LR-N10-0198, Supplemental Information for License Amendment Request: Emergency Diesel Generators (EDG) a and B Allowed Outage Time (AOT) Extension

Supplemental Information for License Amendment Request: Emergency Diesel Generators (EDG) a and B Allowed Outage Time (AOT) Extension

ML101590514
Person / Time
Site:	Hope Creek
Issue date:	05/28/2010
From:	Jamila Perry Public Service Enterprise Group
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
LR-N10-0198
Download: ML101590514 (28)

Text

John F. Perry PSEG Nuclear LLC Hope Creek Site Vice President P. 0. Box 236, Hancocks Bridge, NJ 08038 - 0236 Tel: 856-339-3463, Fax: 856-339-1113 email: john.perry@pseg.com Nuclear LLC MAY 28 2010 10 CFR 50.90 LR-N10-0198 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, DC 20555-0001 Hope Creek Generating Station Facility Operating License No. NPF-57 NRC Docket No. 50-354

Subject:

Supplemental Information - License Amendment Request: Emergency Diesel Generators (EDG) A and B Allowed Outage Time (AOT) Extension

References:

(1) Letter from PSEG to NRC, "License Amendment Request: Emergency Diesel Generators (EDG) A and B Allowed Outage Time (AOT) Extension," dated March 29, 2010 (2) Letter from NRC to PSEG, "Hope Creek Generating Station - Supplemental Information Needed for Acceptance of Requested Licensing Action Re:

Amendment Request Regarding Emergency Diesel Generator Allowed Outage Time Extension (TAC No. ME3597)," dated May 4, 2010 In Reference 1, PSEG Nuclear LLC (PSEG) submitted a license amendment request (H 10-03) for the Hope Creek Generating Station (HCGS). The proposed change would modify TS 3/4.8.1, "AC Sources - Operating"; specifically ACTION b concerning one inoperable Emergency Diesel Generator (EDG). The proposed change would extend the Allowed Outage Time (AOT) for the 'A' and 'B' EDGs from 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> to 14 days. The proposed extended AOT is based on application of the Hope Creek Generating Station (HCGS) Probabilistic Risk Assessment (PRA) in support of a risk-informed extension, and on additional considerations and compensatory actions.

In Reference 2, the NRC requested supplemental information to complete the acceptance review of Reference 1. The PSEG response to the requested information is provided in the Attachments to this letter. No new regulatory commitments are established by this submittal.

If you have any questions or require additional information, please do not hesitate to contact Mr. Jeff Keenan at (856) 339-5429.

Document Control Desk Page 2 LR-N10-0198 I declare under penalty of perjury that the foregoing is true and correct.

Executed on (Date)

Sincerely, John . Perry Site Vice President Hope Creek Generating Station Attachments (2)

S. Collins, Regional Administrator - NRC Region I R. Ennis, Project Manager - USNRC NRC Senior Resident Inspector - Hope Creek P. Mulligan, Manager IV, NJBNE Commitment Coordinator - Hope Creek PSEG Commitment Coordinator - Corporate LR-N10-0198 SUPPLEMENTAL INFORMATION NEEDED AMENDMENT REQUEST REGARDING EMERGENCY DIESEL GENERATOR ALLOWED OUTAGE TIME EXTENSION PSEG NUCLEAR LLC HOPE CREEK GENERATING STATION DOCKET NO. 50-354 By letter dated March 29, 2010 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML100900458), PSEG Nuclear LLC (PSEG, the licensee) submitted a license amendment request for Hope Creek Generating Station (HCGS). The proposed amendment would revise the HCGS Technical Specifications (TSs) to extend the Allowed Outage Time (AOT) for the "A" and "B" Emergency Diesel Generators (EDGs) from 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> to 14 days. The U.S. Nuclear Regulatory Commission (NRC) staff is reviewing the amendment request and has concluded that the information delineated below is necessary to enable the staff to make an independent assessment regarding the acceptability of the proposed license amendment in terms of regulatory requirements and the protection of public health and safety and the environment.

1. PSEG's letter dated March 29, 2010, states that the risk evaluation and deterministic engineeringanalysis supportingthe proposed change have been developed in accordancewith the guidelines establishedin Regulatory Guide (RG) 1.177, "An Approach for Plant-Specific Risk-Informed Decision-making: Technical Specifications,"

and RG 1.174, "An Approach for using ProbabilisticRisk Assessment in Risk-Informed Decisionson Plant-Specific Changes to the Licensing Basis." As discussed in both of these RG's, in implementing risk-informed decision-making,proposed licensing basis changes are expected to meet a set of key principles. One of these principlesis that the proposed change is consistent with the defense-in-depth philosophy.

As discussed on page 9 of Attachment I to PSEG's letter dated March 29, 2010:

To ensure that the risk associatedwith extending the AOT for an EDG is minimized, and consistent with the philosophy of maintaining defense in depth, compensatorymeasures will be applied when removing an EDG from service as describedin Section 4.5.1. These measures will ensure the risks associatedwith removing an EDG from service are managed to minimize the increasein risk during the out of service time.

The compensatory measures shown in Section 4.5.1 of Attachment I of the letter dated March 29, 2010, consist of a number of regulatorycommitments to perform various administrativecontrols to minimize the risk during the extended 14-day AOT. As discussed in Regulatory Position 2.2.1, "Defense in Depth" in RG 1.177, consistency with the defense-in-depth philosophy is maintained,in part,by avoiding over-relianceon programmaticactivitiesto compensate for weaknesses in plant design. The NRC staff believes that the proposed amendment relies too heavily on the compensatory measures in light of the fact that the HCGS design does not creditan alternate 1 of 7 LR-N10-0198 alternatingcurrent (AAC) source for station blackout (SBO) as discussed in Section 3.2, of Attachment 1 to PSEG's letter dated March 29, 2010.

The licensee should modify the proposed amendment to reduce over-relianceon programmaticactivities. One approach would be to enhance defense-in-depth by creditingan AAC source, with the capability of handling SBO and loss-of-offsite power loads, to supplement the existing EDGs during the extended 14-day AO T.

The NRC staff notes that the 4 precedent license amendments cited in Section 5.3 of Attachment 1 to PSEG's letter dated March 29, 2010, all included AAC sources as part of the basis for accepting the EDG AO T extension.

RESPONSE TO REQUEST #1 The robust HCGS design features discussed in LAIR H10-03, and the limited key programmatic actions identified, are consistent with the guidance in RG 1.177, Section 2.2.1. This is supported by the fact that the HCGS SBO analysis, consistent with RG 1.155, does not require AAC, and that the key programmatic controls are limited to operability of the remaining EDG in the same division and availability of HPCI/RCIC during the 14 day LCO. Consequently PSEG does not believe the analysis provided in the LAR relies too heavily on compensatory measures or that the compensatory measures cited were to compensate for weaknesses in plant design.

The compensatory measures were provided, consistent with RG 1.177, to add additional margin to the PRA results; with no compensatory measures the RG acceptance guidelines are met with some minor exceptions.

PSEG recognizes the most recent precedents (cited in the LAR to reflect similar improved PRA quality) did also credit an AAC source, consistent with their plant design and SBO analyses (there is no regulatory guidance requiring an AAC source; SBO coping ability and AOT extension defense in depth are based on individual plant design). Even though the PSEG LAR is consistent with other historical precedent that also did not require an AAC source for either SBO or EDG extension (e.g., Clinton and LaSalle), PSEG has decided to provide additional defense in depth beyond what has been demonstrated by analysis as adequate to meet RG 1.177. PSEG will credit the existing onsite gas turbine (designated Salem Unit 3) as an AAC source during the requested A and B EDG AOT extension period. Salem Unit 3 can provide an AAC source of power to HCGS during a LOOP or Station Blackout.

Salem Unit 3 consists of two Pratt and Whitney FT 4A-1 1DF Gas Turbine Engines driving an Electric Generator (TP4-9LF TWIN PAC Gas Turbine installation). The Salem Unit 3 Electrical Generator can provide approximately 38 MW to the 13 KV North bus via the Gas Turbine Switchyard. The Gas Turbine Generator may be synchronized to the grid, paralleled with other electric generators already on the line, or operated alone as an isolated power source. Battery power for the Gas Turbine makes it completely independent of an external power source for starting. The Gas Turbine Generator can be automatically synchronized and loaded to full output in approximately 3 minutes after a manual start. The load rate system can be adjusted to a specific rate of loading the generator.

The initial conditions for use of Salem Unit 3 during the A and B EDG AOT extension period are assumed to be a complete LOOP to the PSEG Nuclear site concurrent with a failure of the HCGS EDGs, resulting in all emergency AC power being lost to HCGS. Salem Unit 3 estimated peak output at 11 OF ambient conditions is approximately 38 MW (reference PSEG procedure 2 of 7

Attachment I LR-N10-0198 S3.OP-SO.JET-0002 Exhibit 1). The Salem Unit 3 output capacity exceeds HCGS design EDG loads (reference HCGS Calculation E-9).

The strategy in place establishes a site configuration in both the HCGS and the Salem switchyards and then back-feeds HCGS through the 500KV switchyards using the generating capacity of Salem Unit 3. The Unit 3 combustion turbine is Dead Bus Bootstrap start capable and is periodically tested in this configuration (reference PSEG procedure S3.OP-PT.JET-0001(Q), Dead Bus Bootstrap Start Test).

HCGS operating procedure HC.OP-AB.ZZ-01 35(Q), Station Blackout / Loss of Offsite Power/

Diesel Generator Malfunction, provides the symptoms needed to identify the LOOP condition.

After verifying low or no voltage on transmission lines 5015, 5023 and 5037, HC.OP-AB.ZZ-0135 directs opening all HCGS 500KV Switchyard circuit breakers, opening all HCGS 13 KV Ring Bus circuit breakers and opening all Island Substation circuit breakers. The HCGS 13 KV Ring Bus is then isolated from the HCGS 500KV Section-2 bus 20X by opening the 2T60 and 4T60 Circuit switchers.

HCGS operating procedure HC.OP-AB.ZZ-01 35(Q), Station Blackout / Loss of Offsite Power/

Diesel Generator Malfunction, then directs HCGS Operations to request the Salem Shift Manager to energize the HCGS 500KV Section-2 bus 20X using Salem Unit 3.

2. The licensee discussion of its updated Individual Plant Examination for External Events (IPEEE)external events risk assessments (Appendix A to Attachment 4 to PSEG's letter dated March 29, 2010) does not appearto address the high level attributesof fire and seismic probabilisticrisk assessment (PRA) models identified in Sections 1.2.4 and 1.2.6 of RG 1.200. In the absence of a licensee assessment using the endorsed standards, this information is critical to the NRC staff review of the technical adequacy of these PRA models, and will need to be provided.

RESPONSE TO REQUEST #2 The high level attributes of fire and seismic probabilistic risk assessment (PRA) models identified in Sections 1.2.4 and 1.2.6 of RG 1.200 (Revision 2) are discussed below for the PRA supporting the proposed A and B EDG AOT extension.

The following items are noted with regard to the fire and seismic probabilistic analysis in support of the EDG AOT extension request:

• Appendix A (Attachment 4 of LAR H10-03) included sections on assumptions and limitations. These sections address the attributes in RG 1.200 Revision 2, Sections 1.2.4 and 1.2.6. These discussions have been placed in the format of RG 1.200 Revision 2 attributes and characteristics and provided in tabular format.

" The impact of the fire PRA on this specific PRA application is limited to those fire areas where a fire may cause a loss of offsite power.

• Other fire risk contributors, even though a potentially large fractional CDF contributor, are not relevant for this particular application. (See page A-1 7 of the EDG AOT risk assessment, Attachment 4 of LAR H10-03.)

• Fire risk can be a significant contributor to the base model. However, fire is a relatively small contributor to the EDG A and B AOT extension request.

3 of 7 LR-N10-0198

• Given that only a small fraction of the fire induced core damage sequences lead to a loss of offsite power, the contribution of fire accident sequences to the EDG AOT extension request is relatively small1 .

The evaluation of the fire PRA quantification approach relative to the attributes and characteristics cited by RG 1.200 Revision 2 in Section 1.2.4 is presented in Table 2-1 (Attachment 2 of this submittal)

The evaluation of the seismic PRA quantification approach relative to the attributes and characteristics cited by RG 1.200 Revision 2 in Section 1.2.6 is presented in Table 2-2 (Attachment 2 of this submittal.)

3. The submittal makes commitments to Tier 2 equipment restrictions(reference Appendix D to Attachment 4 to PSEG's letter dated March 29, 2010). It is not clear if these are creditedin the risk analyses. The proposed amendment would incorporate these restrictionsinto plant proceduresand the TS bases, but not into the TS action requirements. There are no sensitivity analyses provided to allow the NRC staff to determine which, if any, of these restrictionsare critical to the acceptance of this change. The licensee will need to provide appropriatesensitivity analyses to permit staff review of the acceptabilityof these restrictionsand their control in procedures and TS bases rather than in the TS actions.

RESPONSE TO REQUEST #3 Six compensatory measures that are judged useful for inclusion in the EDG AOT extension control process have been identified and are commitments within the PSEG processes. These commitments are identified in Table 3.4-1 of the EDG AOT risk assessment (Attachment 4, Page 3-20 of LAR H10-03).

The compensatory measures implemented in the PRA have sensitivity cases developed. See Tables 3.4-3 and 3.4-5 (Attachment 4 of LAR H10-03) for a summary of the sensitivity cases for these compensatory measures.

Table from Section A.3.4 of the EDG AOT risk assessment, Attachment 4 of LAR H10-03 (also see Table 3.5-5).

Risk Metric Hazard ACDF ALERF Internal 90.2% 88.1%

Fire 7.9% 10.0%

Seismic 1.9% 1.9%

4 of 7 LR-N10-0198 Tier 2 Additional Investigations One aspect of the risk assessment is to provide inputs to decision-makers on how to balance limited resources. The status of the Tier 2 items is that they are considered and they are not pursued due to limited additional benefit and the use of limited resources to pursue them with marginal benefit, that is, the risk metrics with the committed compensatory measures (Table 3.4-1) are all within the NRC acceptance guidelines.

The reduction in the risk metrics associated with the Tier 2 identified administrative controls are as follows:

Reduction in Tier 2 Measure ACDF

• Prestage, test, and train on the alignment of -13.3%(1) the DC portable generator

• Minimize Switchyard Work _0.1%(2)

• Testing of Breakers -0.5%(2)

• Testing of EDGs -0.4%(2)

• Verify Battery Voltage -0.1%

• Test SACS valve 2457A -0.2%(3)

(1) Assumes 75% credit for the administrative control.

(2) Assumes 5 0% credit for the administrative control.

(3) Assumes 10% credit for the administrative control.

The % change in CDF risk metric can, as a first approximation, represent the change in ACDF and ICCDP associated with the administrative control changes. The ALERF and ICLERP are not calculated here but are also on the same order of magnitude, i.e., quite low in impact.

The Tier 2 evaluation is presented in Appendix D (Attachment 4 of LAR H10-03) to provide decision makers with additional options to consider if the assessed risk is not acceptably small.

Appendix D (p. D-3) identifies the following conclusions regarding these Tier 2 actions:

The review of the risk significant configurations has identified six additional (i.e., the Tier 2 actions on pp D-2, D-3) compensatory actions that could be considered as part of the EDG AOT extension. None of these actions are currently credited in the risk assessment, nor are any of these actions necessary to meet the acceptance guidelines for RG 1.177 and 1.174.

4. The risk analyses are dependent upon the once per 2 year use of the extended AO T.

The licensee proposes no administrativecontrol for voluntary use of the AO T to once per 2 years. Emergent repairsmay result in additionaluse of the extended AOT, but the risk analyses do not address this. The licensee will need to provide a technicaljustification for this assumption, and will need to provide sensitivity studies to address emergent repairuse of the extended AOT considering the increasedprobabilityof common cause failures (consistentwith Appendix A to RG 1.177).

5 of 7 LR-N10-0198 RESPONSE TO REQUEST #4 The technical justification for this assumption, using a sensitivity study addressing emergent repair use of the extended AOT that considers the increased probability of common cause failures, is provided below.

LAR Submittal Analysis The base calculated ACDF and ALERF risk metrics presented in the LAR are contingent on the assumption of one 14 day outage per 2 years for each EDG plus the nominal additional unavailability using recent operating history. This is consistent with historical experience with the C&D EDGs. The ICCDP and ICLERP risk metrics are not dependent on the once per 2 year use of the extended AOT. Therefore, the evaluation presented in the LAR is conservative. This can be understood, because most 2 year PMs are of relatively short duration (2-4 days). The 5 year PM is longer and may take 7-10 days.

PSEG has shown by historical performance of the C&D EDGs (which have a 14 day AOT) that the AOT is not abused and is used judiciously (See Table 3.4-0 of the EDG AOT risk evaluation, of LAR H10-03.) No 14 day emergent AOTs have been used for the C and D EDGs in the 15 years since approval of their extended AOT.

The common cause failure effects of emergent work are effectively addressed because an immediate test is required. If the failure is a common cause, then at least two (2) EDGs would be unavailable and a 14 day outage is no longer possible because another, more restrictive, TS is entered. See page 3-5 of risk analysis (Attachment 4 of LAR H10-03):

- For emergent corrective maintenance outages, the PSEG practice (and Technical Specification Requirement) is to demonstrate that other similar components are not subject to the same failure, i.e., that there is no common cause link. This is part of the HCGS Technical Specifications 2. Therefore, no model adjustment is made to reflect an increased potential for common cause if one component is OOS for corrective maintenance.

In addition, the HCGS Maintenance Rule program limits the unavailable hours of the A&B EDGs to 325 hours0.00376 days <br />0.0903 hours <br />5.373677e-4 weeks <br />1.236625e-4 months <br /> per cycle.

Addition of Emergent 14 Day AOT Emergent repairs that require a 14 day AOT are anticipated to be rare occurrences, i.e., would not be recurring events, consistent with the operating experience of the C&D EDGs which already have been granted a 14 day AOT (See Table 3.4-0 of the EDG AOT risk evaluation, of LAR H10-03).

Nevertheless, a sensitivity calculation has been developed that addresses the extended EDG AOT for emergent repair of EDG A or B and postulated common cause EDG failures:

2 If the diesel generator became inoperable due to any cause other than an inoperable support system, an independently testable component, or preplanned preventive maintenance or testing, demonstrate the OPERABILITY of the remaining diesel generators by performing Surveillance Requirement 4.8.1.1.2.a.4 separately for each diesel generator within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> unless the absence of any potential common mode failure for the remaining diesel generators is demonstrated.

6 of 7

Attachment I LR-N10-0198 Emergent repairs that impact an EDG and force it into a 14 day AOT are assumed to occur once per cycle. (This is extremely conservative and this frequency of challenge is not supported by historical evidence.)

For this added (added to the assumed planned 14 day planned outage every 2 years) challenge, the most severe case is evaluated:

Assume the emergent issue is a common cause and that this common cause is not discovered by the PSEG engineering process or the required testing. (The common cause failure probabilities of EDGs are therefore reset in the model from their values in the base PRA to the associated common cause factors, assuming that the initial random failure has already occurred. The failure is also conservatively assumed to affect both Fail to Start and Fail to Run common cause basic events in the model.)

This would be the case postulated in Appendix A. 1.3.2 of RG 1.177 (August 1998).

Such a sensitivity case is quite conservative as noted in Appendix A. 1.3.2.3 which identifies that when a component is "tested operable" as would be the case for the HCGS EDGs, then the common cause terms would be set to zero or "false" in the Boolean model. This recommended process by RG 1.177 is more closely associated with the base model calculation for the EDG AOT extension as submitted in the Attachment 4 of LAR H10-03.

The calculated risk metrics for ACDF and ALERF are recalculated for this additional assumed emergent AOT that is conservatively assumed to be a recurring event every cycle. (Compensatory measures 3 through 6 are included in the assessed values.)

The results of this sensitivity case are that the ACDF and ALERF risk metrics meet the acceptance guidelines in RG 1.174:

ACDF = 5.3E-7/Rx yr ALERF = 5.5E-8/Rx yr ICCDP = 3.3E-7 ICLERP = 3.7E-8 These results can be compared with the acceptance guidelines as follows:

] ACCEPTANCE RISK METRIC j TOTAL CHANGE GUIDELINE SOURCE ACDF 5.3E-7/Rx yr 1.OE-6/Rx yr RG 1.174 ALERF 5.5E-8/Rx yr 1.OE-7/Rx yr RG 1.174 ICCDP 3.3E-7 5.OE-7 RG 1.177 ICLERP 3.7E-8 5.OE-8 RG 1.177 All acceptance guidelines continue to be met for this sensitivity case.

7 of 7 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" The fire evaluation was performed on the basis of fire areas which are plant locations completely enclosed by at least two hour rated fire barriers. The fire area boundaries which meet the FIVE fire barrier criteria are assumed to be effective in preventing a fire from spreading from the originating area to another area. The fire area boundaries recognized in the IPEEE fire analysis are identical to those identified in the HCGS UFSAR (1995). In some cases, these fire areas were further subdivided into compartments in the detailed PRA evaluation where it could be

" Global analysis boundary demonstrated that the space was bounded by barriers, where heat and products of captures all plant locations combustion would be substantially confined.

relevant to the fire PRA.

Plant Boundary

• Fire barriers for compartments defined for this analysis were defined in accordance Definition and " Physical analysis units (PAUs) with the EPRI FIVE Method, Paragraph 5.3.6. The fire compartments which met the Partitioning are identified by credited FIVE criteria covered the turbine building, reactor building, control/diesel building, partitioning elements that are radwaste building, service water intake structure, and yard.

capable of substantially confining fire damage behaviors. The design and plant layout of Hope Creek make fire propagation to multiple compartments unlikely compared to the fire risk in individual compartments. An explicit multi-compartment review was not performed.

EDG AOT:

The HCGS IPEEE boundary definition and partitioning is used in the EDG AOT fire hazard quantification. Specifically, the HCGS Fire PRA Analysis includes the Control/Diesel Building in the global analysis boundary, which houses the EDGs. The EDG rooms are identified as unique PAUs.

1 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" Room inventory is developed by a review of the UFSAR, MMIS lists, pre-fire plans, and as witnessed during walkdowns. A Fire Compartment Interaction Analysis (FCIA) Data Sheet was created along the lines of the FIVE methodology. A FCIA sheet was completed for each compartment and required the contents of the compartment, along with sources of this information, whether the equipment in the compartment could cause a plant trip, and whether the compartment contains Appendix R safe shutdown equipment.

" Equipment is selected for " At the time of the Hope Creek IPEEE, the treatment of MSOs was rudimentary. As inclusion in the plant response noted in Section A.3.3, "hot shorts" were considered for selected fires, e.g., fires model that will lead to a fire- affecting ISLOCA, spurious ADS, SORV, and LOCAs.

induced plant initiator, or that is needed to respond to such an The assessment of "hot shorts" considered the possibility of hot shorts for each initiator (including equipment scenario and commented on the possibility under the heading Initiating Event(s) subject to fire-induced spurious within the Fire Scenario Analysis worksheets. Only the control room, lower control Equipment actuation that affects the plant equipment room, and switchyard blockhouse were found susceptible to hot short Selection response). actuation of equipment. The occurrence of hot shorts might cause an SORV (LOCA),

LOOP, or Loss of SWS/SACS. These effects were considered during the calculation of

" The number of spurious core damage frequency.

actuations to be addressed increases according to the This assessment used a value of 3 0 %; that is, given a fire scenario in which a hot significance of the consequence short might cause unwanted effects, the likelihood of those effects is 30% of the (e.g., interfacing systems LOCA). likelihood of the fire scenario. The remaining 70% of the fire scenario is treated as if

" Instrumentation and support hot short did not occur.

equipment are included.

Fire induced LOCAs were found to occur only because of hot shorts, as described above, in cabinets that contain control wiring for SRVs or ADS. This can occur only in the control room and lower control equipment room. Using the above cited value of the conditional probability of hot shorts, the total core damage frequency associated with fire induced LOCAs was found to be approximately 4E-07/yr . (It is noted that the dominant contributors to ACDF and ALERF for the EDG AOT extension analysis do not contain LOCA events.)

2 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

Desirable TeWchnca Chara~cteristics Eeetand Attribuites HCGS Response

(~As noted in RG 1.200 Rev. 2)

An analysis of the interfacing high to low pressure systems was performed for the HCGS PRA. The analysis was reviewed for applicability to fire scenarios. No high to low pressure interface is susceptible for fire scenarios, with one exception. This is because all boundaries are protected by at least two diverse, closed isolation valves, one of which is a check valve or stop check valve. Even if a sustained hot short opened an MOV, the check valves are not susceptible to opening by fire scenarios.

The one exception to this is the RHR shutdown cooling suction lines which are isolated by two closed MOVs. For this case, the shutdown cooling suction valve (BC-HV-F008) is disabled at the circuit breaker by a key switch to prevent inadvertent opening during fires.

Equipment . All support systems are included in the analysis.

Selection (cont'd) 0 Instrumentation availability is addressed in the Control Room fire evaluations.

EDG AOT:

The HCGS IPEEE model for equipment selection is used in the EDG AOT fire hazard quantification.

Equipment included in the HCGS Fire PRA analysis in the EDG rooms includes all cables and equipment in the rooms. The treatment of hot shorts remains as described above.

3 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

Room inventory for inclusion in the model is developed by a review of the UFSAR, MMIS lists, pre-fire plans, and as witnessed during walkdowns. A Fire Compartment Interaction Analysis (FCIA) Data Sheet was created along the lines of the FIVE methodology. A FCIA sheet was completed for each compartment and required the contents of the compartment, along with sources of this information, whether the equipment in the compartment could cause a plant trip, and whether the compartment contains Appendix R safe shutdown equipment.

All fire damage calculations assume cables are unprotected even if they are in conduit, protected by a cable tray bottom, or protected by an enclosed cable tray.

Furthermore, if any cable in a stack of trays was calculated to be damaged, all of Cables that are required to the cables in the stack were assumed to be damaged. In other words, neither support the operation of fire PRA Cable Selection shielding nor delayed fire growth from tray to tray were considered in the fire equipment (defined in the damage calculations.

equipment selection element) are identified and located. Lack of knowledge about the termination points (i.e., functions) of specific cables in a compartment was treated as causing failure of the entire channel in which the cable belongs, if one cable was calculated as damaged.

Selected credit for Balance of Plant (BOP) was included in the 2003 PRA fire update to reflect BOP availability for fires that are in areas of the plant that obviously do not affect the BOP availability.

EDG AOT:

The HCGS IPEEE cable search to support the operation of PRA equipment used in a fire is also used in the EDG AOT fire hazard quantification. All cables located in the EDG rooms are included in the analysis, as well as cables and buses for offsite power.

HCGS IPEEE:

Qualitative S A qualitative screening analysis was not performed for the HCGS IPEEE; that is, no Screening Screened out physical analysis compartments within the defined plant analysis boundary were eliminated from (Optional Element)contributions units to risk negligible represent and are consideration owing to qualitative factors alone.

Element) considered no further. EDG AOT:

The EDG rooms are not screened out.

4 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

The fire PRA model for the IPEEE was formulated to be compatible with the internal events PRA model for system failures and accident sequence logic.

" At the time of the Hope Creek IPEEE, the treatment of MSOs was rudimentary. As noted in Section A.3.3, "hot shorts" were considered for selected fires, e.g., fires affecting ISLOCA, spurious ADS, SORV, and LOCAs. See discussion under the "Equipment Selection" element for further details.

Based upon the internal events

• CCDPs are calculated using the post-initiator operator actions modeled in the PRA PRA, the logic model is adjusted model with human error probabilities (HEPs) unmodified from the internal event to add new fire-induced initiating values. The CCDP calculations of the Fire PRA took advantage of only two recovery events and modified or new actions: 1) recovery of alternate ventilation following a loss of 1E Panel Room HVAC, accident sequences, operator and 2) control of the plant from the remote shutdown panel following a fire that Fire PRA Plant actions, and accident compromises the ability of operators to completely control the plant from the Response Model progressions (in particular those control room.

from spurious actuations).

Inapplicable aspects of the EDG AOT:

internal events PRA model are The model changes to ensure as-built, as-operated fidelity of the PRA model used in the bypassed. EDG AOT extension application necessitated the use of the latest PRA models.

The internal fire probabilistic risk assessment for the EDG AOT extension analysis is based on the latest HCGS internal events model (2008B) which incorporates the fire analysis developed as part of the IPEEE and updated in 2003, i.e., the latest system and accident sequence models.

The 2008B PRA system models and accident sequence models were used in the model.

Fire initiating events are included directly in the model.

5 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

• The fire PRA added new initiating events to the PRA that caused a reactor scram challenge and also defeated those systems that are directly impacted by the fire (i.e., SSCs or cables).

" The PRA accident sequences were adapted to reflect the impacts of the fire (e.g.,

Fire PRA Plant loss of offsite AC power, MSIV closure).

Response Model (cont'd) " Internal Events mitigation systems that are failed by the fire are failed in the fire PRA quantification.

The FPIE PRA model is appropriate for EDG room fire scenarios. A fire in a single EDG room will fail that EDG and the cables exposed to the fire in that room, and they will not be recovered.

6 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" A fire scenario is defined as a unique source, fire intensity, target, and initiating event combination. Fire damage calculations and fire damage time versus suppression time calculations supported the probabilistic analysis. Fire growth and

" Fire scenarios are defined in propagation was considered.

terms of ignition sources, fire growth and propagation, fire The approach taken for the fire PRA was to perform a scenario-by-scenario analysis of detection, fire suppression, and unscreened compartments accounting for the relative location of ignition sources and cables and equipment ("targets") targets. Fire damage calculations were performed to determine the extent of potential damaged by fire. damage from each postulated fire source. Openings in walls as well as open active fire dampers were included in the assessment of the extent of fire damage.

" The effectiveness of various fire protection features and systems Fire barriers for compartments defined for this analysis were defined in accordance is assessed (e.g., fixed with the EPRI FIVE Method, Paragraph 5.3.6.

Fire Scenario suppression systems). Fire damage calculations were used to assess the spread of damage, owing to a hot Selection and

• Appropriate fire modeling tools gas layer, through openings in walls. In these calculations, all walls in the source Analysis are applied. room, below the level of the opening, were assumed non-existent.

" The technical basis is established " Fire suppression was assumed to fail, as well as manual fire suppression efforts and for statistical and empirical fire brigade response.

models in the context of the fire

" The technical basis of the HCGS fire IPEEE was a PRA performed in a manner scenarios (e.g., fire brigade consistent with the guidance in NUREG/CR-2300 and NUREG/CR-4840. The PRA is response). preceded by: 1) a fire compartment interaction analysis (FCIA) per EPRI FIVE

" Scenarios involving the fire- guidance, and 2) a quantitative screening analysis also performed in a manner induced failure of structural steel consistent with FIVE guidance.

are identified and assessed (at

" Fire damage calculations were performed using a modified version of the least qualitatively).

formulation found in the Fire Screening Methodology User Guide (EPRI FIVE).

" Fire-induced failure of structural steel was not considered.

7 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

EDG AOT:

The HCGS IPEEE fire scenario development is used in the EDG AOT fire hazard quantification.

Fire scenarios in the EDG rooms are defined in terms of appropriate ignition sources Fire Scenario located within those rooms. Fire detection and suppression are not credited in the EDG Selection and rooms. Whole room fires within the EDG rooms will fail all components and unprotected Analysis cables in the rooms.

(cont'd)

The exceptions are the A and B EDG rooms. The 10A108 and 10A109 electrical buses connect the station service (offsite power) transformers to the four 1E 4kV switchgear divisions, and these two buses run through all four EDG rooms. These buses are protected by fire wrap in the A and B rooms. Therefore, a fire in the A or B EDG room will maintain offsite power due to the fire wrap protection.

8 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" A fire ignition frequency, using the method of FIVE, was developed for each of the 209 fire compartments. This method was implemented using a Fire Compartment Ignition Source Data Sheet (ISDS) for each compartment.

" Fire ignition frequencies are established for ignition sources and for compartments.

While the screening fire frequency was developed by summing the frequencies of all fire ignition sources in a compartment, the fire PRA was performed on a source-by-source basis. That is, the fire ignition frequency of each scenario was developed

" Frequencies are established for separately for each identified source in a compartment. This was easily derived ignition sources and from the ISDS analysis of each compartment, by simply using the frequency of the consequently for physical individual sources.

analysis units. " The PRA included transient combustible sources as individual scenarios. The

" Transient fires should be method to obtain the frequency of transient combustible fire ignition frequencies Fire Ignition was derived from FIVE. Even though transient combustibles, of sufficient quantity Frequencies postulated for all physical analysis units regardless of to damage cables, were not found in any compartment at the HCGS, a thorough administrative controls. transient combustible analysis was performed. Each compartment included consideration of transient combustibles. This analysis assumed that transient

" Appropriate justification must be combustibles could be located anywhere in the plant.

provided to use nonnuclear experience to determine fire " Nonnuclear experience was not used for fire ignition frequency.

ignition frequency. EDG AOT:

The HCGS IPEEE ignition frequencies were updated in 2003 using NRC reevaluation of the fire events database to reflect later data. Other assumptions and approaches that were adopted in the IPEEE are preserved.

Fire frequencies for equipment and components within the EDG rooms are established, as well as frequencies for the EDG room PAUs. Transient fires are also postulated for EDG rooms. Nonnuclear experience was not used for fire ignition frequency determination.

9 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" A screening process was implemented to avoid a detailed PRA on all of the 209 compartments identified from the Fire Compartment Interaction Analysis and the transformer array in the yard.

" Quantitative screening for fire compartments used a conservative, screening core damage frequency (SCDF). The screening assessment first developed a fire ignition frequency for the compartment, and then assumed that all equipment and cables in the compartment are failed due to the fire. Next the screening identified a o Physical analysis units that are conservative initiating event (reactor trip transient). Finally, the screening process screened out from more refined used a screening conditional core damage probability (SCCDP) from the HCGS IPE quantitative analysis are retained model.

Quantitative to establish CDF and LERF/LRF.

The SCDF was the product of the fire ignition frequency and the SCCDP. The Screening " Typically, those fire PRA compartment was removed from further consideration (screened) if the SCDF for contributions to CDF and that compartment was found to be less than 1E-06/yr.

LERF/LRF that are established in the quantitative screening phase " Each of the unscreened compartments was subjected to a detailed scenario-by-are conservatively characterized. scenario probabilistic analysis. A fire scenario is defined as a unique source, fire intensity, target, and initiating event combination. The total core damage frequency of each compartment was evaluated considering the range of potential interactions of fire sources, targets, intensities, and initiating events.

EDG AOT:

The HCGS IPEEE quantitative screening was retained for the EDG AOT fire hazard quantification. Fires in compartments that could lead to a loss of offsite AC power were not screened.

The EDG rooms were not screened out.

10 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

All fire damage calculations assume cables are unprotected even if they are in conduit, protected by a cable tray bottom, or protected by an enclosed cable tray.

Furthermore, if any cable in a stack of trays was calculated to be damaged, all of the cables in the stack were assumed to be damaged. In other words, neither shielding nor delayed fire growth from tray to tray were considered in the fire damage calculations.

Lack of knowledge about the termination points (i.e., functions) of specific cables in a compartment was treated as causing failure of the entire channel in which the cable belongs, if one cable was calculated as damaged.

At the time of the Hope Creek IPEEE, the treatment of MSOs was rudimentary. As noted in Section A.3.3, "hot shorts" were considered for selected fires, e.g., fires affecting ISLOCA, spurious ADS, SORV, and LOCAs. See discussion under the The conditional probability of "Equipment Selection" element for further details.

Circuit Failure occurrence of various circuit Analysis failure modes given cable The explicit identification and modeling of instrumentation required to support PRA damage from a fire is based credited operator actions is not addressed. The industry treatment for this task is upon cable and circuit features. still being developed.

EDG AOT:

° The HCGS IPEEE treatment of circuit failures is adopted for the EDG AOT fire hazard quantification.

All EDG cables are considered unprotected with the exception of the fire wrapped power buses in the A and B EDG rooms.

The HCGS IPEEE screening was adopted for the EDG AOT fire hazard quantification.

The change in risk metrics for the EDG AOT extension evaluation is related to fires in areas that can lead to a loss of offsite AC power. The areas that could cause a loss of offsite AC power were not screened using this process.

11 of 19 LR-N 10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" Operator actions from the internal events PRA are retained in the fire PRA and are assessed for fire effects. However, only two recovery actions are considered to be affected, as described below.

• Operator actions and related post-initiator HFEs, conducted " CCDPs are calculated using the post-initiator operator actions modeled in the PRA both within and outside of the model with human error probabilities (HEPs) unmodified from the internal event main control room, are values. The CCDP calculations of the Fire PRA took advantage of only two recovery addressed. actions: 1) recovery of alternate ventilation following a loss of 1E Panel Room HVAC, and 2) control of the plant from the remote shutdown panel following a fire that

• The effects of fire-specific compromises the ability of operators to completely control the plant from the procedures are identified and control room. Post 'initiator operator actions, such as inhibit of ADS, used HEP incorporated into the plant values unmodified from the internal event values unless control room abandonment response model. resulted from the scenario. For scenarios involving control room abandonment, the Postfire Human only human action considered was failure to continue operating the plant using the Reliability Analysis

• Plausible and feasible recovery actions, assessed for the effects alternate shutdown procedure with the remote shutdown panel and local manual of fire, are identified and controls which explicitly considered the fire performance shaping factors.

quantified. " Fire-specific procedures or recovery actions are not identified and are not

• Undesired operator actions incorporated into the plant response model except as noted above.

resulting from spurious " Undesired operator actions resulting from spurious indications are not considered.

indications are addressed.

EDG AOT:

" Operator actions from the internal events PRA that are The treatment of operator actions in the EDG AOT fire hazard quantification is the same as retained in the fire PRA are that adopted in the IPEEE. This is acceptable because fires in the locations that could affect assessed for fire effects. the EDG AOT extension request (i.e., EDG compartments, switchgear rooms, and transformers rooms) do not affect the actions modeled in the PRA.

Loss of an EDG due to a fire is not recovered in the model. Operator actions postulated in the Full Power Internal Events (FPIE) model are retained as applicable.

12 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

• For each fire scenario, the fire risk results are quantified by combining the fire ignition frequency, the probability of fire damage and the conditional core damage probability from the fire PRA plant response model.

" For each fire scenario, the fire risk results are quantified by For LERF, Supplement 4 of Generic Letter 88-20 states that the evaluation of the combining the fire ignition containment performance of external events should be directed toward a systematic frequency, the probability of fire examination 1) to determine the existence of containment failure modes owing to damage and the conditional core fire induced sequences that are distinctly different from sequences found in the IPE damage probability (and internal events evaluation and 2) to determine if fires can contribute significantly to CLRP/CLERP) from the fire PRA direct functional failure of the containment which is not a result of a core damage plant response model sequence. The conclusion of this evaluation is that there are no fire induced Fire Risk containment failure modes that are significantly different from those treated in the Quantification " Total fire-induced CDF and HCGS IPE. Therefore, no further containment performance analysis is needed.

LERF/LRF are calculated for the plant and significant contributors " Total fire-induced CDF is calculated for the plant and significant contributors identified identified. See discussion above for LERF.

" The contribution of quantitatively

• The contribution of quantitatively screened scenarios is not added to the total risk.

screened scenarios (from the EDG AOT:

quantitative screening element) is added to yield the total risk " The EDG AOT extension fire hazard analysis uses the full 2008B PRA for both the values CDF and LERF calculations. The fire initiating events are modeled to fail the appropriate SSCs and cables associated with the fire for all non-screened events.

" The screened events do not include areas that could lead to a loss of offsite AC power or to the failure of diesels.

13 of 19 LR-N10-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

" The Seismic/fire interaction was evaluated for three issues: (1) the potential for seismically, induced fires, (2) the potential for seismically-induced actuation of fire suppression systems, and (3) the potential for seismically-induced degradation of fire suppression systems.

" Seismic-Fire interaction walkdowns were performed. No risk contributors that warranted quantification were identified.

" During the seismic walkdown, the team focused on equipment whose failure could be a fire source that would damage equipment important to seismic safety. No credible failures were found. The emergency diesel generator fuel oil day tanks and storage tanks were found to be seismically rugged. All piping associated with the above equipment was found to be sufficiently seismically rugged not to pose a Potential interactions resulting significant fire risk. Both 1E and non-lE cabinet anchorages were included in the from an earthquake and a seismic walkdown and assessment. All non-lE cabinet anchorages were either resulting fire that might screened out or found to have median capacities in excess of 1.5g. Therefore, contribute to plant risk are seismic interactions of non-lE cabinets and 1E equipment is not a significant fire Seismic Fire reviewed qualitatively risk.

Interactions Qualitative assessment verifies A low ruggedness relay evaluation was performed for the HCGS. A total of 12 that such interactions have been panels were identified that contained low ruggedness relays. In addition another 38 considered and that steps are miscellaneous low ruggedness relays were identified. None of these were in the fire taken to ensure that the protection or detection systems. It is concluded that seismic actuation of fire potential risk contributions are suppression systems does not pose a significant risk of flood or a significant mitigated likelihood of disabling safety related equipment.

However, the fire water pumps are located in a Fire Water Pump House which is a block wall structure that is not seismically qualified. The fire water tanks are located outside of this structure and are not seismically qualified. The limiting seismic failure of the fire water system is failure of the tanks. The seismic core damage frequency assessments did not take credit for the fire water system because of its perceived lack of robustness against earthquakes. The fire core damage frequency assessment did not take credit for fire water suppression systems. It is concluded that the unavailability of fire water after an earthquake is the principal mode of seismically induced fire suppression system degradation.

EDG AOT:

The IPEEE approach to seismic-fire interactions is adopted for the EDG AOT extension.

14 of 19 LR-NiO-0198 Table 2-1 Comparison of HCGS FPRA Analysis to RG 1.200 Revision 2 Table 5 (Section 1.2.4)

HCGS IPEEE:

• Uncertainty in quantitative fire PRA results because of

• Sources of uncertainty regarding fire ignition data-and estimation of CDF were are evaluated.

parameter uncertainties evaluated

• Although suppression systems were not credited, a suppression system Uncertainty and Sensitivity " Model uncertainties as well as effectiveness study was included.

the potential sensitivities of the EDG AOT:

results to associated assumptions are identified and Uncertainty in the overall model was evaluated.

characterized The NUREG-1855 process was followed to identify modeling uncertainties.

15 of 19 LR-N10-0198 Table 2-2 Comparison of HCGS Seismic Analysis to RG 1.200 Revision 2 Table 7 (Section 1.2.6)

Seismic hazard analysis HCGS IPEEE:

- establishes the frequency of

• The seismic hazard analysis identifies the sources of earthquakes, evaluates earthquakes at the site earthquake history in the region, develops attenuation relationships, and

- site-specific determines the frequency of exceedance.

- examines all credible sources of damaging earthquakes " The hazard estimate depends on uncertain estimates of attenuation, upper bound

- includes current information magnitudes, and the geometry of the postulated sources. Such uncertainties are

- based on comprehensive data, included in the hazard analysis by assigning probabilities to alternative hypotheses including about these parameters. A probability distribution for the frequency of occurrence

- geological, seismological, and is thereby developed. The annual frequencies for exceeding specified values of the geophysical data ground motion parameter are displayed as a family of curves with different

- local site topography probabilities; they are presented in terms of median, mean, 15 percentile and

- historical information 85 percentile curves.

- reflects the composite Some differences in the shapes of 85 percentile and median uniform hazard spectra distribution of the informed Probabilistic at 10,000 year return period were observed leading one to suspect that there is technical community.

Seismic Analysis - level of analysis depends on uncertainty in the spectral shape. However, the effect of this uncertainty was considered to be small for the following reasons: 1) the difference in the spectral application and site complexity shape seen for the 5% damped spectra may not be relevant to Hope Creek because Aleatory and epistemic of the high composite soil-structure damping (over 10%) present for Hope Creek, 2) uncertainties in the hazard analysis the uncertainty arises because of the uncertainty in the spectral attenuation (in characterizing the seismic relationships which is already considered in the PGA attenuation relationships, and sources and the ground motion 3) the effect of adding the spectral shape uncertainty is not significant considering propagation) the overall uncertainties in the fragilities and hazard curves.

- properly accounted for

• Slope stability, lateral spreading, and soil liquefaction are evaluated.

- fully propagated EDG AOT:

- allow estimates of

> fractile hazard curves, The HCGS seismic hazard analysis is adopted for use in the EDG AOT seismic hazard

> median and mean hazard quantification.

curves,

> uniform hazard response Seismic model characteristics of this element are not dependent on EDG configuration, spectra components, or operability.

16 of 19 LR-N10-0198 Table 2-2 Comparison of HCGS Seismic Analysis to RG 1.200 Revision 2 Table 7 (Section 1.2.6)

Spectral shape used in the seismic PRA

- based on a site-specific evaluation

- broad-band, smooth spectral shapes for lower-seismicity sites acceptable if shown to be appropriate for the site Probabilistic - uniform hazard response Seismic Analysis spectra acceptable if it reflects (cont'd) the site-specific shape

• Need to assess whether for the specific application, other seismic hazards need to be included in the seismic PRA, such as

- fault displacement

- landslide,

- soil liquefaction

- soil settlement HCGS IPEEE:

• Seismic fragilities of structures and equipment were estimated using the procedures described by Kennedy and Reed in the IPEEE. Seismic fragilities in this study have been developed in terms of the peak ground acceleration capacity of structures and Seismic fragility estimate

- plant-specific equipment. As such, the three fragility parameters Am, BR and BU have been calculated for each screened-in component in its significant failure modes. A brief

- realistic

- includes all systems that description of the methods used to calculate the fragility parameters and the results Seismic Fragility are given in the report by EQE.

participate in accident Analysis sequences included in the The process of seismic fragility evaluation can be described by the following steps:

seismic-PRA systems model

- basis for screening of high 1. Based on the preliminary systems analysis and on previous seismic PRAs, a set capacity components is fully of structures and equipment (about 100 items) is selected for fragility described evaluation.

2. Plant design and seismic qualification information is collected.

3. Probabilistic floor and structural response are developed by analysis or by appropriate extrapolation of the design information.

17 of 19

I

'U LR-N10-0198 Table 2-2 Comparison of HCGS Seismic Analysis to RG 1.200 Revision 2 Table 7 (Section 1.2.6)

" Seismic fragility evaluation performed for critical SSCs based Plant walkdowns were performed to search for seismic vulnerabilities, to assist in on

- review of plant design screening out high capacity components and to collect additional data on components needing detailed fragility analysis. Procedures for seismic walkdowns documents

- earthquake experience data are given in the EPRI seismic margin assessment methodology report.

Seismic Fragility Analysis - fragility test data EDG AOT:

(cont'd) - generic qualification test data (use is justified) The HCGS IPEEE seismic analysis is adopted for the EDG AOT seismic hazard

- walkdowns quantification.

" Walkdowns focus on The Control/Diesel Building, which houses the EDGs, is included as a critical SSC that

- anchorage was evaluated.

- lateral seismic support

- Dotential systems interactions 18 of 19 LR-N10-0198 Table 2-2 Comparison of HCGS Seismic Analysis to RG 1.200 Revision 2 Table 7 (Section 1.2.6)

HCGS IPEEE:

- seismic-caused initiating events

• Traditional event tree techniques were used to delineate the potential combinations

- seismically induced SSC of seismic-induced failures, and resulting seismic scenarios, which were termed "seismic damage states." The frequencies of these seismic damage states were failures

- nonseismically induced quantified by convoluting the earthquake hazard curve with the structure and unavailabilities, equipment seismic fragility curves.

- other significant failures For those scenarios that required additional non-seismic failures to occur to result in (including human errors) that core damage, the PRA internal events model (event trees and fault trees) was used can lead to CDF or LERF to develop conditional core damage probabilities, with appropriate changes given the seismic damage state. These calculations incorporate random failures of The seismic PRA models Seismic Plant equipment and operator actions.

- adapted to incorporate Response Analysis seismic-analysis aspects that

• The event and fault tree models developed for the HCGS internal events PRA have are different from been used as the starting point for the seismic IPEEE models. Traditional event tree corresponding aspects found techniques were used to delineate the potential combinations of seismic-induced in the at-power, internal failures, and resulting seismic scenarios, which were termed "seismic damage events PRA model states."

- reflects the as-built and as-operated plant being analyzed " The seismic event tree (SET) is used to delineate the potential successes and failures that could occur due to a seismic event, based on the structures and Quantification of CDF and LERF components and their fragilities. Boolean equations were developed for each of the integrates SET top events, based on the logic and seismic fragility information. Each seismic

- the seismic hazard sequence equation represents the Boolean logic associated with its corresponding

- the seismic fragilities seismic damage state (SDS).

- the systems analysis EDG AOT:

The EDG AOT extension analysis for seismic effects has adopted the IPEEE seismic analysis into the CAFTA framework and has updated all system and event tree logic to reflect the as-built, as-operated plant.

HRA: The HEPs are modified to reflect the increased probability of failure under seismic events (e.g., increased stress, increased work load,- limitations in access).

It is noted that a seismic event that fails one EDG is a perfectly correlated failure mode (i.e., if a seismic event fails one EDG, then all EDGs are assumed failed). Therefore, no impact of the EDG unavailability is seen for these more severe seismic events.

19 of 19