NL-17-0218, Fukushima Near-Term Task Force Recommendation 2.1: Seismic Probabilistic Risk Assessment

Fukushima Near-Term Task Force Recommendation 2.1: Seismic Probabilistic Risk Assessment

ML17088A130
Person / Time
Site:	Vogtle
Issue date:	03/27/2017
From:	Hutto J Southern Nuclear Operating Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
NL-17-0218
Download: ML17088A130 (128)

Text

4' Southern Nuclear March 27, 2017 Docket Nos.:

• 50-424 50-425 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, D. C. 20555-0001 J. J. Hutto Regulatory Affairs.Director 40 Inverness Center Parkway Post Office Box 1295 Birmingham, 35242 205 992 5872 tel 205 992 7601 fax jjhutto@southernco.com NL-17-0218 Vogtle Electric Generating Plant Units 1 and 2 Fukushima Near-Term Task Force Recommendation 2.1: Seismic Seismic Probabilistic Risk Assessment

References:

1. NRC Letter, Request for Information Pursuant to 10 CFR 50.54(f) Regarding Recommendations 2.1, 2.3, and 9.3 of the NTTF Review of Insights from the Fukushima Dai-ichi Accident, dated March 12, 2012.

2. EPRI Report 1025287, "Seismic Evaluation Guidance: Screening, Prioritization and Implementation Details (SPID) for the Resolution of Fukushima Near-Term Task Force Recommendation 2.1: Seismic." ML12333A170.

3. Letter to the NRC, Vogtle Electric Generating Plant Units 1 and 2, Seismic Hazard and Screening Report for CEUS Sites, dated March 31, 2014. ML14092A019.

4. NRC Letter, Vogtle Electric Generating Plant Units 1 and 2, Staff Assessment of Seismic Hazard Reevaluations Pertaining to Recommendation 2.1, dated April 20, 2015. ML15054A296.

5. NRC Letter, Final Determination of Licensee Seismic Probabilistic Risk Assessments, dated October 27, 2015. ML15194A015.

Ladies and Gentlemen:

On March 12, 2012, the Nuclear Regulatory Commission (NRC) issued a request for information pursuant to 1 O CFR 50.54(f) associated with the recommendations of the Fukushima Near-Term Task Force (NTTF) (Reference 1). Enclosure 1 of Reference 1 requested each licensee to reevaluate the seismic hazards at their sites using present-day NRC requirements and guidance, and to identify actions taken or planned to address plant-specific vulnerabilities associated with the updated seismic hazards.

Reference 2 contains industry guidance developed by EPRI that provide the screening, prioritization and implementation details for the resolution of Fukushima Near-Term Task Force Recommendation 2.1: Seismic. The SPID (Reference 2) was used to compare the reevaluated seismic hazard to the design basis hazard. The Vogtle Electric Generating Plant (VEGP), Units 1 and 2 reevaluated seismic hazard (Reference 3) concluded that the ground motion response spectrum (GMRS) exceeded the design basis seismic response spectrum in the 1 to 1 O Hz range, and therefore a seismic probabilistic risk assessment was required.

U. S. Nuclear Regulatory Commission NL-17-0218 Page 2 Reference 4 contains the NRC Staff Assessment of the VEGP Units 1 and 2 seismic hazard submittal which concluded that the reevaluated seismic hazard prepared for VEGP is suitable for other activities associated with the NRC Near-Term Task Force Recommendation 2.1: Seismic.

Reference 5 contains the NRC letter "Final Determination of Licensee Seismic Probabilistic

• Risk Assessments." In that letter (Table 1 a - Recommendation 2.1 Seismic - Information Requests) the NRC instructed VEGP Units 1 and 2 to submit an SPRA by March 31, 2017. of this letter contains the VEGP Units 1 and 2, Seismic Probabilistic Risk Assessment (SPRA) Summary Report which provides the information requested in, Item (8) B. of the 10 CFR 50.54(f) letter.

In accordance with PWROG-14001, "PRA Model for the Generation Ill Westinghouse Shutdown Seal," a PWR Owners' Group project (PA-RMSC-1423) was initiated to evaluate the Generation Ill Westinghouse Reactor Coolant Pump (RCP) SHIELD Passive Thermal Shutdown Seal (Generation Ill SOS). Coupling the current Emergency Operating Procedures with the results of the project, cold leg temperatures which could adversely impact the operation of the Generation Ill SOS would not be reached at VEGP.

This letter contains no NRC commitments. If you have any questions, please contact John Giddens at 205.992.7924.

Mr. J. J. Hutto states he is the Regulatory Affairs Director for Southern Nuclear Operating Company, is authorized to execute this oath on behalf of Southern Nuclear Operating Company and, to the best of his knowledge and belief, the facts set forth in this letter are true.

Respectfully submitted, Ejt?liC>

J. J. Hutto Regulatory Affairs Director JJH/JMG/GLS Sw n.to and subscri* ed before me this J,J__ day of JY1 evt ~ '2017.

My commission expires: /O

• 8' - Z O I 1

Enclosure:

Vogtle Electric Generating Plant - Units 1 and 2 Seismic Probabilistic Risk Assessment Summary Report

,. '~.

)*

U.S. Nuclear Regulatory Commission NL-17-0218 Page 3 cc:

. Southern Nuclear Operating Company Mr. S. E. Kuczynski, Chairman, President & CEO Mr. D. G. Bost, Executive Vice President & Chief Nuclear Officer Mr. M. D. Meier, Vice President - Regulatory Affairs Mr. B. K. Taber, Vice President - Vogtle 1 & 2 Mr. B. J. Adams, Vice President - Engineering Mr. D. D. Sutton, Regulatory Affairs Manager - Vogtle 1 & 2 RType: CVC7000 U. S. Nuclear Regulatory Commission Ms. C. Haney, Regional Administrator Mr. M. D. Orenak, NRR Project Manager - Vogtle 1 & 2 Mr. M. F. Endress, Senior Resident Inspector - Vogtle 1 & 2 State of Georgia Mr. R. E. Dunn, Director - Environmental.Protection Diyision

Vogtle Electric Generating Plant - Units 1 and 2 Fukushima Near-Term Task Force Recommendation 2.1: Seismic Seismic Probabilistic Risk Assessment Enclosure Vogtle Electric Generating Plant - Units 1 and 2 Seismic Probabilistic Risk Assessment Summary Report

VEGP Units 1 and 2 10 CFR 50.54(f) N'TIF 2.1 Seismic PRA Submittal Version 0 - March 2017,

VOGTLE ELECTRIC GENERATING PLANT UNITS 1AND2 SEISMIC PROBABILISTIC RISK ASSESSMENT IN RESPONSE TO 10 CFR 50.54(f) LETTER WITH REGARD TO NTTF 2.1 SEISMIC VERSIONO MARCH2017

SUMMARY

REPORT Page 1of124

VEGP Units 1 and 2 Contents 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 VEGP UNITS 1 AND 2 SEISMIC PROBABILISTIC RISK ASSESSMENT

SUMMARY

REPORT Executive Summary.................................................................................................................................... 3 1.0 Purpose and Objective.................................................................................................................... 4 2.0 Information Provided in This Report.............................................................................................. 5 3.0 VEGP Seismic Hazard and Plant Response..................................................................................... 9 4.0 Determination of Seismic Fragilities for the SPRA....................................................................... 17 5.0 Plant Seismic Logic Model............................................................................................................. 32 6.0 Conclusions.................................................................................................................................... 58 7.0 References..................................................................................................................................... 59 8.0 Acronyms....................................................................................................................................... 62 Appendix A Summary of SPRA Peer Review and Assessment of PRA Technical Adequacy for Response to NTIF 2.1 Seismic 50.54(f) Letter......................................................................................... 65 Page 2of124

' I.

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Executive Summary In response to the 10 CFR 50.54(f) letter issued by the NRC on March 12, 2012, a seismic PRA (SPRA) has been developed to perform the seismic risk assessment for Plant Vogtle Units 1 and 2.

The SPRA shows that the point estimate seismic Core Damage Frequency (SCDF) is 2.8x10-6/yr and the seismic Large Early Release Frequency (SLERF) is 3.3x10-7 /yr. Sensitivity studies were performed to identify critical assumptions, test the sensitivity to quantification parameters and the seismic hazard, and identify potential areas to consider for the reduction of seismic risk.

These sensitivity studies demonstrated that the model results were robust to the modeling and assumptions used. No seismic hazard vulnerabilities were identified, and no plant actions have been taken or are planned given the insights from the seismic risk assessment.

Page 3of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 1.0 Purpose and Objective Following the accident at the Fukushima Dai-ichi nuclear power plant resulting from the March 11, 2011, Great Tohoku Earthquake and subsequent tsunami, the Nuclear Regulatory Commission (NRC) established a Near Term Task Force (NTTF) to conduct a systematic review of NRC processes and regulations and to determine if the agency should make additional improvements to its regulatory system. The NTTF developed a set of recommendations intended to clarify and strengthen the regulatory framework for protection against natural phenomena.

Subsequently, the NRC issued a 10 CFR 50.54(f) letter on March 12, 2012 [1], requesting information to assure that these recommendations are addressed by all U.S. nuclear power plants. The letter (commonly referred to as the "50.54(f) letter") requests that licensees and holders of construction permits under 10 CFR Part 50 reevaluate the seismic hazards at their sites against present-day NRC requirements and guidance.

A comparison between the reevaluated seismic hazard and the design basis for Vogtle Electric Generating Plant (VEGP) Units 1 and 2 has been performed, in accordance with the guidance in EPRI 1025287, "Screening, Prioritization and Implementation Details (SPID) for the Resolution of Fukushima Near-Term Task Force Recommendation 2.1: Seismic" [2], and previously submitted to NRC [3]. That comparison concluded that the ground motion response spectrum (GMRS),

which was developed based on the reevaluated seismic hazard, exceeds the design basis seismic response spectrum in the 1 to 10 Hz range, and a seismic risk assessment is required. A seismic PRA (SPRA) has been developed to perform the seismic risk assessment for VEGP Units 1 and 2 in response to the 50.54(f) letter, specifically item (8) in Enclosure 1 of the 50.54(f) letter.

This report describes the seismic PRA developed for VEGP Units 1 and 2 and provides the information requested in item (8)(B) of Enclosure 1 of the 50.54(f) letter and in Section 6.8 of the SPID [2]. The SPRA model has been peer reviewed (as described in Appendix A) and found to be of appropriate scope and technical capability for use in assessing the seismic risk for VEGP Units 1 and 2, identifying which structures, systems, and components (SSCs) are important to seismic risk, and describing plant-specific seismic issues and associated actions planned or taken in response to the 50.54(f) letter.

This report provides summary information regarding the SPRA as outlined in Section 2.

The level of detail provided in the report is intended to enable NRC to understand the inputs and methods used, the evaluations performed, and the decisions made as a result of the insights gained from the VEGP Units 1 and 2 seismic PRA. For clarification, throughout the remainder of this report, there are some references to VEGP. While the site will eventually be a four-unit site, for this report, VEGP means Vogtle Electric Generating Plant Units 1 and 2.

Page 4of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 2.0 Information Provided in This Report The following information is requested in the 50.54(f) letter [1], Enclosure 1, "Requested Information" Section, paragraph (8)8, for plants performing a SPRA.

(1) The list of the significant contributors to seismic.core damage frequency (SCDF) for each seismic acceleration bin, including importance measures (e.g., Risk Achievement Worth, Fussell-Vesely and Birnbaum)

(2) A summary of the methodologies used to estimate the SCDF and seismic large early release frequency (SLERF), including the following:

i.

Methodologies used to quantify the seismic fragilities of SSCs, together with key assumptions ii.

SSC fragility values with reference to the method of seismic qualification, the dominant failure mode(s), and the source of information iii.

Seismic fragility parameters iv.

Important findings from plant walkdowns and any corrective actions taken

v.

Process used in the seismic plant response analysis and quantification, including the specific adaptations made in the internal events PRA model to produce the seismic PRA model and their motivation vi.

Assumptions about containment performance (3) Description of the P.rocess used to ensure that the SPRA is technically adequate, including the dates and findings of any peer reviews (4) Identified plant-specific vulnerabilities and actions that are planned or taken Note that 50.54(f) letter Enclosure 1 paragraphs 1 through 6, regarding the seismic hazard evaluation reporting, also apply, but have been satisfied through the previously submitted VEGP Seismic Hazard Submittal [3].

Further, 50.54(f) letter Enclosure 1 paragraph 9 requests information on the Spent Fuel Pool. This information is being submitted separately.

Table 2-1 provides a cross-reference between the 50.54(f) reporting items noted above and the location in this report where the corresponding information is discussed.

The SPID [2] defines the principal parts of an SPRA, and the VEGP SPRA has been developed and documented in accordance with the SPID. The main elements of the SPRA performed for VEGP in response to the 50.54(f) Seismic letter correspond to those described in Section 6.1.1 of the SPID, i.e.:

Seismic hazard analysis Seismic structure response and SSC fragility analysis Systems/accident sequence (seismic plant response) analysis Risk quantification Table 2-2 provides a cross-reference between the reporting items noted in Section 6.8 of the SPID

[2], other than those already listed in Table 2-1, and provides the location in this report where the corresponding information is discussed.

Page 5of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 The VEGP SPRA and associated documentation has been peer reviewed against the ASME/ANS PRA Standard [4] in accordance with the process defined in NEI 12-13 [5], as documented in the VEGP SPRA Peer Review Report. The VEGP SPRA, complete SPRA documentation, and details of the peer review are available for NRC review.

This submittal provides a summary of the SPRA development, results and insights, and the peer review process and results, sufficient to meet the 50.54(f) information request in a manner intended to enable NRC to understand and determine the validity of key input data and calculation models used, and to assess the sensitivity of the results to key aspects of the analysis.

The content of this report is organized as follows:

Section 3 provides information related to the VEGP seismic hazard analysis.

Section 4 provides information related to the determination of seismic fragilities for VEGP SSCs included in the seismic plant response.

Section 5 provides information regarding the plant seismic response model (seismic accident sequence model) and the quantification of results.

Section 6 summarizes the results and conclusions of the SPRA, including identified plant seismic issues and actions taken or planned.

Section 7 provides references.

Section 8 provides a list of acronyms used.

Appendix A provides an assessment of SPRA Technical Adequacy for Response to NTIF 2.1 Seismic 50.54(f) Letter, including a summary of VEGP SPRA peer review.

Page 6of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Table 2-1 Cross-Reference for 50.54(f) Enclosure 1 SPRA Reporting 50.54(f) Letter Reporting Item Description Location in this Report 1

List of the significant contributors Section 5 to SCDF for each seismic acceleration bin, including importance measures 2

Summary of the methodologies Sections 3, 4, 5 used to estimate the SCDF and SLERF 2i Methodologies used to quantify Section 4 the seismic fragilities of SSCs, together with key assumptions 2ii SSC fragility values with Tables 5.4-4 and 5.5-2 provide fragilities reference to the method of (Am and beta), failure mode information, seismic qualification, the and method of seismic quantification of dominant failure mode(s), and fragilities for the top risk significant SSCs the source of information based on the Fussell-Vesely (F-V) risk importance measure. Seismic qualification reference is not provided as it is not relevant to development of SPRA.

2iii Seismic fragility parameters Tables 5.4-4 and 5.5-2 provide fragilities (Am and beta) information for the top risk significant SSCs based on the Fussell-Vesely (FV) risk importance measure.

2iv Important findings from plant Section 4.2 addresses walkdowns and walkdowns and any corrective walkdown insights actions taken 2v Process used in the seismic plant Sections 5.1 and 5.3 provide this response analysis and information quantification, including specific adaptations made in the internal events PRA model to produce the seismic PRA model and their motivation 2vi Assumptions about containment Sections 4.3, 4.4, and 5.5 address performance containment and related SSC performance Page 7of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 2-1 Cross-Reference for 50.54(f) Enclosure 1 SPRA Reporting S0.54(f) Letter Reporting Item Description Location in this Report 3

Description of the process used App. A describes the assessment of SPRA to ensure that the SPRA is technical adequacy for the 50.54(f) technically adequate, including submittal and results of the SPRA peer the dates and findings of any review peer reviews 4

Identified plant-specific Section 6 addresses this topic. No vulnerabilities and actions that vulnerabilities were identified or actions are planned or taken planned as a result of the SPRA.

Table 2-2 Cross-Reference for Additional SPID [2] Section 6.8 SPRA Reporting SPID Section 6.8 Item Ill Description Location in this Report A report should be submitted to the NRC Entirety of the submittal addresses this.

summarizing the SPRA inputs, methods, and results.

The level of detail needed in the submittal should Entirety of the submittal addresses this be sufficient to enable NRC to understand and and identifies key methods of analysis determine the validity of all input data and and referenced codes and standards calculation models used The level of detail needed in the submittal should Entirety of the submittal addresses this.

be sufficient to assess the sensitivity of the results Results sensitivities are discussed in to all key aspects of the analysis section 5.7, SPRA model quantification sensitivities.

The level of detail needed in the submittal should Entirety of the submittal addresses this.

be sufficient to make necessary regulatory decisions as a part of NTTF Phase 2 activities.

It is not necessary to submit all of the SPRA Entire report addresses this. This report documentation for such an NRC review. Relevant summarizes important information documentation should be cited in the submittal, from the SPRA, with detailed and be available for NRC review in easily retrievable information in lower tier form.

documentation Documentation criteria for a SPRA are identified This is an expectation relative to throughout the ASME/ANS Standard [4]. Utilities documentation of the SPRA that the are expected to retain that documentation utility retains to support application of consistent with the Standard.

the SPRA to risk-informed plant decision-making.

Note (1): The items listed here do not include those designated in SPID Section 6.8 as "guidance".

Page 8of124

VEGP Units 1 and 2 10 CFR S0.54(f} NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 3.0 VEGP Seismic Hazard and Plant Response This Section provides summary site information and pertinent features including location and site characterization. The subsections provide brief summaries of the site hazard and plant response characterization.

VEGP is a dual unit Westinghouse 4-loop pressurized water reactor plant located approximately 15 miles east-northeast of Waynesboro, Georgia and 26 miles southeast of Augusta, Georgia, adjacent to the Savannah River. The regional and site (local) geology is described in additional detail in the VEGP NTIF 2.1 Seismic Hazard submittal [3]. VEGP is a soil site with 88 feet of backfill on top of an in-situ strata identified as Blue Bluff Marl (BBM). The following Seismic Category I structures are founded directly on the BBM: the Auxiliary Building (AB), Nuclear ~ervice Cooling Water (NSCW) towers, and instrumentation cavity of the Containment. The remaining Seismic Category I structures are founded on backfill. The soil profile was developed using the original Vogtle Units 1 and 2 borehole data supplemented with the latest borehole data taken for the Vogtle Units 3 and 4 new construction and the Dry Cask Storage facility.

Additional site description and composite profile development are described in the VEGP NTIF 2.1 Seismic Hazard submittal [3].

3.1 Seismic Hazard Analysis This section discusses the seismic hazard methodology, presents the final seismic hazard results used in the SPRA, and discusses important assumptions and important sources of uncertainty.

The seismic hazard analysis determines the annual frequency of exceedance for selected ground motion parameters. The analysis involves use of earthquake source models, ground motion attenuation models, characterization of the site response (e.g. soil column), and accounts for the uncertainties and randomness of these parameters to arrive at the site seismic hazard. Detailed information regarding the VEGP site hazard was provided to NRC in the seismic hazard information submitted to NRC in response to the NTIF 2.1 Seismic information request [3]. That information was used in development of the VEGP SPRA.

3.1.1 Seismic Hazard Analysis Methodology For the VEGP SPRA, the following method was used.

As reported in the VEGP NTIF 2.1 Seismic Hazard submittal [3], the control point (power block) hazard curves were used to develop uniform hazard response spectra (UHRS) and the ground motion response spectrum (GMRS). The UHRS were calculated using log-log interpolation to determine the spectral acceleration at each spectral frequency for the 10-4 and 10-s per year hazard levels. The GMRS was calculated from the 10-4 and 10-5 UHRS at each spectral frequency. The control point elevation is defined in the VEGP NTIF 2.1 Seismic Hazard submittal [3] as being at plant grade at an elevation of 220 feet mean sea level (MSL), consistent with the Plant Vogtle Units 1 and 2 FSAR. Table 2.4-1 and Figure Page 9of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 2.4-1 in the VEGP NTIF 2.1 Seismic Hazard submittal [3] provide the mean UHRS for 10-4 and 10-5 and GMRS accelerations for a range of spectral frequencies.

The Reference Earthquake used in developing the building response, and subsequently in the fragility evaluation corresponds to the 10-4 UHRS at plant grade. The 10-4 horizontal UHRS at plant grade has a PGA of 0.436g.

UHRS were developed at specific horizons at the 10-4 and 10-s hazard levels, along with the corresponding strain-compatible properties. This information was used in developing input motion for soil-structure interaction {SSI) analysis. Two types of motion were developed, outcrop UHRS and truncated soil column response {TSCR).

Similar to the site response analysis described in the VEGP NTIF 2.1 Seismic Hazard submittal [3], the rock high frequency (HF) and low frequency {LF) spectra at 10-4 and 10-5 hazard levels were applied at bedrock and are propagated through two sets of 60 simulated profiles. The 5% damping outcrop acceleration response spectra {ARS) at specific horizons were computed. The log-mean (median) results including strains, shear-wave velocity, and damping, are calculated, along with the corresponding log-standard deviations. At each hazard level, the arithmetic mean HF and LF ARS for the two soil columns are arithmetically averaged resulting in the uniform hazard response spectra at the considered horizon.

To calculate TSCR at the considered horizon, the soil layers above that horizon were truncated and site response analysis was repeated using the iterated strain-compatible properties, resulting from the site response analysis runs using the full soil column, and without further iterations.

The site response analyses [18, 39] and the. fragility notebook [16] provide the horizontal 1E-4 UHRS at the elevations as shown in Figure 3.1-1. Plant grade, El 220 ft, is identified as 0 ft outcrop; the other two horizons are identified as depth below grade.

The VEGP Seismic Hazard Submittal [3] used Approach 3 as defined in NUREG/CR-6728

[34] to incorporate site amplification factors with the site rock hazard to calculate seismic hazard curves at the seven oscillator frequencies {0.5, 1, 2.5, 5, 10, 25 and PGA {100) Hz) at the ground surface. This seismic hazard approach resulted in the reported uniform hazard response spectra (UHRS) and the ground motion response spectrum {GMRS) at the ground surface (EL 220 ft) [3].

To provide UHRS at the ground surface, as well as at other foundation elevations, for the purpose of subsequent soil-structure interaction {SSI) analysis, Approach 2A [34] was used in order to readily obtain strain compatable soil profiles. In the case of the Vogtle site, the difference between the UHRS and GMRS calculated by the two approaches was determined to be insignificant.

The methodology for obtaining the vertical response spectra is discussed in Section 3.1.4.

Page 10 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 -'March 2017 Horizontal Mean 1 E-4 UHRS at Vogtle 1.20 1 E-4 ARS - 0 fl Outcrop

,/

I/ \\

1E-4ARS - 47 fl Outcrop I

1 E-4 ARS - 66 fl TSCR

\\

I'..

1 00 c:

0 h

I\\ '\\

,, ~ -~~

~ 0 80 O;

a;

~ 0 60

a.

V)

Cl c: *a. 0 40 E..

0

~

~

0.20 0.00 0.1

~ ~Ii.lo

~ I

" v v

f I

~ J r

~# '

Frequency [Hz)

Figure 3.1-1 Horizontal Mean lE-4 UHRS at Vogtle Site 3.1.2 Seismic Hazard Analysis Technical Adequacy

\\./ "'

10

\\

~\\

\\ r---. !'-...

~ -... -

100 The VEGP SPRA hazard methodology and analysis associated with the horizontal response spectra at the control point were submitted to the NRC as part of the VEGP Seismic Hazard Submittal [3], and found to be technically acceptable by NRC for application to the VEGP SPRA (29].

The VEGP hazard analysis was also subjected to an independent peer review against the pertinent requirements in the PRA Standard [4]. The SPRA was peer reviewed relative to Capability Category II for the full set of requirements in the Standard and determined to be acceptable for use in the SPRA.

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A.

3.1.3 Seismic Hazard Analysis Results and Insights Table 3.1-1 provides the final seismic hazard results used as input to the VEGP SPRA, in terms of exceedance frequencies as a function of PGA level for the mean and several fractiles. Information on the vertical hazard is discussed in Section 3.1.4.

Uncertainties in the PSHA result from uncertainties in input models and parameters.

These have been investigated for the VEGP SPRA (17]. As expected, background sources were found to have a large contribution to the 10 Hz spectral acceleration (SA) hazard, Page 11 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 3.1-1 VEGP Mean and Fractile Exceedance Frequencies Exceedance Frequencies (/yr)

PGA (g) 0.16 0.5 MEAN 0.84 0.1 7.23E-04 1.49E-03 1.93E-03 3.09E-03 0.15 3.19E-04 7.45E-04 1.llE-03 1.87E-03 0.3 6.09E-05 1.49E-04 2.87E-04 4.50E-04 0.5 1.36E-05 3.52E-05 6.78E-05 9.51E-05 0.75 2.76E-06 8.35E-06 1.49E-05 2.13E-05 1

6.36E-07 2.16E-06 3.72E-06 5.58E-06 1.5 3.84E-08 1.42E-07 2.SOE-07 3.84E-07 2

4.00E-09 1.00E-08 1.60E-08 2.50E-08 3

1.53E-10 2.19E-10 2.23E-10 4.43E-10 and a repeated-large-magnitude-earthquake (RLME) source (Charleston) has a large contribution to the 1 Hz SA hazard. Note that the high frequency hazard is typically dominated by closer, moderate sized (background) earthquakes, and larger distant RLME events tend to be more important to low frequency hazard. For this reason, sensitivities to background sources were investigated for 10 Hz SA, and sensitivities to RLME sources were investigated for 1 Hz SA [17]. Sensitivity to site amplification model was also investigated [18].

The main contributors to hazard uncertainty are the ground motion prediction equations (GMPEs) used for hazard calculations, and the characteristic magnitude of the Charleston RLME source. The GMPEs contribute to uncertainty at both high and low spectral frequencies, at spectral amplitudes corresponding to mean annual frequencies of 10-4 and 10-s. The characteristic magnitude of the Charleston source contributes to uncertainty primarily for low spectral frequencies, because the Charleston source has a lower contribution to hazard at high frequencies.

A review was performed [17] of the earthquake catalog used by EPRI for the 2012 hazard study [35]. It was determined that from January 1, 2009 through February 29, 2016, four earthquakes of magnitude M2.9 or greater were recorded within 320 km of the site.

Because this is considerably lower than the frequency that would be expected from the mean annual rates of seismicity modeled for seismic sources by the EPRI 2012 study [35],

it was concluded that the EPRI 2012 study [35] rates do not under-predict the seismicity observed during the period subsequent to that study. Furthermore, given the relatively short period of additional time covered by the updated catalog compared to the total period of time covered by the EPRI 2012 catalog [35], extending the catalog and re-computing new seismicity rates would result in only a very slight decrease in the activity rate in the study region. It was therefore concluded that the EPRI 2012 [35] seismicity parameters are adequate for evaluation of the seismic hazard at VEGP.

In the SPRA plant model, described in Section 5, the hazard data in Table 3.1-1 was discretized into 14 intervals, with parameters as listed in Table 3.1-2.

Page 12 of 124

VEGP Units 1 and2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal" Version 0 - March 2017 Table 3.1-2 Acceleration Intervals and Interval Frequencies as Used in SPRA Model Interval Interval Interval Upper Representative Interval Mean Designator Lower Bound Bound Magnitude PGA (g)

Frequency

%G01 0.1 0.15 0.12 8.20E-04

%G02 0.15 0.3 0.21 8.23E-04

%G03 0.3 0.4 0.35 1.52E-04

%G04 0.4 0.5 0.45 6.71E-05

%GOS 0.5 0.6 0.55 3.18E-05

%G06 0.6 0.7 0.65 1.61E-05

%G07 0.7 0.8 0.75 8.64E-06

%G08 0.8 0.9 0.85 4.79E-06

%G09 0.9 1

0.95 2.71E-06

%G10 1

1.1 1.05 1.55E-06

%G11 1.1 1.2 1.15 9.00E-07

%G12 1.2 1.5 1.34 1.02E-06

%G13 1.5 2

1.73 2.34E-07

%G14 2

2.2 1.60E-08 3.1.4 Horizontal and Vertical Response Spectra This section provides the control point horizontal and vertical response spectra.

The 1E-4, 1E-5 and 1E-6 UHRS, along with the GMRS, at the control point are plotted in Figure 3.1-2. The development of the control point response spectra is described in detail in the VEGP NTIF 2.1 Seismic Hazard submittal [3].

The vertical response spectra were developed based on the corresponding horizontal response spectra, by scaling with an appropriate V /H function. The development of the V /H function is documented in the site response analysis [18] and the Vogtle Electric Generating Plant Units 3 and 4 Early Site Permit application (ESP) [36]. The acceptance of the Vogtle 3 and 4 ESP V/H function is provided in the NRC SER, NU REG 1923 [37].

Table 3.1-3 summarizes the horizontal and vertical response spectra at the control point.

Figure 3.1-3 is a plot of the V /H function. Figure 3.1-4 provides a plot of the horizontal and vertical mean 1E-4 UHRS at the control point.

Page 13 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Mean Soil UHRS and GMRS at Vogtle Ge c'

1.

.2 1E-6 UHRS

~

l E-S UHRS

§ G RS ftJ

~

lE-4 UHRS

~ 0.1 VI 0.01 -+----~~~~---~----~~-~------.

0.1 1

10 100 Spectral frequency, Hz Figure 3.1-2 lE-4, lE-5, AND lE-6 UHRS AND GMRS AT THE CONTROL POINT 1.00

[).I()

0.1

~ -*

I I

,,,,, ~

I

~

- H-

• h--

I I()

frequency (Hz)

Figure 3.1-3 VERTICAL TO HORIZONTAL (V/H ) RATIO FUNCTION

_ j_,._

100 Page 14 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 3.1-3 Horizontal and Vertical Response Spectra at the Control Point Frequency Horizontal Horizontal Horizontal V/H Vertical Vertical Vertical lE-4 UHRS lE-5 UHRS GMRS lE-4 lE-5 GMRS (Hz)

(g)

(g)

(g)

Ratio UHRS (g)

UHRS (g)

(g) 100 4.36E-01 8.14E-01 4.36E-01 0.9 3.92E-01 7.33E-01 3.92E-01 90 4.38E-01 8.20E-01 4.38E-01 0.9 3.94E-01 7.38E-01 3.94E-01 80 4.41E-01 8.27E-01 4.41E-01 0.9 3.97E-01 7.44E-01 3.97E-01 70 4.47E-01 8.36E-01 4.47E-01 0.9 4.02E-01 7.52E-01 4.02E-01 60 4.58E-01 8.49E-01 4.58E-01 0.9 4.12E-01 7.64E-01 4.12E-01 50 4.80E-01 8.71E-01 4.80E-01 0.9 4.32E-01 7.84E-01 4.32E-01 40 5.34E-01 9.18E-01 5.34E-01 0.9 4.81E-01 8.26E-01 4.81E-01 35 5.83E-01 9.64E-01 5.83E-01 0.9 5.25E-01 8.68E-01 5.25E-01 30 6.51E-01 1.04E+OO 6.SlE-01 0.9 5.86E-01 9.36E-01 5.86E-01 25 7.48E-01 1.17E+OO 7.48E-01 0.9 6.73E-01 1.05E+OO 6.73E-01 20 8.83E-01 1.36E+OO 8.83E-01 0.9 7.95E-01 1.22E+OO 7.95E-01 15 1.02E+OO 1.65E+OO 1.02E+OO 0.9 9.18E-01 1.49E+OO 9.18E-01 12.5 1.07E+OO 1.82E+OO 1.07E+OO 0.865 9.26E-01 1.57E+OO 9.26E-01 10 1.09E+OO 1.91E+OO 1.09E+OO 0.824 8.98E-01 1.57E+OO 8.98E-01 9

1.09E+OO 1.95E+OO 1.09E+OO 0.806 8.79E-01 1.57E+OO 8.79E-01 8

1.07E+OO 2.00E+OO 1.07E+OO 0.785 8.40E-01 1.57E+OO 8.40E-01 7

1.02E+OO 1.95E+OO 1.03E+OO 0.763 7.78E-01 1.49E+OO 7.86E-01 6

9.36E-01 1.84E+OO 9.64E-01 0.738 6.91E-01 1.36E+OO 7.llE-01 5

8.76E-01 1.77E+OO 9.21E-01 0.709 6.21E-01 1.25E+OO 6.53E-01 4

9.03E-01 1.80E+OO 9.39E-01 0.676 6.lOE-01 1.22E+OO 6.35E-01 3.5 8.33E-01 1.76E+OO 9.09E-01 0.656 5.46E-01 1.15E+OO 5.96E-01 3

7.62E-01 1.67E+OO 8.55E-01 0.635 4.84E-01 1.06E+OO 5.43E-01 2.5 6.69E-01 1.42E+OO 7.31E-01 0.61 4.08E-01 8.66E-01 4.46E-01 2

4.87E-01 1.16E+OO 5.87E-01 0.581 2.83E-01 6.74E-01 3.41E-01 1.5 4.39E-01 8.55E-01 4.49E-01 0.546 2.40E-01 4.67E-01 2.45E-01 1.25 4.06E-01 8.99E-01 4.60E-01 0.525 2.13E-01 4.72E-01 2.42E-01 1

2.21E-01 5.53E-01 2.76E-01 0.5 1.llE-01 2.77E-01 1.38E-01 0.9 2.00E-01 4.81E-01 2.42E-01 0.5 1.00E-01 2.41E-01 1.21E-01 0.8 2.00E-01 4.60E-01 2.33E-01 0.5 l.OOE-01 2.30E-01 1.17E-01 0.7 2.17E-01 4.83E-01 2.47E-01 0.5 1.09E-01 2.42E-01 1.24E-01 0.6 2.42E-01 5.41E-01

2. 76E-01 0.5 1.21E-01 2.71E-01 1.38E-01 0.5 2.lOE-01 5.26E-01 2.62E-01 0.5 1.05E-01 2.63E-01 1.31E-01 0.4 1.68E-01 4.20E-01 2.lOE-01 0.5 8.40E-02 2.lOE-01 1.05E-01 0.35 1.47E-01 3.68E-01 1.84E-01 0.5 7.35E-02 1.84E-01 9.20E-02 0.3 1.26E-01 3.15E-01 1.57E-01 0.5 6.30E-02 1.58E-01 7.85E-02 0.25 1.05E-01 2.63E-01 1.31E-01 0.5 5.25E-02 1.32E-01 6.55E-02 0.2 8.40E-02 2.lOE-01 1.05E-01 0.5 4.20E-02 1.05E-01 5.25E-02 0.15 6.30E-02 1.58E-01 7.87E-02 0.5 3.15E-02 7.90E-02 3.94E-02 0.125 5.25E-02 1.31E-01 6.56E-02 0.5 2.63E-02 6.SSE-02 3.28E-02 0.1 3.36E-02 8.41E-02 4.20E-02 0.5 1.68E-02 4.21E-02 2.lOE-02 Page 15 of 124

VEGP Units 1 and 2 10.00 c:

.2 "l;j 1 00 o;..

~

iii

~

Q.

V>

Cl c: *a. 0 10 E..

0

~

~

0 01 10 CFR 50.54{f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Horizontal and Vertical Mean 1 E-4 UHRS at the Control Point f-

~

~-

Hcx-izonal Mean 1E-4 UHRS

- ~

Vertical Mean 1E-4 UHRS

~

~

~ ~ -

I-p

/ -

~

-V

~

v

........... r--

/

./

Vj

/

./_

77-v 1-

~

0 1 10 100 Frequency [Hz]

Figure 3.1-4 Horizontal and Vertical Mean lE-4 UHRS at the Control Point Page 16 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 4.0 Determination of Seismic Fragilities for the SPRA This section provides a summary of the process for identifying and developing fragilities for SSCs that participate in the plant response to a seismic event for the VEGP SPRA. The subsections provide brief summaries of these elements.

4.1 Seismic Equipment List For the VEGP SPRA, a seismic equipment list (SEL) was developed that includes those SSCs that are important to achieving safe shutdown following a seismic event, and to mitigating radioactivity release if core damage occurs, and that are included in the SPRA model. The methodology used to develop the SEL is generally consistent with the guidance provided in the EPRI Seismic PRA Implementation Guide [10].

4.1.1 SEL Development The following is a summary of items considered in developing the SEL [8].

The first step in developing the SEL was to determine the potential initiating events that could occur as a result of a seismic event. Initiating events considered could occur either directly as a result of the earthquake or due to random or consequential events that occur subsequent to the earthquake. The process of identification of potential initiating events used the internal events PRA for guidance.

Based on the internal events PRA and review of other potential seismic initiators, the primary seismic initiators identified were loss of offsite power (LOSP), loss of coolant accidents (LOCAs), reactor pressure vessel failure, secondary line break, and station blackout (SBO). The scope of the SPRA is power operation, therefore low power and shutdown states were not considered.

The safety functions that would be required to respond to the initiating events identified above were determined based on EPRI NP 6041-SL [7] and NU REG 1407 [15]. These safety functions are:

Reactivity control Reactor coolant system pressure control Reactor coolant system inventory control Decay heat removal Containment isolation and integrity The frontline systems used to meet the five safety functions were identified from the VEGP internal events PRA. In addition to the frontline systems, the required support systems were identified. However, unlike the internal events PRA, only systems that do not require offsite power were selected. Because the offsite power grid, switchyard insulators, and large transformers have relatively low seismic capacity, they cannot be Page 17of124

VEGP Units 1 and 2

• 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Versron 0 - March 2017 relied on to provide power after a major earthquake. Only systems that can be supported by the onsite emergency AC power sources are considered. For Vogtle 1&2, the IPEEE Seismic Margin Analysis (SMA) safe shutdown equipment list (SSEL) was used as the initial list of equipment that would be used to mitigate seismic events. The SSEL already considered the five safety functions listed above, and contains useful information such as the equipment building and elevation, the normal and desired position for active components, the official mark number, the equipment drawing reference, and the equipment category.

Enhancements to the SSEL were made for the following reasons:

Systems and equipment have been revised since the SMA SSH was developed.

The scope of the SMA was limited to consideration of two success paths, while the SPRA considers the broader accident sequence paths and associated systems and equipment.

Additional seismic initiators must be considered, such as larger LOCAs, since the SMA only considered loss of offsite power and small LOCA.

The enhancements were identified by using system P&IDs and electrical diagrams to ensure that all necessary components are on the SEL.

The following types of equipment were added to the SEL:

Components required to maintain pressure boundary integrity of the modeled systems.

Active valves (and other components) that may have been screened from the SSEL or internal events PRA model but which could be transferred to an undesired state due to seismic-induced relay chatter.

Reactor coolant system components, including: Reactor pressure vessel (and supports); Reactor internals; Control rods; Steam generators; Reactor coolant pumps (for RCS integrity, since they would not have power); Pressurizer; and Main RCS piping.

Distribution systems (i.e., piping, HVAC ducting, and cable trays), treated as single distributed system entries in the SEL.

Electrical panels, cabinets, and instrument racks need to provide emergency power and control for components on the SEL, including main control room bench boards and reactor protection system (RPS) cabinets.

Equipment or instrumentation that would be required per the plant emergency procedures after an earthquake.

In addition, the plant areas in which operators would need to perform seismic response actions were reviewed for accessibility and evaluated for potential impact.

Page 18 of 124

VEGP Units 1 and 2

• 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Components required for maintaining containment integrity were included in the SEL.

These include SSCs related to containment isolation (such as containment isolation signals and valves) and containment pressure suppression and heat removal (such as the containment fan cooler units and containment sprays).

The structures associated with the SEL equipment are the following:

Auxiliary Feedwater (AFW) Pump House Auxiliary Building Containment (Reactor Building)

Control Building Diesel Generator Building Fuel Handling Building Nuclear Service Water Cooling Towers The following types of equipment were not included on the SEL based on their having very high seismic capacity, and their passive nature:

Check valves and backdraft dampers Manual valves and dampers, including fire dampers Small spring-operated relief valves Small passive in-line filters that are supported only by the piping or ducting Heat tracing In addition, instrumentation that is not required for mitigation of the seismic accident sequence (generally local instrumentation that is not part of a plant procedure that would be implemented during a seismic event) was not included in the SEL.

Equipment that is captured through "rule-of-the-box" considerations, e.g., equipment contained on a skid or in a cabinet, that can be subsumed into the major skid equipment or into the cabinet, was also not explicitly included on the SEL. For such equipment, the seismic fragilities for the containing equipment consider all of the equipment in the "box."

As a check on the SEL, the list of basic events in the internal events PRA was reviewed to identify additional systems and equipment that should be included in the SPRA, and the SEL. Systems and components that rely on offsite power were excluded.

The resulting SEL for each unit includes approximately 950 components for each unit.

Page 19 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal

' Version 0 - March 2017 4.1.2 Relay Chatter/Spurious Breaker Trip Evaluation During a seismic event, vibratory ground motion can cause relays to chatter. The chattering of relays potentially can result in spurious signals to equipment. Most relay chatter is either acceptable (does not impact the associated equipment), is self-correcting, or can be recovered by operator action.

An extensive relay chatter evaluation was performed for the VEGP SPRA, in accordance with SPID [2] Section 6.4.2 and ASME/ANS PRA Standard [4] Section 5-2.2. The evaluation resulted in most relay chatter scenarios screened from further evaluation for reasons such as no impact to component function, other components in the circuit that would prevent undesired impact, a self-correcting condition in which a signal would restore the proper position, or seismic qualification as part of the equipment containing the relay.

The unscreened relays in each unit were considered in the SPRA fragility and evaluated for inclusion in the model. The relays that were ultimately included in the SPRA model are listed in table 4.1-1.

A systematic evaluation of spurious trips of breakers was also performed for low and medium voltage switchgear. The functionality of breakers was evaluated either through the EPRI NP 6041-SL [7] value or through test response spectra evaluation. The major types of breakers at the plant are vacuum, air and molded case circuit breakers. Molded case circuit breakers inherently have high seismic capacity. The switchgear which houses air and vacuum type breakers are evaluated through EPRI NP 6041-SL [7] proxy evaluation. The seismic capacities used in this evaluation are the same or lower than EPRI generic equipment ruggedness spectra (GERS) given for low and medium voltage switchgears in EPRI NP-5223-SL [25].

This evaluation meets the intent of the high frequency screening requirement in Section 3.4.1 of the SPID [2]. Section 4.4.2 of this report provides discussion on seismic fragility evaluation of the critical relays.

Table 4.1-1 Summary of Unscreened Relays from Each Unit Included in SPRA Model Relay Function Disposition AFW ADV Trip Relay Trips the AFW Turbine Modeled in fault tree for seismic failure Trip and throttle valve with separation of variables (SOV) fragility and operator recovery Emergency Diesel Trips the EOG Modeled in fault tree for seismic failure Generator (EOG) Engine with SOV fragility Protective Relays Page 20 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 4.2 Walkdown Approach This section provides a summary of the methodology and scope of the seismic walkdowns performed for the SPRA [8]. Walkdowns were performed by personnel with appropriate qualifications. as defined in the SPID [2] Table 6-5 and the associated requirements in the PRA Standard (4). The seismic review team (SRT) was comprised of several seismic engineering experts with extensive experience in fragility assessment. Walkdowns of those SSCs included on the seismic equipment list were performed, as part of the development of the SEL, to assess the as-installed condition of these SSCs for use in determining their seismic capacity and performing initial screening, to Identify potential II/I spatial interactions and look for potential seismic-induced fire/flood interactions. The fragilities walkdowns were performed in accordance with the criteria provided in EPRI NP 6041-SL [7] and SQUG guidance [30).

The information obtained was used to refine the SEL, and provide input to the fragilities analysis and SPRA modeling (e.g.,

regarding correlation and rule-of-the-box considerations).

The seismic fragility walkdowns were conducted on a mixture of both Unit 1 and Unit 2 equipment. The fragility walkdowns included the evaluation of seismic interactions, including the effects of seismic-induced fires and flooding. The SRT was comprised of several seismic engineering experts with extensive experience in fragility assessment.

In addition to evaluating individual components and associated systems on the SEL, the walkdown reviewed the fire protection system. The fire protection piping was found to be well supported and not susceptible to anchorage failures.

The fire detection equipment was found to be ruggedly mounted and no concerns with seismically induced inadvertent initiation were identified.

The major concern for seismic-induced fires is from flammable liquids and gases. Thus, the walkdown focused on these sources and their proximity to components on the SEL.

Potential fires in the turbine building (and yard areas), hydrogen cylinders outside the MSIV area, transformers in SEL buildings, and lube oil for the diesel generators are examples of scenarios that were evaluated.

The potential for seismically-induced flooding was also evaluated. During the walkdowns, potential spray and flooding scenarios from piping systems and SEL components was reviewed.

Particular emphasis was placed upon threaded or jointed piping.

Flood sources, including the fire-protection system, the turbine building, large tanks, were evaluated.

While most of the SEL components were robust, it was identified that anchorage failure could lead to subsequent flooding scenarios. Identified scenarios that were included in the model are the following:

Seismic failure of the essential service water (ESW) chillers causing a NSCW flood in the control building; Page 21 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Seismic failure of the auxiliary component cooling water (ACCW) heat exchangers causing a NSCW flood in the auxiliary building; Seismic failure of the aux cooling units or containment cooling units causing a NSCW flood in the containment.

4.2.1 Significant Walkdown Results and Insights Consistent with the guidance from NP 6041-SL [7], no significant findings were noted during the VEGP seismic walkdowns.

Components on the SEL were evaluated for seismic anchorage and interaction effects in accordance with SPID [2] guidance and ASME/ANS PRA Standard [4] requirements. The walkdowns also assessed the effects of component degradation, such as corrosion and concrete cracking, for consideration in the development of SEL fragilities. In addition, walkdowns were performed on operator pathways, and seismic-induced fire and flooding scenarios were assessed, and potential internal flood scenarios were incorporated into the VEGP SPRA model. The walkdown observations were used in developing the SSC fragilities for the SPRA.

4.2.2 Seismic Equipment List and Seismic Walkdowns Technical Adequacy The VEGP SPRA SEL development and walkdowns were subjected to an independent peer review against the pertinent requirements (i.e., the relevant SFR and SPR requirements) in the PRA Standard [4].

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the VEGP SPRA SEL and seismic walkdowns are suitable for this SPRA application.

4.3 Dynamic Analysis of Structures This section summarizes the dynamic analyses of structures that contain systems and components important to achieving a safe shutdown.

4.3.1 Fixed-base Analyses Since VEGP is a deep soil site, fixed-base analyses were not applicable.

4.3.2 Soil Structure Interaction (SSI) Analyses VEGP is a deep soil site where all safety-related structures are founded on or embedded in the soil where significant soil-structure-interaction (SSI) effects are expected [16]. The effects of uncertainty in SSI are dominated by the soil response. Uncertainty in the VEGP seismic analysis was accounted for by evaluating the SSI model for three soil columns, best estimate (BE), upper bound (UB), and lower bound (LB) (best estimate, upper bound, and lower bound) in accordance with NRC Standard Review Plan (SRP) Section 3.7.2 [26].

These three soil columns account for the measured variation in site-specific soil Page 22 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 properties and account for most of the uncertainty in seismic response of the VEGP structures.

The ground motion input was developed using the site-specific BE, UB, and LB shear wave velocity profiles, and strain-compatible damping ratio profiles.

The reference level earthquake ground motion corresponds to the lE-4 UHRS at plant grade, which has a horizontal PGA of 0.436g. The SASSl2010 analysis code was used to perform the SSI analysis and the depth of soil considered was at least three times the maximum foundation dimension below the foundation {ASCE 4-98 [22]). Cutoff frequency for the SSI analysis was chosen to be 30 Hz, as all input motion spectra show that the energy content decreases above 20 Hz and completely fades at frequencies above 30 Hz. The SSI models were sufficiently refined to transmit frequencies up to 30 Hz through the soil-foundation interface.

Preliminary SSI analyses with the assumption of uncracked section for concrete elements were performed using three soil conditions {BE, UB, LB). In accordance with ASCE 4-98

[22], response spectrum or time history analysis was performed to confirm locations of cracked concrete. If the predicted stresses exceeded ASCE 4-98 limits, then the highly stressed regions were re-analyzed with reduced stiffness to simulate concrete cracking.

The SSI analysis for the surface founded structures utilized the SASSI Direct Method. The SSI analysis for the deeply embedded structures relied on the SASSI Modified Subtraction Method {MSM) or Extended Subtraction Method 2 {ESM). The size of the model did not permit the use of the SASSI direct method for the embedded structures. In addition, sensitivity studies were performed to confirm the accuracy of the MSM and ESM results.

4.3.3 Structure Response Models This section summarizes the Seismic Structure Response and Soil Structure Interaction Analysis methodology used, discusses significant/ limiting seismic structure response and structure fragility results for the SSCs modeled in the SPRA, discusses important assumptions and important sources of uncertainty, and describes any particular fragility-related insights identified.

The seismic structure response analysis considers the impact of seismic events on the response of site structures containing systems and components important to achieving a safe shutdown. VEGP is characterized as a deep soil site based on the site-specific best-estimate lE-4 UHRS shear wave velocity profile, which does not exceed 3,000 fps for a depth of 1,000 feet below grade elevation, see Figure 4.3-1 [16].

The in-structure response spectra {ISRS) for structures considered in the seismic PRA were developed using time-history analysis. Both horizontal and vertical ISRS were computed from time-history motions at various floors or other important locations. In-structure response was generated by applying five sets of input motions, each set was applied to simulate fault normal conditions and then applied to simulate fault parallel conditions.

At representative node locations, various damping acceleration response spectra (ARS) in Page 23 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 the three orthogonal directions are calculated for each of the three directions of the input ground motion. Selection of the locations at which the responses were calculated was based on the equipment location within the building. The ARS are calculated at 301 frequency points equally distributed on the logarithmic scale at the range of frequency from 0.1 Hz to 100 Hz. The responses obtained for the three directions of the input ground motion are combined using the square root sum of the squares (SRSS) method as follows:

where ARS(m)(n) are the SASSI ARS results for the response in "n" direction due to earthquake in "m" direction.

Median and g4th percentile seismic demands (ISRS) are computed using the following procedure:

Step 1: For each soil case and time history set, the ARS from the three earthquake components are combined using the SRSS method as explained above.

Step 2: The SRSS combined demands for each of the time history sets corresponding to a single soil case are averaged to get the median demands for that particular soil case.

Step 3: The median demands for the three soil cases are averaged to get the final median seismic demands. The median demands for the three soil cases are enveloped to get the final 84th percentile seismic demands.

ISRS with highly amplified narrow frequency content was clipped for comparison to broad-banded test response spectra; typical of most NPP components. The guidance in EPRI TR-103959 [21] was performed for the peak clipping process.

The seismic models were based on recent NRC guidance (SRP [26]) and industry codes and standards (ASCE 4-98 [22], and ASCE 43-05 [23]). Building models were developed and median centered response analyses including soil-structure-interaction effects were performed to determine seismic response of SSCs for the lE-4 uniform hazard response spectra input motion.

Page 24 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Vogtle - SSI Profiles Shear-Wave Velocity [ft/sec]

o 1000 2000

JJOO 4000 5000 0000 1000 rooo 0000 10000 0

500

\\'4

~

LB 1E-4

~'::.**

BE1E-4 J i":t i;.

• UB1E-4 Lf;, ~'-:-.

LB 1E-5 I ~ "

BE1E-5

'rl

• I ~

I

• UB 1E-5 I

\\1,,

I ~ j I 1000

..&. T":! =- --

,,,__ ~

""t.....

......... = ~ K

-l

'L.

..,_ Ii....,~ 11

• ,\\

~

I

~

l l

-5 0..

13 1500

(

I 2000

'. d 2500 Figure 4.3-1 Vogtle SSI Profile Simple structures, such as tanks, fuel handling building, and auxiliary feedwater pump house were modeled with lumped-mass stick models (LMSM). More complex structures, such as the Control Building, Auxiliary Building, Diesel Generator Building, Containment Building, and NSCW tower were all modeled with three-dimensional finite element models (FEMs). These detailed building models were sufficiently refined to capture building torsion, out-of-plane floor response, and in-plane floor diaphragm stiffness.

Consideration of secondary system masses (point loads and assumed live load distribution) was performed in accordance with ASCE 4-98 [22].

The compressive strength for concrete material was building-specific and ranged from 4,000 psi to 6,000 psi. The yield strength for steel structures was 36,000 psi. FEM analysis model verification was performed by comparing fixed-base fundamental frequencies with those of the design basis LMSM. Static analyses were also performed in which lg acceleration Page 25 of 124

VEGP Units 1 and 2 10 CFR 50.54{f) NlTF 2.1 Seismic PRA Submittal Version 0 - March 2017

• forces were independently imposed in each of the orthogonal directions and the results of these analyses were reviewed to confirm that the model reasonably represents the fixed-base structure behavior.

Consideration of concrete cracking and structural damping was performed in accordance with ASCE 4-98 (22] and ASCE 43-05 (23]. The effects of concrete cracking and reduced stiffness were addressed by checking for stresses that exc.eeded code limits and then reducing the stiffness of those elements. In addition, material damping in cracked regions was increased from 4% to 7%.

Reviewing transfer functions is an important task when performing SSI analysis. The results of the SSI analysis were validated by carefully reviewing the behavior of transfer functions in all directions and soil cases. The functions were reviewed to confirm low frequency response (i.e., transfer function values should approach 1.0 at low frequency) and to confirm the reasonableness of amplification with increased building elevation. Soil column modes, predicted by the SASSI analysis, were verified by comparing against soil column frequency relationships in ASCE 4-98 (22].

Table 4.3-1 summarizes the type of analysis and model used for each of the structures modeled in the SPRA.

Table 4.3-1 Description of Structures and Dynamic Analysis Methods for VEGP SPRA Foundation Type of Analysis Method Comments/Other Structure Condition Model Information Containment Soil FEM Deterministic SSI LB, BE, U B cases, 5 sets of Building

~ime histories {T-H), used Auxiliary Building Soil FEM Deterministic SSI LB, BE, UB cases, 5 sets of T-H used Control Building Soil FEM Deterministic SSI LB, BE, U B cases, 5 sets of T-H used Fuel Handling Soil LMSM Deterministic SSI LB, BE, UB cases, Building 5 sets of T-H used Diesel Generator Soil FEM Deterministic SSI LB, BE, UB cases, Building 5 sets ofT-H used Auxiliary Feedwater Soil LMSM Deterministic SSI LB, BE, UB cases, Pumphouse 5 sets ofT-H used NSCWTower Soil FEM Deterministic SSI LB, BE, UB cases, 5 sets ofT-H used Condensate Storage Soil LMSM Deterministic SSI LB, BE, UB cases, Tanks 5 sets of T-H used Refueling Water Soil LMSM Deterministic SSI LB, BE, UB cases, Storage Tank 5 sets of T~H used Reactor Make-up Soil LMSM Deterministic SSI LB, BE, UB cases, Water Storage Tank 5 sets of T-H used Page 26 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - Ma*rch 2017 4.3.4 Seismic Structure Response Analysis Technical Adequacy The VEGP SPRA Seismic Structure Response and Soil Structure Interaction Analysis were subjected to an independent peer review against the pertinent requirements in the PRA Standard [4].

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the VEGP SPRA Seismic Structure Response and Soil Structure Interaction Analysis are suitable for this SPRA application.

4.4 SSC Fragility Analysis The SSC seismic fragility analysis considers the impact of seismic events on the probability of SSC failures at a given value of a seismic motion parameter, such as peak ground acceleration (PGA), peak spectral acceleration, floor spectral acceleration, etc. The SSC seismic fragility evaluations performed for VEGP anchors the probability of SSC failures to the horizontal PGA of 0.436g, which corresponds to the lE-4 UHRS at plant grade. The fragilities of the SSCs that participate in the SPRA accident sequences, i.e., those included on the seismic equipment list (SEL) are addressed in the model. Seismic fragilities for the significant risk contributors, i.e., those which have an important contribution to plant risk, are intended to be generally realistic and plant-specific based on actual current conditions of the SSCs in the plant, as confirmed through the detailed walkdown of the plant.

This section summarizes the fragility analysis methodology, presents a tabulation of the fragilities (with appropriate parameters i.e., Am, ~r, ~u), and the calculation method and failure modes for those SSCs determined to be sufficiently risk important, based on the final SPRA quantification (as summarized in Section 5). Important assumptions and important sources of uncertainty, and any particular fragility-related insights identified, are also discussed.

4.4.1 SSC Screening Approach The seismic logic model (described in Section 5) was developed in parallel with the fragility analyses, starting with the initial judgments of seismic capacity based on the rough estimation (rugged, high, medium, low) of the walkdowns.

Many of the components on the walkdown list were screened out from explicit seismic modeling in the quantification based on their rugged seismic capacity. For example, virtually all of the instrumentation was assessed qualitatively to have seismically rugged anchorage, such that no quantitative evaluation was needed. The initial logic model included the seismic failures of the low and medium capacity equipment, including most of the electrical switchgear, cabinets, and panels, which were initially judged to have medium capacity.

As the fragility analyses proceeded, a screening criterion was established based on potential contribution to SCDF. It was determined that a component (or correlated group of components) with median capacity of 2.5g (assuming ~c of 0.3) could contribute about 2E-07 /yr to SCDF if its failure led directly to core damage, which is conservative for many Page 27of124

VEGP Units 1 and 2 "10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 components. With respect to the final SCDF, the maximum contribution for screening at a seismic capacity of 2.5g is 7%. Since this contribution is significant, the screening value was adjusted until the maximum contribution was 2% of the final SCDF, with the result being a 3g screening level. The equipment with seismic capacities between 2.5g and 3g were evaluated to determine if there was the potential for significant improvement in realistic fragilities through additional refinements in fragility calculations. Equipment in this group that could not refined beyond 3g were included in the model or screened out from the model on a system response basis. Exceptions to this screening were made for components that were known to be important (such as the diesel generators), or potentially had significant SLERF impacts. Another exception was made for the structures housing SEL equipment. These were all included in the seismic logic model, except for the AFW pumphouse. The seismic fragility analysis demonstrated that the AFW pumphouse had very high seismic capacity (> 5g), and could be screened from the logic model.

4.4.2 SSC Fragility Analysis Methodology Seismic fragility evaluations were performed for VEGP SSCs contributing to core damage and large early release. The SSC fragility analysis was performed in accordance with Section 6.4.3 of the SPID [2] and the requirements defined in Section 5-2.2 of the ASME/ANS PRA Standard [4]. For fragility evaluation guidance, the SPID recommends Seismic Fragility Applications Guide Update (EPRI 1019200 [19]), Seismic Fragility Application Guide (EPRI 1002988 [20]), Methodology for Developing Seismic Fragilities (EPRI TR-103959 [21]), and A Methodology for Assessment of Nuclear Plant Seismic Margin (EPRI NP 6041-SL [7]). The VEGP fragility analysis is based on these documents, among other industry codes and standards.

VEGP fragility parameters for SSCs were developed based on the following:

Plant-specific design information.

Use of conservative generic fragilities (e.g., EPRI proxy methods).

The hybrid method outlined in the Seismic Fragility Application Guide (EPRI 1002988

[20]) and in Section 6.4.1 and Table 6-2 of the SPID [2].

The more-detailed separation of variables approach outlined in Methodology for Developing Seismic Fragilities (EPRI TR-103959 [21]).

Critical failure modes were identified and seismic fragility calculations were performed to estimate three important fragility parameters: median capacity (Am), and logarithmic standard deviations for randomness and uncertainty (~rand ~u). These three parameters provide sufficient information to construct a family of fragility curves for use in the SPRA logic model. In instances where a fragility estimate resulted in the SSC's contribution to SCDF and/or SLERF being significant, refinement was performed to better estimate the median capacity.

Page 28 of 124

I VEGP Units 1 and 2

• 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 A detailed evaluation was performed for VEGP Units 1 and 2 in order to determine how similar the two units are. The evaluation concluded that the two units are sufficiently identical and Unit 1 results can be applicable to Unit 2.

Structures The VEGP Seismic Category I structures evaluated are:

• Containment Building Auxiliary Building

• Control Building Fuel Handling Building NSCW Tower and Valve House Diesel Generator Building

• Auxiliary Feedwater Pumphouse Condensate Storage Tanks Refueling Water Storage Tank, and Reactor Make-up Water Storage Tank.

Structural demands (member forces and acceleration response) required for fragility analysis were derived from seismic models based on recent NRC guidance (SRP [26]) and industry codes and standards (ASCE 4-98 [22], and ASCE 43-05 [23)). These seismic models were detailed, three-dimensional, finite element models or LMSM (see Table 4.3-1) based on plant-specific information and used the lE-4 uniform hazard as input motion. As VEGP is a deep soil site, these models accounted for the effects of soil-structure-interaction. The effects of building stability, such as sliding and overturning, were evaluated as well as the potential for differential displacement between buildings.

These effects, along with earthquake-induced settlement and liquefaction were evaluated and shown not to be significant factors in the fragility evaluation.

For cylindrical shell structures, such as tanks and containment, two failure modes (tangential shear failure and flexural failure) were evaluated in accordance with EPRI NP 6041-SL [7], EPRI 103959 [21], and ACI 349 [24). For shear wall structures, three failure modes (diagonal shear cracking, flexure and shear friction) were evaluated in accordance with EPRI NP 6041-SL [7], EPRI 103959 [21) and ACI 349 [24). Inelastic energy absorption, which accounts for additional capacity due to ductile design detailing, was considered in accordance with EPRI NP 6041-SL Rl [7] and ASCE 43-05 [23).

Components The VEGP component fragilities were derived using a multi-step approach. The EPRI Proxy Method described in EPRI 1019200 [19) was used to develop and assign fragilities to the Page 29 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 components as the first step, which included some conservative simplifying assumptions.

The EPRI Proxy Method uses the capacity based on EPRI NP 6041-SL [7] and plant-specific demands. The fragility parameters for certain risk-significant components (i.e., important contributors to SCDF and/or SLERF) were then refined to become more plant-specific and realistic. When EPRI Proxy Method fragilities were utilized for mechanical and electrical components, the EPRI NP 6041-SL [7] equipment caveats were confirmed to be satisfied.

Realistic component failure modes included anchorage, functional failures, and failure due to seismic interactions. Anchorage capacities typically were calculated based on standard practice and functional capacities were extracted from existing quantification reports. The seismic demands for both anchorage and functional evaluations come from the in-structure response spectra (ISRS). The ISRS is component specific and depends on the location of the component within the building, and is generated from the seismic analysis building models.

Seismic fragility calculations for critical relays were performed, and made use of GERS [25, 32, 33, 40, 41] for seismic capacities. It was confirmed that the relay vintage and model numbers were consistent with each GERS equipment class. The GERS capacities used are lower than, or the same as, the capacities of these relays in the high frequency range [31, 42]. VEGP is a deep soil site and due to the associated soil-structure interaction effects the predominant seismic demand occurs in the low frequency range. Therefore, this evaluation addresses fragility for high frequency sensitive components as discussed in Section 6.4.2 of the SPID [2].

The nuclear steam supply system (NSSS) was evaluated for fragility variables. The primary system includes the reactor vessel, the steam generators, the reactor coolant pumps, a pressurizer, and the piping that connects these components to the reactor vessel. The fragility evaluation of these components was based on scaling of the existing safety analysis results, in accordance with SPID [2] guidance.

Correlation Correlation of components (or common cause failure) was considered in accordance with the ASME/ANS PRA Standard [4]. For the VEGP SPRA, if the equipment was similar in design, with similar anchorage, and located in the same building on the same elevation, then the equipment was assumed to be fully-correlated. In some cases, detailed model results were used to develop location-specific fragilities. These results were used to refine fragility estimates for similar components located on the same floor.

In order to model the potential correlated failures of like components during an earthquake, the following general correlation rule was used:

If the equipment is similar in design, with similar anchorage, and located in the same building on the same elevation, then it is treated as a correlated failure. That is, all of the similar equipment is modeled to fail with the same likelihood from a given challenge. For example, if one 4-kv emergency switchgear fails given a particular Page 30 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 seismic initiator, then the other also fails. In the PRA model, as discussed in Section 5, this one seismic failure would fail both trains of the switchgear.

Otherwise, there is no correlation. For example, the 4-kv switchgear failures are not correlated with the 480-v motor control center failures.

However, there were a few exceptions to this general correlation rule. Because detailed finite element models of the structur_es were developed, the seismic demand at different nodes of the buildings could be determined. Since the seismic fragility of a component is a function of the component seismic capacity and the seismic demand at the component location, similar components at different locations could have different demands, and thus different fragilities. If the difference between fragilities was small, then the components were correlated using the lower fragility value. However, if there was a significant difference in fragilities, then the higher capacity was used to assign a higher correlated fragility to both components, but the lower capacity component was also assigned a unique seismic capacity that only failed that component. Thus, the lower capacity component could fail by itself, but was guaranteed to fail if the higher capacity component was failed. Detailed individual fragilities were not always calculated for every component, so in some cases this more detailed correlation modeling could not be performed, and the general correlation rule was followed.

4.4.3 SSC Fragility Analysis Results and Insights The final set of fragilities for the risk important contributors to SCDF and SLERF are summarized in Section 5, Table 5.4-4 (for SCDF) and Table 5.5-2 (for SLERF). Detailed (separation of variables, SOV) calculations have been performed for the highest risk significant SSCs, as well as for selected other components.

4.4.4 SSC Fragility Analysis Technical Adequacy The VEGP SPRA SSC Fragility Analysis was subjected to an independent peer review against the pertinent requirements in the PRA Standard [4].

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the VEGP SPRA SSC Fragility Analysis is suitable for this SPRA application.

Page 31of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 5.0 Plant Seismic Logic Model This section summarizes the adaptation of the VEGP internal events at power PRA model to create the seismic PRA plant response (logic) model.

The seismic plant response analysis models the various combinations of structural, equipment, and human failures given the occurrence of a seismic event that could initiate and propagate a seismic core damage or large early release sequence. This model is quantified to determine the overall SCDF and SLERF and to identify the important contributors, e.g., important accident sequences, SSC failures, and human actions. The quantification process also includes an evaluation of sources of uncertainty and provides a perspective on how such sources of uncertainty affect SPRA insights.

5.1 Development of the SPRA Plant Seismic Logic Model The VEGP seismic response model was developed by starting with the VEGP internal events at power PRA model of record as of August 31, 2015, and adapting the model in accordance with guidance in the SPID [2] and PRA Standard [4], including adding seismic fragility-related basic events to the appropriate portions of the internal events PRA, eliminating some parts of the internal events model that do not apply or that were screened-out, and adjusting the internal events PRA model human reliability analysis to account for response during and following a seismic event. The model is developed using the EPRI CAFTA software suite. This model does not credit non-permanently installed FLEX equipment, but does include low leakage reactor coolant pump (RCP) seals. Both random and seismic-induced failures of modeled SSCs are included. The modifications to develop the SCDF fault tree are summarized in Table 5.1-1.

For the VEGP SPRA, the following discussion addresses the methods used to develop the seismic plant response model.

Initiating Events and Accident Sequences The seismic hazard was modeled using 14 discrete hazard intervals (or bins) based on increasing peak ground acceleration. The seismic hazard bins are as listed in Table 3.1-2.

Each bin is treated as a seismic initiator and the SCDF (and SLERF) results are summed over all the bins to obtain the total SCDF (and SLERF). Bin-specific SSC fragilities are used in the accident sequences for each bin.

The SPRA models each seismic event (i.e., each bin) as possibly leading to transients and LOCAs (small, medium, large, and excess LOCA (e.g., reactor pressure vessel failure)), with and without onsite AC power, and with response reflecting impact of the seismic event on mitigating systems.

Page.32of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.1-1 Summary of Modifications to Internal Events CDF Fault Tree to Create Seismic CDF Fault Tree Modification Added seismic-induced large loss of coolant accident (LLOCA), medium loss of coolant accident (MLOCA), small loss of coolant accident (SLOCA), LOSP to their respective internal initiator gate logic; Revised the seismic LOSP gate to include the high capacity seismic initiating events (LLOCA, MLOCA, SLOCA).

Revised SBO initiator FT to include the seismic SBO initiator fault tree; Revised SBO fault tree (FT) to include all failures of 4KV emergency buses rather than only long term bus failures.

Removed recovery of AC power; Removed the ability to use Plant Wilson after seismic event; Removed the ability to crosstie Unit 2 EDGs with Unit 1.

Added direct seismic core damage FT for seismic failures of buildings, excessive LOCAs, and loss of instrumentation and control.

Added logic to incorporate consequential SLOCA given a seismic LOSP with inadvertent safety injection signal.

Added two new seismic SBO sequences to reflect core damage with a 21gpm seal LOCA.

Added seismic initiators that could result in an anticipated transient without trip (ATWT) to the existing ATWT sequences; Revised ATWT logic to fail recovery of ATWT by driving CRDs or manually tripping RX in the case of seismic failures of CRD or RV internals; Added seismic SBO ATWT (assumed core damage with no mitigation).

Added seismic failure of the electrical aux board in main control room (MCR) to existing failures.

Added operator actions to start EDG, close breakers, start equipment for sequencer failure.

Added seismic failure of the NSCW piping to the Control Building ESF chillers causing large flooding on Control Building 260' and propagating to core damage.

Added seismic failure of the ACCW heat exchangers as initiator for a flood event, with small and large flood scenarios, with arid without seismic LOCAs, with associated Operator actions.

Added seismic failure of the ACU and CCUs as flood initiators inside containment, with and without seismic LOCA scenarios, with associated Operator actions.

Added seismic failure ofthe CCUs as a flood initiator inside containment, with and without seismic LOCA scenarios. Operator actions were also added.

Certain structural failures are modeled as leading directly to core damage given the potential for multiple system impacts or distributed system failures. These include seismic failure of:

Containment Auxiliary building Control building NSCW cooling towers and basins Page 33 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 In addition, the following failures of instrumentation and control were included in the model, and assumed to lead directly to core damage:

Seismic failure of the main control board Seismic failure of 125vdc control power panels Although not required by the SPID [2], the potential for seismically-induced internal fires and internal floods was evaluated based on walkdown observations and several internal flooding scenarios were developed for inclusion in the SPRA model. These are:

Seismic failure of the ESF chillers causing a NSCW flood in the control building; Seismic failure of the ACCW heat exchangers causing a NSCW flood in the auxiliary building; Seismic failure of the aux cooling units or containment cooling units causing a NSCW flood in the containment.

Modeling of Correlated Components Treatment of correlation of modeled components is discussed in Section 4.4.2. Fully correlated components were assigned to correlated component groups so that all components in the group fail with the same probability based on the seismic magnitude for each hazard bin. The model assumes fully correlated response of same or very similar equipment in the same structure, elevation, and orientation. Correlated component groups were developed for all redundant components in the model that met these correlation criteria. For correlated groups where there was a significant difference in fragilities, then the higher capacity was used to assign a higher correlated fragility to both components, but the lower capacity component was also assigned a unique seismic capacity that only failed that component. Thus, the lower capacity component could fail by itself, but was guaranteed to fail if the higher capacity component was failed.

Modeling of Human Actions Human error probabilities (HEP) for operator actions in the SPRA model are developed using the same methodology as in the internal events PRA. The EPRI Human Reliability Analysis (HRA) Calculator software was used to develop and document the HEPs for the internal events actions and for the limited set of seismic response operator actions. HEPs were then adjusted as a function of seismic magnitude using a performance shaping factor approach consistent with the EPRI seismic HRA methodology [9].

Several seismic specific human actions were also identified in response to seismically induced flooding. The seismic response operator actions are listed in Table 5.1-2.

In the peer reviewed model, several operator actions were found to be risk-significant.

However, this was based on conservative fragility estimates in the peer reviewed model.

Subsequently, in addressing peer review findings, many of the fragilities were refined to reduce conservatism.

These changes in fragility estimation resulted in the seismic operator actions no longer being risk-significant.

Page 34 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.1-2 Seismic Response Operator Actions Basic Event Description HEP S-OA-BKR-LOCAL Failure of Operator action to locally reclose breaker after 2.2E-03 seismic event S-OA-DG-ST ART Operator fails to start and load DG if sequencer fails 6.2E-03 S-OA-TDAFW-Operator fails to reset the TDAFW trip throttle valve after 1.lE-03 RELAY relay chatter due to a seismic event S-OA-ISOL-CIV Operator fails to manually isolate containment isolation 1.6E-02 valves with loss of l&C S-OA-ACCW-1-45 Operator fails to isolate flood and recover NSCW in 45 1.7E-01 minutes (1 hr available)

S-OA-ACCW-1-90 Operator fails to isolate flood and recover NSCW in 90 1.0E+OO minutes (1 hr available)

S-OA-ACCW-2-45 Operator fails to isolate flood and recover NSCW in 45 4.6E-02 minutes (2 hr available)

S-OA-ACCW-2-90 Operator fails to isolate flood and recover NSCW in 90 8.SE-02 minutes (2 hr available)

S-OA-ACCW-4-45 Operator fails to isolate flood and recover NSCW in 45 4.lE-02 minutes (4 hr available)

S-OA-ACCW-4-90 Operator fails to isolate flood and recover NSCW in 90 4.lE-02 minutes (4 hr available)

S-OA-CU-150 L-15 Operator fails to isolate NSCW to CCU in <15min 1.2E-02 A complete dependency analysis was performed on all human actions (including both seismic-specific actions and actions included in the internal events model on which the SPRA is based) required for a response to a seismic event. The results of this dependency demonstrated that those combinations of actions that were identified as dependent were not risk significant (less than 0.5% contribution to SCDF).

SLERF Model The additional seismic initiating events, and their associated accident sequences, added to the core damage model were also added to the seismic LERF model. Each new seismic core damage sequence was mapped to the appropriate SLERF groups based on the mapping in the internal events level 2 PRA. Most core damage sequences went to several SLERF groups depending on failures in the Level 2 event trees from the internal events PRA. Some of the new sequences, such as failure of the containment or steam generators, were directly mapped to SLERF. Others, such as a SBO with 21gpm seal LOCA, were mapped based on similar core damage sequence mapping, using the level 2 event trees in the internal events PRA. The modifications to develop the seismic LERF fault tree are summarized in Table 5.1-3.

Page 35 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.1-3 Modifications to Internal Events LERF Fault Tree to Create Seismic LERF Fault Tree Modification Added LERF gate inputs for sequences that were added to SCDF logic model:

structural failures, l&C failures, 21gpm seal LOCA after SBO, SBO with ATWT, seismic reactor vessel rupture, Steam Generator failure, ACCW floods, ACU and CCU floods Additional SPRA model and quantification assumptions, including treatment of loss of offsite power and seismic induced reactor coolant system leakage, are listed in Section 5.3.2.

5.2 SPRA Plant Seismic Logic Model Technical Adequacy The VEGP SPRA seismic plant response methodology and analysis were subjected to an independent peer review against the pertinent requirements in the PRA Standard [4].

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the VEGP SPRA seismic plant response analysis is suitable for this SPRA application.

5.3 Seismic Risk Quantification In the SPRA risk quantification the seismic hazard is integrated with the seismic response analysis model to calculate the frequencies of core damage and large early release of radioactivity to the environment. This section describes the SPRA quantification methodology and important modeling assumptions.

5.3.1 SPRA Quantification Methodology For the VEGP SPRA, the following approach was used to quantify the seismic plant response model and determine seismic CDF and LERF.

The EPRI FRANX software code was used to discretize the seismic hazard into the 14 seismic initiators, and quantify to produce cutsets and estimate the mean SCDF. The EPRI ACUBE code was then utilized to estimate the CDF/LERF more accurately by calculating the exact probability on the entire set of SCDF/SLERF cutsets. ACUBE does not use the rare events approximation as is utilized in CAFTA's min cut upper bound estimation calculation and so ACUBE provides a more accurate solution. Additional details can be found in the following sections, along with descriptions of sensitivity studies, uncertainty estimations and a more complete description on the insights from top contributors to SCDF/SLERF.

Page 36 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 I

5.3.2 SPRA Model and Quantification Assumptions The following assumptions are important to the seismic PRA model development and quantification:

1. Offsite power cannot be recovered within the 24-hour mission time.

2. Diesel generators and the 4KV emergency switchgear cannot be shared between units.

3. Plant Wilson has a low seismic capacity, and is not available during a seismic event.

4. The potential impacts of aging on equipment are not included in the SPRA.

However, the potential impacts identified in the walkdowns, such as corrosion or concrete cracking, are included when judged to be significant.

5. The potential for a seismic-induced SGTR is assessed to be very low, based on a detailed assessment, and failure of the steam generators with potential LERF is dominated by failure of the steam generator (SG) supports.

6. In a seismic LOSP with failure of the CROM or RV internals, it was assumed that the control rods would not insert, and an ATWT would occur. The internal events PRA did not question ATWT for LOSP sequences. In a LOSP (0.3g), the control rods would be released immediately, so they may insert before the failure of the higher capacity failures of the CROM (2.2g) or RV internals (>4g).

7. It was conservatively assumed that the normal reactor trip signals and reactor trip breakers must work, even after a LOSP. In part, this was to account for potentially delayed LOSP.

8. The seismic capacity for a small-small LOCA is evaluated to be very high, based on walkdown observations, and is not included in the baseline SPRA.

9. The seismic capacity for small LOCA was estimated using the fragility for the seismic capacity reactor coolant pump, which is also used for the medium, large and excess LOCA fragility. This is conservative since it essentially over estimates the effect of failure of the pump.

10. For the NSCW flooding scenarios for failure of the ACCW heat exchangers, it was assumed that the NSCW pumps would have to be stopped in order for the operators to close the manual isolation valves next to the ACCW HX's. These are large valves near the expected rupture flange, and are the only valves that can be used to isolate the ACCW HX's from the NSCW pumps.

11. The NSCW flooding scenario for failure of the ESF chillers in the control building conservatively assumes that the flooding causes loss of instrumentation and control, since the flood would propagate to the control building basement, flooding the electrical rooms.

12. Seismic failure of the auxiliary building is conservatively assumed to result in core damage and large early release. The calculated seismic capacity is based on failure of the entire first story of the building. The story failure would fail the containment penetration areas. The penetrations are actually connected to the containment, and the auxiliary building does not have walls around the containment, just sealant where walls touch. Therefore, these penetrations would be failed when Page 37 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 the auxiliary building collapses. Although an interior wall has lower capacity, which could potentially fail the A train RHR pump and A train pipe chase, the failure of the interior wall would not lead to core damage.

13. A detailed functional fragility analysis could not be performed for the TDAFW pump based on the available information, so the detailed anchorage fragility was used for this pump. This fragility was similar to the controlling fragilities for other large pumps, such as the RHR and SI pumps.

14. The two motor-driven AFW pumps (MDAFWP) were assumed to have a correlated failure based on similar equipment in the same location. The turbine-driven AFW pump (TDAFWP) is located in the transverse direction, and has an entirely different driver, with anchorage differences. Therefore the MDAFWP failure was not correlated with the TDAFWP for the baseline CDF analysis.

5.4 SCDF Results This section presents the base SCDF results, a list of the SSCs that are significant contributors, including risk importance measures, and a discussion of significant sequences/cutsets and their relative SCDF contributions. A discussion of sensitivity studies is provided in Section 5.7.

The VEGP SCDF is 2.8x10-6/yr. Table 5.4-1 presents the 14 most dominant cutsets that each contribute at least 1% or more to seismic core damage frequency. Note that these cutsets have been combined across all the hazard bin intervals. Therefore, they are a summation of all cutsets with the same failures but with different seismic initiators.

The dominant 14 cutsets represent approximately 45% contribution to SCDF. Most ofthe remaining significant cutsets are variations of the top 14 where all the same sequences are represented with varying failures of components leading to the loss of the same function. This is discussed in the following descriptions of the dominant cutsets. Note that the percentage contributions represent the sum of contributions over all the seismic hazard bins for the particular cutset.

The most dominant cutset, representing about 16% of the SCDF, is correlated seismic failure of all four of the 125 VDC lE Distribution Panels which leads to the failure of vital instrumentation and control.

Given this failure, the operators would not have any indication of RCS level, temperature or pressure, so failure of these panels is assumed to lead to failure of automatic and manual response, leading directly to core damage. A sensitivity study has been performed where the median seismic capacity of these panels has been increased to determine if additional insights are masked due to this modeling assumption. Additional failures, such as the Main Control Board (lACBD-MCB), represent an additional 5% contribution to overall SCDF.

The next most dominant cutset (#2) involves anticipated transient without trip (ATWT) sequences where offsite power is lost due to seismic failure, the control rods fail to drop due to seismically-induced mechanical displacement issues, and the operator fails to Page 38 of 124

____ J

VEGP Units 1 and 2 10 CFR S0.54{f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 perform emergency boration in time to reduce reactor power. The sequence represents an inability to control reactivity; leading to core damage. This cutset is approximately 7%

of SCDF and mostly comes from the higher PGA intervals (%G10 and higher). In these high PGA scenarios, the operator action to borate is modeled as guaranteed to fail. Nine of the top 14 cutsets (3, 5, 6, & 9-14) are ATWT sequences where the CRDM seismically fails following LOSP, with a 3rd failure that includes one of the following:

Seismic failure of vital inverter (1ACIV-120-AB220-LC-B)- 4.7%

Seismic failure of diesel generator building fans (1DGFN-FAN) - 2.1%

Seismic failure of diesel generator lube oil (1DGHE-LUBEOIL) -1.8%

Seismic failure of motor drive AFW pumps (1AFPM-MDP) - 1.3%

Seismic failure of Diesel Generator components - 4.8%

The fourth most dominant cutset includes the seismic support failure of all the RCPs. The failure of all the pumps is assumed to lead to a LOCA beyond mitigation capability (excessive LOCA) and core damage.

The contribution from this cutset is approximately 4%.

The seventh and eighth most dominant cutsets involves the seismic failure of the control building ESF Chillers causing an NSCW flood in the control building, and the seismic failure of the main control board, respectively.

These failures lead to the failure of vital instrumentation and control.

The operators will not have any indication on level, temperature or pressure, so failure of these components is assumed to lead directly to core damage. However, the contribut~on from these cutsets is only 1.5% and 1.3%,

respectively.

Additional cutsets involving smaller LOCAs are not represented in the top 14 cutsets but contribute approximately 15% to SCDF. Examples include a seismically induced LOCA with failure of the following components:

Seismic failure of Diesel Generator components Vital AC Inverters Vital DC Buses Page 39 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.4-1 Dominant SCDF Cutsets

• CDF Input 1 Input 2 Input 3 Input 4 Input 5 1

16.2%

4.65E-07 S 1DCBS-PN-CB180-1E SEQ_ DAMAGE 2

7.1%

2.02E-07 PLL S CRDM-ATWT S SEISMIC-LOSP SEQ_ATWT-GT40-7 OA-OBR-------H 3

4.7%

1.35E-07 PLL S_lACIV-120-S_CRDM-ATWT S_SEISMIC-LOSP SEQ_ATWT-GT40-13 AB220-LC-B 4

4.0%

1.13E-07 S_SEISMIC_RCP _ALL SEQ_ DAMAGE 5

2.1%

6.0SE-08 PLL S 1DGFN-FAN S CRDM-ATWT S_SEISMIC-LOSP SEQ_ATWT-GT40-13 6

1.8%

5.19E-08 S_1DGHE-LUBEOIL S CRDM-ATWT S SEISMIC-LOSP SEQ_SEIS-SBO-ATWT 7

1.5%

4.17E-08 S CB-CHLR-NSCW-SEQ_ DAMAGE FLOOD 8

1.3%

3.69E-08 S_1ACBD-MCB SEQ_ DAMAGE 9

1.3%

3.64E-08 PLL S 1AFPM-MDP S CRDM-ATWT S SEISMIC-LOSP SEQ_ATWT-GT40-13 10 1.0%

2.83E-08 PLL S 1DGDM-VENT S CRDM-ATWT S SEISMIC-LOSP SEQ_ATWT-GT40-13 3

11 1.0%

2.83E-08 PLL S 1DGHE-LUBEOIL S CRDM-ATWT S_SEISMIC-LOSP SEQ_ATWT-GT40-13 12 1.0%

2.78E-08 S CRDM-ATWT S DG-BLDG S SEISMIC-LOSP SEQ_SEIS-SBO-ATWT 13 1.0%

2.70E-08 PLL S 1DG S CRDM-ATWT S SEISMIC-LOSP SEQ_ATWT-GT40-13 14 1.0%

2.68E-08 S 1SWFN-NSCW-FANS S_CRDM-ATWT S_SEISMIC-LOSP SEQ_SEIS-SBO-ATWT

• Frequencies are point estimates that do not reflect quantification refinement using ACUBE; however they are valid for relative sequence evaluation; each cutset also includes a plant availability factor basic event (0.9, not shown) to reflect the fraction of time at power.

Descriptions of each basic event listed in the cutsets are provided in Table 5.4-la.

Page 40 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.4-la Basic Event Description Table EVENT DESCRIPTION OA-OBR-------H Seismically induced failure of Operator Fails to perform emergency boration S 1ACBD-MCB Seismically induced failure of Main Control Board S_1ACIV-120-AB220-LC-B Seismically induced failure of Vital AC Inverter 1BD1112 S_lAFPM-MDP Seismically induced failure of BOTH AFW MOP S 1DCBS-PN-CB180-1E Seismically induced failure of 125 VDC 1E Distribution Panel - CB180 S 1DG Seismically induced failure of Both Diesel Generators S lDGDM-VENT-1-3 Seismically induced failure of DG Vent Damper For Fans 1&3 S_lDGFN-FAN Seismically induced failure of DG BLDG ESF Supply Fan S_lDGHE-LUBEOIL Seismically induced failure of DG Lube Oil HX S_lSWFN-NSCW-FANS Seismically induced failure of NSCW Tower Fans S CB-CHLR-NSCW-FLOOD Seismic Failure Of CB ESF Chillers Cause NSCW Flood On CB 260 S CRDM-ATWT Seismically induced failure of ATWT Due To CROM Fail To Drop S DG-BLDG Seismically induced failure of Diesel Buildings S_SEISMIC_RCP _ALL Seismic Failure Of All RCPS S_SEISMIC-LOSP Seismic Loss Of Offsite Power SEQ_ATWT-GT40-13 Seismically induced failure Sequence Label SEQ_ATWT-GT40-7 Seismically induced failure Sequence Label SEQ_ DAMAGE Seismically induced failure Sequence Label SEQ_SEIS-SBO-A TWT Seismically induced failure Sequence Label PLL Fraction of time above power level specified in ATWS model Page 41 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.4-2 shows the contribution from seismically induced initiators that accounts for 90% of total SCDF.

These initiators are described earlier in Section 5.1 and were developed specifically for the SPRA.

Table 5.4-2 Contribution of Seismic Initiators to SCDF Seismic Initiator CDF Seismic LOSP (without other seismic initiators) 9.18E-07 33 Seismic Failure of Instrumentation and Control 5.93E-07 22 (direct core damage)

Seismically Induced ATWT 4.41E-07 16 Seismically Induced LOCA (all sizes) 3.35E-07 19 Table 5.4-3 presents the percentage of the core damage frequency that derives from each interval in the seismic hazard curve.

Also shown are the conditional core damage probability (CCDP) forthe bin, the percent of total SCDF, and the cumulative SCDF. As can be seen, the majority of the contribution comes from seismic events with PGA greater than 0.8g. Bin %G12 contributes the most to the overall SCDF with 22%. From 0.8g to 2g, the contribution totals about 85%, where below 0.8g the contribution is only about 15%.

Importance analyses were performed for both SCDF and SLERF, using the ACUBE code.

From the ACUBE output, Fussell-Vesely (FV) values were determined for each basic event (BE) in the model. Since seismically induced failures require a unique BE for each seismic interval, the FV values for seismic failures for each interval were summed together for each seismic fragility group.

Table 5.4-4 provides the important SCDF contributors, sorted by FV, for SSCs with SCDF FV;:::: 0.02. Because the VEGP SCDF is very low, it was judged that this is a sufficiently low importance target for consideration of important contributors. This table also indicates the seismic fragility parameters for the significant contributors, including median capacity (Am), uncertainty parameters (~r and

~u), failure mode, and method of fragility calculation.

The FV listing shows the top individual contributors to SCDF as seismically induced LOSP, due to the low median seismic capacity assumed for offsite power failure following a seismic event. The fragility for LOSP is a generic value and considered reasonably representative for VEGP.

The next highest contributor is seismically induced correlated failure of all 125 VDC lE Distribution Panels, discussed earlier as a contributor to the dominant seismic cutsets.

Other important contributors are seismically induced failure of the control rod drive mechanism (CROM) resulting in failure to drop the control rods; seismically induced dislocation failure of the RCPs leading to an excessive LOCA; and seismic failures of RCS piping leading to various sizes of LOCA.

The remaining significant components all have relatively low FV contributions.

Page 42 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.4-3 Contribution to SCDF by Acceleration Interval Seismic IE Bin Frequency CCDP SCDF Cumulative SCDF

%G01- {0.lg to <0.15g) 8.20E-04 1.08E-06 8.86E-10 0%

8.86E-10

%G02- {0.15g to <0.3g) 8.23E-04 1.28E-05 1.0SE-08 0%

1.14E-08

%G03- {0.3g to <0.4g) 1.52E-04 9.lSE-05 1.39E-08 0%

2.53E-08

%G04- (0.4g to <0.Sg) 6.71E-05 6.22E-04 4.17E-08 1%

6.70E-08

%GOS- (0.Sg to <0.6g) 3.18E-05 3.00E-03 9.54E-08 3%

1.62E-07

%G06- (0.6g to <0. 7g) 1.61E-05 1.0SE-02 1.69E-07 6%

3.31E-07

%G07- (0.7g to <0.8g) 8.64E-06 3.00E-02 2.59E-07 9%

5.90E-07

%G08- (0.8g to <0.9g) 4.79E-06 6.78E-02 3.25E-07 12%

9.lSE-07

%G09~ (0.9g to <lg) 2.71E-06 1.33E-01 3.60E-07 13%

1.28E-06

%G10- (lg to <1.lg) 1.SSE-06 2.31E-01 3.58E-07 13%

1.63E-06

%G11- (1.lg to <1.2g) 9.00E-07 3.52E-01 3.17E-07 11%

1.95E-06

%G12- (1.2g to <1.Sg) 1.02E-06 6.0lE-01 6.12E-07 22%

2.56E-06

%G13- (1.Sg to <2g) 2.34E-07 8.97E-01 2.lOE-07 8%

2.77E-06

%G14- (>2g) 1.60E-08 9.59E-01 1.53E-08 1%

2.79E-06 Page 43 of 124

VEGP Units 1 and 2 10 CFR 50..54-(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.4-4 SCDF Importance Measures and Fragility Parameters Ranked by FV Component/

Description and Failure Mode FV Am (g)

J3r J3u Failure Fragility Fragility Group Mode Method S_SEISMIC-LOSP SEISMIC LOSS OF OFFSITE POWER 0.33 0.30 0.40 0.27 Generic Generic S_1DCBS-PN-CB180-1E SEISMIC FAILURE OF 125 VDC lE 0.21 2.07 0.25 0.37 Functional sov DISTR. PANEL - CB180 S_CRDM-ATWT SEISMIC FAILURE OF ATWT DUE 0.16 2.42 0.26 0.45 Anchorage sov TO CROM FAIL TO DROP S_SEISMIC_RCP _ALL SEISMIC FAILURE OF ALL RCPS 0.07 2.54 0.30 0.33 Anchorage sov S_SEISMIC-LLOCA SEISMIC INDUCED LLOCA 0.04 2.54 0.30 0.33 Anchorage sov (1)

S_SEISMIC-MLOCA SEISMIC INDUCED MLOCA 0.04 2.54 0.30 0.33 Anchorage sov (1)

S_SEISMIC-SLOCA SEISMIC INDUCED SLOCA 0.04 2.54 0.30 0.33 Anchorage SOV(l)

S_lACBD-MCB SEISMIC FAILURE OF MAIN 0.04 2.86 0.32 0.33 Functional sov CONTROL BOARD SDS-F RCP SHUTDOWN SEAL FAILS TO 0.04 N/A N/A N/A Random N/A ACTIVATE AND SEAL FOR 24 HRS.

S_lAFPM-MDP SEISMIC FAILURE OF BOTH AFW 0.03 1.42 0.32 0.32 Functional EPRI NP 6041-SL MDP RCP SEAL LEAK 21 GPM/PUMP RCPSL-21GPM AFTER 13 MIN. TOTAL LOSS OF 0.03 N/A N/A N/A Random N/A SEAL COOLING S_RX-TRIP-BKRS-SEIS SEISMIC FAILURE OF REACTOR 0.02 3.00 0.32 0.32 Functional EPRI NP 6041-SL TRIP BREAKERS S_lFC-CCU-FLD SEISMIC FAILURE OF MULTIPLE 0.02 1.92 0.28 0.28 Anchorage sov CCU WITH NSCW FLD S_RV-INT-ATWT SEISMIC INDUCED FAILURE OF RX 0.02 3.08 0.31 0.34 Anchorage sov VESSEL INTERNALS S_1DCBS-SGR-CB180 SEISMIC FAILURE OF 125 VDC 0.02 1.98 0.32 0.32 Functional EPRI NP 6041-SL SWITCHGEAR CB180 Note 1: Based on seismic capacity of RCP coolant pump supports. All other NSSS equipment that lead to a LOCA had a greater capacity.

Page 44of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 5.5 SLERF Results This section presents the seismic large early release frequency (SLERF) results, a list of the SSCs that are significant contributors, including risk importance measures, and a discussion of significant sequences/cutsets and their relative SLERF contributions.

Seismic LERF is defined consistent with the internal events model. A 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> time period after event initiation is assumed to allow for evacuation. This time period is considered to be valid for Vogtle seismic events, particularly due to the very low population density in the area. Other characteristics, such as bypass and scrubbing, are the same for seismic as for internal events. The logic for the internal events LERF model is very straightforward, with sequ~nces from the SCDF model ANDed with the appropriate fault tree that models failures leading to bypass of containment.

The VEGP SLERF is 3.3x10-7 /yr. Table 5.5-1 lists the dominant SLERF cutsets aggregated across all the seismic hazard bins. The most dominant SLERF cutset, representing about 8% of the SLERF, is failure of the steam generator piping connections, leading to a LOCA beyond mitigation capability (excessive LOCA) and modeled as a direct bypass due to the assumption of damage to the containment penetrations when the SGs are displaced. The majority of this SLERF contribution comes from the seismic bins %G09 and higher.

Cutsets 2 and 4 are representative of the dominant LERF sequence, and are seismically induced LOCA with seismically induced failure of vital inverters. Cutsets 3, 5, and 6 are similar, but involve failure of the DC switchgear instead of the inverters. This leads to the failure of engineered safety features actuation system (ESFAS) signal to automatically close air-operated containment isolation (Cl) valves. The inverter failure also fails the operator's ability to close the air operated valves (AOVs) manually from the control room.

These AOVs are designed to fail closed on loss of DC power or instrument air, but both could remain available for some time following the seismic event. Therefore, it is assumed that the AOVs will remain open if they fail to receive the isolation signal. This is a potentially conservative modeling assumption, as there are additional ways to verify those valves should be manually closed by the operator. There is also a backup power source, 480V AC power, which could potentially power the ESFAS panels to allow the signal to be sent. Neither of these compensatory measures were included in the model and a sensitivity has been performed to show that the current model is potentially conservative. Cutsets 8 and 9 are similar in impact to Cutsets 2 and 4, except involving failures of AC panels leading to the failures noted above. There are additional cutsets containing higher capacity components that fail the same functions leading to the same accident sequence not represented in the top contributing cutsets. Various combinations of the failure of the vital AC inverters, DC buses and diesel generator components all contribute in lesser amounts to add up to the 60% contribution.

Cutset 7 {3%) is failure of containment, modeled as tangential shear cracking failure, leading to direct bypass. This is dominated by failures in the bins that represent the highest seismic ground motion.

Page 45 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Cutsets 10 through 12 are similar to the scenario in cutsets 2 through 6, 8 and 9, except that for small LOCA, failures of the EDGs are also required.

The listed cutsets contribute approximately 53% of total SLERF. The remaining cutsets constitute small individual SLERF contributions.

SLERF importance measures were calculated in the same manner as for SCDF, and the results are listed in Table 5.5-2.

The FV listing for SLERF indicates that top contributors to SLERF are seismically induced LOCAs of all sizes. In addition, the seismic failure of the 125 VDC Switchgear, CB180 AC Inverter, 120 VAC Panel, battery charger and battery have a high FV values. These SSCs are modeled as leading to a loss of ESFAS signal to close containment isolation AOVs.

The next highest contributing SSCs are the seismically induced failure of all four Steam Generators and seismically induced failure of the containment at high seismic levels. The failure of either of these SSCs would lead to direct failure of containment (penetrations in the steam generator failure case) and cause a large early release.

Table 5.5-3 presents the percentage of the SLERF that derives from each interval in the seismic hazard curve. Also shown are: the hazard bin conditional large early release probability (CLERP), i.e., the probability that a large early release occurs given that a core damage event occurs; the percent of total SLERF; and the cumulative SLERF. As can be seen, 90% of the contribution comes from seismic events with PGA greater than 1.0g. The relatively low CLERP results even at high seismic magnitudes (e.g., %G12) is an indication of the robust containment capability.

Page 46 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.5-1 Dominant SLERF Cutsets

• CDF Input 1 Input 2 Input 3 Input 4 Input 5 Input 6 1

8.21%

2.68E-08 S_SEISMIC-SG SEQ_SEIS-LERF-DI R-1 2

5.71%

l.86E-08 S _ 1ACIV-120-CB180 S_SEISMIC-LLOCA SEQ_LERF-08 SEQ_LL-3 3

5.26%

l.71E-08 S_1DCBS-SGR-CB180 S_SEISMIC-LLOCA SEQ_LERF-08 SEQ_LL-3 4

5.26%

l.71E-08 S _ lACIV-120-CB 180 S_SEISMIC-MLOCA SEQ_LERF-08 SEQ_ML-5 5

4.82%

l.57E-08 S_1DCBS-SGR-CB180 S_SEISMIC-MLOCA SEQ_LERF-08 SEQ_ML-5 6

4.51%

l.47E-08 S_1DCBS-SGR-CB180 S_SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-9 7

3.34%

l.09E-08 S_CONTAINMENT SEQ_SEIS-LERF-DI R-1 S_1ACBS-120PN-8 2.97%

9.69E-09 CB180 S_SEISMIC-LLOCA SEQ_LERF-08 SEQ_LL-3 S_1ACBS-120PN-9 2.65%

8.65E-09 CB180 S_SEISMIC-MLOCA SEQ_LERF-08 SEQ_ML-5 10 1.96%

6.38E-09 S_1ACIV-120-CB180 S_lDGHE-LUBEOIL S SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-8 S_1ACBS-120PN-11 1.35%

4.42E-09 CB180 S_lDGHE-LUBEOIL S_SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-8 12 1.25%

4.06E-09 S_1ACIV-120-CB180 S_lDG S_SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-8 S_lFCMO-CCU-S_SEISMIC-SEQ_LERF-13 1.16%

3.80E-09 S _ 1ACIV-120-CB180 S lCCHE-4 6FANS SLOCA 08 SEQ_SL-2 S_lSWFN-NSCW-14 1.15%

3.75E-09 S_1ACIV-120-CB180 FANS S_SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-8 S _ 1ACIV-120-AB220-S_SEISMIC-SEQ_LERF-15 1.12%

3.65E-09 LC-B S_1ACIV-120-CB180 S_lDGDM-VENT-1-3 SLOCA 08 SEQ_SL-8 16 1.05%

3.42E-09 S_1ACIV-120-CB180 S_DG-BLDG S_SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-8 17 1.05%

3.42E-09 S_1ACIV-120-CB180 S_lDGPN-ENG S_SEISMIC-SLOCA SEQ_LERF-08 SEQ_SL-8

• Frequencies are point estimates that do not reflect quantification refinement using ACUBE; however they are valid for relative sequence evaluation; each cutset also includes a plant availability factor basic event (not shown) to reflect the fraction of time at power.

Descriptions of each basic event listed in the cutsets are provided in Table 5.5-la.

Page 47 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.5-la Basic Event Description Table Event Description S SEISMIC-SG Seismic Failure Of Steam Generators S 1ACIV-120-CB180 Seismically induced failure of AC Inverter CB180 S 1DCBS-SGR-CB180 Seismically induced failure of 125 VDC Switchgear CB180 S CONTAINMENT Seismically induced failure of Containment S 1ACBS-120PN-CB180 Seismically induced failure of 120 VAC Panel CB 180 S 1ACIV-120-AB220-LC-B Seismically induced failure of Vital AC Inverter 1BD1112 S SEISMIC-LLOCA Seismic Induced LLOCA S SEISMIC-MLOCA Seismic Induced MLOCA S SEISMIC-SLOCA Seismic Induced SLOCA S lDGHE-LUBEOIL Seismically induced failure of DG Lube Oil HX s lDG Seismically induced failure of Both Diesel Generators S lCCHE-4 Seismically induced failure of CCW Heat Exchanger S lSWFN-NSCW-FANS Seismically induced failure of NSCW Tower Fans S DG-BLDG Seismically induced failure of Diesel Buildings S lDGPN-ENG Seismically induced failure of DG Engine Control Panel S 1FCMO-CCU-6FANS Seismically induced failure of Containment Fan Cooler Units-6,3,4,1,5,8 S lDGDM-VENT-1-3 Seismically induced failure of DG Vent Damper For Fans 1&3 SEQ_SEIS-LERF-DIR-1 Direct Bypass Sequence Tag SEQ LERF-08 Large Early Release Sequence Tag SEQ_LL-3 Core Damage Sequence Tag SEQ ML-5 Core Damage Sequence Tag SEQ SL-9 Core Damage Sequence Tag SEQ_SL-8 Core Damage Sequence Tag Page 48 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Table 5.5-2 SLERF Importance Measures Ranked by FV Final Equipment Description FV Am J3r J3u Mode Method S SEISMIC-LLOCA SEISMIC INDUCED LLOCA 0.21 2.54 0.3 0.33 Anchorage SOV(l)

S_SEISMIC-MLOCA SEISMIC INDUCED MLOCA 0.21 2.54 0.3 0.33 Anchorage SOV(l)

S_SEISMIC-SLOCA SEISMIC INDUCED SLOCA 0.20 2.54 0.3 0.33 Anchorage SOV(l)

SEISMIC FAILURE OF 125 voe EPRI NP S 1DCBS-SGR-CB180 0.19 1.98 0.32 0.32 Functional SWITCHGEAR CB180 6041-SL SEISMIC FAILURE OF AC INVERTER EPRI NP S_1ACIV-120-CB180 0.19 1.98 0.32 0.32 Functional CB180 6041-SL S_SEISMIC-SG SEISMIC FAILURE OF SG 0.14 2.75 0.24 0.26 Anchorage CDFM SEISMIC FAILURE OF 120 VAC PANEL CB S 1ACBS-120PN-CB180 0.12 2.16 0.31 0.31 Functional sov 180 S CONTAINMENT SEISMIC FAILURE OF CONTAINMENT 0.08 2.90 0.23 0.25 Structural CDFM SEISMIC FAILURE OF BATIERY CHARGER Functional EPRINP S 1DCBC-CB180 0.02 1.98 0.32 0.32 CB180 6041-SL SEISMIC FAILURE OF 125 voe BATIERY S 1DCBV-CB180 0.02 2.26 0.25 0.35 Functional sov CB180 Notes:

1: Based on seismic capacity of RCP coolant pump supports. All other NSSS equipment that lead to a LOCA had a greater capacity.

Page 49 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table 5.5-3 Contribution to SLERF by Acceleration Interval Seismic IE Bin Frequency CLE RP SLERF Cumulative SLERF

%G01- {O.lg to <0.lSg) 8.20E-04 Note 1 Note 1 0%

Note 1

%G02- (O.lSg to <0.3g) 8.23E-04 Note 1 Note 1 0%

Note 1

%G03- (0.3g to <0.4g) 1.52E-04 3.lSE-08 4.79E-12

<1%

4.79E-12

%G04- (0.4g to <0.Sg) 6.71E-05 7.31E-07 4.90E-11

<1%

5.38E-11

%GOS- (O.Sg to <0.6g) 3.18E-05 9.00E-06 2.86E-10

<1%

3.40E-10

%G06- (0.6g to <0. 7g) 1.61E-05 7.71E-05 1.24E-09

<1%

1.58E-09

%G07- (0.7g to <0.8g) 8.64E-06 4.60E-04 3.98E-09 1%

5.56E-09

%G08- (0.8g to <0.9g) 4.79E-06 1.96E-03 9.40E-09 3%

1.SOE-08

%G09- (0.9g to <lg) 2.71E-06 6.45E-03 1.75E-08 5%

3.25E-08

%G10- (lg to <1.lg) 1.SSE-06 1.72E-02 2.67E-08 8%

5.92E-08

%G11- {1.lg to <1.2g) 9.00E-07 3.79E-02 3.41E-08 10%

9.33E-08

%G12- (1.2g to <1.Sg) l.02E-06 1.18E-01 1.20E-07 37%

2.13E-07

%G13- (1.Sg to <2g) 2.34E-07 4.27E-01 9.99E-08 31%

3.13E-07

%G14- (>2g) 1.60E-08 7.66E-01 1.23E-08 4%

3.25E-07 Note 1: Contribution is insignificant in this interval.

Page 50 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NlTF 2.1 Seismic PRA Submittal Version 0 - March 2017 5.6 SPRA Quantification Uncertainty Analysis Parameter uncertainty in seismic PRA results comes from seismic hazard curve uncertainty, the SSC fragility uncertainties, and uncertainties in the human interaction and random failure calculations. SPRA model parameter uncertainty was quantified using the EPRI UNCERT code. The results are provided in Table 5.6-1, and Figures 5.6-1 and 5.6-2, each of which shows the curves of cumulative probability and probability density function.

Table 5.6-1 Parameter Uncertainty Analysis Results Point Estimate Mean 5%

Median 95%

Standard Skewness Deviation Mean Seismic CDF Mean 2.79E-06 3.57E-06 3.34E-07 l.92E-06 l.18E-05 5.69E-06 7.12 Seismic LERF Mean 3.25E-07 4.29E-07 3.17E-08 2.13E-07 l.50E-06 6.90E-07 5.97 The UNCERT runs were performed using the Monte Carlo method of sampling and a total of 20,000 samples. Both SCDF and SLERF runs solved 20,000 cutsets using ACUBE. The distribution for both SCDF and SLERF appears generally uniform. The uncertainty is generally dominated by the hazard uncertainty. Since much of the seismic risk comes from higher seismic intervals (greater ground motion), the failure probabilities at this ground motion are generally very high and therefore will not contribute much in the way of uncertainty. The point estimate mean is calculated for each acceleration interval using mean values for the seismic hazard frequency, mean values for the seismic fragilities, and mean values for the random failures and human error probabilities. These acceleration interval point estimate means are then summed for the total SCDF and SLERF point estimate means. Comparison with the point estimate values indicates that the point estimates provide a reasonable approximation of the mean values.

Model uncertainty is introduced when assumptions are made in the SPRA model and inputs to represent plant response, when there may be alternative approaches to particular aspects of the modeling, or when there is no consensus approach for a particular issue. For the VEGP SPRA, the important model uncertainties are addressed through the sensitivity studies described in Section 5.7 to determine the potential impact on SCDF or SLERF.

Page 51of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seism ic PRA Submittal 08

/ -

1

/

I v

t O*

I i t

~

~

~ l 02 0.....

08 08 0 4 0 2 0.....

I

,_/

1E-07

,/

_/.....

~

I

\\ I\\

I

.\\

\\

+ Mean

• Median

~

Figure 5.6-1 SCDF - 20,000 Sample Monte Carlo OB 06 04 02 0

1E*10 OB 06 04 02 0

lE-09 lE-08 I

I v J

/

lE-07 v-

/

lE-06

+ Mean

• Median l E-05 1E*10 lE-09 lE-08 lE-07 lE-06 lE-05 Figure 5.6-2 SLERF - 20,000 Sample Monte Carlo Version 0 - March 2017 Page 52 of 124

'VEGP Units 1 and 2 10 CFR 50.54{f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Completeness uncertainty relates to potential risk contributors that are not in the model.

The scope of the VEGP SPRA is for at-power operation, and does not include risk contributors from low power-shutdown operation, or for spent fuel pool risk. In addition, there may be potential issues related to factors that are not included, such as the impact of aging on equipment reliability and fragility. Note that any significant degradation identified during the plant walkdown was included in the fragility calculations. Other potential issues include impacts of plant organizational performance on risk, and unknown omitted phenomena and failure mechanisms. By their very nature, the impacts on risk of these types of uncertainties are not known.

5.7 SPRA Quantification Sensitivity Analysis Several sensitivity studies were performed to examine different input information and assumptions on the VEGP SPRA results. Among the sensitivity studies examined were the following:

Model Truncation and Convergence Small-Small LOCA Control Rod Insertion Offsite Power Impact with Improved Plant Wilson fragility Human Reliability Increased Seismic Capacity of all Four 125 VDC 1E Distribution Panels Preventing ESFAS Signal Failure on Loss of Vital AC Inverters Auxiliary Building Failure Table 5.7-1 provides a summary of the sensitivity studies. It also describes the sensitivity and lists any change in median seismic capacity (if applicable). The results for the base case and the sensitivity are shown for both SCDF and SLERF. A percent difference is shown to illustrate more important cases.

5.7.1 Model Truncation and Convergence The baseline SPRA was quantified at lE-11 for seismic initiators %G01-%G11 and 5E-10 for %G12-%G14. The reason for the two different truncation levels is that the higher bins have a CCDP of approximately 1.0 and the quantification time drastically increases below 5E-10 for no added benefit. The SLERF model is similar where %G01-%G11 was quantified with a truncation of lE-12 and %G12-%G14 was set at 5E-10. Model convergence per the criteria in the PRA Standard was achieved at these levels.

5.7.2 Small-Small LOCA As discussed in the PRA Standard [4] and the EPRI SPRA Implementation Guide [10], the SPRA must consider the potential occurrence of a small-small LOCA. For VEGP the seismic capacity walkdowns evaluated small piping and tubing inside the containment, with the conclusion that the fragility for a small-small LOCA was about the same as for a SLOCA, which is included in the SPRA. A sensitivity study with conservative assumptions was performed by reducing the SLOCA median capacity to 1.02g, which corresponds to Page 53 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 assuming that the Small-Small LOCA high confidence of low probability of failure (HCLPF) is about two times the design basis earthquake (DBE) of 0.2g. The impact on CDF was an increase of more than 105% and LERF increased more than 229%.

This sensitivity study confirms.the model's reliance on the Small-Small LOCA modeling assumptions. However, there are two factors that over-emphasize this apparent sensitivity. First, the HCLPF value used in this sensitivity is much lower than anticipated based on the SPRA walkdowns. Second, the model includes a conservative assumption that failure of any of the Class lE Inverters will lead to the failure of ES FAS to send a safety injection (SI) signal, thus failing to auto-start AFW in CDF sequences. In LERF sequences, failure of the inverters will also lead to the failure to send a Cl signal and several AOVs will remain open leading to a large early release. During a small-small SLOCA, the time to available to restart AFW is more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and credit for an operator action would substantially reduce the small small LOCA contribution to SCDF and SLERF.

These conservative assumptions directly affect all SLOCA sequences where ESFAS is relied upon for response to prevent core damage and subsequently a large early release. Since the VEGP SPRA SCDF and SLERF results are already relatively low, no further model refinement was deemed necessary.

5.7.3 Control Rod Insertion The model assumes that given a seismic LOSP with failure of the CRDM or RV internals, the control rods would not insert due to mechanical interference, and an ATWT would occur. (The internal events PRA does not question ATWT for LOSP sequences, since the rods would drop on loss of power to the gripper coils and mechanical binding is unlikely).

In a LOSP (0.3g), the control rods would be released immediately, so they may insert before the failure of the higher capacity CRDM (2.2g) or RV internals (>4g). A sensitivity study provides an assessment by assuming that only 10% of the seismic failures would result in failure of the control rods to insert. SLERF was not affected, but SCDF was reduced by about 6%. Given the relatively low VEGP SPRA SCDF, no further refinement to the model was made.

5.7.4 Offsite Power Impact with Improved Plant Wilson Fragility Allen B. Wilson Combustion Turbine Plant (Plant Wilson) is owned by Southern Company and is located in Waynesboro, Georgia. Plant Wilson provides a backup source of offsite power to VEGP. Plant Wilson has several black-start combustion turbine generators, and a dedicated underground line to a small transformer in the Vogtle Units 1 through 4 switchyard. During the walkdown of Plant Wilson, several seismic vulnerabilities were identified, such as unanchored starting batteries. If these vulnerabilities were modified, the seismic capacity of offsite power could be improved. For study purposes, the seismic LOSP capacity in the VEGP SPRA was increased to 0.5g median. The results showed that LERF was not significantly impacted, but SCDF was reduced by about 2%. This adjusted median capacity would mostly benefit the lower seismic hazard scenarios, and the overall Page 54 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 result would not change much since these scenarios do not contribute much to SCDF.

Based on this result, no plant changes are warranted.

5. 7.5 Human Reliability To examine the uncertainty inherent in the calculation of operator action human error probabilities (HEPs), three sensitivity studies were performed. For the first case, all of the internal event PRA HEPs were increased by a factor of 3, with a cap of 1.0 (failure). SCDF increased by 9% and SLERF increased by 1%. The risk insights from the sensitivity, however, did not change. The action modeling operators failing to borate and start feed and bleed became slightly more important, consistent with the increase in failure probability of those events.

For the second case, all of the seismic-specific operator action HEPs were increased by a factor of 3, with a cap of 1.0 (failure). Neither SCDF nor SLERF was significantly impacted by this increase. This is expected because the major contribution to SCDF and SLERF come from high PGA scenarios where the HEPs are already set to 1.0 (guaranteed failure).

The third case assumed no credit for operator action in all bins above 0.8g. This again did not impact SCDF or SLERF because most of these actions were already set to (or close to) 1.0 in the base model quantification.

The results indicate that the model is not overly sensitive to operator response credit.

5.7.6 Increased Seismic Capacity of all Four 125 VDC lE Distribution Panels A sensitivity study was performed to determine the potential reduction in SCDF from the top contributing SSC (correlated failure of the four 125 VDC lE Distribution Panels). In this case the panel Am value was set to 5g (i.e., assume very high seismic capacity could be achieved). A reduction of about 22% was realized in SCDF; however, the SLERF reduction was only about 1%. The SCDF reduction is related to the conservative modeling of failure of ESFAS due to failure of the panels. The smaller impact to SLERF is because there are also other failures that can fail the containment isolation signal to close AOVs preventing a release. The next section provides a related sensitivity case.

Given the relatively low VEGP SPRA SCDF, no plant changes are warranted.

5.7.7 Preventing ESFAS Signal Failure on Loss of Vital AC Inverters A potentially conservative portion of the SLERF model fails the ES FAS signal on the failure of Vital AC inverters. This sensitivity evaluates the impact of the conservatism. The dominant sequences in which this scenario arises is following a seismically-induced LOCA (any size) where AFW also fails to start and the containment isolation AOVs (HV-0780, HV-0781) do not close because DC power and instrument air for the valve operators is still available for some time following the event. The model requires the ESFAS signal to close the valves and no operator action is credited for manually closing the valves following a LOCA. Another potential recovery is restoring power to the panels from 480V AC to allow the ESFAS signal to be sent.

Page SS of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Table 5.7-1 Summary of Sensitivity Study Results Baseline Values 2.SE-06 SCDF 1 3.3E-07 SLERF 1 Sensitivity Item Description Small-Small LOCA Impact Reduce Seismic Capacity of SLOCA Decrease the frequency of failure of CRD Insertion rods to 10% of challenges Internal Events PRA HEPs Increase Multiply all FPIE HEPs by 3 SPRA HEPs Increase Multiply all seismic specific HEPs by 3 HEPs above 0.8g Increase all HEPs above 0.8g to 1.0 125voe1E 1DCBS-PN-CB180-1E median fragility Distribution Panels increased to 5.0g

• Added Plant Wilson Credit to the Plant Wilson Credit Model Aux Building Leads Direct to Core Aux Building Failure Damage Notes: 1. ACUBE reported value NC= No change Am Am Original Sensitivity 2.54 1.02 2.07 5

0.3 0.3 4.6 2.8 Version 0 - March 2017 SCDF 1 Change SLERF 1 Change 5.7E-06 105%

1.lE-06 229%

2.6E-06

-6%

3.2E-07

<-1%

3.0E-06 9%

3.3E-07 1%

2.8E-06 NC 3.3E-07 NC 2.8E-06 NC 3.3E-07 NC 2.2E-06

-22%

3.2E-07

-1%

2.8E-06

-1%

3.2E-07

-1%

2.9E-06 4%

3.3E-07 NC Page 56 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 The sensitivity was performed by preventing the failure of the AOVs to close for seismic events. This reduced LERF by a factor of about 3, to lxl0-7/yr.

5.7.8 Auxiliary Building Failure The Auxiliary Building (AB) has walls that could fail seismically and damage equipment. Since a detailed study was not performed to determine the specific impacts of the wall failures, this sensitivity study replaced a direct-to-core-damage event with a surrogate (non-SCDF end state) for the failure of the AB walls. The median capacity was set at 2.8g. The results indicate that SCDF would be increased by about 4% and the impact to SLERF was insignificant. This demonstrates that the current modeling technique is adequate for the treatment of the AB failures.

5.8 SPRA Logic Model and Quantification Technical Adequacy The VEGP SPRA risk quantification and results interpretation methodology were subjected to an independent peer review against the pertinent requirements in the ASME/ANS PRA Standard [4].

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the VEGP SPRA seismic plant response analysis is suitable for this SPRA application.

Page 57of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 6.0 Conclusions A seismic PRA has been performed for Vogtle Electric Generating Plant Units 1 and 2 in accordance with the guidance in the SPID [2]. The SPRA shows that the point estimate seismic CDF is 2.8x10-6/yr and the seismic LERF is 3.3x10-7 /yr. Uncertainty, importance, and sensitivity analyses were performed. Sensitivity studies were performed to identify critical assumptions, test the sensitivity to quantification parameters and the seismic hazard, and identify potential areas to consider for the reduction of seismic risk. These sensitivity studies demonstrated that the model results were robust to the modeling and assumptions used.

The SPRA as described in this submittal reflects the as-built/as-operated Vogtle Electric Generating Plant Units 1and2 as of the SPRA freeze date, August 31, 2015. An assessment is included in Appendix A of the impact on the results of plant changes not included in the model. No seismic hazard vulnerabilities were identified, and no plant actions have been taken or are planned given the insights from this study.

Page 58 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 7.0 References

1)

NRC (E Leeds and M Johnson) Letter to All Power Reactor Licensees et al., "Request for Information Pursuant to Title 10 of the Code of Federal Regulations 50.54(f)

Regarding Recommendations 2.1, 2.3 and 9.3 of the Near-Term Task Force Review of Insights from the Fukushima Dai-lchi Accident," U.S. Nuclear Regulatory Commission, March 12, 2012.

2)

EPRI 1025287, Seismic Evaluation Guidance: Screening, Prioritization and Implementation Details {SPID} for the Resolution of Fukushima Near-Term Task Force Recommendation 2.1: Seismic. Electric Power Research Institute, Palo Alto, CA: February 2013.

3)

Southern Nuclear Operating Company, NL-14-0344, Vogtle Electric Generating Plant - Units 1 and 2, Seismic Hazard and Screening Report for CEUS Sites, March 31, 2014. NRC Adams ML14092A019.

4)

ASME/ANS RA-S-2008, Standard for Level 1/Large Early Release Frequency Probabilistic Risk Assessment for Nuclear Power Plant Applications, including Addenda B, 2013, American Society of Mechanical Engineers, New York, September 30, 2013.

5)

NEl-12-13, External Hazards PRA Peer Review Process Guidelines, Revision 0, Nuclear Energy Institute, Washington, DC, August 2012.

6)

PWROG-15004-P, "Peer Review of the Vogtle Units 1 & 2 Seismic Probabilistic Risk Assessment," Westinghouse Electric Company, February 2015.

7)

EPRI NP 6041-SL, A Methodology for Assessment of Nuclear Power Plant Seismic Margin, Rev. 1., Electric Power Research Institute, Palo Alto, CA, August 1991.

8)

SNC Calculation V-RIE-SEIS-U00-002-001, Vogtle Units 1 and 2, Seismic Equipment List and SEL Walkdown Report - Seismic PRA.

9)

EPRI 3002008093, An Approach to Human Reliability Analysis for External Events with a Focus on Seismic, Electric Power Research Institute, Palo Alto, CA, December 2016.

10)

EPRI 30020007091, Seismic PRA Implementation Guide, Electric Power Research Institute, Palo Alto, CA, December 2013.

11)

Regulatory Guide 1.200, Revision 2, "An Approach For Determining The Technical Adequacy Of Probabilistic Risk Assessment Results For Risk-Informed Activities,"

U.S. Nuclear Regulatory Commission, March 2009.

12)

EPRI 1009684, CEUS Ground Motion Project Final Report, Electric Power Research Institute, Palo Alto, CA, December 2004.

13)

NUREG-1855, "Guidance on the Treatment of Uncertainties Associated with PRAs in Risk-Informed Decision Making," Rev. 0, U.S. Nuclear Regulatory Commission, March 2009.

Page 59 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017

14)

EPRI 1016737, Treatment of Parameter and Modeling Uncertainty for Probabilistic Risk Assessments, Electric Power Research Institute, Palo Alto, CA, December 2008.

15)

NUREG 1407, "Procedural and Submittal Guidance for the Individual Plant Examination of External Events for Severe Accident Vulnerabilities," U.S. Nuclear Regulatory Commission, June 1991.

16)

SNC Calculation V-RIE-SEIS-U00-006-001, Fragility Notebook-SPRA Vogtle Units 1 and 2, Ver. 2.

17)

Lettis Consultants International, Inc., "Seismic Hazard and Sensitivity Results for Vogtle Nuclear Plant, Vogtle Peer Review Response," VAPOOl-AQPR-01-Rev.1, prepared for Southern Nuclear, February 2017.

18)

SNC Calculation X2CFS129, Ver 3, "Site Response Analysis for the Vogtle Site."

19)

EPRI 1019200, Seismic Fragility Applications Guide Update, Electric Power Research Institute, Palo Alto, CA, December 2009.

20)

EPRI 1002988, Seismic Fragility Application Guide, Electric Power Research Institute, Palo Alto, CA, December 2002.

21)

EPRI TR-103959, Methodology for Developing Seismic Fragilities, Electric Power Research Institute, Palo Alto, CA, July 1994.

22)

ASCE 4-98, "Seismic Analysis of Safety-Related Nuclear Structures and Commentary," American Society of Civil Engineers, Reston, VA.

23)

ASCE 43-05, "Seismic Design Criteria for Structures, Systems, and Components in Nuclear Facilities," American Society of Civil Engineers, Reston, VA.

24)

ACI 349, "Code Requirements for Nuclear Safety-Related Concrete Structures and Commentary," American Concrete Institute.

25)

EPRI NP-5223-SL Rl, Generic Seismic Ruggedness of Power Plant Equipment (Revision 1}, Electric Power Research Institute, Palo Alto, CA, August 1991.

26)

NUREG-0800, "Standard Review Plan", Section 3.7.2, Seismic System Analysis, Revision 3, U.S. Nuclear Regulatory Commission, March 2007.

27)

ASME/ANS RA-S-2002, Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications, including Addenda B, 2005, American Society of Mechanical Engineers, New York, December 2005.

28)

Regulatory Guide 1.200, Revision 1, "An Approach For Determining The Technical Adequacy Of Probabilistic Risk Assessment Results For Risk-Informed Activities,"

U.S. Nuclear Regulatory Commission, January 2007.

29)

"Vogtle Electric Generating Plant, Units 1 and 2 - Staff Assessment of Information Provided Pursuant to Title 10 of the Code of Federal Regulations Part 50, Section 50.54{F), Seismic Hazard Reevaluations Relating to Recommendation 2.1 of the Page 60 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Near-Term Task Force Review of Insights From the Fukushima Dal-lchi Accident (TAC NOS. MF3770 AND MF3771)", NRC Adams ML15054A296.

30)

Generic Implementation Procedure for Seismic Verification of Nuclear Plant Equipment", Rev. 2, Seismic Qualification Utility Group, (SQUG), September 1990.

31)

EPRI 3002004396, High Frequency Program Application Guidance for Functional Confirmation and Fragility Evaluation, Final Report, Electric Power Research Institute, Palo Alto, CA, July 2015.

32)

EPRI NP-7147-SL, Seismic Ruggedness of Relays, Electric Power Research Institute, Palo Alto, CA, August 1991.

33)

EPRI TR-105988-V2, GERS Formulated Using Data from the SQURTS Program, Electric Power Research Institute, Palo Alto, CA, April 1999.

34)

McGuire, R.K., W. J. Silva, and C. J. Costantino. "Technical Basis for Revision of Regulatory Guidance on Design Ground Motions, Hazard-and Risk-Consistent Ground Motion Spectra Guidelines",

prepared for Nuclear Regulatory Commission, NUREG/CR-6728, 2001.

35)

EPRI, USDOE, USNRC, 2012, "Central and Eastern United States Seismic Source Characterization for Nuclear Facilities," U.S. Nuclear Regulatory Commission Report NUREG-2115.

36)

Southern Nuclear Operating Company, Vogtle Early Site Permit Application Part 2, Site Safety Analysis Report (SSAR) Revision 5 December 2008.

37)

Nuclear Regulatory Commission, "NUREG-1923 - Safety Evaluation Report for an Early Site Permit (ESP) at the Vogtle Electric Generating Plant (VEGP) ESP Site,"

July 2009, ML092290630 and ML092290650.

38)

Southern Nuclear Operating Company, Vogtle Electric Generating Plant Units 3 and 4, COL Application Part 2, Final Safety Analysis Report, Revision 5, Section 19.55.6.3, Site Specific Seismic Margin Analysis, March 2011.

39)

SNC Calculation X2CFS129, Ver 1, Site Response Analysis for the Vogtle Site.

40)

EPRI NP-7147-SL, V2, Addendum 1, Seismic Ruggedness of Relays, Electric Power Research Institute, Palo Alto, CA, Sep 1993.

41)

EPRI NP-7147-SL, V2, Addendum 2, Seismic Ruggedness of Relays, Electric Power Research Institute, Palo Alto, CA, April 1995.

42)

EPRI 3002002997, High Frequency Program: High Frequency Testing Summary, Final Report, Electric Power Research Institute, Palo Alto, CA, Sep 2014.

Page 61 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 8.0 Acronyms AB Auxiliary Building ACCW Auxiliary Component Cooling Water ACI American Concrete Institute ACU Auxiliary Cooling Unit AFW Auxiliary Feedwater System Am Median Seismic Capacity ANS American Nuclear Society AOV Air Operated Valve ARS Acceleration Response Spectra ASCE American Society of Civil Engineers ASME American Society of Mechanical Engineers ATWT Anticipated Transient Without Trip BBM Blue Bluff Marl CCDP Conditional Core Damage Probability CCU Containment Cooling Unit CDFM Conservative Deterministic Failure Model CEUS Central and Eastern United States Cl Containment Isolation CROM Control Rod Drive Mechanism DBE Design Basis Earthquake EOG Emergency Diesel Generator EPRI Electric Power Research Institute ESEP Expedited Seismic Evaluation Program ESF Engineered Safety Features ES FAS Engineered Safety Features Actuation System ESM Extended Subtraction Method ESP Early Site Permit ESW Essential Service Water FEM Finite Element Model FIRS Foundation Input Response Spectra FT Fault Tree FV Fussell-Vesely (risk importance measure)

GERS Generic Equipment Ruggedness Spectra GMPE Ground Motion Prediction Equation GMRS Ground Motion Response Spectra HCLPF High Confidence of Low Probability of Failure Page 62 of 124

VEGP Units 1 and 2 HEP HF HRA HVAC IE IPEEE ISRS LF LLOCA LMSM LOCA LOSP MDAFWP MFFF MLOCA MSL MSM NEI NRC NSCW NSSS NTIF PGA PSHA RCP RCS RHR RLME RPS RV SA SBO SCDF SEL SFP SFR SG SHA 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Human Error Probability High Frequency Human Reliability Analysis Heating Ventilation and Air Conditioning Initiating Event Individual Plant Examination for External Events In-Structure Response Spectra Low Frequency Large Loss of Coolant Accident Lumped Mass Stick Model Loss of Coolant Accident Loss of Offsite Power Motor Drive Auxiliary Feedwater Pump MOX Fuel Fabrication Facility Medium Loss of Coolant Accident Mean Sea Level Modified Subtraction Method Nuclear Energy Institute Nuclear Regulatory Commission Nuclear Service Cooling Water Nuclear Steam Supply System Near Term Task Force Peak Ground Acceleration Probabilistic Seismic Hazard Analysis Reactor Coolant Pump Reactor Coolant System Residual Heat Removal Repeated Large Magnitude Earthquake Reactor Protection System Reactor Vessel Spectral Acceleration Station Blackout Seismic Core Damage Frequency Seismic Equipment List Spent Fuel Pool Seismic Fragility Element Within ASME/ANS PRA Standard Steam Generator Seismic Hazard Analysis Element Within ASME/ANS PRA Standard Page 63 of 124

VEGP Units 1 and 2 SHS SI 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Seismic Hazard Submittal Safety Injection Seismic Large Early Release Frequency Small Loss of Coolant Accident Seismic Margin Assessment Separation of Variables Version 0 - March 2017 SLERF SLOCA SMA sov SPID Screening, Prioritization and Implementation Details SPR SPRA SQUG SRP SRSS SRT Seismic PRA Modeling Element Within ASME/ANS PRA Standard Seismic Probabilistic Risk Assessment SSC SSEL SSI TDAFWP TSCR UHRS UHS USI Seismic Qualification Utility Group Standard Review Plan Square Root of Sum of Squares Seismic Review Team Structure, System or Component Safe Shutdown Equipment List Soil Structure Interaction Turbine Driven Auxiliary Feedwater Pump Truncated Soil Column Response Uniform Hazard Response Spectra Ultimate Heat Sink Unresolved Safety Issue VEGP Vogtle Electric Generating Plant VEGP 1 and 2 Vogtle Electric Generating Plant Units 1 and 2 VEGP 3 and 4 Vogtle Electric Generating Plant Units 3 and 4 WUS Western United States Page 64 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Appendix A Summary of SPRA Peer Review and Assessment of PRA Technical Adequacy for Response to NTTF 2.1 Seismic S0.54(f) Letter This Appendix has two purposes:

1.

Provide a summary of the SPRA peer review

2.

Provide the bases for the technical adequacy of the SPRA for the 50.54(f) response.

The VEGP SPRA was subjected to an independent peer review against the pertinent requirements in Part 5 of Addendum B of the AS ME/ANS PRA Standard [4].

The information presented here establishes that the SPRA has been peer reviewed by a team with adequate credentials to perform the assessment, establishes that the peer review process followed meets the intent of the peer review characteristics and attributes in Table 16 of RGl.200 R2 [11] and the requirements in Section 1-6 of the ASME/ANS PRA Standard [4], and presents the significant results of the peer review.

A.1.

Overview of Peer Review The peer review assessment [6], and subsequent disposition of peer review findings, is summarized here. The scope of the review encompassed the set of technical elements and supporting requirements (SR) for the SHA (seismic hazard), SFR (seismic fragilities),

and SPR (seismic PRA modeling) elements for seismic CDF and LERF. The peer review therefore addressed the full set of SRs identified in Tables 6-4 through 6-6 of the SPID [2].

The VEGP SPRA peer review was conducted during the week of November 17, 2014. As part of the peer review, a walk-down of portions of VEGP Units 1 & 2 was performed on November 17, 2014 by members of the peer review team who have the appropriate SQUG training.

A.2.

Summary of the Peer Review Process The peer review was performed against the requirements in Part 5 (Seismic) of Addenda B of the PRA Standard [4], using the peer review process defined in NEI 12-13 [5]. The review was conducted over a four-day period, with a summary and exit meeting on the morning of the fifth day.

The NEI 12-13 SPRA peer review process [5] involves an examination by each reviewer of their assigned PRA technical elements against the requirements in the Standard to ensure the robustness of the model relative to all of the requirements.

Implementing the review involves a combination of a broad scope examination of the PRA elements within the scope of the review and a deeper examination of portions of the PRA elements based on what is found during the initial review. The supporting requirements Page 65 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 (SRs) provide a structure which, in combination with the peer reviewers' PRA experience, provides the basis for examining the various PRA technical elements. If a reviewer identifies a question or discrepancy, that leads to additional investigation until the issue is resolved or a Fact and Observation (F&O) is written describing the issue and its potential impacts, and suggesting possible resolution.

For each technical element, i.e., SHA, SFR, SPR, a team of two peer reviewers were assigned, one having lead responsibility for that area.

For each SR reviewed, the responsible reviewers reached consensus regarding which of the Capability Categories defined in the Standard that the PRA meets for that SR, and the assignment of the Capability Category for each SR was ultimately based on the consensus of the full review team. The Standard also specifies high level requirements (HLR). Consistent with the guidance in the Standard, capability Categories were not assigned to the HLRs, but a qualitative assessment of the applicable HLRs in the context of the PRA technical element summary was made based on the associated SR Capability Categories.

As part of the review team's assessment of capability categories, F&Os are prepared.

There are three types of F&Os defined in NEI 12-13 [5]: Findings, which identify issues that must be addressed in order for an SR (or multiple SRs) to meet Capability Category II; Suggestions, which identify issues that the reviewers have noted as potentially important but not requiring resolution to meet the SRs; and Best Practices, which reflect the reviewers' opinion that a particular aspect of the review exceeds normal industry practice. The focus in this Appendix is on Findings and their disposition relative to this submittal.

A.3.

Peer Review Team Qualifications The members of the peer review team were Mr. Kenneth Kiper and Dr. Andrea Maioli of Westinghouse, Dr. Martin Mccann of Jack R. Benjamin & Associates, Dr. Richard Quittmeyer of Rizzo Associates, Mr. Steve Eder of Facility Risk Consultants, Mr. William Horstman of Pacific Gas & Electric Company, and Mr. Aaron Quaderer of FirstEnergy Nuclear Operating Company. The peer review team members met the peer review independence criteria in NEI 12-13 [S] and had no involvement in the development of the Vogtle Units 1 &2 SPRA.

Mr. Kiper, the team lead, is a Technical Manager at Westinghouse after a 31-year career at Seabrook Station. He has experience in virtually every aspect of PRA modeling and applications, including upgrading and maintaining the Seabrook seismic PRA.

Dr. Martin Mccann was the lead for the Seismic Hazard Analysis (SHA) technical element.

He has 30-years' experience in engineering seismology including site response analysis Page 66 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 and specification of ground motion. He was assisted in the hazard review by Dr. Richard Quittmeyer, an internationally-recognized expert in seismicity, seismic hazard, and site characterization.

Mr. Steve Eder was the lead for the Seismic Fragility Analysis (SFR) technical element. Mr.

Eder has more than 30-years' experience in the fields of natural hazards risk assessment, seismic fragility analysis, structural performance evaluation, and retrofit design. He was assisted by William Horstman, a senior consulting civil engineer for Diablo Canyon. Mr.

Horstman has more than 30-years' experience in the fields of structural engineering and structural mechanics.

Dr. Andrea Maioli was the lead for the Seismic Plant-Response Analysis (SPR) technical element. Dr. Maioli has over 10-years' experience in the nuclear safety area generally and seismic PRA specifically. Mr. Aaron Quaderer assisted in the review of the Seismic Plant Response technical elements. He has over ten years of experience in plant and design engineering, including responsibility for maintenance and application of the Davis Besse PRA model.

A.4.

Summary of the Peer Review Conclusions The review team's assessment of the SPRA elements is excerpted from the peer review report [6] as follows. Where the review team identified issues, these are captured in peer review findings, for which the dispositions are summarized in the next section of this appendix.

SHA As required by the Standard, the frequency of occurrence of earthquake ground motions at the site was based on a site-specific probabilistic seismic hazard analysis (PSHA). The Senior Seismic Hazard Analysis Committee (SSHAC) process of conducting a PSHA was used to develop the regional seismic source characterization (SSC) model and the ground motion model (GMM) inputs to the analysis. The SSC inputs to the PSHA are based on the recently completed Central and Eastern U.S. (CEUS) seismic source model. The ground motion model inputs to the PSHA are based on the CEUS ground motion update project. The requirements of the SSHAC process satisfy the requirements of the standard. The SSHAC process defines a method for utilizing structured expert elicitation and minimum technical requirements to complete a PSHA.

The "SSHAC level" of a seismic hazard study ensures that data, methods and models supporting the PSHA are fully incorporated and that uncertainties are fully considered in the process at sufficient depth and detail necessary to satisfy scientific and regulatory needs. The level of study is not mandated in the standard; however, both the SSC and the GMM parts of the PSHA were developed as a result of SSHAC level 3 Page 67 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 analyses. In the case of the GMM, a SSHAC level 2 analysis was carried out to update a prior level 3 study. These level 3 studies satisfy the requirements of the standard.

As a first step to performing a PSHA, the Standard requires an up-to-date database, including regional geological, seismological, geophysical data, and local site topography, and a compilation of surficial geologic and geotechnical site properties.

These data include a catalog of relevant historical, instrumental, and paleo-seismic information within 320-km of the site. The CEUS SSC study involved an extensive data collection effort that satisfies the requirements of the standard as it relates to developing a regional-scale seismic source model.

e No effort was made to compile new {relative to the data used in the CEUS SSC study) or local {relative to the regional scale) information beyond what was considered in the development of the CEUS SSC regional scale seismic sources. This includes:

o Updating the earthquake catalog and evaluating the potential impact on the estimate of seismicity parameters, and o

Collecting and evaluating geologic, seismologic and geophysical information to assess whether new information or information at a local scale exists that would indicate that new, local seismic sources or modifications to the CEUS regional scale seismic sources are required.

o In the implementation of the CEUS SSC model for the Vogtle site, all distributed seismic sources in the CEUS SSC within 640 km and all RLME sources within 1000 km were included in the PSHA calculations. By including these seismic sources in the analysis, the contribution of "near-field" and "far-field" earthquake sources to ground motions at Vogtle were considered.

o The seismic hazard analysis for the Vogtle site also took into account the effects of local site response. However, the review team determined that the site response analysis did not fully evaluate and model aleatory and epistemic uncertainties in the site response analysis.

The standard requires that spectral shapes be based on a site-specific evaluation, taking into account the contributions of deaggregated magnitude-distance results of the probabilistic seismic hazard analysis. The PSHA fully accounted for the "near" and "far" source spectral shapes.

The standard requires that sensitivity calculations be performed to document the models and parameters that are the primary contributors to the site hazard. In the Vogtle PSHA, only a limited number of sensitivity calculations and deaggregation results were presented. As a result, this requirement was not met.

The standard requires that a screening analysis be performed to determine whether hazards other than earthquake ground motion pose a hazard to the site. A screening analysis was performed for some but not all of the other seismic hazards. Therefore, this assessment was incomplete.

Both the aleatory and epistemic uncertainties have been addressed in characterizing the seismic sources. In addition, uncertainties in each step of the hazard analysis were propagated and displayed in the final quantification of hazard estimates for the Vogtle site.

Page 68 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 The standard requires that documentation of the PSHA be provided that supports the PRA applications, peer review and upgrades. This requirement establishes a high standard for documentation of the PSHA that allows for examination of the PSHA methodology, its implementation, and the PSHA results to assess that the approach is appropriate, the analysis was performed correctly, and the results are reasonable..

The Vogtle PSHA documentation is minimal and therefore this requirement is not met.

The findings with regard to this requirement point out that the way the CEUS SSC and GMM models were implemented in the PSHA was not described; the most recent documentation is incomplete in the sense that it does not include all of the analyses (i.e., results of sensitivity calculations) performed in earlier versions.

SFR The Standard requires that all the structures, systems and components (SSCs) that play a role in the seismic PRA be identified as candidates for subsequent seismic fragility evaluation. A seismic equipment list was developed. Generic seismically insensitive items as well as seismically rugged items identified by the walkdown team were screened out. A screening level of 2.Sg median capacity was then established for fragility evaluation based on insignificant contribution to risk.

Generic data and conservative simplifying assumptions were utilized to establish

. preliminary seismic fragilities for all SSCs, with the goal of demonstrating capacity above this satisfy level. The separation-of-variables methodology was employed. The fragility evaluations were refined on an as-needed basis as requested by the quantification team. Refinements were incorporated until favorable quantification results were achieved. As a result of this process, the seismic fragilities as a whole are not realistic as required by the standard.

Excess conservatism was noted in essentially all fragility calculations. This approach was possible for Vogtle due to the high seismic capacity of the SSCs.

The Standard requires that the seismic-fragility evaluation be based on realistic seismic response that the SSCs experience at their failure levels. The building response spectra were developed using new 3-D dynamic building models and soil-structure-interaction analyses, and used in the evaluation of seismic fragilities. A deterministic method was employed to establish median-centered response corresponding to structural model properties associated with the 1E-4 uniform hazard response spectrum shaking level. As a result, seismic response is overestimated for higher levels of shaking. Higher shaking levels cause additional cracking, leading to softening of members and increased structural damping. Strain dependent soil properties also change.

A structural response modification factor was not used in the fragility evaluations to adjust for this conservatism.

A series of walkdowns, focusing on the anchorage, lateral seismic support, functional characteristics, and potential systems interactions were conducted and documented appropriately in support of the fragility analysis. The walkdowns also identified the potential for seismic-induced fires and floods. The walkdown observations were subsequently incorporated in the seismic fragility evaluations. Some improvements in documentation were recommended.

Page 69 of 124 J

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 The SPRA identifies the relevant failure modes for the SSCs through a review of plant design documents, earthquake experience data, and walkdowns. Seismic-fragility evaluations were performed for these critical failure modes. The review team found that in general this requirement was satisfied, but noted, that the failure modes associated with sl.oshing of water and soil settlement were overlooked.

The Standard requires that the seismic-fragility parameters be based on plant-specific data supplemented as needed by earthquake experience data, fragility test data, and generic qualification test data.

The review team found that this requirement was satisfied. Use of generic data was justified via the iterative process of refining fragility on an as-needed basis.

SPR The seismic PRA model was developed by modifying the Full Power Internal Events (FPIE) PRA model to incorporate specific aspects of seismic analysis that are different from the FPIE. The logic model appropriately includes seismic-caused initiating events and other failures including seismic-induced SSC failures, non-seismic-induced unreliability and unavailability failure modes (based on the FPIE model), and human errors.

Additional documentation was needed regarding the details of the modification performed on the internal events PRA model to generate the seismic model. For example, some sequences have been added (apparently correctly) but without explaining the process. The reviewers felt that additional documentation would be needed to ensure future re-generation and update of the model.

A systematic identification of all the potential seismic induced initiators is presented and, while a hierarchy tree was not generated, the documentation supports the conclusion that the complete spectrum of seismic-induced scenarios was modeled.

The "Surry method" was applied to the Vogtle SPRA for the modification of existing human actions. Some actions were added to the logic specifically for the seismic logic (e.g., seismic-induced flooding specific actions). The reviewers questioned whether the treatment of seismic-specific performance shaping factors {PSF) for these actions was done in a manner consistent with the multiplier method that was selected. They felt there was not a systematic evaluation of the applicability and impact of the selected multiplier method on the overall results (e.g., sensitivities associated with location of the breaking points and/or the multiplier). In addition, the use of a multiplier method with only one breaking point at 0.8g is being now replaced in the industry by multipliers with more breaking points. While not stating that the Vogtle SPRA should adopt a new method, the reviewers felt there is a need for a more systematic assessment of the sensitivity of the model to the selected multiplier method.

The Vogtle SPRA team relied on the internal events PRA for the evaluation of HEP dependency. The reviewers felt that the dependency evaluation should be revised for seismic specific new potential dependencies.

The Vogtle SPRA adopts a standard full correlation of seismic failures, with a number of notable exceptions, that generate potentially non-minimal cutsets. While it is Page 70 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 apparent that the Vogtle SPRA team considered this effect, more documentation is needed to describe the process and the rationale for accepting the mathematical limitations of the selected approach.

The relay identification and screening was adequately performed, however, at the time of the peer review of the Vogtle SPRA, the effect o~ chatter of those relays that were not screened was not yet included in the logic model. Recent SPRAs have shown that relay chatter is an important contributor to the overall seismic risk profile and therefore this model limitation needs to be addressed.

The Vogtle SPRA has extensive documentation of the screening process used for the inclusion of seismic-induced failure in the model. The only notable exception is on the inclusion of very small LOCA. Uniquely from a number of other seismic PRAs, the Vogtle SPRA relied on actual walkdowns of RCS lines to support not modeling very small LOCA. This is a notable effort but the unique approach of the Vogtle SPRA in treating and screening very small LOCA requires more supporting documentation.

The model provided for peer review was not fully completed in the sense that (apart from the missing relay information) the refinement of fragilities was not yet completed for the most significant contributors. At the time of the peer review, the most significant contributors were associated with seismic-induced floods that were then judged to be extremely conservative. While the fragility analysis of these components is being refined, potential modeling considerations also need to be performed. For example, if the current lead contributors would stand, the modeling of operator actions in those scenarios would not be consistent with the Capability Category II of the associated supporting requirement from Part 2.

Documentation of the Vogt le SPRA seismic equipment list (SEL) was judged to be best practice by the peer review team. The database generated to track SEL items and to link it to relay assessment, fragilities and modeling is extremely powerful and well designed and conceived.

The Vogtle SPRA is quantified with the CAFTA suite of codes (i.e., CAFTA, FRANX, ACUBE, and UNCERT) but also requires some post-processing to address code limitations. The documentation of the post-processing is sometimes missing and will be essential to be able to re-produce the results.

The quality of the documentation of the results and insights is impacted by the model not having reached a steady state (i.e., fragility refinements, relays, etc.) and some documentation details were missing for that reason (e.g., documentation of insignificant cutsets). Nevertheless, the Vogtle SPRA team demonstrated the capability to use the model to guide the refinement of the analysis at different levels.

The Vogtle SPRA team essentially relied on the UNCERT code for the propagation of the uncertainties in the SPRA. There is little explanation or documentation of the meaning of the uncertainties results.

The peer review concluded that the Vogtle seismic PRA model is of good quality and integrates the seismic hazard, the seismic fragilities, and the systems-analysis aspects appropriately to quantify core damage frequency and large early release frequency. The Page 71 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 peer reviewers also noted that the relay modeling for the SPRA model that was reviewed was not yet fully completed but stated that the Vogtle SPRA team demonstrated the ability to use the existing SPRA to obtain results and interpret the insights. Some limitations in the documentation were noted, primarily attributed to the closeness in time of the peer review to the completion date of the model but the Vogtle SPRA team demonstrated good knowledge of the inherent limitations of the model and of the conservatism applied. The SPRA Database developed by the team was recognized as an outstanding tool which demonstrates how the Vogtle SPRA team is managing the model and its insights.

No new methodologies have been incorporated into the SPRA model since the peer review.

A.5.

Summary of the Assessment of Supporting Requirements and Findings Table A-1 presents a summary of the SRs graded as "Not Met" or not "Capability Category II," and lists the Finding F&Os associated with those SRs along with the disposition for each. Table A-2, provided at the end of this document due to its size, presents a summary of all the Finding F&Os and the disposition for each.

Page 72 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table A-1 Summary of SRs Graded as Not Met or Capability Category I for Supporting Requirements Covered by the VEGP SPRA Peer Review SR Assessed Associated Disposition to Achieve Met or Capability Finding F&Os Capability Category II Category SHA SHA-C4 Not Met 12-18, 12-36 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SHA-Hl Not Met 12-18, 12-36 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SHA-11 Not Met 12-15 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SHA-12 Not Met 12-15 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SHA-Jl Not Met 12-1, 12-2, Associated F&O Findings have been resolved as 12-11, 12-16 noted in Table A-2. SR is judged to be Met.

SHA-J3 Not Met 12-8 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SFR SFR-A2 CC-I 14-1, 14-7, Associated F&O Findings have been resolved as 14-10 noted in Table A-2. SR is judged to be Met at CC-II.

SPR SPR-B2 Not Met 16-4, 16-6 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SPR-B4 Not Met 16-1 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

SPR-Fl Not Met 12-31, 16-5 Associated F&O Findings have been resolved as noted in Table A-2. SR is judged to be Met.

Page 73 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 A.6.

Summary of Technical Adequacy of the SPRA for the 50.54(f) Response The set of supporting requirements from the ASME/ANS PRA Standard [4] that are identified in Tables 6-4 through 6-6 of the SPID [2] define the technical attributes of a PRA model required for a SPRA used to respond to implement the 50.54(f) letter. The conclusions of the peer review discussed above and summarized in this submittal demonstrates that the VEGP SPRA model meets the expectations for PRA scope and technical adequacy as presented in RG 1.200, Revision 2 [11] as clarified in the SPID [2].

The main body of this report provides a description of the SPRA methodology, including:

o Summary of the seismic hazard analysis (Section 3) o, Summary of the structures and fragilities analysis (Section 4) o Summary of the seismic walkdowns performed (Section 4) o Summary of the internal events at power PRA model on which the SPRA is based, for SCDF and SLERF (Section 5) o Summary of adaptations made in the internal events PRA model to produce the seismic PRA model and bases for the adaptations (Section 5)

Detailed archival information for the SPRA consistent with the listing in Section 4.1 of RG 1.200 Rev. 2 is available if required to facilitate the NRC staff's review of this submittal.

The Vogtle Electric Generating Plant Units 1 and 2 SPRA reflects the as-built and as-operated plant as of the cutoff date for the SPRA, August 31, 2015. The SPRA model does not credit portable or offsite FLEX capabilities for response to extended loss of offsite power or loss of ultimate heat sink response. Certain aspects of FLEX are permanently installed and operational without operator intervention, i.e., notably improved RCP seals, and these are reflected in the internal events PRA model, and therefore the SPRA model.

The peer review observations and conclusions noted in Section A.4, the F&O finding dispositions noted in the discussion in.Section A.5, and the discussion in Section A.7 demonstrate that the VEGP SPRA is technically adequate in all aspects for this submittal.

Subsequent to the SPRA peer review, the peer review findings have been appropriately dispositioned, and the SPRA model has been updated to reflect these dispositions and further refine several fragility values. The results presented in this submittal reflect the updated model as of January 2017. No changes were made in updating the model that would require a subsequent focused peer review.

Page 74 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NITF 2.1 Seismic PRA Submittal Version 0 - March 2017 A.7.

Summary of SPRA Capability Relative to SPID [2] Tables 6-4 through 6-6 The Owners Group performed a full scope peer review of the VEGP internal events PRA and internal flooding PRA that forms the basis for the SPRA to determine compliance with ASME PRA Standard, RA-S-2002, including Addenda RA-Sb-2005 (27] and RG 1.200 Rev. 1 (28] in in May 2009. This review found that all but 3 supporting requirements (SRs) met at least Capability Category II. All of the internal events and internal flooding PRA peer review findings that may affect the SPRA model have been addressed. Since the time of the peer review, the internal events PRA has been maintained in accordance with the requirements for model configuration control in the PRA Standard.

The PWR Owners Group peer review of the VEGP SP RA was conducted in November 2014.

The results of this peer review are discussed above, including resolution of SRs not assessed as meeting Capability Category II by the peer review, and resolution of peer review findings pertinent to this submittal. The peer review team expressed the opinion that the VEGP seismic PRA model is of good quality and integrates the seismic hazard, the seismic fragilities, and the systems-analysis aspects appropriately to quantify core damage frequency and large early release frequency. The general conclusion of the peer review was that the VEGP SPRA is judged to be suitable for use for risk-informed applications.

Table A-1 provides a summary of the disposition of SRs judged by the peer review to be not met, or not meeting Capability Category II.

Table A-2 provides a summary of the disposition of the SPRA peer review findings.

Table A-3 provides an assessment of the expected impact on the results of the VEGP SPRA of any SRs and peer review Findings.

Of the peer review finding-level Facts and Observations (F&Os) listed in Table A-2, most were associated with PRA Standard supporting requirements (SRs) that were deemed by the peer reviewers to be either "Met" or met at "Capability Category II." This indicates, as can be seen from the finding details, that these findings deal with relatively focused issues that have been adequately dispositioned within the reviewed methodologies, for the SPRA and for future risk-informed application. Many of these were documentation-related.

The remaining finding-level F&Os are associated with SRs deemed by the peer reviewers to be "Not Met", or to not meet "Capability Category II." These are as listed in Table A-3.

Page 75 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 SR SHA-C4 SHA-Hl SHA-11 SHA-12 SHA-Jl SHA-J3 SFR-A2 SPR-B2 SPR-B4 SPR-Fl Table A-3 Findings associated with Not Met/CC-I SRs Findings Summary of Issue Not Impact on SPRA Results Fully Resolved 12-18, 12-36 Finding issues are resolved.

No impact on SPRA results.

12-18, 12-36 Finding issues are resolved.

No impact on SPRA results.

12-15 Finding issues are resolved.

No impact on SPRA results.

12-15 Finding issues are resolved.

No impact on SPRA results.

12-1, 12-2, 12-11, 12-16 Finding issues are resolved.

No impact on SPRA results.

12-8 Finding issues are resolved.

No impact on SPRA results.

14-1, 14-7, 14-10 Finding issues are resolved.

No impact on SPRA results.

16-4, 16-6 Finding issues are resolved.

No impact on SPRA results.

16-1 Finding issues are resolved.

No impact on SPRA results.

12-31, 16-5 Finding issues are resolved.

No impact on SPRA results.

As this list indicates, there were only 10 Not Met I Capability Category I SRs associated with the finding F&Os.

Of these, 6 are seismic hazard-related SRs, for which the findings were associated with: (a) inadequate docum,entation of the hazard analysis performed; (b) demonstration that sufficient consideration has been given to more recent geologic events and associated modeling; or (c) sensitivity calculations for the models and parameters used in the site hazard. The identified issues have been addressed, as noted in the dispositions for the affected findings in Table A-2.

One of the SRs is fragilities-related. Two of the 3 findings associated with this SR deal with conservatisms that the reviewers noted, which have now been addressed within the analytical methodology that the peer reviewers found acceptable. The remaining finding is associated with a specific polar crane fragility issue, which has also been addressed within the reviewed methodology.

Three of the SRs are PRA modeling-related. Three of the findings associated with this SR are related to implementation of the seismic performance shaping factor approach in the human reliability analysis. The comments in those findings have been addressed and implemented in the SPRA model, within the reviewed methodology, without significant impact on the results. One finding was related to the relay chatter evaluation, for which the model update resolves the finding.

The last finding was related to the SPR documentation, which has been updated to resolve the finding.

The SPID [2] defines the principal parts of an SPRA, and the VEGP SPRA has been developed and documented in accordance with the SPID. The information in the tables Page 76 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 identified above demonstrates that the VEGP SPRA is of sufficient quality and level of detail for the response to the NTIF 2.1 Seismic SPRA submittal.

A.8.

Identification of Key Assumptions and Uncertainties Relevant to the SPRA Results.

The PRA Standard [4] includes a number of requirements related to identification and evaluation of the impact of assumptions and sources of uncertainty on the PRA results.

NUREG-1855 [13] and EPRI 1016737 [14] provide guidance on assessment of uncertainty for applications of a PRA.

As described in NUREG-1855 [13], sources of uncertainty include '.'parametric" uncertainties, "modeling" uncertainties, and "completeness" (or scope and level of detail) uncertainties.

Parametric uncertainty was addressed as part of the VEGP SPRA model quantification (see Section 5 of this submittal).

Modeling uncertainties are considered in both the base internal events PRA and the SPRA. Assumptions are made during the PRA development as a way to address a particular modeling uncertainty because there is not a single definitive approach. Plant-specific assumptions made for each of the VEGP SPRA technical elements are noted in the SPRA documentation that was subject to peer review, and a summary of important modeling assumptions and associated sensitivity evaluations is included in Section 5.

Completeness uncertainty addresses scope and level of detail. Uncertainties associated with scope and level of detail are documented in the PRA but are only considered for their impact on a specific application. No specific issues of PRA completeness were identified in the SPRA peer review.

A summary of potentially important sources of uncertainty in the VEGP SPRA is listed in Table A-4.

Page 77 of 124

VEGP Units 1 and 2 10 CFR 50.54(f} NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table A-4 Summary of Potentially Important Sources of Uncertainty PRA Summary of Treatment of Sources of Potential Impact on SPRA Element Uncertainty per Peer Review Results Seismic The VEGP SPRA peer review team noted that With regard to aleatory and Hazard both the aleatory and epistemic uncertainties epistemic uncertainties in site have been addressed in characterizing the response analysis, there is an seismic sources. In addition, uncertainties in abundance of site-specific data from VEGP Units 3&4 that each step of the hazard analysis were reduces epistemic uncertainty propagated and displayed in the final to an insignificant level. The quantification of hazard estimates for the documentation has been VEGP site.

expanded to demonstrate that The review team commented that the site the current analysis adequately response analysis did not fully evaluate and represents the Vogtle site.

model aleatory and epistemic uncertainties in the site response analysis.

The characterization of the seismic hazard reasonably reflects sources of uncertainty.

Seismic No specific peer review team comments on Several of the sensitivity studies Fragilities sources of uncertainty in fragilities.

described in Section 5.7 of this report evaluate the impact of changes to fragilities on the SPRA results as one means of assessing the impact of fragilities uncertainties on the SPRA results. No changes to the model were recommended based on these results.

Seismic The peer review team commented that the The discussion of uncertainty PRA Vogtle SPRA team relied on the UNCERT code has been expanded in the SPRA Model for the propagation of the Quantification (QU) report, uncertainties in the SPRA with little including a discussion of explanation or documentation of the meaning sources of model uncertainty, of the uncertainties results.

and potentially important sources have been addressed in the sensitivity analysis. A characterization of the mean SCDF and SLERF is provided in Section 5.6 of this report.

Several sources of model uncertainty are discussed in Section 5.7, and sensitivities performed to evaluate the impact of possible changes to address these.

Page 78 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 A.9.

Identification of Plant Changes Not Reflected in the SPRA The VEGP SPRA reflects the plant as of the cutoff date for the SPRA, which was August 31, 2015. Table A-5 lists provides a summary of plant changes not included in the model and provides a qualitative assessment of the likely impact of those changes on the SPRA results and insights. There are no significant plant changes that have not been reflected in the current SPRA model.

Table A-5 Summary of Significant Plant Changes Since SPRA Cutoff Date Description of Plant Change Impact on SPRA Results Safety-related battery chargers are no An assessment of this change on the VEGP longer operated in a load-share internal events PRA model indicated no configuration. Instead, a single charger significant impact. Further, the battery chargers will be in service and if it fails, the other are modeled as seismically correlated. Thus, charger will be placed in service by modeling of the change in the SPRA would not operator action.

affect the SPRA results.

Permanently installed and portable FLEX Credit for such equipment is likely to improve equipment other than low leakage RCP the SPRA results (SCDF and SLERF) but the seals have not been modeled in the SPRA.

impact is difficult to quantify without detailed modeling.

Page 79 of 124

_I

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Table A-2 Summary of Finding F&Os and Disposition Status 1 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SHA-E2 11-3 While variability in the mean To maintain hazard-consistent Expand documentation to There is an abundance base-case Vs profile is ground motion hazard at the demonstrate that a single of site-specific Vs data incorporated in the site control point, the site response base-case Vs profile from VEGP Units 3&4, response analysis, no analysis needs to incorporate adequately represents the which reduces epistemic epistemic uncertainty in the appropriate epistemic uncertainty Units 1 &2 site. Or if that is not uncertainty to an base-case profile is and aleatory variability in its the case, include epistemic insignificant level.

represented. Documentation inputs. The Vs profile for the uncertainty in the of the justification for this Vogtle Units 1 &2 site is characterization of Vs profile Additional discussion of assessment should be represented by a single Vs profile, and evaluate the impact on the rationale for use of a expanded.

indicating there is no epistemic control point ground motions.

single base-case Vs uncertainty in the mean base-profile for the site has (This F&O originated from SR case profile. Documentation of been included in the SHA-E2) this assessment needs to be documentation. The expanded.

added discussion demonstrates that a Discussion with staff indicates single base-case shear-that consideration of the wave velocity (Vs) profile combined data for the Vogtle site adequately represents (Units 1 &2, Units 3&4, ISFSI) the Vogtle site, based on provides sufficient confidence that the availability of Vs a single mean base-case profile data, which reduces the characterizes the site. This epistemic uncertainty for conclusion is based on the this particular parameter.

quantity and quality of the This finding has been combined data and an evaluation showing the site is relatively resolved with no uniform with respect to Vs. For significant impact to the some depth ranges, data from the SPRA results or nearby Savannah River Site conclusions.

(SRS) are used to support the 1 In Table A-2, all but the last column are extracted directly from the Peer Review report. The last column provides the disposition for the Findings.

Page 80 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description profile interpretation.

Bechtel Document 23162-000-G65-GEK-00010 (SNC #SVO-GB-X?R-011-001) presents summaries of velocity data, but does not provide sufficient information to support the lack of epistemic uncertainty at the Units 1 &2 site over the complete depth range of the Vs profile. This would typically require multiple measurements throughout the depth range that provide a consistent picture of natural variability about a single mean base-case profile. The technical basis and justification that a single base-case profile is appropriate should be provided in more detail.

This should include the basis for applying conclusions from other Vogtle locations to the Units 1 &2 site.

[A related Suggestion 11-2 addresses specifically potential epistemic uncertainty in the Blue Bluff Marl stratum.]

Page 81 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SHA-E2 11-8 Upper crustal site attenuation Calculation X2CFS129 Ver2 Provide a basis in the A discussion of the of ground motion (kappa) is, notes that the damping documentation for range of possible values generally, an uncertain associated with the base-case representing base-case kappa of deep soil damping has parameter. Thus, to maintain profile corresponds to a total at the site by a single value.

been included in the hazard-consistent ground kappa value for the soil column of The basis might include documentation.

motion at the control point, 0.01 sec. The report does not sensitivity analyses to show this uncertainty should be address epistemic uncertainty in the impact of epistemic A sensitivity study on the incorporated in the site kappa.

uncertainty in kappa.

epistemic uncertainty of response analysis, or the deep soil damping has basis for not including it In discussion with staff during the been performed using should be provided. In either peer review, it was noted that median, lower range, case, the technical basis and randomization of the damping and upper range justification should be associated with the profile layers alternatives for deep documented.

represents both random variability rock damping. Site and epistemic uncertainty. It was response analysis was also noted that kappa was performed using 1 E-4 expected to be small for the HF and LF rock input (This F&O originated from SR Vogtle site and uncertainties in motion. The resulting SHA-E2) that small value would not be amplification functions expected to have a significant and Jog-standard impact on site amplification. Staff deviation were weight-also noted that the approach used averaged and compared had been reviewed by the NRC to the original base case for the Vogtle ESP and COLA.

for each of BBM High Pl and BBM Low Pl soil The SPID (EPRI, 2013) provides columns. It was guidance accepted by the NRC concluded that the for response to NTIF 2.1 inclusion of alternative Recommendation: Seismic that base cases for deep soil indicates kappa is difficult to damping to account measure and thus subject to large explicitly for the uncertainty (SPID Section B-epistemic uncertainty 5.1.3.2).

associated with site kappa does not have Documentation of the technical any significant effects on Page 82 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution basis for kappa characterization the resulting seismic should be expanded.

hazard curves and UHRS.

The sensitivity study has been added to the SPRA documentation.

This finding has been resolved with no significant impact to the SPRA results or conclusions.

SHA-J1 12-1 As part of the PSHA The approach that was taken to Documentation should be A PSHA report has been implementation, the analyst model earthquakes in the PSHA provided that describes how prepared that describes has different alternatives for calculation was not identified.

seismic sources are modeled how earthquake events modeling the earthquake There are two basic alternatives in the PSHA (i.e., how the were modeled for area occurrences in the that can be used to model SSC and GMMs) were sources in the PSHA calculations. The PSHA earthquake events; as extended implemented in the Vogtle calculations. This was documentation does not fault ruptures, or as point sources.

PSHA.

by modeling each earth-describe the approach that The approach that is used quake as a point source, was used to model influences how the CEUS ground and using correction earthquakes.

motion model is implemented.

factors for distance and ground motion uncertain-No documentation is provided on ty that modify the ground either of these subjects motion estimate to in-(This F&O originated from SR (earthquake source modeling and elude the effect of a SHA-J1) use of the ground motion closer distance to a fault attenuation models). From rupture (because the questions posed to the PSHA rupture may be closer to analysts, it is our understanding the site than the single that earthquakes were modeled point used to represent as point sources and the that event) and the un-appropriate ground motion certainty in ground Page 83 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description aleatory uncertainty was used in motion because the the calculation.

azimuth of the rupture is unknown. These correc-tion factors were pub-lished in EPRI (2004)

(12].

This finding has been resolved with no significant impact to the SPRA results or conclusions.

SHA-J1 12-11 As part of the PSHA The PSHA analysts were asked to Provide a description of the A PSHA report has been implementation, the analyst describe the approach that was earthquake modeling prepared that describes has alternatives for modeling used to model earthquakes in the approach that was used to how pseudo-faults were the earthquake occurrences Charleston RLME seismic source.

model the Charleston RLME implemented to in the calculations. The PSHA The response indicated that seismic source and how the represent the Charleston documentation does not earthquakes in the Charleston approach was implemented.

RLME source. This describe the approach that RLME source were modeled includes: 1. A description was used to model using 'pseudo faults'.

of the pseudo-faults. 2. A earthquakes in RLME definition of pseudo-sources.

The PSHA report does not:

faults as constructed

1. Describe that a 'pseudo fault' faults that represent approach was used to model possible sources of earthquakes in the Charleston future large earthquakes.

(This F&O originated from SR RLME source.

3. Implementation of the SHA-J1)

2. Provide a definition of 'pseudo pseudo-faults including faults'.

spacing and limits at the

3. Describe how the 'pseudo fault' borders of the approach was implemented for Charleston source. 4.

the Charleston RLME seismic Documentation of the source (e.g., what was the fault rupture area, length, and spacing that was used; how was width that were the earthquake rate distributed to estimated for possible Page 84 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution the faults, etc.).

future earthquakes. 5. A

4. Document the fault rupture description of how model that was used.

earthquake ruptures are

5. Describe how earthquake distributed on the faults.

events are distributed on the faults.

This finding has been resolved with no significant impact to the SPRA results or conclusions.

SHA-11, 12-15 A screening assessment was A screening analysis was not A screening analysis for other This evaluation was performed for soil liquefaction performed for hazards such as seismic hazards should be done for the Vogtle 3&4 SHA-12 and is described in seismic settlement, fault displacement, performed and documented COLA [38] and is noted fragility calculation (PRA-BC-tsunami, seiche, etc.

as part of the PSHA and in the ESP SAR [36].

V-14-025).

SPRA.

The Vogtle 3&4 It is anticipated these other evaluation is applicable A screening assessment was seismic hazards will be screened It is expected that information to, and has been cited not performed for other out.

in the FSAR for Vogtle 1 & 2 in, the Vogtle 1 &2 SPRA potential seismic hazards.

and in the COLA for Units 3 &

Fragility report.

4 can be used to support this (This F&O originated from SR requirement.

This finding has been SHA-11) resolved with no significant impact to the SPRA results or conclusions.

SHA-J1 12-16 The Vogtle PSHA has gone The documentation of the PSHA Prepare a complete and up-A PSHA report has been through a number of changes is provided in a collection of to-date PSHA document that prepared that includes and revisions since 2012 due documents that were prepared in includes all results, sensitivity hazard results, to changes in models, input the 2012-2014 time frame. There calculations, deaggregation uncertainties in hazard, data, etc. As new calculations does not exist a single document results, etc. that is based on and sensitivities to input were performed and reports that contains a set of results that the current model.

uncertainties; this generated, sensitivity results, is based on the current PSHA summarizes hazard were not carried forward. As a model.

results for the Vogtle result, there does not exist a site.

current report that includes all Page 85 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description PSHA results, This finding has been deaggregations, etc. that is resolved with no based on the current PSHA significant impact to the model.

SPRA results or conclusions.

(This F&O originated from SR SHA-J1)

SHA-82, 12-18 The Vogtle PSHA is based on As part of a site-specific PSHA, A data gathering effort should A detailed study of new SHA-C4, the CEUS SSC seismic there is a need to gather, review be undertaken to identify new geological, SHA-H1 source model which was and evaluate new geological, information that post-dates the seismological, and completed in 2012. The SSC seismological, or geophysical CEUS SSC data collection geophysical information model was a developed at a information or information that is effort. The data gathering was conducted, to regional scale that was based defined at a scale that was not effort should also look for determine if any on data gathered up until considered in the development of information local to the Vogtle information subsequent about 2010. (Note, the date the CEUS SSC model. As part of site region that was not to the EPRI SSC model when data was gathered the Vogtle SPRA, no effort was considered, or at a scale that (EPRI, 2012 [35]) is varied; for example the made to gather up-to-date and was not addressed as part of available that should be earthquake catalog was local (local to the Vogtle site) the CEUS SSC regional incorporated into the complete through 2008.) In information to evaluate whether evaluation.

seismic hazard results the sense that the CEUS SSC any new information has become for Vogtle. This study is model was not specifically available on active faulting and/or Some of this information may described in the SPRA performed as a site-specific the development new seismic be available in the COLA for documentation. While PSHA for the Vogtle site.

sources or the revision of sources Vogtle Units 3 & 4.

the area around the site in the CEUS SSC model in the continues to be studied (This F&O originated from SR vicinity of the Vogtle plant.

by many earth scientists, SHA-82) there was no new Since up-to-data was not information identified gathered, consideration of that would change the alternatives could not be estimate of seismic addressed.

hazard for Vogtle.

This finding has been resolved with no significant impact to the Page 86 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SPRA results or conclusions.

SHA-J1 12-2 The method that is used in For soil sites, the soil hazard is The documentation should The methodology used the Vogtle PSHA to estimate generally (though not exclusively, include a description of the for the surface hazard the soil site hazard is not since other methods could be methodology that is used to calculation has been described or referenced.

used) determined in two steps; combine the rock hazard described in detail, and a probabilistic rock hazard results results and the site comparison made are estimated which are then amplification factors to between the GMRS combined with probabilistic determine the soil hazard at using the two (This F&O originated from SR estimates of the site response.

the Vogtle site.

approaches 2A and 3.

SHA-J1)

The method used in the Vogtle Approach 2A was used PSHA to estimate the soil hazard for the calculation of SSI is not described.

input motions at foundation elevations and Approach 3 was used for the calculation of surface hazard and GMRS at the ground surface, as defined in NUREG/CR-6728 [34]. It was concluded that the use of Approach 2A USHRS as input to the SSI analysis of the Vogtle plant is considered acceptable and does not present any significant inconsistency with the seismic hazard curve and GMRS at the ground surface, which were Page 87 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Number Finding Basis Resolution Disposition ment(s)

Description calculated using Approach 3.

This finding has been resolved with no significant impact to the SPRA results or conclusions.

SHA-E2 12-22 The site response Calculation The site response calculation A framework and approach for A description of the X2CFS129 Ver. 1 (2012) and does not present a clear evaluating and modeling methodology used to Ver. 2 (2014) does not description of how aleatory and uncertainties in the site account for epistemic describe a framework for epistemic uncertainties are response should be and aleatory evaluating and characterizing identified and evaluated. As a developed and implemented.

uncertainties in soil sources of aleatory and result it is difficult to track the The site response calculation hazard has been added epistemic uncertainty and propagation of uncertainties is documentation should fully to the documentation.

how the approach was carried out in the site response describe the methodology and implemented.

analysis.

its implementation.

This finding has been resolved with no (This F&O originated from SR It is worth noting that there is significant impact to the SHA-E2) some epistemic site response SPRA results or uncertainty that is accounted for conclusions.

in the rock GMPEs.

SPR-E5 12-23 The quantification process Table 5-1 presents the results of Develop and document an Additional detail has has included the uncertainties three different uncertainty understanding of the earlier been added to the SPRA in the seismic hazard, fragility calculations for GDF and LERF. In point estimate results for GDF Quantification report to and systems-analysis addition, point estimates for GDF and LERF (as reported in document the elements of the SPRA. The and LERF are calculated and Sections 3 and 4) and of uncertainty, importance, results in Table 5.1 are reported in Section 5.1.1. Thus uncertainty results.

and sensitivity analyses internally inconsistent and are the table reports two estimates of and relate the inconsistent with the results the mean GDF and LERF uncertainty analysis reported in Sections 3 and 4 respectively from different mean GDF and LERF to for GDF and LERF, uncertainty calculations and a the point estimate respectively.

'Point Estimates' result for each.

values.

Page 88 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description All of these results are different (This F&O originated from SR than the point estimate This finding has been SPR-E5)

(approximate mean) reported in resolved with no Sections 3 and 4 for CDF and significant impact to the LERF, respectively. The SPRA results or documentation in the report does conclusions.

not describe the basis (inputs) for these calculations, or offer an interpretation of the results.

SPR-E5 12-24 The Quantification report The uncertainty analysis is Provide documentation of the Additional detail has does not provide presented in Section 5.1 with the uncertainty analysis that been added to the documentation of the results reported in Table 5.1. The describes the results, how documentation of the uncertainty analysis results.

report provides limited discussion they are being interpreted and seismic plant response of the results and the insights that the insights that are derived model, model might be gained from them.

from them.

implementation, and quantification in the QU (This F&O originated from SR The two sets of results that are report. In addition, the SPR-E5) reported in Table 5-1 are not uncertainty, importance, discussed in terms of their and sensitivity analyses relationship to each other. For are described in more instance the mean values should detail.

be the same (but are not). The uncertainty estimates provide This finding has been insight to the total uncertainty and resolved with no the contribution of the basic event significant impact to the uncertainty to the total.

SPRA results or conclusions.

In addition, neither Table 5.1 or the discussion identifies what is the 'final' uncertainty result that includes the propagation of uncertainties of all elements of the Page 89 of 124 L

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SPRA to the estimates of CDF and LERF.

SPR-E5 12-26 There are differences in the The report does not present the Document the results of Updated Monte Carlo results for CDF and LERF results of sensitivity calculations sensitivity calculations on the uncertainty runs have that are reported in Table 5.1.

with regard to the number of number of Monte Carlo been performed with A possible contributor to Monte Carlo simulations that are simulations required to 20,000 iterations for these differences may be due needed to produce stable results.

produce stable results.

SCDF and SLERF. This to the number of Monte Carlo is a sufficiently high simulations that were It is our understanding from number of simulations to performed.

discussion with the PRA staff that produce a stable result.

these types of sensitivity The SPRA (This F&O originated from SR calculations were performed.

documentation has been SPR-E5) updated to clearly indicates the results.

This finding has been resolved with no significant impact to the SPRA results or conclusions.

SPR-F2 12-27 Documentation should be The current quantification Provide clear and complete The QU report provided that describes how document does not provide a documentation of the documentation has been the plant model analysis is clear description of the how the approach used to quantify the updated to describe the quantified.

plant model is quantified. For seismic plant response model, quantification process, example the discussion does not to perform the risk including the technique identify how calculations are quantification, uncertainty for combining cutsets performed, what the limitations of analysis, and importance over the 14 acceleration (This F&O originated from SR these quantifications are and how analysis.

intervals, and obtaining SPR-F2) they affect the results.

the importance measures.

This finding has been resolved with no significant impact to the Page 90 of 124

I I

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SPRA results or conclusions.

SPR-E2 12-29 The Quantification report There is limited documentation of Document the process and Additional detail has provides limited the process and the numerical methods that were used to been added to the QU documentation of the process methods that were used to perform the uncertainty report to document the and methods that were used perform the uncertainty analysis.

analysis. Where appropriate uncertainty, importance, to perform the uncertainty Based on the documentation that document where and sensitivity analyses analysis.

is provided and discussions with consistencies and potential and relate the the PRA staff there is limited but inconsistencies in results uncertainty analysis not complete understanding of the might be expected.

mean SCDF and SLERF methods that were used and the to the point estimate (This F&O originated from SR relationship of these methods to values.

SPR-E2) the results were obtained (reported in Table 5.1 ).

This finding has been resolved with no In some cases (as described in significant impact to the the documentation) the results SPRA results or from the uncertainty analysis conclusions.

(Table 5.1) are not the same as the results reported in Sections 3 and 4 for GDF and LERF (though this connection is not clearly stated in the report). However, it would seem the results in Table 5.1 should be internally consistent.

SPR-F1 12-31 The standard requires a level There is limited documentation Documentation should be Additional detail has of documentation that that describes the seismic plant provided in sufficient detail been added to the provides an understanding of response analysis and that describes the seismic documentation of the the seismic plant response quantification; how the model was plant model, how it is seismic plant response model and the quantification.

implemented, how the implement and quantified.

model, model This requirement is not met.

quantification was performed and implementation, and a discussion of the analysis quantification in the QU Page 91 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution results.

report. In addition, the uncertainty, importance, (This F&O originated from SR To meet this requirement, the and sensitivity analyses SPR-F1) documentation must be in are described in more considerable detail in order to detail.

support the review process and future updates. Part of the This finding has been documentation should include a resolved with no detailed discussion of the results, significant impact to the sensitivity calculations, and the SPRA results or uncertainty analysis.

conclusions.

SPR-F3 12-32 The documentation of the The purpose of this supporting Document and discuss the The documentation of sources of model uncertainty requirement is that documentation contribution of the different the uncertainty analysis and a description of the should be presented that sources of uncertainty that are has been expanded in analysis assumptions is not addresses the sources of modeled in the SPRA.

the Quantification report.

complete in the SPRA epistemic (knowledge) uncertainty A discussion of sources quantification report. In that are modeled and their of model uncertainty has addition, there is not a clear contribution to the total been added to the description of the uncertainty uncertainty in GDF and LERF.

report, and potentially analysis and the contributors important sources have to the total uncertainty beyond In addition, the documentation been addressed in the a simple report from should discuss elements of the sensitivity analysis.

UN CERT.

seismic plant model where there may be latent sources of This finding has been (This F&O originated from SR uncertainty that are not modeled resolved with no SPR-F3) and assumptions that are made in significant impact to the performing the analysis.

SPRA results or conclusions.

Page 92 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SHA-83, 12-36 As part of a site-specific As part of the Vogtle PSHA an An up-to-date earthquake An update to the SHA-C4, PSHA, an up-to-date effort was not made to gather catalog for the Vogtle site earthquake catalog was SHA-H1 earthquake catalog should be data on earthquakes that region should be developed to prepared from the time used. The CEUS SSC study occurred since 2008. As such, the assess whether modifications of the CEUS SSC involved the development of a analysts did not assess whether to the seismic source catalog (through 2008) comprehensive earthquake more recent seismicity is recurrence parameters or through February 2016.

catalog based on data consistent with the required. The updated The rate of occurrence through 2008. The Vogtle characterization parameters catalog, resources used in of earthquakes within site-specific PSHA should estimated as part of the CEUS compiling the update and the 320 km of the Vogtle site consider the impact SSC of SSC study (NRC, 2012).

results of the evaluation was compared to the any additional seismicity since should be documented as part rate of earthquakes 2008 up to the time the study We note that as part of the Vogtle of the PSHA. If more recent represented by the started.

PSHA, calculations were seismicity is not consistent CEUS SSC seismic performed to recompute the with the existing CEUS SSC source model for that

{This F&O originated from SR seismic hazard at the site to take seismic source parameters, same area, this SHA-C4) into account changes in the the parameters should be comparison being made CEUS SSC earthquake catalog updated and the PSHA should for M>2.9. It was found through 2008 that were made be updated.

that the updated catalog following the completion of the implied a rate of CEUS SSC study. These changes earthquakes that is lower reflect the identification of than the mean rate from reservoir induced seismicity the CEUS SSC seismic earthquakes and the re-sources. Therefore, interpretation of the location of incorporating the effects some earthquakes in the of a updated catalog on Charleston, SC area that occurred the hazard at Vogtle in the 1880's (EPRI, 2014).

would decrease the hazard slightly, and was References not undertaken. This comparison is EPRI (2014). Review of EPRI documented in the 1021097 Earthquake Catalog for SPRA documentation.

RIS Earthquakes in the Southeastern U. S. and This finding has been Earthquakes in South Carolina resolved with no Page 93 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description Near the Time of the 1886 significant impact to the Charleston Earthquake SPRA results or Sequence, transmitted by letter conclusions.

from J. Richards to R. McGuire on March 5, 2014.

SHA-J3 12-8 A foundational element of The documentation of the sources The resolution to this finding Sources of uncertainty in PSHA as it has evolved over of model uncertainty analysis and could involve:

the seismic hazard the past 30 years is the a description of the analysis analysis for Vogtle are development and implemen-assumptions is not complete in

1. Documentation and discussed in the updated tation of methods to identify, the PSHA report in its current discussion of the contribution SPRA documentation.

evaluate, and model sources form such that a clear of different sources of These include of epistemic (model and understanding of the contribution uncertainty that are modeled uncertainty in seismic parametric) uncertainty in the of individual sources of in the PSHA. The source model (for estimate of ground motion uncertainty to the estimate of documentation of the background earthquake hazards. As such fairly hazard are understood. Limited contribution of different sources and for the rigorous analyses are carried information on the contribution of sources of uncertainty can be Charleston RLME), in out (SSHAC studies) to seismic sources to the total mean shown by means of 'tornado maximum magnitude for quantitatively address model hazard is presented, but plots' that quantify the background seismic uncertainties.

information on the contributors to sensitivity of the hazard at sources and for the the uncertainty is not provided.

different ground motion levels Charleston RLME, in At the same time there is to the various branches in the ground motion prediction within any analysis sources of With respect to addressing model logic tree. These plots show equation, in smoothing uncertainty that are not uncertainties and associated which sources of epistemic assumptions for directly modeled and assumptions there are some uncertainty are most seismicity parameters in assumptions that are made examples that can be identified in important. It should include background sources, for pragmatic or other the Vogtle PSHA. For example, in the source model uncertainty, and in site amplification reasons. There are also the site response analysis the ground motion model model. "Tornado plots" sources of model uncertainty assumption is made that the 1 D uncertainty, and site response are included in the that are embedded in the equivalent linear model (SHAKE uncertainty. Currently, the updated SPRA context of current practice type) to estimate the site total uncertainty is shown by documentation that show that are 'accepted' and amplification and ground motion the hazard fractiles, but it is the contribution to total typically not subject to critical input to plant structures is not broken down to provide uncertainty in seismic review. For instance, in the appropriate.

understanding as to what is hazard from source PSHA it is standard practice most important.

model uncertainty, Page 94 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution to assume that the temporal maximum magnitude occurrence of earthquakes is

2. Identification and uncertainty, ground defined by a Poisson process.

discussion of model motion prediction This assumption is well assumptions that are made.

equation uncertainty, accepted despite the fact that smoothing assumptions it violates certain funda-for seismicity parameters mentally understanding of in background sources, tectonic processes (strain and site response accumulation). A second uncertainty. These plots practice is the fact that are presented for 1 O Hz earthquake aftershocks are and 1 Hz spectral not modeled in the PSHA, acceleration, for ground even though they may be motion amplitudes significant events (depending corresponding to mean on the size of the main event).

annual frequencies of exceedance of 1 E-4 and In the spirit of the standard it 1 E-5. These "tornado seems appropriate that plots" show that ground sources of model uncertainty motion prediction that are modeled as well as equation is the major sources of uncertainty and contributor to seismic associated assumptions as hazard uncertainty for they relate to the site-specific both 10 Hz and 1 hz analysis should be identified/

spectral acceleration, discussed and their influence and maximum on the results discussed.

magnitude of the Charleston RLME As SPRA reviews and the use source is an important of the standard has evolved, it contributor for 1 Hz would seem the former spectral acceleration.

interpretation is reasonable, but potentially incomplete. It is The use of equivalent reasonable from the perspec-linear one-dimensional tive that document-tation of site response analysis, the sources of model uncer-and its associated Page 95 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description tainty and their contribution to assumptions, and its the site-specific hazard adequacy for the Vogtle results is a valuable product site are documented in that supports the peer review the hazard calculation.

process and assessments in the future as new information This finding has been becomes available). Similarly, resolved with no documenting assumptions significant impact to the provides similar support for SPRA results or peer reviews and future conclusions.

updates.

The notion that model uncertainties and related assumptions that are not addressed in the PSHA is at a certain level an extreme requirement that may not be readily met and may not be particularly supportive of the analysis that is performed.

For purposes of this review, the following approach is taken with regard to this supporting requirement:

1. The documentation should present quantitative results and discussion the sources of epistemic uncertainty that are modeled and their contribu-tion to the total uncertainty in the seismic hazard.

2. The documentation should discuss elements of the Page 96 of 124

VEGP Units 1 and 2 10 CFR 50.54{f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description PSHA model where their may be latent sources of model uncertainty that are not modeled and assumptions that are made in performing the analysis.

(This F&O originated from SR SHA-J3)

SFR-A2 14-1 The conservatisms that exist SFR-A2 requires that seismic Account for conservatism in Evaluation of anchorage in structural demand were not fragilities be based on plant-the building response has been updated to properly accounted for in the specific data and that they are analyses in the structure include clipping of in-estimation of component and realistic and median centered with response factor for structure response structure fragilities.

reasonable estimates of component fragility spectra, and the uncertainty.

evaluations.

methodology is documented in the The structural response factor Use clipped spectra for fragility notebook.

(This F&O originated from SR used in all component fragilities assessing anchorage SFR-A2).

reviewed is reported as 1.0. This capacities.

Structure response is factor will be greater than 1.0 dominated by the soft because of the conservatism soil on which Vogtle 1 introduced in the demand through and 2 structures are the structural analysis. Because founded. This would of this, the component and cause higher damping at structural fragilities are biased lower hazard frequency low.

levels and lead to stress similar to the stress The fragilities developed for calculated for the structures and components that buildings at 1 E-4. As a are mounted in those structures result the structural will be biased low because the response factor is close input structural demands include to 1 and is accounted for conservatisms. Time histories appropriately in the used for the SSI analysis have fragility evaluations.

been processed such that each Page 97 of 124

VEGP Units 1 and 2 10 CFR 50.54(f} NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution record envelopes the target The input time history UHRS. This will introduce some motion at the control level of conservatism. The input point in the SSI analysis motion at the control point has has been modified to been scaled to produce resultant reasonably match the FIRS that envelopes the FIRS corresponding 1 E-4 coming out of the site-consistent UHRS from the site-input motion analysis. In structure consistent input motion response spectra coming out of analysis.

the SSI analyses were not peak clipped when computing This finding has been anchorage demands. Structure resolved.

response at the calculated equipment fragility levels is considerably higher than the 1 E-4 UHRS considered in the building response analyses. The structure will have additional cracked shear walls and higher associated levels of damping at these higher ground motions.

SFR-A2 14-10 Significant conservatisms In the fragility calculations of heat Realistic nozzle loads should The CCW and ACCW were noted in several exchangers (PRA-BC-V-14-009 be determined for fragility heat exchanger sampled fragility calculations.

Appendix A), nozzle loads evaluation of heat capacities have been significantly contribute to the exchangers.

updated to reflect seismic demands which form the realistic nozzle loads.

basis for the median capacities.

The equipment fragilities (This F&O originated from SR Based on in-plant walkdowns by have been updated to SFR-A2) the peer review teams and also The equipment capacity factor account for appropriate noted in the walkdown report, the should be based on the frequency, and piping is well supported in all frequency range of interest.

uncertainty has been directions and will not impose That frequency range of considered in these significant nozzle loads during a interest is centered at the updates.

seismic event. The CCW and fundamental frequency of the Page 98of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution ACCW capacities are below the pump, and considers some This finding has been 2.5g screening level and are uncertainty in that frequency.

resolved.

significant contributors to risk so more realistic fragilities are required.

Battery rack 11806B3BN3 in calculation PRA-BC-V-14-01 O Appendix J2 is governed by GERS capacity. The GERS capacity is taken to be 1g, which corresponds to a frequency of 1 Hz. This is not realistic. The actual capacity is about 4g. The median capacity reported in the calculation is well below the 2.5g screening level and is not realistic.

The median capacity reported for the Turbine Driven Auxiliary Feedwater Pump is reported in Calculation PRA-BC-V-14-008 as 1.56g. This fragility is based on the seismic qualification document. The frequency range of interest for the fragility evaluation should be centered around the fundamental frequency of the assembly and not consider the entire frequency range.

Page 99 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SFR-G2 14-14 The iterative process used for In review of the seismic fragility Add a description of the The description of the developing realistic fragilities calculation for the safety features iterative process for iterative process for is not well documented.

sequencer (11821 U3001), it was computing the component computing fragilities has discovered that an iterative fragilities in the SPRA been documented.

process was used. The initial documentation fragility is based on EPRI 6041 This finding has been (This F&O originated from SR screening methodology and an resolved with no SFR-G2) equipment capacity factor that is significant impact to the equal to the EPRI 6041 median SPRA results or capacity divided by the peak in conclusions.

structure demand. If this value is less than the screening capacity (2.5g), then the fragility may be refined by examining the component fundamental frequency. The fragility may be further refined by examining component specific qualification test reports. However, the fragility used in the logic tree by the systems analyst is generally the highest of these computed. This is reasonable and appropriate, however, this process is not described in the fragility notebook or fragility calculations.

SFR-02 14-17 Inconsistencies and errors in Fragilities for the Vogtle 1 &2 Update SNC calculation no.

The following changes NSSS fragility development.

Nuclear Steam Supply System PRA-BC-V-14-015 to have been made: NSSS (NSSS) are based on the results incorporate corrections and fragility calculations have of the Westinghouse analysis of enhancements.

been updated to reflect record (AOR) associated with the Westinghouse-provided (This F&O originated from SR safe shutdown earthquake (SSE).

critical loads and support SFR-02)

In general, fragilities are capacities represented in developed through scaling of the the critical failure modes; Page 100 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SSE demands to the RLE and the effect of inelastic using the AOR seismic margins.

energy absorption is Various deficiencies were noted in factored in and the development of the fragilities documented in fragility associated with these calculation as components.

appropriate; the Reactor Basis: The NSSS Seismic Coolant Pump fragility fragility evaluation (SNC has been updated to calculation no. PRA-BC-V reflect the failure of the 015) includes detail calculations pump associated with for each of the major NSSS LOCA; the reactor components. It indicates that the internals fragility has critical failure modes for the been updated in the components are controlled by the calculation; and the new support capacities.

fragilities have been During the Peer Review, the team reflected in the updated members discussed these issues SPRA model.

with SNC staff to obtain insights and develop potential resolution This finding has been paths. Key issues included:

resolved.

(a) Basis for assumption that the support capacities represented the critical failure mode was not documented. SNC indicated that this was based on input from Westinghouse and NUREG-3360 and will update the fragility evaluation of provide this information.

(b) Inelastic energy absorption was not credited to increase the median capacities - this does not result in realistic median capacities (overly conservative).

(c) Reactor Coolant Pump fragility Page 101 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description was based on consideration of the failure of the attached CCW piping, due to an assumption that a small-break/RCP seal LOCA was critical. It was learned during the Peer Review that failure in the system model was linked to a large-break LOCA, so the failure mode considered in the fragility evaluation is not consistent with the system model - SNC indicated that they will revise the fragility evaluation.

(d) Reactor Internal fragility evaluation determined the demand based an average spectral acceleration over the range of 2 to 3 Hz, rather than using the peak acceleration in this range of the ISRS, and did not consider the contribution of higher modes. SNC indicated that this was done to avoid an overly conservative capacity, but agreed that the contribution of higher modes should be addressed, and will revise the calculation.

(f) Control Rod Drive Mechanism fragility evaluation assumed that material stresses were the critical failure mode, and did not address the potential impact of deflections on rod drop. SNC indicated that information provided by Westinghouse (based on a Page 102 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution Japanese testing program) indicated that the deflection levels associated with seismic loading does not impact rod drop, and agree to add this discussion to the calculation.

SFR-E4, 14-20 Seismic induced fire The only mention for seismic Seismic induced fire is an The seismic-induced fire SPR-89 evaluations are not induced fire evaluation is important element of the and flood evaluations documented in the walkdown contained in the quantification fragility evaluation process have been updated, and report or fragility calculations.

notebook. Based on discussions and this should be clearly documented in the during the peer review, it is documented.

fragility and understood that seismic induced quantification report.

fire was a key consideration This includes the details (This F&O originated from SR during the walkdowns. However, of the walkdown SFR-E4) detail of the walkdown procedure procedure used to for fire following earthquake is evaluate the potential for missing. The write up should seismically induced fires, include team composition, including the methodology, screening criteria, methodology, screening and results, criteria and results.

This finding has been resolved with no significant impact to the SPRA results or conclusions.

Page 103 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SFR-D1 14-4 A potential for sloshing SFR-D1 requires that realistic Evaluate the potential for flood The evaluation for induced inundation of the failure modes of structures and induced failure of the NSCW potential flood induced NSCVV Pumps (11202P4007, equipment that interfere with the Pumps or NSCW discharge failure of the NSCW 11202P408) and associated operation of that equipment be MOVs.

pumps or the NSCW discharge motor operated identified.

discharge MOVs has valves (1 HV11600, 11606, been performed and 11607, 11613) in the NSCW The potential for earthquake documented in the exists and was not identified induced sloshing of the water fragility calculation for either in the walkdowns or within the NSCW tower exists.

the NSCW tower. There subsequent analysis.

From field walkdowns of the wasnos~nffica~impact NSCW it was observed that there on the pump or MOV (This F&O originated from SR is a potential for sloshing of fragilities.

SFR-D1) contents to potentially splash onto or flood the pumps and or motor This finding has been operated valves on the attached resolved with no discharge piping.

significant impact to the SPRA results or conclusions.

SFR-D1 14-5 The potential for seismically-Vogtle 1 &2 is a soil site, with Develop estimates of the Documentation has been induced differential engineered fill from the rock differential settlements updated to include the settlements between interface to the finished grade.

between adjacent structures effects of earthquake structures was not addressed.

The in-scope Seismic Category I and assess the fragility of induced settlement; no structures have foundations with commodities based on their significant differential varying embedment depths, ability to accommodate the settlements were ranging from surface founded associated differential computed between the (This F&O originated from SR (elev. 220 ft.) to a foundation displacements.

structures.

SFR-D1) embedment of 110 ft. (elev. 11 O ft.). Since soils, including This finding has been engineered fill, will resolved with no consolidate/settle to some extent significant impact to the when subjected to high level SPRA results or earthquake ground motion, and conclusions.

the amount of settlement is proportional to the thickness of Page 104 of 124

VEGP Units 1 and 2 10 CFR 50.54{f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution the soil layer under the foundation, the settlement of one structure relative to another structure is dependent on the depth of the foundation embedment.

The Fragility Notebook (PRA-BC-V-14-025) does not address the potential differential settlement between buildings, or the potential effect on commodities (e.g.,

piping, electrical raceways, HVAC ducts, etc.) that cross the separation between adjacent structures. During the performance of the Peer Review, SNC personnel indicated that the consideration of differential settlements was not required, since the structures were founded on engineered fill.

SFR-G2 14-6 The results of the seismic The walkdown guidance provided Provide documentation of the As noted in the Finding gap/shake space walkdowns in Appendix F (Checklists and results of the seismic gap basis, inspection of the are not documented.

Walkdown Data Sheets) of EPRI walkdowns.

seismic gaps was NP-6041 includes attributes of included in the seismic seismic gaps between structures walkdowns. Piping which should be addressed in the across seismic gaps is (This F&O originated from SR performance of the walkdowns.

designed with adequate SFR-G2)

These include the clearance flexibility to between adjacent structures and accommodate building the ability of any subsystems motions, and pipe (e.g., piping, cable trays, HVAC sleeves provide ducts) spanning the gap to adequate gaps for piping Page 105 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution accommodate the differential movement. The seismic displacements.

documentation has been updated to reflect the The Seismic Walkdown Report inspections performed (PRA-BC-V-14-005) does not during the walkdowns.

include documentation of the results/findings/observations This finding has been associated with the inspection of resolved with no the seismic gaps between significant impact to the structures or the subsystems SPRA results or spanning the gap. During the conclusions.

performance of the Peer Review, SNC personal indicated that inspection of the seismic gaps was included in the seismic walkdowns, but not explicitly described in the report. The ability of components to accommodate potential differential movement at the building separations is implied in the discussion of rugged components (piping, cable trays, and HVAC ducts) in Section 2.1 (Rationale for Screening) of the report. In addition, information from the Vogtle IPEEE Report (page 3.1-37) indicated that the seismic gaps had been inspected during the IPEEE.

Page 106 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SFR-A2, 14-7 The fragility evaluation for the The determination of the Update the fragility evaluation The fragility evaluation of SFR-F4 Containment Polar Crane (in fundamental frequency of for the polar crane to address the polar crane has been fragility notebook) did not structures and components potential uncertainty in the updated to address address the impact of involves a certain degree of fundamental frequency and potential uncertainty in variation in the fundamental uncertainty. This uncertainty the contribution of higher the fundamental frequency on the applicable must be accounted for in the modes.

frequency and seismic demand.

determination of the seismic contribution of higher accelerations from the applicable modes.

in-structure response spectra (ISRS).

This finding.has been (This F&O originated from SR resolved.

SFR-A2)

Section 7.4 (Vogtle 1 and 2 Polar Crane) of the Fragility Notebook (SNC calculation no. PRA-BC-V-14-025) evaluates the polar crane as a potential seismic interaction source relative to the reactor vessel and other NSSS components inside the containment structure. In the determination of the vertical spectral acceleration applicable to the polar crane, the computed fundamental frequency falls within a valley in the applicable ISRS, on the low frequency side of the primary spectral peak.

Uncertainty in the calculated frequency, and the contribution of high modes, could result in an increase in the applied vertical acceleration. During the performance of the Peer Review, SNC personnel provided a written response indicating that it is Page 107 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description appropriate to increase the applied acceleration by 50%,

which will result in a 20%

decrease in the median capacity of the polar crane.

SFR-F3 14-8 Relay fragility calculations The relay evaluation for the Perform more realistic relay The relay fragilities have include conservative turbine driven auxiliary feedwater fragility evaluations.

been updated using the assumptions.

pump control panel in calculation appropriate response PRA-BC-V-14-008 is based on a and in-cabinet generic capacity for motor starters amplification factors, and and contactors (intended for are realistic.

(This F&O originated from SR motor control centers) and an SFR-F3) amplification factor associated This finding has been with center of door panel resolved.

response. Based on walkdown observations the relay is not mounted on the door panel so is likely on an internal bracket. The median capacity of 0.627g is well below the screening level and is not realistic.

The relay evaluations in calculation PRA-BC-V-14-009 are governed by response in the vertical direction, and the in-cabinet amplification factors used in the calculation are associated with horizontal response. The resulting median capacities of 0.762g (Appendix M1) and 1.026g (Appendix M2) are well below the screening level and are not realistic.

Page 108 of 124

VEGP Units 1 and 2 10 CFR 50.54{f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SFR-02 14-9 The seismic walkdown report The summary of the seismic Perform resolution of open The noted walkdown includes a number of open walkdowns documents a number items and provide issues have been items that are not are not of issues identified during the documentation of the evaluated and reflected traceable to a resolution performance of the walkdowns resolution associated with in the revised that required follow-up actions each of the issues, either in documentation:

(31 ). These include spatial the Fragility Notebook or the

- potential piping interaction issues, housekeeping SPRA Database.

interaction; (This F&O originated from SR issues, anchorage issues, valves

- the difference in SFR-D2) having configurations that do not inverter anchorage meet the EPRI guidelines, configuration; configuration issues, installation

- potential interaction errors, etc.

concerns with the overhead heater; this The Seismic Walkdown Report evaluation is in the (PRA-BC-V-14-005) does not fragility notebook in document how the issues section 3.4.2.

identified during the walkdowns have been addressed, either in Valve operator heights &

the field (e.g., correction of weights that were installation errors, resolution of outside EPRI guidelines housekeeping issues) or in the have been taken into fragility evaluations (e.g., valve account in the fragility configurations, anchorage analysis for these issues). During the performance components.

of the Peer Review, the Peer Review Team provided a list of The Diesel Generator the walkdown issues to SNC Exhaust Silencer was re-personnel, and SNC provided a evaluated to the as-summary of how they were operated condition.

addressed. Most issues had

~

been adequately addressed The fragility analysis for during the development of the these components has SPRA, but it was determined that been completed for the the following would require further as built condition.

effort for resolution:

Page 109 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution (a) Potential interaction between This finding has been piping and deluge valve (page 19) resolved.

- follow-up walkdowns required.

(b) Anchorage configuration on inverter (page 40) - follow-up revision to fragility evaluation required (c) Overhead heater poses potential interaction issue (page

60) - follow-up walkdown required.

(d) Valve operator heights/weights outside of EPRI guidelines (page

74) - follow-up walkdown required.

(e) Diesel Generator exhaust silencer anchor bolt nuts (page

96) - not addressed in fragility evaluation, further evaluation required.

(f) Valve operator heights outside of EPRI guidelines and potential lack of yoke support (page 105) -

these valves are part of the unfinished scope described in the Fragility Notebook, which will be completed in the future.

(g) Valve operator heights outside of EPRI guidelines (page 107) -

further evaluation required.

SFR-F3, 16-1 The model presented for peer Relay chatter is consistently being Complete the analysis and The approach to SPR-84, review did not incorporate the observed as a significant incorporate the effects of relay screening and modeling SPR-E5 effects of relay chatter as the contributor to risk profile in of seismically-induced analysis was not yet recently peer reviewed S-PRAs relay failures and chatter Page 110 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution complete.

and it is therefore realistic to chatter and similar devices in was provided to the peer expect that relay chatter is a the PRA logic model.

review team and potential significant contributor.

determined to have been During the peer review it was performed appropriately; (This F&O originated from SR discussed that the SPRA team only the incorporation SPR-84) does not believe relays will be a into the model of the significant contributors but it was impacts of relay chatter also said that this conclusion/

from unscreened relays expectation is based on was not complete. The potentially crediting operator final screening resulted actions. Thus, the effects of relay in only 2 relays being chatter per se may be significant incorporated into the (and provide some insights) while model, with one having the combination of relays and a an operator action.

number of HEP may not be.

Relay chatter fragilities and impacts have been incorporated into the seismic model, in a manner consistent with that used for other failures.

This finding has been resolved.

SPR-86 16-10 The documentation about the There is only a short sentence More detailed documentation Walkdown walkdowns in support to supporting the discussion on is suggested to support the documentation on seismic impact on HRA alternative access pathways.

conclusion on accessibility, accessibility for operator appear limited.

alternative route, availability of actions, including tools/keys, clear identification photos, has been of equipment manipulated in improved. Potential each local action.

failure of block walls has (This F&O originated from SR been reviewed and SPR-86)

Obviously, the goal of the documented. Required enhanced documentation is tools and equipment, Page 111 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description not to convince the peer such as ladders, have reviewer that the walkdowns been identified with were performed but rather to locations when needed.

ensure that the analyst is fully The documentation convinced of the conclusions.

supports the seismic HRA assumptions and Past SPRAs have shown modeling.

examples of equipment needed for the HFE that was This finding has been not in the SEL, or that has resolved with no different actuators when significant impact to the manually actuated, or that SPRA results or needed ladders that were not conclusions.

easily accessible or that were close to block walls (or under ceiling that could collapse) that were not considered an issue because the block walls were not near safety related equipment (and therefore not addressed in the rest of the SPRA work). In this perspective, a more systematic documentation of the feasibility and accessibility analysis for each of the HFE credited in the SPRA is suggested.

Page 112 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SPR-E2 16-11 Missing review of the potential It is understood that the As this exercise was A detailed quantitative for additional dependencies investigation performed in internal apparently performed for the HRA dependency introduced by the SPRA events to identify potential H FE Fire PRA (as discussed during analysis based on using models (QU-C1&2) dependency has been relied upon the peer review), it is the HRA calculator was in the Vogtle SPRA.

suggested that a review of the performed and potential for unforeseen documented. There was The SPRA logic may identify dependencies trends is no significant impact on (This F&O originated from SR additional dependencies trends performed.

results since human SPR-E2) that were not identified in the actions are not internal events.

As it is understood that the significant contributors in plan is to transition to a the Vogtle SPRA.

different dependency analysis method (based on HRA This finding has been calculator), this may be resolved.

addressed within the same transition as it is realistic to expect that not too many (if any) new dependencies would be identified.

SPR-E2 16-12 Missing documentation of the It is an industry expectation (as It is understood that the SPRA The QU report has been review of non significant discussed in NEI peer review task documentation will be revised updated to document the cutsets QU-05.

force meetings) that review of the to incorporate explicitly the review of both dominant non significant cutsets is explicitly two reviews discussed in the cutsets and non-documented.

basis for this F&O. It is also significant cutsets for recommended to document both CDF and LERF.

(This F&O originated from SR Based on discussion during the the review of cutsets following SPR-E2) peer review, two reviews were guidance from the NEI peer This finding has been performed to validate the overall review task force.

resolved with no model and cutsets. The first was significant impact to the a random review of cutsets at SPRA results or midpoints and low significance for conclusions.

each of the %Gxx initiators to verify that the cutsets are valid cutsets, and that the patterns are Page 113 of 124

VEGP Units 1 and 2 10 CFR S0.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution appropriate. That is, if one cutset is valid, then another cutset with slightly different seismic failures (or random failures) should also be nearby.

The second review, more importantly, lowered the median seismic capacity for each of the seismic initiators and some of the other seismic failures to ensure that the model would properly generate valid cutsets. For example, the LLOCA fragility was reduced to 0.5g to generate LLOCA cutsets. For ATWT, the fragility of the CRDs and RV internals were reduced to 0.5g to verify that valid ATWT cutsets were generated.

SPR-E6, 16-15 Documentation of LERF The current documentation does Expand the documentation to The LERF SPR-F2 model applicability review.

not explain what are the basis for ensure that the criteria used to documentation in the QU retaining the LERF logic and retain the LERF analysis in report was expanded to analysis unchanged within the the SPRA is explained so that describe the review of SPRA logic.

the same applicability review applicability of the (This F&O originated from SR can be performed following internal events PRA SPR-F2)

During the peer review the future potential revisions of LERF analysis to the following explanation was the LERF modeling.

seismic PRA.

provided by the SPRA team:

This finding has been "The internal events Level 2 resolved with no notebook (Chapter 9) was significant impact to the reviewed to ensure that the definition of LERF would be Page 114 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description appropriate for seismic events.

SPRA results or Section 9.2 provides the LERF conclusions.

definition, including the use of a 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> time period for release after event initiation, to allow for evacuation. This time period is considered to be valid for Vogtle seismic events, particularly due to the very low population density in the area. Other characteristics, such as bypass and scrubbing, are the same for seismic as for internal events.

The logic for the internal events LERF model is very straightforward, with sequences from the CDF model ANDed with the appropriate LERF fault tree.

This logic is also appropriate for seismic events."

SPR-88 16-18 Very small LOCA have been The DB has a specific entry for To the peer review team Additional information on screened from the analysis the incore thermocouples and knowledge Vogtle is the only the walkdown for very based on walkdowns but little provides pictures of them. Still, in-plant that has elected to small LOCA has been documentation exists of such core thermocouple tubing is not perform dedicated walkdowns added to fragility report walkdowns.

the only possible source of very in support of not modeling to provide the basis for small LOCA that is envisioned very small LOCA. This would the VSLOCA screening.

and the only documentation of be a best practice but it also addressing the other potential behooves to the SPRA team This finding has been (This F&O originated from SR sources is in section 2.3.3 of the to provide detailed resolved with no SPR-88) quantification notebook:

documentation of such significant impact to the walkdowns and how they SPRA results or "For Vogtle 1 &2, the seismic supported a systematic walkdowns inspected and Page 115 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version O - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution photographed a large sample of evaluation of the potential conclusions.

the small piping and tubing lines sources of very small LOCA.

connected to the primary system in order to identify any weaknesses. The piping was judged to be rugged."

SFR-C1, 16-2 Fragilities were not corrected The 2014 hazard was only used During the peer review the The fragilities have been SPR-E1 to reflect the 2014 hazard as input to FRANX for the final SNC staff answered a recalculated based on used for quantification. (This quantification. It is understood question on this topic by the 2014 hazard [3] and F&O originated from SR SPR-that the fragility estimates have performing an initial limited the new values E1) been performed based on the investigation of the effect on incorporated into the 2012 hazard. While it is not fragilities correction to reflect SPRA model and expected nor recommended to the 2014 hazard and quantification.

regenerate all the fragility work concluded that the effect of with the new hazard, some this scaling is not insignificant This finding has been consideration on the possible (especially for LERF). It is resolved.

change in fragility due to the use recommended to continue and of the newer hazard should be expand this investigation to made.

make the quantification fully consistent with the fragility values.

SPR-82 16-4 The effect of seismic impact There is no assessment of the While it is recognized that the The methodology used on performance shaping effect of changing the breaking industry is still developing for the seismic HRA factors is considered in the points in the Surry method. The methods in support to this analysis is based on analysis by the usage of the Surry method is based on particular topic (e.g., recently defining PSFs as a Surry method.

methods used in the past at published EPRI HRA method function of seismic SONGS and Diablo Canyon and for external events), some

• hazard level (bins),

the 0.8g breaking point was additional considerations which is consistent with developed for California should be done to understand the EPRI seismic HRA (This F&O originated from SR earthquakes. In the Vogtle the effect of HEPs in the guidance in EPRI SPR-82) analysis there is no indications on model rather than simply 3002008093 [9]. The whether the breaking point at 0.8g implementing the Surry Integrated PSFs and is also applicable to Vogtle. There method as is.

bins (breaking points)

Page 116 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment{s)

Description are also no sensitivity analyses have been updated with that would support whether a Three examples for additional breaking change in the breaking points is addressing this finding may be points and integrated significant or not.

the following:

PSFs to reflect seismic

1. Perform sensitivities on the binning applicable to values of the multipliers and Vogtle, in accordance the g levels where the with this finding and breaking point happens.

consistent with the EPRI

2. Use a different multipliers guidance. The updated method with more breaking values have been points.

applied to both internal

3. Apply the impact of seismic events HFEs and specific PSF at the individual seismic-unique HFEs PSF level (i.e., timing, stress, within the plant response etc.) in the HRA calculator.

model.

There was no significant impact on the SPRA results.

This finding has been resolved.

SPR-81, 16-5 LOCA modeling and fragility The selection of the fragility data Documentation on the use of LOCA basis has been SPR-F1 selection not clearly used for all LOCA is discussed in fragility in support to LOCA re-evaluated and documented.

Appendix B.2 of the quantification should be clarified to better updated. This was notebook but is confusing in the represent the rationale partially due to seismic mapping of selected fragilities selected and potentially fragility update and with specific failures.

addresses the modeling partially a matter of (This F&O originated from SR uncertainties associated with adding amplifying SPR-F1)

It appears that the fragility this selection.

information to the LOCA selected to represent LOCA basis. The quantification sequences are coming from While this finding is expected report includes updated specific components but then they to be addressed via documentation. Although are used to represents sort of documentation, some LOCAs are a significant Page 117of124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution surrogate events for potential additional suggestions are contributor to the SPRA failures along the piping network.

provided, such as:

results, the VEGP SCDF and SLERF are Using localized events as

1. Perform a sensitivity to sufficiently small that surrogate for pipe network failure show that the modeling further LOCA modeling is probably conservative and may approach described is not sensitivity beyond what not be fully consistent with the significantly skew the results has been provided in the system success criteria and for seismic; updated model modeling in the internal events quantification is not modeling. For example, the

2. Modify the logic by mapping warranted.

seismic-induced MLOCA fragility the seismic-induced MLOCA seems to be based on failure of to a different position in the This finding has been the pressurizer surge line, which logic (e.g., a dummy event resolved.

is a localized failure. The seismic-can be entered in the model to induced MLOCA initiator is provide a target for the mapped to the internal events FRANX injection).

MLOCA initiator. The internal events logic for MLOCA has a split fraction that divides MLOCA (and LLOCA) in four 25%

contributors impacting all four CL/HL. Since the seismic-induced MLOCA is a localized failure, the internal events logic is not fully applicable (probably slightly conservative).

Because the documentation is potentially leading to a misunderstanding of the selected approach (thus impacting ease on update), this F&O is considered a finding against the documentation SR.

Page 118 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTTF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution SPR-82 16-6 The effect of seismic impact The Vogtle SPRA elected to use Expand the IPSF approach to The methodology used on performance shaping Integrated Performance Shaping all the operator actions for the seismic HRA factors is not considered for Factors (IPSF) multipliers. While credited in the SPRA.

analysis is based on any action that was explicitly this approach was used for the defining PSFs as a added for the SPRA (e.g.,

HEPs that were carried over from function of seismic flood isolation or DG output internal events, it was hazard level (bins),

breaker closure).

systematically not done for all the which is consistent with actions explicitly added for the EPRI seismic HRA seismic.

guidance in EPRI 3002008093 [9]. The Based on discussion during the Integrated PSFs and (This F&O originated from SR peer review, the analyst believed bins (breaking points)

SPR-82) that having designed these have been updated to actions for specific scenarios reflect seismic binning following a seismic event, the applicable to Vogtle, in impact of seismic specific PSF is accordance with this already included.

finding and consistent with the EPRI The objection to this conclusion is guidance. The updated that the seismic specific PSF values have been should realistically change with applied to both internal the magnitude of the event. This events HFEs and change addresses the change in seismic-unique HFEs the overall context of the plant within the plant response when a small seismic event model.

happens as opposed to when a very large seismic event happens.

There was no significant This seems not to be captured by impact on the SPRA the approach selected for the results.

Vogtle SPRA. One example of this is that an action that has a 30 This finding has been minute Tsw (S-OA-8KR-LOCAL) resolved.

maintains an HEP of 1.60E-03 at all g levels, including the %G14 interval (i.e., >2g).

Page 119 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description It is understood that this is not expected to be quantitatively significant because failure of the recovered equipment is taken care by the logic model.

SPR-E2 16-7 Base case seismic LERF Both CDF and LERF are LERF at 1 E-11 truncation LERF truncation, which does not meet the truncation truncated at 1.0E-09 with 1000 meets the QU-B3 truncation was already considered requirements from QU-B3.

cutsets managed by ACUBE. This requirement. Rename LERF in sensitivity studies, has meets the QU-B3 requirement for at 1 E-11 as the base case for been revised CDF but not for LERF.

LERF.

appropriately to meet QU-B3. A new LERF (This F&O originated from SR truncation limit has been SPR-E2) established consistent with the LERF results.

Quantification is at 1 E-12, which is a suitably low value.

This finding has been resolved.

SPR-E2 16-8 Missing documentation of Section 3.1 is the only description While it is understood that the The QU report has been cutsets review (cfr. QU-D1) of the most important scenarios Draft. B version of the updated to document the but there is no cutset-by-cutset quantification notebook is still review of both dominant review.

somewhat a work in process, cutsets and non-it is expected that when the significant cutsets for (This F&O originated from SR model reaches a more stable both CDF and LERF.

SPR-E2) state documentation of the review of the cutsets is going This finding has been to be part of the resolved with no documentation.

significant impact to the SPRA results or conclusions.

Page 120 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description SPR-81, 16-9 Screening values used for the At the time when the An appropriate resolution of The seismic HRA SPR-84b HEPs that (at the time of the documentation was provided for this F&O is pending the analysis has been provided documentation) peer review, the most significant current evolution of the model revised to be consistent were in the most significant operator actions (i.e., flood and the importance of with the EPRI seismic cutsets.

isolation of ACCW HX) were all operator actions in the SPRA.

HRA guidance in EPRI screening values, which would Given the expectation that 3002008093 [9]. The only meet CCI for HR-G1 (directly operator actions will be original screening HEPs called through SPR-81).

needed to mitigate the have been updated (This F&O originated from SR importance of relay chatter using the HRA SPR-81)

In addition, there is little (not yet included in the SPRA Calculator, consistent documentation or supporting logic model) this F&O was with the approach used evidence to justify screening provided to ensure care is in the VEGP internal values as low as 3.00E-2 used in the generation of events PRA. The HEPs if they appear in Documentation has been important cutsets and also to updated. Operator provide more justification for response to relay chatter screening values less than has been addressed and 1.00E-1 because a low evaluated within the screening value may indeed same process, and not skew the actual importance of found to be important.

the newly generated HEP.

This finding has been resolved.

SPR-81 17-1 The documentation does not The modeling approach injected A separate section in the The discussion of specifically address the seismic fragilities into fault trees documentation that accident sequences and applicability of the internal that were modified from the specifically addresses success criteria has events accident sequences internal events PRA model. It can accident sequences and been expanded, and and success criteria to the be inferred from this approach, success criteria is needed to specific descriptions of SPRA model, and does not and it was verified by discussions collect the information in one the flooding scenarios properly document the with the staff, that the internal logical place, and is needed to has been added. This accident sequences created events sequences and success support effective peer reviews finding is documentation specifically for the SPRA criteria were considered to be and future model updates.

only and does not impact model.

applicable to the SPRA model.

Seismic PRA model This was not specifically stated in results.

Page 121 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution the documentation.

This finding has been (This F&O originated from SR Further, several additional seismic resolved with no SPR-81) flooding sequences were added significant impact to the to the fault tree. These SPRA results or sequences are not discussed conclusions.

from an accident sequence and success criteria perspective.

Inspection of the fault tree and discussions with the staff indicate that the sequences were appropriately developed with specific success criteria that is different from other internal events sequences. The development of these sequences needs to be included in the documentation. Including event trees for these sequences would also aid in a reader's understanding.

SPR-E2, 17-2 The processes used to create Examples include:

Expand the documentation to Documentation for QU SPR-F2 the presented quantification clearly explain the post-results has been im-results are not fully The top cutsets shown in table 3-processing of the results proved to describe the documented.

1 of the quantification report are generated by CAFTA and processes used to ag-produced by combining the FRANX. Examples include:

gregate results over the cutsets from all the seismic 14 hazard intervals. The interval cutsets in a process that

- Explain how the cutsets importance calculations (This F&O originated from SR is not documented.

generated by FRANX are have been re-quantified SPR-F2) combined into g-level-and the method for While the process used to obtain independent cutsets.

presentation the importance measures in documented.

section 5.2 of the quantification

- Explain the post-processing notebook is documented in that used to generate importance This finding has been Page 122 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition Number Basis Resolution ment(s)

Description section, discussions with the PRA measures, especially focusing resolved with no staff indicated that importances on the deviation from a normal significant impact to the for some of the basic events were practice that is currently only SPRA results or obtained in a different manner mentioned in the notebook.

conclusions.

(setting to one or zero and requantifying). This is not documented in the notebook.

SPR-83, 17-3 Subdividing correlation To account for similar equipment The impact of the retention of The non-minimal cutsets SPR-E4 groups based on that has different fragilities due to these non-minimal cutsets on in the peer reviewed weaker/stronger components different building locations, certain CDF/LERF and importance model were identified resulted in retention of non-correlation groups were measures should be assessed and reviewed for impact, minimal cutsets in some subdivided to assign a seismic and the results documented, and determined to be cases, which could impact capacity to a weaker component or a method to remove the non-significant to risk.

CDF/LERF results as well as that only failed that component.

non-minimal cutsets should be The results were very model importance measures.

The higher capacity was then devised. Each subdivided slightly conservative due The magnitude and assigned to both components, correlation group should be to these non-minimal acceptability of these impacts and was effectively the correlated investigated for similar effects.

cutsets. The issue has was not documented.

failure of both components. This been addressed in the can result in the retention of non-updated model, such minimal cutsets in some cases.

non-minimal cutsets no For example, for the Containment longer appear.

Fan Cooler Units there are cutsets in which, due to other This finding has been (This F&O originated from SR failures, only one containment fan resolved with no SPR-E4) cooler needs to seismically fail to significant impact to the cause core damage. Inspection SPRA results or of the cutsets shows that two conclusions.

otherwise identical cutsets are retained: one in which the 1 Fan

'group' occurs, and one in which the 4Fans group occurs. The 4Fans cutset is not minimal, and should not be included in the results. Discussions with the staff Page 123 of 124

VEGP Units 1 and 2 10 CFR 50.54(f) NTIF 2.1 Seismic PRA Submittal Version 0 - March 2017 Supporting Finding Finding Suggested Finding Require-Finding Disposition ment(s)

Number Description Basis Resolution indicated that these non minimal cutsets were noted during the quantification review process, but were thought to not greatly impact overall results. No formal assessment was done, however, and no record of the informal assessment was included in the documentation.

SPR-E6 17-4 No quantitative analysis of the A quantitative analysis is required Perform the analysis and The quantitative analysis relative contribution to LERF to meet CCII for LE-F1 & LE-G3, include the results in the

• of significant LERF plant from Plant Damage States which are directly called from quantification notebook.

damage states and and Significant LERF SPR-E6.

contributors has been contributors from Table 2-2.8-performed. A table and 9 was presented in the associated discussion of quantification results.

plant damage states and significant contributors (This F&O originated from SR has been added to the SPR-E6)

LERF QU documentation to resolve this finding.

This finding has been resolved.

Page 124 of 124