that the cost for SAMA 10 includes 1) $50K to perform a feasibility study to confirm that there is no technical basis preventing implementation of a more rapid cooldown on loss of RCP seal cooling and 2) $150K to revise the emergency operating procedures (EOPs), which are more expensive to revise and require more extensive training than other plant procedures (PSEG 2010a). PSEG also explained that the cost for SAMA 11 includes 1) $50K to perform a feasibility study to confirm that there is no technical basis preventing PDP cross-tie when RCP seal cooling is lost, 2) $50K to revise the plant procedures, and 3) $50K for each unit to involve plant licensing staff. The total of $200K for both SAMAs is divided evenly between the two units. The NRC staff considers the bases for the estimated costs for these SAMAs reasonable.

The NRC staff concludes that the cost estimates provided by PSEG are sufficient and appropriate for use in the SAMA evaluation.

F.6 Cost-Benefit Comparison PSEG's cost-benefit analysis and the NRC staff's review are described in the following sections.

F-37

F.6.1 PSEG's Evaluation The methodology used by PSEG was based primarily on NRC's guidance for performing cost-benefit analysis, i.e., NUREG/BR-01 84, Regulatory Analysis Technical Evaluation Handbook (NRC 1997a). The guidance involves determining the net value for each SAMA according to the following formula:

Net Value = (APE + AOC + AOE + AOSC) - COE, where APE = present value of averted public exposure ($)

AOC = present value of averted offsite property damage costs ($)

AOE = present value of averted occupational exposure costs ($)

AOSC = present value of averted onsite costs ($)

COE = cost of enhancement ($)

If the net value of a SAMA is negative, the cost of implementing the SAMA is larger than the benefit associated with the SAMA and it is not considered cost-beneficial. PSEG's derivation of each of the associated costs is summarized below.

NUREG/BR-0058 has recently been revised to reflect the agency's policy on discount rates.

Revision 4 of NUREG/BR-0058 states that two sets of estimates should be developed, one at 3 percent and one at 7 percent (NRC 2004). PSEG provided a base set of results using the 3 percent discount rate and a sensitivity study using the 7 percent discount rate (PSEG 2009).

Averted Public Exposure (APE) Costs The APE costs were calculated using the following formula:

APE = Annual reduction in public exposure (Aperson-rem/year) x monetary equivalent of unit dose ($2,000 per person-rem) x present value conversion factor (15.04 based on a 20-year period with a 3-percent discount rate)

As stated in NUREG/BR-01 84 (NRC 1997a), it is important to note that the monetary value of the public health risk after discounting does not represent the expected reduction in public health risk due to a single accident. Rather, it is the present value of a stream of potential losses extending over the remaining lifetime (in this case, the renewal period) of the facility.

Thus, it reflects the expected annual loss due to a single accident, the possibility that such an accident could occur at any time over the renewal period, and the effect of discounting these potential future losses to present value. For the purposes of initial screening, which assumes elimination of all severe accidents, PSEG calculated an APE of approximately $2,350,000 for the 20-year license renewal period (PSEG 2009).

Averted Offsite Property Damage Costs (AOC)

The AOCs were calculated using the following formula:

F-38

AOC = Annual CDF reduction x offsite economic costs associated with a severe accident (on a per-event basis) x present value conversion factor.

This term represents the sum of the frequency-weighted offsite economic costs for each release category, as obtained for the Level 3 risk analysis. For the purposes of initial screening, which assumes elimination of all severe accidents caused by internal events, PSEG calculated an AOC of about $306,000 based on the Level 3 risk analysis. This results in a discounted value of approximately $4,600,000 for the 20-year license renewal period.

Averted Occupational Exposure (AOE) Costs The AOE costs were calculated using the following formula:

AOE = Annual CDF reduction

" occupational exposure per core damage event x monetary equivalent of unit dose x present value conversion factor PSEG derived the values for averted occupational exposure from information provided in Section 5.7.3 of the regulatory analysis handbook (NRC 1997a). Best estimate values provided for immediate occupational dose (3,300 person-rem) and long-term occupational dose (20,000 person-rem over a 10-year cleanup period) were used. The present value of these doses was calculated using the equations provided in the handbook in conjunction with a monetary equivalent of unit dose of $2,000 per person-rem, a real discount rate of 3 percent, and a time period of 20 years to represent the license renewal period. For the purposes of initial screening, which assumes elimination of all severe accidents caused by internal events, PSEG calculated an AOE of approximately $31,000 for the 20-year license renewal period (PSEG 2009).

Averted Onsite Costs Averted onsite costs (AOSC) include averted cleanup and decontamination costs and averted power replacement costs. Repair and refurbishment costs are considered for recoverable accidents only and not for severe accidents. PSEG derived the values for AOSC based on information provided in Section 5.7.6 of NUREG/BR-0184, the regulatory analysis handbook (NRC 1997a).

PSEG divided this cost element into two parts - the onsite cleanup and decontamination cost, also commonly referred to as averted cleanup and decontamination costs (ACC), and the replacement power cost (RPC).

ACCs were calculated using the following formula:

ACC = Annual CDF reduction x present value of cleanup costs per core damage event x present value conversion factor F-39

The total cost of cleanup and decontamination subsequent to a severe accident is estimated in NUREG/BR-0184 to be $1.5 x 109 (undiscounted). This value was converted to present costs over a 10-year cleanup period and integrated over the term of the proposed license extension.

For the purposes of initial screening, which assumes elimination of all severe accidents caused by internal events, PSEG calculated an ACC of approximately $965,000 for the 20-year license renewal period.

Long-term RPCs were calculated using the following formula:

RPC = Annual CDF reduction x present value of replacement power for a single event x factor to account for remaining service years for which replacement power is required x reactor power scaling factor PSEG based its calculations on a SGS net output of 1115 megawatt electric (MWe) and scaled up from the 910 MWe reference plant in NUREG/BR-0184 (NRC 1997a). Therefore PSEG applied a power scaling factor of 1115/910 to determine the replacement power costs. For the purposes of initial screening, which assumes elimination of all severe accidents caused by internal events, PSEG calculated an RPC of approximately $335,000 and an AOSC of approximately $1,300,000 for the 20-year license renewal period.

Using the above equations, PSEG estimated the total present dollar value equivalent associated with completely eliminating severe accidents from internal events at SGS to be about $8.28M.

Use of a multiplier of 2 to account for external events increases the value to $16.56M and represents the dollar value associated with completely eliminating all internal and external event severe accident risk for a single unit at SGS, also referred to as the Single Unit Maximum Averted Cost Risk (MACR).

PSEG's Results If the implementation costs for a candidate SAMA exceeded the calculated benefit, the SAMA was considered not to be cost-beneficial. In the baseline analysis contained in the ER (using a 3 percent discount rate and considering the impact of external events), PSEG identified 11 potentially cost-beneficial SAMAs. PSEG performed additional analyses to evaluate the impact of parameter choices (alternative discount rates and variations in MACCS2 input parameters) and uncertainties on the results of the SAMA assessment and, as a result of this analysis, identified five additional potentially cost-beneficial SAMAs. PSEG also performed an analysis on a less costly alternative to SAMA 5 (SAMA 5A) and found it to be potentially cost-beneficial.

The potentially cost-beneficial SAMAs for SGS are the following:

• SAMA 1 - Enhance Procedures and Provide Additional Equipment to Respond to Loss of Control Area Ventilation

• SAMA 2 - Re-configure Salem 3 to Provide a More Expedient Backup AC Power Source for Salem 1 and 2 F-40

• SAMA 3 - Install Limited EDG Cross-tie Capability Between Salem 1 and 2 0 SAMA 4 - Install Fuel Oil Transfer Pump on "C" EDG & Provide Procedural Guidance for Using "C" EDG to Power Selected "A" and "B" Loads

• SAMA 5 - Install Portable Diesel Generators to Charge Station Battery and Circulating Water Batteries & Replace PDP with Air-Cooled Pump

• SAMA 5A - Install Portable Diesel Generators to Charge Station Battery and Circulating Water Batteries

• SAMA 6 - Enhance Flood Detection for 84' Aux Building and Enhance Procedural Guidance for Responding to Service Water Flooding

• SAMA 7 - Install "B" Train AFWST Makeup Including Alternate Water Source

• SAMA 8 - Install High Pressure Pump Powered with Portable Diesel Generator and Long-term Suction Source to Supply the AFW Header

• SAMA 9 - Connect Hope Creek Cooling Tower Basin to Salem Service Water System as Alternate Service Water Supply

" SAMA 10 - Provide Procedural Guidance for Faster Cooldown on Loss of RCP Seal Cooling

• SAMA 11 - Modify Plant Procedures to Make Use of Other Unit's PDP for RCP Seal Cooling

• SAMA 12 - Improve Flood Barriers Outside of 220/440VAC Switchgear Rooms

• SAMA 14 - Expand AMSAC Function to Include Backup Breaker Trip on RPS Failure

• SAMA 17 - Enhance Procedures and Provide Additional Equipment to Respond to Loss of EDG Control Room Ventilation

" SAMA 24 - Provide Procedural Guidance to Cross-tie Salem 1 and 2 Service Water Systems

• SAMA 27 -In Addition to the Equipment Installed for SAMA 5, Install Permanently Piped Seismically Qualified Connections to Alternate AFW Water Sources PSEG indicated that they plan to further evaluate these SAMAs for possible implementation using existing action-tracking and design change processes (PSEG 2009).

F-41

The potentially cost-beneficial SAMAs, and PSEG's plans for further evaluation of these SAMAs, are discussed in detail in Section F.6.2.

F.6.2 Review of PSEG's Cost-Benefit Evaluation The cost-benefit analysis performed by PSEG was based primarily on NUREG/BR-0184 (NRC 1997a) and discount rate guidelines in NUREG/BR-0058 (NRC 2004) and was executed consistent with this guidance.

SAMAs identified primarily on the basis of the internal events analysis could provide benefits in certain external events, in addition to their benefits in internal events. To account for the additional benefits in external events, PSEG multiplied the internal event benefits for all but one internal event SAMA (SAMA 20, discussed further below) by a factor of 2, which is approximately the ratio of the total CDF from internal and external events to the internal event CDF (PSEG 2009). As discussed in Section F.2.2, this factor was based on a seismic CDF of 9.5 x 10.6 per year, plus a fire CDF of 3.8 x 10-5 per year, plus the screening values for high winds, external flooding, transportation, detritus, and chemical release events (1 x 10.6 per year for each). The external event CDF of 5.3 x 10-5 per year is thus about 110 percent of the internal events CDF used in the SAMA analysis (5.0 x 10-5 per year). The total CDF is 2.1 times the internal events CDF and this was rounded to 2. Eleven SAMAs were determined to be cost-beneficial in PSEG's analysis (SAMAs 1, 2, 4, 6, 9, 10, 11, 12, 14, 17, and 24 as described above).

PSEG did not multiply the internal event benefits by the factor of 2 for three SAMAs that specifically address fire risk (SAMAs 21, 22, and 23). Doubling the internal event estimate for SAMAs 21, 22, and 23 would not be appropriate because these SAMAs are specific to fire risks and would not have a corresponding benefit on the risk from internal events.

For all but one internal event SAMA also having benefits in fire and seismic risk (i.e., SAMAs 1, and 8 for fire and SAMAs 5, 5A, and 27 for seismic), PSEG separately quantified the benefits for fire and seismic events and added these results to the benefits from internal events and external events developed from applying the factor of 2 (as discussed in Section F.4 above). The NRC staff noted that this process appeared to be double counting the benefits from external events and requested clarification (NRC 2010a). In response to the RAI, PSEG acknowledged that this process results in "double counting" of some external event contributions to the total averted cost risk and stated that this approach was retained, unless it resulted in a gross misrepresentation of a SAMA's benefit, in order to avoid underestimating the external events averted cost risk (PSEG 201 Oa). PSEG further concluded that this process does not impact the conclusions of the SAMA analysis. Since the process that PSEG used over-estimates the benefits from external events and therefore results in conservative estimates of the SAMA benefits, the NRC staff considers the process PSEG used acceptable for the SAMA evaluation.

For SAMA 20, "Fire Protection System to Provide Make-up to RCS and Steam Generators,"

PSEG multiplied the estimated benefits for internal events by a factor of 2.0 to account for external events in the Phase I analysis. In the Phase II analysis, PSEG separately quantified the internal event, fire event, and seismic event benefits, as described in Section F.4 above, and F-42

to account for the additional benefits in other (non-fire/non-seismic) external events, PSEG multiplied the internal event benefits by a factor of 1.1, which is the ratio of the total CDF from internal and other external events to the internal event CDF (based on an HFO CDF of 5.0 x 10 6 per year and an internal events CDF of 5.0 x 105 per year used in the SAMA analysis). The estimated SAMA benefits for internal events with the factor of 1.1 applied to account for other external events, fire events, and seismic events were then summed to provide an overall benefit. Since the methodology PSEG used accounts for both internal events and external events, the NRC staff considers the methodology PSEG used for SAMA 20 acceptable for the SAMA evaluation.

PSEG considered the impact that possible increases in benefits from analysis uncertainties would have on the results of the SAMA assessment. In the ER, PSEG presents the results of an uncertainty analysis of the internal events CDF which indicates that the 95th percentile value is a factor of 1.64 times the point estimate CDF for SGS. Since the one Phase I SAMA that was screened based on qualitative criteria was screened due to not being applicable to SGS, a re-examination of the Phase I SAMAs based on the upper bound benefits was not necessary.

PSEG considered the impact on the Phase II screening if the estimated benefits were increased by a factor of 1.64 (in addition to the multiplier of 2 for external events). Four additional SAMAs became cost-beneficial in PSEG's analysis (SAMAs 5, 7, 8, and 27 as described above).

PSEG noted that the 9 5 th percentile value for CDF may be underestimated because uncertainty distributions are not applied to all basic events in the SGS PRA model. Based on this, PSEG used a factor of 2.5 times the point estimate CDF to represent the 95th percentile value, which is stated to be typical of most light water reactor CDF uncertainty analyses. PSEG considered the impact on the Phase II screening if the estimated benefits were increased by a factor of 2.5 (in addition to the multiplier of 2 for external events). One additional SAMA became cost-beneficial (SAMA 3). The NRC staff notes that while the factor of 2.5 does not represent an upper bound, it is typical of factors used in prior SAMA analyses, is higher than the factor calculated for other Westinghouse 4-loop plants and used in prior SAMA analysis, and is therefore considered by the NRC staff to be appropriate for use in the SAMA sensitivity analyses.

PSEG provided the results of additional sensitivity analyses in the ER, including use of a 7 percent discount rate and variations in MACCS2 input parameters. These analyses did not identify any additional potentially cost-beneficial SAMAs (PSEG 2009).

The NRC staff noted that the ER reported that the licensed thermal power for SGS Unit 1 is 3,459 MWt, which equates to a net electrical output of 1,195 MWe when operating at 100 percent power, while 1,115 MWe was used to calculate long-term replacement power costs for the SAMA analysis (NRC 2010a). In response to the RAI, PSEG clarified that 1,115 MWe used in the.SAMA analysis was incorrect and provided a revised replacement power cost estimate of

$359,000 using the correct 1,195 MWe, which is an approximately 7 percent increase over that used in the SAMA analysis (PSEG 2010a). PSEG also provided a revised MACR of $16.61M, which is an increase of about 0.3 percent over the MACR used in the SAMA analysis and concluded that the small error would have a negligible impact on the conclusions of the SAMA analysis. The NRC staff agrees with this assessment.

F-43

As indicated in Section F.3.2, in response to an NRC staff RAI, PSEG extended the review of Level 1 and Level 2 basic events down to an RRW of 1.006, which equates to a benefit of about

$47,000, using SGS PRA MOR Revision 4.3 (PSEG 2010a). The review identified the following three additional SAMAs associated with new basic events added to the importance lists: 1)

SAMA 30, "Automatic Start of Diesel-Powered Air Compressor," 2) SAMA 31, "Fully Automate Swapover to Sump Recirculation," and 3) SAMA 32, "Enhance Flood Detection for 100-foot Auxiliary Building and Enhance Procedural Guidance for Responding to Internal Floods." Each of these new SAMAs is included in Table F-6. PSEG performed a Phase II evaluation using results for SGS PRA MOR Revision 4.3 and the process described above. PSEG stated that the release frequency for MOR Revision 4.3 is 2.2 x 10- per year, a decrease of over 50 percent from MOR Revision 4.1, and that the 95th percentile value for CDF is a factor of 2.1 times the point estimate CDF. Based on information provided in the RAI response, the NRC staff estimated, for the MOR Revision 4.3 results, the total present dollar value equivalent associated with completely eliminating severe accidents from internal events at SGS to be about $2.3M, a revised external event multiplier of about 3.4, and a revised MACR of about

$7.9M. These results represent a decrease of more than 50 percent compared to the SGS PRA MOR 4.1 results reported in the ER. PSEG's analysis determined that none of the three SAMA candidates was cost-beneficial in either the baseline analysis or the uncertainty analysis.

Based on these results for MOR Revision 4.3 and the changes in the importance lists, the NRC staff asked PSEG to assess the impact on the SAMA evaluation of the PRA model changes made since MOR Revision 4.1 (NRC 2010b). In response to the RAI, PSEG re-evaluated each potentially cost-beneficial SAMA using MOR Revision 4.3 and determined that SAMA benefits both increased (up to 42 percent) and decreased (up to 99 percent) from the results using MOR Revision 4.1 and that five SAMA candidates (SAMA 3, 5, 11, 14, and 27) would no longer be cost-beneficial (PSEG 2010b). PSEG also qualitatively evaluated each SAMA determined to not be cost-beneficial and concluded that none would become cost-beneficial using MOR Revision 4.3 based on the following:

• The implementation cost is greater than the revised MACR even after accounting for uncertainty (SAMA 13).

• For SAMAs that address fire events only, the maximum averted cost risk for external events decreased, which would result in a corresponding decrease in the maximum calculated benefit for these SAMAs (SAMAs 21, 22, and 23).

• The cost of implementation was sufficiently greater than the MOR Revision 4.1 benefit that changes in MOR Revision 4.3 would not be expected to overcome the difference (SAMAs 15, 16, 18, and 19). The NRC staff notes that this difference, even after accounting for uncertainty, is on the order of 50 percent or more for all of these SAMAs and agrees that it is unlikely that a revised evaluation would result in a change to the cost-beneficial status for these SAMAs.

• The cost of implementation is greater than the revised MACR (SAMA 20). The NRC staff notes that MOR Revision 4.1 results indicate that the fire and seismic events contributors to the MACR are four times the internal events contribution and, since the F-44

maximum averted cost risk for external events has decreased with MOR Revision 4.3, agrees that it is unlikely that a revised evaluation would result in a change to cost-beneficial status for this SAMA.

As indicated in Section F.3.2, the NRC staff asked the licensee to evaluate several potentially lower cost alternatives to the SAMAs considered in the ER (NRC 201 Oa), as summarized below:

Operating the AFW AF1 1/21 valves closed in lieu of SAMA 8, "Install High Pressure Pump Powered with Portable Diesel Generator and Long-term Suction Source to Supply the AFW Header." In response to the RAI, PSEG stated that the AF1 1 valves on the discharge side of the motor-driven AFW pumps are already operated closed, leaving only the AF21 valves on the discharge side of the turbine-driven AFW pump operating open (PSEG 201 Oa). Steam binding of the common suction line to all three AFW pumps could therefore only occur as a result of high temperature water leaks through three check valves in series on the discharge to the turbine-driven AFW pump. PSEG concluded that the proposed improvement would not be feasible because 1) industry data used to represent common-cause steam binding of all three AFW pumps appears to be conservative relative to the SGS configuration, thereby overstating the risk significance of this failure at SGS, 2) operating all of the AF1 1/21 valves closed could actually provide a negative risk benefit based on a new failure event to represent common-cause failure of the valves to open, and 3) operating all of the AF1 1/21 valves closed could have a potentially adverse effect on the SGS design basis because design-basis calculations and assumptions would need to be modified to reflect this change in AFW strategy.

• Install improved fire barriers in the 460V switchgear rooms to provide separation between the three power divisions in lieu of SAMA 20, "Fire Protection System to Provide Make-up to RCS and Steam Generators." In response to the RAI, PSEG explained that the configuration of Fire Area 1 FA-AB-84A, addressed by SAMA 20, is significantly more complex than Fire Area 1 FA-AB-64A, addressed by SAMA 23, "Install Fire Barriers and Cable Wrap to Maintain Divisional Separation in the 4160V AC Switchgear Room" (PSEG 2010a). The SAMA 23 estimated implementation cost of

$975K only addresses the risk associated with preventing the spread of transient fires between divisions and did not address the entire fire risk in the fire area, which would include protecting the overhead cables. PSEG estimates that the cost of addressing the entire fire risk in Fire Area 1 FA-AB-64A would be at least an order of magnitude greater than the estimated implementation cost for SAMA 23. PSEG further estimates that the cost of addressing the fire risk in Fire Area 1FA-AB-84A could be double that for Fire Area 1 FA-AB-64A. The maximum benefit of the proposed SAMA, which assumes elimination of all fire risk for Fire Area 1 FA-AB-84A, is estimated to be $2.OM in the baseline analysis, or $5.1M accounting for uncertainties, using the MOR Rev. 4.1 PRA model. Furthermore, PSEG determined that the maximum benefit would be about 30 percent lower if the MOR Rev. 4.3 PRA model were used. Because the estimated implementation cost is significantly greater than the maximum potential benefit, PSEG concluded that the proposed SAMA would not be cost-beneficial.

F-45

• Install improved fire barriers to provide separation between the AFW pumps in lieu of SAMA 8, "Install High Pressure Pump Powered with Portable Diesel Generator and Long-term Suction Source to Supply the AFW Header." In response to the RAI, PSEG estimated the cost to implement the proposed SAMA to be $750K (PSEG 201 Oa).

Failure of multiple AFW pumps accounted for about 67 percent of the Fire Area 1 FA-AB-84B fire risk. The maximum benefit of the proposed SAMA, which assumes elimination of all of this fire risk, is estimated to be $120K in the baseline analysis, or $290K accounting for uncertainties, using the MOR Rev. 4.1 PRA model. Furthermore, PSEG determined that the maximum benefit would be about 30 percent lower if the MOR Rev.

4.3 PRA model were used. Because the estimated implementation cost is significantly greater than the maximum potential benefit, PSEG concluded that the proposed SAMA would not be cost-beneficial.

PSEG indicated that the 17 potentially cost-beneficial SAMAs (SAMAs 1, 2, 3, 4, 5, 5A, 6, 7, 8, 9, 10, 11, 12, 14, 17, 24, and 27) will be considered for implementation through the established Salem Plant Health Committee (PHC) process (PSEG 2009). In response to an NRC staff RAI, PSEG described the PHC as being chaired by the Plant Manager and includes as members the Plant Engineering Manager and the Directors of Operations, Engineering, Maintenance, and Work Management (PSEG 2010a). The PHC is chartered with reviewing issues that require special plant management attention to ensure effective resolution and, with respect to each of the potentially cost-beneficial SAMAs, will decide on one of the following courses of actions: 1) approve for implementation, 2) conditionally approved for implementation pending the results of requested evaluations, 3) not approved for implementation, or 4) tabled until additional information needed to make a final decision is provided to the PHC. Additional information requested may include 1) making corrections to the original SAMA analysis, 2) examining an alternate solution, 3) performing sensitivity studies to determine the effect of implementing a sub-set of SAMAs, already approved SAMAs, or already approved non-SAMA design changes on the SAMA, or 4) coordinating the SAMA with related Mitigating System Performance Index (MSPI) margin recovery activities. If approved or conditionally approved for implementation, the SAMA will be ranked with respect to priority and assigned target years for implementation.

The NRC staff concludes that, with the exception of the potentially cost-beneficial SAMAs discussed above, the costs of the other SAMAs evaluated would be higher than the associated benefits.

F.7 Conclusions PSEG compiled a list of 27 SAMAs based on a review of: the most significant basic events from the plant-specific PRA and insights from the SGS PRA group, insights from the plant-specific IPE and IPEEE, Phase II SAMAs from license renewal applications for other plants, and the generic SAMA candidates from NEI 05-01. A qualitative screening removed SAMA candidates that: (1) are not applicable to SGS due to design differences, (2) have already been implemented at SGS, (3) would achieve results that have already been achieved at SGS by other means, and (4) have estimated implementation costs that would exceed the dollar value associated with completely eliminating all severe accident risk at SGS. Based on this screening, 2 SAMAs were eliminated leaving 25 candidate SAMAs for evaluation. One F-46

additional SAMA candidate was identified and evaluated as a sensitivity case. Three additional SAMA candidates were identified and evaluated in response to an NRC staff RAI.

For the remaining SAMA candidates, including the sensitivity case SAMA and three SAMAs added in response to the NRC staff RAI, a more detailed design and cost estimate were developed as shown in Table F-6. The cost-benefit analyses in the ER and RAI response showed that 11 of the SAMA candidates were potentially cost-beneficial in the baseline analysis (Phase II SAMAs 1, 2, 4, 6, 9, 10, 11, 12, 14, 17, and 24). PSEG performed additional analyses to evaluate the impact of parameter choices and uncertainties on the results of the SAMA assessment. As a result, five additional SAMA candidates (SAMA 3, 5, 7, 8, and 27) were identified as potentially cost-beneficial in the ER. The ER also showed that the sensitivity case SAMA (SAMA 5A) was potentially cost-beneficial. PSEG has indicated that all 17 potentially cost-beneficial SAMAs will be considered for implementation through the established Salem Plant Health Committee process.

The NRC staff reviewed the PSEG analysis and concludes that the methods used and the implementation of those methods was sound. The treatment of SAMA benefits and costs support the general conclusion that the SAMA evaluations performed by PSEG are reasonable and sufficient for the license renewal submittal. Although the treatment of SAMAs for external events was somewhat limited, the likelihood of there being cost-beneficial enhancements in this area was minimized by improvements that have been realized as a result of the IPEEE process, and inclusion of a multiplier to account for external events.

The NRC staff concurs with PSEG's identification of areas in which risk can be further reduced in a cost-beneficial manner through the implementation of the identified, potentially cost-beneficial SAMAs. Given the potential for cost-beneficial risk reduction, the NRC staff agrees that further evaluation of these SAMAs by PSEG is warranted. However, these SAMAs do not relate to adequately managing the effects of aging during the period of extended operation.

Therefore, they need not be implemented as part of license renewal pursuant to Title 10 of the Code of FederalRegulations, Part 54.

F.8 References American Society of Mechanical Engineers (ASME). 2005. "Addenda to ASME RA-S-2002, Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications." ASME RA-Sb-2005, December 2005.

BEA (Bureau of Economic Analysis). 2008. Regional Economic Accounts, accessed June 20 at http://www.bea.qov/re-qional/reis/.

Electric Power Research Institute (EPRI). 1989. "Probabilistic Seismic Hazard Evaluations at Nuclear Plant Sites in the Central and Eastern United States; Resolution of the Charleston Earthquake Issues." EPRI NP-6395-D, EPRI Project P101-53. Palo Alto, CA. April 1989.

F-47

Electric Power Research Institute (EPRI). 1991. "A Methodology for Assessment of Nuclear Power Plant Seismic Margin," Implementation Guide NP-6041, Revision 1. Palo Alto, CA.

August 1991.

Electric Power Research Institute (EPRI). 1993. "Fire Induced Vulnerability Evaluation (FIVE)

Methodology." TR-100370, Revision 1, Palo Alto, CA. September 19, 1993.

KLD Associates, Inc. (KLD). 2004. Salem /Hope Creek Nuclear GeneratingStations Development of Evacuation Time Estimates. KLD TR-356. February 2004.

Nuclear Energy Institute (NEI). 2005. "Severe Accident Mitigation Alternative (SAMA) Analysis Guidance Document." NEI 05-01 (Rev. A), Washington, D.C. November 2005.

Nuclear Energy Institute (NEI). 2007. "Process for Performing Follow-on PRA Peer Reviews using the ASME PRA Standard (Internal Events)." NEI 05-04, Rev. 1, Washington, D.C.

December 2007.

Public Service Electric and Gas Company (PSEG). 1993. Letter from Stanley LaBruna, PSEG, to NRC Document Control Desk.

Subject:

"Generic Letter 88-20; Individual Plant Examination (IPE) Report, Salem Generating Station, Unit Nos. 1 and 2, Docket Nos. 50-272 and 50-311,"

Hancocks Bridge, New Jersey. July 30, 1993. Accessible at ML080100047.

Public Service Electric and Gas Company (PSEG). 1995. Letter from E. Simpson, PSEG, to NRC Document Control Desk.

Subject:

"Response to Generic Letter 88-20 Individual Plant Examination for Sever Accident Vulnerabilities - 1 OCFR50.54 (f) Request for Additional Information Salem Generating Station, Unit Nos. 1 and 2 Facility Operating License Nos. DRR-70 and DPR-75 Docket Nos. 50-272 and 50-311," Hancocks Bridge, New Jersey. August 01, 1995. Accessible at ML080100021.

Public Service Electric and Gas Company (PSEG). 1996. Letter from E. C. Simpson, PSEG, to NRC Document Control Desk.

Subject:

"Response to Generic Letter No. 88-20, Supplement 4, Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilities, Salem Generating Station Units Nos. 1 and 2, Facility Operating License Nos. DPR-70 and DPR-75, Docket Nos. 50-272 and 50-311," Hancocks Bridge, New Jersey. January 29, 1996.

Accessible at ML080100023.

PSEG Nuclear, LLC (PSEG). 2009. Salem Nuclear Generating Station --- License Renewal Application, Appendix E: Applicant's EnvironmentalReport; OperatingLicense Renewal Stage.

Hancocks Bridge, New Jersey. August 18, 2009. Accessible at ML092400532.

PSEG Nuclear, LLC (PSEG). 2010a. Letter from Paul. J. Davison, PSEG, to NRC Document Control Desk.

Subject:

"Response to NRC Request for Additional Information dated April 12, 2010, related to the Severe Accident Mitigation Alternatives (SAMA) review of the Salem Nuclear Generating Station, Units 1 and2," Hancocks Bridge, New Jersey. May 24, 2010.

Accessible at ML101520326.

F-48

PSEG Nuclear, LLC (PSEG). 2010b. Letter from Christine T. Neely, PSEG, to NRC Document Control Desk.

Subject:

"Supplement to RAI responses submitted in PSEG Letter LR-N10-0164 dated May 24, 2010, related to the Severe Accident Mitigation Alternatives (SAMA) review of the Salem Nuclear Generating Station, Units 1 and 2," Hancocks Bridge, New Jersey. August 18, 2010. Accessible at ML102320211.

U.S. Department of Agriculture (USDA). 2004. 2002 Census of Agriculture - Volume 1, Geographic Area Series, Part 8 (Delaware), Part 20 (Maryland), Part 30.

U.S. Nuclear Regulatory Commission (NRC). 1988. Generic Letter 88-20, "Individual Plant Examination for Severe Accident Vulnerabilities." November 23, 1988. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1989. Fire Risk Scoping Study. NUREG/CR-5088. January 1989. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1990. Severe Accident Risks: An Assessment for Five U.S. Nuclear Power Plants. NUREG-1150. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1991 a. Generic Letter No. 88-20, Supplement 4, "Individual Plant Examination of External Events for Severe Accident Vulnerabilities." NUREG-1407. Washington, D.C. June 28,1991.

U.S. Nuclear Regulatory Commission (NRC). 1991b. "Procedural and Submittal Guidance for the Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilities." NUREG-1407. Washington, D.C. June 1991.

U.S. Nuclear Regulatory Commission (NRC). 1994. Revised Livermore Seismic Hazard Estimates for Sixty-Nine Nuclear Plant Sites East of the Rocky Mountains. NUREG-1488, April 1994. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1996. Letter from Leonard Olshan, U.S. NRC, to Leon Eliason, PSEG.

Subject:

Individual Plant Examination (IPE) Submittal - Internal Events, Salem Nuclear Generating Station, Units 1 and 2 (TAC No. M74461). March 21, 1996.

U.S. Nuclear Regulatory Commission (NRC). 1997a. Regulatory Analysis Technical Evaluation Handbook. NUREG/BR-0184. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1997b. Individual Plant Examination Program:

Perspectives on Reactor Safety and Plant Performance. NUREG-1560. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1998. Code Manual for MACCS2: User's-Guide.

NUREG/CR-6613, Volume 1, May 1998. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 1999. Letter from Patrick D. Milano, U.S. NRC to Harold W. Keiser, PSEG.

Subject:

Generic Letter 88-20, Supplement 4, "Individual Plant Examination for External Events for Severe Accident Vulnerabilities," Salem Nuclear Generating Station, Unit Nos. 1 and 2 (TAC Nos. M83669 and M83670). May 21, 1999.

F-49

U.S. Nuclear Regulatory Commission (NRC). 2003. SECPOP2000:Sector Population,Land Fraction,and Economic Estimation Program. NUREG/CR-6525, Rev. 1. Sandia National Laboratories. August 2003.

U.S. Nuclear Regulatory Commission (NRC). 2004. Regulatory Analysis Guidelines of the U.S.

Nuclear Regulatory Commission. NUREG/BR-0058, Rev. 4. Washington, D.C.

U.S. Nuclear Regulatory Commission (NRC). 2007. "An Approach for Determining the Technical Adequacy of Probabilistic Risk Assessment Results for Risk-Informed Activities."

Regulatory Guide 1.200, Revision 1. January 2007.

U.S. Nuclear Regulatory Commission (NRC). 2010a. Letter from Charles Eccleston, U.S. NRC, to Thomas Joyce, PSEG.

Subject:

Request for Additional Information, Regarding Severe Accident Mitigation Alternatives for the Salem Nuclear Generating Station, Units 1 and 2. April 12, 2010. Accessible at ML100910252.

U.S. Nuclear Regulatory Commission (NRC). 2010b. Teleconference held on July 29, 2010 between NRC staff and PSEG Nuclear, LLC staff regarding clarifications to the RAI responses provided by PSEG via letter dated April 12, 2010. Accessible at ML102220012.

F-50