ML120030437

Fpra Pr Position Paper Final

ML120030437
Person / Time
Site:	Waterford
Issue date:	01/03/2012
From:	Entergy Operations
To:	Division of Operating Reactor Licensing
Kalyanam N, NRR/DLPM, 415-1480
References
TAC ME7602
Download: ML120030437 (14)

Text

Waterford 3 (WF3) Fire PRA Peer Review Position Paper Page 1 of 4

Background

The NRC requested additional information on the Waterford 3 Fire PRA peer review to determine if changes to the Waterford-3 Fire PRA since the peer review would require a focused-scope peer review.

Summary The Waterford 3 Fire PRA Peer Review was performed in November 2010 [1]. The peer review evaluated the Fire PRA against all technical elements in Section 4 of the ASME/ANS PRA Standard with the exception of the Qualitative Screening and Quantitative Screening elements, which are not applicable to Waterford-3. These sections are not applicable to WF3 because WF3 did not screen any fire Plant Analysis Units (PAUs) from the Fire PRA based on qualitative or quantitative criteria (e.g., no automatic or manual scram or low CDF values).

The Waterford 3 Fire PRA Peer Review report [1] states:

At the time of the peer review, the WSES Fire PRA was complete but on-going enhancements to reduce the Core Damage Frequency (CDF) and Large Early Release Frequency (LERF) were in progress. This review was conducted against the completed WSES PRA as of November 2010.

The ongoing enhancements to reduce CDF and LERF involved review of fire impact on component failures for important fire scenarios and were performed within the methodology reviewed by the peer review team. In addition, the peer review findings were addressed to meet Capability Category II of the Fire PRA standard requirements [2]

with the exception of seven Fire Scenario supporting requirements (SRs) where the conservative analysis for Capability Category I provides acceptable results for Waterford

3. These seven requirements are discussed in the attached Table V-2.

Appendix 1-A of the ASME/ANS PRA Standard [2] discusses the types of changes that would require an additional peer review including multiple examples of changes that would or would not require a peer review. PRA model upgrades generally satisfy one of three criteria: (1) new methodology, (2) change in scope that impacts the significant accident sequences or the significant accident progression sequences, and (3) change in capability that impacts the significant accident sequences or the significant accident progression sequences. None of the changes made for the Waterford-3 Fire PRA since the Peer Review would be considered PRA upgrades as defined by the PRA Standard [2] or NEI 07-12 [4].

Attached is a detailed table providing an explanation for how each of the Fact &

Observations (F&Os) findings was closed or why this issue does not affect the NFPA 805 results (Note that Suggestions from the Peer Review were excluded from this Table).

Waterford 3 (WF3) Fire PRA Peer Review Position Paper Page 2 of 4 From the discussion on 11/23/11 with the NRC, the primary areas of interest by the NRC were associated with the completeness of LERF, HRA, and Uncertainty elements of the Fire PRA at the time of the Peer Review. The F&Os associated with these three elements are summarized in the following paragraphs.

Large Early Release Frequency (LERF) Element Four of the findings identified were associated with LERF calculations.

The first issue is associated with bypass paths missing from the fault tree. These bypass paths had been reviewed and screened as a large release path. However, the penetration screening was not sufficiently detailed in the report documentation at the time of the peer review and was corrected based on the peer review comment.

The next two findings involve the reasonableness review of the internal events LERF and Fire CDF and LERF. The internal events LERF cutsets were reviewed for reasonableness and the LERF results were compared with the other Entergy PWRs, but additional documentation is needed to clarify that the reviews were performed, and this can be done when the Fire PRA documents are next updated.

The Fire PRA cutsets were reviewed for CDF and LERF at the time of the peer review. However, Entergy was planning on enhancements to reduce Fire CDF and LERF. Multiple cutset reviews and uncertainty evaluations were performed on the updated fire results following the Peer review. These enhancements are consistent with the ASME/ANS PSA Standard HLR-QU-D and HLR-FQ-E requirements.

The final finding was associated with review of LERF sources of model uncertainty. At the time of the Peer Review, a quantitative assessment of Fire LERF uncertainty had not been performed. This addition for Fire LERF Uncertainty follows the methodology used for Internal Events which has already been reviewed by other Peer Reviews and therefore does not represent a change in the methodology used to perform the LERF evaluation.

Human Reliability Analysis (HRA) Element Five of the findings identified were associated with Human Reliability Analysis.

The first finding involved a self-identified HRA F&O. A conservative screening value for select HRA events was initially used. However, cutset reviews indicated a detailed HRA evaluation was needed. This HRA was subsequently evaluated (after the Peer Review) using the same HRA methodology as was used for other HRA events.

The second finding involved a review of fire procedures to identify fire-specific HRAs. The Operator Manual Actions (OMAs) in the Fire Abnormal Operating Procedures (AOPs) were reviewed and none are included in the Fire PRA as necessary to reduce significant fire scenarios for the transition from Appendix R to NFPA-805 (See WF3 805 LAR Attachment G).

Waterford 3 (WF3) Fire PRA Peer Review Position Paper Page 3 of 4 The third finding involves the more detailed review of instrumentation for impact of fire on credited operator actions. The instrumentation review was enhanced after the Peer Review with simulator runs and operator reviews to more realistically address the impact of instruments. If the failure of an instrument prevents a cue required to perform an operator action, then the operator action would be considered failed in the fire scenarios involving that instrument. This issue affected only the credit for HRAs in areas where the instrumentation is affected by the fire and therefore does not represent a Fire PRA Methodology change.

The next finding involves a feasibility review of operator actions that are primarily performed in the control room but have some ex-control room actions.

Additional discussions were conducted with plant operators to verify feasibility before the Peer Review but were added to the documentation after the Peer Review.

The final finding involves new combinations of operator actions that had not been evaluated at the time of the peer review. The Entergy HRA methodology ensures that a conservative HRA value is applied so that these new combinations are not missed. This same method is also used for internal event HRAs and has been determined by previous Peer Review teams to meet the PSA Standard.

Uncertainty Element Five of the findings identified were associated with uncertainty analyses. Uncertainty analysis consists of two parts: parametric uncertainty and modeling uncertainty.

Parametric uncertainty involves running a Monte Carlo code to determine the upper and lower bounds of CDF or LERF given the uncertainty in basic event probabilities.

Modeling uncertainty involves the impact of changing certain assumptions in various model elements on CDF or LERF. A change in assumptions would not be classified as a change in Methodology.

The first finding involved the discussion of uncertainties associated with plant partitioning. A qualitative uncertainty evaluation was added after the Peer Review to the Fire PRA documentation to address this comment. This qualitative review of the potential impact of an assumption on the Fire PRA results does not represent a methodology change.

The next two findings involved the discussion of uncertainties associated with the fire modeling. Waterford included a qualitative discussion of uncertainty for the generic fire modeling technique used. Adding this discussion is only a qualitative evaluation and does not represent a methodology change.

The fourth finding involved the documentation of the truncation sensitivity. A draft of this process was provided for the peer review team. The final convergence evaluation results were added to the documentation after the Peer Review as part of the documentation to support the LAR [3].

The final finding involved a review of model assumptions for impact of these assumptions on the Fire PRA results. Most of these assumptions were qualitatively evaluated. The most important assumptions were reviewed quantitatively in the final model documentation for the LAR. The additional

Waterford 3 (WF3) Fire PRA Peer Review Position Paper Page 4 of 4 assumptions or sensitivity cases added would not be considered a methodology change.

Conclusions The Waterford 3 Fire PRA model was reviewed per the ASME Standard and F&Os from the review have been addressed or dispositioned for the NFPA 805 application per the ASME PRA Standard [2] and the NEI 07-12 peer review guidelines [4]. Each of the subject findings at Waterford 3 are considered to be Maintenance items. A focused scope peer review is therefore not needed for these changes per the ASME PRA Standard [2]

and the NEI 07-12 peer review guidelines [4].

References

1. Attachment to LTR-RAM-II-11-003, Fire PRA Peer Review of Waterford Steam Electric Station Unit 3 Fire Probabilistic Risk Assessment against the Fire PRA Standard Supporting Requirements from Section 4 of the ASME/ANS Standard, February 23, 2011

2. ASME/ANS RA-Sa-2009, Addenda to ASME/ANS RA-S-2008, Standard for Level 1 / Large Early Release Frequency Probabilistic Risk Assessment for Nuclear Power Plant Applications.

3. W3F1-2011-0074, License Amendment Request to Adopt NFPA 805 Performance-Based Standard for Fire Protection for Light Water Reactor Generating Plants (2001 Edition), November 17, 2011.

4. NEI 07-12, Revision 1, Fire Probabilistic Risk Assessment (FPRA) Peer Review Process Guidelines, May 2010.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change CS-A3-01 Cables Associated Closed The instruments and cables associated with permissives and The mapping of all items on the SSEL was reexamined with Maintenance This issue involves mapping of instrument and cable faults to PRA with Interlocks interlocks do not appear to have been comprehensively addressed particular attention to instruments to ensure that their failure modes. Instrument and cable mappings are performed Other Affected SR in the PRA. Starting interlocks for pumps and breaker closure or consequential impacts have been properly linked to the PRA based on discussions with Fire Protection and Electrical ES-B4 tank interlocks that open or close valves or flow switches that start model. The methodology and results of the analysis are Engineers. Additional mapping are performed in accordance with pumps all could have fire effects that would adversely affect the documented in the Waterford 3 Component and Cable Selection the peer reviewed methodology.

success of various system functions. Report (R0247070001.02, Revision 2).

Specific rational of impacts on permissives and interlocks is documented for several components/ cables in the MSO Expert Panel discussions (Appendix I).

CS-B1-01 Electrical Closed Electrical coordination is addressed in the scenario development The Supplemental Coordination Evaluation has been added to Maintenance This issue is primarily a documentation issue and therefore does Coordination report (R0247070001.06 Appendix E). Appendix E of Appendix E of the Waterford fire PRA Scenarios Report not represent a methodology change. Electrical coordination of R0247070001.06 provides information concerning electrical (R0247070001.06, Revision 2, Appendix E). the SSD buses was include in the report at the time of the peer coordination. However, it is incomplete because the supplemental review. Supplemental electrical coordination of non-SSD buses coordination evaluation is missing from the document. The Waterford 3 FPRA meets the Category II/III requirements for was available at the time of the peer review, but the Preliminary coordination review has been performed and exists in CS-B1. documentation was missing from the report. The method for an email (though not formally documented). reviewing the non-SSD buses is the same as the SSD buses. This gap has now been closed.

Complete Appendix E of R0247070001.06 by adding the supplemental coordination analysis discussion. When this is completed, the category for this SR will be category II/III.

ES-A2-01 Documentation of Closed Breaker coordination does not appear to be properly modeled in The Supplemental Coordination Evaluation has been added to Maintenance This issue is primarily a documentation issue and therefore does Breaker the PRA. The SSD analysis is credited to ensure there is breaker Appendix E of the Waterford FPRA Scenarios Report not represent a methodology change. Electrical coordination of Coordination coordination, but the actual fire induced failure of the circuits that (R0247070001.06, Revision 2, Appendix E). the SSD buses was include in the report at the time of the peer ensure the coordination do not appear to be modeled. The review. Supplemental electrical coordination of non-SSD buses configuration of the plant requires the concurrent failure of 2 DC The Waterford 3 FPRA meets the SR requirements for ES-A2. was available at the time of the peer review, but the busses for this loss of coordination; however, the documentation documentation was missing from the report. The method for of this is insufficient. A failure of the over-current trip relays on a reviewing the non-SSD buses is the same as the SSD buses. This bus could allow a fault to transfer to the upstream bus. gap has now been closed.

The documentation should explicitly discuss the breaker coordination with respect to the PRA model and how the process used ensures comprehensive review.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change ES-A3-02 Inadequate Closed The loss of DC does not appear to be adequately addressed in the The impact of loss of DC power on the ability to trip the RCP Maintenance This issue addresses the discussion of Fire PRA treatment of DC Evaluation of Loss fire PRA. For example a failure of DC to supply control power to breakers is addressed in the Waterford 3 FPRA. The details are Control Power, particularly for RCPs breaker tripping. A detailed Other affected SR of DC the RCP breakers would inhibit the operator action to trip the documented in the Waterford 3 Fire Probabilistic Risk Assessment review of the power cables and overcurrent protection found that CS-A3 Reactor Coolant Pumps (RCPs) in a loss of seal cooling scenario. Fire Scenarios Report (R0247070001.06, Revision 2). Appendix no modifications to existing fire scenarios are needed to address This was compensated for by a spurious start of the RCPs which G in the report - Loss of Overcurrent Protection Review this finding. Therefore, the impact of this finding is primarily on the would affect the same state in the model. Similarly, a loss of DC specifically addresses the impact of losing DC power. This review documentation of this review and does not represent a power could potentially transfer a fault due to inhibition of in Appendix G specifically addresses FPRA Treatment of methodology change.

coordination. The plant has redundant DC supplies to the two Switchgear DC Control Power.

breakers which makes this failure less probable. However, additional documentation is required to clarify the issue. The fire effects on DC could adversely affect coordination as well as remote operation of breakers.

ES-B1-01 Inclusion of High Closed The process and documentation did not demonstrate that all high The Component and Cable Selection Report (R0247070001.02, Maintenance This issue addresses cable routing for risk important components Risk-Importance risk-importance components identified in the internal events PRA Revision 2) discusses the method of Identifying additional that are not included in the SSEL. Components without specific Components had been considered. components from the internal events PRA (Section 2.7). cable routing are assumed failed for the Fire PRA. Credit for Additionally, a sensitivity analysis was completed to evaluate the these components is performed on an as needed basis to Demonstrating and documenting that all high risk-importance impact of components with unknown cable routing (typically enhance the fire scenarios in accordance with the peer reviewed components identified in the internal events PRA had been components in the PRA model, but not in the SSL). methodology. This does not represent a methodology change.

considered systematically.

The methodology applied and the sensitivity analyses together provide adequate justification that all high risk components have been included in the FPRA.

ES-B2-01 LERF/Bypass logic Closed Section 2.4 of the Equipment Selection notebook and Appendix C The Component and Cable Selection Report (R0247070001.02, Maintenance This issue is primarily a documentation issue. All penetrations Missing from Fault documents the review of containment penetrations to identify Revision 2) provides sufficient detail on penetrations to disposition listed in the Finding are screened out but were not so noted in the Tree potential containment bypass paths for LERF evaluation. Valves this finding. Appendix C of the report is a Containment documentation. Therefore this issue does not affect the results.

not screened were to be added to the Fire PRA model. The Fire Penetration Review. In this review, the penetrations listed in This does not represent a methodology change.

PRA fault tree (WF3-L2-Toplogic_Mgd_Fire-f.caf) was reviewed to Finding have been screened.

determine if valves associated with un-screened penetrations were included. The review determined that valves associated with un- The details provided in the Component and Cable Selection screened penetrations 27, 32, 33, 34A/B, 40, 41 and 44 were not Report support a Category II classification for ES-B2.

included in the Fire PRA fault tree. The self assessment noted that this may be because the documentation is not up to date with the current PRA fault tree, but if that is the case no documentation of these changes have been noted. Multiple paths for LERF exist from spurious actuations which have are not included in the Fault Tree (FT) and could impact LERF analysis in fire scenarios.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change ES-C1-01 Instruments for Closed The instrumentation used by the operators has been addressed by In detailed HRA analysis and development for the Fire PRA, each Maintenance This issue is primarily a documentation issue and therefore does Human Actions utilizing instruments in the SSD analysis and relying on action was reviewed. During this review cues were identified for not represent a methodology change. The instrumentation redundancy of instruments to ensure that proper cues are utilized. each credited action. For each credited action, directly available associated with annunciators and indications required for operator This approach could impede the operator's ability to instruments were mapped for each HFE in the model. actions were reviewed to determine that redundant instruments simultaneously combat a fire and transient. are available to allow for credit for the operator actions. No single Appendix D of the Waterford3 Fire Probabilistic Risk Assessment instrument failure was identified that would prevent the operators For significant operator actions, the knowledge of what Quantification Model Preparation and Database Development from performing an action modeled in the PRA. Therefore this instrumentation fails allows the operator to more readily assess Report (R0247070001.03 - Revision 2) contains the relevant HFE issue does not affect the results.

the transient. development details, including instruments credited for cueing operators.

ES-C2-01 Documentation of Closed Component and Cable Selection Report R0247070001.02, Details of the simulator review and additional operator insights are Maintenance This issue is primarily a documentation issue and therefore does Operator Review of Revision 0 in Section 2.6 states, "An instrumentation review was documented in Appendix E Operator Interview Results of the not represent a methodology change. The Fire Quantification Other Affected SR Single Instrument conducted using the simulator and operators to identify single Waterford 3 Fire Probabilistic Risk Assessment Quantification report reviewed by the peer review discussed the performance of HRA-A3, HRA-A4 Vulnerabilities. instrument reliance and single indication/instruments whose Model Preparation and Database Development Report operator interviews and simulator observations. However, the malfunction would cause operators to take action that would result (R0247070001.03 - Revision 2). This Appendix outlines the documentation did not include the information gained from these in un-recoverable states. No single instrument vulnerabilities were purpose, methods, and results of detailed discussions with several reviews. Appendix E was added to the Fire Quantification report identified." operators and the documentation of a simulator observation. that includes the information from the interviews and observations.

No model or methodology changes were required to address this However, there is no documentation or discussion of this activity. The details provided in the listed attachment support a Category II finding.

Engineering standard EN-FP-S-008- Multi has a process for classification for ES-C2.

reviewing indication needs for post fire in the simulator in 5.3.4, but this does not specifically address spurious indications that cause unwanted actions.

What is needed to meet Category II for this SR is to develop a process for how various indications are reviewed and screened and then considered for inclusion into the FPRA model. There is a sample process that the PWROG did for ERGs for Westinghouse sites, this process is more detailed than required for meeting this SR for this application, but does show the process. No evidence other than statement that a simulator walkdown was performed.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change FQ-A3-01 Self Healing Circuit Closed Self Healing circuit failure likelihoods are applied in various fire The FPRA Fire Scenarios Report (R0247070001.06, Revision 2) Maintenance This issue is primarily a documentation issue, and therefore does Failures scenarios in the Fire Scenario Selection Notebook. The use of discusses credited application of self-healing hot shorts (Section not represent a methodology change. No self-healing of hot shots Other Affected SR self healing factors in every scenario should be evaluated for 10.5). was added or removed following the peer review. The CF-A1 success criteria and documented why the use of the self healing is documentation was updated to clearly state the two applications of applicable. However, there is no documentation for the scenarios The listed section provides details for each case (only 2 scenarios self-healing and the basis for the timing. No fault tree or evaluated for the cables used. This is also true for manual credit self healing). The details in the revised Fire Scenario Report quantification changes were made as a result of credit for self-exclusion of components noted - exclusion notation is often include judgments used in applying method and calculated time healing. Therefore, no methodology changes were made to inadequate. Requirement for developing a basis for all impacts to available for given scenarios. address this finding.

the fire scenarios quantification factors. This has a clear impact on scenarios included and non-conservative factors could be The details provided in the listed document support a Category II/II applied to the scenario. classification for SR CF-A1. The listed details on circuit failure also meet all listed SR requirements for SR FQ-A3.

Review where self healing and exclusion is used and document clearly why each case is correct for the scenario. Example - a Reactor Head Vent valve when self healed would return to closed

- reference analysis on why the time to self healing with the vent valve open will not affect success criteria for that scenario. While basis for exclusion is noted cryptically, it needs a more robust discussion to allow understanding of basis.

FQ-B1-01 Truncation Closed The Fire PRA Summary Report does not include documentation of Uncertainty and sensitivity methods and results as well as formal Maintenance A convergence study was provided to the peer review team but Sensitivity the process of truncation sensitivity performed. Back Referenced convergence evaluation are documented in the Waterford 3 Fire was not included in the Summary Report. Model convergence SRs QU-B3 and QU-B4 (referenced in FQ-B1) require a truncation Probabilistic Risk Assessment Summary Report (Report involves quantifying the model at various truncations limits (1E-11, sensitivity to be documented. 0247070001.07, Revision 2). Section 2.14 of the Summary Report 1E-12, 1E-13) until the difference between the CDF or LERF is documents the results of the convergence evaluation and displays sufficiently small to indicate that the further reduction in truncation Document the apparent draft truncation sensitivity was provided to CDF and LERF values at a range of truncation settings. would not impact the result. The quantification methodology does the review team on EXCEL Spreadsheet. not change as the truncation is reduced. The convergence study The details provided in the Summary Report meet all listed was updated after the CDF and LERF were updated for the final requirements for SR FQ-B1. report and therefore does not represent a methodology change.

FQ-C1-01 Dependencies on Closed The WSES3 Fire Probabilistic Risk Assessment Quantification The Quantification Model & Database report (Report Maintenance Some combinations of HRAs were not evaluated at the time of the Combinations of Model Preparation and Database Development report 0247070001.03, Revision 2) has been updated. The updated peer review. However, the methodology for performing these HFEs (R0247070001.03), Revision 0 section 5 uses verbs in the future document does not use future tense verbs, and clearly explains evaluations is identical to the method evaluated in the internal tense implying that certain analysis will be done at some future how completed analysis items were done and documented. events PRA and evaluated for the HRA combinations in the Fire date. One of these future actions is the analysis of combinations Section 5.2.2 or the report and Appendix B address joint human PRA at the time of the peer review. If an HRA combination has in the scenario cutsets. To date, the dependency analysis applies failure actions and their associated dependencies. not been explicitly evaluated, the conservative HEP associated various factors to combinations that appeared in the internal with a subset of the HRAs is applied.

events PRA only. Doing a dependency analysis is required to The details provided in R0247070001.03 Revision 2 fulfill the listed meet the ASME/ANS standard. Supporting Requirements for FQ-C1.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change FQ-D1-02 Internal Events Closed There has been no detailed assessment of LERF cutsets from the A reasonableness review was completed on the internal events Maintenance This issue is primarily a documentation issue and therefore does LERF internal events LERF model per peer review finding LE-F1-01 and LERF results in response to the internal events peer review not represent a methodology change. The Entergy PSA staff has LE-F3-01 note that no review has been documented. This would (though it was not well documented in time for the Fire RPA performed a detailed review of the LERF results for both the include documentation of a review of LERF cutsets. Without this review). As documented in response to LE-F1b-01 in Attachment internal events model and the Fire PRA. The detailed review review it cannot be confirmed per the standard the results of the U of this LAR, the Waterford PRA documentation on this issue will included a review of the top cutsets for each sequence and some LERF model are valid. The resolution noted is a comparison with be expanded and clarified. However, the inclusion of additional review of non-significant cutsets to ensure that the cutsets a similar plant results, but this does not meet the intent of the documentation for the PRA does not impact the FPRA results. represent the as-build, as-operated plant. However, the internal Finding from the internal peer review. events LERF review had not been included in the LERF Review the internal events cutsets and document reasonableness Additionally, Appendix D of the Summary Report (Report documentation. The methods for reasonableness review for LERF of the cutsets. If changes are required per this review to the LE 0247070001.07, Revision 2) provides details on sources of are the same as the review for CDF.

model update that used for Fire. uncertainty for each individual task in the FPRA development.

This Appendix also contains a parametric uncertainty analysis of the FPRA CDF results (including calculated uncertainty bounds).

FQ-E1-01 Reasonableness Closed There has been no detailed review of reasonableness of CDF or Following completion of the quantifications, a series of Entergy Maintenance Entergy was planning on updating the Fire PRA documentation to review LERF cutsets whether significant or non-significant outside of review meetings was conducted to review the cutsets for selected incorporate peer review comments. Therefore, some reviews reducing high CDF cutsets. This was acknowledged by the fire initiating events. The results are documented in the Summary were not documented at the time of the peer review. This has Entergy team. The reasonableness review needs to be performed Report (0247070001.07, Revision 2), Section 2.14 and Appendix been done. The method for reviewing cutsets for Fire PRA is the and documented. D. same as the reasonableness review for internal events CDF and LERF. No new methodologies were employed.

FQ-E1-02 LERF Uncertainty Closed The LERF sources of model uncertainty and related assumptions Appendix D of the Summary Report (Report 0247070001.07, Maintenance LERF uncertainty for Fire PRA is included in the current Summary have not been assessed at this time Revision 2) provides details on sources of uncertainty for each Report. The methodology for performing parameteric and individual task in the FPRA development. This Appendix also modeling uncertainty for internal events and Fire PRA are identical contains a parametric uncertainty analysis of the FPRA CDF and therefore does not represent a methodology change.

results (including calculated uncertainty bounds). Additionally, section 3.2 of the Summary Report includes a qualitative analysis of uncertainty associated with LERF.

FSS-B2-01 Main Control Room Closed Main Control Room (MCR) abandonment has been modeled with The Waterford 3 Fire Probabilistic Risk Assessment Fire No Change NUREG/CR-6850 allows for a conservative point estimate of abandonment eleven scenarios (RAB1A CRA1M through CRA5M, CRA1S Scenarios Report (R0247070001.06, Revision 2) Sections 13.2.1 CCDP for MCR abandonment. Since MCR abandonment cases modeling through CRA5S, and CRA5T). Fire modeling using CFAST was and 13.2.2 discuss the use of a 0.1 CCDP for MCR abandonment. were not dominant fire scenarios, Waterford used the conservative used to determine abandonment times. A CCDP of 0.1 is assumed A detailed human reliability analysis for shutdown outside of the approach for MCR abandonment. A detailed MCR abandonment for all of the abandonment scenarios. This is justified based on MCR does not exist. Shutdown outside the MCR was judged to evaluation could be considered a PRA Upgrade if Waterford were combination of the HEPs and random failures associated with the result in a relatively high CCDP. This judgment was due to the to change from the point estimate for future revisions. See Table remaining plant capability following MCR abandonment and is combination of the HFEs, random failures associated with the V-2 of the Waterford NFPA-805 LAR for the acceptability of CC-I judged to result in a relatively high CCDP. The non-suppression of remaining plant capability, and the damage to other systems by for this Supporting Requirement.

fire in the MCR is also credited for these scenarios. This fire. Applying a 0.1 CCDP for such cases is acceptable for methodology is judged to result in bounding results. Bounding addressing MCR abandonment and bound this scenario.

analysis rather than realistic analysis has been performed.

Provide realistic rather than bounding analysis to meet Category 2.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change FSS-C1-01 Multi-Point Heat Closed Two points or a range of heat release values were not assigned to The use of multi-point heat release rates (to meet the Category II No Change Waterford uses a generic fire modeling technique from Hughes Release Rates the ignition sources. SR) has not been performed for Waterford 3. Category I is and Associates that provides a slightly conservative fire damage Other Affected SR A lower overall CDF will likely be achieved by using a two point acceptable for the application. These results are comparable to assessment. Change to detailed fire modeling could be FSS-C2, FSS-C3 analysis or additional fire modeling to represent HRR profiles from the results of more detailed fire modeling. considered a PRA Upgrade and if done, may warrant a re-review.

ignition thru burnout and the corresponding probabilities of See Table V-2 of the Waterford NFPA-805 LAR for the damage. Section 14 of the Fire Scenarios Report discusses the use of acceptability of CC-I for this Supporting Requirement.

generic fire modeling versus detailed fire modeling and FSS-D7-01 Plant specific Non- Closed As noted in Section 8.0 or R024707001.06 Rev 0, when applied to Section 8.1 of the Fire Scenarios Report (R0247070001.06, Maintenance Waterford reviewed the maintenance history of the automatic suppression a scenario NUREG-6850 non-suppression probability values were Revision 2) documents a plant specific analysis of fire suppression suppression systems credited in the Fire PSA. This review did not Probability used in the analysis. system failure probabilities. This analysis includes a review of identify any excessive maintenance that would impact the generic maintenance history and unavailability. This section provides values from NUREG/CR-6850. If the probabilities were To move from CC-I to CC-II, specific WSES maintenance history adequate details to the meet the Category II SR requirements for significantly different, a Bayesian update could be performed review to assess outlier behavior is to be documented. FSS-D7. similar to the methods for plant-specific data in the internal events PSA.

FSS-E3-01 Quantitative Closed Only qualitative discussions were provided with respect to the Complete discussion of uncertainty and sensitivity is included in No Change Waterford is using generic fire modeling which provides a Uncertainty uncertainty intervals for the fire modeling parameters. the Summary Report (0247070001.07, Revision 2). However, conservative assessment of fire damage. The NFPA-805 consistent with industry discussions with ACRS, quantitative methods allow for qualitative uncertainty in the fire damage Provide quantitative uncertainty intervals. treatment of uncertainty intervals for fire modeling parameters is scenarios. See Table V-2 of the Waterford NFPA-805 LAR for the not required at this time. acceptability of CC-I for this Supporting Requirement.

FSS-F2-01 Structural Collapse Closed Criteria for structural collapse or non-collapse was not provided. The Fire Scenarios Report (R0247070001.06, Revision 2) Maintenance This issue is primarily a documentation issue. Fire Scenario Analysis Only judgment statements were provided and these statements do provides details on the potential for structural collapse. Section documentation was expanded to discuss structural collapse. The Other Affected SR not appear to reflect reality. 10.3 of the report has been updated since the peer review with methodology and analysis were performed in accordance with FSS-F3 expanded discussion and a quantification screening value. NUREG/CR-6850.

Re-perform the analysis to address the situation where a turbine building collapse occurs due to a large turbine lube oil fire. Revise The details provided in the listed document support a Category the documents to eliminate the implication that failure of structural II/III classification for both FSS-F2 and FSS-F3.

steel is not a credible event.

FSS-H2-01 Plant specific Closed Section 4.0 details damage criteria used in the Fire Scenario Plant specific damage thresholds were not developed for the No Change Waterford uses a generic fire modeling technique from Hughes damage criteria Report. No cases of where plant specific thresholds or damage WSES FPRA. This is in line with the application of the Hughes and Associates. Change to detailed fire modeling could be mechanisms were used. generic fire modeling approach (no detailed fire modeling was considered a PRA Upgrade and if done in the future, may warrant done). a re-review.

Perform plant specific evaluations of the basis for target damage, including suppression considerations where fire sprinkler systems The Waterford 3 FPRA is acceptable as a Category I SR for FSS-are installed. H2.

FSS-H5-01 Fire Scenario Closed Documentation SR FSS-H5 requires analysis of parametric The Waterford 3 FPRA uses generic fire modeling for individual No Change Waterford uses a generic fire modeling technique from Hughes Uncertainty uncertainty for results of Fire Scenarios results. No fire scenarios. With this approach, the Waterford FPRA only and Associates. Change to detailed fire modeling could be Other Affected SR documentation of this type of analysis exists. meets Category I of SR FSS-H5. This approach is based on the considered a PRA Upgrade and if done in the future, may warrant FQ-E1 Hughes generic fire modeling approach. a re-review. See Table V-2 of the Waterford NFPA-805 LAR for the acceptability of CC-I for this Supporting Requirement.

Section 14 of the Fire Scenarios Report (R0247070001.06, Revision 2) provides the basis for the approach and the

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change FSS-H10-01 Walkdown Closed Walkdown documentation is weak. Many references are made in The documentation of the walkdowns is provided in the Waterford Maintenance This issue is primarily a documentation issue and therefore does Documentation the reports to walkdowns, so a significant improvement in the 3 Fire Probabilistic Risk Assessment Plant Partitioning and Fire not represent a methodology change. Walkdown documentation Other Affected SR walkdown documentation is highly recommended. Ignition Frequency Development Report (R0247070001.01, was enhanced based on the peer review comment.

FSS-D10, FSS-D11 Revision 1). This document includes specific walkdown Provide walkdown documentation that has been recorded using summaries of ignition sources with an originator and reviewer consistency in the level of detail, transcribed using a QC process, including dates as transcribed from field notes. Additional detail on and verified as to accuracy. the methodology used for the walkdowns is also documented in Appendix A of the Fire Scenarios Report (R0247070001.06, Revision 2).

HRA-A2-01 HRA Closed Section 5 of R0247070001.03 discusses the Human Reliability The Waterford 3 Fire Probabilistic Risk Assessment Quantification Maintenance This issue is primarily a documentation issue. RHFPUMPOFP did Documentation Analysis for the WSES fire PRA. As noted in 5.2.3, only one fire Model Preparation and Database Development (R0247070001.03, not have a detailed HRA evaluation since it did not occur in Other Affected SR specific HFE was identified, RHFPUMPOFP. This is an action Revision 2) includes the HRA analysis for the Waterford 3 FPRA. internal event cutsets. The Fire PRA updated the probability this HRA-B2, HRA-D2 called out in procedure OP-901-524. This HFE had already been This HRA analysis has been updated to include reevaluation of event based on fire. No other fire specific HRAs were needed for included in the Full Power Internal Events (FPIE) model with a feasibility and specific calculations for credited HFEs. The NFPA-805. The methodology used to develop this operator action screening value. However, no HRA calculation sheets could be updated document specifically includes details on RHFPUMPOFP and address fire impacts on other HRAs is consistent with the located in the FPIE or in the FPRA documentation. development. methodology used for the internal events HRA.

Note that this issue was self-identified by WSES but they have not resolved it yet.

HRA-A4-01 Operator/Training Closed No documentation could be found (e.g., Operator Interview Documentation of HRA development has been updated to include Maintenance This issue is primarily a documentation issue. No fire-specific Review of Credited Sheets) that shows a review of the procedures associated with interviews with plant operations personnel. Appendix E Operator HRAs were identified. The impact of fires on operator actions had HRAs actions identified in SRs HRA-A1 and HRA-A2 has occurred with Interview Results has been added to the Waterford 3 Fire been discussed with operations but was not included in the Fire plant operations or training personnel to confirm that the Probabilistic Risk Assessment Quantification Model Preparation PRA reports. Closing this gap after the peer review does not affect interpretation is consistent with plant operational and training and Database Development (R0247070001.03, Revision 2). the results for this application. The methodology used to update practices. Documentation of Operations review of applicable the HRA worksheets based on fire impacts is consistent with the FPRA HFEs is necessary to assure proper application of HEP internal events HRA.

values.

Document a review of the Fire PRA actions identified in SRs HRA-A1 and A2 has occurred with Operations and Training.

HRA-C1-01 Fire Impacts on Closed The detailed analysis of the fire affected HFEs should be The FPRA HRA development was reviewed and updated to Maintenance This issue involves an operator action that is not needed for the HFE Actions developed more addressing the fire effects on the action. assure that sufficient cues are available for all credited operator fire scenarios. Documentation was updated to clarify the HRAs Particular attention should be focused on the required actions (HFES). No new instruments were identified that need to required for fire scenarios and therefore does not represent a instrumentation. Also, the screening of "EHFSTRBATP" states be added. The Waterford 3 Fire Probabilistic Risk Assessment methodology change.

that SBO is not applicable to fires and sets the probability to "0" for Quantification Model Preparation and Database Development the HFE. A fire induced SBO is applicable and the screening (R0247070001.03, Revision 2) report provides the HFE should be corrected. It is not clear if more instruments need to be development details.

cable traced and added to the model. With respect to EHFSTRBATP, the event is not present in the portion of the PRA model used for the Fire PRA. As such, its probability has no impact on the results.

HRA-D1-01 HRA Feasibility Closed Table D-1 in R0247070001.03, Revision 0 lists five HFEs which As part of the HRA review update (documented in the Waterford 3 Maintenance This issue is primarily a documentation issue and therefore does had detailed analysis applied. Some of these HFEs have ex-main Fire Probabilistic Risk Assessment Quantification Model not represent a methodology change. The feasibilities of the Other Affected SR control room actions embedded. There is no evidence that a Preparation and Database Development R0247070001.03, HRAs for fire had been performed before the peer review but had PRM-B11 feasibility analysis has been performed on these actions. The Revision 2) operator interviews were conducted to verify feasibility. not been documented.

demonstration of feasibility is a requirement of the ASME HFEs were reviewed to identify embedded Actions. Appendix E of standard. the documents contains the details of the operator interviews and Table E-2 documents the feasibility of each credited action.

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change HRA-E1-01 Conflicting Closed Documentation issues with R0247070001.03, Revision 0 section The Waterford 3 Fire Probabilistic Risk Assessment Quantification Maintenance This issue is primarily a documentation issue and therefore does Documentation of 5: Page 5-8 has a paragraph explaining that ex-mcr HFEs are set Model Preparation and Database Development Report not represent a methodology change. The write-up was not clear HRA Methods to true and then the risk significant HFEs are analyzed in more (R0247070001.03, Revision 2) has been updated (twice) since the as to the method being performed. The actual method was detail. If the HFE is set to true, then it would not show up in the peer review. The updated document provides a clearer discussed with the peer team, and they agreed that the method cutsets. This paragraph needs a rewrite. Page 5-9 has a table description of the applied methodology for HFEs. used was appropriate.

explaining various treatments of HFEs in the model. Two of the Page 5-8 explains how screening HFEs are included in the model columns conflict; one recommends a course of action and the (and specifically how the FRANCs software uses them). The table resolution takes another course with no explanation of the on the following pages explains how specific HFE issues (not all differences. Also there should be some discussion about thermal HFEs) were treated.

hydraulic analysis on any new sequences.

IGN-A10-01 Uncertainty Closed Documentation of sources of uncertainty covers uncertainties A detailed uncertainty evaluation is documented in Appendix D of Maintenance This issue is primarily a documentation issue and therefore does Analysis associated with NUREG/CR-6850 Bin elements, and uncertainties the Summary Report (0247070001.07, Revision 2). This appendix not represent a methodology change. The preliminary uncertainty associated with the Bayesian update. No discussion on provides adequate technical and qualitative detail to satisfy the analysis had been performed but the documentation was not uncertainties associated with partitioning and weighting factor Category III Supporting Requirement for IGN-A10. complete at the time of the peer review. The uncertainty applications. The uncertainty analysis is incomplete. associated with partitioning is limited based on the multi-Documentation of weighting factor impact on uncertainty is not compartment analysis, which found no unscreened MCA and HGL Provide either a numerical uncertainty analysis or qualitative required by the standard. A discussion of partitioning impacts on scenarios.

discussion of other sources of uncertainty as required by the uncertainty is included in Table D-1 in Appendix D.

standard.

PP-A1-01 Partitioning Closed The station's switchyard, transformer yard, and Turbine building The switchyard has been incorporated into the existing YARD Maintenance This issue is primarily a documentation issue. The bases for Element Selection that houses switchyard relays, are in close proximity to each other physical analysis unit (PAU). The wall between the transformer partitioning these areas was discussed but not included in the Other Affected SR and Definition with no defined fire barrier between them. There is little discussion yard (XFMR YARD) and the Turbine Building (TGB) Switchgear documentation. The partitioning methodology was performed PP-B1 concerning why these three areas should be separate. The Fire Room is of concrete construction and is rated for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> (Ref consistent with NUREG/CR-6850.

Hazard Analysis does not discuss the TBG, or either Yard Dwg G1370). There are no fixed ignition sources directly adjacent boundaries as well. to the wall with the switchgear more than 8 feet away and transformers are more than 15 ft. away. Transient sources would

Table V-1 Fire PRA Peer Review - Findings and Observations SR Topic Status Finding/Observation Disposition Classification of Explanation Change PP-B7-01 PAU Barrier Closed Appendix D of H0247000017.01, R1 contains copies of the PAU Discussions of barrier observation were performed as part of the Maintenance This issue is primarily a documentation issue and therefore does Verification walkdown sheets. However, the walkdown which generated these FPRA effort and are documented in the Partitioning Report not represent a methodology change. The bases for PAUs has Walkdowns sheets was ignition source counting and did not look at confirming (R0247070001.01, Revision 1). In addition, efforts completed as been updated to discuss the barriers used for defining the PAU.

the location and characterization of the credited barriers. part of the Deterministic -NFPA-805 Transition work have been The procedures and methods used for establishing the PAUs was Discussions with WSES personnel indicated that the barrier referenced (and added to reference section). Additional discussed with the peer team. The team acknowledged that the confirmation walkdowns had been conducted as part of the safe procedures for continuing inspection as part of plant operation barriers were valid but the procedures needed to be included in shutdown program development. Furthermore, there were have also been added to the reference section including that for the documentation.

inspection procedures that verified barrier conditions on a routine Fire Rated Walls, Floors, and Ceilings, Fire Wrap Barriers, and basis. However, neither was found to be referenced in Fire Dampers.

H0247000017.01, R1. A check of the safe shutdown analyses did not immediately show documentation of the walkdowns of interest.

WSES needs to directly reference any procedures that they are using as the basis for the conditions and characteristics of the credited partitioning elements. For example, ME-003-006 is the procedure for of inspection for penetrations. Equivalent procedures cover fire doors, dampers and other credited partitioning elements. These procedures and their results should be referenced.

UNC-A1-01 Uncertainty in Closed QU-E1 and QU-E2 requires identification of sources of model Appendix D of the Summary Report (Report 0247070001.07, Maintenance Limited uncertainty evaluations were performed for the Fire PRA at Assumptions, CDF, uncertainty and assumptions. In general, WSES had an Revision 2) provides details on sources of uncertainty for each the time of the peer review. However, the method for determining

& LERF Results assumption section in each report. However, a simple search on individual task in the FPRA development. This Appendix also the assumptions and sensitivity cases is similar to the internal "assum" showed that there were many more assumptions than contains a parametric uncertainty analysis of the FPRA CDF events model. The method for identifying and performing were listed in the assumption sections. At the CC-I level, QU-E3 results (including calculated uncertainty bounds). Additionally, uncertainties is consistent with the method reviewed by the peer requires estimation of the uncertainty interval of the overall CDF section 3.2 of the Summary Report includes a qualitative analysis team for internal events. Besides, this addition does not represent results. WSES does not provide an estimation of the uncertainty of uncertainty associated with LERF results (i.e. a reasonableness a methodology change.

interval for CDF. QU-E4 requires that for each source of model review).

uncertainty and related assumption identified in QU-E1 and QU- Entergy concludes that the Waterford 3 FPRA contains sufficient E2, respectively; IDENTIFY how the PRA model is affected. details and analysis to address the uncertainty bounds for this fire WSES only identifies how the model is impacted for some of the induced CDF finding.

assumptions and sources of uncertainty. The Uncertainty and Sensitivity Matrix in Appendix D of R0247070001.07. Per LE-F2, WSES should review LERF contributors for reasonableness (e.g.,

to assure excessive conservatisms have not skewed the results, level of plant specificity is appropriate for significant contributors, etc.). There is no evidence that such a review was performed As a minimum, WSES needs to provide an estimate of the uncertainty bounds for the fire-induced CDF. WSES should also make an effort to capture all assumptions for each

Waterford 3 Attachment V - Fire PRA Quality Table V-2 Fire PRA- Category I Summary SR Topic Status FSS-B2 Main Control Room Abandonment The Waterford 3 Fire Probabilistic Risk Assessment Fire Scenarios Report (PRA-W3-05-006, Revision

0) Sections 13.2.1 and 13.2.2 discuss the use of a 0.1 CCDP for MCR abandonment. A detailed human reliability analysis for shutdown outside of the MCR does not exist. Shutdown outside the MCR was judged to result in a relatively high CCDP. This judgment was due to the combination of the HFEs, random failures associated with the remaining plant capability, and the damage to other systems by fire. The approach for applying a bounding 0.1 CCDP for is judged to be appropriate for these cases.

A Capability Category 1 is considered acceptable for the Fire PRA application.

FSS-C1 Use of Multi-point Heat Release Rate Section 14 of the Fire Scenarios Report discusses the use of generic fire as opposed to more detailed Treatment fire modeling. Waterford 3 applied the Hughes Associates Generic Fire Modeling Treatment. This treatment offers a means for incorporation of fire modeling into the fire PRA in a manner that eliminates the need for separate scenario specific analyses. While the results are slightly more conservative, they are consistent with the results of the more detailed fire modeling.

A Capability Category I is acceptable for this application.

FSS-C2 Peak Heat Release Rates Versus Time The use of fire growth curves are not part of the Generic Fire Modeling Treatments used at Waterford Dependent Fire Growth. 3. Section 14 of the Fire Scenarios Report discusses the use of generic fire modeling versus detailed fire modeling and justifies the approach for the Waterford 3 Fire PRA application.

A Capability Category I is considered acceptable for this application.

FSS-C3 Fire Development - Burnout/Growth/Decay The use of fire growth curves are not part of the Generic Fire Modeling Treatments used at Waterford

3. Section 14 of the Fire Scenarios Report discusses the use of generic fire modeling versus detailed fire modeling and justifies the approach for the Waterford 3 Fire PRA application.

A Capability Category I is considered acceptable for this application.

FSS-E3 Quantitative Uncertainty Complete discussion of uncertainty and sensitivity is included in the Summary Report (PRA-W3 007, Revision 0). However, consistent with industry discussions with ACRS, quantitative treatment of uncertainty intervals for fire modeling parameters is not required at this time.

FSS-H2 Plant Specific Damage Criteria Plant specific damage thresholds were not developed for the Waterford 3 FPRA. This is in line with the application of the Hughes generic fire modeling approach (no detailed fire modeling was done).

A Capability Category I for FSS-H2 is considered acceptable for the FPRA application.

FSS-H5 Fire Scenario Uncertainty The Waterford 3 FPRA uses generic fire modeling for individual fire scenarios. With this approach, the Waterford 3 FPRA only meets Category I of SR FSS-H5. Section 14 of the Fire Scenarios Report (PRA-W3-05-006, Revision 0) provides the basis for this approach and the justification for its use in the FPRA methodology.

A Capability Category I for FSS-H5 is acceptable for the FPRA application.

Page 16 Rev. 1