RA15-032, Response to NRC Request for Additional Information on Severe Accident Mitigation Alternatives Review, Dated April 30, 2015 Regarding the License Renewal Application Environmental Review

Response to NRC Request for Additional Information on Severe Accident Mitigation Alternatives Review, Dated April 30, 2015 Regarding the License Renewal Application Environmental Review

ML15149A370
Person / Time
Site:	LaSalle
Issue date:	05/29/2015
From:	Gallagher M Exelon Generation Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
RA15-032
Download: ML15149A370 (83)

Text

Michael P. Gallagher Vice President, License Renewal Exelon Generation Exelon Nuclear 200 Exelon Way Kennett Square. PA 19348 610 765 5958 Office 610 765 5956 Fax www.exeloncorp.com michaelp.gallagher@exeloncorp.com 10 CFR 50 10 CFR 51 10 CFR 54 RA15-032 May 29, 2015 U.S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555-0001.

LaSalle County Station, Units 1 and 2 Facility Operating License Nos. NPF-11 and NPF-18 NRC Docket Nos. STN 50-373 and STN 50-374

Subject:

Response to NRC Request for Additional Information on Severe Accident Mitigation Alternatives Review, Dated April 30, 2015, Regarding the LaSalle County Station, Units 1 and 2 License Renewal Application Environmental Review

References:

1. Letter from Michael P. Gallagher, Exelon Generation Company LLC *

(Exelon) to NRC Document Control Desk, dated December 9, 2015, "Application for Renewed Operating Licenses"

2. Letter from David Drucker, US NRC to Michael P. Gallagher, Exelon, dated April 30, 2015, "Request for Additional Information on Severe Accident Mitigation Alternatives Regarding the LaSalle County Environmental Review (TAC Nos. MF5567 AND MF5568)

In the Reference 1 letter, Exelon Generation Company, LLC (Exelon) submitted the License Renewal Application (LRA) for the LaSalle County Station, Units 1 and 2. In the Reference 2 letter, the NRC requested additional information to support the Staff's review of the LRA.

The Enclosure to this letter contains the responses to the request for additional information in the Reference 2 letter.

May 29, 2015 Page 2 of 2 There are no new or revised regulatory commitments contained in this letter.

If you have any questions, please contact Ms. Nancy Ranek, Environmental Lead, Exelon License Renewal, at 610-765-5369.

I declare under penalty of perjury that the foregoing is true and correct.

Respectfully,

~e-£~1----. .

Michael P. Gallagher Vice President - License Renewal Projects Exelon Generation Company, LLC

Enclosure:

Exelon Generation Company, LLC Response to NRC Request for Additional Information on Severe Accident Mitigation Alternatives Review, Dated April 30, 2015, Regarding the LaSalle County Station, Units 1 and 2 License Renewal Application Environmental Review cc: Regional Administrator - NRC Region Ill NRC Project Manager (Safety Review), NRR-DLR NRC Project Manager (Environmental Review), NRR-DLR NRC Project Manager, NRR-DORL- LaSalle County Station NRC Senior Resident Inspector, LaSalle County Station Illinois Emergency Management Agency - Division of Nuclear Safety

RA15-032 May 29, 2015 ENCLOSURE Exelon Generation Company, LLC Response to NRC Request for Additional Information on Severe Accident Mitigation Alternatives Review, Dated April 30, 2015, Regarding the LaSalle County Station, Units 1 and 2 License Renewal Application Environmental Review

RA15-032 ENCLOSURE TABLE OF CONTENTS Acronyms ..................................................................................................................................................... iii QUESTION 1.a ............................................................................................................................................... 1 QUESTION 1.b.i ............................................................................................................................................. 3 QUESTION 1.b.ii ............................................................................................................................................ 4 QUESTION 1.b.iii ........................................................................................................................................... 5 QUESTION 1.c................................................................................................................................................ 6 QUESTION 1.d ............................................................................................................................................... 8 QUESTION 1.e ............................................................................................................................................... 9 QUESTION 1.f .............................................................................................................................................. 10 QUESTION 1.g ............................................................................................................................................. 11 QUESTION 1.h ............................................................................................................................................. 12 QUESTION 2.a ............................................................................................................................................. 14 QUESTION 2.b ............................................................................................................................................. 15 QUESTION 2.c.............................................................................................................................................. 16 QUESTION 2.d.i ........................................................................................................................................... 18 QUESTION 2.d.ii .......................................................................................................................................... 19 QUESTION 2.e ............................................................................................................................................. 23 QUESTION 2.f .............................................................................................................................................. 24 QUESTION 2.g ............................................................................................................................................. 25 QUESTION 2.h ............................................................................................................................................. 28 QUESTION 3.a ............................................................................................................................................. 30 QUESTION 4.a ............................................................................................................................................. 32 QUESTION 4.b ............................................................................................................................................. 33 QUESTION 4.c.............................................................................................................................................. 34 QUESTION 4.d ............................................................................................................................................. 35 QUESTION 4.e ............................................................................................................................................. 38 QUESTION 4.f .............................................................................................................................................. 39 QUESTION 4.g ............................................................................................................................................. 40 QUESTION 4.h ............................................................................................................................................. 41 QUESTION 5.a.i ........................................................................................................................................... 42 TOCi

RA15-032 ENCLOSURE QUESTION 5.a.ii .......................................................................................................................................... 43 QUESTION 5.a.iii.......................................................................................................................................... 48 QUESTION 5.a.iv.......................................................................................................................................... 49 QUESTION 5.b.i ........................................................................................................................................... 51 QUESTION 5.b.ii .......................................................................................................................................... 52 QUESTION 5.c.............................................................................................................................................. 53 QUESTION 5.d ............................................................................................................................................. 55 QUESTION 5.e ............................................................................................................................................. 56 QUESTION 5.f.i ............................................................................................................................................ 58 QUESTION 5.f.ii ........................................................................................................................................... 65 QUESTION 5.f.iii .......................................................................................................................................... 66 QUESTION 5.f.iv .......................................................................................................................................... 67 QUESTION 6.a ............................................................................................................................................. 68 QUESTION 6.b ............................................................................................................................................. 69 QUESTION 6.c.............................................................................................................................................. 70 QUESTION 6.d ............................................................................................................................................. 71 QUESTION 6.e ............................................................................................................................................. 73 QUESTION 6.f .............................................................................................................................................. 74 QUESTION 7.a ............................................................................................................................................. 75 TOCii

RA15-032 ENCLOSURE Acronyms AB Auxiliary Building AC Alternating Current ADS Automatic Depressurization System ATWS Anticipated Transient Without Scram BOC Break Outside Containment BWR Boiling Water Reactor BWROG Boiling Water Reactor Owners' Group CCF Common Cause Failure CDF Core Damage Frequency CET Containment Event Tree CPI Consumer Price Index CRD Control Rod Drive CSCS Core Standby Cooling System CST Condensate Storage Tank DAEC Duane Arnold Energy Center DC Direct Current DG Diesel Generator DHR Decay Heat Removal DLOOP Dual Unit Loss of Offsite Power DW Drywell EALs Emergency Action Levels ECCS Emergency Core Cooling System EDG Emergency Diesel Generator EFPD Effective Full Power Days EOPs Emergency Operating Procedures EPGs Emergency Procedure Guidelines EPRI Electric Power Research Institute ER Applicants Environmental Report - Operating License Renewal Stage, LaSalle County Station ESF Engineered Safety Feature FLEX Diverse and Flexible Coping Strategies (as described in NEI 12-06, Rev. 0)

FPIE Full Power Internal Event TOCiii

RA15-032 ENCLOSURE Acronyms FPS Fire Protection System F-V Fussell - Vesely FW Feedwater HEP Human Error Probability HFE Human Failure Event HPCS High Pressure Core Spray HVAC Heating Ventilating Air Conditioning IPE Individual Plant Examination IPEEE Individual Plant Examination - External Events JHEP Joint Human Error Probability LERF Large Early Release Frequency LGA LaSalle County Station General Abnormal Procedure LOCA Loss of Coolant Accident LOOP Loss of Offsite Power LP Low Pressure LPCI Low Pressure Coolant Injection LPCS Low Pressure Core Spray LSCS LaSalle County Station MAAP Modular Accident Analysis Program MACCS2 MELCOR Accident Consequences Code System, Version 2 MACR Maximum Averted Cost-Risk MCR Main Control Room MSCWLL Minimum Steam Coolant Water Level Limit NRC U.S. Nuclear Regulatory Commission NTTF Near Term Task Force on Fukushima PRA Probabilistic Risk Assessment PACR Potential Averted Cost Risk PSA Probabilistic Safety Assessment RCIC Reactor Core Isolation Cooling RHR Residual Heat Removal RHRSW Residual Heat Removal Service Water RMIEP Risk Methods Integration & Evaluation Program RPS Reactor Protection System TOCiv

RA15-032 ENCLOSURE Acronyms RPV Reactor Pressure Vessel SAMA Severe Accident Mitigation Alternative SBO Station Blackout SECPOP Sector Population and Economic Estimator SPC Suppression Pool Cooling SRV Safety Relief Valve SW or WS Service Water TB Turbine Building URE Updating Requirement Evaluation TOCv

RA15-032 ENCLOSURE QUESTION 1.a Section F.2 of the Environmental Report (ER) indicates that the LSCS PRA is a Unit 2 model and is considered applicable to Unit 1 since the two units are nearly identical, unless otherwise noted. Provide a brief description of the differences between the units, particularly those differences that might impact internal flooding and their impacts on the unit risk.

RESPONSE

The major difference between the Units is that the common or 0 diesel generator (0 diesel generator) is located in the Unit 1 Auxiliary Building (AB) and the 0 diesel generator cooling water pump is located in the Unit 1 Division 1 core standby cooling system (CSCS) pump room.

A flood event that initiates in the Unit 1 Division 1 CSCS pump room does not propagate from this room as the surrounding areas are protected by a water tight door and walls. This specific flood scenario is a negligible contributor to risk on both units because it only impacts one division of CSCS equipment. The Unit 1 core damage frequency (CDF) contribution from a flood in this area would be slightly higher than Unit 2 because the internal flood will also impact the Unit 1 RHRSW pumps. The Unit 1 RHRSW pumps do not have a function for Unit 2; therefore, loss of these pumps has no quantitative impacts on the Unit 2 CDF values. Since these flood scenarios have negligible CDF impact and have limited impact to plant equipment and operator response, there is judged to be no impact on the SAMA evaluation. The general plant layout and other internal flooding scenarios are discussed in detail below.

The Unit 1 and Unit 2 turbine building and reactor buildings are open to each other and because of this open space, internal flood impacts are similar for each unit. For example, a turbine building flood on one unit will propagate into the other unit and behave similarly on both units.

The 0 diesel generator is located on the ground floor and there are no credible floods that disable any diesel generators directly. However, the diesel generator cooling water pumps and the RHRSW pumps are located in the CSCS rooms (one room for each division). The CSCS rooms are located below grade and are subject to flooding from flood sources in the turbine building and reactor building. The CSCS rooms are provided with flood protection, but they are not protected from all floods.

On Unit 1, the Division II CSCS pump room is accessed from the Auxiliary Building (AB) stairwell. This room has a flood door; however, the door is only designed for floods that are initiated in that room to prevent them from propagating to the turbine building. This flood door does not protect the room from floods initiated outside of this room and it is assumed that the door will eventually fail with the pressure of flood waters from the AB side of the door. Access to the Division I and Division III CSCS pumps rooms is via the Division II CSCS pump room.

The flood door between Division I and Division II CSCS pump rooms is water tight in both directions. Therefore, a flood in Division II or the Turbine Building will not impact Division I.

The flood door between the Division II and Division III CSCS rooms is also water tight in both directions, so that a flood from Division II will not impact Division III. However, a flood from the Turbine Building will impact Division III because there is a water tight door in the Division III switchgear room (also connected by the AB stairwell) that is not watertight in the direction of the flood (i.e., from the AB stairwell). This will allow a flood in the Turbine Building (TB) to also propagate to the Division III CSCS pump room. This configuration is similar on both units; however, there is no 0 diesel generator cooling water pump on Unit 2. Nevertheless, due to the 1 of 75

RA15-032 ENCLOSURE open nature of the TB floor plan, a TB flood impacts both units, so there is no asymmetry in the response.

The other asymmetries between units include the power supplies for the shared systems. For example, the common service water pump is powered from Unit 2 and the common instrument air compressor is powered from Unit 1. These electrical alignment differences result in minor differences between the units in CDF when considering a loss of offsite power, These issues have negligible impact on CDF and they are evaluated for plant specific applications involving these systems (e.g., maintenance configurations to assess online maintenance, notices of enforcement discretion). The power supply differences for common equipment results in a slightly smaller CDF for Unit 1.

None of the minor differences are judged to have an impact on the SAMA evaluation.

2 of 75

RA15-032 ENCLOSURE QUESTION 1.b.i Relative to the contributions to the internal events Core Damage Frequency (CDF):

i. Provide the contributions to internal events CDF from Station Blackout (SBO).

RESPONSE

The contribution to internal events CDF from SBO is 6.37E-07/yr. This value is the sum of CDF for accident classes IBE and IBL (i.e., early station blackout and late station blackout respectively) listed in Table F.2-3 of Appendix F of the LaSalle County Station Environmental Report.

3 of 75

RA15-032 ENCLOSURE QUESTION 1.b.ii Relative to the contributions to the internal events Core Damage Frequency (CDF):

ii. Table F.2-3 indicates the CDF contribution due to Anticipated Transient Without Scram (ATWS) is 4.87 x 10-7/year; discuss the reasons why the CDF due to ATWS is such a relatively high contribution to the total internal events CDF.

RESPONSE

Joint human error probabilities (JHEPs) for operator actions that must be performed in short time windows are prevalent in the ATWS cutsets and are the dominant contributors to CDF from this accident class. Some short time windows for response are due to the plant design and the lack of a redundant, high pressure, high volume injection system for use in these scenarios.

The only high pressure, high volume injection system available in an ATWS is feedwater. The emergency operating procedures require the High Pressure Core Spray System to be secured and the Reactor Core Isolation Cooling System (RCIC) is a low volume system. The capability of feedwater is limited by the rate at which water can be made up to the condenser.

The ATWS contribution for LSCS is about 19% of the internal events CDF, which is the same as that reported for the Columbia Generating Station (another BWR 5) based on PRA model revision 7.1 (NRC 2012). However, the Columbia Generating Station internal events CDF is larger than the LSCS internal events CDF, so the corresponding ATWS CDF is also larger (1.4E-06/yr). The DAEC Environmental Report (FPL 2008) identifies that the ATWS CDF is 3.15E-06/yr, which corresponds to 28.9% of the internal events CDF, while the Pilgrim Environmental Report (Entergy 2006) indicates that the 5.3E-08/yr ATWS CDF represents only 0.83% of the internal events CDF. In general, ATWS CDF and the percent contribution of ATWS to the CDF vary significantly in the industry, and values can be found both above and below those documented for LSCS. The specific reasons for the differences cannot be characterized without a detailed analysis of multiple PRAs.

REFERENCES:

Entergy 2006 Entergy (Entergy Nuclear Generation Company and Entergy Nuclear Operations, Inc.). Pilgrim Nuclear Power Station. 2006. License Renewal Application. Environmental Report. Attachment E: Severe Accident Mitigation Alternatives Analysis. January.

FPL 2008 FPL (FPL Energy Duane Arnold, LLC). Duane Arnold Energy Center. 2008. License Renewal Application. Environmental Report.

Appendix F: SAMA Analysis. September.

NRC 2012 NRC (U.S. Nuclear Regulatory Commission). 2012. Generic Environmental Impact Statement for License Renewal of Nuclear Plants, Supplement 47, Regarding Columbia Generating Station.

NUREG-1437. Final Report. Office of Nuclear Reactor Regulation.

April.

4 of 75

RA15-032 ENCLOSURE QUESTION 1.b.iii Relative to the contributions to the internal events Core Damage Frequency (CDF):

iii. Clarify what is meant by the statement all ATWS events are modeled as a turbine trip in Section F.2.3.1 of the ER.

RESPONSE

On page F-12 of the ER, the following statement is made; The turbine trip initiating event is important to note because it also represents the ATWS frequency (i.e., all ATWS events are modeled as a turbine trip). The statement in parenthesis, which is meant to clarify the first statement, is misleading. A better way to describe the relationship of the turbine trip initiator to the ATWS is that ATWS sequences are a large contributor to the percentage of CDF attributed to the turbine trip initiating event. It should be noted that other initiating events can also result in ATWS sequences.

5 of 75

RA15-032 ENCLOSURE QUESTION 1.c Identify the plant systems shared or that could be cross-tied between units and describe the modeling in the PRA, including the treatment of unavailability during outages or events involving the other unit.

RESPONSE

The most significant piece of equipment that is shared between units is the common or 0 diesel generator (0 diesel generator). The 0 diesel generator is physically located on Unit 1. It is shared between the Unit 1 and Unit 2 Division 1 AC power divisions. The modeling of this component assumes that it loads automatically to the modeled unit 50% of the time when demanded. The model includes an operator action to manually align the 0 diesel generator to the modeled unit if the 0 diesel generator does not automatically align to the unit. The model also includes an operator action related to the failure to control the loading of the diesel generator when it is aligned to both units.

The Division 1 and Division 2 AC safety related 4 kV buses can be cross-tied between units, which provides a method by which to power these buses via the opposite unit offsite power sources. The PRA model includes the opposite unit offsite power sources, divisional cross-tie breakers and the operator action necessary to complete this alignment. Note that Division 3 does not have the capability to be cross tied between units.

Because the division 2 safety related 4 kV buses can be cross-tied between units, the Division 2 diesel generators can also be shared by the other unit. For example, if there is a failure of the Unit 2 Division 2 diesel generator (2A diesel generator), power can be supplied to the Unit 2 Division 2 bus from the Unit 1 Division 2 diesel generator (1A diesel generator). In this sense, the Division 2 diesel generators can also be shared between units. This is modeled with an operator action to manually align the opposite unit diesel generator to the unit of interest.

The Division 1 and Division 2 DC safety related buses also have the capability to be cross-tied between units. However, this is not included in the PRA model because the reliability of the DC system is such that it is not a risk significant configuration. Additionally, it is a somewhat complicated action, and based on operator interviews, this configuration is not preferred as the DC system is not designed to be operated in this manner.

The instrument air system is a common system that is shared between units and modeled as such. The service water system is also shared and it normally operates as one system that provides service water to both units.

The fire protection is a shared system that can be used on either unit.

For all shared systems, system unavailability during outages is considered in the development of system maintenance unavailability basic events. Other than the 0 diesel generator, cross-tie breakers and AC power supplies, the model does not explicitly consider the impacts at one unit on the other unit. Service water and instrument air have the capability to support both units simultaneously during events on both units. With respect to the fire protection system, it could be aligned to both units and could be shared for injection. However, the PRA model does not account for this condition probabilistically.

6 of 75

RA15-032 ENCLOSURE With respect to the 0 diesel generator and electric plant, the PRA model considers the impacts of a single unit loss of offsite power (LOOP) and a dual unit loss of offsite power. The systems discussed in this response have the ability to support both units while both units are in an accident or transient condition.

7 of 75

RA15-032 ENCLOSURE QUESTION 1.d Section F.2.1 of the ER indicates that no significant plant modifications affecting the risk profile were performed since the PRA model freeze date. Other than the hard pipe vent, identify any planned changes or modifications to the LSCS design, operation, reactor power level, fuel cycle, or fuel design that might impact the SAMA analysis and describe their expected impacts.

RESPONSE

The PRA open items tracking database (commonly referred to as the URE database) was reviewed and no outstanding modifications were identified with the exception of the necessary plant modifications to meet the NRC order related to the Near Term Task Force (NTTF) on Fukushima. In particular, these changes will be made related to mitigating strategies (EA 049) and installation of the hardened containment vent (EA-12-050).

The LSCS conceptual modifications are described in NRC submittals dated February 28, 2013 (RS-13-021) as well as 6 month updates dated August 28, 2013 (RS-13-121) and February 28, 2014 (RS-14-011). These submittals to the NRC address Beyond Design Basis External Events (BDBEE) Order Number EA-12-049 dated March 12, 2012.

The LSCS response to the NRC regarding plans for installation of a reliable hardened containment vent was submitted on February 28, 2013 (RS-13-112). This submittal contains the conceptual modifications for the reliable hardened containment vent.

The conceptual design modifications continue to evolve from those described in the regulatory submittals. A sensitivity study was performed to determine the impact of these modifications including the effect of the hardened containment vent on CDF. This PRA sensitivity study was performed using the best available information regarding the plant design as of May 1, 2014.

The study showed that CDF decreased approximately 5% due to FLEX modifications, excluding the hardened containment vent system. The FLEX equipment has the largest impact on Class IBE, Class IBL and Class II accident sequences.

Other than FLEX modifications, no other future plant design changes, procedure changes or fuel cycle changes have been identified at this time that might impact the SAMA analysis.

8 of 75

RA15-032 ENCLOSURE QUESTION 1.e Table F.2-1 of the ER includes a number of footnote citations, but the content of the footnotes are not included. Provide the footnotes.

RESPONSE

Notes to Table F.2-1:

(1) Mean CDF value (does not include internal flooding) based on the internal events CDF from integrated quantification as reported in NUREG/CR-4832, Volume 2.

(2) The Risk Methods Integration & Evaluation Program (RMIEP) study (i.e.,

NUREG/CR-4832) did not calculate a large early release frequency (LERF) metric comparable to the current LERF definition.

(3) CDF value based on information in NUREG-1560, Volume 3, Table B-1. Note that this NUREG Volume did not contain specific details on the CDF values reported.

(4) A LERF value for the 2000C model is not available.

9 of 75

RA15-032 ENCLOSURE QUESTION 1.f Section F.2.2.4 of the ER states that the 2000A model included a seismic PRA model (removed when the 2006A model was developed). Confirm that the seismic risk is not included in the results given in Table F.2-1 of the ER. If included, provide the results without the seismic risk.

RESPONSE

The CDF values provided in Table F.2-1 did include the CDF due to the seismic initiators because these initiating events were included in the base PRA model. To provide the results without the seismic initiating event, each seismic event was set to false in the model cutsets and the results are reported below. Seismic initiators had only a minor impact on the base CDF.

Model Revision CDF Reported in Table F.2-1 CDF without seismic initiator (i.e.,

(includes seismic initiating Events) set to false) 2000A 5.90E-06 5.80E-06 2000B 5.90E-06 5.80E-06 2000C 8.20E-06 8.10E-06 2001A 5.70E-06 5.60E-06 2003A 6.64E-06 6.52E-06 10 of 75

RA15-032 ENCLOSURE QUESTION 1.g 1.g Provide a summary of the more significant changes incorporated in the 2006A PRA model.

RESPONSE

The 2006A PRA model (1) included the most recent plant and industry data, (2) addressed PRA open issues identified through the use of the model for applications, and (3) included an updated internal flooding model.

Significant changes that were incorporated in the 2006A PRA are listed below. Identification of the more significant changes was based on judgment because the changes to risk were not quantified on a per-change level and no quantitative information is available for any individual model update task.

The emergency operating procedures for LSCS (referred to as LGAs) do not direct automatic depressurization system (ADS) inhibit unless a failure to scram occurs (or power is unknown). The PRA model is modified to reflect the LGAs which differ from the generic BWROG EPGs.

An update of the TB flooding accident sequences was performed.

Allocation of loss of coolant accident (LOCA) frequencies on a location and size specific basis. [The LOCA locations have been subdivided for more accurate assessments of their consequences.]

Data were updated to use the most current industry data and plant specific data.

11 of 75

RA15-032 ENCLOSURE QUESTION 1.h Changes incorporated into the 2011A model included assignment of mitigated ATWS scenarios from Accident Class IV to Accident Class II in Section F.2.2.11 of the ER. However, Table F.2-3 of the ER does not appear to include any Class II ATWS scenarios. Provide more detail on the assignment of ATWS scenarios, including any impacts from the 2013A Level 2 model upgrade.

RESPONSE

The ATWS Level 1 event trees are generally structured with the following nodes:

RPS- MECH: NO MECHANICAL CCF OF CONTROL ROD INSERTION RPS-ELEC: NO ELECTRICAL CCF OF CONTROL ROD INSERTION RPT-SRV: OVERPRESSURE PROTECTION ADEQUATE ARI: ALTERNATE ROD INSERTION AVAILABLE MC: MAIN CONDENSER AVAIL (ATWS)

MCFW: FEEDWATER AVAILABLE SL1: EARLY SLC AND LEVEL CONTROL AVAILABLE SL2: LATE SLC AND LEVEL CONTROL AVAILABLE LP-OIADS: LP SYS AVAIL; INHIBIT ADS; AND LVL CONTROL HP-LP: HI OR LO PRES SYS AVAIL W: CONTAINMENT HEAT REMOVAL ADEQUATE The ATWS event trees have the following endstates given the success or failure of the various nodes.

OK Class IC Class II Class IV The endstates assigned to Accident Class IC are characterized by success of either SL1 or SL2, success of LP-OIADS and then subsequent failure of HP-LP. In these sequences, the reactor is successfully shut down by standby liquid control (SL1 or SL2) prior to core damage, ADS is inhibited and level is properly controlled. However, high pressure or low pressure injection fails.

This sequence is best characterized as a loss of coolant injection at low pressure (Class IC).

Endstates assigned to Accident Class II are characterized by success of either SL1 or SL2, success of LP-OIADS and HP-LP, but failure at the W node. In these sequences, the reactor is successfully shut down by standby liquid control (SL1 or SL2) prior to core damage, ADS is inhibited and level is properly controlled, and depressurization and low pressure injection are properly controlled. However, containment heat removal fails. This sequence is best characterized as a loss of decay heat removal (Class II).

All other accident sequences are either classified as OK or Class IV.

These changes were made as part of the Level 1 2011A PRA model. The Level 1 PRA model was not changed as part of the 2013A Level 2 model. The more accurate assessment of the accident classes in the Level 1 model provides for a proper transition from the Level 1 model to the Level 2 containment event trees.

12 of 75

RA15-032 ENCLOSURE With respect to Table F.2-3, the Class II CDF is reported as one value (1.03E-06) and not further broken down by the event tree or accident sequences. This value includes all Class II accident sequences including the ATWS sequences. The following table lists the ATWS sequences that contribute to the Class II accident class. This table was generated using the Level 1 (CDF) cutset importance report.

Class II ATWS Sequences F-V CDF Sequence CDF RCVSEQ-ATW1-007 3.66E-04 2.58E-06 9.43E-10 RCVSEQ-ATW1-011 1.55E-04 2.58E-06 4.00E-10 RCVSEQ-ATW1-016 1.14E-06 2.58E-06 2.94E-12 RCVSEQ-ATW1-030 5.67E-03 2.58E-06 1.46E-08 RCVSEQ-ATW1-034 1.81E-02 2.58E-06 4.67E-08 RCVSEQ-ATW1-039 1.53E-04 2.58E-06 3.94E-10 RCVSEQ-ATW1-043 3.85E-04 2.58E-06 9.92E-10 RCVSEQ-ATW2-003 2.85E-05 2.58E-06 7.35E-11 RCVSEQ-ATW2-007 8.25E-05 2.58E-06 2.13E-10 RCVSEQ-ATW2-013 1.34E-04 2.58E-06 3.45E-10 RCVSEQ-ATW2-017 4.52E-04 2.58E-06 1.17E-09 RCVSEQ-ATW4-004 5.76E-06 2.58E-06 1.48E-11 RCVSEQ-ATW4-008 5.70E-06 2.58E-06 1.47E-11 RCVSEQ-ATW4-015 6.54E-05 2.58E-06 1.69E-10 RCVSEQ-ATW4-019 2.19E-04 2.58E-06 5.65E-10 RCVSEQ-ATW6-004 5.31E-06 2.58E-06 1.37E-11 RCVSEQ-ATW6-008 3.29E-06 2.58E-06 8.48E-12 RCVSEQ-ATW6-015 6.03E-05 2.58E-06 1.55E-10 RCVSEQ-ATW6-019 1.54E-04 2.58E-06 3.97E-10 TOTAL CDF Due to Class II ATWS Sequences 6.71E-08 13 of 75

RA15-032 ENCLOSURE QUESTION 2.a Section F.2.2.11 of the ER is titled 2013A Upgrade. Describe the changes in the Level 2 model that are considered upgrades as opposed to updates. Also, describe the steps taken to assure technical adequacy of the 2013A Level 2 model.

RESPONSE

The word upgrade in the sub-heading 2013A Upgrade in Section F.2.2.11 of the ER was intended to have its common meaning rather than the meaning ascribed to it by the ASME/ANS PRA Standard. In retrospect, this was misleading. No change to the large early release frequency (LERF) analysis or methodology that would be considered an upgrade by the ASME/ANS PRA Standard definition was made during the model update that created the 2013A model. Accordingly, the 2013A model did not require a peer review.

Prior to the 2013A model, the Level 2 model was a LERF model only (i.e., containment event trees end states of H/E or OK). The update that created the 2013A model expanded the scope of the model to a full Level 2 assessment and incorporated the assessment of both timing and release categories for all endstates in the containment event trees.

Exelon has implemented procedures and processes to assure the technical adequacy of the full power internal event (FPIE) PRA models. Model development and technical adequacy are outlined in ER-AA-600-1015, FPIE PRA Model Update. Among other requirements, ER-AA-600-1015 requires detailed review of cutsets (both dominant and non-dominant), importance measures, accident sequences, and systems modeling. ER-AA-600-1015 also requires an assessment of whether or not the changes made to the model during an update require a peer review per ASME/ANS RA-S-2009, Standard for Level 1/Large Early Release Frequency Probabilistic Risk Assessment for Nuclear Power Plant Applications. The 2013A model was reviewed and approved in accordance with this procedure.

14 of 75

RA15-032 ENCLOSURE QUESTION 2.b Provide a description of the interface between the Level 1 CDF analysis and the Level 2 analysis, including the use of plant damage states or core damage accident classes and how the two models are linked.

RESPONSE

In the LSCS Level 1 PRA, accident sequences are postulated that lead to core damage and potentially challenge containment. The LSCS Level 1 PRA has identified discrete accident sequences that contribute to the core damage frequency and represent the spectrum of possible challenges to containment. These accident sequences are then placed into easily understood functional sequence groupings (i.e., accident classes). The accident classes are based on the studies presented in WASH-1400, Reactor Safety Study: An Assessment of Accident Risks on U.S. Commercial Nuclear Power Plants.

Functional accident sequences can be defined to group similar systemic accident sequences.

The definition of the functional accident sequences are derived to show the relationship of these LSCS specific sequences to: (a) the BWR critical safety functions; (b) the event tree development; and (c) the need for information transfer to the Level 2 portion of the PRA.

The Level 1 accident class bins are then used as the starting point for the Level 2 PRA containment event tree (CET) analysis. The Level 1 accident sequences and resulting cutsets are incorporated directly into the Level 2 analysis to maintain the propagation of dependencies.

Each node in the CET is evaluated using the nodal functional fault trees which include the following:

Fault tree models from the Level 1 analysis for the system or function Any Level 2 limitations in timing, procedures, access, or dependencies Phenomenological effects Environmental impacts on equipment or operator actions Therefore, when the CET is evaluated any equipment or operator failures that have already failed in the Level 1 sequence are automatically treated in the analysis, i.e., the dependencies are explicitly handled. The CETs and the end state assignments are derived based on the deterministic core melt progression calculations.

The Level 1 and Level 2 models are directly linked using a single top logic structure and quantified using the EPRI software, CAFTA.

15 of 75

RA15-032 ENCLOSURE QUESTION 2.c Provide additional information on the containment event trees (CETs) utilized in the Level 2 analysis including: the number of CETs, the sequences handled by each CET, the events included in each CET, and how the loss of offsite power (LOOP) and SBO sequences are addressed.

RESPONSE

Several types of CETs are necessary to characterize various containment challenges. The LSCS PRA directly links the Level 1 to Level 2 portions of severe accident sequences through linked event trees that are ultimately converted to a large fault tree model for quantification.

These trees convey the support system conditions throughout the Level 1 and level 2 trees and include considerations of preventive or mitigative features as well as timing considerations.

Three different general containment event tree types are used:

CET1: Class I and III CETs: Containment initially intact. These sequences are characterized by an initial loss of coolant makeup to the reactor vessel that leads to core damage. The attempts to arrest the melt progression in-vessel and ex-vessel are assessed along with containment integrity during the challenge. In all cases, the entry point to the containment event tree is at the time that the core is initially damaged.

CET2: Class II, and IV CETs: Containment is initially failed or subject to incipient failure before core damage. For these classes of accidents, the primary containment boundary would generally fail before the molten core penetrates the reactor vessel. In Class II accident sequences, the inability to remove heat from the containment results in a gradual heat-up of the suppression pool. For Class IV accidents, the amount of energy transferred to the suppression pool exceeds its heat removal capacity.

CET3: Class V: CET3 is used to evaluate several distinct core damage scenarios: (1) LOCAs outside containment for which coolant makeup to the reactor vessel has failed lead to a core melt event with a direct release pathway from the vessel to the reactor building; and (2) an interfacing LOCA or drywell bypass.

The LSCS model uses 15 CETs in the Level 2 analysis and these CETs address all of the Level 1 accident sequences that lead to core damage.

The following is a list of the LSCS specific Level 2 CETs and the sequences that are handled by each CET:

Class IA (Class IA accident sequences)

Class IBE (Class IBE accident sequences)

Class IBL (Class IBL accident sequences)

Class IC (Class IC accident sequences)

Class ID (Class ID accident sequences)

Class II (Class II accident sequences that do not result in an early release)

Class IIE (Class II accident sequences that result in an early release) 16 of 75

RA15-032 ENCLOSURE Class IIV (Class IIV accident sequences that do not result in an early release)

Class IIVE (Class IIV accident sequences that result in an early release)

Class IIIA (Class IIIA accident sequences)

Class IIIB (Class IIIB accident sequences)

Class IIIC (Class IIIC accident sequences)

Class IIID (Class IIID accident sequences)

Class IV (Class IV accident sequences)

Class V (Class V accident sequences)

LSCS Containment Event Trees (CETs) are developed to provide the link between: (1) the Level 1 event tree core damage end-states and (2) safe shutdown or radionuclide release end-states that describe release magnitude and timing. The CET is used to map out the possible containment conditions affecting the radionuclide releases associated with a given core damage sequence. These CETs describe the various potential radionuclide release paths to the environment and provide estimates of their relative likelihoods. This process is iterative, and requires feedback and interactions among the analysts involved in the systemic event trees, the CET, and the plant response evaluation. The explicit link using the Level 1 sequence logic allows explicitly accounting for the dependencies between initiating events, system failures, and containment mitigation systems.

The functional event nodes of the LSCS Level 2 CET are the following:

Containment Isolated (IS)

RPV Depressurization (OP)

Core Melt Arrested In-Vessel (RX)

Combustible Gas Venting Initiated (GV)

Containment Remains Intact (CZ)

Injection Established to RPV or Drywell (TD)

Containment Flooding Occurs with Drywell Vent (FC)

Containment Heat Removal (HR)

Containment Vent (CV)

Suppression Pool Bypass (SP)

No Large Containment Failure (NC)

Inventory Make up Available (MU)

Drywell Intact (DI)

Wetwell Airspace Failure (WW)

Reactor Building Effectiveness (RB)

LOOP/SBO accident sequences are modeled in Level 1 in the DLOOP and LOOP event trees.

The LSCS probabilistic analysis considers the availability of safety and non-safety systems for LOOP and DLOOP events. The regulatory station blackout analysis is tailored to address certain pre-specified failures (i.e., high pressure core spray (HPCS) may not be considered failed under SBO Analysis Requirements). The probabilistic analysis considers the potential for the emergency diesel generators (Unit 1, Unit 2 and Unit 0 [common] EDGs) to fail. In addition, the probabilistic analysis considers the potential for common cause failures among all diesel generators (e.g., the EDGs and the HPCS diesel generator). These common cause effects are treated in the fault trees. The Level 2 model explicitly accounts for recovery of AC power in both the RX and TD nodes. Each DLOOP and LOOP core damage accident sequence is grouped in a discrete core damage accident class and incorporated into the corresponding Level 2 CETs.

17 of 75

RA15-032 ENCLOSURE QUESTION 2.d.i Table F.2-6 of the ER includes Accident Classes II, IIE, IIV and IIVE. Classes IIE and IIVE are not defined in Table F.2-3.

i. Define these accident classes.

RESPONSE

Class IIE accidents are the portion of the Class II accidents that result in an early release.

Similarly, Class IIVE accidents are the portion of the Class IIV accidents that result in an early release. For convenience in modeling, the Class II accident classes are separated into two containment events trees; one representing the early releases and the other representing intermediate or late releases.

The analysis of the timing of a general emergency declaration has been evaluated from a probabilistic standpoint for Class II accident scenarios. This evaluation was performed because of the potential differences in interpretation of the Emergency Action Levels (EALs) and the probability that the Emergency Director could delay the declaration of the General Emergency resulting in an early release.

Interviews of key emergency response personnel were performed to determine the best estimate probability that the interpretation of the EALs would be such that the General Emergency declaration would be delayed resulting in an early release as opposed to an intermediate or late release. These interviews consisted of case studies and a discussion of EALs. The mean probability of failure to declare a General Emergency with adequate time to take protection measures for the general public prior to containment failure is 5% or 0.05 based on the expert opinion information. As a result, the LSCS Level 2 model is structured such that 5% of the Class II accidents can result in an Early release, but 95% of the Class II accident releases are non-early.

18 of 75

RA15-032 ENCLOSURE QUESTION 2.d.ii Table F.2-6 of the ER includes Accident Classes II, IIE, IIV and IIVE. Classes IIE and IIVE are not defined in Table F.2-3.

ii. Table F.2-6 also includes a significant intact frequency for the Class II accident classes which is normally considered and defined in Table F.2-3 to include sequences where core damage occurs after containment failure. Discuss the reasons for these results, how the intact frequency was determined, and the potential for impacting all the release category frequencies.

RESPONSE

The Class II accident class is defined, as noted in the question, as sequences that occur after core damage. The intact frequency is calculated by summing all of the release category frequencies and then subtracting that total frequency from the core damage frequency.

Therefore, the intact frequency is determined mathematically; it is not calculated by model quantification.

Investigation into the reason for the intact frequency revealed that the model quantification was truncating low frequency Class II scenarios. As discussed in response to Question 2.d.i in this Enclosure, the Class II accident classes are split into two timing categories; early and non-early.

There is a separate Class II event tree for each category. For the early release class, the total Class II frequency is multiplied by 0.05 when the CDF value is transferred into the Level 2 model. Then the remaining event tree nodes are evaluated. This results in truncation of low frequency sequences.

A sensitivity study was performed to determine the impact of this truncation issue. The study was performed by setting the PRA model basic event related to the General Emergency declaration to 1.0 and requantifying the model. After quantification, the 95% and 5% split between not early and early were calculated using an Excel spreadsheet. The results demonstrated that the low frequency sequences had previously been truncated from the final results. It was noted in the re-quantified results that there was still a very small impact on Level 2 release results due to truncated sequences (approximately 5% of CDF). This release frequency was divided proportionally among each release category applicable to that accident class (i.e., Class II, IIE, IIV and IIVE). Model truncation is a known source of uncertainty in model results and while the proportional redistribution of the truncated cutsets does not provide an exact solution to this quantification issue, it is considered to provide a reasonable approximation of how these low contributors should be allocated to the release categories. Any deviations between the true release category allocations of the truncated frequency and those resulting from the proportional distribution of the frequency would correlate to very small changes to the cost benefit results and it is estimated that they would not impact the conclusions of the SAMA analysis.

The Class II accident class is the only accident class where the CDF values were multiplied by a split fraction before transferring into the Level 2 fault trees. Therefore, the larger impact observed for this accident class is not applicable to other accident classes. Other truncation issues would be small and are bounded by the uncertainty in other inputs, such as those documented in Section F.7.3 of the ER related to population estimates. Therefore, the SAMA evaluation would not be impacted by model truncation issues.

19 of 75

RA15-032 ENCLOSURE The attached revision to Table F.2-6 provides the re-calculated results. This sensitivity study shows the impact on all of the release frequencies.

Based on these revised results, the impact on the SAMA analysis was evaluated. Based on a new maximum averted cost-risk (MACR) of $6,073,600 (a 7.4% increase), one SAMA that was previously screened during phase I would be retained (SAMA 26).

The updated baseline cost benefit analysis shows that two additional SAMAs would be considered cost beneficial in the baseline analysis; specifically, SAMAs 3 and 4. In the 95th percentile sensitivity, no additional SAMAs become cost beneficial that were not before.

SAMAs 3 and 4 are now cost beneficial in the baseline case rather than in the 95th percentile sensitivity case. The updated results are provided in the table titled Summary of the Impact of Using the 95th Percentile PRA Results - REVISED below.

20 of 75

RA15-032 ENCLOSURE Table F.2 REVISED Summary Of LSCS Level 2 Release Categories (/Yr) (1), (2), (3)

Total Class CDF Intact LL/E LL/I LL/L L/E L/I L/L M/E M/I M/L H/E H/I H/L Release IA 8.46E-08 1.18E-08 N/A 0.00E+00 N/A 3.68E-08 1.08E-08 N/A 1.10E-09 1.67E-08 N/A 7.34E-09 2.05E-10 N/A 7.29E-08 IBE 3.43E-07 2.96E-07 N/A 0.00E+00 N/A 1.85E-08 1.28E-08 N/A 3.07E-09 7.53E-09 N/A 4.53E-09 7.40E-10 N/A 4.72E-08 IBL 2.94E-07 1.57E-07 N/A 0.00E+00 N/A N/A 8.33E-08 N/A N/A 3.72E-08 N/A N/A 1.63E-08 N/A 1.37E-07 IC 1.67E-07 1.44E-07 N/A 0.00E+00 N/A 1.08E-08 9.11E-09 N/A 1.60E-09 5.63E-11 N/A 1.57E-09 0.00E+00 N/A 2.31E-08 ID 3.53E-08 2.98E-09 N/A 0.00E+00 N/A 2.72E-08 0.00E+00 N/A 0.00E+00 4.92E-09 N/A 2.13E-10 0.00E+00 N/A 3.23E-08 (4)

II 8.77E-07 0.00E+00 N/A 0.00E+00 N/A N/A 0.00E+00 N/A N/A 8.75E-07 N/A N/A 1.97E-09 N/A 8.77E-07 (4)

IIE 4.62E-08 0.00E+00 0.00E+00 N/A N/A 0.00E+00 N/A N/A 4.61E-08 N/A N/A 1.04E-10 N/A N/A 4.62E-08 (4)

IIV 1.05E-07 0.00E+00 N/A N/A N/A N/A 2.94E-08 N/A N/A 7.51E-08 N/A N/A 1.80E-10 N/A 1.05E-07 (4)

IIVE 5.51E-09 0.00E+00 N/A N/A N/A 1.55E-09 N/A N/A 3.95E-09 N/A N/A 9.50E-12 N/A N/A 5.51E-09 IIIA 9.62E-10 2.37E-10 N/A 0.00E+00 N/A 0.00E+00 6.99E-10 N/A 8.82E-12 0.00E+00 N/A 1.73E-11 N/A N/A 7.25E-10 IIIB 1.49E-08 1.33E-08 N/A 0.00E+00 N/A 0.00E+00 1.32E-09 N/A 1.49E-11 0.00E+00 N/A 2.57E-10 N/A N/A 1.59E-09 IIIC 9.98E-09 0.00E+00 N/A 0.00E+00 N/A 6.16E-09 4.58E-10 N/A 3.09E-09 2.63E-10 N/A 1.85E-10 N/A N/A 1.02E-08 IIID 2.68E-08 0.00E+00 N/A N/A 0.00E+00 0.00E+00 N/A N/A 0.00E+00 N/A N/A 2.69E-08 N/A N/A 2.69E-08 IV 4.88E-07 0.00E+00 0.00E+00 N/A N/A 2.93E-07 N/A N/A 1.77E-07 N/A N/A 1.83E-08 N/A N/A 4.88E-07 V 8.32E-08 0.00E+00 N/A N/A N/A N/A N/A N/A N/A N/A N/A 8.32E-08 N/A N/A 8.32E-08 Total 2.58E-06 6.25E-07 0.00E+00 0.00E+00 0.00E+00 3.94E-07 1.48E-07 0.00E+00 2.36E-07 1.02E-06 N/A 1.43E-07 1.94E-08 N/A 1.96E-06 Notes to Table F.2 REVISED:

(1) Based on results of PRAQuant results at the sequence level. Level 2 quantified at a truncation value of 1E-12/yr.

(2) N/A indicates that the accident class did not contribute to release of that specific category.

(3) Numerical differences in column totals may occur due to rounding.

(4) Due to truncation issues, Class II, IIE, IIV and IIVE are manually adjusted to reflect containment failure and a release category. The small portion of CDF that is truncated from quantification (approximately 5%) is divided proportionally among the release categories.

21 of 75

RA15-032 ENCLOSURE Summary of the Impact of Using the 95th Percentile PRA Results - REVISED1 Change Averted Averted Net Value in SAMA Implementation Net Value Cost Risk Cost Risk (95th Cost ID Cost (per unit) (Base) (95th (Base) Percentile) Effective-Percentile) ness?

No SAMA 1 $12,940,000 $2,507,762 -$10,432,238 $5,366,611 -$7,573,389 SAMA 2 $400,000 $1,434,784 $1,034,784 $3,070,438 $2,670,438 No SAMA 3 $1,000,000 $1,026,189 $26,189 $2,196,044 $1,196,044 No SAMA 4 $635,242 $674,034 $38,792 $1,442,433 $807,191 No SAMA 5 $400,000 $386,266 -$13,734 $826,609 $426,609 Yes SAMA 6 $2,900,000 $86,788 -$2,813,212 $185,726 -$2,714,274 No SAMA 7 $962,403 $50,508 -$911,895 $108,087 -$854,316 No SAMA 8 $400,000 $258,846 -$141,154 $553,930 $153,930 Yes SAMA 9 $115,000 $223,818 $108,818 $478,971 $363,971 No SAMA 10 $260,219 $1,324,565 $1,064,346 $2,834,569 $2,574,350 No SAMA 11 $217,415 $20,561 -$196,854 $44,001 -$173,414 No SAMA 12 $4,401,674 $1,070,763 -$3,330,911 $2,291,433 -$2,110,241 No SAMA 14 $489,277 $443,654 -$45,623 $949,420 $460,143 Yes SAMA 15 $1,370,000 $3,383,546 $2,013,546 $7,240,788 $5,870,788 No SAMA 16 $475,000 $866,362 $391,362 $1,854,015 $1,379,015 No SAMA 18 $649,194 $608,722 -$40,472 $1,302,665 $653,471 Yes SAMA 19 $2,900,000 $3,431,860 $531,860 $7,344,180 $4,444,180 No SAMA 20 $1,150,000 $113,381 -$1,036,619 $242,635 -$907,365 No SAMA 21 $1,481,002 $754,468 -$726,534 $1,614,562 $133,560 Yes SAMA 22 $205,000 $33,706 -$171,294 $72,131 -$132,869 No SAMA 23 $1,370,000 $3,018,954 $1,648,954 $6,460,562 $5,090,562 No SAMA 24 $1,824,084 $366,787 -$1,457,297 $784,924 -$1,039,160 No SAMA 25 $112,000 $41,725 -$70,275 $89,292 -$22,709 No SAMA 26 $5,984,407 $2,283,117 -$3,701,290 $4,885,870 -$1,098,537 No No SAMA 27 $512,000 $196,045 -$315,955 $419,536 -$92,464 1

These results reflect corrections for SAMA 25 as documented in the response to Question 6.d (see p. 72 of this Enclosure).

22 of 75

RA15-032 ENCLOSURE QUESTION 2.e It is noted in Table F.2-6 of the ER that the majority of Class IV (ATWS) sequences result in a low/early (L/E) release and, as indicated in Table F.3-17 for the majority of these sequences, core melt is arrested in the reactor vessel. Discuss the reasons and basis for these results.

RESPONSE

The release magnitude (L) is directly related to the success of RX (core melt arrested in the reactor vessel). See the response in this Enclosure to Question 2.f, which discusses further the assignment of the release category for Class IV sequences with core melt arrested in the reactor vessel.

The release timing (E) is related to early containment failure for Class IV sequences. The containment is assumed to fail for Class IV scenarios as the heat into containment from continued power production exceeds the heat removal capability of the RHR system Success at the RX node (core melt arrested in reactor vessel) has a relatively high probability for Class IV sequences. The node examines the availability of the low pressure injection system after core damage. In ATWS accident sequences, reactor pressure vessel (RPV) injection is intentionally terminated per the emergency operating procedures to lower reactor power. Lowering RPV level too low can be one cause of core damage. Similarly, after depressurization, failure to restart injection can cause core damage. After core damage, the availability of these systems is again queried in the RX node. Since in many sequences, low pressure injection was not failed in Level 1 due to equipment failure, there is a higher probability low pressure injection is available in the RX node. In contrast, for Class I sequences, low pressure injection is failed in Level 1 model; therefore, there is a lower likelihood it will be recovered and successful in the Level 2 RX node.

23 of 75

RA15-032 ENCLOSURE QUESTION 2.f In Table F.3-17 of the ER, the reference Modular Accident Analysis Program (MAAP) case for Sequence IV-04, which is the dominant contributor to the low/early (L/E) release category frequency, is stated not to have been used for the release fractions because this MAAP case does not model the core melt arresting in the reactor vessel. Another MAAP case, with a release fraction 1/65th of that for the ATWS case, was used to determine the release fractions for the L/E release category. Provide additional support for the use of these release fractions.

RESPONSE

It should be noted that the reference MAAP cases in the Level 2 analysis are not necessarily exact models of the sequence, but are instead used along with the Level 2 Release Category rules to assign an appropriate end state to the Level 2 sequence.

A specific MAAP scenario modeling sequence IV-04 was not developed for the LSCS SAMA analysis. Rather, expert judgement and review of similar MAAP cases were used to assign an L/E end state to the sequence. Sequence IV-04, modeling success of in-vessel retention of the core (Level 2 Event Tree Node RX), is conservatively modeled if using the reference MAAP scenario for radionuclide release fractions since the reference MAAP scenario does not model this in-vessel retention. In-vessel retention would be expected to reduce radionuclide release fractions significantly for the following reasons:

RPV injection post-core damage prevents further heat up of radionuclides The suppression pool is not bypassed (i.e., fission products are retained in the pool instead of being released directly to the environment)

Note that Tables F.3-17 and F.3-18 of the ER indicate that MAAP case LS130533B was used to represent the L/E release category and that this scenario has a 40 hour4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> run time. However, LS130524B was actually used to represent the L/E release category as indicated in Table F.3-

19. Use of the LS130524B scenario in place of the LS130533B scenario is conservative since the LS130524B release fractions bound the release fractions from the LS130533B scenario and the LS130524B scenario results in a substantial release of radionuclides several hours earlier than LS130533B scenario (containment breach in LS130524B vs. containment vent in LS130533B). Both scenarios are containment flooding scenarios.

Sequence IV-04 represents an ATWS sequence with a wetwell airspace failure, successful RPV depressurization, in-vessel retention, and no suppression pool bypass. LS130524B represents an ATWS sequence with wetwell airspace failure, successful RPV depressurization, core spray available after core damage at vessel breach, and containment flooding. The use of MAAP scenario LS130524B to model sequence IV-04 as a surrogate for the L/E release category is judged to be reasonable since LS130524B models containment flooding shortly after vessel breach which mimics the actions of in-vessel retention by both cooling debris and retaining fission products in the local water inventory (i.e., scrubbing the release).

24 of 75

RA15-032 ENCLOSURE QUESTION 2.g The run time for several of the LSCS MAAP cases in Table F.3-18 of the ER is less than 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> past the declaration of a general emergency when effective offsite resources might be available for mitigating the accident. While it is stated that the representative MAAP cases were run until plateaus of the cesium iodide (CsI) and cesium hydroxide (CsOH) release fractions were achieved, provide more information to justify the adequacy of the MAAP release fraction results for the SAMA analysis for those cases with run times less than 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after the general emergency.

RESPONSE

Table 2.g-1 provides justification of the MAAP scenarios with durations less than 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after the General Emergency is declared.

Table 2.g-1 MAAP Scenario Adequacy Assessment General Release Category / Emergency Scenario Representative MAAP Declaration End Time Scenario End Time Justification Scenarios Time (Hours)

(Hours)

The H/E-BOC MAAP case results in catastrophic failure of containment that results in the plateau of release fractions by t=20 hours. The release fractions of CsI (0.94) and CsOH (0.87) indicate most of the CsI and CsOH has been released by this time. From t=20 H/E-BOC / LS130528 0.5 40 hours to t=40 hours there is negligible increase in any radionuclide release fraction. Given the very large release of CsI and CsOH and the stable trend of the radionuclide release fractions after 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br />, no additional significant releases would occur out to 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after the General Emergency.

25 of 75

RA15-032 ENCLOSURE Table 2.g-1 MAAP Scenario Adequacy Assessment General Release Category / Emergency Scenario Representative MAAP Declaration End Time Scenario End Time Justification Scenarios Time (Hours)

(Hours)

The H/E MAAP scenario models lower pedestal wall failure for corium-concrete interactions that results in the inundation of the lower pedestal by the wetwell water inventory prior to the end of the scenario. Given this inundation, the debris in the lower pedestal is H/E / LS130521x 0.5 40 quenched by a large volume of water.

Additionally, debris located in the upper pedestal and drywell has stabilized at relatively low temperatures (< 1800°F).

Given these conditions, no additional significant radionuclide releases would occur out to 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after the General Emergency declaration.

The H/I MAAP scenario models lower pedestal wall failure for corium-concrete interactions that results in the inundation of the lower pedestal by the wetwell water inventory prior to the end of the scenario. Given this inundation, the debris in the lower pedestal is H/I / LS130536x 5.6 48 quenched by a large volume of water.

Additionally, debris located in the upper pedestal and drywell has stabilized at relatively low temperatures (< 1800°F).

Given these conditions, no additional significant radionuclide releases would occur out to 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after the General Emergency declaration.

26 of 75

RA15-032 ENCLOSURE Table 2.g-1 MAAP Scenario Adequacy Assessment General Release Category / Emergency Scenario Representative MAAP Declaration End Time Scenario End Time Justification Scenarios Time (Hours)

(Hours)

Tables F.3-17 and F.3-18 of the ER indicate that MAAP case LS130533B was used to represent the L/E release category and that this scenario has a 40 hour4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> run time. However, LS130524B was actually used to represent the L/E release category as indicated in Table F.3-19. Use of the LS130524B scenario in place of the LS130533B scenario is conservative since the LS130524B release fractions bound the release fractions from the LS130533B scenario. Additionally, the L/E / LS130524B 0.5 40 LS130524B scenario results in a substantial release of radionuclides several hours earlier than LS130533B scenario. Both scenarios are containment flooding scenarios.

Given containment is being flooded and core debris is quenched for both LS130533B and LS130524B at t=40 hours no additional significant radionuclide releases would occur out to 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after the General Emergency declaration.

27 of 75

RA15-032 ENCLOSURE QUESTION 2.h In Table F.3-18 of the ER, the time of 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> for declaration of a general emergency for the moderate/intermediate (M/I) release category is indicated to have been determined probabilistically. Describe how this time was determined, how it was utilized in the Level 2 and Level 3 analysis and the impact on the SAMA cost-benefit analysis if the minimum steam coolant water level limit (MSCWLL) time of 27.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> is used in the analysis.

RESPONSE

The M/I release category is dominated by Class II sequences. The analysis of the timing of a General Emergency declaration has been evaluated from a probabilistic standpoint for Class II accident scenarios. This evaluation was performed because of the potential differences in interpretation of the Emergency Action Levels (EALs) and the probability that the Emergency Director could delay the declaration of the General Emergency resulting in an early release.

Interviews with key emergency response personnel were performed to determine the best estimate probability that the interpretation of the EALs would be such that the General Emergency declaration would be delayed resulting in an early release as opposed to an intermediate or late release. These interviews consisted of case studies and a discussion of EALs.

The EAL key parameter alerting the Emergency Director to containment failure and the need for a General Emergency declaration is the containment design pressure of 62 psig. From an emergency classification perspective, the containment design pressure indicates containment failure is imminent. However, actual (mean) containment failure is not estimated to occur until 140 psig is reached. This 140 psig represents a best estimate of containment performance under quasi-static, extreme temperature (< 500°F) conditions as documented in the LSCS Level 2 Notebook. As a result, operators would likely declare the General Emergency hours before containment failure actually occurs.

The mean probability of failure to declare a General Emergency with adequate time to take protection measures for the general public prior to containment failure is 5% or 0.05 based on the expert opinion information. As a result, the LSCS Level 2 model is structured such that 5%

of the Class II accidents can result in an Early release, but 95% of the Class II accident releases are non-early.

The representative MAAP case for the M/I release category (LS130516) indicates that containment pressure reaches 62 psig at t=16.4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> and that containment failure pressure (140 psig) is exceeded at t=27.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.

Given that evacuation of the 10-mile region surrounding LSCS is estimated to take no more than 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, evacuation would be completed prior to containment failure based on these setpoints. For this MAAP case, the assumption that the General Emergency declaration time occurs at t=4 hours is inconsequential to the release timing since evacuation is expected to be completed many hours prior to containment failure.

If the MSCWLL time of 27.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> was used as the General Emergency declaration timing for the M/I endstate, evacuation would not be completed prior to containment failure (@ 140 psig).

However, vessel breach does not occur until after t=35 hours at which time evacuation will have 28 of 75

RA15-032 ENCLOSURE been completed. Prior to t=35 hours, radionuclide release fractions are very low. The CsI release fraction (0.003) and CsOH release fraction (0.002) at t=35 hours indicate that impacts on offsite dose risk metrics would be minimal. Offsite cost risk, onsite cost risk, and onsite dose risk metrics would not be impacted by a change in evacuation timing.

29 of 75

RA15-032 ENCLOSURE QUESTION 3.a Section F.5.1.6.2.1 indicates that the impact of using the LSCS 2013 seismic hazard curves on the Risk Methods Integration and Evaluation Program (RMIEP) analysis has been investigated.

Provide more information on the source of these hazard curves including a characterization of how they compare with the LSCS hazard curves included in the industry's response to the NRC's Near Term Task Force (NTTF) on Fukushima Recommendation 2.1. Also, provide the LSCS seismic CDF using the updated hazard curves and the simplified methodology of Generic Issue (GI)-199.

RESPONSE

The LSCS 2013 seismic hazard curves used in the SAMA analysis are the same as the seismic hazard curves developed to address NTTF Recommendation 2.1; therefore, there is no difference between the hazard curves used in the SAMA analysis and those used to respond to NTTF Recommendation 2.1. The seismic hazard curves were developed as described in the LSCS Seismic Hazard and Screening Report [1] and the details of that process have not been reproduced here. The source document for the LSCS 2013 seismic hazard curves is identified as Reference 23 in the LSCS Seismic Hazard and Screening Report [1].

The simplified GI-199 methodology [2] was used in conjunction with the 2013 LSCS seismic hazard curves to generate CDF values for the plant level fragility data available for LSCS Unit 1 and Unit 2. Table C-2 of GI-199 provides plant level fragility data used to estimate the plant CDF. The Unit 2 results are the same as those for Unit 1.

The table below shows a summary of the CDF results for LSCS. The first column presents the plant name as provided in Table C-2 of GI-199. The second and third columns present the plant level fragility median seismic capacity (C50) and the composite logarithmic standard deviation (C) as provided in Table C-2 of GI-199. The last column provides the CDF estimated using the GI-199 methodology based on the LSCS 2013 hazard curve.

PGA Fragility Data PGA Core Damage Frequency Plant C50 C

(/yr)

LSCS 1 1.32 0.4 2.25E-06 LSCS 2 1.32 0.4 2.25E-06 REFERENCES

[1] Exelon Generation Company, LLC letter to the NRC, Exelon Generation Company, LLC, Seismic Hazard and Screening Report (Central and Eastern United States (CEUS)

Sites), Response to NRC Request for Information Pursuant to 10 CFR 50.54(f)

Regarding Recommendation 2.1 of the Near-Term Task Force Review of Insights from the Fukushima Dai-ichi Accident, LaSalle County Station, Units 1 and 2, Correspondence No. RS-14-068, dated March 31, 2014.

30 of 75

RA15-032 ENCLOSURE

[2] U.S. Nuclear Regulatory Commission, Implications of Updated Probabilistic Seismic Hazard Estimates in Central and Eastern United States on Existing Plants Generic Issue 199 (GI-199), Safety Risk Assessment, Washington, DC, August, 2010.

31 of 75

RA15-032 ENCLOSURE QUESTION 4.a Indicate the Sector Population and Economic Estimator (SECPOP2000) version used in the analysis. Confirm that manual entry of economic values was sufficient to prevent known SECPOP2000 errors from influencing the SAMA results. The known errors are described in NRC Request 4.b on page E1-24 of the response letter to NRC from another licensee (ADAMS Accession No. ML102100588).

RESPONSE

The LSCS SAMA analysis utilized SECPOP 2000 version 3.12. The identified errors impact economic inputs and outputs of the SECPOP 2000 code. No SECPOP 2000 related economic data were used for the LSCS SAMA analysis. Economic data input was calculated outside of the SECPOP 2000 code. This approach negates the potential for formatting errors in the SECPOP 2000 economic database, including gaps in the database file. Additionally, the LSCS population and economic data contained in the SAMA base case and sensitivity MACCS2 Site Input Files were formatted manually and have been checked to ensure that they conform to the formatting requirements specified in NUREG/CR-6613, Vol. 1 (Code Manual for MACCS2:

Users Guide).

32 of 75

RA15-032 ENCLOSURE QUESTION 4.b Compared to higher (total) estimates in Table F-3.2 for an assumed uniformly distributed population, provide supporting justification that 2010 population estimates in Table F-3.5 are more representative of the actual 2010 population distribution.

RESPONSE

The LSCS MACCS2 model population data from the SECPOP 2000 code is based on block-level census data that provides a much higher resolution on population distribution in a region relative to whole county-level data since block regions typically encompass small sections of a county. The use of these census blocks removes much of the population overestimation obtained from using county-level census data since population information is more finely binned.

Specifically, the use of block-level data avoids the over-estimation of population in counties that comprise portions of large metropolitan areas where population densities may vary significantly (e.g., moving from rural to suburban to urban). In the case of LSCS, where the Chicago metropolitan area spans several of the outlying counties in the 50-mile region, county based uniform population estimates skew the 50-mile population higher since many of the higher population density areas lie outside of the 50-mile region surrounding LSCS. Within the 50-mile region of LSCS are small sections of these highly populated counties (e.g., Cook County). The portions of these counties situated closest to LSCS are typically much less populated (i.e., more rural and suburban) than those county regions situated closest to Chicago (i.e., more urban).

Figure 3.1-1 of the ER illustrates this point. As the figure indicates, the 50-mile region surrounding LSCS only skirts the perimeter of the Chicago metropolitan area and only captures the less populated edges of some of the counties in the metropolitan area. Based on this information, use of the 2010 block-level population estimates in Table F-3.5 of the ER, rather than the assumed uniformly distributed population estimates in Table F-3.2, is reasonable because the 50-mile region is comprised of farms and smaller cities as indicated by the 2010 block-level population data and the figure.

33 of 75

RA15-032 ENCLOSURE QUESTION 4.c Indicate the source of the county population growth rates in Table F.3-1 of the ER.

RESPONSE

The county growth rates were originally obtained from the following source:

State of Illinois, Department of Commerce and Economic Opportunity, 2000-2030 Population Projections, http://www.ildceo.net/dceo/Bureaus/Facts_Figures/Population_Projections/

However, the reference link is no longer active. The same information can now be found at the following internet web site:

http://www2.illinoisbiz.biz/popProj/

34 of 75

RA15-032 ENCLOSURE QUESTION 4.d Compare the more recent 2012 Census of Agriculture data to values used in the analysis from the 2007 Census of Agriculture data with any consumer price index updates.

RESPONSE

Table 4.d-1 provides a comparison of the farm wealth values calculated in the LSCS SAMA analysis using the 2007 Census of Agriculture (updated to July 2013 using the CPI) and the same parameters using the 2012 Census of Agriculture (also updated to July 2013 using the CPI). As the table indicates, the fraction of farmland and fraction of dairy farms for the surrounding region is relatively stable between 2007 and 2012. The annual sale of farm products and farm wealth values are typically larger considering the 2012 Census of Agriculture data.

Table 4.d-2 provides the breakdown of offsite economic costs based on farm-dependent costs.

As the table identifies, farm-dependent costs encompass only 5% of the total offsite economic cost risk (2.70E+03 / 5.34E+04). Given the relatively minor contribution from farm-dependent cost risk, use of the 2012 Census of Agriculture data would not have a significant impact on the SAMA analysis.

35 of 75

RA15-032 ENCLOSURE Table 4.d-1 2012 and 2007 Census of Agriculture Farm Wealth Input Parameter Comparisons Farm Property Farm Sales Fraction Farm Fraction Dairy Value

($/hectare)

($/hectare)

County 2012 2007 2012 2007 2012 2007 2012 2007 Bureau 0.809 0.860 0.001 0.002 2,306 1,566 18,438 11,275 Cook 0.014 0.014 0.036 0.036 3,106 4,601 26,282 28,720 DeKalb 0.984 0.918 0.005 0.013 2,949 2,013 19,736 12,885 DuPage 0.035 0.038 0.000 0.000 3,415 4,374 17,459 20,877 Ford 0.992 0.871 0.000 0.002 1,537 1,331 19,065 11,055 Grundy 0.811 0.805 0.004 0.003 1,467 1,206 19,122 11,556 Iroquois 0.936 0.948 0.002 0.006 1,821 1,525 17,103 11,217 Kane 0.506 0.578 0.019 0.018 2,875 2,544 22,275 13,552 Kankakee 0.791 0.891 0.003 0.008 2,072 1,562 16,610 12,053 Kendall 0.633 0.814 0.000 0.008 1,962 1,532 21,844 12,032 LaSalle 0.829 0.886 0.000 0.001 1,884 1,263 19,852 11,680 Lee 0.795 0.852 0.001 0.002 2,416 1,338 19,061 11,992 Livingston 0.982 0.941 0.010 0.009 1,542 1,378 18,079 11,538 Marshall 0.845 0.828 0.004 0.005 1,608 1,215 17,408 11,262 McLean 0.914 0.893 0.031 0.053 1,787 1,339 20,553 11,633 Ogle 0.775 0.755 0.015 0.015 3,194 1,744 17,415 12,608 Peoria 0.632 0.654 0.002 0.013 1,849 1,203 17,384 10,798 Putnam 0.587 0.613 0.005 0.008 3,354 2,557 15,798 10,971 Tazewell 0.812 0.793 0.008 0.011 1,931 1,387 18,702 11,230 Will 0.437 0.412 0.012 0.013 1,783 1,427 19,652 15,683 Woodford 0.956 0.854 0.004 0.005 1,840 1,521 19,485 11,812 36 of 75

RA15-032 ENCLOSURE Table 4.d-2 LSCS Farm-Dependent Offsite Cost Risk Offsite Farm Total Total Farm Dependent Offsite Release Frequency Offsite Dependent Offsite Cost Category (per yr) Economic Economic Economic Risk Cost ($) Cost Risk Cost ($) ($/yr)

($/yr)

H/E-8.32E-08 8.68E+10 2.59E+09 7.22E+03 2.15E+02 BOC H/E 5.93E-08 4.66E+10 1.74E+09 2.76E+03 1.03E+02 H/I 1.9E-08 5.02E+10 1.84E+09 9.54E+02 3.50E+01 M/E 2.14E-07 4.39E+10 1.94E+09 9.39E+03 4.15E+02 M/I 9.27E-07 3.53E+10 1.94E+09 3.27E+04 1.80E+03 L/E 3.88E-07 3.19E+08 1.78E+08 1.24E+02 6.91E+01 L/I 1.45E-07 1.22E+09 4.68E+08 1.77E+02 6.79E+01 INTACT 7.45E-07 8.57E+05 8.57E+05 6.38E-01 6.38E-01 TOTAL 5.34E+04 2.70E+03 37 of 75

RA15-032 ENCLOSURE QUESTION 4.e Section F.3-7 of the ER mentions greater plume washout and radionuclide deposition during precipitation events. Estimate the sensitivity of the offsite population dose and offsite economic cost risk to atmospheric conditions with and without precipitation events.

RESPONSE

Three years (2010, 2011, and 2012) of LSCS meteorological data were analyzed using the LSCS MACCS2 model. Table 4.e-1 provides yearly precipitation, offsite dose risk, and offsite economic cost risk comparisons for these meteorological datasets (2012 is the base case for the SAMA analysis). As the table indicates, even though the year 2012 had less total precipitation relative to years 2010 and 2011, year 2012 had higher offsite dose and cost risk results. These sensitivities indicate that precipitation is only one of the atmospheric conditions that influence offsite dose and cost risk. Other notable influencing atmospheric conditions include wind direction and wind speed.

Table 4.e-1 LSCS MACCS2 Meteorological Sensitivity Summary Total Population Dose Delta Cost Risk Delta Case Precipitation Risk 50 mile Base (%) ($/year) Base (%)

(inches) (Person-rem/year)

Year 2012 Meteorological 25.3 7.1 -- 5.34E+04 --

Data (SAMA Base Case)

Year 2011 Meteorological 29.2 7.0 -0.9% 5.01E+04 -6.1%

Data Year 2010 Meteorological 26.7 6.8 -4.4% 4.86E+04 -8.8%

Data 38 of 75

RA15-032 ENCLOSURE QUESTION 4.f Specify the software codes and versions used for calculating the core inventory.

RESPONSE

ORIGEN version 2.1 was used to calculate the LSCS core inventory.

39 of 75

RA15-032 ENCLOSURE QUESTION 4.g NEI 05-01 lists several radionuclides to be considered in the core inventory for the SAMA analysis. Describe how the core inventories for Co-58 and Co-60 were calculated and justify the appropriateness of inherent assumptions to LSCS Units 1 and 2.

RESPONSE

The Co-58 and Co-60 inventories used in the LSCS SAMA (i.e., 2.15E+16 Bq and 2.36E+16 Bq, respectively) were calculated based on the BWR core inventory provided in NUREG/CR-4551, Vol 2, Rev. 1, Part 7, page A-7 because these two inventories were not calculated in the plant specific calculation. The values from NUREG/CR-4551 (i.e., 2.024E+16 Bq for Co-58, and 2.423E+16 Bq for Co-60) were scaled based on core thermal power levels between LSCS (3489 MWth) and the NUREG/CR-4551 BWR (3578 MWth), equating to a factor of 0.975 (3489/3578). In the scaling calculation, the inventory for Co-58 was erroneously taken as 2.204E+16 Bq rather than 2.024E+16 Bq due to a transcription error, thereby increasing the core inventory for Co-58 by about 9%.

The use of these Co-58 and Co-60 inventories is reasonable since offsite dose and economic costs are typically dominated by CsI and CsOH. The core inventory of Co-58 and Co-60 are typically one or more orders of magnitude less than that of Cs in terms of activity and have a marginal impact on offsite consequences.

40 of 75

RA15-032 ENCLOSURE QUESTION 4.h Section F.2 of the ER indicates the PRA used in the SAMA analysis is a model for Unit 2, unless otherwise noted. Section F.3.5 indicates the plant-specific calculation of core inventory represents bounding isotopic values. Clarify if the calculation of radionuclides shown in Table F.3-11 of the ER relates to LSCS Unit 1, Unit 2, or both units. Justify why the calculation is bounding for either unit.

RESPONSE

The LSCS core inventory calculation is based on the equilibrium core design developed for Unit 1 and is representative of Unit 2. This calculation is bounding for both units for the following reasons:

Fuel burned in more than one cycle ignores time spent in refueling outages.

Extending the LSCS cycle length from 680 Effective Full Power Days (EFPD) to 711 EFPD without raising initial enrichments of the fuel is conservative since longer cycles generally result in higher activities for a given power.

The ORIGEN code calculated inventories at 100 EFPD and at end of cycle. The inventory values chosen for each isotope for the MACCS2 analysis represent the higher (i.e., bounding) of the 100 EFPD and end of cycle values.

While small differences may exist between the Unit 1 and Unit 2 reactors (e.g., actual power histories), these small differences would result in marginal differences to core inventory calculations.

The core design process for each unit ensures that each core inventory remains within the single design basis source term and is therefore applicable to both units.

41 of 75

RA15-032 ENCLOSURE QUESTION 5.a.i The turbine trip with bypass initiating event (%TT) is indicated to have a frequency of 0.8 per year, which implies that a number of turbine trips with bypass have occurred over the LSCS operating history. Do any of these occurrences suggest possible cost-beneficial SAMAs?

RESPONSE

As described in Table F.5-1 of the ER, most of the risk associated with a turbine trip event is associated with ATWS scenarios. The SAMAs identified to address these events (SAMAs 4 and 5) were both shown to be potentially cost beneficial.

The LSCS turbine trip initiating event frequency of 7.98E-01 per year is based on a Bayesian update of a generic prior for the period of 2006-2010, which includes two plant specific events (one for each unit). This initiating event frequency is almost identical to the generic frequency of 8.30E-01 per year from NUREG/CR-6928. Review of the two plant-specific events used in the Bayesian update indicates that the causes of the events are not related to one another, and that the site has taken corrective action to address the root causes of the events. These conditions do not imply any outlier behavior at LSCS or indicate that there is a potential to identify cost-beneficial SAMAs related to turbine trip events.

42 of 75

RA15-032 ENCLOSURE QUESTION 5.a.ii The discussion of potential SAMAs for a significant number of events (RCVCL-1BE, %TIA, RCVSEQ-DLOP-041, BWTOPWTHXSTBYH--, and others) involve water hammer induced loss of coolant accidents (LOCAs) related to the generation of a LOCA signal on high drywell pressures when an actual LOCA does not exist and are mitigated by SAMA 7. Provide more detail on the water hammer scenarios in general and more specifically those involving the above identified events including: how loss of instrument air initiated events involve water hammer induced LOCAs, why alignment of the standby turbine building closed cooling water heat exchanger train is needed, why water hammer LOCAs are not important for other non-LOCA transients, and the potential for preventing water hammer by preventing residual heat removal (RHR) pump start when water hammer conditions exist or preventing draining of the RHR line.

Also, clarify the SAMA 7 statement in Table F.5-4 of the ER that the LOOP-delayed LOCA scenario is not specifically modeled in the PRA.

RESPONSE

There are four typical cases associated with water hammer events at LSCS.

Case 1: Transients that do not cause an immediate LOCA signal but nevertheless may induce a LOOP event at t=0 (time of scram).

Case 2: LOCA sequences that cause an immediate scram and LOCA signal. This condition has been linked by the NRC to a higher conditional LOOP probability than Case 1.

Case 3: Transients that may induce a delayed LOCA signal (i.e., High Drywell (DW) pressure) and a conditional LOOP event at approximately 30 minutes following scram. These transients are different than Case 1 because:

1) The RHR system may be realigned during the 30 minutes prior to the LOCA signal and be in suppression pool cooling. Probability =

1.0

2) The conditional LOOP probability when the LOCA signal occurs and the emergency core cooling system (ECCS) pumps auto start is assessed as significantly higher than Case 1 consistent with the conditional probability assigned in Case 2.

Case 4: LOOP or DLOOP initiating events which disable the systems at t=0 and then the systems will run only on the diesel generator until offsite AC power can be restored (i.e., Case 3 failure modes are not applicable).

The Case 1 scenario below could lead to a postulated water hammer and degradation of the plant mitigating systems:

1. One RHR train is operating in suppression pool cooling mode or test mode.

2. A plant transient occurs.

43 of 75

RA15-032 ENCLOSURE

3. If no LOOP occurs, the RHR train continues to operate as designed and there is no additional risk associated with operating in the suppression pool cooling mode. If a LOOP is the plant initiator or a plant trip results in a LOOP the scenario continues.

4. With a LOOP, the RHR discharge piping drains through the full flow test valve to the suppression pool. A void is created from the outboard isolation valve in the Reactor Building down to an elevation supported by the vacuum.

5. An emergency diesel generator (EDG) is started and power is restored to the electrical bus feeding the RHR train.

6. The operators fail to perform one of the following actions to prevent a water hammer event:

1) Fail to recognize the low discharge pressure in the RHR discharge piping and place the affected RHR pump in pull-to-lock to prevent it from starting; or, failing this,

2) Fail to properly fill and vent the RHR discharge piping prior to RHR pump start (time constraints may preclude refill of the RHR lines).

7. The RHR pump is started, either through operator action or as result of a low pressure coolant injection (LPCI) automatic start signal. The LPCI signal (i.e., the ECCS start signal) is from either low-low (Level 1) RPV water level or a high drywell pressure setpoint being reached.

8. Water hammer in the RHR system occurs with one of the following possible outcomes:

The RHR train remains available; or RHR piping leakage occurs and causes train to be unavailable and induces potential flood effects; or RHR piping is crimped and plugs.

For Case 2, the initiator is a LOCA, which results in a higher condition of LOOP probability than in Case 1.

1. In these cases, a train of RHR is running in suppression pool cooling mode when the LOCA initiating event occurs.

2. A LOCA signal is registered (high DW pressure and/or reduced RPV level).

3. A conditional LOOP occurs. With the LOOP, the RHR discharge piping drains through the full flow test valve to the suppression pool. A void is created from the outboard isolation valve in the Reactor Building down to an elevation supported by the vacuum.

4. An emergency diesel generator (EDG) is started and power is restored to the electrical bus feeding the RHR train.

44 of 75

RA15-032 ENCLOSURE

5. Initiators that lead to a quick increase in drywell pressure (within a few minutes of the initiating event) are assumed to be such that operator action (to place the previously operating pump into PTL) cannot be accomplished in time to prevent system start up and a potential water hammer. These initiators are: %S2 (Small LOCA); %S1 (Medium LOCA); %A (Large LOCA);

and, %R (Excessive LOCA).

6. Water hammer in the RHR system occurs with the same potential outcomes described for Case 1.

The Case 3 scenario below could lead to a postulated water hammer and degradation of the plant mitigating systems:

1. One RHR train is initiated in SPC mode or injection mode after a plant transient occurs and before an induced LOCA signal or LOOP occurs. HPCS, LPCS, or RHR C could also be in this configuration (i.e., injection or minimum flow modes).

2. A LOCA signal (high DW pressure) is induced because of the nature of the initiating event (e.g., loss of SW). Reaching the high DW pressure setpoint would occur if the operators fail to initiate administrative containment pressure control using the 2-inch vent/purge line (this is included in the fault tree logic). The LOCA signal causes ECCS pumps to start and a LOOP is induced).

3. With a LOOP, the RHR discharge piping drains through the full flow test valve to the suppression pool. A void is created from the outboard isolation valve in the Reactor Building down to an elevation supported by the vacuum. If the systems are operating in injection mode or min flow mode, the only possibility of a draindown event occurs if the pump discharge check valve or minimum flow line check valve fails to close creating a drain path back to the suppression pool.

4. An emergency diesel generator (EDG) is started and power is restored to the electrical bus feeding the RHR train.

5. The operators fail to perform one of the following actions to prevent a water hammer event (short time frame 20-30 seconds):

1) Fail to recognize the low discharge pressure in the RHR discharge piping and place the affected RHR pump in pull-to-lock to prevent it from starting; or, failing this,

2) Fail to properly fill and vent the RHR discharge piping prior to RHR pump start (time constraints may preclude refill of the RHR lines).

6. The RHR pump is started as result of a low pressure coolant injection (LPCI) automatic start signal. The LPCI signal (i.e., the ECCS start signal) is from either low-low-low (Level 1) RPV water level or a high drywell pressure setpoint being reached.

7. Water hammer in the RHR system occurs with one of the following possible outcomes:

The RHR train remains available; 45 of 75

RA15-032 ENCLOSURE or RHR piping leakage occurs and causes train to be unavailable and induces potential flood effects; or RHR piping is crimped and plugs.

The Case 4 scenario is similar to Case 1 except the initiator is a guaranteed LOOP event at t=0.

If a LOCA signal occurs at 30 minutes, the running ECCS systems are not shed.

The events %TIA and BWTOPWTHXSTBYH-- are relevant to these scenarios because they are related to loss of Drywell Cooling, which makes it necessary for the operators to perform the action to initiate the 2-inch containment vent to avoid the 2 psig high drywell pressure signal (the cause of the LOCA signal).

Sequence flag RCVSEQ-DLOP-041 is related to LOOP events, which, as documented in the case descriptions above, are linked to pipe voiding and the potential for subsequent water hammer events.

Accident class flag RCVCL-1BE is related to the water hammer events because the LOOP initiating events first tie them to the LOOP event tree, and the subsequent logic of the event tree bins them as early station blackout (SBO) events due to the water hammer induced failures even though emergency 4kV diesel power may be available. This is because the failures caused by the water hammer event share the same path as other LOOP/DLOOP scenarios that are generally SBO events. Specifically, the water hammer event fails the ECCS systems due to either the loss of the suppression pool suction source or due to the consequences of flooding from the water hammer event, and finally AC power recovery fails, all of which occur in the early stages of the scenario.

Water hammer is not important for non-LOCA transients that do not result in loss of drywell cooling because the high drywell pressure signal would not be generated.

The prevention of RHR pump start to preclude water hammer is already a proceduralized response and it is accounted for in the PRA. For non-LOCA initiators in which a LOOP initiating event occurs (or a consequential LOOP occurs with the initiator), the RHR pumps are shed from the emergency bus and they are not re-loaded when the buses are re-energized by the EDG.

There would be about 30 minutes before the high drywell pressure signal would be generated for loss of drywell cooling conditions and the scenarios in which the operators fail to prevent RHR restart are smaller contributors to risk due to the relatively long time available to perform the proceduralized action. Water hammer can also occur in these cases when the operators fail to properly fill and vent the RHR system before re-start, but this is also a proceduralized task. For scenarios in which a LOCA signal induces a LOOP, only about 20 seconds are available before the ECCS pumps are reloaded onto the bus, and the probability of failing to block the pump start is 1.0 because it is assumed that there is not enough time to place the pumps in Pull-to-Lock. While the action to prevent pump start is proceduralized, no credit is taken for this action.

46 of 75

RA15-032 ENCLOSURE Options were considered to include a fast-acting air operated or solenoid operated valves in the RHR discharge line that would fail closed on loss of power to prevent draining the pipe inventory to the suppression pool; however, in addition to being relatively high cost changes, they would introduce the potential for undesired flow blockages, and these changes were not pursued as SAMAs.

Once a LOCA signal occurs due to the loss of drywell cooling and failure to use the 2-inch containment vent, there are a series of operator actions required to address the automated actuations that occur, but these actions are not explicitly included in the PRA model because the dominant contributors to core damage are considered to be the water hammer events. It is this portion of the scenario that is the focus of the statement in the Table F.5.4 description of SAMA 7 regarding the LOOP-delayed LOCA scenario that is not modeled. Including these scenarios in the PRA model would result in a very small change in plant risk that would impact neither the SAMA identification process nor the conclusions of the SAMA analysis.

47 of 75

RA15-032 ENCLOSURE QUESTION 5.a.iii For event 2ADRX-TRANS--H-- (p. F-244 of the ER), failure to manually depressurize the reactor pressure vessel, the action is indicated to be required for about 90 percent of the scenarios because failure to initiate suppression pool cooling (SPC) results in containment failure. Discuss these scenarios in more detail including the reason for this high percentage compared to other causes of failure of high-pressure injection.

RESPONSE

In these scenarios, high pressure injection is initially successful (high pressure core spray operates), but the operators fail to initiate suppression pool cooling (SPC) and the suppression pool heats up, resulting in a procedurally driven direction to depressurize the RPV. The operators fail to manually depressurize the RPV and the HPCS system continues to run. The operators fail to initiate SPC to prevent containment pressurization, and they fail to vent the containment, which results in containment failure due to over-pressurization. The release of steam into the reactor building results in an adverse environment, which fails the ECCS pumps and prevents access to the areas required to align alternate injection systems. Core damage ensues.

In the cases where this combination of operator action failures exist, failure of high pressure injection is another potential path to core damage, but because the same failures of SPC initiation and containment venting will ultimately lead to failure of the low pressure injection systems, they are non-minimal scenarios.

The operator action 2ADRX-TRANS--H-- is applicable only to manual depressurization for transient events. It is not applicable to LOCAs or other situations that involve a loss of reactor inventory. Therefore, for these scenarios, HPCS, motor driven feedwater and RCIC would all be required to fail the high pressure injection function. The probability of such system failure combinations relative to the conditional JHEP of the SPC initiation and venting failures after the failure to manually depressurize post ADS inhibit is relatively low.

48 of 75

RA15-032 ENCLOSURE QUESTION 5.a.iv Regarding event 2VYFNNWVY01--X-- (p. F-271 of the ER), VY NW corner room (RHR A) cooling fan 2VY01C fails to run, previous LSCS evaluations could not demonstrate that portable fans would provide adequate cooling for the reactor building corner rooms when the normal cooling system failed. Discuss this previous evaluation and if sufficient cooling can be achieved if the portable fans can provide air movement through the room coolers assuming cooling water remains available.

RESPONSE

Prior to the Core Standby Cooling System (CSCS) valve replacement project at LSCS (September 2004), Design Engineering performed room cooling calculations using GOTHIC code to attempt to demonstrate continued operability of A RHR room cooling without cooling water from CSCS (i.e., fans only).

The calculations were performed using a number of assumptions including the removal of floor plugs to increase air flow and heat exchange with the rest of the building. With the floor plugs removed, the temperatures on the first floor of the Reactor Building near the floor plugs were causing safety related buses to be considered inoperable in that configuration. With the floor plugs in, the temperatures in the A RHR room exceeded the operability limits. Based on these calculations and engineering judgement, it was concluded that fans alone (i.e., portable fans with similar flow to the permanently installed fans) would be insufficient to provide adequate room cooling. These evaluations demonstrated that cooling water is required to maintain temperature in the room. These calculations were not finalized because they were not successful in demonstrating operability, and LSCS eventually submitted a License Amendment Request to extend the allowed outage time to complete the CSCS valve replacement.

If cooling water remains available, it is conceivable that the room coolers could be modified such that a portable fan could be used in conjunction with a temporary duct connection to provide flow through the room coolers for alternate room cooling. However, as stated in the ER, SAMA 1 addresses a large portion of the risk associated with the scenarios in which the ECCS room cooling fans fail. In order to estimate the potential impact of this type of alternate room cooling strategy, the cutsets were modified to eliminate reactor building corner room cooling fan failures (assumes the SAMA eliminates all risk associated with fan failures). Specifically, the following fan failure events were set to 0.0 in the cutsets associated with the plant configuration in which the reliable hard pipe containment vent (SAMA 1) has been implemented:

2VYFN2VY02C--A--: VY SW CORNER ROOM (HPCS) FAN 2VY02C FAILS TO START 2VYFN2VY02C--X--: VY SW CORNER ROOM (HPCS) FAN 2VY02C FAILS TO RUN 2VYFN2VY03C--A--: VY SE CORNER ROOM (RHR B & C) COOLING FAN 2VY03C FAILS TO START 2VYFNCSNWVY01A--: VY NW CORNER ROOM (RHR A) COOLING FAN 2VY01C FAILS TO START 2VYFNNWCORNERM--: VY NW CORNER ROOM (RHR A) COOLING / VENTILATION MUA 2VYFNNWVY01--X--: VY NW CORNER ROOM (RHR A) COOLING FAN 2VY01C FAILS TO RUN 49 of 75

RA15-032 ENCLOSURE 2VYFNSECORNERM--: VY SE CORNER ROOM (RHR B & C) COOLING /

VENTILATION MUA 2VYFNSEVY03CBX--: VY SE CORNER ROOM (RHR B & C) COOLING FAN 2VY03C FAILS TO RUN 2VYFNSWCORNERM--: VY SW CORNER ROOM (HPCS) COOLING MUA When the 95th percentile PRA results are considered, the averted cost risk associated with these changes is only about $42,000. The implementation cost for SAMA 16, which is a similar type of change related to the use of portable fans and temporary ductwork for alternate room cooling, is $475,000 (over 10 times larger than the averted cost-risk). This SAMA would not be cost beneficial for LSCS.

50 of 75

RA15-032 ENCLOSURE QUESTION 5.b.i For event 2TDOP-RECLPS2H-- (p. F-291 of the ER), operator fails to recover low-pressure systems, the discussion indicates that low-pressure systems would not have the power to function. Consider the potential for utilizing a fire truck or other portable self-powered pumps for injection into the containment.

RESPONSE

SAMAs 1 and 8 are proposed to address this contributor in the ER. SAMA 1, the reliable hardened pipe containment vent, is a commitment that will be implemented for reasons not related to the SAMA analysis. SAMA 8, the portable 480V AC generator, was determined to be potentially cost beneficial in the ER.

The combination of these SAMAs would help prevent core damage by providing a means of containment heat removal/pressure control via the reliable hardened pipe containment vent, and a means of maintaining RPV injection using the existing fire protection system connection by providing long term support for the safety relief valves (SRVs) so that they do not reclose and cause RPV re-pressurization.

The addition of a portable self-powered injection pump without SAMA 8 for these contributors would limit the benefit of such a pump to preventing drywell failure from core-concrete interactions post core damage. The benefit of this approach would be less than that of SAMA 8, there would be no apparent reduction in the implementation cost, and it would serve the less desirable function of mitigating a core damage event rather than preventing core damage.

The use of a portable self-powered pump is not considered to be the optimal choice for mitigating the risk associated with event 2TDOP-RECLPS2H--.

51 of 75

RA15-032 ENCLOSURE QUESTION 5.b.ii Event 2GVPHCMBSTGASF-- (pp. F-286, F-300, and F-316), Combustible gas venting fails, appears in all three Level 2 importance lists. Discuss this event and if any SAMAs are possible to directly mitigate this failure.

RESPONSE

The event 2GVPHCMBSTGASF-- represents the probability that containment venting to remove combustible gases does not occur. The failure of the venting function, however, does not necessarily imply containment failure. The containment failure probabilities are determined by separate containment event tree nodes that take into account the status of the combustible gas venting node.

The reasons for combustible gas venting failure vary, depending on the scenario/accident class (containment event tree) in which it is included. For example, in ATWS scenarios (Class IV),

the venting pathway is not capable of keeping up with combustible gas generation even if it is open, and it is always considered to be failed. For SBO scenarios (Class IBE/IBL), the support systems required to open the vent pathway are not available and combustible gas venting is always failed. Similarly, for loss of decay heat removal (DHR) scenarios, support system failures dominate the causes of combustible gas venting failures (combustible gas venting always fails when the support systems are not available).

The SAMAs proposed to address this event in the ER are considered to directly mitigate these failures. The reliable hardened pipe containment vent (SAMA 1) can be operated without support systems, so it precludes the failure modes that dominate the SBO and loss of DHR scenarios. For ATWS events, the ATWS sized hardened pipe vent (SAMA 17) precludes the dominant failure mode by providing a vent path of adequate size.

52 of 75

RA15-032 ENCLOSURE QUESTION 5.c The discussion in Section F.5.1.3.6 of Grand Gulf SAMA 59 states that Rather than cycling large pumps in scenarios where the cooling system is lost, a more effective means of maintaining injection with the Emergency Core Cooling System (ECCS) pumps is considered to be through the use of portable/temporary cooling alignment, which is addressed in the LSCS importance list review by SAMA 16." Consider a SAMA similar to Grand Gulf's SAMA 59, Increase operator training for alternating operation of the low pressure emergency core cooling system pumps (low-pressure coolant injection and low pressure core spray) for loss of standby service water scenarios, for rooms where the use of portable fans may not be effective.

RESPONSE

For the cases in which the portable fans would not provide benefit (the scenarios in which the fans in the ECCS rooms fail when there is cooling flow), the potential risk reduction associated with a procedure change to direct the cycling of ECCS pumps is very limited due to both planned plant changes and human reliability analysis factors. The following is a summary of these issues for one of the room cooling fan failure events (2VYFNNWVY01--X--) for which the pump cycling procedure change is suggested:

31.5% of the scenarios are related to failure to recover instrument air to support containment venting. Implementation of SAMA 1 (i.e., installation of a reliable hardened pipe containment vent), which is already a planned plant change, removes the containment vent dependence on instrument air. Thus, these cases will no longer be significant contributors.

A separate 26.5% of the scenarios are related to failure to align the diesel fire pump for injection given a harsh environment caused by vent path rupture in the turbine steam tunnel (0.5). SAMA 1 will essentially eliminate these vent path failures, and the reliability of the human failure event (HFE) to align the diesel fire pump for injection would be significantly improved.

A separate 23.7% of the scenarios are related to cases in which the operators fail to initiate containment venting. If the operators have failed to perform the relatively simple task to vent containment, it may be difficult to justify credit for the more complex task of cycling pumps that are located in a reactor building with an adverse environment caused by containment failure.

A separate 6.1% of the contributors are scenarios in which offsite power failures fail the instrument air system (for venting support), which will be non-contributors once SAMA 1 has been implemented.

About 40% of the remaining contributors are ATWS scenarios for which crediting ECCS pump cycling would be extremely difficult, leaving about 7.3% of the fan failure scenarios as potential candidates for improvement. This correlates to about a $20,000 averted cost risk when the 95th percentile results are considered.

Fan failure scenarios for other divisions are similar in nature to the event described above.

Other failures, such as loss of cooling water pumps to the room cooling, would also be similar in nature to the event described above.

53 of 75

RA15-032 ENCLOSURE The other contributors to loss of ECCS room cooling (i.e. loss of cooling flow to the ECCS room coolers due to unavailability of the pump) for which SAMA 16 would not be effective, would also mostly be mitigated by SAMA 1, and therefore, the residual benefit of a SAMA to cycle large ECCS pumps would be small.

In summary, once the commitment to install the reliable hardened pipe containment vent has been satisfied, the potential benefit of a procedure change to cycle the ECCS pumps for the fan failure scenarios is small, and it is not considered to address a dominant risk contributor for LSCS.

54 of 75

RA15-032 ENCLOSURE QUESTION 5.d While the Commonwealth Edison Individual Plant Examination (IPE) submittal did not identify any vulnerability, as discussed in the NRCs IPE Safety Evaluation Report, the IPE did cite NUREG/CR4832, Vol. 3, Part 1, for insights about potential improvements. These are: (i)

Eliminate the sneak circuit in the reactor core isolation cooling (RCIC) isolation logic that results in the RCIC steam line inboard isolation valve closing when offsite alternating current (AC) power is lost and the appropriate diesel generator starts; (ii) Change the RCIC room temperature isolation logic so that, in cases where AC power from train A has failed but AC power from train B is available, this isolation logic does not isolate if no other emergency core cooling system is working; and (iii) Change the venting procedure so that venting does not result in severe environments in the reactor building. Discuss the current status of these potential improvements.

RESPONSE

Potential Improvement 1: Subsequent review of the RCIC sneak circuit issue resulted in the conclusion that the postulated RCIC isolation scenario would not occur for the LSCS configuration. However, modifications were performed to install a replacement relay (relay K11B) in the RCIC isolation circuitry with a time delay to energize setting of 1 second, such that this postulated sequence of events would definitively be precluded. This change has been implemented.

Potential Improvement 2: The current design of RCIC is such that the leak detection system will not isolate RCIC upon loss of AC power.

Potential Improvement 3: Design improvements were considered and dispositioned at the time of the IPE. No design changes were made to install a hardened piped vent and prevent the rupture of the containment vent piping. Rather, procedure changes were made to acknowledge the design deficiency and initiate alignment of alternate injection systems prior to containment venting, if at all possible. Changes were also made to procedures to control containment pressure within a specific range during the venting evolution. The potential for containment vent path rupture/failure will be addressed through the installation of the reliable hardened pipe containment vent.

55 of 75

RA15-032 ENCLOSURE QUESTION 5.e The RMIEP Summary (NUREG/CR-4832, Volume 1) identifies the common cause failures of the core standby cooling system (CSCS) cooling water pumps as the dominant events in the seismic risk reduction importance assessment. Discuss these events and their importance in the seismic analysis used for SAMA identification.

RESPONSE

There are differences between the seismic results descriptions in Volumes 1 and 8 of NUREG/CR-4832 which suggest that they may reflect different iterations of the seismic model.

For example:

Volume 1 is dated July, 1992 while Volume 8 is dated November, 1993.

In Section 4.5.1 of Volume 1, the third paragraph states that the only explicitly seismic events appearing in the final cutsets are the seismic initiating event frequencies for each level and the seismically induced loss of offsite power (LOOP) conditional probabilities. In Volume 8, Section 11.1 indicates that offsite power is assumed to be lost, and that the seismically induced failure of the condensate storage tank (CST) is a contributor to the results.

Section 4.5.2 of Volume 1 indicates that non-seismic common cause failure (CCF) of the diesel generator CSCS cooling water pumps is in the dominant cutsets for each of the seven seismic levels. The results section of the executive summary in Volume 8 identifies the dominant component failures (diesel generator failure to start, run), but the CCF of the CSCS cooling water pumps is not mentioned.

Subsequent to the completion of the RMIEP analysis, Commonwealth Edison performed a review of that study and largely used the results as the basis for the combined IPE and IPEEE submittal. It should be noted that one of the conclusions of the review effort, which is documented in the executive summary of the April 1994 IPE/IPEEE, is that Commonwealth Edison considered the Beta factor common cause failure process used in the RMEIP analysis to be overly conservative.

Based on this information and the other descriptions of results contained in Volume 8 of NUREG/CR-4832, it does not appear that the CCFs of the CSCS cooling water pumps are driving the results of seismic risk at LSCS, and no SAMAs were generated to directly address these specific failures. However, the consequences of the CCF of the CSCS cooling water pumps is essentially a long term station blackout, and these scenarios are addressed by SAMAs evaluated in the LSCS analysis. SAMA 27 mitigates cases in which RCIC remains available (not failed by seismically induced CST failure), and SAMA 26 reduces the risk associated with scenarios in which RCIC is not available. These SAMAs are considered to adequately address any potential contributions from CCF of the CSCS cooling water pumps.

In addition, review of the LOOP and dual unit LOOP contributors from the LSCS 2013A PRA model (used for the SAMA analysis) indicates that common cause failures of the CSCS pumps are low contributors to those scenarios. For single unit LOOP, the largest Fussell-Vesely value of any CSCS pump CCF combination is less than 5.0E-04. For Dual Unit LOOP events, the largest Fussell-Vesely value of any CSCS pump CCF combination is less than 6.0E-03. If, as 56 of 75

RA15-032 ENCLOSURE stated in the fourth paragraph of Section 4.5.1 of NUREG/CR-4832 Volume 1, the dominant seismic sequences are all seismically induced losses of offsite power and look exactly like the equivalent internally initiated sequences except that no credit is given for recovering offsite power, then the conclusion would be that CCF of the CSCS cooling water pumps are not significant contributors to seismic risk.

57 of 75

RA15-032 ENCLOSURE QUESTION 5.f.i As discussed in Section F.5.1.6.1 of the ER, the lower cutoff value of 5 percent of the total fire CDF is used for considering SAMAs. Using this cutoff value results in potentially not identifying cost-beneficial SAMAs below a cost of $142,000. Consider additional fire sequences below this threshold in order to identify potential low cost procedure enhancement SAMAs.

RESPONSE

As discussed in Section F.5.1.1 of the ER, the SAMA Guidance Document (NEI 05-01), which has been endorsed by the NRC, describes the SAMA identification process as one which should identify plant-specific SAMA candidates by reviewing dominant risk contributors (to both CDF and population dose) in the Level 1 and Level 2 Probabilistic Safety Assessment (PSA) models. While it may be possible to identify potentially cost beneficial procedure changes for fire zones contributing 2.5% of the total CDF or less, such an effort is not consistent with the intent of the SAMA process. Further, the LSCS fire PRA is a preliminary or interim model. It is considered possible to obtain general insights from a model in this state, but because multiple model development tasks are still in progress to better meet the ASME/ANS PRA Standard, as identified in Section F.5.1.6.1 of the ER, the benefit of developing SAMAs that would yield changes to the CDF that are less than 3.0E-07 per year is questionable.

Expanding the fire related SAMA identification process to identify procedure changes that may be cost beneficial requires the cost of such a procedure change to be established to bound the review. The Oyster Creek SAMA submittal (AMERGEN 2005) approximates the cost of a procedure change to be $50,000 for a single unit. For a dual unit site, some cost sharing is likely, but because fire PRAs are unit specific and the cable routing is different for the two units, the Unit 1 and Unit 2 procedures would likely not be the same. For this analysis, it is assumed the second units procedure can be developed at half the cost of the first unit such that the site cost would be $75,000 ($50,000 + 25,000 = $75,000). This translates to a cost of $37,500 per unit.

The following tables summarize the un-reviewed fire zones at LSCS with potential averted cost-risk (PACR) values larger than $37,500:

Unit 1 Fire Zone CDF PACR 2F-2 3.36E-07 $141,586.62 12 3.11E-07 $131,051.90 4D1-1 2.77E-07 $116,724.69 4E1-1 2.40E-07 $101,133.30 4E3-1 2.06E-07 $86,806.08 2I3 1.50E-07 $63,208.31 58 of 75

RA15-032 ENCLOSURE Unit 2 Fire Zone CDF PACR 13 3.20E-07 $134,844.40 4E4-1 1.98E-07 $83,434.97 3I3 1.49E-07 $62,786.93 5C11 1.33E-07 $56,044.71 3I5 1.26E-07 $53,094.98 4F1 1.24E-07 $52,252.21 3G-2 1.20E-07 $50,566.65 3I4 1.05E-07 $44,245.82 4E2-1 1.01E-07 $42,560.26 The contributors to the fire risk for each of these zones have been reviewed to determine if there are any procedure changes that could be potentially cost beneficial. The results of this review are presented below on a zone-by-zone basis.

Unit 1 U1 Zone 2F-2: UNIT 1 - ELEVATION 740'0" - NORTH SIDE RB About 98% of the zone 2F-2 contribution is from the D scenario, which is a fire initiated in motor control center (MCC) 1AP62E (and 1AP75E) that leads to multiple fire induced failures, including RCIC, ADS, RHR train A, and the containment vent paths. The core damage scenarios include random failure of the remaining RHR train and in over 90% of the cases, failure of the remaining ECCS injection pumps occurs either due to harsh environmental conditions or from the consequences of a large containment failure. Implementation of the reliable hardened pipe containment vent (SAMA 1) will mitigate these scenarios by providing a heat removal/containment vent path.

For the remaining 10% of the contributors, various random failures such as suction strainer clogging or HPCS/support system failures result in the loss of high pressure core spray (HPCS) injection. No procedure changes have been identified that would address these scenarios, although some previously identified SAMAs would be relevant for a small set of scenarios, such as SAMA 16 for alternate core standby cooling system room cooling.

U1 Zone 12: UNIT 1 TRANSFORMERS These scenarios are related to fires that are initiated in, and fail, the Unit Auxiliary Transformer, the Main Power Transformer, and the System Auxiliary Transformer. The dominant contributors are containment failures leading to loss of injection. In some cases (about 20%), operator error leads to failure of suppression pool cooling initiation and containment venting. These cases are addressed by SAMA 2 (auto SPC) or SAMA 3 (passive vent path).

Most other scenarios are related to loss of support systems for the containment vent and random failures of the RHR system (or its support systems). The reliable hardened pipe containment vent will provide a means of removing heat in these scenarios that will not cause a harsh reactor building environment and allow for continued ECCS injection, or injection with the fire protection system.

59 of 75

RA15-032 ENCLOSURE No potentially cost beneficial procedure changes have been identified for this fire zone.

U1 Zone 4D1-1: UNIT 1 - CABLE SPREADING ROOM - DIV. I CABLE RISER ROOM These scenarios are related to transient fires from hot work that lead to multiple fire induced failures, including RCIC, RHR train A, and the containment vent paths. The core damage scenarios and potential SAMAs are similar to those described for Unit 1 fire zone 2F-2.

The potential transient ignition sources are limited in this zone and the fire analysis methodology already estimates the risk from these types of sources as the lowest possible level for fire zones that are accessible during plant operation (no additional reductions are possible through procedure changes). As described in the response to RAI 5.f.iii for the Auxiliary Electrical Equipment Room, plant procedures are currently in place to require a fire watch to be posted with portable fire extinguishers when hot work is being performed.

No potentially cost beneficial procedure changes have been identified for this fire zone.

U1 Zone 4E1-1: UNIT 1 - AUXILIARY EQUIPMENT ROOM - SOUTH AER ROOM Over 85% of the risk for this fire zone is related to fires initiated in panels that support RHR B and the containment vent paths. The core damage scenarios and potential SAMAs are similar to those described for Unit 1 fire zone 2F-2, but for this fire zone, the random failures are associated with train A of RHR.

There are some additional contributors for this fire zone associated with HPCS failures and vacuum breaker failures that would not be addressed by SAMA 1 alone. For the failures related to loss of HPCS, an independent injection system capable of providing makeup early in the scenario would be required. These cases would be addressed by SAMA 18 (Improve the Connection Between the Fire Protection and Feedwater Systems). The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

No potentially cost beneficial procedure changes have been identified for this fire zone.

U1 Zone 4E3-1: UNIT 1 - DIVISION 2 ESSENTIAL SWITCHGEAR ROOM - DIV. I CABLE RISER ROOM These scenarios are related to transient fires that lead to multiple fire induced failures, including RCIC, RHR train A, and the containment vent paths. The core damage scenarios and potential SAMAs are similar to those described for Unit 1 fire zone 2F-2.

The potential transient ignition sources are limited in this zone and the fire analysis methodology already estimates the risk from these types of sources as the lowest possible level for fire zones that are accessible during plant operation (no additional reductions are possible through procedure changes). As described in the response to RAI 5.f.iii for the Auxiliary Electrical Equipment Room, plant procedures are currently in place to require a fire watch to be posted with portable fire extinguishers when hot work is being performed.

No potentially cost beneficial procedure changes have been identified for this fire zone.

60 of 75

RA15-032 ENCLOSURE U1 Zone 2I3: UNIT 1 - RHR PUMP B&C CUBICLE These scenarios are related to fires that initiate in the B or C RHR pump and result in a full zone burn-up. The fire induced failures include RHR B, C, and RCIC. About half of the core damage scenarios are related to pressure suppression bypass events caused by vacuum breaker failures. The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

The remaining core damage scenarios include random failures that lead to loss of containment heat removal capability, generally failures of RHR train A (or its support systems) and a loss of offsite power (which fails the containment vent paths). Containment failure leads to harsh environmental conditions and subsequent injection system failures. In some scenarios, containment venting is successful, but the vent path ruptures and leads to a harsh reactor building environment that fails RPV makeup systems. Implementation of the reliable hardened pipe containment vent (SAMA 1) will mitigate these scenarios by providing a viable heat removal/containment vent path.

No potentially cost beneficial procedure changes have been identified for this fire zone.

Unit 2 U2 Zone 13: UNIT 2 TRANSFORMERS These scenarios are related to fires that are initiated in, and fail, the Unit Auxiliary Transformer, the Main Power Transformer, and the System Auxiliary Transformer. The dominant contributors are containment failures leading to loss of injection. In some cases (about 20%), operator error leads to failure of suppression pool cooling initiation and containment venting. These cases are addressed by SAMA 2 (auto SPC) or SAMA 3 (passive vent path).

Most other scenarios are related to loss of support systems for the containment vent and random failures of the RHR system (or its support systems). The reliable hardened pipe containment vent will provide a means of removing heat in these scenarios that will not cause a harsh reactor building environment and allow for continued ECCS injection, or injection with the fire protection system.

No potentially cost beneficial procedure changes have been identified for this fire zone.

U2 Zone 4E4-1: UNIT 2 - DIVISION 2 ESSENTIAL SWITCHGEAR ROOM - DIV. I CABLE RISER ROOM These scenarios are related to transient fires that lead to multiple fire induced failures, including RCIC, RHR train A, and the containment vent paths. The core damage scenarios and potential SAMAs are similar to those described for Unit 1 fire zone 2F-2.

The potential transient ignition sources are limited in this zone and the fire analysis methodology already estimates the risk from these types of sources as the lowest possible level for fire zones that are accessible during plant operation (no additional reductions are possible through procedure changes). As described in the response to RAI 5.f.iii for the Auxiliary Electrical 61 of 75

RA15-032 ENCLOSURE Equipment Room, plant procedures are currently in place to require a fire watch to be posted with portable fire extinguishers when hot work is being performed.

No potentially cost beneficial procedure changes have been identified for this fire zone.

U2 Zone 3I3: UNIT 2 - RHR PUMP B&C CUBICLE These scenarios are related to fires that initiate in the B or C RHR pump and result in a full zone burn-up. The fire induced failures include RHR B, C, fire protection, and RCIC. About half of the core damage scenarios are related to pressure suppression bypass events caused by vacuum breaker failures. The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

The remaining core damage scenarios include random failures that lead to loss of containment heat removal capability, generally failures of RHR train A (or its support systems) and a loss of offsite power (which fails the containment vent paths). Containment failure leads to harsh environmental conditions and subsequent injection system failures. In some scenarios, containment venting is successful, but the vent path ruptures and leads to a harsh reactor building environment that fails RPV makeup systems. Implementation of the reliable hardened pipe containment vent (SAMA 1) will mitigate these scenarios by providing a viable heat removal/containment vent path.

No potentially cost beneficial procedure changes have been identified for this fire zone.

U2 Zone 5C11: TURBINE BUILDING GROUND FLOOR GENERAL AREA One of the larger scenarios in the turbine building general area is a fire initiated in equipment (mostly electrical cabinets and transformers) that does not impact any appendix R equipment.

In these cases, the dominant contributors to risk include operator errors to initiate suppression pool cooling and containment venting. These cases are addressed by SAMA 2 (auto SPC) or SAMA 3 (passive vent path).

The other top contributors (scenarios F, F1, and J) are fires that lead to the loss of the System Auxiliary Transformer, 0 diesel generator, RCIC, and RHR A. In about half of these contributors, there is a fire induced medium LOCA, but in most cases, the LOCA is initially mitigated (HPCS, LPCS and RHR C are initially available) and core damage ultimately results due to the consequences of containment failure (from failures that eliminate containment heat removal capability). These scenarios are similar to those described for Unit 1 fire zone 2F-2, and the reliable hardened pipe containment vent will provide a means of preventing containment failure such that the risk from these contributors will be significantly reduced.

No potentially cost beneficial procedure changes have been identified for this fire zone.

U2 Zone 3I5: UNIT 2 - RHR PUMP A CUBICLE These scenarios are related to fires that initiate in the A RHR pump room and result in a full zone burn-up. The fire induced failures include RHR A, fire protection, HPCS, CRD, and RCIC. About half of the core damage scenarios are related to pressure suppression bypass events caused by vacuum breaker failures. The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

62 of 75

RA15-032 ENCLOSURE The remaining core damage scenarios include scenarios where venting is performed successfully, but failure of the vent path ducts results is harsh conditions that lead to ECCS pump failures (about 25% of the total fire zone CDF). These failures would be mitigated by the availability of the reliable hardened pipe containment vent (SAMA1).

Of those scenarios not addressed by SAMAs 1 and 20, about half are related to the failure to close the turbine driven pump discharge valves to prevent hotwell depletion/RPV overfill. These failures are addressed by SAMA 10 (Change the Logic to Close the Turbine Driven Feedwater Pump Discharge Valves When the Pumps are Not Running).

No potentially cost beneficial procedure changes have been identified for this fire zone.

U2 Zone 4F1: UNIT 1 - DIVISION 1 ESSENTIAL SWITCHGEAR ROOM These scenarios are primarily related to fires that initiate in the Unit 1 4.16KV 141Y switchgear or the Unit 1 6.9KV 151 switchgear. Fire induced failures include switchgear 241-Y, 0 diesel generator, RHR A, RCIC, and the containment vent paths. The core damage scenarios and potential SAMAs are similar to those described for Unit 1 fire zone 2F-2.

No potentially cost beneficial procedure changes have been identified for this fire zone.

U2 Zone 3G-2: UNIT 2 - ELEVATION 710'6" - NORTH SIDE RB These fires originate in 480V bus 235Y and result in the loss of 0 diesel generator, RHR A, the suppression pool vent path, CRD, Feedwater, fire protection system, and LPCS.

About 30% of the core damage scenarios are related to pressure suppression bypass events caused by vacuum breaker failures. The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

In about 13% of the scenarios, containment venting is performed successfully through the drywell vent, but failure of the vent path ducts results is harsh conditions that lead to ECCS pump failures (loss of injection). These failures would be mitigated by the availability of the reliable hardened pipe containment vent (SAMA1).

In about 40% of the scenarios, the drywell containment vent path fails due to support system failures. These failures would also be mitigated by the availability of the reliable hardened pipe containment vent (SAMA1).

For the remaining 17% of the contributors, various random failures, such as suction strainer clogging or HPCS/support system failures, result in the loss of RPV injection. No procedure changes have been identified that would address these scenarios, although some previously identified SAMAs would be relevant for a small set of scenarios, such as SAMA 16 for alternate core standby cooling system room cooling.

U2 Zone 3I4: UNIT 2 - LPCS/RCIC PUMP CUBICLE These scenarios are based on the full fire zone ignition frequency and are assumed to result in a full room burn up. The fire induced failures include RCIC, fire protection, and LPCS. About 63 of 75

RA15-032 ENCLOSURE 70% of the core damage scenarios are related to pressure suppression bypass events caused by vacuum breaker failures. The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

About 7.5% of the contribution is related to operator failures to either vent and/or initiate SPC.

These cases are addressed by SAMA 2 (auto SPC) or SAMA 3 (passive vent path).

The remaining core damage scenarios include different evolutions that lead to loss of RPV makeup, such as cases where venting is performed successfully, but failure of the vent path ducts results is harsh conditions that lead to ECCS pump failures, failures of containment vent support systems that lead to loss of RPV injection after containment failure, and suction strainer clogging. No procedure changes have been identified that would address these scenarios, although some previously identified SAMAs would be relevant for a small set of scenarios, such as SAMA 6 for creating an ECCS suction backwash capability with RHRSW.

U2 Zone 4E2-1: UNIT 2 - AUXILIARY EQUIPMENT ROOM - WEST AER ROOM Over 77% of the risk for this fire zone is related to electrical panel and cable fires that support RHR A, RCIC, and 0 diesel generator (and a subset of these fires also fail RHR B).

About 22.5% of the core damage scenarios are related to pressure suppression bypass events caused by vacuum breaker failures. The vacuum breaker failures are addressed by SAMA 20 (Improve Vacuum Breaker Reliability by Installing Redundant Valves in Each Line).

Loss of containment vent support systems account for at least an additional 44% of the fire zone risk, which would be mitigated by SAMA 1.

No potentially cost beneficial procedure changes have been identified for this fire zone.

In summary, the LSCS fire zones for which it was considered possible to identify potentially cost beneficial procedure changes were reviewed; however, no such procedure changes were identified as part of this effort.

REFERENCES AMERGEN 2005 AmerGen (AmerGen Energy Company, LLC). 2005. License Renewal Application - Oyster Creek Generating Station, Appendix E Environmental Report, Appendix F Severe Accident Mitigation Alternatives. July.

64 of 75

RA15-032 ENCLOSURE QUESTION 5.f.ii As indicated in Section F.5.1.6.1.1 of the ER, fires in the Division 1 and Division 2 essential switchgear rooms for each unit make up 60% of the total fire CDF. The only SAMA proposed for mitigating these fires is SAMA 1, installing a reliable hardened pipe vent. SAMA 1 only mitigates the adverse consequences of venting and does not mitigate the direct impact of the fire. Discuss these fire scenarios and the potential for other SAMAs to directly mitigate the fire at an earlier stage of the scenario or by mitigating events in the cutsets other than the adverse conditions due to venting.

RESPONSE

For fires in the Division 1 and 2 essential switchgear rooms, the main contributors are either cases in which injection is initially available, but subsequently fails due to harsh environmental conditions after containment failure, or are cases in which a combination of fire induced and random failures lead to loss of injection. A more detailed review of the cutsets identifies that about 79% of the contributors are related to the first of these types of scenarios. The unavailability of the containment vent is due to fire induced failures associated with vent valve support systems, with the vent valve control cables, or both. The reliable hardened pipe vent (SAMA 1) eliminates support system requirements and would provide a means of preventing the containment failures and vent path failures that lead to loss of injection capability.

In about 21% of the contributors, hardware failures result in the unavailability of injection systems such that even if a containment vent path were available, core damage would still occur. The hardware failures are comprised of both failures of the 4kv AC power sources that support the injection systems and the failures of the injection system components themselves.

In order to mitigate these scenarios, it would be necessary to provide an independent RPV makeup source that does not rely on existing electrical support systems and that would be available in fire scenarios. LSCS already has the B.5.b pump, which can provide this RPV injection capability for the long term scenarios that are characteristic of these contributors, but it is not currently credited in the PRA model. LSCS operators have been trained to use the B.5.b pump and the current emergency operating procedures include the B.5.b pump as a potential alternate RPV injection source that could be used to mitigate the Division 1 and Division 2 essential switchgear room fire scenarios. There are also procedures with specific instructions for the setup and use of the B.5.p pump. If the B.5.b pump were credited, the risk associated with these scenarios would be reduced and the potential to identify additional cost beneficial enhancements would be very low.

In addition, FLEX is intended to provide a separate set of portable makeup pumps that could also be used to perform this function. While the details of the FLEX changes are not finalized, they would further improve the capability of LSCS to respond in these types of scenarios.

65 of 75

RA15-032 ENCLOSURE QUESTION 5.f.iii Section F.5.1.6.1.3 indicates that the largest contributing fire scenario to the Auxiliary Electrical Equipment Room fire risk is a bounding cable fire caused by hot work. Discuss the potential for a SAMA to mitigate this risk.

RESPONSE

The transient fire ignition frequency calculated for a fire compartment considers the potential for ignition due to hot work (cutting, grinding, or welding tasks), among other sources. Based on the 2009 LSCS Fire PRA, unless the area is inaccessible during plant operation such that this type of activity would be positively precluded, the fire compartment is assigned an influence factor that characterizes the frequency of hot work activities that range from low (work performed on an annual basis) to very high (performed daily). The hot work influence factor for the Auxiliary Electrical Equipment Room is low, which is the lowest possible factor for this area.

Review of the LSCS Fire PRA indicates that procedures are already in place that require a fire watch to be posted (with portable extinguishes available) and for nearby equipment to be protected when hot work is being performed. These practices are already accounted for in the fire risk evaluation for the Auxiliary Electrical Equipment Room. Based on these considerations, no potential SAMAs related to hot work limitations are available that would yield a measurable change in Auxiliary Electrical Equipment Room fire risk.

Because the hot work is associated with a bounding transient fire scenario that is assumed to lead to failure of all equipment in the fire compartment, no details are available that would help generate SAMAs to mitigate the fire beyond those associated with the initiating event frequency.

66 of 75

RA15-032 ENCLOSURE QUESTION 5.f.iv The RMIEP Summary (NUREG/CR-4832, Volume 1) identifies the three most important events in the fire analysis risk reduction importance assessment to be in the control room abandonment scenario. Discuss this scenario and its importance in the current fire analysis and the potential for a cost-beneficial SAMA to mitigate this scenario.

RESPONSE

In Volume 9 of NUREG/CR-4832, which includes the details of the fire analysis, main control room (MCR) fire scenarios are described as follows: a fire starts in some cabinet, the fire is not suppressed before smoke requires abandonment of the main control room, the fire damages sufficient equipment to require manual operation, and the operators fail to recover injection and containment heat removal from the remote shutdown panel. The core damage frequency is calculated as the product of the control room fire ignition frequency, the probability that the fire is not suppressed before smoke requires control room abandonment, and the probability that the operators fail to recover the plant from the remote shutdown panel. It appears that the probability that the fire damages sufficient equipment to require manual operation is assumed to be 1.0.

The LSCS Fire PRA approach is similar, but NUREG-6850 methodology uses different ignition frequencies and non-suppression probability calculations. In addition, the failure to control the plant from the remote shutdown panel is characterized in the Fire PRA as a conditional core damage probability given that the MCR must be abandoned rather than the more narrow definition of operator error alone. The difference in the CDF estimates is mostly tied to the non-suppression probabilities. The Fire PRA ignition frequencies and the conditional probability to fail to reach safe shutdown at the remote shutdown panel are larger than their NUREG/CR-4832 counterparts, but the NUREG-6850 based non-suppression/severity factor probability is about 3.0E-03 while it is 1.0E-01 in NUREG/CR-4832.

The Fire PRA CDF contribution from MCR abandonment scenarios is only about 1.5E-07/year, which correlates to a cost-risk of less than $65,000. This suggests a very limited potential for the identification of cost beneficial SAMAs, and it essentially rules out hardware modifications.

Given that the ignition frequency is dominated by the main control board and electrical panel fires, which are in turn driven by the number of panels required to house the control cables that are required to be in the MCR, there are no procedure changes that would impact the fire ignition frequency. For fire suppression reliability, the MCR is constantly manned and the operators are trained on fire suppression. Changes to LSCS procedures would not impact the non-suppression probability. Finally, the conditional core damage probability is not based on any specific characteristics of LSCS that could be improved to further reduce the CDF.

In summary, while the RMIEP analysis identified that MCR abandonment scenarios were large contributors to risk, the current LSCS Fire PRA indicates that the risk associated with MCR abandonment scenarios is low, and no low cost changes have been identified that would have a measurable impact on the core damage frequency calculated for these scenarios at LSCS.

67 of 75

RA15-032 ENCLOSURE QUESTION 6.a Clarify if maintenance costs are included in the estimated implementation costs.

RESPONSE

The costs of testing and maintaining the SAMA equipment during the license renewal period are not included in the implementation cost estimates.

68 of 75

RA15-032 ENCLOSURE QUESTION 6.b Discussions of SAMA 8 in Section F.6.8 of the ER and SAMA 14 in Section F.6.13 of the ER include the statement Flow from the fire protection system, in its current configuration, is only adequate in cases where RCIC.... Discuss what is meant by current configuration and the potential for a SAMA to address the fire protection system configuration and make its use possible without prior RCIC operation.

RESPONSE

There are at least two major issues that preclude the use of the fire protection system (FPS) for RPV makeup in early time frames to prevent core damage. The first is related to the low flow rate of the makeup path, and the second is related to the relatively long time that is required to align the FPS for injection.

The connection that is currently used to provide RPV makeup from the FPS consists of fire hoses that are manually aligned between the FPS header and the Feedwater injection lines.

Analysis of this flow path indicates that RPV pressure must be reduced below 75 psig for Unit 1 and 60 psig for Unit 2 in order to achieve a flow rate of 200 gpm into the RPV. LSCS thermal hydraulic analysis (MAAP analysis) indicates that this alignment would be successful after 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> following initial high pressure injection and subsequent RPV depressurization. For SBO scenarios, this would correspond to cases with successful RCIC injection. Prior to 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, decay heat levels would be such that 200 gpm would be insufficient for RPV make-up.

Operator interviews indicate that about 40 minutes are required to travel to the FPS header location, obtain and connect the fittings to the fire hoses that are required for the connection, manually connect the fire hoses to the FPS and Feedwater piping, and open the valves required to establish the flow path. From the time when it has been determined that RCIC has failed and the FPS would be needed for RPV makeup (the cue time), it would not be possible to align this injection path even if adequate flow could be obtained from the FPS to make up for boil-off.

SAMA 18, Improvement of the Connection Between the Fire Protection and Feedwater Systems, was developed to reduce the alignment time such that it would be possible to perform the alignment in cases where RCIC injection fails. One of the assumptions made in the SAMA 18 assessment was that the improved connection would also improve the flow rate of FPS injection to a point where it could provide adequate makeup in loss of all injection scenarios.

This may be an optimistic assumption given that SAMA 18 does not reduce the elevation differences between the lower lying FPS pumps and the Unit 1 and Unit 2 RPVs, which imposes a constant pressure head that will continue to impact the RPV injection flow rate.

69 of 75

RA15-032 ENCLOSURE QUESTION 6.c Discussion of SAMA 14 (Provide a portable DC source to support RCIC and SRV operation) in Section F.6.13 of the ER indicates that the PRA model was changed to include a lumped event to represent the 480V AC power source that feeds the Division 1 battery chargers. The SAMA description and the basic events cited to be mitigated by this SAMA indicated that a direct current (DC) power source to directly supply an ESF DC distribution panel is needed. Discuss the inclusion of an AC power supply in the model.

RESPONSE

The description of the PRA model changes related to the lumped event that was added to represent the 480V AC power source that feeds the Division 1 battery charger is erroneous.

The text should read a lumped event was added to represent the 125V DC generator that would supply 125V ESF DC distribution panel 1(2)11Y.

The detailed PRA model change descriptions documented in Section F.6.13 accurately describe how the lumped event (basic event ID SAMA14) was incorporated into the PRA model for SAMA 14. For example, the third bullet states that basic event SAMA14 was included under an AND gate (SAMA14-AC) with existing gate 2DC08E-PWR-AC (FAULTS AFFECTING POWER FROM DC BUS 2DC11E), which allows event SAMA14 to mitigate, among other things, failures of 125V DC bus 2A, battery charger failures, and failures of the 480V AC power supply to the battery chargers.

Note that the PRA nomenclature may introduce confusion on this issue as well since the PRA basic events include the designator AC in the name. However, based on examination of the PRA model, these basic events are related to the DC bus and DC distribution of Bus 1(2)11Y.

70 of 75

RA15-032 ENCLOSURE QUESTION 6.d SAMA 7 (Water hammer prevention) and SAMA 25 (Periodic training on water hammer scenarios resulting from a false LOCA signal) both are intended to mitigate the water hammer scenarios involving SPC operation interrupted by a LOOP. The changes made to the model and the impact of the changes on person-rem and offsite economic cost risk are significantly different for the two SAMAs. Discuss these SAMAs and their analyses in more detail to justify these differences.

RESPONSE

SAMA 7 and SAMA 25 are different approaches to addressing the risk associated with water hammer events at LSCS. SAMA 7 is a change to the LOCA signal generation logic that is intended to preclude the onset of the conditions that can lead to water hammer events in scenarios in which SPC is placed into service after the initiating event and a consequential loss of offsite power occurs after a LOCA signal is generated on high drywell pressure. The reliability of SAMA 7 is tied to the hardware associated with the revised LOCA signal logic, and because the failure rate for the logic would be low, it was assumed that SAMA 7 completely eliminated the risk associated with these scenarios.

SAMA 25 was developed from an industry SAMA that promoted the general idea of increasing operator training on systems and operator actions determined to be important from the Probabilistic Safety Assessment. Review of the important LSCS operator actions identified that the action to perform drywell venting to prevent the generation of a high drywell pressure signal was not a highly trained action at LSCS. Implementation of SAMA 25 was assumed to make the operators highly familiar with the scenario and the steps required to prevent the high drywell pressure signal, which was reflected in the human reliability analysis evaluation of the drywell venting action through application of the lower bound accident sequence evaluation program non-response probability curve in place of the median curve. The result was a 70% reduction of the CDF for the scenarios that included this operator action (or one of the joint HFEs associated with it). If these HFEs were set to 0.0 rather than the values documented in Section F.6.23 of the ER, the CDF would match the CDF documented for SAMA 7 (2.39E-06/yr).

The larger reductions to the dose-risk and the offsite economic cost risk documented in the ER for SAMA 25 compared to SAMA 7 resulted from a recently discovered error in a supporting spreadsheet calculation. Corrected calculations for SAMA 25 have been made and are tabulated below:

CDF Dose-Risk OECR Base Value 2.58E-06 7.64 $57,655 SAMA Value 2.44E-06 7.61 $57,437 Percent Change 5.4% 0.4% 0.4%

71 of 75

RA15-032 ENCLOSURE Release Category Freq.BASE Freq.SAMA Dose-RiskBASE Dose-RiskSAMA OECRBASE OECRSAMA H/E-BOC 8.32E-08 8.32E-08 1.34E+00 1.34E+00 $7,222 $7,223 H/E 5.98E-08 5.87E-08 3.16E-01 3.10E-01 $2,787 $2,735 H/I 1.94E-08 1.94E-08 1.10E-01 1.10E-01 $974 $974 M/E 2.36E-07 2.35E-07 1.74E+00 1.74E+00 $10,360 $10,312 M/I 1.02E-06 1.02E-06 3.94E+00 3.92E+00 $36,006 $35,889 L/E 3.94E-07 3.92E-07 8.71E-02 8.66E-02 $126 $125 L/I 1.48E-07 1.46E-07 1.05E-01 1.03E-01 $181 $178 INTACT 6.20E-07 4.89E-07 1.35E-03 1.06E-03 $1 $0 Total 2.58E-06 2.44E-06 7.64E+00 7.61E+00 $57,655 $57,437 The averted cost risk corresponding to these results is provided below2:

Base Case Revised Averted Unit Cost-Risk Cost-Risk Cost-Risk LSCS Unit 2 $6,073,600 $6,031,875 $41,725 Based on a $112,000 cost of implementation for LSCS, the net value for this SAMA is -$70,275

($41,725 - $112,000), which indicates this SAMA is not cost-beneficial. When the 95th percentile PRA results are used, the averted cost-risk is increased by a factor of 2.14 to

$89,292, which still yields a negative net value ($89,292 - $112,000 = -$22,709). Therefore, this SAMA is also not cost-beneficial in the 95th percentile sensitivity analysis, which is a change from the conclusion in Section F.7.2.3 of the ER.

In summary, SAMA 7 prevents the need for operator action and is assumed to be highly reliable due to the design of system logic and instrumentation. SAMA 25 is designed to improve the reliability of operator actions; however, operator performance is still not assumed to be perfect.

The larger reduction in dose risk and offsite economic risk documented in the ER for SAMA 25 relative to SAMA 7 resulted from a recently discovered error in a supporting spreadsheet calculation. After correction of the error, the dose risk and offsite economic risk for SAMA 25 are comparable to those for SAMA 7, and the classification of SAMA 25 has been changed from potentially cost beneficial in the 95th percentile sensitivity analysis, to not cost beneficial..

2 These results are the basis for the values presented for SAMA 25 in the table titled Summary of the Impact of Using the 95th Percentile PRA Results REVISED in the response to Question 2.d.ii (see p. 22 of this Enclosure).

72 of 75

RA15-032 ENCLOSURE QUESTION 6.e A few SAMAs (e.g., 8, 14, and 27) involve use of equipment that may be available because of the B.5.b Program. Discuss this further and the impact on the cost-benefit analysis.

RESPONSE

The B.5.b program at LSCS includes a small generator that is used to support two individual SRV solenoids to hold the SRVs open after 125V DC battery depletion, but the generator does not power the station battery chargers and it is not designed to support the RCIC system through direct DC feeds. Because of these limitations, the B.5.b generator is not a viable substitute for the generators that have been proposed for SAMAs 8, 14, and 27; therefore, the availability of the B.5.b generator would not reduce the implementation costs for these SAMAs.

73 of 75

RA15-032 ENCLOSURE QUESTION 6.f SAMA 9 (Develop flood zone specific procedures) and SAMA 11 (Provide the capability to trip the FPS pumps) both address internal flooding whose principal contributor is a fire protection system (FPS) pipe rupture in the reactor building. The assumption for evaluating SAMA 9, that the risk from all internal floods will be eliminated, results in a 9 percent reduction in CDF and this SAMA being cost-beneficial. The assessment of SAMA 14 [sic] results in less than a 2 percent reduction in CDF; while as stated in Section F.6.11, SAMA 11 is designed to eliminate the FPS's flow. Discuss the FPS design and the FPS pipe break scenarios and associated modeling to support the above results.

RESPONSE

The primary reason for the significant difference between SAMA 9 and 11 risk reductions is that they target different portions of flooding risk. SAMA 9 targets all flooding initiators while SAMA 11 targets only a subset of FPS flooding scenarios.

SAMA 9 is designed to improve operator response to all significant internal flooding events by providing explicit, pre-planned steps to mitigate the flooding events. The SAMA was conservatively modeled by assuming that it eliminated all internal flooding risk. The 9%

reduction in CDF is consistent with the CDF values reported for the internal flooding initiators in Table F.2-2 of the ER.

SAMA 11 provides a means of tripping the FPS pumps in the main control room (MCR), which helps reduce the risk of FPS flooding events that require a rapid response (i.e., large breaks),

but would only have a small impact on the longer term FPS flooding scenarios. In the longer term FPS flooding scenarios, the time required to travel to the Lake Screen House to trip the FPS pumps is not driving the human error probability (HEP) associated with this action, and reducing the time required to perform the pump trip action would not significantly reduce the HEP. SAMA 11 was modeled by changing the FPS pump trip HEP for the FPS flooding scenarios requiring a rapid response to reflect the availability of MCR pump controls. The reduction in HEP from 4.1E-01 to 3.2E-02 reduced the CDF by about 1.6%, which is consistent with the 1.015 risk reduction worth value for the HEPs basic event (2FPOPMANTRIP1H--).

Since SAMA 11 only improves the reliability of one human action in the PRA model, only a small improvement in CDF is expected. SAMA 9 improves the reliability of all human response actions to all flood scenarios, and therefore, has a larger overall impact on the model.

74 of 75

RA15-032 ENCLOSURE QUESTION 7.a Consider the potential for changes to the suppression pool cooling operating procedures/practices to reduce the chance of the 2 psi high drywell signal being reached for normal transients as an alternative to SAMA 7.

RESPONSE

Changes to the LSCS operating procedures would not reduce the chance that the 2 psig high drywell pressure signal would be reached in the scenarios of interest. LSCS procedures are already structured in a way that will result in the start of suppression pool cooling (SPC) shortly after the initiating event. In fact, the water hammer scenarios are the result of RHR being in operation for the purposes of containment heat removal (i.e., when the LOCA induced LOOP occurs, the piping for the operating SPC train(s) voids). The issue is that in the scenarios of interest (initiators that result in the loss of containment cooling), SPC does not have the capacity to remove enough heat to prevent the containment pressure from exceeding 2 psig without venting containment in accordance with normal operating procedures to maintain containment pressure less than the LOCA signal initiation pressure setpoint.

Any changes to the suppression pool cooling system operating procedures intended to increase the time available between the initiating event and the 2 psig high drywell pressure signal would have a very small impact on risk. Human error probabilities (HEPs) are influenced by a number of factors, including time available for response. In this case, the HEP for the operator action to perform the 2-inch containment vent to prevent the high drywell signal is not significantly impacted by timing considerations, and increasing the time available to the operators would correlate to very small averted cost risk values.

The more difficult mitigating actions for water hammer scenarios are associated with the approximate 20 second time interval between the occurrence of the loss of the running pump on the LOCA induced LOOP and the time when the pumps are automatically started and reloaded onto the emergency bus. This short time period is not associated with the operation of suppression pool cooling; therefore a SAMA would have no benefit to these scenarios. In these scenarios, the operators must put the pump in pull-to-lock to prevent restart and then fill and vent the system to ensure a water hammer does not occur.

75 of 75