DCL-15-119, Response to NRC Request for Additional Information - National Fire Protection Association Standard 805, Enclosure, Attachment 1
| ML15316A008 | |
| Person / Time | |
|---|---|
| Site: | Diablo Canyon  |
| Issue date: | 10/15/2015 |
| From: | Allen B Pacific Gas & Electric Co |
| To: | Document Control Desk, Office of Nuclear Reactor Regulation |
| Lingam S | |
| References | |
| DCL-15-119, TAC MF2333, TAC MF2334 | |
| Download: ML15316A008 (36) | |
Text
The Enclosure, Attachment 2 Contains Security-Related Information -Withhold Under 10 CFR 2.390 Enclosure PG&E Letter DCL-1 5-119 Response to NRC Request for Additional Information - National Fire Protection Association Standard 805
- 1. PRA RAI-03, PRA RAI-01.h.01, & Follow-on Responses to NRC Request for Additional Information-National Fire Protection Association Standard 805
- 2. REFERENCES ATTACHMENTS:
- 1. List of acronyms used in this response for convenience.
- 2. Update to Attachments C, D, G, and W of the License Amendment Request (2.390).
- 3. Updates to the License Amendment Request as a Result of Updated Analyses, Editorial Changes, and Reference Corrections.
1 The Enclosure, Attachment 2 Contains Security-Related Information - Withhold Under 10 CFR 2.390 When separated from these attachments, this document is decontrolled.
Enclosure PG&E Letter DCL-15-119
- 1. PRA RAI-03. PRA RAI-01.h.01. & Follow-up Responses to NRC Request for Additional Information - National Fire Protection Association Standard 805 On June 26, 2013, PG&E submitted a LAR to adopt NFPA 805 (Reference 1).
On July 31, 2014, the NRC provided RAI's (Reference 2) regarding LAR 13-03, herein referred to as "LAR" or "the LAR." PG&E provided responses to NRC RAIs.
In Reference 3, 4, and 5, PG&E submitted 60-day, 90-day, and 120-day RAI responses, respectively, to the NRC RAIs (Reference 2).
Responses to the NRC questions on PRA RAI-03 and PRA RAI-01.h are provided below with follow-up responses directly following for RAIs impacted by the integrated analysis results..*
Attachment I of this enclosure provides a list of acronyms used in this response for convenience. of this enclosure provides an update to Attachments C, D, G, and W previously submitted in Reference 1, as a result of RAI responses including PRA RAI 3. of this enclosure provides an update to the LAR as a result of updated ana~lyses, editorial changes, and reference corrections. The changes also include those revisions previously included in PG&E responses to other RAIs (References 3, 4, and 5).
NRC PRA RAI-03.a Section 2.4.4.1 of NFPA -805 states that the change in public health risk arising from transition from the current fire protection program to an NFPA -805 based program, and all future plant changes to the program, shall be acceptable to the NRC. RG 1.174, "An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, "provides quantitative guidelines on core damage frequency, large early release frequency, and identifies acceptable changes to these frequencies that result from proposed changes to the plant's licensing basis and describes a general framework to determine the acceptability of risk-informed changes. The NRC staff review of the information in the MAR has identified additional information that is required to fully characterize the risk estimates.-
The PRA methods currently under review in the LAR include:
- PRA RAI l.a regarding removal of conservatism due to lack of cable routing 2
Enclosure PG&E Letter DCL-1 5-1.19 PRA RAI 1.b regarding modeling of common enclosure related circuits PRA RAI 1.fregarding credit taken for new RCP seal PRA RAI 1.g regarding exclusion of electrical distribution panels from the FPRA PRA RAI 1.h regarding MCR abandonment due to habitability PRA RAI 1.k regarding the inclusion of SOKC for internal and fire event related factors PRA RAI 1.1 regarding applying a minimum joint probability for HFEs PRA RAI 2. a regarding systems or actions needed for safe and stable state at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> PRA RAI 2.b regarding complete treatment of pre-initiators
-* PRA RAI 5 regarding Heat Release Rates lower than 317 kW for transient sources PRA RAI 6 regarding other disclosed deviations from acceptable PRA methods PRA RAl 8 regarding MCB modeling PRA RAI 9 regarding incipient detection PRA RAI 10 regarding fire damage effects from the opposite unit PRA RAI 11 regarding screening junction boxes as non-damaging ignition sources PRA RAI 18 regarding update to internal events PRA as a result of F&O dispositions PRA RAI 22 regarding CCDP developed for loss of control scenarios FM RAl 1.i regarding calculation of time to abandonment Please provide the following:
a) Results of an aggregate analysis that provides the integrated impact on the fire risk (i.e., the total transition CDF, LERF, ACDF, ALERF) of replacing specific methods identified above with alternative methods which are acceptable to the NRC. In this aggregate analysis, for those cases where the individual issues have a synergistic impact on the results, a simultaneous analysis must be performed. For those cases where no synergy exists, a one-at-a-time analysis may be done. For those cases that have a negligible impact, a qualitative evaluation may be done. It should be noted that this list 3
Enclosure PG&E Letter DCL-15-119 may expand depending on NRC's review of the responses to other RAIs in this document.
PG&E RESPONSE PRA RAI 3.a The results of the aggregate analysis that provides the integrated impact on the fire risk (i.e. the total transition CDF, LERF, ACDF, ALERF) reflecting PG&E's responses to and resolution of all RAIs listed above are provided in the update to Attachment W in the transmittal letter that accompanies this RAI response. The aggregate analysis also includes other RAIs as stated in PG&E Letters DCL-14-1 10, "One-Hundred-Twenty-Day Response to NRC Request for Additional Information - National Fire Protection Association Standard 805," dated November 26, 2014, and DCL-15-032, "Response to NRC Request for Additional Information - National Fire Protection Association Standard 805," dated February 25, 2015. The following RAIs were identified as having potential impact on the fire PRA model and were also incorporated into the aggregate analysis:
PRA RAI-01.c regarding screening of unground 3-phase hot short failures PRA RAI-01.d regarding circuit failure probabilities of Thermoset versus Thermoplastic cable PRA RAI-01.g.01 regarding discrepancy between the fire ignition Bin 15 count and per-cabinet frequency PRA RAI-14.c regarding VFDRs associated with loss of process instrumentation PRA RAI-1 5.01 regarding update of the risk-offset values in Tables W-4 and W-5 with the GEN Ill SDS FM RAI-01.c.01 regarding the flame spread rate for trays with mixed cables FM RAI-0i.h regarding the C FAST analysis and input parameters in Fire Zone 14-D All of the changes resulting from the issues stated in this RAI response were evaluated simultaneously in the aggregate analysis.
NRC PRA RAI-03.b Section 2.4.4.1 of NEPA -805 states that the change in public health risk arising from transition from the current fire protection program to an NEPA -805 based program, and all future plant changes to the program, shall be acceptable to the NRC.
RG 1.174, "An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis," "provides quantitative guidelines on core damage frequency, large early release frequency, and identifies acceptable changes to these frequencies that result from proposed changes to the 4
Enclosure PG&E Letter DCL-1 5-119 plant's licensing basis and describes a general framework to determine the acceptability of risk-informed changes. The NRC staff review of the information in the LAR has identified additional information that is required to fully characterize the risk estimates.
The PRA methods currently under review in the LAP include:
PRA RAI l.a regarding removal of conservatism due to lack of cable routing PRA RAI 1.b regarding modeling of common enclosure related circuits PRA RAI 1.f regarding credit taken for new RCP seal PRA RAI 1.g regarding exclusion of electrical distribution panels from the FPRA PRA PAl 1.h regarding MCR abandonment due to habitability
- PRA PAl 1.k regarding the inclusion of SOKC for internal and fire event related factors
- PRA RAI 2. a regarding systems or actions needed for safe and stable state at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />
- PRA PAl 2.b regarding complete treatment of pre-initiators
- PRA RAI 2. f regarding fire-induced flooding or sprays PRA RAI 4 regarding treatment of sensitive electronics PRA RAI 5 regarding Heat Release Rates lower than 317 kW for transient sources PRA RAI 6 regarding other disclosed deviations from acceptable PRA methods PRA PAl 8 regarding MCB modeling PRA RAI 9 regarding incipient detection PRA PAl 10 regarding fire damage effects from the opposite unit PRA PAl 11 regarding screening junction boxes as non-damaging ignition sources PRA PAl 18 regarding update to internal events PRA as a result of F&O dispositions PRA PAl 22 regarding CCDP developed for loss of control scenarios FM PAl 1.i regarding calculation of time to abandonment b) For each method (i.e., each bullet) above, explain how the issue will be addressed in 1) the final aggregate analysis results provided in support of the 5
Enclosure PG&E Letter DCL-1 5-119 LAR, and 2) the PRA that will be used at the beginning of the self-approval of post-transition changes. In addition, provide a method to ensure that all changes will be made, that a focused-scope peer review will be performed on changes that are PRA upgrades as defined in the PRA standard, and that any findings will be resolved before self-approval of post-transition changes.
PG&E RESPONSE to PRA RAI-03.b An explanation of how each RAI item is addressed is provided below for each method (i.e., each bullet) listed above in 3.b for: (1) the final aggregate analysis results provided in support of the LAR, and (2) the PRA that will be used at the beginning of the self-approval of post-transition changes. None of the changes made to the aggregate analysis model or the post-transition model as responses to these RAls are PRA upgrades, as defined in the PRA standard, and, therefore, no new (either focused or full scope) peer review will be required.
PRA RAI l.a Removal of conservatism due to lack of cable routing Disposition in the final aggregate analysis results (1) The following systems or functions are determined to have no impact or minimal impact on both the compliant plant and VFDR models and' therefore are not credited in the aggregate analysis: Makeup to the Refueling Water Storage Tank from the Spent Fuel Pool, 500 kV Back-feed, 12 kV non-essential power, and AMSAC.
(2) An assumption of "guaranteed failure" of the following systems in the compliant plant model could overestimate the risk of the compliant plant and therefore, underestimate the fire ACDF and ALERF risk of VFDRs; instrument air, main feedwater and condensate systems, containment fan cooler units, and containment spray system. The assumption of "guaranteed failure" of these systems is removed from the compliant plant model. No change to the variant model is made; this conservative assumption is maintained in the fire PRA model.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes (1) Makeup to the refueling water storage tank from the spent fuel pool, 500 kV back-feed, 12 kV non-essential power, and AMSAC provide no support function or reasonable alternate shutdown path to the credited SSD paths. Not crediting these systems or functions therefore, should have no or minimal impact on the risk characterization of a post-transition FP Program.
(2) This assumption will be maintained in the post-transition fire PRA model.
6
Enclosure PG&E Letter DCL-15-119 PRA RAI I1.b Modeling of common enclosure related circuits Disposition in the final aggregate analysis results Scenario development for four potential common enclosure cables previously identified (Reference PG&E Calculation 1 34A-DC-T, Revision 1) indicates that there is no credible fire which can cause a secondary fire with the potential damage to fire PRA cables or equipment beyond that included in the fire PRA circuit analysis and fire PRA model. Therefore, no new fire scenario is introduced in the fire PRA model or in the aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The disposition of PRA RAl 1.b does not require any change to the post-transition fire PRA model.
PRA RAI Il.f Credit taken for new RCP seal Disposition in the final aggregate analysis results At the time of this response, PG&E is committed to install the Westinghouse GEN Ill SDS into the RCPs. The GEN Ill SDS is modeled in the fire PRA model used for the aggregate analysis and its modeling is based on guidance in PWROG-14001-P, Revision 1.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes Once the Westinghouse GEN Ill RCP SDS are installed and operable, the post-transient PRA model will be verified in its modeling with the as-approved version of PWROG-14001-P before it is used in self-approval of post-transition changes.
The other hazard groups (i.e., Internal Events, Seismic Events, and Internal Floods) also will credit the as-built/as-operated GEN Ill RCP SDS.
PRA RAI l.g Exclusion of electrical distribution panels from the FPRA Disposition in the final aggregate analysis results During a confirmatory walkdown in response to PRA RAl I.g, two electrical cabinets could not be screened as robust as initially assumed. A new set of fire scenarios 7
Enclosure PG&E Letter DCL-15-119 originating from these cabinets were developed and included in the final aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include these new fire scenarios.
PRA RAI 1.hMCR abandonment due to habitability Disposition in the final aggregate analysis results (1) A new fault tree representing the MCR smoke purge system (function) was developed that includes the following: an operator action to switch the ventilation mode from Normal to Purge mode, unavailability of the purge system due to maintenance and testing, and failures of the SS0s required for the Smoke Purge mode of operation. This fault tree is integrated into the fire PRA model and is used in the aggregate analysis.
(2) The results of the updated MCR abandonment analysis performed in PG&E's response to FM RAI-01.i (Reference 5) confirmed that there are no cases where the abandonment criteria are reached for a transient fire in the MCR.
(3) The MCR abandonment scenarios (from a non-transient fire either due to loss of habitability or loss of control) are included in the final aggregate analysis.
Disposition in the PRA that will be used at the beginning of theself-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the MCR abandonment scenarios caused by either loss of habitability or loss of control.
PRA RAI Il.k Inclusion of SOKC for internal and fire event related factors Disposition in the final aggregate analysis results SOKC is addressed using the fire PRA model prepared for the aggregate analysis.
A comparison of the mean values with consideration of the SOKO and the point estimate values used in support of the LAR shows that the difference between the mean value of the ODE and LERF is slightly higher (less than 1 percent) than the point estimate CDF and LERF.
8
Enclosure PG&E Letter DCL-15-119 The SOKC has minimum impact on the mean value of the fire induced total CDF and LERF and therefore, the use of point estimate CDF and LERF as a surrogate for the mean value of the risk metrics in Attachment W of the NFPA 805 LAR is a reasonable approach.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes As demonstrated via a SOKC analysis, the SOKC effect on risk matrix is low and therefore, the point estimate values from the post-transition fire PRA model without any quantitative adjustments will be used for self-approval of post-transition changes.
PRA RAI 1.1 Aplyllingc a minimum joint probability for HFEs Disposition in the final aggregate analysis results The results of the fire HRA dependen~cy analysis are included in the aggregate analysis. The minimum joint HEP analysis, following the dependency analysis, did not identify any additional change to the fire PRA model used for the aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the results of the fire HRA dependency analysis. No change related to the minimum joint HEP is required.
PRA RAI 2.a Systems or actions needed for safe and stable state at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> Disposition in the final aggregate analysis results The final aggregate analysis includes the requirement for a supplemental water supply (i.e., raw water reservoir or fire water storage tank) to the AFW to achieve and maintain stable plant condition for a 24-hour mission time.
9
Enclosure PG&E Letter DCL-1 5-119 Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include a new top event modeling the long-term (i.e., 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> mission time) AFW cooling water supply requirement.
PRA RAI 2.b Complete treatment of pre-initiators Disposition in the final aggregate analysis results The final aggregate analysis includes all pre-initiators HFEs identified in response to F&Os SY-A16-01, SY-B15-01, HR-A1-01, HR-A3-01, HR-C3-01, and D3-01.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes As there are no excluded pre-initiators in the fire PRA model, which will be used for self-approval of post-transition changes, it is not necessary to establish a method or process of incorporating "excluded" pre-initiators into the fire PRA model before using the model for self-approval.
PRA RAI 2.c Actuation logic permissives and interlocks Disposition in the final aggregate analysis results The final aggregate analysis includes those permissives and interlocks identified in PG&E's response to NRC PRA RAI 2.c (Reference 4) as having potential risk impact.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the failures of those permissives and interlocks (Reference 4).
PRA RAI 2.f Fire-induced floodinq or sprays
- 1. Disposition in the final aggregate analysis results The final aggregate analysis includes fire-induced flooding scenarios identified in PG&E's response to NRC PRA RAl 2.f (Reference 4).
10
Enclosure PG&E Letter DCL-1 5-119
- 2. Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the fire-induced flooding scenarios.
PRA RAI 4 Treatment of sensitive electronics Disposition in the final aggregate analysis results As responded in Reference 4, PG&E concluded that the treatment of sensitive electronics in the current fire modeling analysis is consistent with FAQ 13-004.
Therefore, no change to the approach is necessary in the final aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will not include any change to the current PG&E's approach of treating the sensitive electronics.
PRA RAI 5 Heat Release Rates lower than 317 kW for transient sources Disposition in the final aggregate analysis results As responded in Reference 5, PG&E provided justifications for all reduced HRRs credited in the fire PRA. The response to NRC PRA RAI 5 does not impact the final aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will not include any change to the current PG&E's approach of selecting reduced HRRs for a fire area.
PRA RAI 6 Other disclosed deviations from acceptable PRA methods Disposition in the final aggregate analysis results As stated in the responses to PRA RAI 6 in Reference 3, and in the responses to other NRC RAls related to fire PRA methods or approaches in References 4 and 5, PG&E concluded that the DCPP PRA model used to support the LAR submittal was 11
Enclosure PG&E Letter DCL-15-119 consistent with the "NRC Accepted Fire PRA Methods" (i.e., NUREG/CR-6850, NRC approved FAQs,' NUREGs, etc.).
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, does not contain any NRC UAMs. There will be no change in modeling methods or approaches in the post-transition fire PRA model as a result of NRC PRA RAI 6.
PRA RAI 8 MCB modeling Disposition in the final aggregate analysis results In Reference 4, PG&E provided clarifications or justifications of treatment of MCB fires that were within the approaches prescribed in Appendix L and in other sections of NUREG/CR-6850.
No change to the treatment of MCB fires is necessary in the final aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes No change is required, in response to NRC RAI 8, to the post-transition fire PRA model, which will be used for self-approval of post-transition changes.
PRA RAI 9 Incipient detection Disposition in the final aggregate analysis results As stated in Reference 3, PG&E does not take credit for area-wide incipient detection or beyond what is described in FAQ 08-0046. The incipient detection system, which will be installed in the CSR and the SSPS room in each unit, will be credited in the fire PRA and in the final aggregate analysis according to the guidance in FAQ 08-0046.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes Once the proposed incipient detection is installed in the CSR and SSPS rooms and operable, the post-transition fire PRA model crediting the incipient detection for the CSR and SSPS cabinet fires will be used for self-approval of post-transition changes.
12
Enclosure PG&E Letter DCL-1 5-1 19 PRA RAI 10 Fire damage effects from the opposite unit Disposition in the final aggregate analysis results As stated in Reference 3, no additional change to the fire PRA resulted from the assessment of the risk contribution from either opposite unit or common areas.
Therefore, no change to the fire PRA model is introduced as part of the aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes No change is required in response to NRC PRA RAI 10, to the post-transition fire -
PRA model, which will be used for self-approval of post-transition changes.
PRA RAI 11 Screening iunction boxes as non-damaging ignition sources Disposition in the final aggregate analysis results PG&E provided detailed justification (Reference 4 and Reference 8) that the treatment of junction box in the fire PRA modeling was consistent with the method described in Section 6.5.6 of FAQ 13-006 for frequency estimate and conservative from a fire impact perspective. No change to the treatment of the junction box in the fire PRA modeling was necessary and therefore no change ot the fire PRA model is introduced as part of the aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes No change is required in response to NRC PRA RAI 11, to the post-transition fire PRA model, which will be used for self-approval of post-transition changes.
PRA RAI 18 Update to internal events PRA as a result of F&O dispositions Disposition in the final aggregate analysis results The final aggregate analysis includes impacts on fire PRA based on improvements made to the Internal Event PRA as a result of responses to the Internal Event F&Os.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes 13
Enclosure PG&E Letter DCL-15-119 The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include improvements to the Internal Events PRA as a result of F&O dispositions.
PRA RAI 22 CCDP developed for loss of control scenarios Disposition in the final aggregate analysis results The MCR abandonment analysis is updated to consider the full range of possible damage scenarios (loss of control and loss of habitability scenarios), providing an assessment of human actions and fire induced equipment failures including spurious operations. The results of the updated MCR abandonment analysis is included in the aggregate analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the updated MCR abandonment analysis.
FM RAIi 1.i Calculation of time to abandonment Disposition in the final aggregate analysis results The MCR abandonment fire modeling analysis is updated incorporating the updated MCR volume, the updated transient fire growth profile, and the updated abandonment time for propagating panel fires, as responded in Reference 5.
The final aggregate analysis includes the MCR abandonment scenarios reflecting the updated abandonment fire modeling analysis.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the updated MCR abandonment fire scenarios.
NOTE: PG&E identified, as listed in PG&E Response PRA PAl 3.a, seven additional RAIs that could potentially impact the results of the fire PRA and therefore the aggregate analysis. The impacts of these additional RAls on the aggregate analysis and the PRA to be used for the post-transition self-approval of change evaluations are provided below as part of response to PRA RAl 3.b 14
Enclosure PG&E Letter DCL-15-119 PRA RAI 1.c Screening of unground 3-phase hot short failures Disposition in the final aggregate analysis results Per NUREG/CR-7150, proper polarity hot shorts on three-phase power cabling are classified as "incredible and not to be considered by the fire PRA." PG&E updated the CFMLA based on the guidance in NUREG/CR-7150. Spurious actuation of RHR Valves 8701 and 8702 remains "excluded" from the fire PRA.
The final aggregate analysis does not include the spurious actuation (i.e., opening) of normally closed RHR Valves 8701 and 8702.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes No change is required in response to NRC PRA RAI 1.c, to the post-transition fire PRA model, which will be used for self-approval of post-transition changes.
PRA RAI l.d Circuit failure probabilities of Thermoset versus Thermoplastic cables Disposition in the final aggregate analysis results PG&E updated the CFMLA based on the guidance in NUREG/CR-7150, which includes consideration of insulation materials as established in the new spurious operation probability tables.
The final aggregate analysis includes the results of the updated CFMLA.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the results of the updated CFMLA.
PRA RAI 1.g.01 Discrepancy between fire ignition Bin 15 count and per-cabinet fire frequency Disposition in the final aggregate analysis results Some of the electrical cabinets were initially screened from fire modeling consideration based on guidance in FAQ 08-042, but counted as fire ignition sources in the Bin 15 population. This inconsistency created the effect of reducing the per-cabinet fire frequency and underestimation of risk contribution from electrical cabinet 15
Enclosure PG&E Letter DCL-15-119 fires. Those electrical cabinets screened from the fire modeling consideration based on guidance in FAQ 08-042 were removed from the Bin 15 population. The per-cabinet fire frequency was recalculated based on the corrected Bin 15 population.
The final aggregate analysis includes the corrected per-cabinet fire frequency.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition Changes, will include the updated fire modeling of electrical cabinets with the corrected per-cabinet fire frequency.
PRA RAI 14.c VFDRs associated with loss of process instrumentation Disposition in the final aggregate analysis results PG&E's response to NRC PRA RAI 14.c (Reference 5) stated that "..., to account for the additional risk, these VFDRs will be evaluated quantitatively and will be included in the risk estimates to be provided in response to PRA RA-03." However, given that the planned HSDP modification will provide alternate process instrumentation, the*
following VFDRs associated with loss of process instrumentation in the MCR and CSRs are eliminated from the FRE of Fire Areas CRI(UI), CR1I(U2), 7A and 7B:
1-CR1-024 1-CR1-031 1-CR1-032 1-CR1-033 1-CR1-034 1-7A-023 1-7A-030 1-7A-032 1-7A-033 1-7A-034 2-CR1 -024
- 2-CR1 -031 2-CR1 -032 2-CR1-033
- 2-CR1-034
- 2-7B-023
- 2-7B-030
- 2-7B-032 2-7B-033 16
Enclosure PG&E Letter DCL-1 5-119
- 2-7B-034 No change to the aggregate analysis models (both the compliant plant model and the variant plant model) and therefore, to the final aggregate analysis is necessary.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition, does not need to be changed to reflect the response to PRA RAI 14.c PRA RAI 15.01 Update of the risk-offset values in Tables W-4 and W-5 with the GEN Ill SDS Disposition in the final aggregate analysis results As stated in PG&E response to NRC PRA RAl 15.01 (Reference 8), no change to the fire PRA models (both compliant plant model and variant plant model) related to the fire modeling of the CSR was necessary.. However, the risk-offset values in Tables W-4 and W-5 of the LAR, which were based on an older version of the Westinghouse RCP SDS modeling, were updated as part of the final aggregate analysis to reflect the proposed installation of the Westinghouse GEN Ill RCP SDS seals.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes No change is required in response to NRC PRA RAls 15 or 15.01, to the post-transition fire PRA model, which will be used for self-approval of post-transition changes.
The modeling of the Westinghouse GEN Ill RCP SOS is discussed in response to PRARPAI 1.f.
FM RAI 1.c.01 Flame spread rate for trays with mixed cables Disposition in the final aggregate analysis results The final aggregate analysis includes the results of the updated fire modeling analysis, which incorporates the guidance provided in NUREG/CR-7010 for the flame spread rate for trays with mixed cables; that is trays with 50 percent or more thermoplastic cable will assume the thermoplastic cable flame spread rate.
17
Enclosure PG&E Letter DCL-15-119 Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the updated fire modeling reflect~ing the corrected flame spread rate for trays carrying predominantly thermoplastic cable.
FM RAI 1.h CFAST analysis and input parameters in Fire Zone 14D Disposition in the final aggregate analysis results In response to NRC FM RAI 1.h (Reference 5), PG&E re-performed the fire modeling analysis of Fire Zone 14D using an FDS analysis because the CFAST analysis for Fire Zone 14D is no longer used. The results of the FDS analysis of Fire Zone 14D revealed that a catastrophic TG fire can potentially generate temperatures in excess of the critical failure temperature of exposed structural steel members, which can lead to a failure of the turbine building.
The final aggregate analysis includes an update to the fire modeling of the TG oil fire in Fire Zone 14D.
Disposition in the PRA that will be used at the beginning of the self-approval of post-transition changes The post-transition fire PRA model, which will be used for self-approval of post-transition changes, will include the updated fire modeling of a catastrophic TG fire that include a range of impacts such as the collapse of the turbine building and damage to PRA SSCs housed within the building.
NRC PRA RAI 3.c In the response, explain how the RG 1.205 risk acceptance guidelines are satisfied for the aggregate analysis. If applicable include a description of any new modifications or operator actions being credited to reduce delta risk as well as a discussion of the associated impacts to the fire protection program.
PG&E RESPONSE to PRA RAI-03.c The results of the aggregate analysis are provided in Table PRA RAI 3-1 for CDF and LERF, and in Table PRA RAI 3-2 for transition risks ACDF and ALERF. As shown in Tables W-4 and W-5 and discussed in Section W 2.2 of the NFPA 805 LAR submittal, there is a net decrease in CDF and LERF associated with the transition of the FP Program to NFPA 805. Regardless of the results of the aggregate analysis shown in Tables PRA RAI 3-1 and 3-2 below, the transition of 18
Enclosure PG&E Letter DCL-15-119 the FP Program to NFPA 805 satisfies the relevant principle of risk-informed regulation with respect to CDF and LERF, as stated in RG 1.174, Revision 2.
PRA RAI 3-1 Results of Aggregate Analysis - CDF and LERF HaadGopUnit 1 Unit 2 HaadGop CDF LERF CDF LERF Fire 4.83E-05 2.45E-06 5.24E-05 2.17E-06 Internal Events 1.13E-05 1.70E-06 1.13E-05 1.70E-06 Internal Floods 7.91E-06 2.93E-07 5.59E-06 2.14E-07 Seismic Events 2.62E-05 2.71 E-06 2.62E-05 2.71E-06 Total Plant Risk 9.37E-05 7.14E-06 9.55E-05 6.78E-06 PRA RAI 3-2. Results of Aggregate Analysis - ACDF and ALERF Unit I Unit 2 ACDF ALERF ACDF ALERF Change in Risk 9.20E-06 6.iI1E-07 1.13E-05 6.69E-07 NRC PRA RAI 3.d If any unacceptable methods or weaknesses will be retained in the PRA that will be used to estimate the change in risk of post-transition changes to support self-approval, explain how the quantification results for each future change will account for the use of these unacceptable methods or weaknesses.
PG&E RESPONSE TO PRA RAI 3.d As documented in previous RAI correspondences (Reference 3, 4, 5, and 8) for each PRA method listed under NRC PRA RAI 3 and PG&E's response to PRA RAI 3.a above, PG&E does not consider any of them as an unacceptable method or weakness. These PRA methods are included in the final aggregate analysis and the PRA models to be used for self-approval of post-transition changes, as explained in PG&E responses to PRA RAI-3.b. Therefore the quantification results for each future change does not need to provide further assessment or explanation for the use of the methods listed in NRC PRA RAI 3 and other RAIs identified in PG&E's response to PRA RAI-03.a.
19
Enclosure PG&E Letter DCL-1 5-119 UPDATES to 60.) 90. and 120-DAY RAI RESPONSES IMPACTED BY PRA RAI-03 (FOLLOW-UP RAI QUESTIONS)
The following also includes updates to RAI responses previously submitted in the 60, 90, and 120-day response letters that were subsequently impacted by the final results of PRA RAI-03 Integrated Analysis. Enclosed are PG&E's responses to these follow-up RAI questions.
NRC PRA RAI 01.h.01 (MCR Abandonment Scenario)
The response to PRA RAI 01.h, dated November 26, 2014, for MCR abandonment due to both loss of habitability and control, the response states: "the full range Of possible damage scenarios associated with abandonment, from scenarios where there is essentially no fire-induced damage (which may occur for some loss of habitability scenarios) up to and including scenarios where the damage may be such that successful remote shutdown is not feasible." Describe the range of CCDPs associated with these MCR abandonment scenarios for both loss of habitability and control. Discuss this CCDP range for loss of habitability separately from the range due to loss of control. If the maximum CCDP in each case is not 1.0O, then discuss the general reasons that a CCDP of one is inappropriate.
PG&E's RESPONSE to PRA RAI 01.h.01 PG&E refined the MGR abandonment analysis fo)r both LOG and LOH as described in Reference 5, considering the full range of possible damage scenarios associated with abandonment.
Scenarios involving LOG occur due to ignition sources in both the GSR and MGR.
LOH scenarios involve only certain MCR ignition sources and result in environmental abandonment conditions.
Control Room Abandonment - Loss of Control The GCDP for scenarios that resulted in LOG were evaluated using a fault tree node in the Riskman event tree model which includes contributions from HFEs and failures of necessary plant equipment. Modele~d HFEs include the operator decision to abandon the control room, to transfer control to the HSDP and to recover functions that have been lost (See PG&E response to PRA RAI-22.01 for a more detailed discussion of LOG HFE development). Gredit for LOG control room abandonment was taken for fire scenarios in which one or more of the following fire induced impacts occurs:
- 1. Loss of AFW
- 2. Loss of AG Power 20
Enclosure PG&E Letter DCL-15-119
- 3. Spurious PORV actuation
- 4. Loss of RCS makeup capability through centrifugal charging pumps Depending on the extent of fire induced damage, the CCDP evaluation includes all necessary operator actions and equipment needed to reach a stable condition. The following table describes the range of CCDPs used and the scenarios for which they are used.
Split Fr mat/eurdRcvr cin rbblt Fraction Fr mat/eurdRcvr cin rbblt MCR1 AFW, Makeup, recirculation required 6.40E-02 MCR1 P AFW, Makeup, PORV Open, recirculation required 6.82E-02 MCR1AC AFW, Makeup, AC Power, recirculation required 6.80E-02 AFW, Makeup, AC Power, PORV Open, MCR1 PA recirculation required 7.21E-02 MCR3 Makeup 4.49E-02 MCR3P Makeup, PORV Open 4.92E-02 MCR3AC Makeup, AC Power 4.89E-02 MCR3PA Makeup, AC Power, PORV Open 5.32E-02 MCR7 Makeup, recirculation required 5.19E-02 MCR7P Makeup, recirculation required, PORV Open 5.61E-02 MCR7AC Makeup, AC Power, recirculation required 5.59E-02 Makeup, AC Power, PORV Open recirculation MCR7PA required 6.01E-02 MCR9 AC Power 3.65E-02 In addition to the CCDPs displayed in the table above, there are certain scenarios where a CCDP of 1.0 is assessed. For example if the fire results in an ISLOCA, mitigation from the HSDP is assumed to be unsuccessful and no credit for recovery is taken in the PRA. Also not shown are scenarios where functional abandonment criteria is not reached (AFW, AC power and RCS makeup available with no spurious PORV operation) and the MCR abandonment modeling is bypassed.
Control Room Abandonment - Loss of Habitability For LOH scenarios, diagnosis of the need to abandon is assumed to be successful when the habitability limit criteria are met. All other success criteria used in the LOG modeling are applicable to LOH scenarios. The same fault tree used for LOG modeling is used for LOH modeling except that different boundary conditions are used to calculate the LOH split fractions. For LOH scenarios, operators are directed to abandon the MCR due to habitability concern, not due to loss of control.
Therefore at the time of the MCR abandonment, the status of the plant condition 21
Enclosure PG&E Letter DCL-15-119 (i.e., degree of LOG) is not known. This uncertainty on the degree of LOG is modeled by conservatively requiring recovery of all control functions (i.e., AFW, RCS Makeup, AC Power, Spuriously Open PORVs, Containment Recirc) outside of the MCR.
SplitImat/eurdRcvrAcinPrbbly FractionImat/eurdRcvrAcinPrbily AFW, Makeup, AC Power, PORV Open, recirculation MCR8 required 4.53E-02 The CCDP (Split Fraction MCR8) for LOH is lower than those for LOG since the dominant contributor, diagnosis of the need to abandon, is not included in the calculation.
The maximum CCDP for all LOH and LOG scenarios is not equal to 1.0 except for certain accident sequences where the HSDP is assumed incapable of recovery (i.e.
ISLOCA), since all scenarios involve impacts that are recoverable through the use of the HSDP and/or local actions. This is intuitive since the design of the HSDP is such that fire induced damage in the MCR or CSR is isolable and, once operators take action to transfer control from the control room, functionality is available at the HSDP to reach a safe and stable plant condition.
NRC PRA RAI 01.k.01 - State of Knowledge Correlation (F&Os)
Please clarify the following dispositions to fire F&Os and Supporting Requirement (SR) assessment identified in LAR Attachment V that have the potential to impact the Fire PRA (Fire PRA) results and do not appear to be fully resolved:
k) FSS-E3-O01-2008 (State of Knowledge Correlation)
This F&O cites the lack of qualitative or quantitative uncertainty analysis associated with fire modeling and accident sequence analysis. The F&O disposition explains that since the 2008 peer review, the analysis has been completed that provides qualitative and quantitative characterization of uncertainty and that the analysis discusses the state of knowledge correlation (SOKC) and indicates that SOKC needs to be considered for fire ignition frequencies. The analysis also states that the SOKC was taken into account for" fire ignition frequencies and also indicates that the uncertainty of internal events component failure probabilities, circuit failure likelihood, and non-suppression probabilities were treated quantitatively, but does not indicate that SOKC was taken into consideration. If SOKC for these parameters was not accounted for in the Fire PRA quantification, then address SOKC for these parameters in the integrated analysis performed in response to PRA RAl 3.
22
Enclosure PG&E Letter DCL-15-119 PG&E's RESPONSE to PRA RAI 1.k.01 As stated in PG&E response to PRA RAI 1.k in Reference 5, PG&E performed the uncertainty analysis SOKC based on the fire PRA model developed in response to PRA RAI-03. The SOKC analysis considered for all parameters in the Fire PRA model with which uncertainty distributions were available such as the fire ignition source frequency, circuit failure probabilities, component unavailability, human error probabilities, etc. Distributions for some parameters such as the non-suppression factor/probability were not available and were therefore not considered in the uncertainty analysis.
The analysis was performed using the Monte Carlo Calculation function of RISKMAN. The method used in RISKMAN essentially samples all the parametric distributions in the model and applies each sample value of each parameter to all the basic events which have the same parameter and to all fire scenarios that has the same Fire "Bin" or ignition source. The system/top event failure probabilities are next re-quantified using the sampled value from all the parameters (including.
component failure rates and unavailability values, human error probability, hot-short probability, etc.). The fire induced accident sequences are next quantified using the updated (based on the sampled parameter values) fire scenario frequency and system/top event split fraction values to obtain a sample CDF and LERF value. The process is repeated 2000 times to calculate the distributions for the fire-induced CDF and LERF and their mean values.
A comparison of the mean values with consideration of the SOKC and the point estimate values used in support of the LAR shows that the difference between the mean value of the CDF and LERF is slightly higher (less than 1 percent) than the point estimate CDF and LERF.
The SOKC is expected to have minimum impact on the mean value of the fire induced total CDF and LERF and therefore, the use of point estimate CDF and LERF as a surrogate for the mean value of the risk metrics is a reasonable approach.
NRC PRA RAI 01.1.01-Minimum joint Human Error Probability (F&Os)
Please clarify the following dispositions to fire F&Os and Supporting Requirement (SR) assessment identified in LAR Attachment V that have the potential to impact the Fire PRA (Fire PRA) results and do not appear to be fully resolved:
I) HRA-CI-01-2010 (Minimum joint Human Error Probability)
The disposition to this F&O explains that since the 2010 peer review, an updated human reliability analysis (HRA) dependency analysis was completed. The NRC staff noted that the analysis does not indicate that a 23
Enclosure PG&E Letter DCL-1 5-1 19 minimum joint Human Error Probability (HEP) was applied to the dependency analysis. Per guidance in NUREG -1921, "EPRI/NRC-RES Fire Human Reliability Analysis Guidelines Final Report," the Human Failure Event (HIFE) dependency analysis should consider the minimum joint value (i.e., floor) for multiple HFEs occurring in the same cutset. Explain whether a "floor" was applied in the HFE dependency analysis, and if so, identify the floor used and justify any value used less than 1E-5. If a "floor" was not used, apply a floor to the HFE dependency analysis as part of the integrated analysis performed in response to PRA RAI 3, and justify the value used.
PG&E's RESPONSE to PRA RAI 1.1.01 As stated in PG&E response to PRA RAI 1.1 (Reference 5), the fire HRA dependency analysis was reperformed. Following the dependency analysis, joint HEP values were evaluated and compared to a floor value of I1E-05. Except for one class of scenarios, the joint probability for all HEP combinations is greater than 1IE-05.
Because the sequences obtained from Riskman event tree quantification are non-minimal, once the HEP combinations are identified, these combinations must be reviewed in order to identify the minimal set of HEPs necessary and sufficient for core damage to occur. After identification of the minimal HEP combinations, the joint HEP was calculated and compared to a floor criteria of 1 E-05.
The HFEs for establishing RHR closed loop cooling (ZHEMU1) and alignment of the alternate water supply to AFW (ZHEAW9) are the only two HFEs that result in a joint combination of less than 1 E-05. The first action (establishing RHR closed loop cooling) occurs approximately 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after the initiating event, long after the fire would have been extinguished. The following reasons exist for assessing a low or zero dependency between these two actions:
- The time frame between these two actions is greater than 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> which would allow sufficient time for a shift change in between the actions. The oncoming crew will be much less likely to make a dependent error in diagnosis or execution of the subsequent action to establish an alternate water supply to AFW.
- In addition, a variety of cues are available to diagnose the need for an alternate AFW water supply. These cues are completely independent from the cues used to diagnose the n'eed for RHR cooling.
- The procedures used for diagnosis and execution for these actions are different.
24
Enclosure PG&E Letter DCL-15-119 Because of the long time frame between actions, difference in cues, and separate diagnosis/execution procedures, a dependency level of zero is justified between these two actions.
NRC PRA RAI 14.01 - Calculation of VFDR ACDF and ALERF LAR Attachment W, Section W.2.1I provides a description of how the ACDF and ALERF, associated with VFDRs is determined. Please supplement the description by:
a) Describing the types of model adjustments that were made to remove different types of VFDRs from the compliant case FPRA. In addition, identify any major changes made to the FPRA models and data for evaluating VFDRs.
b) Describing how VFDR and non-VFDR plant modifications are credited in the compliant and post transition PRA models.
c) Describing of the type of VFDRs identified, and pro viding justification that any VFDRs identified but not modeled in the FPRA do not impact the risk estimates.
PG&E's INITIAL RESPONSE to PRA RAI 14.a No change required from the initial response to PRA RAl 14.a referenced above.
PG&E's Initial Response as provided in Reference 5 remains unchanged as initially submitted.*.
PG&E's INITIAL RESPONSE to PRA RAI 14.b No change required from the initial response to PRA RAl 14.b referenced above.
PG&E's Initial Response as provided in Reference 5 remains unchanged as initially submitted.
PG&E's RESPONSE TO PRA RAI 14.c.01 PG&E's response to NRC PRA RAl 14.c (Reference 5) stated that "----, to account for the additional risk, these VFDRs will be evaluated quantitatively and will be included in the risk estimates to be provided in response to PRA RAl-03." For VFDRs associated with normally closed Manual Valves MU-0-1 557, MU0-280 and MFW-1/2-FCV-437, to account for the additional risk, these VFDRs were evaluated quantitatively and are included in the risk estimates provided in response to RAl-03.
However, given that the planned HSDP modification will provide alternate process instrumentation, VFDRs associated with loss of process instrumentation in the MCR 25
Enclosure PG&E Letter DCL-15-119 and CSRs (VFDRs 1-CR1-024, 1-CR1-031, 1-CRI-032, 1-CR1-033, 1-CR1-034, 1-7A-023, 1 -7A-030, 1 -7A-032, 1 -7A-033, 1 -7A-034, 2-CR1 -024, 2-CR1 -031, 2-CRI1-032, 2-CR1 -033, 2-CR1-034, 2-7B-023, 2-7B-030, 2-7B-032, 2-7B-033, and 2-7B-034) are eliminated from the FRE of Fire Areas CR1I(U 1), CR1I(U2), 7A and 7B.
Therefore, no change to the aggregate analysys models (both compliant and variant plant) or final aggregate analysis is necessary.
NRC PRA RAI 15.01 - Larqe Reduction Credit for Modifications NRC staff notes that based on the response to PRA RAI 15, dated November 26,
-2014, the most dominant scenario for the compliant plant model for both units (by a wide margin) is the whole-room burn-up scenario for the Cable Spreading Room (CSR), representing most of risk reduction credit realized in the "~risk offset' values presented in Tables W-4 and W-5 of the LAR. Whole-room burn-up modeling implies that this fire area was conservatively modelled. If this fire area was conservatively modeled, then it appears that this scenario contributes to overestimation of the compliant plant risk and there fore underestimation of the change-in-risk estimates.
In light of potential conservatism in the compliant plant modeling, demonstrate that the net change-in-risk ('i.e., total risk decrease associated with non-VFDR risk reduction modifications and total risk increase associated with unresolved VFDRs) associated with the integrated analysis presented in response to PRA RAI 3 meet Regulatory Guide 1.174 CDF and LERF risk guidelines.
PG&E's RESPONSE PRA RAI 15.01 PG&E's INITIAL Response to PRA RAI 15 as provided in Reference 8 "As stated in PG&E response to PRA RAI 15 in PG&E Letter DCL-15-032, dated February 25, 2015, PG&E provided justification that the modeling of the whole-room burnup scenario for the Cable Spreading Room (CSR) does not result in an underestimation 6f the change-in-risk estimate. The response concluded that no change to the compliant plant model, the PRA RAI-03 aggregate model, or variant plant model is necessary as part of the response to this RAI. However due to DCPP's plan to install the Westinghouse GEN Ill RCP SDS instead of an older version of the Westinghouse SDS, it was necessary to update the risk-off set values in Tables W-4 and W-5 of the LAR."
PG&E's UPDATED RESPONSE to PRA RAI 15.01:
The risk-offset values were updated as part of response to PRA RAI-03 using the fire PRA model developed for the aggregate analysis, which includes modeling of the Westinghouse GEN Ill RCP SDS. The updated risk-offset values are provided in 26
Enclosure PG&E Letter DCL-1 5-119 Tables W-4 and W-5. These updated risk-offset values also' include fire scenarios that were not analyzed for risk offset in the original Tables W-4 and W-5. The risk offset values provided in the original Tables W-4 and W-5 compared only fire scenarios that caused VFDRs as determined by the NSCA. To improve consistency, the updated risk offset values based on the final aggregate analysis compare all non-screened fire scenarios with and without the RCP SDS modification, regardless of impact to VFDRs.
NRC PRA RAI 21.01 - Sensitivity Analysis on FAQ 08-0048 Fire Bin Frequencies LAR Attachment V, Table V-I presents the results of a sensitivity study on the updated fire ignition bin frequencies provided in NUREG/CR-6850, Supplement I (i.e., FAQ 08-0048) using the mean of the fire frequency bins contained in Section 6 of NUREG/CR-6850 for those bins with an alpha value less than or equal to one. It is not clear why the percent increase for A* CDF and A* LERF is higher than the percent increase for the total fire CDF and LERF (e.g., the percent increase in CDF for Unit 2 is 57% while the percent increase in A CDF for Unit 2 is 74%), given that fire ignition frequencies impact both the compliant and post-transition plant case accident sequences the same, and do not affect CCDP and CLERP values.
Please provide the following:
a) An explanation of the anomaly cited above and whether the reported change-in-risk values in LAR Attachment V, Table V-I are correct.
b) An updated sensitivity study based on the integrated analysis performed in response to PRA RAI 3. Include in the sensitivity study any adjustments needed to correct anomalous results from the initial sensitivity study c) An indication of whether the acceptance guidelines of RG 1.174 may be exceeded when this sensitivity study with respect to FAQ 08-004 8 is applied to the integrated study of PRA RAI 3. If these guidelines may be exceeded, provide a description of fire protection, or related, measures that can be taken to provide additional defense depth, as discussed in FAQ 08-0048.
27
Enclosure PG&E Letter DGL-1 5-119 PG&E's RESPONSE to PRA RAI 21.a (Excerpt provided in Reference 5)
"The difference in the percent increase between the total fire CDF (and LERF) and the transition ACDF (and ALERF) reported in Table V-i, Attachment V of the DCPP NEPA 805 LAR submittal is not an anomaly. The reported change-in risk values in Table V-I, which were based on the Fire PRA model used for the LAR in June 2013, are correct....In summary, a higher percent increase for ACDF and ALERF is accounted for and is not an anomaly. There is no adjustment or correction to be made to the aggregate model in response to PRA RAI-21 (a)."
No change is required from the initial response to PRA RAl 21.a referenced above.
PG&E's Initial Response as provided in Reference 5 remains unchanged as initially submitted.
PG&E's INITIAL RESPONSE to PRA RAI 21.b (Provided in PG&E Letter DCL-14-110. dated November 26, 2014)
Once the results (i.e., the total CDF, LERF, ACDF, ALERF) of the aggregate analysis in response to PRA RAI-03 is available, a sensitivity study on the updated fire ignition bin frequencies provided in NUREG/CR-6850, Supplement 1 (i.e., FAQ 08-0048) will be performed. A review of the results from the initial sensitivity study in response to PRA RAI-21 (a) shows that a higher percent increase for ACDF and ALERF reported in the initial sensitivity study as compared to CDF and LERF is not an anomaly. This review identified no correction or adjustment to be made to the aggregate analysis or the sensitivity study.
PG&E's RESPONSE to PRA RAI 21.b.01 A sensitivity study on the updated fir-e ignition bin frequencies provided in NUREG/CR-6850, Supplement I (i.e., FAQ 08-0048) was re-performed based on the results of the aggregate analysis in response to PRA RAI-03.
The updated Table V-I of the NFPA 805 LAR submittal is provided as part of the response to PRA RAl 21.b.01 as shown below.
Table V-I Ignition Frequency Sensitivity Unit I Unit 2 28
Enclosure PG&E Letter DCL-1 5-119 Resultant, eutn
- -Increase Rslat Increase Case CDFILERF Change CDFILERF Change (y)
CDFILERF (y)
CDF/LERF CDF - Base 4.83E-05 5.24E-05 CDF -IGF Sensitivity 6.94E-05 2.11 E-05 44%
7.62E-05 2.38E-05 45%
ACDF - Base 9.22E-06 1.12E-05 ACDF - IGF Sniiiy 1.61 E-05 6.84E-06 74%
1.89E-05 7.71 E-06 69%
LERF -Base 2.45E-06 2.17E-06 LERF-IGF Sensitivity 3.65E-06 1.20E-06 49%
3.19E-06 1.02E-06 47%
ALERF - Base 6.12E-07 6.64E-07 AEF-GE 1.16E-06 5.44E-07 89%
1.21 E-06 5.46E-07 82%
Sensitivity SPG&E's INITIAL RESPONSE to PRA RAI 21.c (as provided in Reference 5)
The results of the updated sensitivity study in response to PRA RAI-21 (b) will be compared to the acceptance guidelines of RG 1.174. If these guidelines are exceeded,I the results of the sensitivity study will be reviewed to obtain risk insights such as dominant fire areas, fire scenarios or accident sequences, and differences in risk profile between the sensitivity study cases. Based on these risk insights, the existing defense-in-depth measures will be reviewed if they provide adequate measures to counter potential risk increase, or additional defense-in-depth measures will be identified if the existing measures are determined to be inadequate.
PG&E's RESPONSE TO PRA RAI 21.c.01 Tables RAI 21.c -1 and RAI 21.c -2 below provide a comparison between the acceptance guidelines of RG 1.174 and the results of the updated sensitivity study performed per FAQ 08-0048 in response to PRA RAI-21.b. The total CDF and LERF are the aggregated values of all hazard groups, while ACDF and ALERF represent change-in-risk.
29
Enclosure PG&E Letter DCL-15-119 Table RAI 21.c -1, Comparison between RG 1.174 guidelines and Unit 1 Sensitivity Results Exceed RG Figure-Of-Based on Based on RG 1.1741.7 Merit (per Integrated Sensitivity Guidelines Gudln year)
Analysis Analysis (Region II)
(YIN)
Total CDF 9.37E-05 1.15E-04
<1 E-04 Y
Total LERF 7.14E-06 8.34E-06
<1E-05 N
ACDF 9.22E-06 1.61E-05
<1E-05 Y
ALERF 6.12E-07 1.16E-06
<1E-06 Y
Table RAI 21.c-2, Results Comparison between RG 1.174 guidelines and Unit 2 Sensitivity Exceed RG Figure-Of-Based on Based on RG 1.1741.7 Merit (per Integrated Integrated Guidelines Gudln year)
Analysis Analysis (Region II)
(YIN)
Total CDF 9.55E-05 1.19E-04
<1E-04 Y
Total LERF 6.78E-06 7.80E-06
<1E-05 N
ACDF 1.12E-05 1.89E-05
<1E-05 Y
ALERF 6.64E-07 1.21E-06
<1E-06 Y
As indicated in Tables RA! 21.c-1 and RAI 21.c-2, the total CDF, ACDF, and ALERF values for both units exceed when the sensitivity study, with respect to FAQ 08-0048, is applied to the integrated study of PRA RAI-03. However, given a net decrease in the transition risk (that is negative ACDF and ALERF) after applying the risk-offset credit from the installation of the Westinghouse GEN III RCP seals, as discussed in Attachment W of the NFPA 805 LAR and based on the guidance provided in RG 1.174, which states that "~If the application clearly shows.a decrease in CDF (and LERF), the change will be considered to have satisfied the relevant principle of risk-informed regulation with respect to CDF." It is not required to establish additional defense-in-depth measure to compensate for the RG 1.174 exceedance as indicated above. The results of the sensitivity study are reviewed and the following risk insights are provided only for the sake of completeness of the sensitivity study.
- For both units, the results of the integrated analysis are sensitive mostly to changes in Bin 15.1 (Electrical Cabinet) and Bin 4 (Main Control Room Main Control Board) generic frequencies. At DCPP, over 90 percent of risk change 30
Enclosure PG&E Letter DCL-15-119 between the integrated analysis and the sensitivity analysis is attributed to Bins 15.1 and 4.
- For both units, fire scenarios of electrical cabinet fires (Bin 15.1) in the SSPS rooms and the Battery Charger Rooms are dominant scenarios affected by the sensitivity study from a CDF perspective.
- For both units, fire scenarios of electrical cabinet fires (Bin 15.1) in the SSPS rooms, the Battery Charger Rooms, the CSR, and 230kV SWGR rooms, and fire scenarios of MCB fires (Bin 4) are dominant scenarios affected by the sensitivity study from a ACDF and ALERF perspective.
NRC PRA RAI 22.01 - HEP for CR Abandonment Scenarios The responses to PRA RAI 22 and PRA RAI 01.h, dated November 26, 2014, explain that detailed Human Reliability Analysis (HRA) based on guidance from NUREG-1921 was used to model actions for alternate shutdown, and indicate that scenario-specific timing based on damage and spurious actions induced by fire was addressed. How were the time windows used in evaluating Human Failure Events adjusted for different scenarios?
PG&E's RESPONSE PRA RAI 22.01 As indicated in Reference 5, PG&E updated the MCR abandonment modeling, including a more complete assessment of human actions, fire induced equipment failures, and spurious operations. The updated MCR abandonment HRA was developed using timing information from thermal-hydraulic analysis for the full range of abandonment scenarios. MCR abandonment scenarios that stem from a fire induced LOG as a result of a CSR or MCR fire include one or more of the following impacts:
Loss of AFW Loss of AC Power Spurious PORV actuation Loss of RCS makeup capability through centrifugal charging pumps For each of the different LOG scenarios, an individual and detailed assessment of the CCDP is made based on the operator actions and equipment required to reach a safe and stable plant condition.
31
Enclosure PG&E Letter DCL-1 5-119 Assessment of Timing The updated MCR abandonment HRA contains an evaluation of the cognitive HFE for the decision to abandon the control room. This decision to abandon is based on the availability of cues which indicate that one or more of the above impacts have occurred. The time available for operators to diagnose a LOC is based on the most limiting T-H case in which the reactor has tripped on a loss of main feedwater and a fire has prevented actuation of AFW and charging injection. The time available for diagnosis is reduced by the time needed to transfer control to the HSDP and establish AFW and charging injection. The timing needed to perform these functions is estimated through the use of simulator based JPMs.
It was estimated that operators would reach the abandonment decision step in the CSR fire response procedure within 10 minutes based on the time it would take to enter the appropriate.procedure and perform steps prior to the abandonment decision step. At this point, each of the above fire induced impact cues would be evident through status checks of steam generator level, AC Power, RCS pressure and charging status and a decision to abandon could be made.
Comparison of Results The original analysis used in support of the DCPP NFPA 805 LAR contained a much simpler assessment of human actions taken in response to a fire which required control room abandonment. In the original analysis, only a simple diagnosis action was, modeled and the time window for this action was not based on T-H analysis, but instead on the time taken by operator crews in simulator scenarios. In addition, the fault tree analysis which calculates the CCDPs for various abandonment scenarios did not include detailed system modeling of equipment required to reach a safe and stable condition.
The updated analysis which is used to provide a response to RAI-3 contains all diagnosis and execution HFE elements that are required to reach a safe and stable plant condition.
Operator actions that are modeled in the updated control room abandonment PRA*
analysis include:
- 1. Recovery of RCS makeup capability
The timeline for these actions use T-H analysis to determine the time available for diagnosis/execution as well as detailed JPM information to estimate the time 32
Enclosure PG&E Letter DCL-15-119 required by operators to perform various functions. The updated modeling.also includes system modeling of all equipment need to reach a safe and stable plant condition. With the updated model, a range of probabilistic results have been obtained for each of the different fire scenarios and their unique impacts.
The updated MCR abandonment HRA is incorporated into a fault tree that includes the random failure of equipment needed to reach a safe and stable plant condition.
More detailed information regarding this fault tree and associated CCDPs is a~vailable in the response to PRA RAI 01h.01. These more detailed results are in contrast to the PRA used in support of the LAR since the original PRA included only a single generalized CCDP for all control room abandonment scenarios.
- 2. REFERENCES
- 1) PG&E Letter DCL-13-065, "License Amendment Request 13-03, License Amendment Request to Adopt NFPA 805 Performance-Based Standard for Fire Protection for Light Water Reactor Electric Generating Plants (2001 Edition),"
dated June 26, 2013
- 2) NRC Letter, "Diablo Canyon Power Plant, Units 1 and 2 - Request for Additional Information Re: License Amendment Request to Adopt National Fire Protection Association Standard 805 (TAC Nos. MF2333 and MF2334)," dated July 31, 2014
- 3) PG&E Letter DCL-14-085, "Sixty-Day Response to NRC Request for Additional Information - National Fire Protection Association Standard 805," dated September 29, 2014
- 4) PG&E Letter DCL-14-098, "Ninety-Day Response to NRC Request for Additional Information - National Fire Protection Association Standard 805," dated October 29, 2014
- 5) PG&E Letter DCL-14-110, "One Hundred Twenty Day Response to NRC Request for Additional Information - National Fire Protection Association Standard 805," dated November 26, 2014
- 6) NRC Email from S. Lingam, Project Manager (NRR/DORL/LPL4-1), "Diablo Canyon, Units 1 and 2 - Requests for Additional Information (RAIs) for Fire Protection (NFPA-805) License Amendment Request (LAR) (TAC Nos. MF2333 and MF2334)," dated February 4, 2015
- 7) PG&E Letter DCL-1 5-031, "Proposed Commitment for Response to NRC Request for Additional Information Regarding Submittal Date of the Integrated 33
Enclosure PG&E Letter DCLI15-119 Analysis - National Fire Protection Association Standard 805," dated February 25, 2015
- 8) PG&E Letter DCL-I15-032, "Response to NRC Request for Additional Information
- National Fire Protection Association Standard 805," dated February 25, 2015 34
Enclosure PG&E Letter DCL-1 5-119 List of Acronyms Used in This Response AC.
Alternating Current AFW Auxiliary Feedwater system AMSAC ATWVS Mitigation System Actuation Circuitry CCDP Conditional Core Damage Probability CDF Core Damage Frequency ACDF Delta Core Damage Frequency C FAST Consolidated Model of Fire and Smoke Transport CFMLA Circuit Failure Mode Likelihood Analysis CLERP Conditional Large Early release Probability CSR Cable Spreading Room DCPP Diablo Canyon Power Plant FAQ Frequently Asked Question FDS Fire Dynamics Simulator F&O Findings and Observations FP Fire Protection FPRA Fire Probabilistic Risk Assessment FRE Fire Risk Evaluation HEAF High Energy Arcing Fault HEP Human Error Probability HFE Human Failure Event HRA Human Reliability Analysis HRR Heat Release Rate HSDP Hot Shut Down Panel IGF Ignition Frequency ISLOCA Interfacing System Loss of Coolant Accident J.PM Job Performance Measures
[AR License Amendment Request LERF Large Early Release Frequency 35
Enclosure PG&E Letter DCL-1 5-119 ALERF LOG LOCA LOH MOB MCR NFPA NRC NSCA PG&E PORV PPRA RAl RCP RCS RG RHR RISKMAN SBO SDS SOKC SR SSC SSD SSPS SWGR TG T-H UAM Delta Large Early Release Frequency Loss of Control Loss of Coolant Accident Loss Of Habitability Main Control Board Main Control Room National Fire Protection Association Nuclear Regulatory Commission Nuclear Safety Capability Assessment Pacific Gas and Electric Company Power-Operated Relief Valve Probabilistic Risk Assessment Request for Additional Information Reactor Coolant Pump Reactor Coolant System Regulatory Guide Residual Heat Removal Riskman Software Station Blackout' Shutdown Seals State Of Knowledge Correlation Supporting Requirements Structure, System, or Component Safe Shutdown Solid State Protection System Switchgear Turbine Generator Thermal Hydraulic Un-reviewed Analysis Methods 36