ML11263A033

E-mail with Attachment from B. Pham, NRR, to B. Harris, NRR, on Appendices for Salem & HCGS SAMA

ML11263A033
Person / Time
Site:	Salem, Hope Creek
Issue date:	09/29/2010
From:	Bo Pham NRC/NRR/DLR/RARB
To:	Harris B License Renewal Projects Branch 1
References
FOIA/PA-2011-0113
Download: ML11263A033 (104)

Text

Pham, BQ From:

Sent:

To:

Cc:

Subject:

Hi Brian, Pham, Bo Wednesday, September 29, 2010 8:29 AM Harris, Brian Perkins, Leslie Appendices for Salem & HCGS SAMA Attached are the Thanks.

Appendix F V 1.docx Appendices that go with Chapter 5 of the DSEIS. Leslie asked me to fwd them to you.

Appendix G.V. 1.docx Bo Pham Chief, Projects Branch 1 Division of License Renewal Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission 301-415-8450 25

/

1 Appendix F 2

U.S. Nuclear Regulatory Commission Staff Evaluation of 3

Severe Accident Mitigation Alternatives for 4

Salem Nuclear Generating Station Units 1 and 2 5

In Support of License Renewal Application Review 6

1 This Page Intentionally Left Blank

Appendix F 1

F. U.S. Nuclear Regulatory Commission Staff Evaluation of Severe 2

Accident Mitigation Alternatives for Salem Nuclear Generating Station 3

Units land 2 in Support of License Renewal Application Review 4

F.1 Introduction 5

PSEG Nuclear, LLC, (PSEG) submitted an assessment of severe accident mitigation 6

alternatives (SAMAs) for the Salem Nuclear Generating Station (SGS) as part of the 7

environmental report (ER) (PSEG 2009). This assessment was based on the most recent 8

Salem probabilistic risk assessment (PRA) available at that time, a plant-specific offsite 9

consequence analysis performed using the MELCOR Accident Consequence Code System 2 10 (MACCS2) computer code, and insights from the Salem individual plant examination (IPE) 11 (PSEG 1993) and individual plant examination of external events (IPEEE) (PSEG 1996). In 12 identifying and evaluating potential SAMAs, PSEG considered SAMAs that addressed the major 13 contributors to core damage frequency (CDF) and release frequency at SGS, as well as SAMA 14 candidates for other operating plants that have submitted license renewal applications. PSEG 15 initially identified 27 potential SAMAs. This list was reduced to 25 unique SAMA candidates by 16 eliminating SAMAs that are not applicable to Salem due to design differences, have already 17 been implemented at SGS, would achieve the same risk reduction results that had already been 18 achieved at SGS by other means, or have excessive implementation cost. PSEG assessed the 19 costs and benefits associated with each of the potential SAMAs and concluded in the ER that 20 several of the candidate SAMAs evaluated are potentially cost-beneficial.

21 Based on a review of the SAMA assessment, the U.S. Nuclear Regulatory Commission (NRC) 22 staff issued a request for additional information (RAI) to PSEG by letter dated April 12, 2010 23 (NRC 2010a) and, based on a review of the RAI responses, a request for RAI response 24 clarification by teleconference dated July 29, 2010 (NRC 2010b). Key questions concerned:

25 discussing internal and external review comments on the PRA model, including the impact of 26 the Pressurized Water Reactor (PWR) Owner's Group PRA peer review comments on the 27 SAMA analysis results; clarifying the development bases and assumptions for the Level 2 PRA 28 model; additional details on the quality and implementation status of the SGS fire risk model; the 29 SAMA screening process and additional potential SAMAs not previously considered; and further 30 information on the costs and benefits of several specific candidate SAMAs. PSEG submitted 31 additional information by a letters dated May 24, 2010 (PSEG 2010a) and August 18, 2010 32 (PSEG 2010b). In the responses, PSEG provided: a listing of open gaps and "key findings" 33 from the 2008 PRA peer review and an assessment of their impact on the SAMA analysis; 34 clarification of Level 2 PRA modeling details and assumptions; further details on the SGS fire 35 PRA model; analyses of additional SAMAs; and additional information regarding several specific 36 SAMAs. The licensee's responses addressed the NRC staff's concerns.

September 2010 F-1 Draft NUREG-1437, Supplement 45

Appendix F 1

An assessment of SAMAs for SGS is presented below.

2 F.2 Estimate of Risk for Salem 3

PSEG's estimates of offsite risk at SGS are summarized in Section F.2.1. The summary is 4

followed by the NRC staff's review of PSEG's risk estimates in Section F.2.2.

5 F.2.1 PSEG's Risk Estimates 6

Two distinct analyses are combined to form the basis for the risk estimates used in the SAMA 7

analysis: (1) the SGS Level 1 and 2 PRA model, which is an updated version of the IPE (PSEG 8

1993), and (2) a supplemental analysis of offsite consequences and economic impacts 9

(essentially a Level 3 PRA model) developed specifically for the SAMA analysis. The SAMA 10 analysis is based on the most recent SGS Level 1 and Level 2 PRA model available at the time 11 of the ER, referred to as the Salem PRA (Revision 4.1, September 2008 model of record 12 (MOR)). The scope of this Salem PRA does not include external events.

13 The SGS CDF is approximately 4.8 x 10-5 per year for internal events as determined from 14 quantification of the Level 1 PRA model at a truncation of 1 x 10-11 per year. When determined 15 from the sum of the containment event tree (CET) sequences, or Level 2 PSA model, the 16 release frequency (from all release categories, which consist of intact containment, late release, 17 and early release) is approximately 5.0 x 10-5 per year, also at a truncation of 1 x 10.11 per year.

18 The latter value was used as the baseline CDF in the SAMA evaluations (PSEG 2009). The 19 CDF is based on the risk assessment for internally initiated events, which includes internal 20 flooding. PSEG did not explicitly include the contribution from external events within the SGS 21 risk estimates; however, it did account for the potential risk reduction benefits associated with 22 external events by multiplying the estimated benefits for internal events by a factor of 2. This is 23 discussed further in Sections F.2.2 and F.6.2.

24 The breakdown of CDF by initiating event is provided in Table F-1. As shown in this table, 25 events initiated by loss of control area ventilation, loss of offsite power, and loss of service water 26 are the dominant contributors to the CDF. PSEG identified that Station Blackout (SBO) 27 contributes 8 x 10-6 per year, or 17 percent, to the total internal events CDF (PSEG 201 Oa).

28 Table F-I. SGS Core Damage Frequency for Internal Events CDF1

% 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

September 2010 F-2 Draft NUREG-1437, Supplement 45

Appendix F Transients 4.0 x 10- 6 8

Steam Generator Tube Rupture (SGTR) 2.7 x 10-6 6

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 10-5 100

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

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

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

1 The Level 2 Salem PRA model that forms the basis for the SAMA evaluation is essentially a 2

complete revision of the original IPE Level 2 model and conforms to current industry guidance.

3 The Level 2 model utilizes a single CET containing both phenomenological and systemic 4

events. The Level 1 core damage sequences are binned into accident classes which provide 5

the interface between the Level 1 and Level 2 CET analysis. The CET is linked directly to the 6

Level 1 event trees and CET nodes are evaluated using supporting fault trees and logic rules.

7 The result of the Level 2 PRA is a set of 11 release or source term categories, with their 8

respective frequency and release characteristics. The results of this analysis for SGS are 9

provided in Table E.3-6 of ER Appendix E (PSEG 2009). The categories were defined based 10 on the timing of the release, the initiating event, whether feedwater is available, and the 11 containment failure mode. The frequency of each release category was obtained by summing 12 the frequency of the individual accident progression CET endpoints binned into the release 13 category. Source terms were developed for each of the 11 release categories using the results 14 of Modular Accident Analysis Program (MAAP Version 4.0.6) computer code calculations 15 (PSEG 2010a).

16 The offsite consequences and economic impact analyses use the MACCS2 code to determine 17 the offsite risk impacts on the surrounding environment and public. Inputs for these analyses 18 include plant-specific and site-specific input values for core radionuclide inventory, source term 19 and release characteristics, site meteorological data, projected population distribution (within a 20 50-mile radius) for the year 2040, emergency response evacuation modeling, and economic 21 data. The core radionuclide inventory corresponds to the end-of-cycle values for SGS operating 22 at 3632 MWt, which is five percent above the current licensed power level of 3,459 MWt. The 23 magnitude of the onsite impacts (in terms of clean-up and decontamination costs and 24 occupational dose) is based on information provided in NUREG/BR-0184 (NRC 1997a).

September 2010 F-3 Draft NUREG-1437, Supplement 45

Appendix F 1

2 3

4 5

6 7

8 9

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.

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 Total3 78.2 100

'one person-rem = 0.01 person-Sv 2Derived from Table E.3-7 of the ER 3Column totals may be different due to round off.

10 11 F.2.2 Review of PSEG's Risk Estimates 12 PSEG's determination of offsite risk at the SGS is based on the following three major elements 13 of analysis:

14 0

the Level 1 and 2 risk models that form the bases for the 1993 IPE submittal (PSEG 15 1993), and the external event analyses of the 1996 IPEEE submittal (PSEG 1996),

16 0

the major modifications to the IPE model that have been incorporated in the SGS PRA, 17 including a complete revision of the Level 2 risk model, and 18 e the MACCS2 analyses performed to translate fission product source terms and release 19 frequencies from the Level 2 PRA model into offsite consequence measures.

September 2010 F-4 Draft NUREG-1437, Supplement 45

Appendix F 1

2 3

4 5

6 7

8 9

10 11 12 13 14 15 16 17 18 19 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 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.'

8/1996 Model 2.0 Enhanced the service water system and reactor coolant pump (RCP) seal 5.2 x 10-5 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 September 2010 F-5 Draft NUREG-1437, Supplement 45

Appendix F 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 105 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) 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 10s 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

'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.

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

September 2010 F-6 Draft NUREG-1437, Supplement 45

Appendix F 3lThe internal flooding contribution is not included in the reported CDF.

1 2

The CDF values from the 1993 IPE (6.4 x 10-5 per year for Unit 1 and 6.0 x 10s per year for Unit 3

2) are in the middle range of the CDF values reported in the IPEs for Westinghouse four-loop 4

plants. Figure 11.6 of NUREG-1560 shows that the IPE-based total internal events CDF for 5

Westinghouse four-loop plants ranges from 2 x 10-6 per year to 2 x 10- per year, with an 6

average CDF for the group of 6 x 10.5 per year (NRC 1997b). It is recognized that other plants 7

have updated the values for CDF subsequent to the IPE submittals to reflect modeling and 8

hardware changes. The current internal events CDF results for SGS (4.8 x 10-5 per year) are 9

comparable to that for other plants of similar vintage and characteristics.

10 PSEG explained in the ER that the Salem PRA model is representative of Unit 1, that 11 differences in system configuration and success criteria between Units 1 and 2 are minimal, and 12 that plant-specific data are averaged between the two units. In response to an NRC staff RAI 13 (PSEG 201 Oa), PSEG further clarified that there are currently no differences between Units 1 14 and 2 that are believed to be important from a risk perspective. The specific design differences 15 are 1) the recirculation switchover on unit 1 is strictly manual whereas on Unit 2 it is semi-16 automatic and 2) one component cooling heat exchanger on Unit 1 is of a different design than 17 its counterpart on Unit 2. PSEG also stated that future plant modifications that make the risk 18 profile significantly different between the two units will be addressed by the PRA maintenance 19 and update process. The NRC staff concurs that the design differences between Units 1 and 2 20 are not likely to impact the results of the SAMA evaluation and that use of Revision 4.1 of the 21 Salem PRA model to represent Unit 2 is reasonable.

22 The NRC staff considered the peer reviews performed for the SGS PRA, and the potential 23 impact of the review findings on the SAMA evaluation. In the ER (PSEG 2009) and in response 24 to an NRC staff RAI (PSEG 201 Oa), PSEG described two industry peer reviews of the SGS 25 PRA. The first, conducted by the Westinghouse Owners Group in February 2002, reviewed 26 PRA Model Revision 3.2a. The second, conducted by the PWR Owners Group in November 27 2008, reviewed PRA Model Revision 4.1.

28 PSEG stated in the ER that all Level A and B (extremely important and important, respectively) 29 facts and observations (F&Os) from the Westinghouse Owners Group peer review have been 30 addressed (PSEG 2009).

31 The 2008 peer review of Model Revision 4.1 was performed using the Nuclear Energy Institute 32 peer review process (NEI 2007) and the ASME PRA Standard (ASME 2005) as endorsed by the 33 NRC in Regulatory Guide 1.200, Rev. 1 (NRC 2007). The final report for this peer review had 34 not been completed when the SAMA analysis was performed. In response to an NRC staff RAI, 35 PSEG provided a listing and discussion of eight "key" findings from the 2008 PWR Owners 36 Group peer review (PSEG 201 Oa). A finding is an observation that is necessary to address to 37 ensure 1) the technical adequacy of the PRA, 2) the capability/robustness of the PRA update 38 process, and 3) the process for evaluating the necessary capability of the PRA technical September 2010 F-7 Draft NUREG-1 437, Supplement 45

Appendix F 1

elements (NEI 2007). Four of the findings were determined to have no impact on the SAMA 2

analysis because it was either a documentation issue (one finding), the current treatment in the 3

PRA model was determined to be conservative (one finding), the finding was determined to be 4

in conflict with other requirements in the PRA standard which were met by the PRA (one 5

finding), or no change to the model was determined to be necessary based on additional 6

analysis (one finding). The other four findings were determined to have a non-significant impact 7

on the SAMA analysis for the following reasons:

8 Component availability did not include a contribution from surveillance testing. PSEG 9

explained that component availability is based on Mitigating Systems Performance 10 Index (MSPI) and Maintenance Rule data, which is believed to be accurate, and that 11 any changes in failure rates resulting from a comparison of this data with expected 12 unavailability due to test procedures and maintenance is expected to be non-significant.

13 Events that occurred at conditions other than at-power operation or which resulted in 14 controlled shutdown were not considered. PSEG explained that identification of 15 initiating events did include a review of events other than at-power operations and that 16 events occurring during shutdowns and non-power conditions which could have 17 occurred at power were not excluded from the review.

18 The SBO success paths following offsite power recovery do not address recovery and 19 operation of required safety systems. PSEG explained that the likelihood of LOOP, 20 followed by SBO, followed by successful recovery of offsite power, and then followed by 21 multiple equipment failures preventing long-term safe shutdown is very small and that, 22 therefore, the current treatment of SBO is sufficient for the SAMA analysis.

23 Omission of failure modes for the EDGs due to the use of only MSPI data and not all 24 plant-specific data. PSEG explained that component availability is based on MSPI and 25 Maintenance Rule data, which is believed to be reliable, and that any changes in failure 26 rates resulting from a validation with other plant-specific data is expected to be non-27 significant.

28 In response to another NRC staff RAI, PSEG provided a listing and discussion of the resolution 29 of the 72 supporting requirements (SRs) that did not meet Capability Category II or higher and 30 that remain open in SGS PRA MOR Revision 4.3 (PSEG 2010b). Capability Category II is 31 described as follows (ASME 2005): 1) the scope and level of detail has resolution and 32 specificity sufficient to identify the relative importance of significant contributors at the 33 component level including human actions, as necessary, 2) plant-specific data/models used for 34 significant contributors, and 3) departures from realism will have small impact on the 35 conclusions and risk insights as supported by good practices. PSEG evaluated each of the 72 36 SRs for impact on the SAMA evaluation and concluded the following:

September 2010 F-8 Draft NUREG-1437, Supplement 45

Appendix F 1

° PSEG determined that 63 SRs were documentation issues and have no impact on the 2

SAMA analysis.

3 0 Three issues were determined to have no impact on the SAMA analysis because: 1) the 4

finding is principally a documentation issue and the one event cited by the peer reviewer 5

as being mis-classified was determined by PSEG to be appropriately classified (SR IE-6 A3), 2) PSEG determined that they made appropriate approximations for certain 7

component/failure models where data were lacking (SR SY-A21), and 3) the finding has 8

to do with a conservative modeling issue that does not impact the SAMA analysis (SR 9

IE-C3).

10 Six issues were determined to have minimal impact on the SAMA analysis because: 1) 11 the referenced event is bounded by the current PRA model (SR IE-A1), 2) the issue 12 relates to how initiating events are grouped (SRs IE-B3 and AS-A5), 3) the issue impacts 13 only one specific human failure event (HFE) (SR SY-A16), or 4) the un-modeled pre-14 initiator human errors are viewed as having a low risk contribution (SRs HR-C3 and SY-15 B16).

16 PSEG further states that, overall, resolution of the SRs will have a minimal impact on the SAMA 17 evaluation and is well within the uncertainty analysis discussed in Section F.6.2, and that all of 18 the identified SRs that did not meet Capability Category II or higher will be reviewed for 19 consideration during the next periodic update of the PRA model.

20 The NRC staff considers PSEG's disposition of the peer review findings to be reasonable and 21 that final resolution of the findings is not likely to impact the results of the SAMA analysis.

22 PSEG also stated that there have not been any further reviews of the SGS internal events PRA 23 since the 2008 peer review of PRA Model Revision 4.1.

24 The NRC staff asked PSEG to identify any changes to the plant, including physical and 25 procedural modifications, since Revision 4.1 of the Salem PRA model that could have a 26 significant impact on the results of the SAMA analysis (NRC 2010). In response to the RAI 27 (PSEG 2010a), PSEG explained that one design change and one procedural change have been 28 made since PRA Model Revision 4.1 that have the potential to significantly change the PRA 29 results. The design change was to allow use of two small non-engineered safety feature (ESF) 30 diesel generators to provide power for control and operation of switchyard breakers and to 31 provide a backup source of power to station battery chargers. The procedure change included 32 new procedural steps to provide forced flow of large quantities of outside air to areas supplied 33 by the control area ventilation system. These plant changes resulted in a reduction in the SGS 34 CDF. While the CDF for the updated SGS PRA model, designated as model of record Revision 35 4.3, was not provided in the RAI response, PSEG did provide the updated SGS release 36 frequency of 2.2 x 10-5 per year, which is more than a 50 percent reduction from the 5.0 x 10-September 2010 F-9 Draft NUREG-1437, Supplement 45

Appendix F 1

per year used in the SAMA analysis. The impact of this change on the SAMA analysis is 2

discussed in Sections F.3.2 and F.6.2.

3 In the ER, PSEG explains that, in addition to peer reviews, other measures to ensure, validate, 4

and maintain the quality of the SGS PRA include a formal qualification program for PRA staff, 5

use of procedural guidance to perform PRA tasks, and a program to control PRA models and 6

software. PSEG concludes that based on this quality control process, use of PRA Model 7

Revision 4.1 for the SAMA evaluation was deemed appropriate.

8 Given that the PSEG internal events PRA model has been peer-reviewed and the peer review 9

findings were judged to have minimal impact on the results of the SAMA analysis, and that 10 PSEG has satisfactorily addressed NRC staff questions regarding the PRA, the NRC staff 11 concludes that the internal events Level 1 PRA model is of sufficient quality to support the 12 SAMA evaluation.

13 As indicated above, the current SGS PRA does not include external events. In the absence of 14 such an analysis, PSEG used the SGS IPEEE to identify the highest risk accident sequences 15 and the potential means of reducing the risk posed by those sequences, as discussed below 16 and in Section F.3.2.

17 The SGS IPEEE was submitted in November 1995 (PSEG 1996), in response to Supplement 4 18 of Generic Letter 88-20 (NRC 1991a). The submittal included a seismic PRA, a fire PRA, and a 19 screening analysis for other external events. While no fundamental weaknesses or 20 vulnerabilities to severe accident risk in regard to the external events were identified, several 21 potential enhancements were identified as discussed below. In a letter dated May 21, 1999, 22 (NRC 1999) NRC staff concluded that the submittal met the intent of Supplement 4 to Generic 23 Letter 88-20, and that the licensee's IPEEE process is capable of identifying the most likely 24 severe accidents and severe accident vulnerabilities.

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

26 The seismic PRA included: a seismic hazard analysis, a seismic fragility assessment, a seismic 27 systems analysis, and quantification of seismic CDF.

28 The seismic hazard analysis estimated the annual frequency of exceeding different levels of 29 ground motion. Seismic CDFs were determined for both the EPRI (EPRI 1989) and the 30 Laurence Livermore National Laboratory (LLNL) (NRC 1994) hazard assessments. The seismic 31 fragility assessment utilized the walkdown and screening procedures in EPRI's seismic margin 32 assessment methodology (EPRI 1991). Fragility calculations were made for about 100 33 components and, using a screening criteria of median peak ground acceleration (pga) of 1.5 g 34 which corresponds to a 0.5 pga high confidence low probability of failure (HCLPF) capacity, a 35 total of 27 components remained after screening. The seismic systems analysis defined the 36 potential seismic induced structure and equipment failure scenarios that could occur after a 37 seismic event and lead to core damage. The SGS IPE event tree and fault tree models were September 2010 F-10 Draft NUREG-1437, Supplement 45

Appendix F 1

used as the starting point for the seismic analysis but an explicit seismic event tree (SET) was 2

used to delineate the potential successes and failures that could occur due to a seismic event.

3 Quantification of the seismic models consisted of considering the seismic hazard curve with the 4

appropriate structural and equipment seismic fragility curves to obtain the frequency of the 5

seismic damage state. The conditional probability of core damage given each seismic damage 6

state was then obtained from the IPE models with appropriate changes to reflect the seismic 7

damage state. The CDF was then given by the product of the seismic damage state probability 8

and the conditional core damage probability.

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

11 Both utilized the IPE internal events PRA, with a CDF of 6.4 x 10s per year for quantification of 12 non-seismic failures. While the IPEEE indicated that the EPRI results were believed to be more 13 realistic PSEG assumed a seismic CDF of 9.5 x 10.6 per year based on the LLNL seismic 14 hazard curve in the development of the external events multiplier for purposes of the SAMA 15 evaluation (PSEG 2009). In the ER, PSEG provided a listing and description of the top seven 16 seismic core damage contributors. The dominant seismic core damage contributors for the 17 LLNL seismic hazard curve, representing about 95 percent of the seismic CDF, are listed in 18 Table F-4. The largest contributors to seismic CDF are seismic-induced LOOP caused by 19 failure of the switchyard ceramic insulators combined with random failure of the EDGs and 20 seismic-induced LOOP and failure of battery trains A and B caused by failure of the masonry 21 block walls around the batteries. The NRC staff agrees that the seismic CDF of 9.5 x 10.6 per 22 year is reasonable for the SAMA analysis.

23 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 September 2010 F-1 1 Draft NUREG-1437, Supplement 45

Appendix F 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) 2 The SGS IPEEE did not identify any vulnerabilities due to seismic events but did identify three 3

improvements to reduce seismic risk. These improvements are 1) procedural change to ensure 4

long term alternate ventilation for the Auxiliary Building, 2) replacement of identified low 5

ruggedness relays with higher seismic capacity relays, and 3) reinforcement of an 8-foot 6

masonry wall in the 4kV switchgear room. PSEG clarified in response to an NRC staff RAI that 7

the first two improvements have been implemented (PSEG 2010a). The third improvement is 8

discussed further in Section F.3.2.

9 The SGS IPEEE fire analysis employed EPRI's fire-induced vulnerability evaluation (FIVE) 10 methodology (EPRI 1993) followed by a PRA quantification of the unscreened compartments.

11 The fire evaluation was performed on the basis of fire areas which are plant locations 12 completely enclosed by 2-hour rated fire barriers and meeting the FIVE fire barrier criterion 13 related to preventing propagation. Stage 1 consisted of qualitative screening of all plant fire 14 areas to determine whether a fire could cause a plant shutdown or trip, or lead to loss of safe 15 shutdown equipment. Stage 1 also consisted of quantitative screening performed by estimating 16 whether an area's associated fire frequency in combination with the conditional core damage 17 probability given by the loss of functions potentially impacted by the fire was less than the 1 x 18 10e per year. Based on qualitative and quantitative screening all but 38 fire areas were 19 screened out. Stage 2 was to evaluate the remaining fire areas by modeling fire growth and 20 propagation to determine the fire damage state for each fire area. Stage 3 was an evaluation of 21 Sandia Fire Risk Scoping Study issues (NRC 1989) using the tailored walkdown approach 22 provided in the FIVE methodology. Containment performance was also examined to evaluate 23 the performance of containment systems and equipment following core damage resulting from a 24 fire. The final stage was assessment of the functional effects on the plant for each fire damage 25 state by developing explicit fire event trees to probabilistically assess unscreened areas.

26 Probabilistic credit was given for automatic and manual fire suppression systems. Final 27 quantification utilized FIVE fire data and refined conditional core damage probabilities (CCDPs) 28 from the IPE internal events PRA. The resulting fire induced CDF was calculated to be 2.3 x 10 29 5per year.

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

31 The dominant fire core damage contributors, representing about 99 percent of the fire CDF, are 32 listed in Table F-5. The largest contributors to fire CDF are fires in the 460V Switchgear 33 Rooms, Relay Room, and Control Rooms.

34 Subsequent to the IPEEE, SGS replaced the CO 2 suppression systems with water sprinkler 35 systems in the 460V Switchgear Rooms, 4160V Switchgears Rooms, and Lower Electrical September 2010 F-12 Draft NUREG-1437, Supplement 45

Appendix F 1

Penetration Area. In addition, the results of cable wrap tests suggested that the cable wrap 2

would not perform as expected in some areas of the plant and, subsequent to the IPEEE, was 3

removed and replaced. Because of the suppression system changes made to the three areas 4

identified, PSEG did not consider the IPEEE results for these areas valid. PSEG reassessed 5

the fire CDF for these areas using PRA insights from an interim SGS fire model. If the interim 6

SGS fire model showed a higher CDF for any of these three areas, the higher CDF was used for 7

the SAMA analysis. This was the case for the 460V Switchgear Rooms and the Lower 8

Electrical Penetration Area. The fire CDF from the interim SGS fire model for these two fire 9

areas are provided in Table F-5. These insights increased the total fire CDF to 3.8 x 10-5 per 10 year, which was used in the SAMA analysis.

11 The NRC staff asked PSEG to provide additional information about the interim SGS fire model 12 and, specifically, why it was not used for the SAMA analysis beyond the three areas discussed 13 (NRC 201 Oa). In response to the RAI, PSEG explained that after the completion of the IPEEE, 14 there was an effort made to develop a fire PRA. This resulted in a partially complete "interim 15 SGS fire model." However, the interim SGS fire model was never integrated into the internal 16 events PRA model of record (which at the time was Revision 3) and was essentially abandoned 17 because of the forthcoming NUREG/CR-6850 fire PRA development guidance that would 18 render the SGS fire modeling methodology obsolete.

19 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 10-5 34 Relay Room 7.2 x 10 -

19 Control Rooms, Peripheral Room, and 7.0 x 10-6 18 Ventilation Rooms 4160V Switchgear Room 3.4 x 104 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 10-7 2

Service Water Intake 4.2 x 10-7 1

Reactor Plant Auxiliary Equipment Area (1OOC) 2.9 x 107 1

1CDF reported for the 460V Switchgear Rooms and 4160V Switchgear Rooms is from the interim SGS fire model. All other CDFs are from the IPEEE.

20 September 2010 F-1 3 Draft NUREG-1437, Supplement 45

Appendix F 1

The SGS IPEEE did not identify any vulnerabilities due to fire events but did identify two 2

improvements to reduce fire risk. These improvements are 1) procedural change to enhance 3

cooling in the switchgear and control areas in the event of a fire and 2) procedural change for 4

the control of transient combustibles in the turbine building. PSEG clarified in response to an 5

NRC staff RAI that the two suggested improvements have been implemented (PSEG 201 Oa).

6 As discussed previously, PSEG identified in the ER that SGS has replaced CO 2 fire suppression 7

systems with water sprinkler systems in three areas of the plant since the IPEEE and that cable 8

wrap has been removed and replaced in several areas of the plant since the IPEEE. The NRC 9

staff asked PSEG if any other fire-related improvements have been made since the IPEEE 10 (NRC 2010a). In response to the RAI, PSEG indicated that the following improvements had 11 been made since the IPEEE: 1) the ventilation system and strategy for maintaining viable 12 working conditions was revised for the 4160 Switchgear Room and the Upper Electrical and 13 Piping Penetration Areas and 2) the maintenance shop was eliminated in the Turbine and 14 Service Buildings in order to reduce the initiating event frequency of fires that would damage the 15 cables for the emergency 4kV buses.

16 In the ER, PSEG states that an effective comparison between the internal events PRA results 17 and the fire analysis results is not possible because neither the plant response model or the fire 18 modeling methodology used in the IPEEE is up-to-date. PSEG also identified areas where fire 19 CDF quantification may introduce different levels of uncertainty than expected in the internal 20 events PRA and identified a number of conservatisms in the IPEEE fire analysis, including:

21 0

A revised NRC fire events database indicates a trend toward lower frequency and less 22 severe fires than assumed in the SGS IPEEE.

23 Bounding fire modeling assumptions are used for many fire scenarios. For example, all 24 equipment in a cabinet is damaged for any fire within a cabinet, regardless of whether it 25 is suppressed. Other examples are provided in the ER.

26 0

Because of a lack of industry experience with regard to crew performance during the 27 types of fires modeled in the fire PRA, the characterization of crew actions in the fire 28 PRA is generally conservative.

29 PSEG's conclusion is that while there are both conservative and potentially non-conservative 30 factors included in the IPEEE fire model, the IPEEE is judged to have more conservative bias 31 than the internal events model.

32 Although the arguments regarding the conservatisms in the fire analysis are presented in the 33 ER, PSEG used the modified IPEEE fire CDF of 3.8 x 10-5 per year in the SAMA analysis rather 34 than some reduced value. Considering the above discussion, the conservatisms in the IPEEE 35 fire analysis as currently understood, and the response to the NRC staff RAIs, the NRC staff 36 concludes that the fire CDF of 3.8 x 10.5 per year is reasonable for the SAMA analysis.

September 2010 F-14 Draft NUREG-1437, Supplement 45

Appendix F 1

The SGS IPEEE analysis of high winds, floods, and other external (HFO) events followed the 2

progressive screening method defined in NUREG-1407 (NRC 1991b). While SGS is not 3

considered a 1975 Standard Review Plan (SRP) plant, aspects of its licensing basis do conform 4

to the 1975 SRP criteria because SGS is co-located with Hope Creek Generating Station 5

(HCGS), which does meet the 1975 SRP criteria (PSEG 1996). For those events that are 6

based on the location of the site, and not plant-specific features, the 1975 SRP criteria was 7

used for the HFO screening analysis. Progressively more quantitatively based methods were 8

employed for those events that could not be shown to conform to the 1975 SRP criteria. The 9

IPEEE concluded that all HFO events either complied with the 1975 SRP criteria or that their 10 predicted CDF was below the IPEEE screening criteria (i.e. < 1 x 10-6 per year). For the SAMA 11 analysis, PSEG assumed a CDF contribution of 1 x 10e per year for each of high winds, 12 external floods, transportation and nearby facilities, detritus, and chemical releases for a total 13 HFO CDF contribution of 5 x 10-6 per year (PSEG 2009).

14 Although the SGS IPEEE did not identify any vulnerabilities due to HFO events, three 15 improvements to reduce risk were identified. These improvements are 1) modify the circulating 16 water intake structure to protect against detritus (blockage), 2) make improvements to protect 17 against water ingress pathways for external flooding events, and 3) improve the hold downs for 18 hydrogen tanks to protect against tornados. PSEG clarified in response to an NRC staff RAI 19 that the first two suggested improvements have been implemented (PSEG 2010a). The third 20 improvement is discussed further in Section F.3.2.

21 The NRC staff asked about the status and potential impact on the SAMA analysis of a liquefied 22 natural gas (LNG) terminal planned for Logan Township, New Jersey, upstream on the 23 Delaware River from the SGS site (NRC 2010a). In response to the RAI, PSEG discussed the 24 current status of the LNG terminal as well as the regulatory controls for LNG marine traffic and 25 LNG ship design and the safety record of LNG shipping (PSEG 2010a). The LNG terminal 26 remains in the planning stage and no construction has begun. Further, the state of Delaware 27 has denied applications for several required environmental permits and approvals. PSEG 28 concluded that based on the regulatory process and controls for assuring the safety and 29 security of LNG ships, the safety record of LNG ships, and the uncertainty of the planned 30 terminal, consideration of potential SAMAs associated with the possible future terminal is not 31 warranted. The NRC staff agrees with this conclusion.

32 Based on the aforementioned results, the external events CDF is approximately equal to the 33 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 34 year, an HFO CDF of 5.0 x 10-6 per year, and an internal events CDF of 5.0 x 10- per year 35 used in the SAMA analysis). Accordingly, the NRC staff concurred with SGS's conclusion that 36 the total CDF (from internal and external events) would be approximately 2 times the internal 37 events CDF. In the SAMA analysis submitted in the ER, PSEG doubled the benefit that was 38 derived from the internal events model to account for the combined contribution from internal 39 and external events. The NRC staff agrees with the licensee's overall conclusion concerning 40 the multiplier used to represent the impact of external events and concludes that the licensee's September 2010 F-1 5 Draft NUREG-1437, Supplement 45

Appendix F 1

use of a multiplier of 2 to account for external events is reasonable for the purposes of the 2

SAMA evaluation. This is discussed further in Section F.6.2.

3 The NRC staff reviewed the general process used by PSEG to translate the results of the Level 4

1 PRA into containment releases, as well as the results of the Level 2 analysis, as described in 5

the ER and in response to NRC staff RAIs (PSEG 2010a). The current Level 2 model is 6

essentially a complete revision of the IPE Level 2 model. In response to an NRC staff RAI, 7

PSEG stated that the IPE Level 2 model was abandoned, with the exception of LERF, with 8

Revision 3 of the SGS PRA model and that the Level 2 model was recreated incorporating 9

current industry guidance as part of the transition from Revision 3 to Revision 4 of the PRA 10 model (PSEG 2010a).

11 The current SGS Level 2 model utilizes a single CET containing both phenomenological and 12 systemic events. The Level 1 core damage sequences are grouped into core damage accident 13 classes, or plant damage states (PDSs), with similar characteristics. The PDSs are defined 14 based on the following attributes: (1) RCS pressure (high or low), (2) containment isolation 15 status, (3) containment bypass status, (4) containment bypass via an unisolated SGTR, (5) 16 containment bypass via an unisolated, large ISLOCA, (6) containment spray operation mode, 17 (7) containment fan cooler operation, and (8) RWST injection. All of the sequences in an 18 accident class are then input to the CET by linking the level 1 event tree sequences with the 19 level 2 CET. The CET is analyzed by the linking of fault trees that represent each CET node.

20 Whenever possible the fault trees utilized in the Level 1 analysis are utilized in the CET to 21 propagate dependencies. In response to an NRC staff RAI, PSEG states that the Level 1 and 22 Level 2 models are integrated in that the Level 1 sequences are directly passed to the Level 2 23 model in the software through the Level 1 sequence fault trees (PSEG 2010a). Twenty-three 24 distinct CET end states or sequences result.

25 Section E.2.2.3 of the ER describes each of the top events of the CET and states that branch 26 point probabilities for each top event are based on previous SGS Level 2 analyses, recent 27 accident progression research, and similar analyses for other nuclear plants. The NRC staff 28 requested that PSEG describe how the branch point probabilities were developed specifically 29 for top events RCS Depressurization and Containment Heat Removal (NRC 2010a). In 30 response to the RAI, PSEG clarified that top event RCS Depressurization consists of the 31 combination of an existing human action from the HRA and the fault tree for PORV operation 32 (PSEG 2010a). The Containment Heat Removal top event is determined by specific Level 2 33 system models for containment fan cooler units (CFCUs) and containment spray (CS), either of 34 which can be used for containment heat removal at SGS.

35 Each CET end state represents a radionuclide release to the environment and is assigned to a 36 release category based on timing of release, the initiating event, whether feedwater is available, 37 and the containment failure mode. Three general release categories are defined: intact 38 containment, late release, and early release. These are further divided into eleven detailed 39 release categories based on the above attributes, as defined in Section E.2.2.6 of the ER.

September 2010 F-16 Draft NUREG-1437, Supplement 45

Appendix F 1

The frequency of each release category was obtained by summing the frequency of the 2

contributing CET end states. The release characteristics for each release category were 3

developed by using the results of Modular Accident Analysis Program (MAAP Version 4.0.6) 4 computer code calculations (PSEG 2010a). Representative MAAP cases for each release 5

category were chosen to either represent the most likely initiators in the release category (intact 6

containment and late release categories) or to conservatively bound the consequences of the 7

release (early release categories). The NRC questioned why PSEG did not also use 8

representative cases that bound the consequences for the late release categories (NRC 2010a).

9 In response to the RAI, PSEG stated that, because the late release categories take more time 10 to evolve than the early release categories, the late release categories are less affected by the 11 initial accident conditions and so result in more uniform consequences than the early release 12 categories (PSEG 2010a). Since the accident sequences assigned to the late release 13 categories yielded similar consequences, PSEG selected representative MAAP cases that 14 represented the most likely initiators within those release categories. The release categories, 15 their frequencies, and release characteristics are presented in Tables E.3-5 and E.3-6 of 16 Appendix E to the ER (PSEG 2009).

17 The total Level 2 release frequency is of 5.0 x 10-5 per year, which is about 4 percent higher 18 than the internal events CDF of 4.8 x 10s per year. The ER states that this difference is due to 19 truncation of low probability sequences and inclusion of non-minimal Level 1 sequences. The 20 NRC staff considers that use of the release frequency rather than the Level 1 CDF will have a 21 negligible impact on the results of the SAMA evaluation because the external event multiplier 22 and uncertainty multiplier used in the SAMA analysis (discussed in Section F.6.2) have a much 23 greater impact on the SAMA evaluation results than the small error arising from the model 24 quantification approach.

25 The revised SGS Level 2 PRA model was included in the 2008 PWR Owner's Group peer 26 review discussed above. While none of the eight key findings had to do with the Level 2 27 analysis, eight LERF analysis SRs did not meet Capability Category II or higher and remain 28 open in SGS PRA MOR Revision 4.3 (PSEG 2010b). PSEG determined that all eight of these 29 findings were documentation issues that did not impact the SAMA analysis.

30 Based on the NRC staff's review of the Level 2 methodology, that PSEG has adequately 31 addressed NRC staff RAIs, and that the Level 2 model was reviewed in more detail as part of 32 the 2008 PWR Owners Group peer review and there were no findings that impacted the SAMA 33 analysis, the NRC staff concludes that the Level 2 PRA provides an acceptable basis for 34 evaluating the benefits associated with various SAMAs.

35 The NRC staff reviewed the process used by PSEG to extend the containment performance 36 (Level 2) portion of the PRA to an assessment of offsite consequences (essentially a Level 3 37 PRA). This included consideration of the source terms used to characterize fission product 38 releases for the applicable containment release categories and the major input assumptions 39 used in the offsite consequence analyses. The MACCS2 code was utilized to estimate offsite September 2010 F-17 Draft NUREG-1437, Supplement 45

Appendix F 1

consequences. Plant-specific input to the code includes the source terms for each source term 2

category and the reactor core radionuclide inventory (both discussed above), site-specific 3

meteorological data, projected population distribution within an 80-kilometer (50-mile) radius for 4

the year 2040, emergency evacuation modeling, and economic data. This information is 5

provided in Section E.3 of Appendix E to the ER (PSEG 2009).

6 PSEG used the MACCS2 code and a core inventory from a plant specific calculation at end of 7

cycle to determine the offsite consequences of activity release. In response to an NRC staff 8

RAI, PSEG stated that the MACCS2 analysis was based on the core inventory used in the 9

February 2006 NRC-approved Alternate Source Term for SGS (PSEG 2010a). As indicated in 10 the ER, the reactor core radionuclide inventory used in the consequence analysis was based on 11 a thermal power of 3632 MWt, which is 5 percent higher than the current licensed thermal 12 power of 3459 MWt for SGS. In response to an NRC staff RAI, PSEG stated that the higher 13 thermal power was used to provide margin for a future power uprate (PSEG 2010a).

14 All releases were modeled as being from the top of the reactor containment building and at low 15 thermal content (ambient). Sensitivity studies were performed on these assumptions and 16 indicated little or no change in population dose or offsite economic cost. Assuming a ground 17 level release decreased dose risk and cost risk by 8 percent and 7 percent, respectively.

18 Assuming a buoyant plume decreased dose risk and cost risk by 1 percent or less. Based on 19 the information provided, the staff concludes that the release parameters utilized are acceptable 20 for the purposes of the SAMA evaluation.

21 PSEG used site-specific meteorological data for the 2004 calendar year as input to the 22 MACCS2 code. The development of the meteorological data is discussed in Section E.3.7 of 23 Appendix E to the ER. The data were collected from onsite and local meteorological monitoring 24 systems. Sensitivity analyses using MACCS2 and the meteorological data for the years 2005 25 through 2007 show that use of data for the year 2004 results in the largest dose and economic 26 cost risk. Missing meteorological data was filled by (in order of preference): using data from the 27 backup met pole instruments (10-meter), using corresponding data from another level of the 28 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 29 same hour and a nearby day (substitution technique). The 10-meter wind speed and direction 30 were combined with precipitation and atmospheric stability (derived from the vertical 31 temperature gradient) to create the hourly data file for use by MACCS2. The NRC staff notes 32 that previous SAMA analyses results have shown little sensitivity to year-to-year differences in 33 meteorological data and concludes that the use of the 2004 meteorological data in the SAMA 34 analysis is reasonable.

35 The population distribution the licensee used as input to the MACCS2 analysis was estimated 36 for the year 2040 using year 1990 and year 2000 census data as accessed by SECPOP2000 37 (NRC 2003) as a starting point. In response to an NRC staff RAI, PSEG stated that the 38 transient population was included in the 10-mile EPZ, and in the population projection (PSEG 39 2010a). A ten year population growth rate was estimated using the year 1990 to year 2000 September 2010 F-18 Draft NUREG-1437, Supplement 45

Appendix F 1

SECPOP2000 data and applied to obtain the distribution in 2040. The baseline population was 2

determined for each of 160 sectors, consisting of sixteen directions for each of ten concentric 3

distance rings to a radius of 50 miles surrounding the site. The SECPOP2000 census data from 4

1990 and 2000 were used to determine a ten year population growth factor for each of the 5

concentric rings. The population growth was averaged over each ring and applied uniformly to 6

all sectors within each ring. The NRC staff requested PSEG provide an assessment of the 7

impact on the SAMA analysis if a wind-direction weighted population estimate for each sector 8

were used (NRC 2010a). In response to the RAI, PSEG stated that the impacts associated with 9

angular population growth rates on population dose risk and offsite economic cost risk are 10 minimal and bounded by the 30 percent population sensitivity case (PSEG 2010a). This is 11 based on the relatively even wind distribution profile surrounding the site, the tendency for 12 lateral dispersion between sectors, and the use of mean values in the analysis. A sensitivity 13 study was performed for the population growth at year 2040. A 30 percent increase in 14 population resulted in a 30 percent increase in dose risk and a 29 percent increase in cost risk.

15 In response to an NRC staff RAI, PSEG stated that the radial growth rates used in the MACCS2 16 analysis provides a more conservative population growth estimate than using 'whole county' 17 data for averaging (PSEG 2010a). PSEG also identified that the population sensitivity case of 18 30 percent growth was approximately equivalent to adding 6.8 percent to the 10-year growth 19 rate. The NRC staff considers the methods and assumptions for estimating population 20 reasonable and acceptable for purposes of the SAMA evaluation.

21 The emergency evacuation model was modeled as a single evacuation zone extending out 16 22 kilometers (10 miles) from the plant (the emergency planning zone - EPZ). PSEG assumed 23 that 95 percent of the population would evacuate. This assumption is conservative relative to 24 the NUREG-1 150 study (NRC 1990), which assumed evacuation of 99.5 percent of the 25 population within the emergency planning zone. The evacuated population was assumed to 26 move at an average radial speed of approximately 2.8 meters per second (6.3 miles per hour) 27 with a delayed start time of 65 minutes after declaration of a general emergency (KLD 2004). A 28 general emergency declaration was assumed to occur at the onset of core damage. The 29 evacuation speed is a time-weighted average value accounting for season, day of week, time of 30 day, and weather conditions. It is noted that the longest evacuation time presented in the study 31 (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 32 issuance of the advisory to evacuate). Sensitivity studies on these assumptions indicate that 33 there is minor impact to the population dose or offsite economic cost by the assumed variations.

34 The sensitivity study reduced the evacuation speed by 50 percent to 1.4 m/s. This change 35 resulted in a 4 percent increase in population dose risk and no change in offsite economic cost 36 risk. The NRC staff concludes that the evacuation assumptions and analysis are reasonable 37 and acceptable for the purposes of the SAMA evaluation.

38 Site specific agriculture and economic parameters were developed manually using data in the 39 2002 National Census of Agriculture (USDA 2004) and from the Bureau of Economic Analysis 40 (BEA 2008) for each of the 23 counties surrounding SGS, to a distance of 50 miles. Therefore, 41 recently discovered problems in SECPOP2000 do not impact the SGS analysis. The values September 2010 F-19 Draft NUREG-1437, Supplement 45

Appendix F 1

used for each of the 160 sectors were the data from each of the surrounding counties multiplied 2

by the fraction of that county's area that lies within that sector. Region-wide wealth data (i.e.,

3 farm wealth and non-farm wealth) were based on county-weighted averages for the region 4

within 50-miles of the site using data in the 2002 National Census of Agriculture (USDA 2004) 5 and the Bureau of Economic Analysis (BEA 2008). Food ingestion was modeled using the new 6

MACCS2 ingestion pathway model COMIDA2 (NRC 1998). For SGS, less than one percent of 7

the total population dose risk is due to food ingestion.

8 In addition, generic economic data that is applied to the region as a whole were revised from the 9

MACCS2 sample problem input in order to account for cost escalation since 1986, the year that 10 input was first specified. A factor of 1.96, representing cost escalation from 1986 to April 2008 11 was applied to parameters describing cost of evacuating and relocating people, land 12 decontamination, and property condemnation.

13 The NRC staff concludes that the methodology used by PSEG to estimate the offsite 14 consequences for SGS provides an acceptable basis from which to proceed with an 15 assessment of risk reduction potential for candidate SAMAs. Accordingly, the NRC staff based 16 its assessment of offsite risk on the CDF and offsite doses reported by PSEG.

17 F.3 Potential Plant Improvements 18 The process for identifying potential plant improvements, an evaluation of that process, and the 19 improvements evaluated in detail by PSEG are discussed in this section.

20 F.3.1 Process for Identifying Potential Plant Improvements 21 PSEG's process for identifying potential plant improvements (SAMAs) consisted of the following 22 elements:

23 0

Review of the most significant basic events from the current, plant-specific PRA and 24 insights from the SGS PRA group, 25 0

Review of potential plant improvements identified in, and original results of, the SGS IPE 26 and IPEEE, 27 0

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

28 nuclear sites, and 29 0

Review of generic SAMA candidates from NEI 05-01 (NEI 2005) to identify SAMAs that 30 might address areas of concern identified in the SGS PRA.

31 Based on this process, an initial set of 27 candidate SAMAs, referred to as Phase I SAMAs, was 32 identified. In Phase I of the evaluation, PSEG performed a qualitative screening of the initial list 33 of SAMAs and eliminated SAMAs from further consideration using the following criteria:

September 2010 F-20 Draft NUREG-1437, Supplement 45

Appendix F 1

9 The SAMA is not applicable to SGS due to design differences 2

0 The SAMA has already been implemented at SGS, 3

0 The SAMA would achieve results that have already been achieved at SGS by other 4

means, or 5

0 The SAMA has estimated implementation costs that would exceed the dollar value 6

associated with completely eliminating all severe accident risk at SGS.

7 Based on this screening, two SAMAs were eliminated leaving 25 for further evaluation. The 8

results of the Phase I screening analysis is given in Table E.5-3 of Appendix E to the ER. The 9

remaining SAMAs, referred to as Phase II SAMAs, are listed in Table E.6-1 of Appendix E to the 10 ER. In Phase II, a detailed evaluation was performed for each of the 25 remaining SAMA 11 candidates, as discussed in Sections F.4 and F.6 below. To account for the potential impact of 12 external events, the estimated benefits based on internal events were multiplied by a factor of 2, 13 as previously discussed.

14 F.3.2 Review of PSEG's Process 15 PSEG's efforts to identify potential SAMAs focused primarily on areas associated with internal 16 initiating events, but also included explicit consideration of potential SAMAs for important fire 17 and seismic initiated core damage sequences. The initial list of SAMAs generally addressed the 18 accident sequences considered to be important to CDF from risk reduction worth (RRW) 19 perspectives at SGS, and included selected SAMAs from prior SAMA analyses for other plants.

20 PSEG provided a tabular listing of the Level 1 PRA basic events sorted according to their RRW 21 (PSEG 2009). SAMAs impacting these basic events would have the greatest potential for 22 reducing risk. PSEG used a RRW cutoff of 1.01, which corresponds to about a one percent 23 change in CDF given 100-percent reliability of the SAMA. This equates to a benefit of 24 approximately $164,000 (after the benefits have been multiplied by a factor of 2 to account for 25 external events). PSEG also provided and reviewed the Level 2 PRA basic events, down to a 26 RRW of 1.01, for the release categories contributing over 94 percent of the population dose-risk.

27 The Level 2 basic events for the remainder of the release categories were not included in the 28 review so as to prevent high frequency-low consequence events from biasing the importance 29 listing. All of the basic events on the Level 1 and 2 importance lists were addressed by one or 30 more of the SAMAs (PSEG 2009). As a result of the review of the Level 1 and Level 2 basic 31 events, 19 SAMAs were identified.

32 The NRC staff requested PSEG to extend the review of the Level 1 and 2 basic events down to 33 a RRW threshold of 1.003, which equates to a benefit of approximately $50,000, the assumed 34 cost of a procedural change at SGS (NRC 2010a). In response to the RAI, PSEG provided 35 revised Level 1 and Level 2 importance lists using SGS PRA model of record Revision 4.3, 36 which was discussed in Section F.2.2, and extended the review of the basic events down to an September 2010 F-21 Draft NUREG-1 437, Supplement 45

Appendix F 1

RRW of 1.006, which equates to a benefit of about $47,000 using PRA Revision 4.3. The 2

review identified the following three additional SAMAs associated with new basic events added 3

to the importance lists (PSEG 2010a):

4 0

SAMA 30 - Automatic Start of Diesel-Powered Air Compressor 5

0 SAMA 31 - Fully Automate Swapover to Sump Recirculation 6

0 SAMA 32 - Enhance Flood Detection for 100-foot Auxiliary Building and Enhance 7

Procedural Guidance for Responding to Internal Floods 8

A Phase II detailed evaluation was performed for each of these additional SAMAs, which is 9

discussed in Section F.6.2.

10 The NRC staff asked PSEG to clarify the appropriateness of determining importance factors, 11 and SAMAs, for initiators that are identified as flag events having an assigned probability of 1.0 12 (NRC 2010a). PSEG explained in response to the RAI that fault trees were developed for 13 several loss of support system initiating events (PSEG 201 Oa). Those events that lead to the 14 loss of a support system and are responsible for causing the modeled initiating event were 15 identified as flag events. These events are representative of that initiating event's contribution 16 to CDF and were therefore considered appropriate by PSEG for risk ranking. PSEG further 17 clarified that events whose failure leads to the occurrence of the modeled initiating event will 18 also be listed in the importance list ranking and that the flag probability was therefore set to 1.0 19 to determine the appropriate CDF contribution of the cutsets. The RRW calculated for these 20 flag events therefore correctly measures the risk significance of the initiating event modeled in 21 this manner.

22 The NRC staff also asked PSEG to clarify the significance of determining importance factors, 23 and SAMAs, for two split fraction events identified in the importance listing: "RCS-SLOCA-24 SPLIT" and "MFI-UNAVAILABLE" (NRC 2010a). PSEG explained in response to the RAI that 25 the first event, "RCS-SLOCA-SPLIT," is a flag event that indicates those cutsets in which an 26 RCP seal LOCA has occurred and that the second event, "MFI-UNAVAILABLE," is the 27 conditional probability that the main feedwater system is unavailable given that a reactor trip 28 signal has been generated, irrespective of whether an ATWS condition exists (PSEG 201 Oa).

29 Because the first event is a flag event, it was assigned a probability of 1.0. SAMA 6, "Enhance 30 Flood Detection for 84' Auxiliary Building and Enhance Procedural Guidance for Responding to 31 Service Water Flooding," was identified because isolating a service water rupture early could 32 help prevent the conditions that can lead to an RCP seal LOCA. The second event was 33 assigned a conditional probability of 0.3. SAMA 14, "Expand ATWS Mitigation System 34 Actuation Circuitry (AMSAC) Function to Include Backup Breaker Trip on Reactor Protection 35 System (RPS) Failure," was identified to use the AMSAC system to provide a redundant trip 36 signal to help mitigate ATWS events. In over 60 percent of the scenarios in which MFI-September 2010 F-22 Draft NUREG-1437, Supplement 45

Appendix F 1

UNAVAILABLE is a contributor, AMSAC maintenance is also a contributor. By mitigating ATWS 2

events, SAMA 14 also mitigates scenarios having this combination of events.

3 PSEG reviewed the cost-beneficial Phase II SAMAs from prior SAMA analyses for five 4

Westinghouse PWR and one General Electric BWR sites. PSEG's review determined that all of 5

the Phase II SAMAs reviewed were either already represented by a SAMA identified from the 6

Level 1 and 2 importance list reviews, are already addressed by other means, have low 7

potential for risk reduction at SGS, or were not applicable to the SGS design. This review 8

resulted in no additional SAMAs being identified.

9 The NRC staff noted that PSEG's review of these other analyses appeared to have overlooked 10 additional cost-beneficial SAMAs identified during the staff's review of these same SAMA 11 analyses and requested PSEG provide an assessment any additional cost-beneficial SAMAs 12 identified during these reviews for applicability to SGS (NRC 2010a). In response to the RAI, 13 PSEG reviewed the cost-beneficial SAMAs identified in the NRC-issued NUREG-1437 reports 14 for each of the six nuclear sites and concluded the cost-beneficial SAMA either 1) was already 15 identified and evaluated in the ER, 2) was already implemented at SGS, or 3) would not reduce 16 SGS risk (PSEG 2010a). No additional SAMAs were identified from this review.

17 PSEG considered the potential plant improvements described in the IPE in the identification of 18 plant-specific candidate SAMAs for internal events. Review of the IPE lead to no additional 19 SAMA candidates since the three improvements identified in the IPE have already been 20 implemented at SGS (PSEG 2009).

21 As a sensitivity case to SAMA 5, PSEG identified and evaluated SAMA 5A, "Install Portable 22 Diesel Generators to Charge Station Battery and Circulating Water Batteries." This SAMA only 23 addresses cases in which RCP seals remain intact, which occurs in a majority of the SBO 24 scenarios. PSEG performed a Phase II evaluation of SAMA 5A, which is in addition to the 25 Phase II evaluations performed for the 25 SAMAs discussed above that were not screened 26 during the Phase I evaluation.

27 Based on this information, the NRC staff concludes that the set of SAMAs evaluated in the ER, 28 together with those identified in response to NRC staff RAIs, addresses the major contributors 29 to internal event CDF.

30 Although the IPEEE did not identify any fundamental vulnerabilities or weaknesses related to 31 external events, the ER identified three improvements related to external events (PSEG 2009).

32 The NRC staff noted that the IPEEE safety evaluation report (NRC 1999) identified five total 33 improvements related to external events and requested PSEG review these improvements for 34 potentially additional SAMAs (NRC 2010a). In response to the RAI, PSEG reviewed the five 35 suggested improvements and reassessed the three improvements originally evaluated in the ER 36 (PSEG 201 Oa). As a result of this review, two improvements related to fire events, three 37 improvements related to seismic events, and three improvements related to HFO events were September 2010 F-23 Draft NUREG-1437, Supplement 45

Appendix F 1

identified. The two suggested fire-related improvements have been implemented, two of the 2

seismic-related improvements have been implemented, and two of the HFO-related 3

improvements have been implemented. The remaining two improvements that have not been 4

implemented are as follows:

5 Seismic-related improvement - reinforcement of an 8-foot masonry wall in the 4kV 6

switchgear room. PSEG described the results of an evaluation that determined there 7

was no interaction between the wall and the switchgear bus during a seismic event and 8

subsequent implementation of a corrective action to revise the associated calculation to 9

clarify the lack of interaction. Based on this, PSEG concluded that reinforcement of the 10 masonry wall was not necessary and no SAMA is suggested (PSEG 201 Oa).

11 HFO-related improvement - improve hold downs for the hydrogen tanks to protect 12 against tornados. In response to the RAI, PSEG performed a walk down of the 13 hydrogen racks and determined that the IPEEE suggested improvements to the Unit 2 14 racks to make the design consistent with the Unit 1 racks was not implemented as 15 indicated in the ER. PSEG further noted that the IPEEE states that these hydrogen 16 tanks "will not have any significant impact on safety structures." Based on this, PSEG 17 concluded that, while the suggested change was prudent, it would not reduce plant risk 18 and no SAMA is suggested.

19 In the ER PSEG also identified three post IPEEE site changes to determine if they could impact 20 the IPEEE results and possibly lead to a SAMA. From this review, one plant change to replace 21 CO 2 fire suppression with water sprinkler systems was determined to have an impact on fire 22 CDF, which was discussed in Section F.2.2. No additional SAMAs were identified from this 23 review.

24 In a further effort to identify external event SAMAs, PSEG reviewed the top 10 fire areas 25 contributing to fire CDF based on the results of the IPEEE and interim SGS fire PRA models.

26 These areas are all of the SGS fire areas having a maximum benefit equal to or greater than 27 approximately $50,000, which is the approximate value of implementing a procedure change at 28 a single unit at SGS. The maximum benefit for a fire area is the dollar value associated with 29 completely eliminating the fire risk in that fire area, which is discussed in Section F.6.2. SAMAs 30 having an implementation cost of less than that of a procedure change, or $50,000, are unlikely.

31 As a result of this review, PSEG identified five Phase I SAMAs to reduce fire risk. The SAMAs 32 identified included both procedural and hardware alternatives (PSEG 2009). The NRC staff 33 concludes that the opportunity for fire-related SAMAs has been adequately explored and that it 34 is unlikely that there are additional potentially cost-beneficial, fire-related SAMA candidates.

35 For seismic events, PSEG reviewed the top seven seismic sequences contributing to seismic 36 CDF based on the results of the IPEEE seismic PRA model. These areas are all of the SGS 37 seismic sequences having a benefit equal to or greater than approximately $50,000, which is 38 the approximate value of implementing a procedure change at a single unit at SGS. The September 2010 F-24 Draft NUREG-1437, Supplement 45

Appendix F 1

maximum benefit for a seismic sequence is the dollar value associated with completely 2

eliminating the seismic risk for that sequence, which is discussed in Section F.6.2. SAMAs 3

having an implementation cost of less than that of a procedure change, or $50,000, are unlikely.

4 As a result of this review, PSEG identified three additional Phase I SAMAs to reduce seismic 5

risk (PSEG 2009). The NRC staff concludes that the opportunity for seismic-related SAMAs has 6

been adequately explored and that it is unlikely that there are additional potentially cost-7 beneficial, seismic-related SAMA candidates.

8 As stated earlier, other external hazards (high winds, external floods, transportation and nearby 9

facility accidents, release of on-site chemicals, and detritus) are below the IPEEE threshold 10 screening frequency, or met the 1975 SRP design criteria, and are not expected to represent 11 vulnerabilities. Nevertheless, PSEG reviewed the IPEEE results and subsequent plant changes 12 for each of these external hazards and determined that either 1) the maximum benefit from 13 eliminating all associated risk was less than approximately $50,000, which is the approximate 14 value of implementing a procedure change at a single unit at SGS, or 2) only hardware 15 enhancements that would significantly exceed the maximum value of any potential risk 16 reduction were available. As a result of this review, PSEG identified no additional Phase I 17 SAMAs to reduce HFO risk (PSEG 2009). The NRC staff concludes that the licensee's 18 rationale for eliminating other external hazards enhancements from further consideration is 19 reasonable.

20 The NRC staff noted that, while the generic SAMA list from NEI 05-01 (NEI 2005) was stated to 21 have been used in the identification of SAMAs for SGS, it was not specifically reviewed to 22 identify SAMAs that might be applicable to SGS but rather was used to identify SAMAs that 23 might address areas of concern identified in the SGS PRA (NRC 201 Oa). The NRC staff asked 24 PSEG to provide further information to justify that this approach produced a comprehensive set 25 of SAMAs for consideration. In response to the RAI, PSEG explained that, based on the early 26 SAMA reviews, both the industry and NRC came to realize that a review of the generic SAMA 27 list was of limited benefit because they were consistently found to not be cost-beneficial and that 28 the real benefit was considered to be in the development of SAMAs generated based on plant 29 specific risk insights from the PRA models (PSEG 2010a).

30 Furthermore, while the generic list does include potential plant improvements for plants having a 31 similar design to SGS, plant designs are sufficiently different that the specific plant 32 improvements identified in the generic list are generally not directly applicable to SGS, and 33 require alteration to specifically address the SGS design and risk contributors or otherwise 34 would be screened as not applicable to the SGS design. For these reasons, PSEG concludes 35 that the real value of the NEI 05-01 generic SAMA list is as an idea source to generate SAMAs 36 that address important contributors to SGS risk. The NRC staff accepts PSEG's conclusion.

37 The NRC staff questioned PSEG about potentially lower cost alternatives to some of the SAMAs 38 evaluated (NRC 2010a), including:

September 2010 F-25 S Draft NUREG-1437, Supplement 45

Appendix F

.1 0

Operating the AFW AF1 1/21 valves closed.

2 0

Install improved fire barriers in the 460V switchgear rooms to provide separation 3

between the three power divisions.

4 0

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

5 In response to the RAIs, PSEG addressed the suggested lower cost alternatives and 6

determined that they were either not feasible or were not cost-beneficial (PSEG 201 Oa). This is 7

discussed further in Section F.6.2.

8 The NRC staff notes that the set of SAMAs submitted is not all-inclusive, since additional, 9

possibly even less expensive, design alternatives can always be postulated. However, the NRC 10 staff concludes that the benefits of any additional modifications are unlikely to exceed the 11 benefits of the modifications evaluated and that the alternative improvements would not likely 12 cost less than the least expensive alternatives evaluated, when the subsidiary costs associated 13 with maintenance, procedures, and training are considered.

14 The NRC staff concludes that PSEG used a systematic and comprehensive process for 15 identifying potential plant improvements for SGS, and that the set of potential plant 16 improvements identified by PSEG is reasonably comprehensive and, therefore, acceptable.

17 This search included reviewing insights from the plant-specific risk studies, and reviewing plant 18 improvements considered in previous SAMA analyses. While explicit treatment of external 19 events in the SAMA identification process was limited, it is recognized that the prior 20 implementation of plant modifications for fire and seismic risks and the absence of external 21 event vulnerabilities reasonably justifies examining primarily the internal events risk results for 22 this purpose.

23 F.4 Risk Reduction Potential of Plant Improvements 24 PSEG evaluated the risk-reduction potential of the 25 remaining SAMAs and one sensitivity 25 case SAMA that were applicable to SGS. The SAMA evaluations were performed using realistic 26 assumptions with some conservatism. On balance, such calculations overestimate the benefit 27 and are conservative.

28 PSEG used model re-quantification to determine the potential benefits. The CDF, population 29 dose reductions, and offsite economic cost reductions were estimated using the SGS PRA 30 model. The changes made to the model to quantify the impact of SAMAs are detailed in 31 Section E.6 of Appendix E to the ER (PSEG 2009). Table F-6 lists the assumptions considered 32 to estimate the risk reduction for each of the evaluated SAMAs, the estimated risk reduction in 33 terms of percent reduction in CDF and population dose, and the estimated total benefit (present 34 value) of the averted risk. The estimated benefits reported in Table F-6 reflect the combined 35 benefit in both internal and external events. The determination of the benefits for the various 36 SAMAs is further discussed in Section F.6.

September 2010 F-26 Draft NUREG-1437, Supplement 45

Appendix F 1

The NRC staff questioned the assumptions used in evaluating the benefit or risk reduction 2

estimate of SAMA 24, "provide procedural guidance to cross-tie Salem 1 and 2 service water 3

systems" (NRC 2010a). The ER assumed this SAMA did not benefit from a reduction in fire risk 4

yet indicates that this SAMA was identified based on a review of the SGS IPEEE fire PRA 5

model results. In response to an NRC staff RAI, PSEG clarified that this SAMA was actually 6

identified from the review of the internal events importance list, that the procedural guidance 7

suggested in this SAMA to perform the inter-unit service water cross-tie is already in place for 8

fire events and that, therefore, implementation of this SAMA would have no additional benefits 9

in fire events (PSEG 2010a). Based on this, PSEG concluded that this SAMA has been 10 appropriately evaluated.

11 The NRC staff noted that the total of the risk reduction results calculated by summing the 12 individual results for each release category for SAMAs 2, 4, 5A, 18, and 19 was different than 13 the summary results that were used in the SAMA evaluation (NRC 2010a). In response to the 14 RAI, PSEG explained that the release category results provided in the ER for these SAMAs 15 were incorrect, due to typographical errors, and the correct results were provided (PSEG 16 201 Oa). PSEG further explained that the SAMA evaluation reported in the ER used the correct 17 release category results and therefore no re-evaluation of the SAMAs was necessary. The 18 NRC staff accepts PSEG's explanation.

19 For SAMAs that specifically addressed fire events (i.e., SAMA 21, "Seal the Category II and III 20 Cabinets in the Relay Room," SAMA 22, "Install Fire Barriers between the lCCl, 1CC2, and 21 1CC3 Consoles in the Control Room Enclosure (CRE)," and SAMA 23, "Install Fire Barriers and 22 Cable Wrap to Maintain Divisional Separation in the 4160V AC Switchgear Room."), the 23 reduction in fire CDF and population dose was not directly calculated (in Table F-5 this is noted 24 as "Not Estimated"). For these SAMAs, an estimate of the impact was made based on general 25 assumptions regarding: the approximate contribution to total risk from external events relative to 26 that from internal events; the fraction of the external event risk attributable to fire events; the 27 fraction of the fire risk affected by the SAMA (based on information from the IPEEE and interim 28 SGS Fire Model results); and the assumption that SAMAs 21 and 22 completely eliminate the 29 fire risk affected by the SAMA and that SAMA 23 eliminates 95 percent of the fire risk affected 30 by the SAMA. Specifically, it is assumed that the contribution to risk from external events is 31 approximately equal to that from internal events, and that internal fires contribute 72 percent of 32 this external events risk. The fire areas impacted by the SAMA are identified and the portion of 33 the total fire risk contributed by each of these fire areas determined. For SAMAs 21 and 22, the 34 benefit or averted cost risk from reducing the fire risk affected by the SAMA is then calculated 35 by multiplying the ratio of the fire risk affected by the SAMA to the internal events CDF by the 36 total present dollar value equivalent associated with completely eliminating severe accidents 37 from internal events at SGS. For SAMA 23, the benefit or averted cost risk from reducing the 38 fire risk affected by the SAMA is then calculated by multiplying the ratio of 95 percent of the fire 39 risk affected by the SAMA to the internal events CDF by the total present dollar value equivalent 40 associated with completely eliminating severe accidents from internal events at SGS. These 41 SAMAs were assumed to have no additional benefits in internal events.

September 2010 F-27 Draft NUREG-1437, Supplement 45

Appendix F 1

In addition to those SAMAs that only addressed fire events, PSEG evaluated the additional 2

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

3 SAMA 1, "Enhance Procedures and Provide Additional Equipment to Respond to Loss of 4

Control Area Ventilation," SAMA 8, "Install High Pressure Pump Powered with Portable Diesel 5

Generator and Long-term Suction Source to Supply the AFW Header," and SAMA 20, "Fire 6

Protection System to Provide Make-up to RCS and Steam Generators." The benefit or averted 7

cost risk from reducing the fire risk affected by these SAMAs was calculated similar to the 8

method described above with the exception that the fire risk affected by each of these SAMAs 9

were assumed to be reduced based on the same failure probability as was assumed for internal 10 events (i.e., 2.OE-02 for SAMA 1, 1.OE-02 for SAMA 8, and 1.OE-01 for SAMA 20). In other 11 words, SAMA 1 was assumed to eliminate 98 percent, SAMA 8 was assumed to eliminate 99 12 percent, and SAMA 20 was assumed to eliminate 90 percent of the fire risk affected by these 13 SAMAs. The benefit or averted cost risk from reducing the fire risk affected by SAMA 1 is then 14 calculated by multiplying the ratio of 98 percent of the fire risk affected by the SAMA to the 15 internal events CDF by the total present dollar value equivalent associated with completely 16 eliminating severe accidents from internal events at SGS. The benefit from reducing fire risk 17 was calculated similarly for SAMAs 8 and 20. For SAMAs 1 and 8, PSEG added the calculated 18 benefit from reducing fire risk to the benefit from internal events, which was doubled to account 19 for all external events, to obtain the total benefit from internal and external events. This is 20 discussed further in Section F.6.2.

21 PSEG also evaluated the additional benefits from reducing seismic risk for the following SAMAs 22 that also had internal events benefits: SAMA 5, "Enhance Procedures and Provide Additional 23 Equipment to Respond to Loss of Control Area Ventilation," SAMA 5A, "Install Portable Diesel 24 Generators to Charge Station Battery and Circulating Water Batteries," SAMA 20, "Fire 25 Protection System to Provide Make-up to RCS and Steam Generators," and SAMA 27, "In 26 addition to the Equipment Installed for SAMA 5, Install Permanently Piped Seismically Qualified 27 Connections to Alternate AFW Water Sources." For these SAMAs, an estimate of the seismic 28 impact was made based on general assumptions regarding: the approximate contribution to 29 total risk from external events relative to that from internal events; the fraction of the external 30 event risk attributable to seismic events; the fraction of the seismic risk affected by the SAMA 31 (based on information from the IPEEE); and the assumption that these SAMAs would reduce 32 the contribution to the seismic CDF from the impacted seismic sequences by 90 percent.

33 Specifically, it is assumed that the contribution to risk from external events is approximately 34 equal to that from internal events, and that seismic events Contribute 18 percent of this external 35 events risk. The seismic sequences impacted by the SAMA are identified and the portion of the 36 total seismic risk contributed by each of these seismic sequences determined. The benefit or 37 averted cost risk from reducing the seismic risk affected by the SAMA is then calculated by 38 multiplying the ratio of 90 percent of the seismic risk affected by the SAMA to the internal events 39 CDF by the total present dollar value equivalent associated with completely eliminating severe 40 accidents from internal events at SGS. For SAMAs 5, 5A, and 27, PSEG added the calculated 41 benefit from reducing seismic risk to the benefit from internal events, which was doubled to September 2010 F-28 Draft NUREG-1437, Supplement 45

Appendix F 1

account for all external events, to obtain the total benefit from internal and external events. This 2

is discussed further in Section F.6.2.

3 For SAMA 20, PSEG multiplied the benefit from internal events by a factor of 1.1 to account for 4

other (non-fire/non-seismic) events and added this to the benefits or averted cost risk from 5

reducing fire risk and seismic risk to obtain the total benefit from internal and external events.

6 This is discussed further in Section F.6.2.

7 The NRC staff has reviewed PSEG's bases for calculating the risk reduction for the various 8

plant improvements and concludes, with the above clarifications, that the rationale and 9

assumptions for estimating risk reduction are reasonable and generally conservative (i.e., the 10 estimated risk reduction is higher than what would actually be realized). Accordingly, the NRC 11 staff based its estimates of averted risk for the various SAMAs on PSEG's risk reduction 12 estimates.

September 2010 F-29 Draft NUREG-1437, Supplement 45

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)

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.

September 2010 F-30 Draft NUREG-1437, Supplement 45

Appendix F 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 Dose External)

Uncertintye 5 - Install Portable Diesel Generators Modify fault tree to include a new 16 11 3.1M 7.6M 3.3M to Charge Station Battery and basic event, having a failure Circulating Water Batteries and probability of 1.0E-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 I.0E-02 from 2.4E-01 for grid and from I.0E-01 for site/switchyard-related causes and to 3.OE-02 from 2.4E-01 for weather-related causes.

5Ab1 - 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 1.0E-01 for sitelswitchyard-related causes and to 3.OE-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 1.0M 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 4.1 M 2.5M Powered with Portable Diesel basic event, having a failure September 2010 F-31 Draft NUREG-1437, Supplement 45

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

% Risk Reduction Total Benefit ($)

Baseline Baseline With Cost ($)

CDF Dose (internal +

Uncertainty(e)

SAMA Assumptions External)

Generator and Long-term Suction probability of I.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/IFA-TGA-88 and IFA-AB-84B assuming the same failure probability.

9 - Connect Hope Creek Cooling Reduce failure probabilities for all 13 11 1.7M 4.3M 1.2M 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 I.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.

September 2010 F-32 Draft NUREG-1437, Supplement 45

Appendix F 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) 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 1.0E-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.OE-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.

September 2010 F-33 Draft NUREG-1437, Supplement 45

Appendix F Table F-6. SAMA Cost/Benefit Screening Analysis for SGSVa)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Dose (internal +

Uncertainty(e)

SAMA Assumptions DoseExternal)

Uneraitye 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.OE-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 1 FA-AB-84A, 1 FA-EP-78C, 1 FA-AB-64A, 1 FA-AB-84B, and 12FA-SB-100/1FA-TGA-88 assuming the same failure probability of 1.OE-01.

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

22 - Install Fire Barriers between the Eliminate the fire CDF contribution from NOT ESTIMATED 330K 830K 1.6M lCCl, 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.

September 2010 F-34 Draft NUREG-1437, Supplement 45

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

SAMA Assumptions 27 - In addition to the Equipment Installed for SAMA 5, Install Permanently Piped Seismically Qualified Connections to Alternate AFW Water Sources Modify fault tree to include a new basic event, having a failure probability of 1.OE-01, representing 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 sitelswitchyard-related causes and to 3.OE-02 from 2.4E-01 for weather-related causes.

30(c) - Automatic Start of Diesel-Powered The failure probability for the operator 1

<1 40K 83K 100K 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 1.0E-01.

September 2010 F-35 Draft NUREG-1437, Supplement 45

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

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Dose ternal Uncertainty(e)

SAMA 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 $100K assumed in the SAMA evaluation.

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

(e)

Using a factor of 2.5.

September 2010 F-36 Draft NUREG-1437, Supplement 45

1 F.5 Cost Impacts of Candidate Plant Improvements 2

PSEG estimated the costs of implementing the 25 candidate SAMAs through the development 3

of site-specific cost estimates. The cost estimates conservatively did not include the cost of 4

replacement power during extended outages required to implement the modifications (PSEG 5

2009).

6 The NRC staff reviewed the bases for the applicant's cost estimates (presented in Table E.5-3 7

of Attachment E to the ER). For certain improvements, the NRC staff also compared the cost 8

estimates to estimates developed elsewhere for similar improvements, including estimates 9

developed as part of other licensees' analyses of SAMAs for operating reactors.

10 The ER stated that plant personnel developed SGS-specific costs to implement each of the 11 SAMAs. The NRC staff requested more information on the process PSEG used to develop the 12 SAMA cost estimates (NRC 2010a). PSEG responded to the RAI by explaining that the cost 13 estimates were developed in a series of meetings involving personnel responsible for 14 development of the SAMA analysis and the two PSEG license renewal site leads who are 15 engineering managers each having over 25 years of plant experience, including project 16 management, operations, plant engineering, design engineering, procedure support, simulators, 17 and training (PSEG 2010a). During these meetings, each SAMA was validated against the 18 plant configuration, a budget-level estimate of its implementation cost was developed, and, in 19 some instances, lower cost approaches that would achieve the same objective were developed.

20 The SAMA implementation costs were then reviewed by the Design Engineering Manager for 21 both technical and cost perspectives and revised accordingly. PSEG further explained that 22 seven general cost categories were used in development of the budget-level cost estimates:

23 engineering, material, installation, licensing, critical path impact, simulator modification, and 24 procedures and training. For costs that could be shared between the two SGS units, the total 25 estimated cost was evenly divided between the two units to develop a per unit cost. Based on 26 the use of personnel having significant nuclear plant engineering and operating experience, the 27 NRC staff considers the process PSEG used to develop budget-level cost estimates 28 reasonable.

29 In response to an RAI requesting a more detailed description of the changes associated with 30 SAMAs 3, 5, 8, 13, 20, and 23, PSEG provided additional information detailing the analysis and 31 plant modifications included in the cost estimate of each improvement (PSEG 2010a). The staff 32 reviewed the costs and found them to be reasonable, and generally consistent with estimates 33 provided in support of other plants' analyses.

34 The NRC staff also noted that the ER reported an implementation cost for SAMA 3, "Install 35 Limited EDG Cross-Tie Capability Between SGS 1 and 2," of $4.175M in Section E.6.3 and 36

$525K in Section E.5-3 and requested clarification on which was the correct value (NRC September 2010 F-37 Draft NUREG-1437, Supplement 45

Appendix F 1

2010a). SEG responded that $4.175K was the correct value and stated that this value was 2

used in the SAMA evaluation (PSEG 201 Oa).

3 The NRC staff requested PSEG provide justification for the differences in the cost estimates for 4

SAMA 1, "Enhance Procedures and Provide Additional Equipment to Respond to Loss of 5

Control Area Ventilation," having a cost of $475K, and SAMA 17, "Enhance Procedures and 6

Provide Additional Equipment to Respond to Loss of Emergency Diesel Generator (EDG) 7 Control Room Ventilation," having a cost of $200K, which are similar in that each involves 8

opening doors to provide ventilation and using portable fans to enhance natural circulation 9

(NRC 2010a). In response to the RAI, PSEG stated that SAMA I has a higher cost because it 10 is a more complicated modification involving three rooms having differing requirements while 11 SAMA 17 involves four rooms that are basically identical (PSEG 2010a). The NRC staff 12 considers the basis for the differences in cost estimates reasonable.

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

14 and SAMA 22, "Install Fire Barriers between the lCCl, 1CC2, and lCC3 Consoles in the CRE,"

15 are similar in that each involves installing fire barriers to prevent the propagation of a fire 16 between cabinets and requested an explanation for why the estimated cost of $3.23M for SAMA 17 21 to modify 48 cabinets is similar to the estimated cost of $1.6M for SAMA 22 to modify just 18 three consoles (NRC 2010a). PSEG responded that the cost per console ($400K) in SAMA 22, 19 is much higher than the cost per cabinet ($35K - $70K) in SAMA 21 because making the 20 modifications to the Control Room consoles is more complicated than making the modifications 21 to the Relay Room cabinets (PSEG 2010a). Specifically, SAMA 22 requires making ventilation 22 modifications due to the significant heat loads in addition to adding fire barrier materials. The 23 NRC staff considers the basis for the differences in cost estimates reasonable.

24 The NRC asked PSEG to justify the estimated cost of $1 00K for SAMA 10, "Provide Procedural 25 Guidance for Faster Cooldown Loss of RCP Seal Cooling," and SAMA 11, "Modify Plant 26 Procedures to Make use of Other Unit's Positive Displacement Pump (PDP) for RCP Seal 27 Cooling," in light of the statement made in the ER that the minimum expected implementation 28 cost is assumed to be a procedure change at $50K at $1 00K for the site (NRC 2010a). In 29 response to the RAI, PSEG explained that the cost for SAMA 10 includes 1) $50K to perform a 30 feasibility study to confirm that there is no technical basis preventing implementation of a more 31 rapid cooldown on loss of RCP seal cooling and 2) $150K to revise the emergency operating 32 procedures (EOPs), which are more expensive to revise and require more extensive training 33 than other plant procedures (PSEG 2010a). PSEG also explained that the cost for SAMA 11 34 includes 1) $50K to perform a feasibility study to confirm that there is no technical basis 35 preventing PDP cross-tie when RCP seal cooling is lost, 2) $50K to revise the plant procedures, 36 and 3) $50K for each unit to involve plant licensing staff. The total of $200K for both SAMAs is 37 divided evenly between the two units. The NRC staff considers the bases for the estimated 38 costs for these SAMAs reasonable.

39 September 2010 F-38 Draft NUREG-1437, Supplement 45

Appendix F 1

2 The NRC staff concludes that the cost estimates provided by PSEG are sufficient and 3

appropriate for use in the SAMA evaluation.

4 5

6 F.6 Cost-Benefit Comparison 7

8 PSEG's cost-benefit analysis and the NRC staffs review are described in the following sections.

9 10 F.6.1 PSEG's Evaluation 11 12 The methodology used by PSEG was based primarily on NRC's guidance for performing cost-13 benefit analysis, i.e., NUREG/BR-01 84, Regulatory Analysis Technical Evaluation Handbook 14 (NRC 1997a). The guidance involves determining the net value for each SAMA according to 15 the following formula:

16 17 Net Value = (APE + AOC + AOE + AOSC) - COE, where 18 19 APE

= present value of averted public exposure ($)

20 AOC

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

21 AOE

= present value of averted occupational exposure costs ($)

22 AOSC = present value of averted onsite costs ($)

23 COE

= cost of enhancement ($)

24 25 If the net value of a SAMA is negative, the cost of implementing the SAMA is larger than the 26 benefit associated with the SAMA and it is not considered cost-beneficial. PSEG's derivation of 27 each of the associated costs is summarized below.

28 29 NUREG/BR-0058 has recently been revised to reflect the agency's policy on discount rates.

30 Revision 4 of NUREG/BR-0058 states that two sets of estimates should be developed, one at 31 3 percent and one at 7 percent (NRC 2004). PSEG provided a base set of results using the 3 32 percent discount rate and a sensitivity study using the 7 percent discount rate (PSEG 2009).

33 34 Averted Public Exposure (APE) Costs 35 36 The APE costs were calculated using the following formula:

37 38 APE = Annual reduction in public exposure (Aperson-rem/year) 39 x monetary equivalent of unit dose ($2,000 per person-rem) 40 x present value conversion factor (15.04 based on a 20-year period with a 41 3-percent discount rate)

September 2010 F-39 Draft NUREG-1437, Supplement 45

Appendix F 1

As stated in NUREG/BR-0184 (NRC 1997a), it is important to note that the monetary value of 2

the public health risk after discounting does not represent the expected reduction in public 3

health risk due to a single accident. Rather, it is the present value of a stream of potential 4

losses extending over the remaining lifetime (in this case, the renewal period) of the facility.

5 Thus, it reflects the expected annual loss due to a single accident, the possibility that such an 6

accident could occur at any time over the renewal period, and the effect of discounting these 7

potential future losses to present value. For the purposes of initial screening, which assumes 8

elimination of all severe accidents, PSEG calculated an APE of approximately $2,350,000 for 9

the 20-year license renewal period (PSEG 2009).

10 11 1.1.1.1 Averted Offsite Property Damage Costs (AOC) 12 13 The AOCs were calculated using the following formula:

14 15 AOC = Annual CDF reduction 16 x offsite economic costs associated with a severe accident (on a per-event basis) 17 x present value conversion factor.

18 This term represents the sum of the frequency-weighted offsite economic costs for each release 19 category, as obtained for the Level 3 risk analysis. For the purposes of initial screening, which 20 assumes elimination of all severe accidents caused by internal events, PSEG calculated an 21 AOC of about $306,000 based on the Level 3 risk analysis. This results in a discounted value of 22 approximately $4,600,000 for the 20-year license renewal period.

23 24 1.1.1.2 Averted Occupational Exposure (AOE) Costs 25 26 The AOE costs were calculated using the following formula:

27 28 AOE = Annual CDF reduction 29 x occupational exposure per core damage event 30 x monetary equivalent of unit dose 31 x present Value conversion factor 32 PSEG derived the values for averted occupational exposure from information provided in 33 Section 5.7.3 of the regulatory analysis handbook (NRC 1997a). Best estimate values provided 34 for immediate occupational dose (3,300 person-rem) and long-term occupational dose (20,000 35 person-rem over a 10-year cleanup period) were used. The present value of these doses was 36 calculated using the equations provided in the handbook in conjunction with a monetary 37

'equivalent of unit dose of $2,000 per person-rem, a real discount rate of 3 percent, and a time 38 period of 20 years to represent the license renewal period. For the purposes of initial screening, 39 which assumes elimination of all severe accidents caused by internal events, PSEG calculated 40 an AOE of approximately $31,000 for the 20-year license renewal period (PSEG 2009).

41 42 Averted Onsite Costs September 2010 F-40 Draft NUREG-1437, Supplement 45

Appendix F 1

2 Averted onsite costs (AOSC) include averted cleanup and decontamination costs and averted 3

power replacement costs. Repair and refurbishment costs are considered for recoverable 4

accidents only and not for severe accidents. PSEG derived the values for AOSC based on 5

information provided in Section 5.7.6 of NUREG/BR-0184, the regulatory analysis handbook 6

(NRC 1997a).

7 8

PSEG divided this cost element into two parts - the onsite cleanup and decontamination cost, 9

also commonly referred to as averted cleanup and decontamination costs (ACC), and the 10 replacement power cost (RPC).

11 12 ACCs were calculated using the following formula:

13 14 ACC

= Annual CDF reduction 15 x present value of cleanup costs per core damage event i6 x present value conversion factor 17 The total cost of cleanup and decontamination subsequent to a severe accident is estimated in 18 NUREG/BR-0184 to be $1.5 x 109 (undiscounted). This value was converted to present costs 19 over a 10-year cleanup period and integrated over the term of the proposed license extension.

20 For the purposes of initial screening, which assumes elimination of all severe accidents caused 21 by internal events, PSEG calculated an ACC of approximately $965,000 for the 20-year license 22 renewal period.

23 24 Long-term RPCs were calculated using the following formula:

25 26 RPC

= Annual CDF reduction 27 x present value of replacement power for a single event 28 x factor to account for remaining service years for which replacement power is 29 required 30 x reactor power scaling factor 31 32 PSEG based its calculations on a SGS net output of 1115 megawatt electric (MWe) and scaled 33 up from the 910 MWe reference plant in NUREG/BR-0184 (NRC 1997a). Therefore PSEG 34 applied a power scaling factor of 1115/910 to determine the replacement power costs. For the 35 purposes of initial screening, which assumes elimination of all severe accidents caused by 36 internal events, PSEG calculated an RPC of approximately $335,000 and an AOSC of 37 approximately $1,300,000 for the 20-year license renewal period.

38 39 Using the above equations, PSEG estimated the total present dollar value equivalent associated 40 with completely eliminating severe accidents from internal events at SGS to be about $8.28M.

41 Use of a multiplier of 2 to account for external events increases the value to $16.56M and 42 represents the dollar value associated with completely eliminating all internal and external event 43 severe accident risk for a single unit at SGS, also referred to as the Single Unit Maximum 44 Averted Cost Risk (MACR).

September 2010 F-41 Draft NUREG-1437, Supplement 45

Appendix F 1

2 1.1.1,3 PSEG's Results 3

4 If the implementation costs for a candidate SAMA exceeded the calculated benefit, the SAMA 5

was considered not to be cost-beneficial. In the baseline analysis contained in the ER (using a 6

3 percent discount rate and considering the impact of external events), PSEG identified 11 7

potentially cost-beneficial SAMAs. PSEG performed additional analyses to evaluate the impact 8

of parameter choices (alternative discount rates and variations in MACCS2 input parameters) 9 and uncertainties on the results of the SAMA assessment and, as a result of this analysis, 10 identified five additional potentially cost-beneficial SAMAs. PSEG also performed an analysis 11 on a less costly alternative to SAMA 5 (SAMA 5A) and found it to be potentially cost-beneficial.

12 13 The potentially cost-beneficial SAMAs for SGS are the following:

14 15 0

SAMA 1 - Enhance Procedures and Provide Additional Equipment to Respond to Loss 16 of Control Area Ventilation 17 SAMA 2 - Re-configure Salem 3 to Provide a More Expedient Backup AC Power Source 18 for Salem 1 and 2 19 SAMA 3 - Install Limited EDG Cross-tie Capability Between Salem 1 and 2 20 0

SAMA 4 - Install Fuel Oil Transfer Pump on "C" EDG & Provide Procedural Guidance for 21 Using "C" EDG to Power Selected "A" and "B" Loads 22 0

SAMA 5 - Install Portable Diesel Generators to Charge Station Battery and Circulating 23 Water Batteries & Replace PDP with Air-Cooled Pump 24 0

SAMA 5A - Install Portable Diesel Generators to Charge Station Battery and Circulating 25 Water Batteries 26 0

SAMA 6 - Enhance Flood Detection for 84' Aux Building and Enhance Procedural 27 Guidance for Responding to Service Water Flooding 28 0

SAMA 7 - Install "B" Train AFWST Makeup Including Alternate Water Source 29 0

SAMA 8 - Install High Pressure Pump Powered with Portable Diesel Generator and 30 Long-term Suction Source to Supply the AFW Header 31 0

SAMA 9 - Connect Hope Creek Cooling Tower Basin to Salem Service Water System 32 as Alternate Service Water Supply September 2010 F-42 Draft NUREG-1437, Supplement 45

Appendix F 1

0 SAMA 10 - Provide Procedural Guidance for Faster Cooldown on Loss of RCP Seal 2

Cooling 3

0 SAMA 11 - Modify Plant Procedures to Make Use of Other Unit's PDP for RCP Seal 4

Cooling 5

SAMA 12 - Improve Flood Barriers Outside of 220/440VAC Switchgear Rooms 6

SAMA 14 - Expand AMSAC Function to Include Backup Breaker Trip on RPS Failure 7

0 SAMA 17 - Enhance Procedures and Provide Additional Equipment to Respond to Loss 8

of EDG Control Room Ventilation 9

a SAMA 24 - Provide Procedural Guidance to Cross-tie Salem 1 and 2 Service Water 10 Systems 11 0

SAMA 27 -In Addition to the Equipment Installed for SAMA 5, Install Permanently Piped 12 Seismically Qualified Connections to Alternate AFW Water Sources 13 PSEG indicated that they plan to further evaluate these SAMAs for possible implementation 14 using existing action-tracking and design change processes (PSEG 2009).

15 16 The potentially cost-beneficial SAMAs, and PSEG's plans for further evaluation of these 17 SAMAs, are discussed in detail in Section F.6.2.

18 19 F.6.2 Review of PSEG's Cost-Benefit Evaluation 20 The cost-benefit analysis performed by PSEG was based primarily on NUREG/BR-0184 21 (NRC 1997a) and discount rate guidelines in NUREG/BR-0058 (NRC 2004) and was executed 22 consistent with this guidance.

23 SAMAs identified primarily on the basis of the internal events analysis could provide benefits in 24 certain external events, in addition to their benefits in internal events. To account for the 25 additional benefits in external events, PSEG multiplied the internal event benefits for all but one 26 internal event SAMA (SAMA 20, discussed further below) by a factor of 2, which is 27 approximately the ratio of the total CDF from internal and external events to the internal event 28 CDF (PSEG 2009). As discussed in Section F.2.2, this factor was based on a seismic CDF of 29 9.5 x 10-6 per year, plus a fire CDF of 3.8 x 10- per year, plus the screening values for high 30 winds, external flooding, transportation, detritus, and chemical release events (1 x 10-' per year 31 for each). The external event CDF of 5.3 x 10-5 per year is thus about 110 percent of the 32 internal events CDF used in the SAMA analysis (5.0 x 10-5 per year). The total CDF is 2.1 times 33 the internal events CDF and this was rounded to 2. Eleven SAMAs were determined to be cost-September 2010 F-43 Draft NUREG-1437, Supplement 45

Appendix F 1

beneficial in PSEG's analysis (SAMAs 1, 2, 4, 6, 9, 10, 11, 12, 14, 17, and 24 as described 2

above).

3 PSEG did not multiply the internal event benefits by the factor of 2 for three SAMAs that 4

specifically address fire risk (SAMAs 21, 22, and 23). Doubling the internal event estimate for 5

SAMAs 21, 22, and 23 would not be appropriate because these SAMAs are specific to fire risks 6

and would not have a corresponding benefit on the risk from internal events.

7 For all but one internal event SAMA also having benefits in fire and seismic risk (i.e., SAMAs 1, 8

and 8 for fire and SAMAs 5, 5A, and 27 for seismic), PSEG separately quantified the benefits for 9

fire and seismic events and added these results to the benefits from internal events and external 10 events developed from applying the factor of 2 (as discussed in Section F.4 above). The NRC 11 staff noted that this process appeared to be double counting the benefits from external events 12 and requested clarification (NRC 2010a). In response to the RAI, PSEG acknowledged that this 13 process results in "double counting" of some external event contributions to the total averted 14 cost risk and stated that this approach was retained, unless it resulted in a gross 15 misrepresentation of a SAMA's benefit, in order to avoid underestimating the external events 16 averted cost risk (PSEG 2010a). PSEG further concluded that this process does not impact the 17 conclusions of the SAMA analysis. Since the process that PSEG used over-estimates the 18 benefits from external events and therefore results in conservative estimates of the SAMA 19 benefits, the NRC staff considers the process PSEG used acceptable for the SAMA evaluation.

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

21 PSEG multiplied the estimated benefits for internal events by a factor of 2.0 to account for 22 external events in the Phase I analysis. In the Phase II analysis, PSEG separately quantified 23 the internal event, fire event, and seismic event benefits, as described in Section F.4 above, and 24 to account for the additional benefits in other (non-fire/non-seismic) external events, PSEG 25 multiplied the internal event benefits by a factor of 1.1, which is the ratio of the total CDF from 26 internal and other external events to the internal event CDF (based on an HFO CDF of 5.0 x 10 27 6 per year and an internal events CDF of 5.0 x 10s per year used in the SAMA analysis). The 28 estimated SAMA benefits for internal events with the factor of 1.1 applied to account for other 29 external events, fire events, and seismic events were then summed to provide an overall 30 benefit. Since the methodology PSEG used accounts for both internal events and external 31 events, the NRC staff considers the methodology PSEG used for SAMA 20 acceptable for the 32 SAMA evaluation.

33 PSEG considered the impact that possible increases in benefits from analysis uncertainties 34 would have on the results of the SAMA assessment. In the ER, PSEG presents the results of 35 an uncertainty analysis of the internal events CDF which indicates that the 9 5 th percentile value 36 is a factor of 1.64 times the point estimate CDF for SGS. Since the one Phase I SAMA that was 37 screened based on qualitative criteria was screened due to not being applicable to SGS, a re-38 examination of the Phase I SAMAs based on the upper bound benefits was not necessary.

39 PSEG considered the impact on the Phase II screening if the estimated benefits were increased September 2010 F-44 Draft NUREG-1437, Supplement 45

Appendix F 1

by a factor of 1.64 (in addition to the multiplier of 2 for external events). Four additional SAMAs 2

became cost-beneficial in PSEG's analysis (SAMAs 5, 7, 8, and 27 as described above).

3 PSEG noted that the 9 5 th percentile value for CDF may be underestimated because uncertainty 4

distributions are not applied to all basic events in the SGS PRA model. Based on this, PSEG 5

used a factor of 2.5 times the point estimate CDF to represent the 9 5 th percentile value, which is 6

stated to be typical of most light water reactor CDF uncertainty analyses. PSEG considered the 7

impact on the Phase II screening if the estimated benefits were increased by a factor of 2.5 (in 8

addition to the multiplier of 2 for external events). One additional SAMA became cost-beneficial 9

(SAMA 3). The NRC staff notes that while the factor of 2.5 does not represent an upper bound, 10 it is typical of factors used in prior SAMA analyses, is higher than the factor calculated for other 11 Westinghouse 4-loop plants and used in prior SAMA analysis, and is therefore considered by 12 the NRC staff to be appropriate for use in the SAMA sensitivity analyses.

13 PSEG provided the results of additional sensitivity analyses in the ER, including use of a 7 14 percent discount rate and variations in MACCS2 input parameters. These analyses did not 15 identify any additional potentially cost-beneficial SAMAs (PSEG 2009).

16 The NRC staff noted that the ER reported that the licensed thermal power for SGS Unit 1 is 17 3,459 MWt, which equates to a net electrical output of 1,195 MWe when operating at 100 18 percent power, while 1,115 MWe was used to calculate long-term replacement power costs for 19 the SAMA analysis (NRC 2010a). In response to the RAI, PSEG clarified that 1,115 MWe used 20 in the SAMA analysis was incorrect and provided a revised replacement power cost estimate of 21

$359,000 using the correct 1,195 MWe, which is an approximately 7 percent increase over that 22 used in the SAMA analysis (PSEG 2010a). PSEG also provided a revised MACR of $16.61M, 23 which is an increase of about 0.3 percent over the MACR used in the SAMA analysis and 24 concluded that the small error would have a negligible impact on the conclusions of the SAMA 25 analysis. The NRC staff agrees with this assessment.

26 As indicated in Section F.3.2, in response to an NRC staff RAI, PSEG extended the review of 27 Level 1 and Level 2 basic events down to an RRW of 1.006, which equates to a benefit of about 28

$47,000, using SGS PRA MOR Revision 4.3 (PSEG 2010a). The review identified the following 29 three additional SAMAs associated with new basic events added to the importance lists: 1) 30 SAMA 30, "Automatic Start of Diesel-Powered Air Compressor," 2) SAMA 31, "Fully Automate 31 Swapover to Sump Recirculation," and 3) SAMA 32, "Enhance Flood Detection for 100-foot 32 Auxiliary Building and Enhance Procedural Guidance for Responding to Internal Floods." Each 33 of these new SAMAs is included in Table F-6. PSEG performed a Phase II evaluation using 34 results for SGS PRA MOR Revision 4.3 and the process described above. PSEG stated that 35 the release frequency for MOR Revision 4.3 is 2.2 x 10-5 per year, a decrease of over 50 36 percent from MOR Revision 4.1, and that the 9 5 th percentile value for CDF is a factor of 2.1 37 times the point estimate CDF. Based on information provided in the RAI response, the NRC 38 staff estimated, for the MOR Revision 4.3 results, the total present dollar value equivalent 39 associated with completely eliminating severe accidents from internal events at SGS to be September 2010 F-45 Draft NUREG-1437, Supplement 45

Appendix F 1

about $2.3M, a revised external event multiplier of about 3.4, and a revised MACR of about 2

$7.9M. These results represent a decrease of more than 50 percent compared to the SGS PRA 3

MOR 4.1 results reported in the ER. PSEG's analysis determined that none of the three SAMA 4

candidates was cost-beneficial in either the baseline analysis or the uncertainty analysis.

5 Based on these results for MOR Revision 4.3 and the changes in the importance lists, the NRC 6

staff asked PSEG to assess the impact on the SAMA evaluation of the PRA model changes 7

made since MOR Revision 4.1 (NRC 2010b). In response to the RAI, PSEG re-evaluated each 8

potentially cost-beneficial SAMA using MOR Revision 4.3 and determined that SAMA benefits 9

both increased (up to 42 percent) and decreased (up to 99 percent) from the results using MOR 10 Revision 4.1 and that five SAMA candidates (SAMA 3, 5, 11, 14, and 27) would no longer be 11 cost-beneficial (PSEG 201 Ob). PSEG also qualitatively evaluated each SAMA determined to 12 not be cost-beneficial and concluded that none would become cost-beneficial using MOR 13 Revision 4.3 based on the following:

14 0

The implementation cost is greater than the revised MACR even after accounting for 15 uncertainty (SAMA 13).

16 0

For SAMAs that address fire events only, the maximum averted cost risk for external 17 events decreased, which would result in a corresponding decrease in the maximum 18 calculated benefit for these SAMAs (SAMAs 21, 22, and 23).

19 0

The cost of implementation was sufficiently greater than the MOR Revision 4.1 benefit 20 that changes in MOR Revision 4.3 would not be expected to overcome the difference 21 (SAMAs 15, 16, 18, and 19). The NRC staff notes that this difference, even after 22 accounting for uncertainty, is on the order of 50 percent or more for all of these SAMAs 23 and agrees that it is unlikely that a revised evaluation would result in a change to the 24 cost-beneficial status for these SAMAs.

25 The cost of implementation is greater than the revised MACR (SAMA 20). The NRC 26 staff notes that MOR Revision 4.1 results indicate that the fire and seismic events 27 contributors to the MACR are four times the internal events contribution and, since the 28 maximum averted cost risk for external events has decreased with MOR Revision 4.3, 29 agrees that it is unlikely that a revised evaluation would result in a change to cost-30 beneficial status for this SAMA.

31 As indicated in Section F.3.2, the NRC staff asked the licensee to evaluate several potentially 32 lower cost alternatives to the SAMAs considered in the ER (NRC 2010a), as summarized below:

33 Operating the AFW AF1 1/21 valves closed in lieu of SAMA 8, "Install High Pressure 34 Pump Powered with Portable Diesel Generator and Long-term Suction Source to Supply 35 the AFW Header." In response to the RAI, PSEG stated that the AF1 1 valves on the 36 discharge side of the motor-driven AFW pumps are already operated closed, leaving September 2010 F-46 Draft NUREG-1437, Supplement 45

Appendix F 1

only the AF21 valves on the discharge side of the turbine-driven AFW pump operating 2

open (PSEG 201 Oa). Steam binding of the common suction line to all three AFW pumps 3

could therefore only occur as a result of high temperature water leaks through three 4

check valves in series on the discharge to the turbine-driven AFW pump. PSEG 5

concluded that the proposed improvement would not be feasible because 1) industry 6

data used to represent common-cause steam binding of all three AFW pumps appears 7

to be conservative relative to the SGS configuration, thereby overstating the risk 8

significance of this failure at SGS, 2) operating all of the AF1 1/21 valves closed could 9

actually provide a negative risk benefit based on a new failure event to represent 10 common-cause failure of the valves to open, and 3) operating all of the AF1 1/21 valves 11 closed could have a potentially adverse effect on the SGS design basis because design-12 basis calculations and assumptions would need to be modified to reflect this change in 13 AFW strategy.

14 Install improved fire barriers in the 460V switchgear rooms to provide separation 15 between the three power divisions in lieu of SAMA 20, "Fire Protection System to 16 Provide Make-up to RCS and Steam Generators." In response to the RAI, PSEG 17 explained that the configuration of Fire Area 1 FA-AB-84A, addressed by SAMA 20, is 18 significantly more complex than Fire Area 1 FA-AB-64A, addressed by SAMA 23, "Install 19 Fire Barriers and Cable Wrap to Maintain Divisional Separation in the 4160V AC 20 Switchgear Room" (PSEG 2010a). The SAMA 23 estimated implementation cost of 21

$975K only addresses the risk associated with preventing the spread of transient fires 22 between divisions and did not address the entire fire risk in the fire area, which would 23 include protecting the overhead cables. PSEG estimates that the cost of addressing the 24 entire fire risk in Fire Area 1 FA-AB-64A would be at least an order of magnitude greater 25 than the estimated implementation cost for SAMA 23. PSEG further estimates that the 26 cost of addressing the fire risk in Fire Area 1 FA-AB-84A could be double that for Fire 27 Area 1 FA-AB-64A. The maximum benefit of the proposed SAMA, which assumes 28 elimination of all fire risk for Fire Area 1 FA-AB-84A, is estimated to be $2.OM in the 29 baseline analysis, or $5.1M accounting for uncertainties, using the MOR Rev. 4.1 PRA 30 model. Furthermore, PSEG determined that the maximum benefit would be about 30 31 percent lower if the MOR Rev. 4.3 PRA model were used. Because the estimated 32 implementation cost is significantly greater than the maximum potential benefit, PSEG 33 concluded that the proposed SAMA would not be cost-beneficial.

34 Install improved fire barriers to provide separation between the AFW pumps in lieu of 35 SAMA 8, "Install High Pressure Pump Powered with Portable Diesel Generator and 36 Long-term Suction Source to Supply the AFW Header." In response to the RAI, PSEG 37 estimated the cost to implement the proposed SAMA to be $750K (PSEG 201 Oa).

38 Failure of multiple AFW pumps accounted for about 67 percent of the Fire Area 1 FA-AB-39 84B fire risk. The maximum benefit of the proposed SAMA, which assumes elimination 40 of all of this fire risk, is estimated to be $120K in the baseline analysis, or $290K 41 accounting for uncertainties, using the MOR Rev. 4.1 PRA model. Furthermore, PSEG September 2010 F-47 Draft NUREG-1437, Supplement 45

Appendix F I

determined that the maximum benefit would be about 30 percent lower if the MOR Rev.

2 4.3 PRA model were used. Because the estimated implementation cost is significantly 3

greater than the maximum potential benefit, PSEG concluded that the proposed SAMA 4

would not be cost-beneficial.

5 PSEG indicated that the 17 potentially cost-beneficial SAMAs (SAMAs 1, 2, 3, 4, 5, 5A, 6, 7, 8, 6

9, 10, 11, 12, 14, 17, 24, and 27) will be considered for implementation through the established 7

Salem Plant Health Committee (PHC) process (PSEG 2009). In response to an NRC staff RAI, 8

PSEG described the PHC as being chaired by the Plant Manager and includes as members the 9

Plant Engineering Manager and the Directors of Operations, Engineering, Maintenance, and 10 Work Management (PSEG 2010a). The PHC is chartered with reviewing issues that require 11 special plant management attention to ensure effective resolution and, with respect to each of 12 the potentially cost-beneficial SAMAs, will decide on one of the following courses of actions: 1) 13 approve for implementation, 2) conditionally approved for implementation pending the results of 14 requested evaluations, 3) not approved for implementation, or 4) tabled until additional 15 information needed to make a final decision is provided to the PHC. Additional information 16 requested may include 1) making corrections to the original SAMA analysis, 2) examining an 17 alternate solution, 3) performing sensitivity studies to determine the effect of implementing a 18 sub-set of SAMAs, already approved SAMAs, or already approved non-SAMA design changes 19 on the SAMA, or 4) coordinating the SAMA with related Mitigating System Performance Index 20 (MSPI) margin recovery activities. If approved or conditionally approved for implementation, 21 the SAMA will be ranked with respect to priority and assigned target years for implementation.

22 The NRC staff concludes that, with the exception of the potentially cost-beneficial SAMAs 23 discussed above, the costs of the other SAMAs evaluated would be higher than the associated 24 benefits.

25 F.7 Conclusions 26 PSEG compiled a list of 27 SAMAs based on a review of: the most significant basic events from 27 the plant-specific PRA and insights from the SGS PRA group, insights from the plant-specific 28 IPE and IPEEE, Phase II SAMAs from license renewal applications for other plants, and the 29 generic SAMA candidates from NEI 05-01. A qualitative screening removed SAMA candidates 30 that: (1) are not applicable to SGS due to design differences, (2) have already been 31 implemented at SGS, (3) would achieve results that have already been achieved at SGS by 32 other means, and (4) have estimated implementation costs that would exceed the dollar value 33 associated with completely eliminating all severe accident risk at SGS. Based on this 34 screening, 2 SAMAs were eliminated leaving 25 candidate SAMAs for evaluation. One 35 additional SAMA candidate was identified and evaluated as a sensitivity case. Three additional 36 SAMA candidates were identified and evaluated in response to an NRC staff RAI.

37 For the remaining SAMA candidates, including the sensitivity case SAMA and three SAMAs 38 added in response to the NRC staff RAI, a more detailed design and cost estimate were September 2010 F-48 Draft NUREG-1437, Supplement 45

Appendix F 1

developed as shown in Table F-6. The cost-benefit analyses in the ER and RAI response 2

showed that 11 of the SAMA candidates were potentially cost-beneficial in the baseline analysis 3

(Phase II SAMAs 1, 2, 4, 6, 9, 10, 11, 12, 14, 17, and 24). PSEG performed additional analyses 4

to evaluate the impact of parameter choices and uncertainties on the results of the SAMA 5

assessment. As a result, five additional SAMA candidates (SAMA 3, 5, 7, 8, and 27) were 6

identified as potentially cost-beneficial in the ER. The ER also showed that the sensitivity case 7

SAMA (SAMA 5A) was potentially cost-beneficial. PSEG has indicated that all 17 potentially 8

cost-beneficial SAMAs will be considered for implementation through the established Salem 9

Plant Health Committee process.

10 The NRC staff reviewed the PSEG analysis and concludes that the methods used and the 11 implementation of those methods was sound. The treatment of SAMA benefits and costs 12 support the general conclusion that the SAMA evaluations performed by PSEG are reasonable 13 and sufficient for the license renewal submittal. Although the treatment of SAMAs for external 14 events was somewhat limited, the likelihood of there being cost-beneficial enhancements in this 15 area was minimized by improvements that have been realized as a result of the IPEEE process, 16 and inclusion of a multiplier to account for external events.

17 The NRC staff concurs with PSEG's identification of areas in which risk can be further reduced 18 in a cost-beneficial manner through the implementation of the identified, potentially cost-19 beneficial SAMAs. Given the potential for cost-beneficial risk reduction, the NRC staff agrees 20 that further evaluation of these SAMAs by PSEG is warranted. However, these SAMAs do not 21 relate to adequately managing the effects of aging during the period of extended operation.

22 Therefore, they need not be implemented as part of license renewal pursuant to Title 10 of the 23 Code of Federal Regulations, Part 54.

24 F.8 References 25 American Society of Mechanical Engineers (ASME). 2005. "Addenda to ASME RA-S-2002, 26 Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications." ASME RA-27 Sb-2005, December 2005.

28 BEA (Bureau of Economic Analysis). 2008. Regional Economic Accounts, accessed June 20 at 29 http://www.bea.gov/reqional/reis/.

30 Electric Power Research Institute (EPRI). 1989. "Probabilistic Seismic Hazard Evaluations at 31 Nuclear Plant Sites in the Central and Eastern United States; Resolution of the Charleston 32 Earthquake Issues." EPRI NP-6395-D, EPRI Project P101-53. Palo Alto, CA. April 1989.

33 Electric Power Research Institute (EPRI). 1991. "A Methodology for Assessment of Nuclear 34 Power Plant Seismic Margin," Implementation Guide NP-6041, Revision 1. Palo Alto, CA.

35 August 1991.

September 2010 F-49 Draft NUREG-1437, Supplement 45

Appendix F 1

Electric Power Research Institute (EPRI). 1993. "Fire Induced Vulnerability Evaluation (FIVE) 2 Methodology." TR-1 00370, Revision 1, Palo Alto, CA. September 19, 1993.

3 KLD Associates, Inc. (KLD). 2004. Salem / Hope Creek Nuclear Generating Stations 4

Development of Evacuation Time Estimates. KLD TR-356. February 2004.

5 6

Nuclear Energy Institute (NEI). 2005. "Severe Accident Mitigation Alternative (SAMA) Analysis 7

Guidance Document." NEI 05-01 (Rev. A), Washington, D.C. November 2005.

8 Nuclear Energy Institute (NEI). 2007. "Process for Performing Follow-on PRA Peer Reviews 9

using the ASME PRA Standard (Internal Events)." NEI 05-04, Rev. 1, Washington, D.C.

10 December 2007.

11 Public Service Electric and Gas Company (PSEG). 1993. Letter from Stanley LaBruna, PSEG, 12 to NRC Document Control Desk.

Subject:

"Generic Letter 88-20; Individual Plant Examination 13 (IPE) Report, Salem Generating Station, Unit Nos. 1 and 2, Docket Nos. 50-272 and 50-311,"

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

15 Public Service Electric and Gas Company (PSEG). 1995. Letter from E. Simpson, PSEG, to 16 NRC Document Control Desk.

Subject:

"Response to Generic Letter 88-20 Individual Plant 17 Examination for Sever Accident Vulnerabilities - 1 OCFR50.54 (f) Request for Additional 18 Information Salem Generating Station, Unit Nos. 1 and 2 Facility Operating License Nos. DRR-19 70 and DPR-75 Docket Nos. 50-272 and 50-311," Hancocks Bridge, New Jersey. August 01, 20 1995. Accessible at ML080100021.

21 Public Service Electric and Gas Company (PSEG). 1996. Letter from E. C. Simpson, PSEG, to 22 NRC Document Control Desk.

Subject:

"Response to Generic Letter No. 88-20, Supplement 4, 23 Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilities, 24 Salem Generating Station Units Nos. 1 and 2, Facility Operating License Nos. DPR-70 and 25 DPR-75, Docket Nos. 50-272 and 50-311," Hancocks Bridge, New Jersey. January 29, 1996.

26 Accessible at ML080100023.

27 PSEG Nuclear, LLC (PSEG). 2009. Salem Nuclear Generating Station --- License Renewal 28 Application, Appendix E: Applicant's Environmental Report; Operating License Renewal Stage.

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

30 PSEG Nuclear, LLC (PSEG). 2010a. Letter from Paul. J. Davison, PSEG, to NRC Document 31 Control Desk.

Subject:

"Response to NRC Request for Additional Information dated April 12, 32 2010, related to the Severe Accident Mitigation Alternatives (SAMA) review of the Salem 33 Nuclear Generating Station, Units 1 and2," Hancocks Bridge, New Jersey. May 24, 2010.

34 Accessible at ML101520326.

35 PSEG Nuclear, LLC (PSEG). 2010b. Letter from Christine T. Neely, PSEG, to NRC Document 36 Control Desk.

Subject:

"Supplement to RAI responses submitted in PSEG Letter LR-N10-0164 September 2010 F-50 Draft NUREG-1437, Supplement 45

Appendix F 1

dated May 24, 2010, related to the Severe Accident Mitigation Alternatives (SAMA) review of 2

the Salem Nuclear Generating Station, Units 1 and 2," Hancocks Bridge, New Jersey. August 3

18, 2010. Accessible at ML102320211.

4 U.S. Department of Agriculture (USDA). 2004. 2002 Census of Agriculture - Volume 1, 5

Geographic Area Series, Part 8 (Delaware), Part 20 (Maryland), Part 30.

6 U.S. Nuclear Regulatory Commission (NRC). 1988. Generic Letter 88-20, "Individual Plant 7

Examination for Severe Accident Vulnerabilities." November 23, 1988. Washington, D.C.

8 U.S. Nuclear Regulatory Commission (NRC). 1989. Fire Risk Scoping Study. NUREG/CR-9 5088. January 1989. Washington, D.C.

10 U.S. Nuclear Regulatory Commission (NRC). 1990. Severe Accident Risks: An Assessment 11 for Five U.S. Nuclear Power Plants. NUREG-1150. Washington, D.C.

12 U.S. Nuclear Regulatory Commission (NRC). 1991 a. Generic Letter No. 88-20, Supplement 4, 13 "Individual Plant Examination of External Events for Severe Accident Vulnerabilities." NUREG-14 1407. Washington, D.C. June 28,1991.

15 U.S. Nuclear Regulatory Commission (NRC). 1991b. "Procedural and Submittal Guidance for 16 the Individual Plant Examination of External Events (IPEEE) for Severe Accident 17 Vulnerabilities." NUREG-1407. Washington, D.C. June 1991.

18 U.S. Nuclear Regulatory Commission (NRC). 1994. Revised Livermore Seismic Hazard 19 Estimates for Sixty-Nine Nuclear Plant Sites East of the Rocky Mountains. NUREG-1488, April 20 1994. Washington, D.C.

21 U.S. Nuclear Regulatory Commission (NRC). 1996. Letter from Leonard Olshan, U.S. NRC, to 22 Leon Eliason, PSEG.

Subject:

Individual Plant Examination (IPE) Submittal - Internal Events, 23 Salem Nuclear Generating Station, Units 1 and 2 (TAC No. M74461). March 21, 1996.

24 U.S. Nuclear Regulatory Commission (NRC). 1997a. Regulatory Analysis Technical Evaluation 25 Handbook. NUREG/BR-0184. Washington, D.C.

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

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

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

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

30 U.S. Nuclear Regulatory Commission (NRC). 1999. Letter from Patrick D. Milano, U.S. NRC to 31 Harold W. Keiser, PSEG.

Subject:

Generic Letter 88-20, Supplement 4, "Individual Plant September 2010 F-51 Draft NUREG-1437, Supplement 45

Appendix F 1

Examination for External Events for Severe Accident Vulnerabilities," Salem Nuclear Generating 2

Station, Unit Nos. 1 and 2 (TAC Nos. M83669 and M83670). May 21, 1999.

3 U.S. Nuclear Regulatory Commission (NRC). 2003. SECPOP2000: Sector Population, Land 4

Fraction, and Economic Estimation Program. NUREG/CR-6525, Rev. 1. Sandia National 5

Laboratories. August 2003.

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

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

8 U.S. Nuclear Regulatory Commission (NRC). 2007. "An Approach for Determining the 9

Technical Adequacy of Probabilistic Risk Assessment Results for Risk-Informed Activities."

10 Regulatory Guide 1.200, Revision 1. January 2007.

11 U.S. Nuclear Regulatory Commission (NRC). 2010a. Letter from Charles Eccleston, U.S. NRC, 12 to Thomas Joyce, PSEG.

Subject:

Request for Additional Information, Regarding Severe 13 Accident Mitigation Alternatives for the Salem Nuclear Generating Station, Units 1 and 2. April 14 12, 2010. Accessible at ML100910252.

15 U.S. Nuclear Regulatory Commission (NRC). 2010b. Teleconference held on July 29, 2010 16 between NRC staff and PSEG Nuclear, LLC staff regarding clarifications to the RAI responses 17 provided by PSEG via letter dated April 12, 2010. Accessible at ML102220012.

18 September 2010 F-52 Draft NUREG-1437, Supplement 45

1 Appendix G 2

U.S. Nuclear Regulatory Commission Staff Evaluation of 3

Severe Accident Mitigation Alternatives for 4

Hope Creek Nuclear Generating Station 5

In Support of License Renewal Application Review 6

I This Page Intentionally Left Blank

Appendix G 1

G. U.S. Nuclear Regulatory Commission Staff Evaluation of Severe 2

Accident Mitigation Alternatives for Hope Creek Nuclear Generating 3

Station in Support of License Renewal Application Review 4

G.1 Introduction 5

PSEG Nuclear, LLC, (PSEG) submitted an assessment of severe accident mitigation 6

alternatives (SAMAs) for the Hope Creek Generating Station (HCGS) as part of the 7

environmental report (ER) (PSEG 2009). This assessment was based on the most recent 8

HCGS probabilistic risk assessment (PRA) available at that time, a plant-specific offsite 9

consequence analysis performed using the MELCOR Accident Consequence Code System 2 10 (MACCS2) computer code, and insights from the HCGS individual plant examination (IPE) 11 (PSEG 1994) and individual plant examination of external events (IPEEE) (PSEG 1997). In 12 identifying and evaluating potential SAMAs, PSEG considered SAMAs that addressed the major 13 contributors to core damage frequency (CDF) and release frequency at HCGS, as well as 14 SAMA candidates for other operating plants that have submitted license renewal applications.

15 PSEG initially identified 23 potential SAMAs. This list was reduced to 21 unique SAMA 16 candidates by eliminating SAMAs that are not applicable to HCGS due to design differences, 17 have already been implemented at HCGS, would achieve the same risk reduction results that 18 had already been achieved at HCGS by other means, have excessive implementation cost or 19 could be combined with another SAMA candidate. PSEG assessed the costs and benefits 20 associated with each of the potential SAMAs, and concluded in the ER that several of the 21 candidate SAMAs evaluated are potentially cost-beneficial.

22 Based on a review of the SAMA assessment, the U.S. Nuclear Regulatory Commission (NRC) 23 issued a request for additional information (RAI) to PSEG by letter dated May 20, 2010 (NRC 24 201 Oa) and, based on a review of the RAI responses, a request for RAI response clarification by 25 teleconference dated July 29, 2010 (NRC 2010b). Key questions concerned: discussing 26 internal and external review comments on the PRA model, including the impact of the 2008 PRA 27 peer review comments on the SAMA analysis results; the process and criteria used to assign 28 containment event tree (CET) end states to release categories; additional details on the seismic 29 analysis; the SAMA screening process and additional potential SAMAs not previously 30 considered; and further information on the costs and benefits of several specific candidate 31 SAMAs and low cost alternatives. PSEG submitted additional information by a letters dated 32 June 1, 2010 (PSEG 2010a) and August 18, 2010 (PSEG 2010b). In the responses, PSEG 33 provided: a listing of open gaps and findings from the 2008 PRA peer review and an 34 assessment of their impact on the SAMA analysis; additional description of how CET end states 35 were assigned to release categories and how representative sequences were selected for each 36 release category; clarification of certain elements of the seismic analysis and an assessment of 37 the impact of seismic assumptions on the external events multiplier; analyses of additional 38 SAMAs; and additional information regarding several specific SAMAs. PSEG's responses September 2010 G-1 Draft NUREG-1437, Supplement 45

Appendix G 1

addressed the NRC staffs concerns, and resulted in the identification of additional potentially 2

cost-beneficial SAMAs.

3 An assessment of SAMAs for HCGS is presented below.

4 5

G.2 Estimate of Risk for HCGS 6

7 PSEG's estimates of offsite risk at HCGS are summarized in Section G.2.1. The summary is 8

followed by the NRC staff's review of PSEG's risk estimates in Section G.2.2.

9 G.2.1 PSEG's Risk Estimates 10 11 Two distinct analyses are combined to form the basis for the risk estimates used in the SAMA 12 analysis: (1) the HCGS Level I and Level 2 PRA model, which is an updated version of the IPE 13 (PSEG 1994), and (2) a supplemental analysis of offsite consequences and economic impacts 14 (essentially a Level 3 PRA model) developed specifically for the SAMA analysis. The SAMA 15 analysis is based on the most recent HCGS Level 1 and Level 2 PRA model available at the 16 time of the ER, referred to as the HC108B update. The scope of this HCGS PRA does not 17 include external events.

18 The HCGS CDF is approximately 5.1 x 10-6 per year as determined from quantification of the 19 Level 1 PRA model at a truncation of 1 x 10-12 per year. When determining the frequency of the 20 source term categories from the sum of the containment event tree (CET) sequences, or Level 2 21 PRA model, a higher truncation of 5 x 1011 per year was used and the resulting release 22 frequency (from all release categories, which consist of intact containment, late release, and 23 early release) is approximately 4.4 x 10.6 per year. The latter value was used as the baseline 24 CDF in the SAMA evaluations (PSEG 2009). The CDF is based on the risk assessment for 25 internally-initiated events, which includes internal flooding. PSEG did not explicitly include the 26 contribution from external events within the HCGS risk estimates; however, it did account for the 27 potential risk reduction benefits associated with external events by multiplying the estimated 28 benefits for internal events by a factor of 6.3. This is discussed further in Sections G.2.2 and 29 G.6.2.

30 The breakdown of CDF by initiating event is provided in Table G-1 (PSEG 2009). As shown in 31 this table, events initiated by loss of offsite power, loss of service water and other transients 32 (manual shutdown and turbine trip with bypass) are the dominant contributors to the CDF.

33 Anticipated transient without scram (ATWS) sequences account for 3% of the CDF, station 34 blackout accounts for 12% of the CDF (PSEG 201 Oa).

35 Table G-1. HCGS Core Damage Frequency for Internal Events CDF

% Contribution Initiating Event (per year) to CDF1 Loss of Offsite Power 9.3 x 10-7 18 Loss of Service Water (SW) 8.1 x 10-7 15 Manual Shutdown 7.7 x 10-7 15 September 2010 G-2 Draft NUREG-1437, Supplement 45

Appendix G Turbine Trip with Bypass 6.2 x 10-7 12 Small Loss of Coolant Accident (LOCA) - Water 2.8 x 10- 7 5

(Below Top of Active Fuel)

Small LOCA - Steam (Above Top of Active Fuel) 2.3 x 10- 7 4

Loss of Condenser Vacuum 2.0 x 10-7 4

Fire Protection System Rupture Outside Control Room 1.9 x 10-7 4

Isolation LOCA in Emergency Core Cooling System 1.1 x 10-7 2

(ECCS) Discharge Paths Main Steam Isolation Valve (MSIV) Closure 1.1 x 10-7 2

Internal Flood Outside Lower Relay Room 9.7 x 10-8 2

Loss of Feedwater 8.8 x 10- 8 2

Loss of Safety Auxiliaries Cooling System 7.9 x 10-8 2

Reactor Auxiliaries Cooling System (RACS) Common 7.6 x 10-8 1

Header Unisolable Rupture Unisolable SW A Pipe Rupture in RACS Room 5.7 x 10-8 1

Unisolable SW B Pipe Rupture in RACS Room 5.7 x 10-8 1

Others (less than 1% each) 4.1 x 10-7 8

Total CDF (internal events) 5.1 x 10-6 100 1Column totals may be different due to round off.

1 The Level 2 HCGS PRA model that forms the basis for the SAMA evaluation is essentially a 2

complete revision to the IPE model. The Level 2 model utilizes three containment event trees 3

(CETs) containing both phenomenological and systemic events. The Level 1 core damage 4

sequences are binned into accident classes that provide the interface between the Level 1 and 5

Level 2 CET analysis. The CETs are linked directly to the Level 1 event trees and CET nodes 6

are evaluated using supporting fault trees.

7 The result of the Level 2 PRA is a set of 11 release or source term categories, with their 8

respective frequency and release characteristics. The results of this analysis for HCGS are 9

provided in Table E.3-6 of ER Appendix E (PSEG 2009). The categories were defined based 10 on the timing of the release, the magnitude of the release, and whether or not the containment 11 remains intact or fails. The frequency of each release category was obtained by summing the 12 frequency of the individual accident progression CET endpoints binned into the release 13 category. Source terms were developed for each of the 11 release categories using the results 14 of Modular Accident Analysis Program (MAAP 4.0.6) computer code calculations.

15 The offsite consequences and economic impact analyses use the MACCS2 code to determine 16 the offsite risk impacts on the surrounding environment and public. Inputs for these analyses 17 include plant-specific and site-specific input values for core radionuclide inventory, source term 18 and release characteristics, site meteorological data, projected population distribution (within a September 2010 G-3 Draft NUREG-1437, Supplement 45

Appendix G 1

50-mile radius) for the year 2046, emergency response evacuation modeling, and economic 2

data. The core radionuclide inventory corresponds to the end-of-cycle values for HCGS 3

operating at 3917 MWt, which is two percent above the current extended power uprate (EPU) 4 licensed power level of 3,840 MWt. The magnitude of the onsite impacts (in terms of clean-up 5

and decontamination costs and occupational dose) is based on information provided in 6

NUREG/BR-0184 (NRC 1997a).

7 In the ER, PSEG estimated the dose to the population within 80-kilometers (50-miles) of the 8

HCGS site to be approximately 0.23 person-Sievert (Sv) (22.9 person-roentgen equivalent man 9

[rem]) per year. The breakdown of the total population dose by containment release mode is 10 summarized in Table G-2. Releases from the containment within the early time frame (0 to less 11 than 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> following event initiation) and intermediate time frame (4 to less than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> 12 following event initiation) dominate the population dose risk at HCGS.

13 Table G-2. Breakdown of Population Dose by Containment Release Mode 14 Population Dose Percent Containment Release Mode (Person-Rem1 Per Year)

Contribution Early Releases (< 4hrs) 11.9 52 Intermediate Releases (4 to <24 hrs) 9.9 43 Late Releases (>24 hrs) 1.1 5

Intact Containment

<0.1 negligible Total 22.9 100 1One person-rem = 0.01 person-Sv 15 16 G.2.2 Review of PSEG's Risk Estimates 17 PSEG's determination of offsite risk at HCGS is based on the following three major elements of 18 analysis:

19 a

The Level 1 and 2 risk models that form the bases for the 1994 IPE submittal 20 (PSEG1994), and the external event analyses of the 1997 IPEEE submittal (PSEG 21 1997),

22 0

The major modifications to the IPE model that have been incorporated in the HCGS 23 PRA, and 24 The MACCS2 analyses performed to translate fission product source terms and release 25 frequencies from the Level 2 PRA model into offsite consequence measures.

26 Each of these analyses was reviewed to determine the acceptability of PSEG's risk estimates 27 for the SAMA analysis, as summarized below.

September 2010 G-4 Draft NUREG-1437, Supplement 45

Appendix G 1

2 3

4 5

6 7

8 9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 The NRC staffs review of the HCGS IPE is described in an NRC report dated April 23, 1996 (NRC 1996). Based on a review of the IPE submittal and responses to RAIs, the NRC staff concluded that the IPE process is capable of identifying the most likely severe accidents and severe accident vulnerabilities, and therefore, that the HCGS IPE has met the intent of GL 88-20 (NRC 1988).

During the performance of the IPE, transients involving heating, ventilation, and air conditioning (HVAC) failure were determined to contribute inordinately to the CDF. This was labeled a vulnerability and a procedure to provide alternate ventilation was developed. The implementation of this procedure removed this vulnerability. Credit for this procedure was taken in the HCGS IPE submittal. No other vulnerabilities were identified. In the ER, PSEG indicated that there were three improvements identified in the process of performing the IPE. Two of the improvements were performing refined calculations to allow increased credit for existing plant design features. The third was developing a procedure for operation of the Safety Auxiliaries Cooling System in severe accident conditions. All of these improvements are stated to have been implemented (PSEG 2009).

There have been twelve revisions to the IPE model since the 1994 IPE submittal. A listing of the changes made to the HCGS 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 G-3. A comparison of internal events CDF between the 1994 IPE and the current PRA model indicates a decrease of about a factor of ten in the total CDF (from 4.7 x 10-5 per year to 5.1 x 10-6 per year). This reduction can be attributed to significant changes in success criteria, modeling details and removal of conservatism.

Table G-3. HCGS PRA Historical Summary PRA Total CDF1 Version Summary of Changes from Prior Model (per year) 1994 IPE Submittal 4.7 x 10-5 Model 0 Credit taken for beyond design basis performance of Safety Auxiliaries 1.3 x 10-'

9/1994 Cooling System (SACS) and Station Service Water System (SSWS) based on updated success criteria calculations.

Model 1.0 Integrated the Level I and II models 1.9 x 10-5 7/1999 Updated the database Further developed sequence end states Developed fault trees for special initiators Reviewed dependent operator actions Model 1.32 Requantified two important human error probabilities 9.3 x 10' 10/2000 Revised treatment of disallowed maintenance to credit plant procedures and operating practices.

Revised common cause failure assessment Eliminated core spray room cooling dependency on SACS based on review of room heat up calculations Added models for breaks outside containment and manual shutdown Updated AIWS analysis September 2010 G-5 Draft NUREG-1437, Supplement 45

Appendix G PRA Version Model 2003A 8/2003 Summary of Changes from Prior Model Incorporated resolution of 1999 BWROG peer review Facts and Observations (Attachment 14 to PSEG 2005)

Converted from NUPRA to CAFTA software Performed completely new human reliability assessment Revised accident sequence definitions Performed new MAAP calculations for extended power uprate (EPU) conditions Updated data Modified system models Updated common cause failure analysis AA AlU~.

Ha #~

IOU A

AI.t

1. ~

I.~

LOttO tC Total CDF1 (per year) 3.1 x 10-5 Rev. 2.0 Modified 480 VAC dependencies 1.7 x 10.'

10/2004 Modified SACS success criteria Modified SACS-SW Human Error Probabilities Model 2005C 3 Removed conservatism in the SACS-SW success criteria 9.8 x 10' 2/2006 Included more detailed logic for AC power supplies Removed conservatism in operator action human error probabilities (HEPs)

Reduced turbine trip initiating event frequency HC108A BWROG Peer Reviewed 7.6 x 10.6 8/2008 Incorporated seasonal success criteria for SACS and SSWS Updated internal flooding scenarios and initiating event frequencies to be consistent with ASME PRA standard Credited use of portable battery charger for Station Blackout scenarios Reassessed human error probabilities using Electric Power Research Institute (EPRI) human reliability analysis (HRA) calculator Updated evaluation of dependent operator actions HC108B4 Credited procedure changes for local manual manipulation of SSWS valves 5.1 x 10.

12/2008 under LOOP conditions (4.4 x 10')

Removed conservatism in modeling of 120 VAC inverter room cooling logic Updated SACS pump failure probabilities to be consistent with Bayesian update values 1Total CDF includes internal floods. Prior to Model 2003A, IPE internal flood analysis was retained.

2Changes for Model 1.3 includes those for prior intermediate Models 1.1 and 1.2. All changes were considered minor.

3Changes for Model 2005C includes those for prior intermediate Models 2005A and 2005B. All changes to Models 2005A and 2005B were considered minor.

4Model HC108B truncation limit was decreased to 1 x 10-12 per year from 5 x 10'11 per year utilized for the HCI08A and 2005 models. The CDF in parentheses is the result based on the higher truncation limit.

The CDF value from the 1994 IPE (4.7 x 10-5 per year) is in the upper third of the values reported for other BWR 3/4 plants. Figure 11.2 of NUREG-1 560 shows that the IPE-based total 1

2 3

4 September 2010 G-6 Draft NUREG-1437, Supplement 45

Appendix G 1

internal events CDF for BWR 3/4 plants ranges from 9 x 10-8 per year to 8 x 10-5 per year, with 2

an average CDF for the group of 2 x 10-5 per year (NRC 1997b). It is recognized that other 3

plants have updated the values for CDF subsequent to the IPE submittals to reflect modeling 4

and hardware changes. The current internal events CDF results for HCGS (5.1 x 10-6 per year) 5 are comparable to that for other plants of similar vintage and characteristics.

6 The NRC staff considered the peer reviews performed for the HCGS PRA, and the potential 7

impact of the review findings on the SAMA evaluation. In the ER (PSEG 2009) and in response 8

to an NRC staff RAI (PSEG 2010a) and in other unrelated submittals (PSEG 2005), PSEG 9

described three BWROG Peer Reviews for the HCGS PRA. The first was a pilot of the BWROG 10 peer review process conducted in 1996 of PRA Model 0. The second, conducted in 1999, 11-reviewed PRA Model 1.0. The third, conducted in 2008, reviewed the HCl08A Model.

12 The 1999 peer review identified no Level A (extremely important) and 80 Level B (important) 13 Facts and Observations (F&Os). It was stated that these F&Os were resolved and incorporated 14 in the 2003A PRA Model (PSEG 2005).

15 The 2008 peer review of the HC108A model was requested by PSEG because of the significant 16 changes in PRA methods since the prior peer review. This peer review was performed using 17 the Nuclear Energy Institute peer review process (NEI 2007) and the ASME PRA Standard 18 (ASME 2005) as endorsed by the NRC in Regulatory Guide 1.200, Rev. 1 (NRC 2007). In the 19 ER PSEG summarizes the results of the peer review by reporting the number of ASME 20 Standard's supporting requirements (SRs) that were assessed to meet each of the standard's 21 Capability Categories. Of the 301 SRs applicable to HCGS, 286 were found to meet the 22 requirements for Capability Category II or higher, seven met Capability Category I and eight did 23 not meet any Capability Category. Capability Category II is described as follows (ASME 2005):

24

1) the scope and level of detail has resolution and specificity sufficient to identify the relative 25 importance of significant contributors at the component level including human actions, as 26 necessary, 2) plant-specific data/models are used for significant contributors, and 3) departures 27 from realism will have small impact on the conclusions and risk insights as supported by good 28 practices.

29 In the ER, PSEG indicated that the SRs identified as "not met" were addressed in the HC108B 30 model. In response to an NRC staff RAI, PSEG provided a listing and discussion of the 31 resolution of the SRs that only met Capability Category I and of other Peer Review Finding-level 32 F&Os (PSEG 2010a). It should be noted that a Finding-level F&O is essentially equivalent to 33 and replaces the previously used Level A and B F&Os and is defined as an observation that is 34 necessary to address to ensure 1) the technical adequacy of the PRA, 2) the 35 capability/robustness of the PRA update process, and 3) the process for evaluating the 36 necessary capability of the PRA technical elements (NEI 2007).

37 Of the seventeen identified SRs and findings, thirteen were stated to have been resolved as part 38 of the HC108B PRA update and re-assessed as meeting Capability Category II at a minimum as 39 a result of additional investigation, analysis and/or documentation. Four of the SRs and findings 40 remain open. In the discussion of the status and impact of these open items, PSEG concluded 41 that the resolution of each would not impact the conclusions of the SAMA risk assessment. Two 42 of the open items were documentation issues. One issue was related to the need for additional 43 plant-specific data for important events. PSEG indicated that a review of HCGS recent September 2010 G-7 Draft NUREG-1437, Supplement 45

Appendix G 1

experience indicates "no anomalous behavior" and that minor changes to component 2

unavailability and unreliability values would not change the conclusions of the SAMA risk 3

evaluation. The fourth issue was related to the identification, characterization and 4

documentation of model uncertainties. PSEG indicated that a number of sensitivity evaluations 5

were performed and that other areas of the HCGS PRA were investigated for potential impact 6

on the PRA results but none were found to rise to the level of being candidates for modeling 7

uncertainty. PSEG concluded that the resolution of this open item would not impact the 8

conclusions of the SAMA evaluation (PSEG 201 Ga). PSEG further states that the HCGS PRA 9

treatment of model uncertainty is considered to meet the requirements of the latest NRC 10 guidance on model uncertainty, NUREG-1855 (NRC 2009).

11 In the initial response to the NRC staff RAIs (PSEG 2010a) PSEG's discussion of the resolution 12 of the supporting requirements that were not met addressed only six items whereas the initial 13 listing in the ER indicated that there were eight SRs that were not met. In response to the 14 request for clarification PSEG pointed out that the draft peer review report identified eight SRs 15 as not met, while the final review report identified only six SRs as not being met (PSEG 2010b).

16 The NRC staff considers PSEG's disposition of the peer review findings to be reasonable and 17 that final resolution of the findings is not likely to impact the results of the SAMA analysis.

18 The Revision HC108B model reflects the current (as of the date of the ER submittal) HCGS 19 configuration and design. The licensee states that HCGS risk management personnel have 20 reviewed plant modifications and procedure changes since the HC108B model freeze date. No 21 changes were identified that required PRA model updates and therefore the licensee concluded 22 that none of the plant modifications and procedure changes since the HC108B PRA update 23 would impact the conclusions of the SAMA analysis. (PSEG 2010a, PSEG 2010b) 24 In response to an RAI, PSEG described the overall quality assurance program applicable to the 25 HCGS PRA and its updates by providing descriptions of significant governing PSEG 26 procedures. These procedures address the overall risk management program, risk 27 management documentation including quality requirements for preparation, review and 28 approval, configuration control and PRA model updates. The procedures appear to address the 29 appropriate requirements.

30 Given that the HCGS internal events PRA model has been peer-reviewed and the peer review 31 findings were all addressed, and that PSEG has satisfactorily addressed NRC staff questions 32 regarding the PRA, the NRC staff concludes that the internal events Level 1 PRA model is of 33 sufficient quality to support the SAMA evaluation.

34 As indicated above, PSEG does not maintain a current HCGS external events PRA that 35 explicitly models seismic and fire initiated core damage accidents that can be linked with the 36 current Level 2 and 3 PRA. However, the models developed for seismic and fire events in the 37 IPEEE were partially updated in 2003 to utilize revised initiating event frequencies and 38 conditional core damage probabilities based on the 2003A internal events PRA Model. These 39 results were used to identify SAMAs that address important fire and seismic risk contributors, as 40 discussed below in Section G.3.2. The updated seismic and fire core damage results are 41 described in ER Section E.5.1.7 42 The HCGS IPEEE was submitted in July 1997 (PSEG 1997), in response to Supplement 4 of 43 Generic Letter 88-20 (NRC 1991a). The submittal included a seismic PRA, an internal fire PRA, September 2010 G-8 Draft NUREG-1437, Supplement 45

Appendix G 1

and an evaluation of high winds, external flooding, and other hazards. While no fundamental 2

weaknesses or vulnerabilities to severe accident risk in regard to the external events were 3

identified, two potential enhancements were identified as discussed below. In a letter dated July 4

26, 1999 (NRC 1999), the NRC staff concluded that PSEGs IPEEE process is capable of 5

identifying the most likely severe accidents and severe accident vulnerabilities, and therefore, 6

that the HCGS IPEEE has met the intent of Supplement 4 to Generic Letter 88-20.

7 The HCGS IPEEE seismic analysis utilized a seismic PRA following NRC guidance (NRC 8

1991a). The seismic PRA included: a seismic hazard analysis, a seismic fragility assessment, a 9

seismic systems analysis, and quantification of seismic CDF.

10 The seismic hazard analysis estimated the annual frequency of exceeding different levels of 11 ground motion. Seismic CDFs were determined for both the EPRI (EPRI 1989) and the 12 Laurence Livermore National Laboratory (LLNL) (NRC 1994) hazard assessments. The seismic 13 fragility assessment utilized the walkdown procedures and screening caveats in EPRI's seismic 14 margin assessment methodology (EPRI 1991). Fragility calculations were made for about 90 15 components and, using a screening criterion of median peak ground acceleration (pga) of 1.5 g 16 which corresponds to a 0.5 pga high confidence low probability of failure (HCLPF) capacity, a 17 total of 17 components were screened in. The seismic systems analysis defined the potential 18 seismic induced structure and equipment failure scenarios that could occur after a seismic event 19 and lead to core damage. The HCGS IPE event tree and fault tree models were used as the 20 starting point for the seismic analysis. Quantification of the seismic models consisted of 21 convoluting the seismic hazard curve with the appropriate structural and equipment seismic 22 fragility curves to obtain the frequency of the seismic damage state. The conditional probability 23 of core damage given each seismic damage state was then obtained from the IPE models with 24 appropriate changes to reflect the seismic damage state. The CDF was then given by the 25 product of the seismic damage state probability and the conditional core damage probability.

26 The seismic CDF resulting from the HCGS IPEEE was calculated to be 3.6 x 10.6 per year using 27 the LLNL seismic hazard curve and 1.0 x 10.6 per year using the EPRI seismic hazard curve.

28 Both utilized the HCGS Model 0 internal events PRA, with a CDF of 1.3 x 10-5 per year for 29 quantification of non-seismic failures.

30 The HCGS IPEEE did not identify any vulnerability due to seismic events or any potential 31 improvements to reduce seismic risk. The IPEEE noted, however, that fire water tanks are not 32 seismically robust and hence no credit was taken for the fire protection system in the seismic 33 PRA. This is discussed further in Section G.3.2.

34 Subsequent to the IPEEE, PSEG updated the seismic PRA utilizing conditional core damage 35 probabilities from the 2003A PRA model modified to reflect the seismic human reliability 36 assessment that was performed to support the IPEEE, referred to as the HCGS 2003 External 37 Events Update (PSEG 2009). The resulting seismic CDF using the EPRI seismic hazard curves 38 is 1.1 x 10-6 per year. In the ER, PSEG provided a listing and description of the top ten seismic 39 core damage contributors. The dominant seismic core damage contributors with a CDF of 40 1 x 10.8 per year or more are listed in Table G-4. In response to an NRC staff RAI, PSEG also 41 determined the updated seismic CDF using the LLNL seismic hazard curve and the total 42 seismic CDF was determined to be 3.6 x 10-6 per year. The seismic CDF utilizing the LLNL 43 hazard curves for dominant seismic core damage contributors are also listed in Table G-4.

September 2010 G-9 Draft NUREG-1437, Supplement 45

Appendix G 1

Table G-4. Dominant Contributors to the Seismic CDF Based on EPRI Seismic Based on LLNL Seismic Hazard Curves Hazard Curves Contribution Contribution Basic CDF (per to Seismic CDF (per to Seismic Event ID Seismic Sequence Description year)

CDF year)

CDF

%IE-Seismic-Induced Equipment 60 70 SET36 Damage State SET-36 (Impacts -

6.7x 10-7 2.5x 10-6 120V PNL481)

%IE-Seismic-Induced Equipment 27 9

SET18 Damage State SET-18 (Impacts -

3.1 x 10-7 3.3 x 10-7 LOOP)

%IE-Seismic-Induced Equipment 6.8 x 10.8*

6 4.4 x 10-7 12 SET37 Damage State SET-37 (Impacts -

125V)

%IE-Seismic-Induced Equipment 4.6 x 10-8 4

1.6 x 10-7 5

SET35 Damage State SET-35 (Impacts -

120V PNL482, RSP)

%IE-Seismic-Induced Equipment 2.1 x 10-8 2

5.4 x 10-8 2

SET38 Damage State SET-38 (Impacts -

1 E panel room Ventil.)

• In response to an RAI, PSEG indicated that the value reported in the ER page E-99 for this contributor was in error and should be that given in the IPEEE - 6.8 x 10-8 per year (PSEG 2010a).

For both hazard curves, the largest contributor to seismic CDF is a seismic-induced loss of all four divisions of 1 E 120 VAC instrumentation distribution panels that leads directly to core damage. Other significant contributors are: for the EPRI hazard curves, a seismic-induced loss of offsite power which together with non-seismic random failures leads to core damage and, for the LLNL hazard curves, a seismic induced failure of all 125 VDC 1E power to loads that lead directly to core damage. The failure of all four 1E 120 VAC divisions and failure of all 125 VDC occur at a relatively high ground acceleration (a median failure at 1.08g and 1.47g, respectively) while the loss of offsite power occurs at a relatively low ground acceleration (a median failure of 0.31g) (PSEG 1997).

The NRC staff requested the applicant assess the impact the higher seismic CDF resulting from the use of the LLNL hazard curves would have on the external events multiplier and the results of the SAMA analysis as well as the impact of the increased CDF for important seismic sequences on the identification and evaluation of SAMAs for these sequences. This is discussed further below and in Sections G.3.2 and G.6.2.

The HCGS IPEEE fire analysis employed EPRI's fire-induced vulnerability evaluation (FIVE) methodology (EPRI 1993) to perform a fire compartment interaction analysis (FCIA) and a 2

3 4

5 6

7 8

9 10 11 12 13 14 15 16 17 18 September 2010 G-10 Draft NUREG-1437, Supplement 45

Appendix G 1

quantitative screening analysis. This was then followed by a PRA quantification of the 2

unscreened compartments.

3 The FCIA identified 209 fire compartments meeting the FIVE criteria for the entire plant. The 4

quantitative screening utilized a threshold fire ignition frequency obtained using the FIVE 5

methodology and the assumptions that all fires resulted in a reactor trip or more severe transient 6

and that any fire in a compartment damaged all the equipment and cables in the compartment.

7 Using the assessed screening fire frequency and conservatively determined screening 8

conditional core damage probabilities (CCDPs) from the Model 0 internal events PRA resulted 9

in screening out (at a CDF of less than 1 x 10- per year) of all but 38 fire compartments.

10 The analysis for the unscreened areas employed a detailed probabilistic assessment of each 11 possible fire initiator/target combination including intermediate fire growth stages. Fire damage 12 calculations used a modified version of the FIVE fire propagation methodology. No explicit 13 credit was taken for manual or automatic fire suppression. Final quantification utilized FIVE fire 14 data and refined CCDPs from the Model 0 internal events PRA. The resulting fire induced CDF 15 was calculated to be 8.1 x 10-5 per year. A walkdown and verification process was employed to 16 verify that the assumptions and calculations were supported by the physical condition of the 17 plant.

18 The HCGS IPEEE did not identify any vulnerabilities due to internal fires or any potential 19 improvements to reduce internal fire risk.

20 Subsequent to the IPEEE, PSEG updated the fire PRA to incorporate more recent fire initiating 21 event frequencies based on information in the 2002 NRC fire database and conditional core 22 damage probabilities from the 2003A PRA model, referred to as the 2003 HCGS External 23 Events Update. The resulting fire CDF is 1.7 x 10-5 per year.

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

25 The important fire core damage contributors with a CDF of 1 x 10-7 per year or more are listed in 26 Table G-5. As can be seen from these results the fire risk at HCGS is dominated by panel fires 27 in the control room.

28 Table G-5. Important Contributors to Fire CDF Basic Event CDF

% Contribution ID Fire Area Description per year to Fire CDF

%IE-FIRE03 Control Room Fire Scenario Small Cab_3 (Loss of 5.3x 10-6 31 Emer. Bat.)

%IE-FIRE02 Control Room Fire Scenario Small Cab_2 (Loss of 4.4 x 10-6 25 SSWS)

%IE-FIRE01 Control Room Fire Scenario Small CabI (Loss of 3.8 x 10-6 22 SACS)

%IE-FIRE28 Compartment 5339 Fire Scenario 5339_2 7.5x 10-7 4

%IE-FIRE37 DG room (D) Fire Scenario 5304_2 7.0 x 10-7 4

%IE-FIRE20 DG room (C) Fire Scenario 5306_2 6.7 x 10-7 4

September 2010 G-1 1 Draft NUREG-1437, Supplement 45

Appendix G

%IE-FIRE38 Compartment 3425/5401 Fire Scenario 5401_1 5.9-10-7 3

%IE-FIRE06 Control Room Fire Scenario Large Cabjl (MSIV 5.1 x 10-7 3

Closure) 1 2

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

and the fire analysis results is not possible because neither the plant response model or the fire 4

modeling methodology used in the updated fire model is current. PSEG identified in the ER 5

areas where fire CDF quantification may introduce levels of uncertainty different from those 6

expected in the internal events PRA, including a number of conservatisms in the fire modeling, 7

as follows:

8 0

Several system models assume the systems are unavailable or are unrecoverable in a 9

fire. For example, any fire is assumed to result in a plant trip, even if it is not severe.

10 Other examples are provided in the ER.

11 0

Bounding fire modeling assumptions are used for many fire scenarios. For example, all 12 cables are damaged in a fire even if they are enclosed in cable trays or conduit. Other 13 examples are provided in the ER.

14 0

Because of a lack of industry experience with regard to crew performance during the 15 types of fires modeled in the fire PRA, the characterization of crew actions in the fire 16 PRA is generally conservative.

17 PSEG's conclusion is that while some of the conservatisms have been addressed in the 18 updated fire model, the result is still believed to be conservative.

19 Considering the above discussion, the conservatisms in the updated fire PRA model as 20 currently understood, and the response to the NRC staff RAIs, the NRC staff concludes that the 21 fire CDF of 1.7 x 10-5 per year is reasonable for the SAMA analysis.

22 The IPEEE analysis of high winds, floods and other (HFO) external events indicated that each 23 of the events identified in NUREG-1407 (NRC 1991b) had a core damage contribution of less 24 than the screening criterion of 1 x 10-6 per year. This was done by either showing compliance 25 with the 1975 Standard Review Plan criteria or by a bounding analysis that demonstrated that 26 the CDF contribution was less than the screening criterion. For the SAMA analysis, PSEG 27 assumed a CDF contribution of 1 x 10.6 per year for each of high winds, external floods, 28 transportation and nearby facilities, detritus, and chemical releases, for a total HFO CDF 29 contribution of 5 x 10-6 per year (PSEG 2009).

30 Although the HCGS IPEEE did not identify any vulnerabilities due to HFO events, two 31 improvements to reduce risk were identified as described below.

32 For high winds, the HCGS design was compared to the SRP criteria and found to have a CDF 33 contribution less than the screening criterion. A walkdown was performed to evaluate high wind 34 hazards and as a result work was initiated to install a missile shield in front of a door into the 35 Technical Support Center. This improvement has been implemented.

September 2010 G-12 Draft NUREG-1437, Supplement 45

Appendix G 1

For external floods the HCGS was found to be adequately protected from the postulated 2

occurrence of the probable maximum hurricane surge with wave run-up coincident with the 10%

3 exceedance high tide. HCGS was also found to comply with the latest probable maximum 4

precipitation criteria. A walkdown confirmed that there were no severe accident vulnerabilities 5

due to external floods.

6 A review of transportation and nearby facility accidents confirmed that there were no severe 7

accident vulnerabilities from these accidents. During the review it was discovered that in a 8

single year there had been some unauthorized shipments of explosives on the Delaware River 9

in the vicinity of the HCGS. The U.S. Coast Guard (USCG), which controls such shipments, 10 was contacted and procedures were put in place to prevent such shipments in the future. This 11 improvement has been implemented.

12 The NRC staff asked about the status and potential impact on the SAMA analysis of a liquefied 13 natural gas (LNG) terminal planned for Logan Township, New Jersey, upstream on the 14 Delaware River from the HCGS site (NRC 2010a). In response to the RAI, PSEG discussed the 15 current status of the LNG terminal as well as the regulatory controls for LNG marine traffic and 16 LNG ship design and the safety record for LNG shipping (PSEG 2010a). The LNG terminal 17 remains in the planning stage and no construction has begun. Further, the state of Delaware 18 has denied applications for several required environmental permits and approvals. PSEG 19 concluded that based on the regulatory process and controls for'assuring the safety and 20 security of LNG ships, the safety record of LNG ships and the uncertainty of the planned 21 terminal, consideration of potential SAMAs associated with the possible future terminal is not 22 warranted. The NRC staff agrees with this conclusion.

23 As indicated in the ER (PSEG 2009), a multiplier of 6.3 was used to adjust the internal event 24 risk benefit associated with a SAMA to account for external events. This multiplier was based 25 on a total external event CDF of 2.3 x 10s per year. This CDF is the sum of the updated fire 26 CDF of 1.7 x 10-5 per year, the updated seismic CDF of 1.1 x 10-6 per year, and the HFO CDF 27 of 5 x 10-6 per year. The external event CDF is thus approximately 5.3 times the internal events 28 CDF of 4.4 x 10-6 per year used in the SAMA analysis at a truncation of 5 x 10-11 per year. The 29 higher truncation used for determining the multiplier is to be consistent with that used to 30 determine the release category frequencies and that used to evaluate the fire and seismic 31 CDFs. The total CDF is thus 6.3 times the internal events CDF (PSEG 2009).

32 As indicated above, in response to an NRC staff RAI, PSEG determined the seismic CDF based 33 on the LLNL hazard curve to be 3.6 x 10-6 per year (PSEG 2010a). If this is utilized instead of 34 the value using the EPRI hazard curve, the total external events CDF is 2.6 x 10-5 per year and 35 the external events multiplier is 6.8. The impact of this revised multiplier on the SAMA 36 assessment is discussed further in Section G.3.2 and Section G.6.2.

37 The NRC staff reviewed the general process used by PSEG to translate the results of the Level 38 1 PRA into containment releases, as well as the results of the Level 2 analysis, as described in 39 the ER and in response to NRC staff requests for additional information (PSEG 2010a, PSEG 40 2010b). The HCGS Level 2 PRA model is essentially a complete revision of the IPE Level 2 41 model, including completely revised containment event trees and system fault trees and 42 completely updated thermal hydraulic analyses, incorporating the latest emergency operating 43 procedures (EOPs), severe accident guidelines (SAGs), and emergency action level (EAL) and 44 implementation using the CAFTA software.

September 2010 G-13 Draft NUREG-1437, Supplement 45

Appendix G 1

The current Level 2 model utilizes a set of three containment event trees (CETs) containing both 2

phenomenological and systemic events. The Level 1 core damage sequences are grouped into 3

core damage accident classes with similar characteristics. All the sequences in an accident 4

class are then input to one of the three CETs by linking the level 1 event tree sequences with 5

the level 2 CET. The CETs are analyzed by the linking of fault trees that represent each CET 6

node. These fault trees are based on the Level 1 models for the system or function as modified 7

for Level 2 considerations of timing, procedures, access or dependencies including recovery 8

actions as documented in the HCGS emergency Operating Procedures and Severe Accident 9

Management Guidelines.

10 Each CET end state represents a radionuclide release to the environment and is characterized 11 by one of thirteen release bins based on magnitude and timing of release. Magnitude is given 12 by Csl release fraction: High (H) > 10%, Moderate (M) 1% to 10%, Low (L) 0.1% to 1%, Low-13 Low (LL) <0.1% and negligible or no release<< 0.1%. Timing is given by time of initial release 14 from the time of declaration of a General Emergency: Early (E) < 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, Intermediate (I), 4 to 15 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and Late (L) > 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. The assignment of each end state to a given release bin is 16 made on the basis of a MAAP calculation for the accident sequence or a similar MAAP 17 calculated sequence. The thirteen release bins were subsequently refined into eleven release 18 categories for input to the MACCS consequence calculations by dividing the high early release 19 bin into three release categories (high pressure, low pressure and breaks outside containment) 20 and combining several of the end states with Low and Low-Low release magnitudes.

21 The frequency of each release category was obtained by summing the frequency of the 22 contributing CET end states. The release characteristics for each release category were 23 developed by using the results of Modular Accident Analysis Program (MAAP 4.0.6) computer 24 code calculations. A representative MAAP case for each of the release categories was chosen 25 based on a review of the Level 2 cutsets and the dominant types of scenarios that contribute to 26 the results. The MAAP case chosen for each release category was generally the case with the 27 highest consequence (PSEG 2010a). A description of the representative MAAP case for each 28 release or source term category is provided in Table E.3-5 of the ER. The release categories, 29 their frequencies, and release characteristics are presented in Table E.3-6 of the ER (PSEG 30 2009).

31 It is noted for the SAMA analysis the CET end state and release category frequencies were 32 determined using a truncation value of 5 x 10-11 per year. This results in a total CDF of 33 approximately 4.4 x 10-6 per year, which is about 16 percent less that the internal events CDF of 34 5.1 x 10-6 per year obtained when a truncation of 1 x 10-12 per year. The NRC staff considers 35 that use of the release frequency rather than the Level 1 CDF will have a negligible impact on 36 the results of the SAMA evaluation because the external event multiplier and uncertainty 37 multiplier used in the SAMA analysis (discussed in Section G.6.2) have a much greater impact 38 on the SAMA evaluation results than the small error arising from the model quantification 39 approach.

40 The NRC staff review of release category information noted an apparent discrepancy in the 41 release magnitude and release timing assigned for ST5 and ST7 and requested the applicant to 42 clarify the reasons for these discrepancies (NRC 2010a). Both these release categories involve 43 loss of containment heat removal with subsequent containment failure, core damage and fission 44 product release. For ST5 the containment failure is in the wet well while for ST7 the September 2010 G-14 Draft NUREG-1437, Supplement 45

Appendix G 1

containment failure is in the drywell. While the drywell failure would be expected to result in a 2

higher release than a wet well failure, the reverse is true for the results provided in the ER.

3 Further, the release timings were found to be slightly different even though the core damage 4

times were the same. In response to the RAI, PSEG pointed out that the wet well failure for 5

ST5 occurred below the water level and, due to the loss of suppression pool water inventory, 6

resulted in significantly less cesium iodide removal from the safety relief valve (SRV) flow to the 7

suppression pool for ST5 than for the drywell failure case ST7 (PSEG 2010a). The differing 8

release pathways resulted in the slightly different times for the initiation of release to the 9

environment.

10 Based on the NRC staffs review of the Level 2 methodology, the applicant's responses to RAIs 11 and the fact that the Level 2 model was reviewed in more detail as part of the 2008 BWROG 12 peer review and found to be acceptable (except for two documentation related findings which 13 would not impact the SAMA analysis), the NRC staff concludes that the Level 2 PRA provides 14 an acceptable basis for evaluating the benefits associated with various SAMAs.

15 The NRC staff reviewed the process used by PSEG to extend the containment performance 16 (Level 2) portion of the PRA to an assessment of offsite consequences (essentially a Level 3 17 PRA). This included consideration of the source terms used to characterize fission product 18 releases for the applicable containment release categories and the major input assumptions 19 used in the offsite consequence analyses. The MACCS2 code was utilized to estimate offsite 20 consequences. Plant-specific input to the code includes the source terms for each category and 21 the reactor core radionuclide inventory (both discussed above), site-specific meteorological 22 data, projected population distribution within an 80-kilometer (50-mile) radius for the year 2046, 23 emergency evacuation modeling, and economic data. This information is provided in Section 24 E.3 of Appendix E to the ER (PSEG 2009).

25 PSEG used the MACCS2 code and a core inventory from a plant specific calculation at end of 26 cycle to determine the offsite consequences of activity release. In response to an NRC staff 27 RAI, PSEG stated that the MACCS2 analysis was based on the core inventory used in the 28 NRC-approved Alternate Source Term for HCGS (PSEG 2010a).

29 All releases were modeled as being from the top of the reactor containment building and at low 30 thermal content (ambient). Sensitivity studies were performed on these assumptions and 31 indicated little or no change in population dose or offsite economic cost. Assuming a ground 32 level release decreased dose risk and cost risk by 6 percent and 7 percent, respectively.

33 Assuming a buoyant plume decreased dose risk and cost risk by 1 percent. Based on the 34 information provided, the staff concludes that the release parameters utilized are acceptable for 35 the purposes of the SAMA evaluation.

36 PSEG used site-specific meteorological data for the 2004 calendar year as input to the 37 MACCS2 code. The development of the meteorological data is discussed in Section E.3.7 of 38 Appendix E to the ER. The data were collected from onsite and local meteorological monitoring 39 systems. Sensitivity analyses using MACCS2 and the meteorological data for the years 2005 40 through 2007 show that use of data for the year 2004 results in the largest dose and economic 41 cost risk. Missing meteorological data was filled by (in order of preference): using data from the 42 backup met pole instruments (10-meter), using corresponding data from another level of the 43 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 44 same hour and a nearby day (substitution technique). The 10-meter wind speed and direction September 2010 G-15 Draft NUREG-1437, Supplement 45

Appendix G 1

were combined with precipitation and atmospheric stability (derived from the vertical 2

temperature gradient) to create the hourly data file for use by MACCS2. The NRC staff notes 3

that previous SAMA analyses results have shown little sensitivity to year-to-year differences in 4

meteorological data and concludes that the use of the 2004 meteorological data in the SAMA 5

analysis is reasonable.

6 The population distribution the licensee used as input to the MACCS2 analysis was estimated 7

for the year 2046 using year 1990 and year 2000 census data as accessed by SECPOP2000 8

(NRC 2003) as a starting point. In response to an NRC staff RAI, PSEG stated that the 9

transient population was included in the 10-mile EPZ, and included prior to the population 10 projection (PSEG 2010a). A ten year population growth rate was estimated using the year 1990 11 to year 2000 SECPOP2000 data and applied to obtain the distribution in 2046. The baseline 12 population was determined for each of 160 sectors, consisting of sixteen directions for each of 13 ten concentric distance rings to a radius of 50 miles surrounding the site. The SECPOP2000 14 census data from 1990 and 2000 were used to determine a ten year population growth factor for 15 each of the concentric rings. The population growth was averaged over each ring and applied 16 uniformly to all sectors within each ring. The NRC staff requested PSEG provide an 17 assessment of the impact on the SAMA analysis if a wind-direction weighted population 18 estimate for each sector were used (NRC 2010a). In response to the RAI, PSEG stated that the 19 impacts associated with angular population growth rates on PDR and OECR are minimal and 20 bounded by the 30% population sensitivity case (PSEG 2010a). This is based on the relatively 21 even wind distribution profile surrounding the site, the tendency for lateral dispersion between 22 sectors, and the use of mean values in the analysis. A sensitivity study was performed for the 23 population growth at year 2040. A 30 percent increase in population resulted in a 29 percent 24 increase in dose risk and a 30 percent increase in cost risk. In response to an NRC staff RAI, 25 PSEG stated that the radial growth rates used in the MACCS2 analysis provides a more 26 conservative population growth estimate than using 'whole county' data for averaging (PSEG 27 2010a). PSEG also identified that the population sensitivity case of 30 percent growth was 28 approximately equivalent to adding 5.9 percent to the 10-year growth rate. The NRC staff 29 considers the methods and assumptions for estimating population reasonable and acceptable 30 for purposes of the SAMA evaluation.

31 The emergency evacuation model was modeled as a single evacuation zone extending out 16 32 kilometers (10 miles) from the plant (the emergency planning zone - EPZ). PSEG assumed 33 that 95 percent of the population would evacuate. This assumption is conservative relative to 34 the NUREG-1 150 study (NRC 1990), which assumed evacuation of 99.5 percent of the 35 population within the emergency planning zone. The evacuated population was assumed to 36 move at an average radial speed of approximately 2.8 meters per second (6.3 miles per hour) 37 with a delayed start time of 65 minutes after declaration of a general emergency (KLD 2004). A 38 general emergency declaration was assumed to occur at the onset of core damage. The 39 evacuation speed is a time-weighted average value accounting for season, day of week, time of 40 day, and weather conditions. It is noted that the longest evacuation time presented in the study 41 (i.e., full 10 mile EPZ, winter snow conditions, 99th 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 42 issuance of the advisory to evacuate). Sensitivity studies on these assumptions indicate that 43 there is minor impact to the population dose or offsite economic cost by the assumed variations.

44 The sensitivity study reduced the evacuation speed by 50 percent to 1.4 m/s. This change 45 resulted in a 2 percent increase in population dose risk and no change in offsite economic cost September 2010 G-1 6 Draft NUREG-1437, Supplement 45

Appendix G 1

risk. The NRC staff concludes that the evacuation assumptions'and analysis are reasonable 2

and acceptable for the purposes of the SAMA evaluation.

3 Site specific agriculture and economic parameters were developed manually using data in the 4

2002 National Census of Agriculture (USDA 2004) and from the Bureau of Economic Analysis 5

(BEA 2008) for each of the 23 counties surrounding HCGS, to a distance of 50 miles.

6 Therefore, recently discovered problems in SECPOP2000 do not impact the HCGS analysis.

7 The values used for each of the 160 sectors were the data from each of the surrounding 8

counties multiplied by the fraction of that county's area that lies within that sector. Region-wide 9

wealth data (i.e., farm wealth and non-farm wealth) were based on county-weighted averages 10 for the region within 50-miles of the site using data in the 2002 National Census of Agriculture 11 (USDA 2004) and the Bureau of Economic Analysis (BEA 2008). Food ingestion was modeled 12 using the new MACCS2 ingestion pathway model COMIDA2 (NRC 1998a). For HCGS, less 13 than one percent of the total population dose risk is due to food ingestion.

14 In addition, generic economic data that is applied to the region as a whole were revised from the 15 MACCS2 sample problem input in order to account for cost escalation since 1986, the year that 16 input was first specified. A factor of 1.96, representing cost escalation from 1986 to April 2008 17 was applied to parameters describing cost of evacuating and relocating people, land 18 decontamination, and property condemnation.

19 The NRC staff concludes that the methodology used by PSEG to estimate the offsite 20 consequences for HCGS provides an acceptable basis from which to proceed with an 21 assessment of risk reduction potential for candidate SAMAs. Accordingly, the NRC staff based 22 its assessment of offsite risk on the CDF and offsite doses reported by PSEG.

23 G.3 Potential Plant Improvements 24 25 The process for identifying potential plant improvements, an evaluation of that process, and the 26 improvements evaluated in detail by PSEG are discussed in this section.

27 G.3.1 Process for Identifying Potential Plant Improvements 28 29 PSEG's process for identifying potential plant improvements (SAMAs) consisted of the following 30 elements:

31 0

Review of the most significant basic events from the current, plant-specific PRA and 32 insights from the HCGS PRA Group, 33 0

Review of potential plant improvements identified in, and original results of, the HCGS 34 IPE and IPEEE, 35 0

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

36 nuclear sites, and 37 0

Review of generic SAMA candidates from NEI 05-01 (NEI 2005) to identify SAMAs that 38 might address areas of concern identified in the HCGS PRA.

September 2010 G-17 Draft NUREG-1437, Supplement 45

Appendix G 1

Based on this process, an initial set of 23 candidate SAMAs, referred to as Phase I SAMAs, was 2

identified. In this Phase I evaluation, PSEG performed a qualitative screening of the initial list of 3

SAMAs and eliminated SAMAs from further consideration using the following criteria:

4 0

The SAMA is not applicable at HCGS due to design differences, 5

0 The SAMA has already been implemented at HCGS, 6

0 The SAMA would achieve results that have already been achieved at HCGS by other 7

means, or 8

0 The SAMA has estimated implementation costs that would exceed the dollar value 9

associated with completely eliminating all severe accident risk at HCGS.

10 Based on this screening, one SAMA was eliminated, and one additional SAMA was eliminated 11 by subsuming it into another SAMA, leaving 21 SAMAs for further evaluation. The results of the 12 Phase I screening analysis is given in Table E.5-3 of Appendix E to the ER. The remaining 13 SAMAs, referred to as Phase II SAMAs, are listed in Table E.6-1 of Appendix E to the ER. In 14 Phase II a detailed evaluation was performed for each of the 21 remaining SAMA candidates, 15 as discussed in Sections G.4 and G.6 below. To account for the potential impact of external 16 events, the estimated benefits based on internal events were multiplied by a factor of 6.3, as 17 previously discussed.

18 G.3.2 Review of PSEG's Process 19 PSEG's efforts to identify potential SAMAs focused primarily on areas associated with internal 20 initiating events, but also included explicit consideration of potential SAMAs for important fire 21 and seismic initiated core damage sequences. The initial list of SAMAs generally addressed the 22 accident sequences considered to be important to CDF from risk reduction worth (RRW) 23 perspectives at HCGS, and included selected SAMAs from prior SAMA analyses for other 24 plants.

25 PSEG provided a tabular listing of the Level 1 PRA basic events sorted according to their RRW 26 (PSEG 2009). SAMAs impacting these basic events would have the greatest potential for 27 reducing risk. PSEG used a RRW cutoff of 1.006, which corresponds to about a 0.6 percent 28 change in CDF given 100-percent reliability of the SAMA. This equates to a benefit of 29 approximately $100,000 (after the benefits have been multiplied by a factor of 6.3 to account for 30 external events), which is the minimum implementation cost associated with a procedure 31 change. As a result of this review, 11 SAMAs were identified.

32 In the level 1 importance review, PSEG stated for the important initiating events that "This 33 initiator event is a compilation of industry and plant specific data. (No specific SAMA identified)."

34 The NRC staff requested that PSEG provide assurance that for each of these initiating events 35 there is not a dominant contributor for which a potential SAMA to reduce the initiating event 36 frequency or mitigate the impact of the initiator would be viable. In response to this RAI, PSEG 37 discussed each of the initiators and the previously identified SAMAs that would reduce the 38 importance of the initiator by mitigating other failures in the core damage sequences associated 39 with these initiators (PSEG 2010a). In response to a request for clarification PSEG indicated September 2010 G-18 Draft NUREG-1437, Supplement 45

Appendix G 1

that HCGS specific failures that are contributors to the initiating event frequencies that pose a 2

unique vulnerability are typically captured and corrected within existing procedures, e.g., the 3

corrective action program, and can result in procedure changes, plant modifications and training 4

enhancements aimed at reducing further recurrence (PSEG 2010b). Based on this discussion 5

and a review of the latest ten years of HCGS Licensee Event Reports, the NRC staff concludes 6

that it is unlikely that further HCGS data review will identify any additional cost beneficial SAMAs 7

beyond those already identified.

8 The PSEG response to the NRC staff request for clarification provided additional information on 9

initiators modeled utilizing a fault tree approach rather then being based on initiating event data.

10 For the loss of station auxiliaries cooling system initiating event (%IE-SACS), PSEG identified 11 and evaluated SAMA 42, "Installation of SACS Standby Diesel-Powered Pump" (PSEG 2010b).

12 For an event involving the station service water system (NR-IE-SWS, "Nonrecovery of %IE-13 SWS"), the importance review identified two SAMAs as potentially mitigating this event: SAMA 14 3, "Install Back-up Air Compressor to Supply Air-Operated Valves (AOVs)," and SAMA 4, 15 "Provide Procedural Guidance to Cross-Tie Residual Heat Removal (RHR) Trains." In response 16 to an NRC staff RAI to clarify the source and applicability of these SAMAs to this event, PSEG 17 discussed the modeling involving the NR-IE-SWS event and the applicability of the SAMAs in 18 terms of the more general loss of decay heat removal function of which the event is associated 19 and other SAMAs that would mitigate this event (PSEG 2010a). Based on this discussion, the 20 NRC staff concludes that this event is adequately addressed in the SAMA analysis.

21 For a significant number of the Level 1 events reviewed no SAMAs were identified with the 22 reason stated to be that "... based on low contribution to L[evel] 1 risk and engineering 23 judgment, the anticipated implementation costs of hardware mods associated with mitigating 24 this event would likely exceed the expected cost-risk benefit" (PSEG 2009). In response to an 25 NRC staff RAI, PSEG provided a revised assessment of each of these events that showed that 26 each was either already addressed by an existing SAMA or that no effective SAMAs could be 27 identified (PSEG 2010a).

28 The NRC staff also requested PSEG to specifically consider the following proposed SAMAs to 29 address basic events on the Level 1 importance list for which no SAMA was identified (NRC 30 2010a):

31 Install a diverse redundant temperature controller to address basic event SAC-XHE-MC-32 DF01, "dependent failure of miscalibration of temperature controller HV-2457S." In 33 response to the RAI, PSEG explained that this SAMA is not warranted since 1) 34 procedures are already in place to manually control the affected system which, if 35 credited using a failure probability of 0.1, would reduce the RRW for this basic event to 36 1.005, the review threshold, and 2) controller miscalibration would be observed during 37 normal operation (PSEG 2010a).

38 Install flood barriers to address basic event %FL-FPS-5302, "internal flood outside lower 39 relay room." In response to the RAI, PSEG clarified that the ER incorrectly did not 40 identify SAMA 8, "Convert Selected Fire Protection Piping from Wet Pipe to Dry Pipe 41 System," to address this event and further explained that the proposed SAMA is not 42 necessary because the conversion to a dry pipe system was considered preferable to September 2010 G-19 Draft NUREG-1 437, Supplement 45

Appendix G 1

developing flood barriers considering the multiple doors that exist in the corridor outside 2

the relay room (PSEG 2010a).

3 Install a spray shield to address basic event SWS-MOV-VF-SPRAY, "flood - spray 4

causes motor-operated valve (MOV) failure in reactor auxiliaries cooling system (RACS) 5 compartment." In response to the RAI, PSEG explained that the proposed SAMA is not 6

required because the PRA conservatively assumes that all relevant spray events cause 7

failure of the MOVs and that an assumption of 1 in 10 events causing failure would 8

reduce the RRW for this basic event to below the 1.005 review threshold (PSEG 9

2010a).

10 Installation of a passive containment vent to address basic event NR-RHRVENT-INT, 11 "fail to initiate vent given failure to initiate residual heat removal (RHR) in suppression 12 pool cooling (SPC)." This proposed SAMA would also be an alternative to SAMA 4, 13 "Provide Procedural Guidance to Cross-tie RHR Trains." In response to the RAI, PSEG 14 indicated that changing the existing hard pipe venting system to a passive vent design is 15 not considered feasible due to the loss in response flexibility provided by the existing 16 hard pipe venting system and the potential for premature opening of the rupture disks in 17 the passive design (PSEG 2010a). In response to a request for clarification PSEG 18 identified and evaluated SAMA 41, "Installation of Passive Hardened Containment 19 Ventilation Pathway" (PSEG 2010b).

20 In summary, as a result of PSEG's reconsideration of basic events for which no SAMA had 21 been identified in the ER, two new SAMAs were identified: SAMA 41, "Installation of Passive 22 Hardened Containment Ventilation Pathway," and SAMA 42, "Installation of SACS Standby 23 Diesel-Powered Pump." A Phase II cost-benefit evaluation was performed for each of these 24 additional SAMAs, which is discussed in Section G.6.2.

25 In response to an NRC staff RAI, PSEG extended the review down to a RRW of 1.005 to 26 account for a revised external events multiplier of 6.8, which was discussed in Section G.2.2.

27 This extended review identified one additional SAMA as follows: SAMA RAI 5.j-IE1, "Install a 28 Key Lock Switch for Bypass of the MSIV Low Level Isolation Logic" (PSEG 2010a, PSEG 29 2010b). The Phase II cost-benefit evaluation of this SAMA is discussed in Section G.6.2.

30 PSEG also provided and reviewed the Level 2 PRA basic events, down to a RRW of 1.006, for 31 cutsets stated to contribute to large early release. This review did not identify any additional 32 SAMAs. In response to an NRC staff RAI, PSEG revisited this review using only the cutsets 33 from the high and moderate release categories, which contribute over 99 percent of the 34 population dose-risk and offsite economic cost risk (PSEG 2010a). The Level 2 basic events for 35 the remainder of the release categories were not included in the review so as to prevent high 36 frequency-low consequence events from biasing the importance listing. In addition the review 37 was extended down to a RRW of 1.005 to account for a revised external events multiplier of 6.8.

38 The revisited review identified one additional SAMA, not identified in the extended Level 1 39 review discussed above, as follows: SAMA RAI 5p-1, "Install an Independent Boron Injection 40 System." The Phase II cost-benefit evaluation of this SAMA is discussed in Section G.6.2.

September 2010 G-20 Draft NUREG-1437, Supplement 45

Appendix G 1

The NRC staff also requested PSEG to specifically consider the following proposed SAMAs 2

(NRC 2010a):

3

1. Installation of a curb or barrier inside the drywell to prevent early failure of the drywell 4

shell due to shell melt-through. This proposed SAMA addresses basic event CNT-DWV-5 FF-MLTFL, "drywell (DW) shell melt-through failure due to containment failure," for which 6

no SAMA was identified. In response to the RAI, PSEG explained that this proposed 7

SAMA would not be effective in reducing risk because 1) injection is not available and, 8

without cooling, the core debris would degrade the barrier to the point of failure, and 2) 9 an early unscrubbed release pathway is already available as a result of pre-existing 10 containment failures resulting from loss of decay heat removal (PSEG 2010a).

11

2. Replacement of the normally open floor and equipment drain MOVs with fail-closed air-12 operated valves (AOVs). While this proposed SAMA is stated in the ER to be a more 13 costly alternative to SAMA 5, "restore AC power with onsite gas turbine generator," the 14 NRC staff noted in the RAI that it might also be more effective and therefore have a 15 larger benefit. In response to the RAI, PSEG provided a Phase II cost-benefit evaluation 16 of this proposed SAMA, which is discussed in Section G.6.2.

17 One additional SAMA, SAMA 18, "replace a return fan with a different design in service water 18 pump room," was identified in the ER based on a review of PRA insights from the HCGS PRA 19 Group and was identified to address two basic events on the Level 1 basic events importance 20 list.

21 PSEG reviewed the cost-beneficial Phase II SAMAs from prior SAMA analyses for five General 22 Electric BWR and one Westinghouse PWR sites. PSEG's review determined that all but two of 23 the Phase II SAMAs reviewed were either already represented by an existing SAMA, are 24 already implemented at HCGS, have low potential for risk reduction at HCGS, or were not 25 applicable to the HCGS design. This review resulted in two SAMAs being identified by PSEG 26 for HCGS.

27 PSEG's disposition of industry SAMA "auto align 480V AC portable station generator' is stated 28 to be addressed by SAMA 5, "restore AC power with onsite gas turbine generator." The NRC 29 staff noted that the industry SAMA could mitigate events other than those addressed by SAMA 30 5 and requested PSEG to evaluate the industry SAMA (NRC 2010a). In response to an NRC 31 staff RAI PSEG identified and evaluated an additional SAMA to automate the alignment of the 32 portable 480V AC generator (PSEG 201 Oa, PSEG 201 Ob). The cost-benefit evaluation of this 33 additional SAMA is discussed in Section G.6.2.

34 The ER states that an industry SAMA to "develop a procedure to open the door, of the EDG 35 buildings upon the higher temperature alarm" was included in the HCGS SAMA analysis. The 36 NRC staff noted that no such SAMA was evaluated and asked PSEG to clarify this discrepancy 37 (NRC 2010a). In response to the RAI, PSEG explained that this SAMA would not reduce HCGS 38 risk since EDG room cooling issues are small contributors to risk at HCGS and that the 39 statement in the ER is incorrect (PSEG 201 Oa).

40 The NRC asked PSEG to address a SAMA to "increase boron concentration or enrichment in 41 the SLC system," which was determined to be potentially cost-beneficial in the Duane Arnold September 2010 G-21 Draft NUREG-1437, Supplement 45

Appendix G 1

SAMA analysis (NRC 2010a). In response to the RAI, PSEG explained that this SAMA would 2

have a negligible benefit at HCGS because SLC is automatically initiated at HCGS and the 3

basic events the SAMA addresses (related to manual SLC initiation) are not on the importance 4

lists (PSEG 2010a).

5 PSEG considered the potential plant improvements described in the IPE in the identification of 6

plant-specific candidate SAMAs for internal events. Review of the IPE led to no additional 7

SAMA candidates since the three improvements identified in the IPE have already been 8

implemented at HCGS. (PSEG 2009) 9 Based on this information, the NRC staff concludes that the set of SAMAs evaluated in the ER, 10 together with those identified in response to NRC staff RAIs, addresses the major contributors 11 to internal event CDF.

12 Although the IPEEE did not identify any fundamental vulnerabilities or weaknesses related to 13 external events, two improvements related to HFO events were identified. The two 14 improvements have been implemented at HCGS (PSEG 2009). In the ER PSEG also identified 15 three post IPEEE site changes to determine if they could impact the IPEEE results and possibly 16 lead to a SAMA. From this review no additional SAMAs were identified.

17 18 In a further effort to identify external event SAMAs, PSEG identified the top 10 fire scenarios 19 contributing to fire CDF based on the results of the updated HCGS fire PRA model and 20 reviewed the top 8 fire scenarios for potential SAMAs. These 8 scenarios are the only HCGS 21 fire scenarios having a benefit equal to or greater than approximately $100,000, which is the 22 approximate value of implementing a procedure change at a single unit at HCGS. The 23 maximum benefit for a fire area is the dollar value associated with completely eliminating the fire 24 risk in that fire area. SAMAs having an implementation cost of less than that of a procedure 25 change, or $100,000, are unlikely. As a result of this review, PSEG identified six Phase I 26 SAMAs to reduce fire risk. The SAMAs identified included both procedural and hardware 27 alternatives (PSEG 2009). The NRC staff concludes that the opportunity for fire-related SAMAs 28 has been adequately explored and that it is unlikely that there are additional potentially cost-29 beneficial, fire-related SAMA candidates.

30 31 For seismic events, PSEG reviewed the top 10 seismic sequences contributing to seismic CDF 32 based on the results of the 2003 HCGS seismic analysis and initially reviewed the top 2 seismic 33 sequences for potential SAMAs. These two sequences are the only HCGS seismic sequences 34 having a benefit equal to or greater than approximately $100,000, which is the approximate 35 value of implementing a procedure change at a single unit at HCGS. The maximum benefit for 36 a seismic sequence is the dollar value associated with completely eliminating the seismic risk 37 for that sequence. SAMAs having an implementation cost of less than that of a procedure 38 change, or $100,000, are unlikely. As a result of this review, PSEG identified three Phase I 39 SAMAs to reduce seismic risk (PSEG 2009).

40 In response to an NRC staff RAI, PSEG revised the review of seismic sequences to account for 41 the increased maximum benefit of each sequence resulting from the use of the LLNL seismic 42 hazard curve instead of the EPRI curve used initially, as discussed in Section G.2.2. This 43 resulted in two additional seismic sequences having a benefit equal to or greater than the September 2010 G-22 Draft NUREG-1437, Supplement 45

Appendix G 1

$100,000 threshold. As a result of the review of these sequences three additional SAMAs were 2

identified: 1) reinforce 1E 125V DC distribution panels 1AIB/C/D-D-417, 2) reinforce 1E 120V 3

AC distribution panels 1AIB/C/DJ482, and 3) reinforce the 1 E 120V AC 481 distribution panels 4

to 1.Og Seismic Rating (PSEG 2010a, PSEG 2010b). The cost-benefit evaluation of these 5

additional SAMAs is discussed in Section G.6.2.

6 The NRC staff concludes that the opportunity for seismic-related SAMAs has been adequately 7

explored and that it is unlikely that there are additional potentially cost-beneficial, seismic-8 related SAMA candidates.

9 As stated earlier, other external hazards (high winds, external floods, transportation and nearby 10 facility accidents, release of on-site chemicals, and detritus) are below the IPEEE threshold 11 screening frequency, or met the 1975 SRP design criteria, and are not expected to represent 12 vulnerabilities. Nevertheless, PSEG reviewed the IPEEE results and subsequent plant changes 13 for each of these external hazards and determined that either 1) the maximum benefit from 14 eliminating all associated risk was less than approximately $100,000, which is the approximate 15 value of implementing a procedure change at a single unit at HCGS, or 2) only hardware 16 enhancements that would significantly exceed the maximum value of any potential risk 17 reduction were available. As a result of this review, PSEG identified no additional Phase I 18 SAMAs to reduce HFO risk (PSEG 2009). The NRC staff concludes that the licensee's 19 rationale for eliminating other external hazards enhancements from further consideration is 20 reasonable.

21 The NRC staff noted that, while the generic SAMA list from NEI 05-01 (NEI 2005) was stated to 22 have been used in the identification of SAMAs for HCGS, it was not specifically reviewed to 23 identify SAMAs that might be applicable to HCGS but rather was used to identify SAMAs that 24 might address areas of concern identified in the HCGS PRA (NRC 2010a). The NRC staff 25 asked PSEG to provide further information to justify that this approach produced a 26 comprehensive set of SAMAs for consideration. In response to the RAI, PSEG explained that, 27 based on the early SAMA reviews, both the industry and NRC came to realize that a review of 28 the generic SAMA list was of limited benefit because they were consistently found to not be 29 cost-beneficial and that the real benefit was considered to be in the development of SAMAs 30 generated based on plant specific risk insights from the PRA models (PSEG 201 Oa).

31 Furthermore, while the generic list does include potential plant improvements for plants having a 32 similar design to HCGS, plant designs are sufficiently different that the specific plant 33 improvements identified in the generic list are generally not directly applicable to HCGS, and 34 require alteration to specifically address the HCGS design and risk contributors or otherwise 35 would be screened as not applicable to the HCGS design. For these reasons, PSEG concludes 36 that the real value of the NEI 05-01 generic SAMA list is as an idea source to generate SAMAs 37 that address important contributors to HCGS risk. The NRC staff accepts PSEG's conclusion.

38 The NRC staff noted that the 23 Phase I SAMA numbers were not consecutive from 1 to 23, but 39 rather were intermittently numbered between 1 and 40 and requested clarification on the 40 process used to develop the Phase I SAMA list (NRC 2010a). In response to the RAI, PSEG 41 clarified that the original SAMA list was generated from an importance list using the HC108A 42 PRA model, and that review of the subsequent importance list developed using the HC108B September 2010 G-23 Draft NUREG-1437, Supplement 45

Appendix G 1

PRA model determined that certain SAMAs were either no longer applicable or were subsumed 2

into other existing SAMAs (PSEG 2010a). PSEG further clarified that the resulting set of Phase 3

I SAMAs was not renumbered to be consecutive so as to avoid configuration management 4

errors that could occur when working with other documentation and supplemental files. Also, 5

SAMAs identified from the review of external events were given a starting number of 30 so as to 6

avoid overlap with SAMAs developed for internal events.

7 As indicated above two Phase 1 SAMAs were screened out. SAMA 38, "Enhance Fire Water 8

System (FWS) and Automatic Depressurization System (ADS) for Long-term Injection," was 9

screened out on the basis that a procedure has been implemented to address the actions 10 associated with this SAMA. However, as discussed in ER Section E.5.1.7.2.2, this SAMA 11 requires enhancement to the FWS, including strengthening the fire water tanks. In response to 12 an NRC staff RAI, PSEG provided an additional discussion regarding this SAMA and how 13 enhancements to the FWS have been addressed as part of the implementation of the current 14 procedure (PSEG 2010a). The additional discussion indicated that the seismic sequence from 15 which this SAMA originated was a low magnitude earthquake for which there would be a 16 relatively small chance for failure of the FWS. Consequently, strengthening the FWS would 17 have little impact on the sequence and, upon reevaluation, is not needed as part of SAMA 38.

18 PSEG therefore concluded that the procedure implements the remaining requirements of this 19 SAMA.

20 SAMA 14, "Alternate Room Cooling for Service Water (SW) Rooms," was screened out on the 21 basis that it was subsumed into SAMA 4, "cross-tie RHR pump trains." It is described as 22 providing an alternate means of opening Torus Vent Valves, but no basic event in the 23 importance lists is identified as being addressed by this SAMA. In response to an NRC staff 24 RAI, PSEG provided a further discussion of this SAMA and its disposition (PSEG 2010a).

25 SAMA 14 was originally developed to address important containment venting failure events.

26 The importance of these events would be reduced if the need to vent containment is reduced by 27 addressing failure of SW room cooling which leads to loss of containment heat removal. It was 28 subsequently determined that SAMA 4 was the most viable SAMA to address the loss of 29 containment heat removal and SAMA 14 was subsumed into SAMA 4. PSEG also indicated 30 that a loss of SW room cooling could also be addressed by a new SAMA that provides an 31 alternate room cooling strategy for the SW room using procedures and portable fans. A Phase 32 II detailed evaluation was performed for this new SAMA, referred to as SAMA RAI 7.a-1, 33 "enhance procedures and provide additional equipment to respond to loss of all service water 34 pump room supply or return fans" (PSEG 2010a).

35 The NRC staff questioned PSEG about lower cost alternatives to some of the SAMAs evaluated 36 (NRC 2010a), including:

37 0

Establishing procedures for opening doors and/or using portable fans for sequences 38 involving room cooling failures.

39 Extending the procedure for using the B.5.b low pressure pump for non-security 40 events to include all applicable scenarios, not just SBOs.

41 Utilizing a portable independently powered pump to inject into containment.

September 2010 G-24 Draft NUREG-1437, Supplement 45

Appendix G 1

In response to the RAIs, PSEG addressed the suggested lower cost alternatives (PSEG 201 Oa).

2 A new SAMA, SAMA RAI 7.a-1 discussed above, was assessed in a Phase II detailed 3

evaluation for the first item while the other two items are effectively covered by existing 4

procedures. This is discussed further in Section G.6.2.

5 The NRC staff notes that the set of SAMAs submitted is not all-inclusive, since additional, 6

possibly even less expensive, design alternatives can always be postulated. However, the NRC 7

staff concludes that the benefits of any additional modifications are unlikely to exceed the 8

benefits of the modifications evaluated and that the alternative improvements would not likely 9

cost less than the least expensive alternatives evaluated, when the subsidiary costs associated 10 with maintenance, procedures, and training are considered.

11 The NRC staff concludes that PSEG used a systematic and comprehensive process for 12 identifying potential plant improvements for HCGS, and that the set of potential plant 13 improvements identified by PSEG is reasonably comprehensive and, therefore, acceptable.

14 This search included reviewing insights from the plant-specific risk studies, and reviewing plant 15 improvements considered in previous SAMA analyses. While explicit treatment of external 16 events in the SAMA identification process was limited, it is recognized that the prior 17 implementation of plant modifications for fire and seismic risks and the absence of external 18 event vulnerabilities reasonably justifies examining primarily the internal events risk results for 19 this purpose.

20 G.4 Risk Reduction Potential of Plant Improvements 21 22 PSEG evaluated the risk-reduction potential of the 21 remaining SAMAs that were applicable to 23 HCGS, and additional SAMAs identified in response to NRC staff RAIs. The SAMA evaluations 24 were performed using realistic assumptions with some conservatism. On balance, such 25 calculations overestimate the benefit and are, therefore, conservative.

26 PSEG used model re-quantification to determine the potential benefits. The CDF, population 27 dose reductions, and offsite economic cost reductions were estimated using the HCGS PRA 28 model. The changes made to the model to quantify the impact of SAMAs are detailed in 29 Section E.6 of Appendix E to the ER (PSEG 2009). Table G-6 lists the assumptions considered 30 to estimate the risk reduction for each of the evaluated SAMAs, the estimated risk reduction in 31 terms of percent reduction in CDF and population dose, and the estimated total benefit (present 32 value) of the averted risk. The estimated benefits reported in Table G-6 reflect the combined 33 benefit in both internal and external events. The determination of the benefits for the various 34 SAMAs is further discussed in Section G.6.

35 The NRC staff questioned the assumptions used in evaluating the benefit or risk reduction 36 estimate of SAMA 5, "Restore AC Power with Onsite Gas Turbine Generator." The assessment 37 of this SAMA assumed this was equivalent to reducing the probability of failure to cross tie the 38 HCGS emergency diesel generators. This assumption does not provide credit for the gas 39 turbine generator (GTG) in the situation where all the emergency generators are unavailable 40 (NRC 20010a). In response to the RAls, PSEG provided the results of a sensitivity study which 41 the NRC staff subsequently noted did not appear to include credit for the hardware changes 42 included in the cost estimate (NRC 2010b). In response to the request for clarification, PSEG 43 provide the results of a re-evaluation of SAMA 5 that incorporated the additional capability for September 2010 G-25 Draft NUREG-1437, Supplement 45

Appendix G 1

mitigating a more complete set of loss of offsite power initiators consistent with the hardware 2

changes proposed (PSEG 2010b). The revised results are provided in Table G-6.

3 For SAMAs that specifically addressed fire events (i.e., SAMA 30, "Provide Procedural 4

Guidance for Partial Transfer of Control Functions from Control Room to the Remote Shutdown 5

Panel," SAMA 31, "Install Improved Fire Barriers in the Main Control Room (MCR) Control 6

Cabinets Containing the Primary Main Steam Isolation Valve (MSIV) Control Circuits," SAMA 7

32, "Install Additional Physical Barriers to Limit Dispersion of Fuel Oil from Diesel Generator 8

(DG) Rooms," SAMA 33, "Install Division II 480V AC Bus Cross-ties," SAMA 34, "Install Division 9

I 480V AC Bus Cross-ties," and SAMA 35, "Relocate, Minimize and/or Eliminate Electrical 10 Heaters in Electrical Access Room"), the reduction in fire CDF and population dose was not 11 directly calculated (in Table G-6 this is noted as "Not Estimated"). For these SAMAs, an 12 estimate of the impact was made based on general assumptions regarding: the approximate 13 contribution to total risk from external events relative to that from internal events; the fraction of 14 the external event risk attributable to fire events; the fraction of the fire risk affected by the 15 SAMA (based on information from the 2003 HCGS External Events Update); and the 16 assumption that the SAMA eliminates 90 percent (SAMAs 30, 32, 33, and 34), 99 percent 17 (SAMA 35), or all (SAMA 31) of the fire risk affected by the SAMA. Specifically, it is assumed 18 that the contribution to risk from external events is approximately 5.3 times that from internal 19 events, and that internal fires contribute 74 percent of this external events risk. The fire basic 20 events impacted by the SAMA are identified and the portion of the total fire risk contributed by 21 each of these fire basic events determined. For SAMA 31, the benefit or averted cost risk from 22 reducing the fire risk affected by the SAMA is then calculated by multiplying the ratio of the fire 23 risk affected by the SAMA to the internal events CDF by the total present dollar value equivalent 24 associated with completely eliminating severe accidents from internal events at HCGS. For the 25 other fire SAMAs, the benefit or averted cost risk from reducing the fire risk affected by the 26 SAMA is then calculated by multiplying the ratio of 90 percent, or 99 percent (SAMA 35), of the 27 fire risk affected by the SAMA to the internal events CDF by the total present dollar value 28 equivalent associated with completely eliminating severe accidents from internal events at 29 HCGS. These SAMAs were assumed to have no additional benefits in internal events.

30 31 The NRC staff questioned the calculated impact for SAMA 35 which assumed that 90 percent of 32 the fire risk affected by the SAMA was eliminated rather than the 99 percent stated in the ER 33 (NRC 2010a). In response to the RAI, PSEG provided a revised evaluation using 99 percent 34 (PSEG 2010a). The revised results are provided in Table G-6.

35 36 For SAMAs that specifically addressed seismic events (i.e., SAMA 36, "Provide Procedural 37 Guidance for Loss of All 1 E 120V AC Power," and SAMA 37, "Reinforce 1 E 120V AC 38 Distribution Panels") the reduction in seismic CDF and population dose also was not directly 39 calculated. As was done for fire SAMAs, an estimate of the impact of seismic SAMAs was 40 made based on general assumptions regarding: the approximate contribution to total risk from 41 external events relative to that from internal events; the fraction of the external event risk 42 attributable to seismic events; the fraction of the seismic risk affected by the SAMA (based on September 2010 G-26 Draft NUREG-1437, Supplement 45

Appendix G 1

information from the 2003 HCGS External Events Update); and the assumption that the SAMA 2

eliminates 50 percent (SAMA 36) or 90 percent (SAMA 37) of the seismic risk affected by the 3

SAMA. Specifically, it is assumed that the contribution to risk from external events is 4

approximately 5.3 times that from internal events, and that seismic events contribute 5 percent 5

of this external events risk. The seismic basic events impacted by the SAMA are identified and 6

the portion of the total seismic risk contributed by each of these seismic basic events 7

determined. The benefit or averted cost risk from reducing the seismic risk affected by the 8

SAMA is then calculated by multiplying the ratio of 50 percent (SAMA 36), or 90 percent (SAMA 9

37), of the seismic risk affected by the SAMA to the internal events CDF by the total present 10 dollar value equivalent associated with completely eliminating severe accidents from internal 11 events at HCGS. These SAMAs were assumed to have no additional benefits in internal 12 events.

13 14 The NRC staff has reviewed PSEG's bases for calculating the risk reduction for the various 15 plant improvements and concludes, with the above clarifications, that the rationale and 16 assumptions for estimating risk reduction are reasonable and generally conservative (i.e., the 17 estimated risk reduction is higher than what would actually be realized). Accordingly, the NRC 18 staff based its estimates of averted risk for the various SAMAs on PSEG's risk reduction 19 estimates.

September 2010 G-27 Draft NUREG-1437, Supplement 45

Appendix G Table G-6. SAMA Cost/Benefit Screening Analysis for HCGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Dose (Internal +

Uncertainty(e)

SAMA Assumptions External)

I - Remove Automatic The probability that operators fail to 26 29 5.3M 14.9M 200K Depressurization System (ADS) Inhibit inhibit ADS was reduced to 0.1 from from Non-ATWS Emergency Operating 1.0.

Procedures 3 - Install Back-up Air Compressor to The probability that operators fail to 16 16 3.3M 9.4M 700K Supply AOVs restore service water was reduced to 0.5 from 1.0.

4 - Provide Procedural Guidance to The probability that operators fail to 12 21 4.4M 12.4M 100K Cross-Tie RHR Trains recover RHR was reduced to 0.1 from 0.35.

51b) - Restore AC Power with Onsite The probability that operators fail to 9

11 2.2M 6.3M 2.05M Gas Turbine Generator cross-tie the emergency diesel generators (EDGs) was reduced to 0.1 from 1.0. In response to an NRC staff RAI, the GTG failure probability, maintenance unavailability, and human error probability were set to 0.

7 - Install Better Flood Protection The probability that operators fail to 4

2 330K 930K 3.07M Instrumentation for Reactor Auxiliaries isolate locally a service water rupture in Cooling System (RACS) Compartment the RACS compartment was reduced to 0.1 from 1.0.

8 - Convert Selected Fire Protection The probability that operators fail to 4

1 300K 860K 600K Piping from Wet to Dry Pipe System isolate a fire protection header leak was reduced to 0.1 from 1.0.

September 2010 G-28 Draft NUREG-1437, Supplement 45

Appendix G Table G-6. SAMA Cost/Benefit Screening Analysis for HCGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Popuato (internal +

Uncertainty(e)

SAMA Assumptions Dose External) 10 - Provide Procedural Guidance to The probability that operators fail to 1

1 200K 570K 100K use B.5.b Low Pressure Pump for align residual heat removal service Non-Security Events water (RHRSW) for injection into the reactor pressure vessel (RPV) was reduced to 1.OE-02 from 1.OE-01.

15 - Alternate Design of Core Spray The probability that operators fail to 2

1 130K 360K 1.0M System (CSS) Suction Strainer to locally open each of the service water Mitigate Plugging valves was reduced to 8.36E-04 from 8.36E-03.

16 - Use of Different Designs for The probability that FANS AVH401 2

1 130K 370K 400K Switchgear Room Cooling Fans through DVH400 fail-to-start and fail-to-run was set to 0.

17 - Replace a Supply Fan with a The probability that FANS AV503 5

5 960K 2.7M 600K Different Design in Service Water through DV503 fail-to-start and fail-Pump Room to-run was set to 0.

18 - Replace a Return Fan with a The probability that FANS AV504 5

5 960K 2.7M 600K Different Design in Service Water through DV504 fail-to-start and fail-Pump Room to-run was set to 0.

30 - Provide Procedural Guidance for Reduce the fire CDF contribution NOT ESTIMATED 8.6M 24M 100K Partial Transfer of Control Functions from Fire Basic Events %IE-FIRE03, from Control Room to the Remote

%IE-FIRE02, and %IE-FIRE01 by 90 Shutdown Panel percent.

31 - Install Improved Fire Barriers in the Eliminate the fire CDF contribution from NOT ESTIMATED 360K 1.0M 1.2M Main Control Room (MCR) Control Fire Basic Event %IE-FIRE06.

Cabinets Containing the Primary Main Steam Isolation Valve (MSIV) Control Circuits September 2010 G-29 Draft NUREG-1437, Supplement 45

Appendix G Table G-6. SAMA Cost/Benefit Screening Analysis for HCGS(a)

% Risk Reduction Total Benefit ($)

Population Baseline Baseline With Cost ($)

CDF Dose ternal Uncertainty()

SAMA Assumptions External) 32 - Install Additional Physical Reduce the fire CDF contribution NOT ESTIMATED 480K 1.4M 800K Barriers to Limit Dispersion of Fuel Oil from Fire Basic Event %IE-FIRE28 by from Diesel Generator (DG) Rooms 90 percent.

33 - Install Division II 480V AC Bus Reduce the fire CDF contribution from NOT ESTIMATED 450K 1.3M 1.32M Cross-ties Fire Basic Event %IE-FIRE37 by 90 percent.

34 - Install Division I 480V AC Bus Reduce the fire CDF contribution from NOT ESTIMATED 430K 1.2M 1.32M Cross-ties Fire Basic Event %IE-FIRE20 by 90 percent.

35 - Relocate, Minimize andlor Reduce the fire CDF contribution NOT ESTIMATED 410K~c) 1.2M'c) 270K Eliminate Electrical Heaters in from Fire Basic Event %IE-FIRE38 by Electrical Access Room 99 percent.

36 - Provide Procedural Guidance for Reduce the seismic CDF contribution NOT ESTIMATED 240K 680K 270K Loss of All 1 E 120V AC Power from Seismic Basic Event %IE-SET36 by 50 percent.

37 - Reinforce 1E 120V AC Reduce the seismic CDF contribution NOT ESTIMATED 430K 1.2M 500K Distribution Panels from Seismic Basic Event %IE-SET36 by 90 percent.

39 - Provide Procedural Guidance to As provided in response to an NRC 10

<1 130K 380K 120K Bypass Reactor Core Isolation staff RAI, modify fault tree to include Cooling (RCIC) Turbine Exhaust a new operator action, having a Pressure Trip failure probability of 1.0E-02, representing failure of the operator to defeat the HPCI/RCIC back pressure permissive.

September 2010 G-30 Draft NUREG-1437, Supplement 45

Appendix G Table G-6. SAMA Cost/Benefit Screening Analysis for HCGS(a)

% Risk Reduction Total Benefit ($)

Baseline Baseline With Cost ($)

CDF Dose (internal +

Uncertainty(e)

SAMA Assumptions External) 40 - Increase Reliability/Install Manual As provided in response to an NRC staff 1

1 210K 610K 620K Bypass of Low Pressure (LP) Permissive RAI, the probability of common cause mis-calibration of all ECCS pressure transmitters was reduced to 8.0E-06 from 8.OE-05.

41(d) _ Installation of Passive Hardened A completely passive containment vent 15 30 6.2M 18M

>25M Containment Ventilation Pathway system requiring no operator actions is assumed.

42(() - Installation of SACS Standby Reduce the probability of initiating event 2

1 270K 760K 6.2M Diesel-Powered Pump

%IE-SACS to 1.161E-05 per year from 1.16E-04 per year.

(a)

(b)

(c)

(d)

(e)

SAMAs in bold are potentially cost-beneficial.

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

SAMAs 30, 31, and 32 were identified and evaluated in response to an NRC staff RAI (PSEG 2010a). 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.

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

Using a factor of 2.5.

September 2010 G-31 Draft NUREG-1437, Supplement 45

Appendix G 1

G.5 Cost Impacts of Candidate Plant Improvements 2

3 PSEG estimated the costs of implementing the 21 candidate SAMAs through the development 4

of site-specific cost estimates. The cost estimates conservatively did not include the cost of 5

replacement power during extended outages required to implement the modifications, nor did 6

they include contingency costs for unforeseen difficulties (PSEG 2010a). The cost estimates 7

provided in the ER did not account for inflation, which is considered another conservatism.

8 The NRC staff reviewed the bases for the applicant's cost estimates (presented in Table E.5-3 9

of Attachment E to the ER). For certain improvements, the NRC staff also compared the cost 10 estimates to estimates developed elsewhere for similar improvements, including estimates 11 developed as part of other licensees' analyses of SAMAs for operating reactors.

12 13 The ER stated that plant personnel developed HCGS-specific costs to implement each of the 14 SAMAs. The NRC staff requested more information on the process PSEG used to develop the 15 SAMA cost estimates (NRC 2010a). PSEG responded to the RAI by explaining that the cost 16 estimates were developed in a series of meetings involving personnel responsible for 17 development of the SAMA analysis and the two PSEG license renewal site leads who are 18 engineering managers each having over 25 years of plant experience, including project 19 management, operations, plant engineering, design engineering, procedure support, simulators, 20 and training (PSEG 2010a). During these meetings, each SAMA was validated against the 21 plant configuration, a budget-level estimate of its implementation cost was developed, and, in 22 some instances, lower cost approaches that would achieve the same objective were developed.

23 The SAMA implementation costs were then reviewed by the Design Engineering Manager for 24 both technical and cost perspectives and revised accordingly. PSEG further explained that 25 seven general cost categories were used in development of the budget-level cost estimates:

26 engineering, material, installation, licensing, critical path impact, simulator modification, and 27 procedures and training. Based on the use of personnel having significant nuclear plant 28 engineering and operating experience, the NRC staff considers the process PSEG used to 29 develop budget-level cost estimates reasonable.

30 31 The NRC staff requested additional clarification on the estimated cost of $2.05M for 32 implementation of SAMA 5, "Restore AC Power with Onsite Gas Turbine Generator," and on the 33 implementation cost of $270K for implementation of SAMA 36, "Provide Procedural Guidance 34 for Loss of All 1 E 120V AC Power," which are high for what are described as procedure 35 changes and operator training (NRC 2010a). In response to an RAI, PSEG further described 36 the SAMA 5 modification as providing the necessary equipment to connect a dedicated 37 transformer at Salem Unit 3 to HCGS, which is significantly more costly than, and is in addition 38 to, the procedure changes (PSEG 2010a). It was also explained that the SAMA 5 modification 39 assumes that Salem Generating Station (SGS) SAMA 2 to install the dedicated transformer is 40 already implemented and that SAMA 5 is a safety-related permanent plant modification. In 41 response to a different RAI, PSEG explained that the SAMA 36 modification involves the September 2010 G-32 Draft NUREG-1437, Supplement 45

Appendix G 1

development of a group of procedures, not just the revision of existing procedures or the 2

development of a single procedure. In addition, there is a significant effort involved with 3

determining a success path to achieve safe shutdown, to update the simulator to include all 4

necessary components to implement the success path, to test the success path, and to 5

implement the new procedures. Based on this additional information, the NRC staff considers 6

the estimated cost to be reasonable and acceptable for purposes of the SAMA evaluation.

7 8

The NRC staff asked PSEG to justify the estimated cost of $1 00K for SAMA 10, "Provide 9

Procedural Guidance to use B.5.b Low Pressure Pump for Non-Security Events," for what is 10 described as including a new pump when $100K is the estimated cost of a procedure change 11 used in the SAMA analysis (NRC 2010a). PSEG responded that the cost estimate for SAMA 10 12 assumes that an existing pump already installed at HCGS will be made available to implement 13 this SAMA (PSEG 2010a). Based on this additional information, the NRC staff considers the 14 estimated cost to be reasonable and acceptable for purposes of the SAMA evaluation.

15 16 In response to an RAI requesting a more detailed description of the changes associated with 17 SAMA 16, "Use of Different Designs for Switchgear Room Cooling Fans," PSEG provided 18 additional information detailing the cost estimate of this improvement (PSEG 2010a). The staff 19 reviewed the cost estimate and found it to be reasonable, and generally consistent with 20 estimates provided in support of other plants' analyses.

21 22 The NRC staff noted that SAMA 31, "Install Improved Fire Barriers in the Main Control Room 23 (MCR) Control Cabinets Containing the Primary Main Steam Isolation Valve (MSIV) Control 24 Circuits," is similar to SGS SAMAs 21 and 22 in that each involves installing fire barriers to 25 prevent the propagation of a fire between cabinets and requested an explanation for why the 26 estimated cost of $1.2M for SAMA 31 to modify one cabinet is similar to the estimated cost of 27

$1.6M for SGS SAMA 22 to modify three Control Room consoles and is more than one-third of 28 the $3.23M cost for SGS SAMA 21 to modify 48 Relay Room cabinets (NRC 2010a)! PSEG 29 responded that making the modifications to the SAMA 31 Control Room console, which is 30 estimated to be $400K for materials and installation, is more complicated than making 31 modifications to the SGS SAMA 21 Relay Room cabinets, which is estimated to be $35K to 32

$70K for materials and maintenance (PSEG 2010a). Specifically, SAMA 31 requires making 33 ventilation modifications due to the significant heat loads in addition to adding fire barrier 34 materials. PSEG also explained that both SAMA 31 and SGS SAMA 22 assumed the same 35 material and installation cost per console ($400K) and the same engineering cost ($800K) but 36 that the engineering cost was evenly divided between the two units at SGS to arrive at a cost 37 per unit. The NRC staff considers the basis for the differences in cost estimates reasonable.

38 39 The NRC staff noted that the estimated cost of $620K for SAMA 40, "Increase Reliability/Install 40 Manual Bypass of Low Pressure (LP) Permissive," is significantly higher than the estimated cost 41 of $250K for a similar improvement evaluated for the Duane Arnold nuclear power plant license 42 renewal application (NRC 2010a). In response to the RAI, PSEG clarified that SAMA 40 September 2010 G-33 Draft NUREG-1437, Supplement 45

Appendix G 1

involves the installation of six key-lock switches to bypass various low pressure submissives 2

(PSEG 201 Oa). Key-lock switches are used rather than jumpers, as was assumed in the Duane 3

Arnold application, because the benefit of this SAMA cannot be obtained otherwise due to the 4

effort required to install six jumpers, which is a more time intensive action than the time required 5

to operate key-lock switches. Based on this additional information, the NRC staff considers the 6

estimated cost for HCGS to be reasonable and acceptable for purposes of the SAMA 7

evaluation.

8 9

The NRC staff also noted that the estimated cost of $1.32M each for SAMA 33, "Install Division 10 II 480V AC Bus Cross-ties," and SAMA 34, "Install Division I 480V AC Bus Cross-ties," is 11 significantly higher than the estimated cost of $328K to $656K for a similar improvement 12 evaluated for other nuclear power plant license renewal applications, i.e., Wolf Creek and 13 Susquehanna (NRC 2010a). In response to the RAI, PSEG described these modifications as 14 involving the installation of new tie-breakers and cables for the 480V AC bus cross-ties, having 15 a material and installation cost of $400K (PSEG 2010a). The most significant cost was for 16 engineering, which was estimated to be $800K due to the electrical load analysis required to 17 support the cross-ties. Based on this additional information, the NRC staff considers the basis 18 for the estimated cost to be reasonable.

19 20 The NRC staff concludes that the cost estimates provided by PSEG are sufficient and 21 appropriate for use in the SAMA evaluation.

22 G.6 Cost-Benefit Comparison 23 24 PSEG's cost-benefit analysis and the NRC staff's review are described in the following sections.

25 26 G.6.1 PSEG's Evaluation 27 28 The methodology used by PSEG was based primarily on NRC's guidance for performing cost-29 benefit analysis, i.e., NUREG/BR-0184, Regulatory Analysis Technical Evaluation Handbook 30 (NRC 1997a). The guidance involves determining the net value for each SAMA according to 31 the following formula:

32 Net Value = (APE + AOC + AOE + AOSC) - COE, where 33 APE = present value of averted public exposure ($)

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

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

36 AOSC = present value of averted onsite costs ($)

37 COE = cost of enhancement ($)

38 39 If the net value of a SAMA is negative, the cost of implementing the SAMA is larger than the 40 benefit associated with the SAMA and it is not considered cost-beneficial. PSEG's derivation of 41 each of the associated costs is summarized below.

September 2010 G-34 Draft NUREG-1437, Supplement 45

Appendix G 1

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

2 Revision 4 of NUREG/BR-0058 states that two sets of estimates should be developed, one at 3

3 percent and one at 7 percent (NRC 2004). PSEG performed the SAMA analysis using the 4

3 percent discount rate and a sensitivity study using the 7 percent discount rate (PSEG 2009).

5 Averted Public Exposure (APE) Costs 6

The APE costs were calculated using the following formula:

7 APE = Annual reduction in public exposure (Aperson-rem/year) 8 x monetary equivalent of unit dose ($2,000 per person-rem) 9 x present value conversion factor (15.04 based on a 20-year period with a 10 3-percent discount rate) 11 As stated in NUREG/BR-0184 (NRC 1997a), it is important to note that the monetary value of 12 the public health risk after discounting does not represent the expected reduction in public 13 health risk due to a single accident. Rather, it is the present value of a stream of potential 14 losses extending over the remaining lifetime (in this case, the renewal period) of the facility.

15 Thus, it reflects the expected annual loss due to a single accident, the possibility that such an 16 accident could occur at any time over the renewal period, and the effect of discounting these 17 potential future losses to present value. For the purposes of initial screening, which assumes 18 elimination of all severe accidents, PSEG calculated an APE of approximately $688,000 for the 19 20-year license renewal period.

20 Averted Offsite Property Damaqe Costs (AOC) 21 22 The AOCs were calculated using the following formula:

23 AOC = Annual CDF reduction 24 x offsite economic costs associated with a severe accident (on a per-event basis) 25 x present value conversion factor.

26 This term represents the sum of the frequency-weighted offsite economic costs for each release 27 category, as obtained for the Level 3 risk analysis. For the purposes of initial screening, which 28 assumes elimination of all severe accidents caused by internal events, PSEG calculated an 29 AOC of about $155,000 based on the Level 3 risk analysis. This results in a discounted value of 30 approximately $2,332,000 for the 20-year license renewal period.

31 Averted Occupational Exposure (AOE) Costs 32 33 The AOE costs were calculated using the following formula:

34 AOE = Annual CDF reduction 35 x occupational exposure per core damage event 36 x monetary equivalent of unit dose 37 x present value conversion factor September 2010 G-35 Draft NUREG-1437, Supplement 45

Appendix G 1

PSEG derived the values for averted occupational exposure from information provided in 2

Section 5.7.3 of the regulatory analysis handbook (NRC 1997a). Best estimate values provided 3

for immediate occupational dose (3,300 person-rem) and long-term occupational dose (20,000 4

person-rem over a 10-year cleanup period) were used. The present value of these doses was 5

calculated using the equations provided in the handbook in conjunction with a monetary 6

equivalent of unit dose of $2,000 per person-rem, a real discount rate of 3 percent, and a time 7

period of 20 years to represent the license renewal period. For the purposes of initial screening, 8

which assumes elimination of all severe accidents caused by internal events, PSEG calculated 9

an AOE of approximately $2,700 for the 20-year license renewal period (PSEG 2009).

10 Averted Onsite Costs 11 12 Averted onsite costs (AOSC) include averted cleanup and decontamination costs and averted 13 power replacement costs. Repair and refurbishment costs are considered for recoverable 14 accidents only and not for severe accidents. PSEG derived the values for AOSC based on 15 information provided in Section 5.7.6 of NUREG/BR-0184, the regulatory analysis handbook 16 (NRC 1997a).

17 PSEG divided this cost element into two parts - the onsite cleanup and decontamination cost, 18 also commonly referred to as averted cleanup and decontamination costs (ACC), and the 19 replacement power cost (RPC).

20 ACCs were calculated using the following formula:

21 ACC

= Annual CDF reduction 22 x present value of cleanup costs per core damage event 23 x present value conversion factor 24 25 The total cost of cleanup and decontamination subsequent to a severe accident is estimated in 26 NUREG/BR-0184 to be $1.5 x 109 (undiscounted). This value was converted to present costs 27 over a 10-year cleanup period and integrated over the term of the proposed license extension.

28 For the purposes of initial screening, which assumes elimination of all severe accidents caused 29 by internal events, PSEG calculated an ACC of approximately $87,000 for the 20-year license 30 renewal period.

31 32 Long-term RPCs were calculated using the following formula:

33 34 RPC

= Annual CDF reduction 35 x present value of replacement power for a single event 36 x factor to account for remaining service years for which replacement power is 37 required 38 x reactor power scaling factor 39 40 PSEG based its calculations on a HCGS net output of 1287 megawatt electric (MWe) and 41 scaled up from the 910 MWe reference plant in NUREG/BR-0184 (NRC 1997a). Therefore 42 PSEG applied a power scaling factor of 1287/910 to determine the replacement power costs.

43 For the purposes of initial screening, which assumes elimination of all severe accidents caused September 2010 G-36 Draft NUREG-1437, Supplement 45

Appendix G 1

by internal events, PSEG calculated an RPC of approximately $35,000 and an AOSC of 2

approximately $122,000 for the 20-year license renewal period.

3 4

Using the above equations, PSEG estimated the total present dollar value equivalent associated 5

with completely eliminating severe accidents from internal events at HCGS to be about $3.14M.

6 Use of a multiplier of 6.3 to account for external events increases the value to $19.8M and 7

represents the dollar value associated with completely eliminating all internal and external event 8

severe accident risk for a single unit at HCGS, also referred to as the Maximum Averted Cost 9

Risk (MACR).

10 11 PSEG's Results 12 13 If the implementation costs for a candidate SAMA exceeded the calculated benefit, the SAMA 14 was considered not to be cost-beneficial. In the baseline analysis contained in the ER (using a 15 3 percent discount rate, and considering the impact of external events), PSEG identified nine 16 potentially cost-beneficial SAMAs. PSEG performed additional analyses to evaluate the impact 17 of parameter choices (alternative discount rates and variations in MACCS2 input parameters) 18 and uncertainties on the results of the SAMA assessment and, as a result of this analysis, 19 identified four additional potentially cost-beneficial SAMAs.

20 21 The potentially cost-beneficial SAMAs are:

22 23 0

SAMA 1 - remove ADS Inhibit from Non-ATWS Emergency Operating Procedures 24 0

SAMA 3 - Install Back-Up Air Compressor to Supply AOVs 25 0

SAMA 4 - Provide Procedural Guidance to Cross-Tie RHR Trains 26 0

SAMA 8 - Convert Selected Fire Protection Piping from Wet to Dry Pipe System 27 0

SAMA 10 - Provide Procedural Guidance to Use B.5.b Low Pressure Pump for Non-28 Security Events 29 0

SAMA 17 - Replace a Supply Fan with a Different Design in Service Water Pump Room 30 0

SAMA 18 - Replace a Return Fan with a Different Design in Service Water Pump Room 31 0

SAMA 30 - Provide Procedural Guidance for Partial Transfer of Control Functions from 32 the Control Room to the Remote Shutdown Panel 33 0

SAMA 32 - Install Additional Physical Barriers to Limit Dispersion of Fuel Oil from DG 34 Rooms September 2010 G-37 Draft NUREG-1437, Supplement 45

Appendix G 1

0 SAMA 35 - Relocate, Minimize, and/or Eliminate Electrical Heaters in Electrical Access 2

Room 3

0 SAMA 36 - Provide Procedural Guidance for Loss of All 1 E 120V AC Power 4

0 SAMA 37 - Reinforce I E 120V AC Distribution Panels 5

SAMA 39 - Provide Procedural Guidance to Bypass RCIC Turbine Exhaust Pressure 6

Trip 7

PSEG indicated that they plan to further evaluate these SAMAs for possible implementation 8

using existing HCGS Plant Heal Committee processes (PSEG 2009).

9 10 The potentially cost-beneficial SAMAs, and PSEG's plans for further evaluation of these 11 SAMAs, are discussed in detail in Section G.6.2.

12 13 G.6.2 Review of PSEG's Cost-Benefit Evaluation 14 15 The cost-benefit analysis performed by PSEG was based primarily on NUREG/BR-0184 16 (NRC 1997a) and discount rate guidelines in NUREG/BR-0058 (NRC 2004) and was executed 17 consistent with this guidance.

18 SAMAs identified primarily on the basis of the internal events analysis could provide benefits in 19 certain external events, in addition to their benefits in internal events. To account for the 20 additional benefits in external events, PSEG multiplied the internal event benefits for each 21 internal event SAMA by a factor of 6.3, which is the ratio of the total CDF from internal and 22 external events to the internal event CDF. As discussed in Section G.2.2, this factor was based 23 on a seismic CDF of 1.1 x 10-6 per year, plus a fire CDF of 1.7 x 10-5 per year, plus the 24 screening values for high winds, external flooding, transportation, detritus, and chemical release 25 events (1 x 10-6 per year for each). The external event CDF of 2.3 x 10- per year is thus 5.3 26 times the internal events release frequency CDF of 4.4 x 10-6 per year. The total CDF is thus 27 6.3 [(2.3 x 10s + 4.4 x 10-6) / 4.4 x 10-6] times the internal events release frequency CDF (PSEG 28 2009). Seven SAMAs were determined to be cost-beneficial in PSEG's analysis (SAMAs 1, 3, 29 4, 10, 17, 18, and 39 as described above).

30 PSEG did not multiply the internal event benefits by the factor of 6.3 for eight SAMAs that 31 specifically address fire and seismic risk (SAMAs 30, 31, 32, 33, 34, 35, 36, and 37).

32 Multiplying the internal event benefits by 6.3 for these SAMAs would not be appropriate 33 because these SAMAs are specific to fire or seismic risks and would not have a corresponding 34 benefit on the risk from internal events. Two of these SAMAs were found to be cost-beneficial in 35 PSEG's analysis (SAMAs 30 and 35, as described above).

36 PSEG considered the impact that possible increases in benefits from analysis uncertainties 37 would have on the results of the SAMA assessment. In the ER, PSEG presents the results of September 2010 G-38 Draft NUREG-1437, Supplement 45

Appendix G 1

an uncertainty analysis of the internal events CDF which indicates that the 9 5 1h percentile value 2

is a factor of 2.84 times the point estimate CDF for HCGS. Since the two Phase I SAMAs that 3

were screened based on qualitative criteria were screened due to one being subsumed into 4

another SAMA or one having already been implemented at HCGS, a re-examination of the 5

Phase I SAMAs based on the upper bound benefits was not necessary. PSEG considered the 6

impact on the Phase II analysis if the estimated benefits were increased by a factor of 2.84 (in 7

addition to the multiplier of 6.3 for external events). Four additional SAMAs became cost-8 beneficial in PSEG's analysis (SAMAs 8, 32, 36, and 37 as described above).

9 PSEG provided the results of additional sensitivity analyses in the ER, including use of a 7 10 percent discount rate and variations in MACCS2 input parameters. These analyses did not 11 identify any additional potentially cost-beneficial SAMAs (PSEG 2009).

12 PSEG indicated that the 13 potentially cost-beneficial SAMAs (SAMAs 1, 3, 4, 8, 10, 17, 18, 30, 13 32, 35, 36, 37, and 39) will be considered for implementation through the established HCGS 14 Plant Health Committee process (PSEG 2009).

15 As indicated in Section G.3.2, in response to NRC staff RAIs, PSEG considered additional plant 16 improvements to address basic events for which no SAMAs had been identified in the ER.

17 PSEG determined that of the plant improvements considered, two additional SAMAs warrant 18 further consideration: 1) SAMA 41, "Installation of Passive Hardened Containment Ventilation 19 Pathway," and 2) SAMA 42, "Installation of SACS Standby Diesel-Powered Pump." Each of 20 these new SAMAs is included in Table G-6 and were evaluated as described above. PSEG's 21 analysis determined that neither of these SAMA candidates was cost-beneficial in either the 22 baseline analysis or the uncertainty analysis.

23 As indicated in Section G.2.2, PSEG determined that the external events multiplier would be 6.8 24 if the higher seismic CDF obtained using the LLNL hazard curves were used rather than the 25 EPRI hazard curves. As discussed in Section G.3.2, PSEG then reviewed the Level 1 and 26 Level 2 basic events down to an RRW of 1.005 to account for the revised external events 27 multiplier of 6.8. In addition, since the maximum benefit of each seismic sequence increased as 28 a result of using the LLNL hazard curves, PSEG reviewed two additional seismic sequences 29 having a benefit equal to or greater than $100,000, the minimum expected SAMA 30 implementation cost at HCGS. These reviews resulted in the identification and evaluation of 31 five additional SAMAs, as summarized below:

32 SAMA RAI 5.j-IE1, "Install a Key Lock Switch for Bypass of the Main Steam Isolation 33 Valve (MSIV) Low Level Isolation Logic." PSEG estimated the implementation cost for 34 this SAMA to be the same as SAMA 40, "Increase Reliability/Install Manual Bypass of 35 Low Pressure (LP) Permissive," or $620K, which also involved installation of key lock 36 bypass switches (PSEG 201 Oa). The maximum benefit was estimated to be $1 10K in 37 the baseline analysis, and $300K after accounting for uncertainties, which assumed that 38 the risk of the basic event addressed by this SAMA was completely eliminated. Since September 2010 G-39 Draft NUREG-1437, Supplement 45

Appendix G 1

the implementation cost was greater than the estimated benefit accounting for 2

uncertainties, PSEG concluded that SAMA RAI 5.j-IE1 was not cost-beneficial.

3

• SAMA RAI 5p-1, "Install an Independent Boron Injection System." PSEG estimated the 4

implementation cost of this SAMA to be $1.5M based on the estimate for a similar SAMA 5

to install a redundant system evaluated in the Browns Ferry nuclear power plant license 6

renewal application and the estimated cost to install an additional tank (PSEG 2010a).

7 To estimate the risk reduction, PSEG modified the HCGS PRA model fault tree to 8

include a new basic event, having a failure probability of 1.OE-03, representing failure of 9

the redundant system. The benefit was estimated to be $390K in the baseline analysis, 10 and $1.1 M after accounting for uncertainties. Since the implementation cost was greater 11 than the estimated benefit accounting for uncertainties, PSEG concluded that SAMA RAI 12 5p-1 was not cost-beneficial.

13 Reinforce 1E 125V DC distribution panels 1AIBIClD-D-417. PSEG estimated the 14 minimum implementation cost for this SAMA to be the same as SAMA 37, "Reinforce 1 E 15 120V AC Distribution Panels," or $500K, but expects the cost to be higher because 16 these panels have a much higher HCLPF value than the SAMA 37 120V AC panels 17 (PSEG 2010a). To estimate the risk reduction, PSEG assumed that the contribution to 18 risk from external events is approximately 5.8 times that from internal events (based on 19 a revised seismic CDF of 3.58 x 10.6 per year using the LLNL hazard curves), that 20 seismic events contribute 14 percent of this external events risk, and that 50 percent of 21 the fire risk affected by the SAMA is eliminated. The benefit was estimated to be $155K 22 in the baseline analysis, and $440K after accounting for uncertainties. Since the 23 implementation cost was greater than the estimated benefit accounting for uncertainties, 24 PSEG concluded that this SAMA was not cost-beneficial.

25 Reinforce 1E 120V AC distribution panels 1AIB/C/DJ482. PSEG estimated the 26 implementation cost for this SAMA to be the same as SAMA 37, or $500K, which also 27 addresses 120V AC panels (PSEG 2010a). To estimate the risk reduction, PSEG 28 assumed that the contribution to risk from external events is approximately 5.8 times that 29 from internal events (based on a revised seismic CDF of 3.58 x 10.6 per year using the 30 LLNL hazard curves), that seismic events contribute 14 percent of this external events 31 risk, and that all of the seismic risk affected by the SAMA is eliminated. The benefit was 32 estimated to be $1 10K in the baseline analysis, and $320K after accounting for 33 uncertainties. Since the implementation cost was greater than the estimated benefit 34 accounting for uncertainties, PSEG concluded that this SAMA was not cost-beneficial.

35 Reinforce 1 E 120V AC distribution panels to 1.Og Seismic Rating. This SAMA assumes 36 that 1) SAMA 37 is implemented, 2) the HCLPF values for the 120V AC panels are 37 further increased to 1 g as a result of the implementation, 3) the above SAMA to 38 reinforce the 125V DC panels is implemented, and 4) the HCLPF values for the panels 39 are increased from the current 0.57g to 1.0g as a result of the implementation (PSEG September 2010 G-40 Draft NUREG-1437, Supplement 45

Appendix G 1

2010b). SAMA 37 originally was assumed to reduce the risk of seismic basic event %IE-2 SET36, "seismic-induced equipment damage state SET-36 (impacts - 120V PNL481,"

3 by 90 percent while the proposed SAMA to reinforce the 125V DC panels, by itself was 4

originally assumed to reduce the risk of seismic basic event %IE-SET37, seismic-5 induced equipment damage state (impacts - 125V)," by 50 percent. The synergistic 6

benefit of this new proposed SAMA to reinforce the 120V AC panels to a HCLPF value 7

of 1.0g is assumed to be the sum of the benefit to eliminate the remaining 10 percent of 8

the risk of event %IE-SET36 ($176K) and the remaining 50 percent of the risk of event 9

%IE-SET37 ($155K), for a total benefit of $330K in the baseline analysis, and $940K 10 after accounting for uncertainties. PSEG estimated the implementation cost for this 11 SAMA to be $900K, which assumes the panels can be modified and not have to be 12 replaced. Since the estimated benefit is greater than the implementation cost, PSEG 13 determined that this proposed SAMA was potentially cost-beneficial. PSEG stated that 14 this proposed SAMA will be considered for implementation through the established 15 HCGS Plant Health Committee process.

16 The NRC staff notes that SAMA 37 was determined to be cost-beneficial and will be 17 considered by PSEG for implementation through the established HCGS Plant Health 18 Committee process. PSEG concluded, however, that the above originally proposed 19 SAMA to reinforce the 125V DC panels was, by itself, not cost-beneficial, yet it was 20 assumed to be implemented in the evaluation of this new proposed combined SAMA.

21 Because the risk reduction from this new proposed SAMA to reinforce the 120V AC 22 panels to a HCLPF value of 1.0g cannot be obtained without implementation of the 23 proposed SAMA to reinforce the 125V DC panels, the NRC staff concludes that both 24 SAMAs (SAMA 37 and the combined SAMA of reinforcing both the 120 VAC and 125 25 VDC panels) should be considered for implementation.

26 As indicated in Section G.3.2, two plant improvements were identified in the ER but not included 27 in the SAMA evaluation because they were higher cost than the SAMA selected for evaluation.

28 The NRC staff noted however that the two improvements could have larger benefits than the 29 SAMAs evaluated because they could be more effective or could mitigate additional events 30 (PSEG 2010a). In response to the RAIs, PSEG evaluated the two improvements, as 31 summarized below:

32 Replace the normally open floor and equipment drain MOVs with fail-closed AOVs.

33 PSEG estimated the implementation cost of this SAMA to be $2.05M, which is half the 34 estimate for a similar SAMA to replace cooling water system MOVs, which are larger 35 than drain MOVs, with fail-closed AOVs evaluated in the TMI-1 nuclear power plant 36 license renewal application (PSEG 2010a). To estimate the risk reduction, PSEG 37 assumed that the entire release frequency associated with basic event CIS-DRAN-L2-38 OPEN, "valves open automatically for drainage normally open," after adjustment to 39 account for existing procedures that are not credited, was eliminated. The benefit, 40 assuming an external multiplier of 6.8, was estimated to be $710K in the baseline September 2010 G-41 Draft NUREG-1437, Supplement 45

Appendix G 1

analysis, and $2.OM after accounting for uncertainties. Since the implementation cost 2

was greater than the estimated benefit accounting for uncertainties, PSEG concluded 3

the proposed improvement was not cost-beneficial.

4 Auto align 480V AC portable station generator. For HCGS, this improvement is 5

described as requiring permanent installation of an existing portable generator and 6

adding the logic to perform the auto start and load function. PSEG estimated the 7

implementation cost of this SAMA to be at least $1.OM based on an estimate of $1.OM 8

from the Shearon Harris nuclear power plant license renewal application to permanently 9

install a 480V AC generator and pump and an estimate of $3.1M from the TMI-I nuclear 10 power plant license renewal application to automate the start and load of an existing, 11 permanently installed 4KV AC generator (PSEG 2010a, PSEG 2010b). To estimate the 12 risk reduction, PSEG set the failure probabilities of existing operator actions to align the 13 portable generator, and associated joint human error probabilities, to zero. The benefit, 14 assuming an external multiplier of 6.8, was estimated to be $210K in the baseline 15 analysis, and $600K after accounting for uncertainties. Since the implementation cost 16 was greater than the estimated benefit accounting for uncertainties, PSEG concluded 17 the proposed improvement was not cost-beneficial.

18 As indicated in Section G.3.2, for certain SAMAs considered in the ER, there may be 19 alternatives that could achieve much of the risk reduction at a lower cost. The NRC staff asked 20 the applicant to evaluate additional lower cost alternatives to the SAMAs considered in the ER, 21 as summarized below (NRC 201 Oa):

22 Establishing procedures for opening doors and/or using portable fans for sequences 23 involving room cooling failures. In response to the NRC staff RAI, PSEG stated that 24 HCGS already has procedures to implement the suggested alternative on loss of normal 25 Switchgear Room HVAC and that this event is credited in the PRA model (PSEG 26 2010a). However, PSEG did provide an evaluation to implement the suggested 27 alternative in the Service Water Pump Room, which is considered a more practical and 28 cost effective change than SAMA 17, "Replace a Supply Fan with a Different Design in 29 Service Water Pump Room," and SAMA 18, "Replace a Return Fan with a Different 30 Design in Service Water Pump Room," which involve permanent hardware 31 modifications. The cost of implementing an alternate room cooling strategy for this 32 room, identified as SAMA RAI 7.a-1, was estimated to be $150K. The baseline benefit 33 was assumed to be the sum of the estimated benefits for SAMAs 17 and 18, or $1.9M.

34 Accounting for the revised multiplier of 6.8 and uncertainties increases the benefit to 35

$5.9M. Since the estimated benefit is greater than the implementation cost, PSEG 36 determined that SAMA RAI 7.a-1 was potentially cost-beneficial. PSEG also stated that 37 this SAMA will be further evaluated in parallel with cost-beneficial SAMAs 17 and 18 38 since there may be some benefit associated with the permanent hardware modifications 39 considered in these SAMAs.

September 2010 G-42 Draft NUREG-1437, Supplement 45

Appendix G 1

Extending the procedure for using the B.5.b low pressure pump for non-security events 2

to include all applicable scenarios, not just SBOs. In response to the NRC staff RAI, 3

PSEG stated that the estimated benefit for SAMA 10, "Provide Procedural Guidance to 4

use B.5.b Low Pressure Pump for Non-Security Events," already includes the risk 5

reduction for all applicable scenarios (PSEG 2010a). The NRC staff concludes that the 6

suggested alternative has already been addressed.

7 Utilizing a portable independently powered pump to inject into containment. In response 8

to the NRC staff RAI, PSEG explained that the HCGS PRA model already credits use of 9

the diesel fire pump to inject into the RPV and containment and that the addition of 10 another independently powered pump to provide injection would have limited benefit 11 (PSEG 2010a). PSEG further noted that SAMA 10 already evaluated aligning the B.5.b 12 low pressure pump with RHRSW to provide al alternate source of injection. The NRC 13 staff concludes that the suggested alternative has already been addressed.

14 As indicated in Section G.4, the NRC staff questioned PSEG on the risk reduction potential for 15 certain SAMAs (NRC 2010a, NRC 201 Ob), as summarized below.

16 For SAMA 5, "Restore AC Power with Onsite Gas Turbine Generator," PSEG provided a 17 revised estimate of the benefit that included credit for the additional capability for 18 mitigating a more complete set of loss of offsite power initiators that is consistent with 19 the hardware changes proposed (PSEG 201 Oa, PSEG 2010b). This SAMA was 20 determined to be potentially cost-beneficial in PSEG's revised analysis. PSEG stated 21 that SAMA 5 will be considered for implementation through the established HCGS Plant 22 Health Committee process.

23 For SAMA 35, "Relocate, Minimize and/or Eliminate Electrical Heaters in Electrical 24 Access Room", PSEG provided a revised estimate of the benefit assuming 99 percent of 25 the fire risk affected by the SAMA was eliminated (PSEG 201 0a). This SAMA was 26 determined to remain cost-beneficial in PSEG's revised analysis.

27 The NRC staff notes that the 13 cost-beneficial SAMAs (SAMAs 1, 3, 4, 8, 10, 17, 18, 30, 32, 28 35, 36, 37, and 39) identified in PSEG's original baseline and uncertainty analysis, and the three 29 SAMAs and plant improvements determined to be cost-beneficial in response to NRC staff RAIs 30

("establishing procedures for opening doors and/or using portable fans for sequences involving 31 Service Water Pump Room cooling failures," SAMA 5, and "reinforce 1 E 120V AC distribution 32 panels to 1.0g Seismic Rating"), are included within the set of SAMAs that PSEG plans to 33 further consider for implementation through the established Salem Plant Health Committee 34 (PHC) process. The NRC staff suggests that the proposed SAMA to "reinforce the 120V DC 35 panels" also be considered for implementation since it must be implemented to obtain the risk 36 reduction benefits of the SAMA to "reinforce 1 E 120V AC distribution panels to 1.0g Seismic 37 Rating."

September 2010 G-43 Draft NUREG-1437, Supplement 45

Appendix G 1

In response to an NRC staff RAI, PSEG described the PHC as being chaired by the Plant 2

Manager and includes as members the Plant Engineering Manager and the Directors of 3

Operations, Engineering, Maintenance, and Work Management (PSEG 2010a). The PHC is 4

chartered with reviewing issues that require special plant management attention to ensure 5

effective resolution and, with respect to each of the potentially cost-beneficial SAMAs, will 6

decide on one of the following courses of actions: 1) approve for implementation, 2) 7 conditionally approved for implementation pending the results of requested evaluations, 3) not 8

approved for implementation, or 4) table until additional information needed to make a final 9

decision is provided to the PHC. Additional information requested may include 1) making 10 corrections to the original SAMA analysis, 2) examining an alternate solution, 3) performing 11 sensitivity studies to determine the effect of implementing a sub-set of SAMAs, already 12 approved SAMAs, or already approved non-SAMA design changes on the SAMA, or 4) 13 coordinating the SAMA with related Mitigating System Performance Index (MSPI) margin 14 recovery activities.

If approved or conditionally approved for implementation, the SAMA will be 15 ranked with respect to priority and assigned target years for implementation.

16 The NRC staff concludes that, with the exception of the potentially cost-beneficial SAMAs 17 discussed above, the costs of the other SAMAs evaluated would be higher than the associated 18 benefits.

19 G.7 Conclusions 20 21 PSEG compiled a list of 23 SAMAs based on a review of: the most significant basic events from 22 the plant-specific PRA and insights from the HCGS PRA group, insights from the plant-specific 23 IPE and IPEEE, Phase II SAMAs from license renewal applications for other plants, and the 24 generic SAMA candidates from NEI 05-01. A qualitative screening removed SAMA candidates 25 that: (1) are not applicable to HCGS due to design differences, (2) have already been 26 implemented at HCGS, (3) would achieve results that have already been achieved at HCGS by 27 other means, and (4) have estimated implementation costs that would exceed the dollar value 28 associated with completely eliminating all severe accident risk at HCGS. Based on this 29 screening, 2 SAMAs were eliminated leaving 21 candidate SAMAs for evaluation. Nine 30 additional SAMA candidates or plant improvements were identified and evaluated in response to 31 NRC staff RAIs.

32 For the remaining 21 SAMA candidates, a more detailed design and cost estimate were 33 developed as shown in Table G-6. The cost-benefit analyses in the ER and RAI response 34 showed that 9 of the SAMA candidates were potentially cost-beneficial in the baseline analysis 35 (Phase II SAMAs 1, 3, 4, 10, 17, 18, 30, 35, and 39). PSEG performed additional analyses to 36 evaluate the impact of parameter choices and uncertainties on the results of the SAMA 37 assessment. Four additional SAMA candidates (SAMAs 8, 32, 36, and 37) were identified as 38 potentially cost-beneficial in the ER. In response to an NRC staff RAI regarding the 39 assumptions used to estimate the risk reduction potential of certain SAMAs, PSEG identified 40 one additional potentially cost-beneficial SAMA (SAMA 5). In response to NRC staff RAIs 41 regarding the seismic CDF and potential lower cost alternatives, PSEG further identified 42 "establishing procedures for opening doors and/or using portable fans for sequences involving September 2010 G -44 Draft NUREG-1437, Supplement 45

Appendix G 1

Service Water Pump Room cooling failures" and "reinforce 1 E 120V AC distribution panels to 2

1.0g Seismic Rating" as being potentially cost-beneficial enhancements. PSEG has indicated 3

that all 14 potentially cost-beneficial SAMAs, as well as the enhancements "establishing 4

procedures for opening doors and/or using portable fans for sequences involving Service Water 5

Pump Room cooling failures" and "reinforce 1 E 120V AC distribution panels to 1.0g Seismic 6

Rating," will be considered for implementation through the established HCGS Plant Health 7

Committee process. In addition, it is suggested that the plant improvement to "reinforce the 8

120V DC panels" be included in the set of SAMAs to be considered for implementation since it 9

must be implemented to obtain the risk reduction benefits of the plant improvement to "reinforce 10 1 E 120V AC distribution panels to 1.0g Seismic Rating."

11 The NRC staff reviewed the PSEG analysis and concludes that the methods used and the.

12 implementation of those methods was sound. The treatment of SAMA benefits and costs 13 support the general conclusion that the SAMA evaluations performed by PSEG are reasonable 14 and sufficient for the license renewal submittal. Although the treatment of SAMAs for external 15 events was somewhat limited, the likelihood of there being cost-beneficial enhancements in this 16 area was minimized by improvements that have been realized as a result of the IPEEE process, 17 and inclusion of a multiplier to account for external events.

18 The NRC staff concurs with PSEG's identification of areas in which risk can be further reduced 19 in a cost-beneficial manner through the implementation of the identified, potentially cost-20 beneficial SAMAs. Given the potential for cost-beneficial risk reduction, the NRC staff agrees 21 that further evaluation of these SAMAs by PSEG is warranted. However, these SAMAs do not 22 relate to adequately managing the effects of aging during the period of extended operation.

23 Therefore, they need not be implemented as part of license renewal pursuant to Title 10 of the 24 Code of Federal Regulations, Part 54.

25 G.8 References 26 27 American Society of Mechanical Engineers (ASME). 2005. "Addenda to ASME RA-S-2002, 28 Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications." ASME RA-29 Sb-2005, December 2005.

30 BEA (Bureau of Economic Analysis). 2008. Regional Economic Accounts, accessed June 20 at 31 http://www.bea.qov/recqionallreis/.

32 Electric Power Research Institute (EPRI). 1989. "Probabilistic Seismic Hazard Evaluations at 33 Nuclear Plant Sites in the Central and Eastern United States; Resolution of the Charleston 34 Earthquake Issues." EPRI NP-6395-D, EPRI Project P101-53. Palo Alto, CA. April 1989.

35 Electric Power Research Institute (EPRI). 1991. "A Methodology for Assessment of Nuclear 36 Power Plant Seismic Margin," EPRI NP-6041. Palo Alto, CA. August 1991.

37 Electric Power Research Institute (EPRI). 1993. "Fire Induced Vulnerability Evaluation (FIVE) 38 Methodology." TR-100370, Revision 1, Palo Alto, CA. September 19, 1993.

39 KLD Associates, Inc. (KLD). 2004. Salem / Hope Creek Nuclear Generating Stations 40 Development of Evacuation Time Estimates. KLD TR-356. February 2004.

41 Nuclear Energy Institute (NEI). 2005. "Severe Accident Mitigation Alternative (SAMA) Analysis 42 Guidance Document." NEI 05-01 (Rev. A), Washington, D.C. November 2005.

September 2010 G-45 Draft NUREG-1437, Supplement 45

Appendix G 1

Nuclear Energy Institute (NEI). 2007. "Process for Performing Follow-on PRA Peer Reviews 2

using the ASME PRA Standard (Internal Events)." NEI 05-04, Rev. 1, Washington, D.C.

3 December 2007.

4 Public Service Electric and Gas Company (PSEG). 1994. "Hope Creek Generating Station.

5 Individual Plant Examination." April 1994. Accessible at ML080160331.

6 Public Service Electric and Gas Company. (PSEG). 1997. "Hope Creek Generating Station 7

Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilities."

8 July 1997. Accessible at ML080160320.

9 Public Service Electric and Gas Company (PSEG). 2005. Letter from George P. Barnes, 10 PSEG, to NRC Document Control Desk.

Subject:

"Request for License Amendment Extended 11 Power Uprate Hope Creek Generating Station Facility Operating License NPF-57 Docket No:

12 50-354," Hancocks Bridge, New Jersey. Nov. 7, 2005. Accessible at ML053200202.

13 PSEG Nuclear, LLC (PSEG). 2009. Hope Creek Generating Station - License Renewal 14 Application, Applicant's Environmental Report, Operating License Renewal Stage, August 2009.

15 Accessible at ML092430484.

16 PSEG Nuclear, LLC (PSEG). 2010a. Letter from Paul J. Davison, PSEG, to NRC Document 17 Control Desk.

Subject:

"Response to NRC Request for Additional Information dated April 20, 18 2010, related to the Severe Accident Mitigation Alternatives (SAMA) review associated with the 19 Hope Creek Generating Station License Renewal Application," Hancocks Bridge, New Jersey.

20 June 1, 2010. Accessible at ML101550149.

21 PSEG Nuclear, LLC (PSEG). 201 Ob. Letter from Christine T. Neely, PSEG, to NRC Document 22 Control Desk.

Subject:

"Supplement to RAI responses submitted in PSEG Letter LR-N10-0181 23 dated June 1, 2010, related to the Severe Accident Mitigation Alternatives (SAMA) review of the 24 Hope Creek Generating Station," Hancocks Bridge, New Jersey. August 18, 2010. Accessible 25 at ML102320212.

26 U.S. Department of Agriculture (USDA). 2002. "2002 Census of Agriculture - Volume 1, 27 Geographic Area Series, Part 8 (Delaware), Part 20 (Maryland), Part 30.

28 U.S. Nuclear Regulatory Commission (NRC). 1988. Generic Letter 88-20, "Individual Plant 29 Examination for Severe Accident Vulnerabilities." November 23, 1988.

30 U.S. Nuclear Regulatory Commission (NRC). 1990. Severe Accident Risks: An Assessment 31 forFive U.S. Nuclear Power Plants. NUREG-1150. Washington, D.C.

32 U.S. Nuclear Regulatory Commission (NRC). 1991b. "Procedural and Submittal Guidance for 33 the Individual Plant Examination of External Events (IPEEE) for Severe Accident 34 Vulnerabilities." NUREG-1407. Washington, D.C. May 1991.

35 U.S. Nuclear Regulatory Commission (NRC). 1991a. "Individual Plant Examination of External 36 Events (IPEEE) for Severe Accident Vulnerabilities." Generic Letter No. 88-20, Supplement 4.

37 Washington, D.C. June 28, 1991.

38 U.S. Nuclear Regulatory Commission (NRC). 1994. Revised Livermore Seismic Hazard 39 Estimates for Sixty-Nine Nuclear Plant Sites East of the Rocky Mountains. NUREG-1488, April 40 1994. Washington, D.C.

September 2010 G-46 Draft NUREG-1437, Supplement 45

Appendix G 1

U.S. Nuclear Regulatory Commission (NRC). 1996. Letter from David H. Jaffe, U.S. NRC, to 2

Leon R. Eliason, PSEG.

Subject:

"NRC Staff's Evaluation of the Individual Plant Examination 3

(IPE) Submittal (NLR-N94070) Hope Creek Generating Station (TAC No. M74421)." April 23, 4

1996.

5 U.S. Nuclear Regulatory Commission (NRC). 1997a. Regulatory Analysis Technical Evaluation 6

Handbook. NUREG/BR-0184. Washington, D.C.

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

8 Perspectives on Reactor Safety and Plant Performance. NUREG-1 560. Washington, D.C.

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

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

11 U.S. Nuclear Regulatory Commission (NRC). 1999. Letter from Richard B. Ennis, U.S. NRC, to 12 Harold W. Keiser, PSEG.

Subject:

"Review of Individual Plant Examination of External Events 13 (IPEEE) Submittal for Hope Creek Generating Station (TAC No. M83630)".

April 26, 1999.

14 U.S. Nuclear Regulatory Commission (NRC). 2003. SECPOP2000: Sector Population, Land 15 Fraction, and Economic Estimation Program. NUREG/CR-6525, Rev. 1. Sandia National 16 Laboratories. August 2003 17 U.S. Nuclear Regulatory Commission (NRC). 2004. Regulatory Analysis Guidelines of the U.S.

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

19 U.S. Nuclear Regulatory Commission (NRC). 2007. "An Approach for Determining the 20 Technical Adequacy of Probabilistic Risk Assessment Results for Risk-Informed Activities."

21 Regulatory Guide 1.200, Revision 1. January 2007.

22 U.S. Nuclear Regulatory Commission (NRC). 2009. Guidance on the Treatment of 23 Uncertainties Associated with PRAs in Risk-Informed Decision Making. NUREG-1855, 24 Washington, D.C.

25 U.S. Nuclear Regulatory Commission (NRC). 2010a. Letter from Charles Eccleston, U.S. NRC, 26 to Thomas Joyce, PSEG.

Subject:

Revised Request for Additional Information Regarding 27 Severe Accident Mitigation Alternatives for Hope Creek Generating Station. May 20, 2010.

28 Accessible at ML101310058.

29 U.S. Nuclear Regulatory Commission (NRC). 2010b.

Subject:

Summary of Telephone 30 Conference Held on July 29, 2010 between the U.S. Nuclear Regulatory Commission and 31 PSEG Nuclear LLC, Concerning Follow-up Questions Pertaining to the Salem Nuclear 32 Generating Station, Units 1 and 2, and Hope Creek Generating Stations License Renewal 33 Environmental Review. August 13, 3010. Accessible at ML102220012.

34 September 2010 G-47 Draft NUREG-1437, Supplement 45