ML17213A017

Enclosure B: Beaver Valley, Unit 2, Seismic Probabilistic Risk Assessment in Response to 50.54(0 Letter with Regard to NTTF 2.1

ML17213A017
Person / Time
Site:	Beaver Valley
Issue date:	05/11/2017
From:	ABS Consulting, Rizzo Associates
To:	FirstEnergy Nuclear Operating Co, Office of Nuclear Reactor Regulation
Shared Package
ML17213A014	List: L-17-221, Enclosure a: Beaver Valley, Units 1, Seismic Probabilistic Risk Assessment in Response to 50.54(f) Letter with Regard to NTTF 2.1 ML17213A015 ML17213A017
References
2734294-R-036, Rev 0
Download: ML17213A017 (197)

Text

Enclosure B Seismic Probabilistic Risk Assessment in Response to 50.54(0 Letter with Regard to NTTF 2.1, Beaver Valley Power Station Unit No. 2 (196 pages follow)

FIRST ENERGY NUCLEAR OPERATING COMPAh[Y Seismic Prohahilistic Risk Assessment in Response to 50.54(f) Letter with Regard to NTTF 2.I Beaver Valley Power Station Unit 2

Repoffi Namel Date:

Revfuion Nu:

Reviured hy; Reviewrd hy:

Approved by:

Apprrved by; APPROYALS

$eisuric Probabilistic Risk AssEssment in Response to S0.54(Q Letter urith Regard to NTrF 2,1. Beaver velley porryer station - unit l May I l,2017 Revision 0

$r/

5-Zz-lt Rohert trrsek PRA Engineering Date 9 t7 Date Design Engineering Approved by:

{M,*a; s/stln Nathm Walker Date Supervisor, Nucloer MechanicaUstnrctr:ral Engineeriug Date

$uperrrisor, Nuclear Analytical Methods t

E*f-.\\T Moharnmed AIvi Pmject Manager Date t

Pauvlinuh Approved by:

Managcr, Design Engineering

lBSGonsulting tlRlZT.g 2734294-R-036 Revision 0 Beaver Valley Power Statiorl, Unit 2 Seismic Probabilistic Risk Assessment in Response to 50.54(f)

Letter with Regard to NTTF 2.1 May 11,2017 Prepared for:

FirstEnergy Nuclear Operating Company ABSG Consulting Inc.. 300 Commerce Drive, Suite 200. lrvine, California 92602

273429+R-036 Reaision 0 May 1-1-,201.7 Paxe 2 of745 BEAVERVALLEY POWER STATION, UNIT 2 SEISMIC PROBABILISTIC RISK ASSESSMBNT II{ RESPONSE TO 50.54(F) LETTER WITH REGART) TO NTTF 2.1 ABSG CoI*ISULTING INC. Rnronr No. 2734294-R-036 RnvrsroN 0 HIZZO Rnponr No. R12 12-4736 M.+,y llr20ll ABSG CONSULTING INC.

HTZZ,O ASSOCIATES lESGonsufting

(]Rrzzo

273429+R-036 Rwision 0 May 1.1-,201-7 Page 3 of 145 Report Name:

Date:

Revision No.:

Originator:

Reviewer:

Reviewer:

Principal:

Approver:

1. r,,"tl,{rrW Mav 11. 2017 APPROVALS Beaver Valley Power Station, Unit 2 Seismic Probabilistic Risk Assessment in Response to 50.54(f) Letter with Regard to NTTF 2.1 May LI,20l7 0

Bradley Yagla Engineering Associate RIZZO Associates Farzin Beigi, P Consultant P*IZZO Associates Donald J. Wakefield Senior Consultant ABSG f,gncrrltino Tnn Mav 11. 2017 Date Date Mav 11. 20 17 J'li^r-u"fl Date Mav ll.20l7 Nishikant R. Vaidya, Ph.D., P.E.

Vice President P-IZZO Associates Thomas R. Roche, P.E.

Vice President ABSG Consulting [nc.

Date Mav 11. 2017 Date lESGottsulting

{}Rtzzo

273429+R-036 Rruision 0 May 1,1,,2017 Pase 4 0f145 CHANGE MANAGEMENT RECORI}

Rnvrsron No.

D^+.rn DnscruprroNs oF CHANcES/AFFECTED P,tcrs 0

May ll,z0l7 Original Issue lEEGonsulting

(]R.z7.o

273429+R-036 Reaision 0 May LL,201.7 PaRe 5 of145 TABLE OF CONTENTS PAGE PURPOSE AND OBJECTIVE

.,..-..,,,7 INFORMATION PROVIDED IN THIS REPORT

.......8 BVPS.z SEISMIC HAZARD AND PLANT RESPONSE 1.0 2.0 3.0 4.0 Soil Structure Interaction (SSI) Analysis Structure Response Models..........

Seismic Structure Response Analysis Technical Adequacy...........

4,4 SSC Fnacu,rrY ANALYSIS 4.4.1 SSC Screening Approach 4.4.2 SSC Fragility Analysis Methodology......

4.4.3 SSC Fragility Results and Insights 4.4.4 Fragility Analysis Technical Adequacy 5.0 PLANT SEISMIC LOGIC MODEL 5.1 DevBr.opMENT oF THE SPRA Plehrr Sersurc Locrc Monpl 5.1.1 Seismic Initiating Event Impacts 5.1.2 Seismic Event Trees for Large Early Release 5.1.3 Relay Chatter Modeling 5.1.4 Correlation of Fragilities 5.1.5 Human Reliability Analysis 5.1.6 Seismic-Induced Floods 5.1.7 Risk Significant Flood Scenarios

5. I.8 Seismic-Induced Fires 5.2 SPRA Plnur Sprsnarc Loclc Monel TEcHutcAL ADEeUACy 5.3 Ssrsrvrrc Rrsr< QunxrmrcATroN 5.3.1 SPRA Quantification Methodology 5.3.2 SPRA Model and Quantification Assumptions 3.1 Sprsnarc Hnzano ANnr,vsrs..........

3.1.1 Seismic Hazard Analysis Methodology 3.1.2 Seismic Hazard Analysis Technical Adequacy...........

3.1.3 Seismic Hazard Analysis Results and Insights............

3.1,.4 Horizontal and Vertical FIRS DETERMINATION OF SEISMIC FRAGILITIES FOR THE SPRA...

4.1 Serspuc Equrrueur Lrsr 4.1.1 SEL Development 4.1.2 Relay Evaluation 4.2 War-roowu AppnoACH 4.2.1 Significant Walkdown Results and Insights........

4.2.2 Seismic Equipment List and Seismic Walkdowns Technical Adequacy 4.3 Dvller\\ltc ANar,vsrs oF SrRucruRns 4.3.1 Fixed-Base Analyses t2 t2 t2 28 29 35 4t 4l 4l 46 46 50 5l 5l 51 55

.51 4.3.2 4.3.3 4.3.4 59 59 s9 62 74 74 75 75 79 84 84 86 88 9l 92 92 98 98 99 99 lffiGonsulting

()Rtzzo

273429+R-036 Rwision 0 May 11,2017 Page 6 of 1.45 TABLE OF CONTENTS (coNTII\\ruED) 5.4 SCDF Rrsulrs 5.5 SLERF Resur,rs 5.6 SPRAQuerurmrcATroNUr{cnnrArNTyANar.vsls 5.6.1 Model Uncertainty........

5.6.2 Understood and Accepted Generic Uncertainties...

5.6.3 Generic Sources of Model Uncertainty 5.6.4 Plant-Specific Sources of Model Uncertainty 5.6.5 Completeness Uncertainty........

5.7 SPRA QununrrcArroN Smqsrrrvrry ANALysrs.....

5.7.1 Seismic-Related Sensitivity Cases 5.8 SPRA Locrc Mopel AND QunuurrcArroN Tncnhrrcel Anrquacv PAGE 99 6.0 7.0

8.0 CONCLUSION

S..

108 116 119 120 120 120 tzl L2l t22 128 129 130 137 REFERENCES LIST OF ACRONYMS AND ABBREVIATIONS APPENDICES:

APPENDIX A SPRA TECHNICAL ADEQUACY ASSESSMENT AND PEER REVIEW lESGonculting

(]Rtzzo

273429+R-036 Reaision 0 May 1L,201-7 7

1,45 BEA\\TER VALLEY POWER STATION, UNIT 2 SEISMIC PROBABILISTIC RISK ASSESSMENT rN RESPONSE TO 50.54(F) LETTER WrrH REGART) TO NTTF 2.1 1.0 PTJRPOSE AFID OBJECTIVE Following the accident at the Fukushima Dai-ichi nuclear power plant resulting from the March 1 1, 201 I, 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 50.54(f) letter on March 12,2012 (Reference 1), requesting information to assure that these recorrmendations are addressed by all U.S. nuclear power plants.

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 Beaver Valley Power Station, Unit 2 (BVPS-2) has been performed, in accordance with the guidance in Electris Power Research Institute (EPRI) 10252&7, "screening, Prioritization and Implementation Details (SPID) for the Resolution of Fukushima NTTF Recommendation 2.1: Seismic" (Reference2), and previously submitted to NRC (Reference 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 inthe I to l0 Hertz (Hz) range, and a seismic risk assessment is required. A seismic PRA (SPRA) has been developed to perform the seismic risk assessment for BVPS-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 SPRA developed for BVPS-2 and provides the information requested in Item (8)B of Enclosurs 1 of the 50.54(f) letter and in Section 6.8 of the SPID. 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 BVPS-2, identiffing 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 50.5a(f) letter.

This Report provides summary information regarding the SPRA as outlinedin Section 2.0.

The level of detail provided in the Report is intended to enable the NRC to understand the inputs and methods used, the evaluations performed, and the decisions made as a result of the insights gained from the BVPS-2 SPRA.

lEEGonsulting

[]Rtzzo

273429+R-036 Reaision 0 May 1.1.,201.7 Page I of 145 2.0 INFORMATION PROVIDED IN THIS REPORT The following information is requested in the 50.54(f) letter (Reference 1), Enclosure 1, "Requested Information" Section, Paragraph (8)8, for plants performing a SPRA.

1.

The list of the significant contributors to SCDF for each seismic acceleration bin, including importance measures (e.9., Risk Achievement Worilr, Fussell-Vesely (FV), and Birnbaum).

2.

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

i.

Methodologies used to quantiff 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), flnd the source of information.

iii.

Seismic fragility parameters.

iv.

Important findings from plant walkdowns and arry 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.

Assumptionsaboutcontainmentperfofinance.

3.

Description of the process 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 Paragraph I through Paragraph 6, regarding the seismic hazard evaluation reporting, also apply, but have been satisfied through the previously submitted BVPS-2 Seismic Hazard Submittal (Reference 3). Further, 50.54(0 letter Enclosure 1 Paragraph 9 requests information on the spent fuel pool. This information has been suhmitted separately (Reference 86).

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 (Reference 2) defines the principal parts of an SPRA, and the BVPS-2 SPRA has been developed and documented in accordance withthe SPID. The main elements of the SPRA performed for BVPS-2 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 tESGoneulting

(}H1z7o

273429+R-036 Raision 0 May 1L,2017 9

145 Table 2-2 provides a cross-reference between the reporting items noted in Section 6.8 of the SPID, other than those already listed in Table 2-1, and provides the location in this Report where the corresponding information is discussed.

The BVPS-2 SPRA and associated documentation has been peer reviewed against the PRA Standard in accordance with the process defined in Nuclear Energy Institute (NEI) 12-13 (Reference 5), as documented in the BVPS-2 SPRA Peer Review Report (Reference 6). The BVPS-2 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 3.0 provides information related to the BVPS-2 seismic hazard analysis.

Section 4.0 provides information related to the determination of seismic fragilities for BVPS-2 SSCs included in the seismic plant response.

Section 5.0 provides information regarding the plant seismic response model (seismic Section 6.0 summarizes the results and conclusions of the SPRA, including identified plant seismic issues and actions taken or planned.

Section 7.0 provides references.

Section 8.0 provides a list of acronyms.

Appendix A provides an assessment of SPRA Technical Adequacy for Response to NTTF 2.1 Seismic 50.54(f) Letter, including a summary of BVPS-2 SPRA peer review.

lESGonsulting rlRtzzo

273429+R-036 Reaision 0 May 1-1.,201.7 Page 1.0 of L45 TABLE 2-1 CROSS-REFERENCE FOR 50.54(F) ENCLOSURE 1 SPRA REPORTING 50.54(f) LErrsn Rnponrrxc ITEM DBscntprrox LocarroN rN THrs Rrronr I

List of the significant contributors to SCDF for each seismic acceleration bin, including importance measures Section 5.0 2

Summary of the methodologies used to estimate the SCDF and LERF Section 3.0, Section 4.0, and Section 5.0 2i Methodologies used to quantiff the seismic fragilities of SSCs, together with key assumptions Section 4.0

^t tt SSC fragility values with reference to the method of seismic qualification, the dominant failure mode(s), and the source of information Table 5-9 provides fragilities (Am and beta) and failure mode information, ffid method of determining fragilities for the top risk significant SSCs based on standard importance measures such as Fussell-Vesely (F-V).

Seismic qualification reference is not provided as it is not relevant to development of SPRA ziii Seismic fragility parameters Table 5-9 provides fragilities (Am and beta) information for the top risk significant SSCs based on standard importance measures such as F.V.

Ziv Important furdings from plant walkdowns and any corrective actions taken Section 4.2 address walkdowns and walkdown insishts 2v Process used in the seismic plant response analysis and quantifi cation, including specific adaptations made in the intemal events PRA model to produce the seismic PRA model and their motivation Section 5.I and Section 5.2 provide this information 2vi Assumptions about containment performance Section 4.3 and Section 5.5 address containment and related SSC performance 1

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

Identified plant-specific vulnerabilities and actions that are planned or taken Section 6.0 addresses this ABSGonsulting

()Rtzzo

273429+R-036 Raision 0 May 11,2017 Page LL of 1-45 TABLE 2.2 CROSS.REFERENCE FOR ADDITIONAL SPII} SECTION 6.8 SPRA REPORTING Note:

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

lESGonculting tlRrzzo SPID Secrrou 6.8 Irru (r) DnscruprroN LoCATIoN IN THIs RnTonT A Report should be submiued to the NRC summarizing the SPRA inputs, methods, and results.

Entirety of the submiual addresses th[s.

The level of detail needed in the submittal should be sufficient to enable NRC to understand and determine the validity of all input data and calculation models used.

Entirety of the submittal addresses this. The template attempts to identiff key methods of analysis and referenced codes and standards.

The level of detail needed in the submittal should be sufficient to assess the sensitivity of the results to all key aspects of the analysis.

Entirety of the submiual addresses this. Results sensitivities are discussed in the following sections:

t Section 5.7(SPRA Model Sensitivities) t Section 4.4 Fraeility Screening (Sensitivitv).

The level of detail needed in the submittal should be sufficient to make necessary regulatory decisions as a part of NTTF Phase 2 activities.

Entirety of the submittal template addresses this.

It is not necessary to submit all of the SPRA documentation for such an NRC review. Relevant documentation should be cited in the submittal, and be available for NRC review in easily retrievable form.

Entire Report addresses this. This Report summarizes important information from the SPRA, with detailed information in lower tier documentation.

Documentation criteria for a SPRA are identified throughout the ASME/ANS Standard (Reference a). Utilities are expected to retain that documentation consistent with the Standard.

This is an expectation relative to documentation of the SPRA that the utility retains to support application of the SPRA to risk-informed plant decision-making.

273429+R-036 Reoision 0 May 1L,201.7 Page L2 of 1,45 3.0 BYPS-2 SEISMIC IJ.AZARI} AND PLAFIT RESPONSE The BVPS is a soil site located in Shippingport Borough on the south bank of the Ohio River in Beaver County, Pennsylvania, in the Appalachian Plateau Province. The bedrock in the area is the Allegheny formation of Pennsylvanian age consisting of shale and sandstone with several interbedded coal seams. The bedrock is overlain by about 100 feet (ft) of alluvial granular terraces that formed during the Pleistocene. Plant grade is elevation (EL) 735 ft and the top of bedrock is at approximate EL 625 ft, Subsequent to the March 2014 submittal, the BVPS seismic hazard for hard-rock site conditions was updated to address SPRA peer review comments; this updated is summarized in Section 3.7.7. The derivation of Foundation Input Response Spectra (FIRS) is completed for several elevations corresponding to the base of the critical structures located at the BVPS Site. The site response geotechnical model used to derive the FIRS is described in Section 3.1.7.2, with site response analysis results described in Section 3.7,7.3. The seismic haeard results used for the SPRA are described in Section 3.7.3, while the derivation of horizontal and vertical FIRS are described in.Tection 3. 7.4.

3.1 Seismic Hazard Analysis This section discusses the seismic hazard methodology, presents the final hard-rock seismic hazard results used in the SPRA, the site geotechnical model used to derive the FIRS, the site response analysis results, 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 a:rive at the site seismic hazard. More detailed information regarding the BVPS Site Probabilistic Seismic Hazard Analysis (PSHA) hazard was provided to NRC in the seismic hazard information submiued to NRC in response to the NTTF 2.1 Seismic information request (Reference 3) and can be found in Reference 23.

3.1.1 Seismic Hazard Analysis Methodolory For the BVPS-2 SPRA, the quantification of the seismic hazard utilizes RIZZO's in-house software,P*IZZO-HAZARD (Reference 19). This software uses the characterization of seismic sources (NRC, 2012b) and ground motion models (GMM) (EPRI 2013a, referred to as the EPRI GMM update) to estimate the annual exceedance frequencies for various levels of pseudo-Sa at different spectral frequencies.

lESGonsulting tlRtzz'o

273429+R-036 Reaision 0 May L1.,2017 Page 1,3 of 1.a5 The final PSHA results reflect the resolution of SPRA peer review interactions as documented in peer review Facts and Observations (F&O). The specific resolution suilrmaries are provided in Appendix A. The final PSHA and supporting documentation includes the following elements addressing the peer review F&Os:

Enhanced discussion of the potential for induced or higgered earthquakes and the impact of these earthquakes on the seismic hazard for the BVPS Site.

r Quantitative assessment of seismicrty that has occurred since the end of 2008, the cut-off date for the earthquake catalog used to assess earthquake recrurence rates and maximum magnitudes (NRC, 2012b).

. Modifications to the scripts used to combine seismic hazard curves for hard-rock site conditions and updating the hard-rock mean and fractile hazard curves. This resulted in essentially no change to the mean haeard, and only minor changes to fractile hazard curves on which the SPRA is based.

o Enhanced assessment of site response amplification factor epistemic uncertainty to define the input for developing the soil hazard curves. Based on this assessment the soil hazard curves (mean and fractiles) were derived and used to develop FIRS at each foundation elevation.

Assessment of the variance contribution to the total variance for each of the seismic hazard input parameters. This assessment quantifies which seismic hazard input parameter(s) dominates the epistemic uncertainty in seismic hazaord for several mean annual frequencies of exseedance.

r Updating the approach used to assess vertical-to-horizontal ground motion ratios resulting in some reduction in the vertical ground motions at each foundation elevation on which the SPRA is based.

3.1.1.1 Hard-Rock PSHA Results The hard-rock PSHA hazard curves at the BVPS Site are obtained for seven response spectral frequencies (100 Hz [equivalentto PGA],25H2,10 Hz, 5Hz,2.5Hz,l Hz, and0.5 Hz). In addition to the mean, the associated fractile (5 percent, l5 percent, 50 percent (median),

85 percent, and 95 percent) hazard curves are also obtained. Figure 3-1 and Table 3-1 present the PGA hard-rock hazard curves; the fulI set of hazard curves at the seven spectral frequencies associated with the hard-rock Ground Motion Model (EPRI 2013a) can be found in Reference 23.

lESGonsulting rlRtz71o

2734294.R-036 Rwision 0 May 11,,20:17 Page M of 745 Nr 5 \\

5\\

\\

I\\

t ii iiliil a

iltl s

\\

.r

\\

q

\\ \\

\\

fi

,a

\\

\\

)\\s I \\

N

\\

l.

I

\\

\\\\

\\

\\.,

ilil ittt

\\

a\\,

\\

\\\\

\\\\

1\\

N

\\)\\\\\\.

l

\\\\ \\

I

\\

\\

IIiltt itit ttri

\\\\\\\\

ir tt!. l

\\

I tI C'Itta ExH "o

F Er E

raaecE ee a,

=

lE-2 lE-3 lE-4 lE-5 lE-6 lE-7 100 Hz

-MCm

.{-5&

or-1.ro 15&

s0&

rf-85&

95fr lE-8 0.01 0.1 I

l0 Accclcrrflotr (d) x'rGuRE 3-1 l00lil.Z Sr MEAI\\ AI\\ID FRACTILE HAZARD CURVES AT TIIE BVPS SITE X'OR HARD.ROCK SITE COI\\DITIONS The events contolling the hard-rock hazard provide the basis to develop smoothed UHRS at hard rock based on the predicted hazafi at the seven spectral frequencies.

These conholling events are obtained by deaggregating the rock hazard for 1E-4, 1E-5, and 1E-6 mean annual frequency of exceedance (MAFE) into magnitude and distance bins following recommendations in Reference 24. T\\e deaggregation results are used to identiff contolling eartlquakes at each MAFE.

rltffi I}HffiT.Q

273429+R-036 Reuision 0 May L1.,201,7 Page 15 of L45 TABLE 3.I 100 HZ Sr MEAN ANI) FRACTILE HAZARI) CURVES AT THE BVPS SITE FOR HARI}.ROCK SITE CONDITIONS Gnouun MOTION LEVEL (g)

Auuunl PRonanrlrry oF ExcEEDAT{cE Mrnn lYo Fnlcrrln 160/0 Fn+,crrr,r s0%

Fnacrrr,r 84o/"

FntcrrlB 95o/"

Fnacrrln 0.01 2.968-03 9.06E-04 1.46E-03 2.438-03 3.93E-03

8. 13E-03 0.02 I.14E-03

3. 16E-04 4.65E-04 8.56E-04 1.60E-03 3.80E-03 0.03 6.38E-04 1.44E-04 2.22E-04 4.50E-04 9.01E-04 2.18E-03 0.04 4.19E-04 7.85E-05 1.31E-04 2.90E-04 6.14E-04 t.478-03 0.05 3.02E-04 s,l0E-05 8.13E-0s 1.9sE-04 4.60E-04 1.06E-03 0.06 2.31E-04 3.778-05 5.91E-05 1.44E-04 3.55E-04 8.00E-04 0.07 1.84E-04 2.71E-05 4.42E-05 1. 16E-04 2.93E-04 6.36E-04 0.08 1.50E-04 2.14E-05 3.74E-05 9.04E-0s 2.528-04 5.29F-04 0.09 1.268-04 1.728-05 3.01E-05 7.53E-05 1.99E-04 4.45E-04 0.1 1.07E-04 1.43E-05 2.55E-05 6.56E-05 1.71E-04 3.678-04 0.2 3.59E-05 4.46E-06 8.1sE-06 2.228-05 5.86E-05 1.17E-04 0.2s 2.478-05 2.94F-06 5.57E-06 1.55E-05 4.02E-05 7.57E-05 0.3 1.80E-05 2.09E-06 4.20E-06 I.13E-05 3.02E-0s s.62E-0s 0.4 1.06E-05 1.20E-06 2.39E-06 6.63E-06 1.79E-05 3.42E-0s 0.5 6.90E-06 7.07E-07 1.51E-06 4.428-06 1.18E-05 2.15E-05 0.6 4.768-06 4.51E-07 9.51E-07 2.948-06 8.04E-06 1.s2E-05 0.t 3.43E-06 2.958-07 6.748-07 2.09E-06 5.85E-06 1.09E-05 0.8 2.ssE-06 2.04E.07 4.748-07 1.53E-06 4.40E-06 8.12E-06 0.9 1.9sE-06 t.468-07 3.66E-07 1. 16E-06 3.40E-06 6.32E-06 I

r.s2E-06 1.05E-07 2.638-07 8.72E-07 2.58E-06 4.928-06 2

2.398-07 8.39E-09 2.50E-08 I. 13E-07 4.11E-07 8.88E-07 J

6.74E-08 1.34E-09 4.738-09 2.58E-08 1.128-07 2.748-07 5

1. 10E-08 7.8 lE-I I 3.63E-t 0 2.98E-09 1.64E-08 4.94E-08 6

5.46E-09 2.61E-11 1.23E-10 1.22E.09 7.628-09 2.54E-08 7

2.958-09 9.07E-12 4.81E-l l s.46E-10 4.02E-09 I.41E-08 I

1.70E-09 3.89E-12 2.09E-11 2.728-t0 2.18E-09 8.16E-09 9

1.04E-09 1.58E-12 9.49E-12 1.43E-10 1.268-09 4.94E-09 10 6.60E-10 7.218-13 4.678-t2 7.41E-tt 7.52E-10 3.2sE-09 lEEGonsulting tlRtzzo

273429+R-036 Reuision 0 May 1.1.,201.7 Page L6 of 1.45 Because there is a significant contribution to hazard at low frequencies from distant earthquakes, the mean magnitude and distance are identified for the overall hazard and broken down by distance less than and greater than 100 km. Tahle 3-2 identifies the controlling events in terms of the respective mean magnitude and distance for each of the distance bands. For the case in which contribution to hazard is examined separately for distance less than and greater than 100 km, the weight providedin Table 3-2 represents the relative contribution to hazard from each distance range.

TABLE 3.2 CONTROLLING EARTHQUAKES FOR THE BVPS SITE Note:

"Weighf is the percent contribution to overall hazard for the given distance range Response spectral shapes for the controlling earthquakes are determined, following recommendations in Reference 77 for Central and Eastern United States (CEUS) earthquakes.

Equally weighted single-and double-corner spectral shapes from ReferenceTT are scaled up to the UHRS to define the controlling earthquake response spectra. Final hard-rock smoothed UHRS are determined by using the controlling earthquake spectral shape to interpolate and extrapolate the UHRS at response spectral frequencies other than those for which the GMM provides values.

lESGonsulting tlRtzzo H.lz,lnn CoNrnollrNc EARTHeUAKE OvnRLr,r, Hlzlnn R>0km HlzeRu Fnopr R< 100 km Hlznnn Fnopr R> 100 km IVlncurtupn (M)

DIsr^tr,{cn (km)

Mecmrrurn (M)

Drsrilccr ftm)

Wrrcur Mlcurruon (M)

Drsurucr (km)

\\ilBrcur IE.4 MAFE 7.4 549 6.3 32 0.0941 7.5 tJt 0.906 0.5 Hz 1E-4 MAFE 6.6 r39 5.9 3l 0.415 7.1, 399 0.585 1.0 Hz - 2.5 Hz 1E.4 MAFE 5.9 46 5.7 3r 0.777 6.4 t76 0.223 5.0 Hz - 10.0 Hz 1E-4 MAFE 5.8 41 5.7 30 0.829 6.3 168 0.171 25 Hz 1E.5 MAFE 7.3 292 6.5

)1 0.252 7,6 651 0.748 0.5 Hz IE-5 MAFE 6.4 43 6.1 2l 0.734 7.2 337 0.266 1.0 Hz -2.5H2 I E.5 MAFE 5.9 t7 5.8 l6 0.967 6.9 163 0.0331 5.0 Hz - 10.0 Hz IE-5 MAFE 5.8 l5 5.8 l4 0.978 6.9 160 0.0221 25 Hz IE-6 MAFE 7.0 70 6.7 23 4.623 7.5 452 0.177 0.5 Hz 1E-6 MAFE 6.4 l8 6.4 l5 0.936 7.3 216 0.053s 1.0 Hz -2.5H2 IE.6 MAFE 6.1 ll 6.0 u

0.994 7.4 157 0.0061 5.0 Hz - 10.0 Hz IE-6 MAFE 6.0 l0 6.0 l0 0.996 7.4 155 0.00394 25 Hz

273429+R-036 Rsoision 0 May 1.1,,20L7 Page 17 of 145 For response frequencies less than 0.5 Hz,the controlling earthquake response spectrum for distances greater than 100 km is used. Similarly, for response frequencies between 0.5 Hz and 2.5 Hz, 2.5 Hz and I 0 Hz, and greater than I 0 lfz the contolling earthquake response specta for 1.75-Hz Se hazard, 7.5-Hz Sl hazard, ard25-Hz Sa hazard are used, respectively. The smoothed UHRS is derived for 36 spectal frequencies, which meets the minimum number of structural frequencies defined in Reference 24. Figure 3-2 shows the smoothed UHRS with a IE.4 MAFE.

FIGT]RE 3.2 BYPS SITE 1E.4 MAF'E SMOOTHED UNIFORM HAZARI)

RESPONSE SPECTRA AT HARD-ROCK 3.1.1.2 Site Response Analysis Geotechnical Model The BVPS Site is located in the Ohio River Valley, a flat-bottomed, steep-walled valley constructed by erosion. Bedrock underlying the BVPS Site and forming the hills, which rise to an elevation of about 1,100 ft adjacent to the BVPS Site to the north and south of the Ohio River Valley, is characterizedby sandstones and shales interbedded with several thin coal seams and occasional thin limestone beds of the Pennsylvanian age Allegheny formation.

The terrace material at the BVPS Site, overlying bedrock, is characteizedby three levels; high, intermediate, and low. The ground surface of the high terrace ranges between elevations (EL) 740 ftto EL 730 ft. The high terrace is composed of granular material mostly gravel and sand with some cobbles and rock fragments. The intermediate terrace ground surface elevation is approximately at EL 700 ft to EL 685 ft. The intermediate terrace is the result of flood contol projects, which lowered the river level during the 1930s. The upper soils of the intermediate terrace consist of medium clays, which extend to about EL 660 ft. The low terrace being the most recent and closest to the river is located atazone having a ground surface EL 675 ft fESConsultlng

()Rrzzo t

-.r'

-smooth UHRS 1e-4 0.25 o,2 a

a0 Y

.E o.1s

• ,gg n1 O

lJrI Ct rJ 0.05 0

r10 Frequency [Hz]

100 0.1

273429+R-036 Reaision 0 May 11,2017 Page L8 of 1-45 towards the north. The shallow soils consist of soft clay and silt sediments of river showing some organic content.

The plant structures are located upon the high terrace of alluvial gravels. The nominal station grade is EL 735 ft. The ground surface grade elevation for the shared BVPS-I and BVPS-2 Intake Structure is EL 675 ft.

The site response analysis is completed for several elevations coffesponding to the bases of the critical structures located at the BVPS Site. Representative foundation elevations are selected for site response analyses considering that: l) foundation elevation varies for some plant structures and?) some plant structures are founded at similar elevations. Therefore, elevations forwhich site response analyses are performed may not coincide exactly with foundation elevations but are within a few feet. The approximations in elevation have a negligible effect on the structural response. These structures and representative foundation elevations are:

. EL 681: BVPS-2 Reactor Containment Building (RCBX)

. EL 723: BVPS-2 Fuel Handling/Decontamination Buildings (FULB) and Service Building (SRVB) r EL 713: BVPS-2 Diesel Generator Building (DGBX), Main Steam Cable Vault (MSCV) and Safeguards Building (SFGB)

. EL 703: BVPS-2 Control Building (CNTB) and Auxiliary Building (AXLB) r EL 637: BVPS-1/BVPS-2 shared Intake Structure (INTS)

The quantification of site amplification of hard-rock motions takes into account the site-specific shear-wave velocity profile and other relevant dynamic properties for the site geologic material.

These are based on available licensing documents and other relevant studies. Aleatory and epistemic uncertainties in the quantification of site amplification are explicitly addressed by defining alternative shear-wave velocity profiles, alternative shear modulus reduction and damping characteristics of the geologic materials, site attenuation (kappa), and the inherent random variation in these parameters.

Two conditions influence the site amplification factors (AF) for the BVPS Site: there is about 1 5 ft of compacted structural backfill surrounding several of the buildings and there is a significant Vs contrast between the soil materials at the site (compacted structural backfill and the terrace deposits) relative to the underlying sedimentary rock. Because of these two conditions, the calculation of the AFs at the various building elevations account for the potential influence of the soil confinement that surrounds the building. Guidance provided by Reference 25 accounts for these conditions.

The site response analysis for most of the structures at the BVPS Site is based on the fulI soil column extending from hard rock to plant grade (EL 73 5). The full set of strain iterated properties are retained for each of the layers modeled. The geologic column is then truncated at the appropriate building elevations and the site response analysis is repeated using the strain iterated properties from the full column, with no further strain iteration permitted. Because the lESGotrsulting

(}Rtzzo

273429+R-036 Reaision 0 May 1L,201.7 Page 19 of745 soil column for the BVPS Site INTS is different, a second soil profile is developed for that structure, ffid the process outlined above is repeated.

The methodology described in Reference 2 guides the site response analysis. A logic tree is used to assess the epistemic uncertainties in site response input parameters, which includes the following:

. Hard-rock input ground motions are developed for two seismic source models with equal weights. The seismic source model is based on the point source model and uses both single-corner and double-corner input assumptions (Tables B-4 and 8-6 of Reference?).

. Use of three alternative hase-case velocity profiles (BE [P1], LR [P2], and UR [P3]) to represent the shear-wave structure of materials underlying the Site.

r For each base-case profile, use of two scenarios to represent potential strain degradation of material properties of the Paleozoic rocks: materials behave nonlinearly in the top 500 ft of rock and linearly below the top 500 ft of rock to the profile base, and materials behave linearly for the whole profile.

The site parameter kappa describes the damping considered in the site response analysis. In the context of Reference 2, kappa is the profile damping contributed hy both intrinsic hysteretic damping, as well as scaffering due to wave propagation in heterogeneous material. The total site kappa consists of the kappa associated with the near-surface profile and kappa for the half-space (i.e., reference rock). The contribution to kappa from the half-space is taken as 0.006 seconds (s), consistent with the GMM. Both the hysteretic intrinsic damping and the scattering damping within the near-surface profile and apart from the crust are assumed frequency independent.

Based on review of available geotechnical data three base-case profiles were developed. The specified Vs profiles were taken as the mean or BE base-case profile (Pl) with LR and UR base-case profiles P2 and P3, respectively. Consistent with the guidance from EPzu (ReferenceZ),the URbase-case profile is constrainedto not exceed Vs of 9,200 fl/s. The BE profile is given aweight of 0.4 while the LR and URprofiles are each givenaweight of 0.3.

This is consistent with the guidance from Reference 2 where the weights are based on a 3-point approxirnation for a normal distribution reflecting the 10th and 90th percentile.

All three base-case profiles extend to a depth of 4,435 ft below the base of the ground surface at the BVPS Site. This depth is taken as the boundary where hard-rock site conditions exist. The basis for this selection considered guidance from Reference 2 which indicates that a sufficient depth should be selected such that hard-rock Vs is reached or the depth is greater than the criteria for no influence on response for spectral frequencies greater than 0.5 Hz. The base-case profiles (P1, P2, and P3) are shown on Figure 3-3 and listedin Table 3-3, and represent the Vs profiles used for the site response analysis for all structures except the INTS.

To account for random variations in Vs beneath structure footprints, 30 randomized Vs profiles are generated utilizing the stochastic model developed from Reference 78. The range of Vs lSSGonsuEing

[iRtTzo

2734294-R-036 Reoision 0 May 1,1,,2017 Page 20 of 145 values for each of the geologic layers was reviewed to ensure that the Vs values modeled are realistic for the types of soils and Paleozoic rocks at the BVPS Site.

0 3000 1'r (fl/sec)

{000 6000 8000 10000 n000 0

s00 1{tl I r lll,rct; lHrt

'rfr i r{l.r lr*,n)

I 000 l 500 3000

500 3000 3500

{000

{500 5000 eftr{r?,lt S}

ffi{}sr*'+l*

fli

-Prrfr*r Fl F\\

J,t-v t

e&$

bt fi, t{fi}

sa--CL e)A-rx0 Profile P2 Profile F3

-Profile Pt FIGURE 3.3 BASE-CASE Vs PROF'ILES, BVPS SITE Consistent with the guidance from Reference2, uncertainty and variability in material dynamic properties are included in the site response analysis. The soils at the BVPS are generally represented as sand and gravel so both the EPRI soil and Peninsular Range curves from Reference 2 arc appropriate. Consideration was also given to the use of dynamic property curves developed for the proposed Bell Bend Nuclear Power Plant (NPP) site on the Susquehanna River in east-central Pennsylvania, which are also appropriate for sand and gravel; the Bell Bend curves are similar to the EPRI soil curves but allow for more significant non-linear site response as represented by higher shear modulus reduction. In summary, the Bell Bend dynamic property curves are associated with the most non-linear behavior (as expressed by the shear modulus reduction versus shear strain curves) while the Peninsular Range dynamic property curves are associated with the least non-linear behavior. Given this observation the selection of the soil dynamic property curves was directly linked to the stiffness (Vs) of each soil profile.

lBSGonsuEing

()Rrzzo

2734294-R-436 Rasision 0 May 1L,20L7 Page 21 of 1,45 Lavnn Elnvarrou (ft)

Pnorn r Pl Pnorrr,n P2 Pnorrr,r P3 Vs(fUs)

DnprH (fo Vs (fUs)

Dnprn (f0 Vs (ftls)

DBpru (f0 735 730 0

635 0

840 0

720 730 t5 635 15 840 l5 720 1,015 l5 883 l5 1,167 l5 681 1,015 54 883 54 1,167 54 681 1,100 54 957 54 1,265 54 665 1,100 70 957 70 1,265 70 665 1,200 70 1,043 70 1,380 70 625 1,200 ll0 1,043 ll0 1,380 110 625 5,000 ll0 4,348 ll0 5,750 110 550 5,000 185 4,348 185 5,750 l8s 550 6,026 185 5,240 185 6,930 185 3s0 6,026 385 5,240 385 6,930 385 350 6,744 385 5,864 385 7,756 385 300 6,744 43s 5,864 435

'1,756 435 300 6,744 43s 5,864 435 7,',l56 435

-120 6,744 8ss 5,864 855 7,756 855

-120 7,112 855 6,194 855 8,179 855

-2994 7,112 3,729 6,184 3,729 8,179 3,729

-2994 6,416 3,729 5,579 3,729 7,378 3,729

-3700 6,416 4,435 5,579 4,435 7,378 4,435 TABLE 3-3 BASE-CASE Vs PROFILES, BVPS SITE For the rock material, uncertainty is represented by modeling the material as either linear or non-linear in its dynamic behavior over the top 500 ft of rock. This material primarily consists of shale and sandstone. The use of the EPRI rock curves from Reference 2, which exhibit a relatively high amount of low-strain damping (*3.2 percent), is limited to the upper 100 ft where the rock is considered as weathered and fractured. For the alternative linear analyses, the low-strain damping from the EPRI rock curves was used as the constant value of damping in the upper 100 ft.

Within the depth range of 100 ft to 500 ft, non-linear dynamic behavior is based on the unweathered shale dynamic properties from Reference 75 for the Y-12 Site at Oak Ridge, Tennessee. For these curves the low-strain damping is about I percent. For the alternative linear analyses, the low-strain damping from the Reference 75 unweathered shale curves were used as the constant damping value from 100 ft to 500 ft. Below a depth of 500 ft, linear material behavior is adopted, with the damping value specified consistent with the kappa estimate for the Site.

Near-surface site damping is described in terms of the parameter kappa. For the BVPS site, kappa was estimated following the guidance in Reference 2 using the approach for cases where the thickness of the sedimentary rock overlying hard-rock is greater than 3,000 ft. There is lESGonsultittg

()Rtzz,o

273429+R-036 Reaision 0 May'1,'1,,201.7 Page 22 of145 confidence, based on deep well sonic log data from the vicinity of the Site, that the hard-rock horizon is more than 4,000 ft below the top of rock. For each Vs profile, kappa was estimated using the equations from Reference 2 for the kappa contribution from the soil and the kappa contribution from the entire bedrock section. The kappa contribution for the Paleozoic rock section is defined as the bedrock kappa minus the kappa contribution from hard-rock (.006s).

The site kappa is used to establish the damping for the Paleozoic rock material below a depth of 500 ft. This is accomplished by using the low-strain damping and the Vs profiles to determine the remaining kappa contribution from the rock layers below a depth of 500 ft within the rock.

Given the remaining kappa contribution for the deep rock layers and the Vs for those layers, the damping for these layers can be defined. The site response analysis is then completed assuming linear behavior for these deeper rock layers with appropriate low-strain damping values.

Using the kappa values obtained for the three velocity profiles and including a kappa of 0.006s for the underlying hard-rock the total site kappa is estimated to be 0.0167s for profile P1, 0.0191 s for profile P2, and 0.0146s for profile P3. To complete the representation of uncertainty in kappa a 50 percent variation to the base-case kappa estimates was added for profiles P2 and P3. For profile P2, the softest profile, the base-case kappa estimate of 0.0191s was augmented with 50 percent increase in kappa to a value of 0.0286s, resulting in two sets of analyses for profile P2. Similarly uncertainty in kappa for profile P3, the stiffest profile, was augmented with a 50 percent reduction in kappa, resulting in kappa values of 0.0146s and 0.0097s. The suite of kappa estimates and associated weights is listedin Table 3-4.

Consistent with the guidance in Reference2, input For:rier amplitude spectra were defined for a single representative earthquake magnitude (M 6.5) using two different models for the shape of the seismic source spectrum (single-corner and double-corner).

TABLE 3-4 KAPPA VALUES AND WEIGHTS USED IN SITE RESPONSE ANALYSIS Vnr.ocrrY PR0FILE PRon'rr.r, WBrcnr Karra (s)

Karpl Wucur PI Base-Case 0.4 0.0 r 67 1.0 P2 Lower Range 0.3 0.01 9l 0.6 0.0286 0.4 P3 Upper Range 0.3 0.0146 0.6 0.0097 0.4 Parallel to the deviation of site response inputs for the power block area, site response inputs were also derived for the shared INTS. Epistemic uncertainty in Vs is modeled using three base-case profiles, the mean or BE base-case profile (P1) with LR and UR base-case profiles P2 and P3, respectively. Uncertainty and variability in material dynamic properties for the Pleistocene Terrace deposits are included in the site response analysis. The kappa for each of the base-case profiles uses the same Vs, layer thickness, and damping for the deeper geologic units, and adds above them the Pleistocene Terrace layers and their respective Vs and thickness values.

lESGoneultlng tlRtzzo

273429+R-036 Reaision 0 May 1.1.,201,7 145 Also, consistent with the site response analysis for the deeper geologic layers, equivalent-linear and linear damping represents the epistemic uncertainty in dynamic properties.

3.1.1.3 Site Response Analysis Results The site response analysis uses an equivalent-linear method that is implemented using the Random Vibration Theory (RVT) approach. This approach utilizes a simple, efficient method for computing site-specific amplification functions and is consistent with Reference24 and Reference 2. The input motion is applied at the top of the half-space as outcrop motion. The free-field peak responses at the top of any sub-layers are solved by using the RVT technique.

The nonlinearity of the shear modulus and damping is accounted for by the use of equivalent-linear soil properties and an iterative procedure to obtain values for modulus and damping compatible with the effective shear strains in each layer.

Most major structures at the BVPS Site are founded in the Pleistocene Terrace deposits at foundation elevations of approximately 681 ft for the RCBX, 703 ft for the AXLB and CNTB, and 713 ft for the MSCV, DGBX and SFGB. There are a few structures founded in the compacted granular structural backfill at approximate foundation EL 723 for the FULB and SRVB. The site response analysis for the BVPS-I and BVPS-2 shared INTS has a different site profile than for the other structures. The approximate foundation elevation for the INTS is 637 ft while the top of the fulI soil column is at EL 675.

The seismic structural analysis will treat all of these structures as surface founded at the foundation levels ignoring the effects of embedment. The approach to developing FIRS for each elevation is based on the guidance provided by Reference 25. Each FIRS is provided as the Truncated Soil Column Response (TSCR). After the strain-compatible soil profiles are developed for the full soil column, the soil layers corresponding to the embedment depth of the structure are removed and a second round of soil column analysis is performed with the truncated soil columns with no further iteration on soil properties. The free surface outcrop motions from the second round truncated soil column analysis correspond to the required TSCR.

The results of the site response analysis consist of AFs that describe the amplification (or de-amplification) of reference hard-rock response spectra (5-percent-damped pseudo-absolute acceleration) as a function of frequency and input reference hard-rock PGA amplitude. AFs are determined for the appropriate control point elevation. Because of uncertainty and variability incorporated in the site response analysis, a distribution of AFs is produced. The AFs are represented by a median (i.e., ln-mean) amplification value and an associated log standard deviation (sigma ln) for each spectral frequency and input rock amplitude. Consistent with Reference 2, median total amplification was constrained to not fall below 0.5 to avoid extreme de-amplification that may reflect limitations of the methodology.

Table 3-5 provides the median site AFs and standard deviation of the logarithm of site AFs (oln(tr)) forthe spectral frequencies of 0.5 Hz, 7H2,2.5H2,5H2,10 H2,25 Hz, and 100 Hz (PGA) for BVPS Site EL 681. Figure 3-4 and Figure 3-5 show the median site AFs and otn(cr) versus Sa for each of the spectral frequencies. The complete set of site response results can be found in Reference 23.

lESGonsulting

()Rtzz.o 23

TABLE 3.5 AMPLIFICATION F'UNCTIONS FOR BVPS SITE AT EL 68I 100 Hz sA [sl MnnlaN AF Srcru.q, Ln(AF) 25Hlz Sr [g]

Mrnranl AF Srcnm Ln(AF) 10 Hz Sr [sI Mrnmn AF Srcnm Ln(AF) 5Hz sA [el Mrumn AF Srcnan Ln(AF) 9.59E-03 2.50E+00 1.10E-01 1.248-02 2.20E+00 9.59E-02 1.94E-02 2.00E+00 1.59E-01 2.21E-02 3.73E+00 2.54E-01 5.13E-02 2.13E+00 9.748-02 1.008-01 1.68E+00 1.60E-01 1.06E-01 1.87E+00 1.92E-01 8.85E-02 3.71E+00 2.44E.01 1.08E-0 r 1.79E+00 9.798-02 2.12F.01 1.42E+00 1.73E-01 1.988-01 1.82E+00 2.04E-01 1.54E-01 3.64E+00 2.36E-01 2.37E-01 1.47E+00 9.58E-02 4.45E-01 1.21E+00 1.75E-01 3.81E-01 1.80E+00 2.06E-01 2.84E-01 3.43E+00 2.35E-01 3.73E-01 1.28E+00 8.95E-02 6.77E.01 1.09E+00 1.75E-01 s.s8E-01 1.78E+00 2.03E-01 4.09E-01 3.24P+00 2.37E-01 5.15E-01 1.16E+00 8.66E-02 9.13E-01 9.98E-01 1.77E-01 7.37E-01 1.75E+00 1.94E-01 s.35E-01 3.08E+00 2.42E-01 6.61E-01 1.08E+00 8.68E-02 l.l5E+00 9.30E-01 1.80E-01 9.17E-01 1.72E+00 1.89E-01 6.61E-01 2.94E+00 2.51E-01 1.03E+00 9.42E.01 9.26E-02 1.75E+00 7.95E-01 1.90E-01 1.37E+00 1.61E+00 1.84E-01 9.748-01 2.60E+00 2.72E-01 1.42E+00 8.48E-0t 9.t8E.02 2.38E+00 6.98E-01 2.02E.01 1.84E+00 1.48E+00 1.96E-01 1.30E+00 2.35E+00 2.83E-01 1.83E+00 7.81E-01 1.05E-01 3.04E+00 6.22F-0r 2.08E-01 2.33E+00 1.35E+00 2.10E-01 1.64E+00 2.18E+00 2.87E-01 2.238+00 7.33E-01 1.20E-01 3.66E+00 s.66E-01 2.I5E-01 2.79E+00 1.25E+00 2.31E-01 1.97E+00 2.08E+00 2.95E-01 i:t ila NO NE AJ l.,lflf E

Ir+tE

.5 Hz Sr [el Mnnran AF Srcnaa Ln(AF) lIJ:z sA [gl Mnnmx AF Srcpra Ln(AF) 0.5 Hz Sa [gl Mnumu AF Srcua Ln(AF-)

0.1 Hz Sn [sl Mrul+,trl AF Srcua Ln(AF) 2.038-02 1.72E+00 2.298-01 1.39E-02 1.32E+00 1.78E-01 7.89E-03 1.26E+00 6.93F.02 3.56E-04 l.l8E+00 9.00E-02 6.46E-02 t.8lE+00 2.63E-01 3.53E-02 1.34E+00 I.8lE-01 1.75E-02 1.27F+00 7.26E-02 6.64E-04 1.23E+00 9.56E-02 1.07E-01 1.87E+00 2.88E-01 5.56E-02 1.36E+00 1.83E-01 2.66F.02 1.27E+00 7.398-02 1.018-03 1.25E+00 9.78E-02 3.528-01 2.16E+00 3.63E-01 1.72E.07 1.41E+00 1.97E-01 7.85E-02 1.29E+00 8.03E-02 3.07E-03 1.28E+00 9.73E-02 4.328-01 2.258+00 3.69E-01 2.10E-01 1.42E+00 2.02F.41 9.54E-02 1.30E+00 8.12E-02 3.75E-03 1.28E+00 9.678-02 6.32E-01 2.48E+00 3.60E-01 3.04E-01 1.46E+00 2.20E-01 1.37E-01 1.30E+00 7.998-02 5.44E-03 1.29E+00 9.66E-02 8.40E-01 2.69E+00 3.43E-01 4.02E-01 1.49E+00 2.30E-01 l.8lE-O1 1.31E+00 7.86E-02 7.21E.03 1.30E+00 9.578-02 1.06E+00 2.84E+00 3.19E-01 5.04E-01 1.52E+00 2.50E-01 2.26E.01 1.32E+00 7.80E-02 9.05E-03 1.31E+00 9.478-02 1.278+00 2.92E+00 2.97F-01 6.02E-01 1.55E+00 2.828-01 2.70E-01 1.32E+00 7.85E-02 1.08E-02 1.31E+00 9.26E-02 N)

RT sE+

\\I >U ]Y r-r$ H F=..1 LH.ry ss I

\\os' N)

N lN fJt

2734294-R-036 Reoision 0 May 11,201,7 Page 25 of 145

\\,

\\

J

\\

-t:

ar)

-i,

\\

d lG'

\\

b

+.

IJ IEIt 5

a-tcot, t-rh a--c E

4.0 3.5 3.0 2.5 2.O 1.5 1.0 0.5 0.0 0.5 Hz SA [g]

1 Hz SA [g]

,- - 2.5 Hz. SA [g]

-. -5Hz SA[g]

- - 10Hz SAtgl

-- --25 Hz SA tSI 100 Hz sA [g]

0.01 0.10 1.00 Spectral Acceleration (gl 10.00 10.00 F'IGURE 3-4 MEDIAN TOTAL AMPLIFICATION FACTORS YERSUS INPUT HARD.ROCK MOTION X'OR BVPS SITE AT EL 681 o.4 0.5 Hz SA [S]

1Hz SA[e]

- 2.5 Hz SA [g]

-. -5Hz SA[g]

- - 10Hz SA[g]

---

25 Hz SA IgI 100 Hz SA [gJ 0.0 0.01 0.10 1.00 Spectral Acceleratlon (g!

A b

+,E 0.3 lL eo f-+,o(,

E o.2 B

E YE E 0.1

.!p vt FIGURE 3.5 srcMA LN (TOTAL AMPLIFICATION FACTORS) VERSUS rNPUT HARD-ROCK MOTION T'OR BVPS SITE AT EL 681 Given the complexity of the logic trees used to represent epistemic uncertainty in the CEUS-SSC model (Reference 2l) andEPzu GMM (Reference 22),the computational demands of propagating all epistemic uncertainty in the site response logic tree into the PSHA is prohibitive.

As a result, an assessment was performed to determine how the site response logic tree could be fBSConsulting

()Rrzzo

/'

l

,I I

]

)

I2

2734294-R-036 Reaision 0 May LL,201-7 Pnop 76 nf74E simplified without loss of accuracy in the hazard fractiles at the RCBX foundation elevation (EL 681 ft).

Sensitivity studies were performed to test the approach to grouping on the resulting surface hazard fractiles. Specifically, sensitivity testing assessed the impact of the AF grouping process, in which a portion of the epistemic uncertainty is transferred to aleatory uncertainty. The sensitivity study shows that the AF grouping approach has minimal impact on the mean hazard and on any of the hazard fractiles above the mean for all levels of ground motion. The steps for development of the surface control point hazard curves and haeard fractiles are:

. At each response frequency, group the site AFs according to the patterns observed in the site response logic tree branch of site AFs, as described below.

. Apply the grouped site AFs to all logic tree branches of the CEUS-SSC model and EPRI GMM used to derive the hard-rock hazard.

Combine the surface hazard branches, using the same combinations as were used to derive the seismic hazard for hard-rock site conditions.

The assessment performed to determine if grouping of AFs was technically justified began with compilation of all AF branches for the seven response frequencies (0.5 Hz, 1.0 Hz, 2.5 Hz, 5 [fz, 10 Hz, 25 Hz, and 100 Hz). For each response frequency, the mean and standard deviation of AFs are saved, consistent with each end-branch of the site response logic tree. Based on the observed pattern in the trend of mean AF over each of the seven response frequencies, three grouped branches are determined for use calculating the control point hazard.

Figure 3-6 and Figure 3-Tdisplaythe 20 individual branches of AF fromthe logic tree together with the recommended three grouped amplification functions for two example response frequencies. On each figure P represents the site profile, M represents the material dynamic properties, K represents kappa, and I ClzC represent the single-corner and double-corner input motions respectively. The fuIl set of AF grouping results are listed in Reference 23.

lESGonsulting

[]R.z7o

2734294-R-036 Reaision 0 May 11,201,7 Page 27 of 1.115

-Pt Ml Kl lC GPtMtKl 2C Pl M2 K1 tC

-PtM2K1 rc P2Ml Kl lC P2Mr Kl 2C P2M1 K2 tC P2Ml K2 2C P2 M2 Kl rC

-P2M2KI 2C P2ir2K2 rC P2 M2 K2 2C PSMt Kl rC p3Mr Kl 2C P3Ml K2 tC P3 Mt K2 2C Pt M2 Kl rC P3 M2 Kl 2C P3 M2 K2 rC P3 M2 K2 2C

-Qcup_l AF r - &anp-2 AF r&anpJAF 0.005 0.05 0.5 5

Acc.hrdonlgl rorAl sEr oF MEA. sIrE ",ffit#ffilfro*

FACroRs x'oR r0 HERrz SPECTRAL ACCELERATION TOGETHER WITH TIIE THREE AMPLIX'ICATION X'ACTORS FOR THE SELECTED GROUPINGS 3

2.5 0.5 0

lr{

51 o.

I 3,5

-Pt ifl Kl tc

-Pt Mt Kt 2C

-Pt M2 Kt tC

-Pt M2Kl 2C

-P2Ml Kl lC

-PzMt Kt 2C

-P2Mt K2 tC

-P2Ul K2?C

-P2M2Kl lC

-P2M2Kl 2C PZ)12 K2 lC P2M2K2rc

-PsMl Kl lC

-P3Ht Kt rc P3 Ml K2 rC P3MI Y.2?c P3 M2 Kl lC P3 M2 Kr aC P3 M2 r(2 tC P3 u2 K2 2C

-Gottp_lAF r - Gorp-2AF eGoup_3AF 3

2.5 lr'{

f,2 oI 1.5 0.5 00.00s 0.05 0.5 s

Acc.hr*lon (gl F'IGURE 3.7 TOTAL SET OF'MEAN SITE AMPLIFICATION FACTORS F'OR 2.5 HERTZ SPECTRAL ACCELERATION TOGETHER WITH THE THREE AMPLItr'ICATION FACTORS T'OR THE SELECTED GROUPINGS lESConsultirq

()Rtzzo

273429+R-036 Rwision 0 May 1,1,20L7 Page 28 of 1.45 3.1.2 Seismic Hazard Analysis Technical Adequacy The BVPS-2 SPRA hazard methodology and analysis associated with the horizontal GMRS were submitted to the NRC as part of the BVPS-2 Seismic Hazard Submittal (Reference 3), and found to be technically acceptable by NRC for application to the BVPS-2 SPRA.

Subsequent to the March 31,2014 (Reference 3) submittal, the seismic hazard was updated and FIRS were generated for each of the foundation elevations associated with critical structures as the BVPS for use inthe SPRA. Figare 3-Spresents the FIRS atthe control pointEL 681 ft and compares this to the GMRS reported in the BVPS-2 March 2014 submittal (Reference 3). The difference is attributed to:

I The material damping used for the rock material over the upper 500 ft. S/hile the GMRS, reported in the March 2014, submittal is based on the low-strain damping of 3.2 percent over a 500-foot depth of bedrock, the FIRS used in the BVPS-2 SPRA limits this damping value to the upper 100 ft where the rock is considered as weathered or fractured. Within the depth range of 100 ft to 500 ft, a damping of 1 percent is used based on the unweathered shale dynamic properties from Stokoe et a1., (Reference 75). Below a depth of 500 ft, linear material behavior is adopted with the damping value of 0.5 percent is specified consistent with the kappa estimate for the site.

The subsurface profile used in the site amplification analysis. While the GMRS, reported in the March 2014, submittal is based on a profile which extends from the bottom of the RCBX foundation to at depth hard rock, the FIRS used in the SPRA develops from the analysis of the full soil column to plant grade, subsequently truncated to the RCBX foundation level, in accordance with NRC guidance (Reference 25).

IESGoneulting

(]Rtzzo

273429+R-036 Reaision 0 May 1,1,2017 Page 29 of 145

- -.fi$ GMRS NTTF 2.1 Submittal

-RB FlRS SPRA t

0.t ot o.7 0.5 05 0,tl 0.3 02 0.r AD U

Eo-,y lB Lo-o C'

C' a!

Ltoo.

vl 0.10 10"o Frcquency (Hz) f,.IGURE 3.8 COMPARISON BETWEEN GMRS AT CONTROL POINT REPORTED IN MARCH 2014 SUBMITTAL AI\\D X'IRS USED AS BASIS X'OR BTIILDING SEISMIC RESPONSE AI\\D I'RAGILITY CALCT'LATION IN BVPS.2 SPRA PROJECT The BVPS-2hazard analysis was also subjected to an independent peer review against the pertinent requirements in the PRA Standard (Reference 4). The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A.

3.1.3 Seismic HazardAnalysis Results and Insights This section provides the final seismic hazardresults used in the BVPS-2 SPRA.

The site AFs obtained from the site response analysis and the hard-rock PSHA curyes are used to develop the seismic hazard curves and FIRS at the elevations of interest. The procedure to develop the seismic hazafi curves follows the methodology described in Reference 2. This procedure, referred to as Approach 3, computes a site-specific contol point hazard curve for a broad range of Sa given the site-specific bedrock hazard curve and site-specific estimates of soil or soft-rock response and associated uncertainties. The FIRS represent the performance-based ground motion used as input to the seismic analysis of the buildings.

The above procedure is executed to generate the mean hazard, curye and the fractiles at EL 681.

Figure 3-9 presents the mean and fractile hazard curves at EL 681 for the specfral frequency of 100 Hz. Table 3-6 presents numerical values of the mean hazxd curve and the fractiles of the hazarddistribution. The full set of hazard curves at EL 681 can be found in Reference 23.

The PSHA results were used to perform an assessment of the totalhazard sensitivity to the epistemic uncertainty in the particular PSHA input variable (i.e., ground motion prediction equation (GMPE), seismicity of distributed sources, maximum magnitude of distributed 10.00 Lq, fBSConsultlng

()Rrzzo

2734294-R-036 Reaision 0 May 11,,201,7 Pase 30 of 1,45 sources, etc.), which is measured by the variance in the totalhazadwith contribution solely from the epistemic uncertainty in the specific input variable, normalizedby the variance in the total ltazard.

The results of this process are shown on Figure 3-l0 which displays the variance deaggregation for the spectral frequency of 100 Hz (PGA) at the RCBX control point at EL 681 ft.

Deaggregation is shown for MAFE ranging from 1.40E-3 to2.34E-8. The dominant confributor to the total variance is the epistemic uncertainty in GMPEs. As the MAFE gets lower, the epistemic uncertainty in maximum magnitude, the three magnifude-range cases used for deriving rectrrrence rates, and the eight recurence raterealizations become more significant.

100.0 Hz 1.ffi-02 1.0E-03 1.0E-04 1.OE-05 1.tr{5 1,.ff.-O7 1.OE-08 ot,tro?oo I

13 UJ E

C'tro3tro L

llt Ir!

=tc 0.01 0.10 Acceleratlon [gl 1.00 10.oo T.IGURE 3.9 100-HZ Sr MEAN AND FRACTILE HAZARD CURVES X'OR BVPS SITE AT EL 681 (BASE OF BVPS-I AND BVPS-2 REACTOR CONTAINMENT BUILDING FOUNDATION)

Note:

ftd indicates the seismic hazard at the RCBX foundation level.

fBtConsulting rlRrzzo Mean

....... Sth_fdn 16_th_fdn

-. - 50th-fdn

- g4th_fdn

. 95_fdn

273429+R-036 Rsuision 0 May 11,,201,7 Page 3L of 145 00e(

8096 7Wo

@!6 50ts

ffi, 30$6 M,

0,6 100 Hz Hazard E

tr

.0tI L

E(,

t09r r*

rIqrr GM Cludr ed DidrbuEd 3 Cmol hr I Rrdizrffmr Mcdirn ModGl Sourcc Mmu Dirtrbubd br Di$rihrtcd Scirmicity Scismicity ShG Amplificetion Brrrdr DiCbubd Sourc Scirmgrnic Dcpilt Oirtrbutad Soreo Modcl r MAFE=I.40E.3 I i4AFEc8.55E-5 r ffif'f=t,04E-5 r MAFEs:I.01E.6 r MAFEs1.20E s MAFE-2,3{E{

I.IGURE 3.10 VARIAI\\ICE DEAGGREGATION OF'THE BVPS SITE PSHA LOGIC TREE INPUTS x'oR THE SPECTRAL X',REQUENCY OX' tooldz TABLE 3.6 100-lXL Sr MEAITI AI\\D X'RACTILE HAZARD X'OR BVPS SITE AT EL 681 (BASE OF BVPS-I Ar\\D BVPS-2 REACTOR CONTATNMENT BUTLDING X'OI]NDATIOI9 AISConsultlng

(}Rrzzp.

SpnCTRAL AccTLERATION tsl ANnu^lL FREeUENCY oF ExcTEDANCE Mn^l..N 5rt 16TH 5OTH 84TH 95r" 0.01 l.l9E-02 3.61E-03 6.728-03 1.028-02 l.l2E-02 2.688-02 0.02 4.33E-03 1.35E-03 2.228-03 3.758-03 5.83E-03 l.l6E-02 0.03 2.238-03 6.678-04 1.05E-03 l.8lE-03 2.968-03 6.548-03 0.04 1.448-03 4.208-04 6.168-04 l.l4E-03 1.99E-03 4.41E-03 0.05 1.03E-03 2.848-04 4.168-04 7.498-04 1.48E-03 3.428-03 0.06 7.698-04 1.85E-04 2.948-04 5.56E-44 1.1 2E-03 2.578-03 0.07 6.028-04 1.328-04 2.038-04 4.278-04 8.528-04 2.038-03 0.08 4.86E-04 9.348-05 1.55E-04 3.398-04 7.13F.-04 1.70E-03 0.09 4.028-04 7.208-05 1.228-04 2.798-04 5.91E-04 l.4lE-03 0.10 3.3 8E -04 5.82E-05 9.238-05 2.228-04 5.18E-04 1.19E-03 0.20 8.92E-05 l.l4E-05 2.08E-05 5.33E-05 1.458-04 2.98E-04 0.25 5.548-05 6.728-06 1.3 1E-05 3.498-05 8.60E-05 1.85E-04 0.30 3.778-05 4.51E-06 8.30E-06 2.31E-05 6.09E-05 1.278-04

273429+R-036 Reaision 0 May 11,2017 Pnop.17 nf 145 TABLE 3-6 100-HZ Sn MEAN AND FRACTILE HAZARI) FOR BVPS SITB AT EL 681 (BASE OF BYPS-I AND BVPS-2 REACTOR CONTAINMENT BUILDING FOUNDATTON)

(coNTTNUED)

Srncrnq.l AccnIT,RATION lel AITInUII FnnQunNCY oF ExCEEI}ANCE Mrau 5"

16TH SOTH 84'H 95TH 0.40 1.95E-05 2.208-06 4.38E-06 t.22E-0s 3.19E-0s 6.13E-05 0.s0 r.09E-05

1. l7E-06 2.328-06 6.66E-06 1.86E-05 3.52E-05 0.60 6.s8E-06 6.3sE-07 1.35E-06

4. 16E-06 1.13E-05 2.09E-05 0.70 4.21E-06 3.68E-07 8.028-07 2.s8E-06

7. I 1E-06 1.38E-05 0.80 2.78E-06 2.20E.-07 5.01E-07 1.68E-06 4.758-06 8.70E-06 0.90 1.87E-06 1.30E-07 3.20E.07 1.08E-06 3.13E-06

6. 16E-06 1.00 1.26E-06 7.42E.08 2.04E-07 6.88E-07 2.20E-06 4.328-06 2.00 9.70E-08 2.248-09 7.378 -09 3.878-08 1.64E-07 3.94F-07 3.00 2.44E-08 2.59E-10 1.1 1E-09 7.66E.-09 3.75E-08 1.0sE-07 5.00 3.6sE-09 1.26E-11 6.77E-tt 7.24E-10 4.89E-09 l.7lE-08 6.00 1.74E-09 3.89E-12 2.28E-1 I 2.798-10 2.21E.-09 8.23E-09 7.00 9.40E-10 1.36E-12 8.228-12 1.27E-10 l.l4E-09 4.62E.09 8.00 5.61E-10 s.3 8E-r 3 3.938-12 6.27F-n 6.25E-10 2.70E.-09 9.00 3.41E-1 0 2.31E-13 1.778-12 3.30E-l I 3.63E-10 1.76E-09 10.00 2.19E-10 1.0s8-13 8.24E-13 1.77E-l I 2.19E-10 1.07E-09 Following the guidelines in Reference24 the FIRS for the control point of interest are developed following a perfonnance-based approach. The foundation level seismic hazard curves and UHRS provide the input to derive the perfoffnance-based FIRS. The perfonnance-based FIRS are developed by scaling the mean lE-4 MAFE UHRS by a design factor that is related to the ratio of the 1E-5 MAFE Seto the corresponding lE-4 MAFE Sa (Reference 24).

Figure 3-ll presents the perfoffnance-based horizontal FIRS at EL 681, and the IE-4 and 1E-5 UHRS. Table 3-Tpresents numerical values of the Se forthe FIRS at EL 681. The horizontal FIRS at all other foundation elevations are presented in Reference 23.

ABSCutsulting

{}Htrz.o

273429+R-036 Reaision 0 May 1L,20L7 Page 33 of 145 I

1.400 1.200 1.000 0.800 0.600 0.400 0.200 0.000

!0-,

tro a-+,

Eg o(,

C' Eft 1A (D - - 1x104 UHRS [g]

1x1O-5 UHRS [gl FIRS

\\

0.10 1.m 10.00 Frequency (Hzl FIGURE 3-11 UHRS AND FIRS AT THE BVPS SITE AT EL 681 TABLE3-7 UHRS AND FIRS AT THE BVPS SITE AT EL 681 100.00 FnneuENCY (ru)

HORIZONTAT, SPPCTUL ACCELERATIOU (E) AT THE FOUNDATION ELEVATION lXlO.A MAFE TIHRS 1X1O-5 MAFE TIHRS FIRS 0.10 0.0028 0.0067 0.0034

0. 13 0.0040 0.0097 0.0048 0.16 0.0058 0.0141 0.0071 0.20 0.0088 0.021 I 0.0106 0.26 0.0136 0.0321 0.0 162 0.33 0.0205 0.047 4 0.0240 0.42 0.0285 0.0642 0.0328 0.50 0.0353 0.0780 0.0399 0.53 0.0352 0.0783 0.0400 0.67 0.0366 0.083 I 0.0423 0.85 0.0459 0.1077 0.0545 1.00 0.0534 0.1264 0.0639 1.08 0.0573 0.13 84 0.0696 1.37 0.0673 0.17 56 0.0870 1.74 0.0829 0.2351 0.1 145 2.21 0.1 1 15 0.3495 0.1669 fSConsultfuU

()Rrzzo

273429+R-036 Reaision 0 May 1.L,2017 Page 34 af 1.45 TABLE 3.7 UHRS AFID FIRS AT THE BVPS SITE AT BL 681 (coNTINUED)

FnneupNCY (Hz)

HomzoNTAL Spncrnnl AccELERATToN (e) ar rHE FoUNDATToN Elnv.+.rrou IXIOA MAFE T]IIRS IX1O.5 MAFE I]HRS FIRS 2.s0 0.13 r 0 0.4378 0.2064 2.81

0. I 654 0.s799 0.2707 3.s6 0.2802 0.9789 0.4s73 4.52 0.4272 1.3039 0.6258 s.00 0.4s74 r.3216 0.6413 5.74 0.4s 10 r.2712 0.6200 7.28 0.3927 1.1447 0.554s 9.24 0.3372 1.1085 0.s242 10.00 0.3429 1.t766 0.5517 11.72 0.3798 r.2689 0.5981 14.87 0.4039 1.1986 0.5785 18.87 0.3727 r.0896 0.527 5 23.95 0.3193 0.9138 0.4443 25.00 0.3092 0.8922 0.4331 30.39 0.2919 0.8157 0.3985 38.57 0.2724 0.7286 0.3591 48.94 0.2575 0.6661 0.3305 62.t0 0.2333 0.5956 0.2963 78.80 0.2026 0.5244 0.2601 100.00 0.1 885 0.5158 0.2530 Note:

MAFE: mean annual frequency of exceedance.

A3$Gortsulting tlRtzz,o

2734294-R-036 Reaision 0 May 11,2017 Page 35 of 1-45 3.1.4 Horizontal and Vertical FIRS This section provides the control point horizontal and vertical FIRS.

Vertical response spectra are developed at each foundation elevation by combining the appropriate horizontal response spectra and vertical-to-horizontal (V/H) response spectral ratios.

The V/H response spectral ratios consider guidance provided in Reference 77 andReferenceT9, which both provide approaches applicable to a range of CEUS or WUS sites.

Forthe BVPS Site three factors influence the approach used to derive V/H ratios: (1) the kappa values estimated for the site are significantly larger than the hard-rock kappa value of 0.006s reported for CEUS hard-rock sites in Reference 77, (2) the site-specific Vsrovalues for the site profiles are best associated with intermediate or soft sites as reflected in Reference 79, and (3) the shape of the horizontal FIRS at each of the foundation elevations peak at spectral frequencies closer to WUS spectral shapes. Given these factors the approach used to derive V/H ratios for the BVPS Site considers the generic V/H ratios from ReferenceTT and the empirical GMPEs as described in ReferenceT9. For each foundation elevation a mean V/H ratio is derived by considering equal weights for WUS and CEUS rock site conditions, and equal weights on the V/H values derived by applying the GMPEs Reference 80 and Reference 8l and the generic V/FI values from Reference 77.

The calculated V/H ratio for the RCBX foundation elevation is shown on Figure 3-12 which displays the results separately for WUS rock conditions and CEUS rock conditions, showing the range of values for the models considered and the overall median V/H ratio from this range. On this figure the bottom plot displays the overall mean V/H ratio for WUS and CEUS rock conditions (from the top two figures) and the recommended V/H ratio based on averaging the mean V/H ratio for WUS and CEUS rock conditions. The vertical FIRS are derived using the V/H ratios and the horizontal FIRS. Figure 3-13 shows the horizontal and vertical FIRS at the RCBX foundation elevation. The horizontal FIRS, the applicable V/H ratios, and the vertical FIRS for the RCBX foundation elevation are displayed on Table 3-8. The full set of V/H ratios and vertical FIRS at other foundation elevations can be found in Reference 23.

lESGansulting tlR.zzo

273429+R-036 Reaision 0 May 1,L,2017 Page 36 of 145 o.60 I

a-f,E-4 1.20 1.m 0.80 0.40 0.20 0.m 1.20 1.m o.m o.40 0.20 o.m 0.1 wrrs Rocf, coiltxclx)il Frequcncy (Hrl cB(N, SS CgH, TH c804, R

- - CBO4, Awrage GA1l SS GA11 N GA11 R

- - GA1l (average)

-NUREG 6'nB WUSV/H ratios O Mean V&l Ratio c804, ss c804, TH c8(N, R

- - C804, Average GA11 SS GA11 N GAll R

- - GAl1 (average)

-NUREG 6728 CEUS V/H ratos O Mean VIH Ratio 1

10 1m 0.60 soETE-0.1 1m Frcqucnc? (Hrl F'IGURE 3.12 VERTICAL.TO-HORIZONTAL RATIOS FROM DIFFERENT MODELS AND THE RECOMMENDED MEDIAN YERTICAL.TO.HORIZONTAL RATIO X'OR BVPS SITE EL 681 1

10 fESConsulting

(}Rrzzo I

I I

CEUS ROCT CONDICTION

'iii;i:

i:r

'i:!i ftrts l

s: t1 f

\\

1 i

I

^/ / I t

I

)

o

( )rl

,t,

.al

\\d_-

N Ei

2734294-R-036 Reoision 0 May 1,1,,20L7 Page 37 of 1,115

-ff;se6rmnded Mcrn Vftl rotio Vftl ratio WUS rock Conditions V&l rrtioCEUS rock Conditions o.1 1

Frqu.nGy(Hr)

FIGTIRE 3.12 (coNTII\\IUED)

VERTICAL-TO.HORIZONTAL RATIOS FROM DIX'TERENT MODELS AIYD TIIE RECOMMENDED MEDIAIT VERTICAL-TO.HORIZONTAL RATIO F'OR BVPS SITE EL 681 nt\\

t\\

I

^-\\

In-t\\\\

ff

\\

/t

?

o.7 o.6

-E E o.s I,Ti o-4

$ o.,

E o

ra o.1 Horizontal

-!Tqrtkal o

o.1 110 Fregrenca tHzl 100 F'IGURE 3-13 HORIZONTAL AI\\D VERTICAL FOUNDATION INPUT RESPONSE SPECTRA AT TIIE BVPS SITE AT FOUIYDATION EL 681 0.9 0.8 o.7 0.6

.E o.s fiE S r.o 0.3 0.2 0.1 1m 10 fSSGonsultltU

()Rrzzo

273429+R-036 Reoision 0 May 1.1,,201.7 Page 38 of MS TABLE 3.8 HORIZONTAL AND VERTICAL FIRS AT THE BVPS SITE AT EL 681 FnreunNCY ffiz)

HoruzoNTAL FIRS (s)

V/H Rq.rro Ynnrrc.l,L FIRS (s) 0.1 00 0.0034 0.7045 0.0024 0.200 0.0106 0.7045 0.0074 0.331 0.0242 0.7045 0.0170 0.s01 0.0399 0.67 54 4.4270 0.676 0.0425 0.6577 0.0280 1.000 0.0639 0.6480 0.0414 1.202 0.0770 0.6270 0.0483 1.413 0.0898 0.6108 0.0549 t.622

0. I 047 0.6003 0.0629 1.820 0.1220 0.5919 0.0722 2.042 0.1464 0.5843 0.085s 2.1 88

0. l 641 0.s812 0.09s4 2.399 0.r91r 0.5803 0.1 1 09 2.630 0.23 l0 0.s806 0.1 34 l 2.818 0.2726 0.s813 0.1s8s 3.020 0.3220 0.s820 0.1 874 3.311 0.3987 0.5829 0.2324 3.631 0.4723 0.5 821 0.2749 3.q81 0.5484 0.5800 0.3 181 4.266 0.5999 0.581I 0.3486 4.571 0.6284 0.5839 0.3669 4.786 0.6375 0.5867 0.3740 5.01 2 0.6413 0.5906 0.3787 s.248 0.6376 0.59s4 0.3797 5.495 0.629r 0.6009 0.3780 5.754 0.61 95 0.6067 0.37s8 6.026 0.6087 0.6 r 26 0.3729 6.457 0.5879 0.621 I 0.3655 6.918 0.5663 0.6300 0.3568 7.413 0.ss l4 0.6376 0.3s 16 7.763 0.s436 0.6425 0.3492 7.943 0.5398 0.64s0 0.348 r 8.51 I 0.s297 0.6s29 0.34s9 8.913 0.s2s4 0.6584 0.3460 9.ss0 0.5318 0.6669 0.3547 r 0.000 0.5517 0.6730 0.3713 lESGoneulting rlRtzzo

2734294-R-036 Reaision 0 May 1L,201.7 Page 39 ofL45 TABLE 3-8 HORIZONTAL AND VERTICAL F'IRS AT THE BYPS SITE AT EL 681 (coNTTNUED)

FnreurNCY (Hz)

HonTzoNTAL FIRS (s)

VfiI Rluo VnRrrcaI FIRS (s) t2.023 0,5979 0.7030 0.4203 t4.125 0.s849 0.7138 0.4175 16.21 I 0.s64s 0.7226 0.4079 18.r97 0.5372 0.7285 0.3 9 l4 20.417 0.502s 0.732s 0.3681 22.387 0.4676 0.734s 0.343s 23.988 0.4438 0.7328 0.32s3 26343 0.4229 0.731 I 0.3092

28. I 84 0.4109 0.7298 0.2998 30,200 0.3995 0.7290 0.2913 34.674 0.37s8 0.7317 0.27 50 39.811 0.3549 0.7392 0.2623 44.668 0.3413 0.7411 0.2s30 s0.l 19 0.3273 0.7383 0.2417 54.954 0.3144 0.t37s 0.23 t I 60.256 0.3007 0.7366 0.2215 70.795 0.2736 0.7324 0.2004 81.283 0.2587 0.7236 0.t872 100.000 0.2530 0.7017 0.1776 lESGonsulting rlR.zzo

273429+R-036 Reaision 0 May 1,1,,201,7 Page a0 of L45 Dynamic properties of soil are degraded due to their non-linear response under a controlling earthquake motion propagated through the soil profile. This degradation is represented by strain-compatible dynamic properties obtained from the output of an equivalent-linear site response analysis. Epistemic and aleatory uncertainty of the input motion, Vs, thickness, damping etc., is included in the site amplification analysis. Three deterministic soil profiles that represent uncertainty in Vs, Vp, damping, ffid thickness are provided. The approach is consistent with Reference 82 and Reference 2.

A fully probabilistic approach is employed to develop the strain-compatible dynamic properties that preserve consistency with the ground motion hazard. Assuming the strain-compatible properties are lognormally distrihuted, this approach is analogous to Approach 3 described in Reference 77. The mean and standard deviation of logarithmic (lne) strain-compatible properties are determined as a function of rock Se for each soil layer in the same manner that a mean and standard deviation of logarithmic site AFs is determined. The soil Se is determined from the soil hazard curve at the MAFE of interest, and the corresponding AFs and associated strain-compatible properties at the soil Sa are used.

Reference 2 considers the variation of the strain-compatible property for different response frequencies of the FIRS. The FIRS is not a response spectrum associated with a single earthquake, so the main contributor at a spectral frequency of 1.0 Hz could produce strains in the soil column different from those produced by the main contributor at a spectral frequency of 100 Hz (assumed to be PGA). To address this, Reference 2 states: "To examine consistency in strain-compatible properties across structural frequency, the entire process is performed at PGA (typically 100 Hz), and again at low frequency, typically I Hz. If the differences inproperties at high-and low frequency are less than 109/0, the high-frequency properties may be used since this frequency range typically has the greatest impact on soil nonlinearity. If the difference exceeds 10% [hazard-consistent strain-compatible properties] thehazard-consistent strain-compatihle properties (HCSCP) developed at PGA and those developed at I Hz may be combined with equal weights."

To implement this requirement, two set of strain-compatible properties are obtained; one for a spectral frequency of I Hz and the other for 100 Hz Se (PGA). If the differences between the means or standard deviations for the two spectral frequencies are larger than 10 percent, then the approach described above is used.

Once the BE strain-compatible shear modulus (G) and shear-wave damping (S) profiles and their standard deviations are determined, the upper and lower bound profiles are determined following Reference 82. The minimum requirement for coefficient of variation (COV) for site material in NRC, (2013) is 0.5 for well-investigated and 1.0 otherwise.

The resulting set of strain-compatible properties for the BVPS Site is provided in Reference 23.

lESConsulting rlRtzza

273429+R-036 Reoision 0 Moy LL,201.7 Page 4L of 1.45 4.0 DETERMINATION OF SEISMIC FRAGILITIES FOR THE SPRA This section provides a srunmary of the process for identifuing and developing fragilities for SSCs that participate in the plant response to a seismic event for the BVPS-2 SPRA. The subsections provide brief summaries of these elements.

4,1 Susruc EeurpuENT Lrsr For the BVPS-2 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 EPRI 3002000709 (Reference 15).

4.1.1 SEL Ilevelopment The BVPS-2 SEL was developed as follows:

Potential seismic-induced initiating events and consequential events were identified based on the internal events PRA and review of other potential seismic initiators. The following is a swnmary of items considered in developing the SEL.

The creation of the BVPS-2 SPRA SEL started with the SSCs listed in the existing BVPS-2 PRA, Internal Events Model. It further considered the list of SSCs developed much earlier for the BVPS-2 individual plant examination of external events (IPEEE [Reference 9J).

The following bases were used in the development of the BVPS-2 SPRA SEL:

1.

The existing Internal Events PRA for BVPS-2 meets the Capability Category II requirements of the American Society of Mechanical Engineers (ASME) PRA Standard for PRA applications and complies with Regulatory Guide 1.200, Revision 1 (Reference 16).

2.

The internal events PRA model was last formally documented in 2015 (Reference 89).

This BV2REV6 model served as the foundation for the latest version of the seismic PRA presented in the interim revision BV2REV6A model. While the seismic PRA evaluated in the December 2014 Seismic Peer Review used an earlier effective Internal Events PRA model as its foundation, the same methodologies were used when incorporating into the latest model.

3.

SSCs located inthe turbine building are included inthe SPRA SEL, althoughthey are not credited in the SPRA sequence models. While the turbine building has some seismic capacity, it also contains numerous non-seismic SSCs that may fail in ways that fail other SSCs within the building and prevent operator access to the turbine building. Future SPRA evaluations may choose to credit the turbine building and these specific SSCs.

Non-seismic electrical equipment which brings offsite power to the essential buses, are not located in the turbine building.

fESConsulting tlRlzzo

273429+R-036 Reaision 0 May 1,'1,,2017 Page a2 of 1.45 4.1.1,1 Use of the Internal Events PRA and IPEEE Lists of SSCs The EPRI guidance document (Reference 2) says that using the previously developed IPEEE SEL as a starting point for listing the SSCs is acceptable. The ASME combined standard (Reference 4) says to use the existing internal events PRA model as the basis for building the seismic PRA logic model. The ASME Standard implies that the SSCs represented in the PRA logic model basic events would make up a starting point for such an SEL. As the IPEEE SEL includes some SSCs originally judged important for seismic risk, but that are not normally found in a PRA logic model for internal events, it was decided to combine the two lists of SSCs as a starting point forthe development of the SPRAT i.e., the original IPEEE SEL and the SSCs from the current internal events PRA. This initial combined list does not mean that all SSCs listed in the IPEEE or PRA SEL lists will be explicitly represented in the seismic PRA. Rather, it means that they will be included for consideration during the seismic walkdown and their impact on plant response in an earthquake will then be considered.

For BVPS-2, the internal events PRA logic model (Reference 26) is well established, having evolved since the original individual plant examination in the early 1990s. For example, in parallel to this effort to construct an SPRA, the BVPS-2 PRA was also revised to update the PRA logic models for internal events, internal flooding and for internal fire initiating events. The effort from these updates is considered in so far as they may impact the SPRA; e.g., especially in the identification of elechical cabinets and panels whose faih.res could impact the plant response in an earthquake and the listing of potential flood sources.

The internal initiating events were also reviewed for applicability to seismic sequences.

Tahle 4-I presents all of the initiating events and how they are treated in the seismic PRA.

fESGoneulting

[]R]zzo

273429+R-036 Reaision 0 May 1.1.,201.7 Page 43 of145 TABLE 4-1 REVIEW OF'INTERNAL INITIATING EVENTS FOR APPLICABILITY TO SEISMIC SEQUENCES IFrrrtnrrNc Evrxr C.rrscoRrrs Monnuuc or lruIruroR FoR SPRA I

Excessive LOCA (reactor vessel failure, not coolable by ECCS)

Seismic failure of reactor vessel included as EQ37, part of Top Event ZLI

2. Larse LOCA (> 5' UP TO DBA) BVPS-2 per Rx Crit Yr Screened out on hieh capacity

3. Medium LOCA (1.5" TO 5") BVPS-2 per Rx Crit Yr MLOCA assigned fragility curve; seismic failure leads to direct core damage via failure of Top Event ZLI

4. Small LOCA. Nonisolable (0.5 inch to I.S-inch diameter)

Fail too event PR and assume CIA and CIB conditions

5.

Small LOC,A, Isolable (pressure-operated relief valves (PORV) train leakase) (0.5" to 1.5")

Screen out, not a seismic failure mode

6. Interfacing Systems LOCA Screen out on hieh seismic capacity

7. Steam Generator Tube Rupture Screen out, not a seismic failure mode

8. Reactor Trip Assuming plant trip for every seismic initiator due to turbine building failures

9. Turbine Trip Assuming plant trip for every seismic initiator due to turbine buildins failures: requirins turbine to trip each time

10. Loss of Condenser Vacuum Assuming condensate lost for all seismic events; and that there is a resulting pressurizer PORV challenge I l. Closure of All Main Steam Isolation Valves (MSIV)

MSIVs not always required to close but likely will due to loss of station air

12. Steam Line Break Unstream of MSIVs

a. Steam Line Break Inside Containment Screened on high seismic capacity

b. Main Steam Relief or Safety Valve Opening Valves modeled for seismic failure to open

c. Steam Line Break in Common Residual Heat Removal System (RHS) Valve Line Screened on high seismic capacity

13. Steam Line Break Downstream of MSIVs (Outside Containment)

Turbine building collapse is assumed to shear the steamlines; fragility curves for MSIVs are assigned to top event ZMS, and failure of this top event would then fail top event MS and result in loss of steam supply to the TDAFW pump. To satisfu seismic PRA peer review F&O #7-1, this is accounted for in the GENTRANS tee STEAM macro, which includes ALLSEISMIC*MS:F logic

14. Inadvertent Safety Iniection Screened on hieh seismic capacify

15. Miscellaneous Transients

a. Total Main Feedwater Loss or Condensate Assumed for all seismic events

b. Partial Main Feedwatff Loss (one loop)

Bounded by total loss of MFW

c. Excessive Feedwater Not possible since MFW failed for all seismic events

d. Closure of One Main Steam Isolation Valve Model only seismic failure to close; valves do fail closed on loss of station air

e. Core Power Excursion Reactor trip always assumed required; Pressurizer PORV assumed challensed anyway

f. Total Loss of Primary Flow (one or more loops)

Pressurizer spray lost anyway due to assumed loss of containment air

g. Main Feedwater Line Break MFW assumed failed anyway; pressurizer PORV assumed challenged

16. Loss of Offsite Power Modeled in response to seismic event by acceleration denendent failure probabilitv in Top Event ZOG

17. Loss of One 125V DC Emergency Bus ISSGoneulting riRrzzo

2734294-R-036 Reaision 0 May 11-,201-7 Page 44 of145 TABLE 4.I REVIBW OF INTERNAL INITIATING EVENTS FOR APPLICABILITY TO SETSI\\ITC SEQUBNCES (coNTrNUEr))

Inrrmrruc Eyrnr CerrcoRrns Monnr,ruc or lNru.rroR ron SPRA

a. 125V DC Bus 2-I, Orange Screened on hieh seismic capaciff

b. l25V DC Bus 2-2. Pumle Screened on hieh seismic capacity

c. l25V DC Bus 2-5, Orange Modeled in response to seismic event by acceleration dependent failure probability in Top Event ZD5

18. Loss of Service Water and Standby Service Water

a. Loss of Service Water Header A Modeled in response to seismic event by acceleration dependent failure probability in combinations of Top Events ZS2,ZSW,ZF*Z, andZB

b. Loss of Service Water Header B Modeled in response to seismic event by acceleration dependent failure probability in combinations of Top Events ZS2, ZSW, ZRz, and ZR3

c. Loss of Both Service Water Headers Modeled in response to seismic event by acceleration dependent failure probability in combinations of Top Events ZS2, ZSW. ZR2. and ZR3

19. Total Loss of Primary Component Cooline Water Assumed failed for all seismic events

20. Loss of One Vital Instrument Bus

a. Loss of Red Vital Bus Screened on hieh seismic capacity

b. Loss of White Vital Bus Screened on high seismic capacitv

c. Loss of Blue Vital Bus Screened on hieh seismic capacity

d. Loss of Yellow Vital Bus Screened on high seismic capacity

21. Loss of One 4.16-kV Emergency Bus

a. Loss of 4.16-kV Bus 2AE, Orange Modeled in response to seismic event by acceleration dependent failure probability in combinations of Top Events ZOG, ZDs, ZDG, and ZR4A

b. Loss of 4.16-kV Bus 2DF, Purple Modeled in response to seismic event by acceleration dependent failure probability in combinations of Top Events ZOG. ZDG. and ZR4B

22. Loss of aNon-Emergency Bus

a. Loss of 4.16-kV Bus 2,4.

Grouped as failing whenever seismically fail offsite power via ZOG

b. Loss of 4.16-kV Bus 2D Grouped as failing whenever seismically fail offsite power via ZOG

23. Loss of Station Instrument Air Assumed failed for all seismic events

24. Loss of Containment Instrument Air Assumed failed for all seismic events lSSGonsulting rlRt77.o

2734294-R-036 Raision 0 May 1,1,2017 Page 45 of 1.45 The details for the development of the final SEL can be found in Reference 32. Discussions are provided therein regarding items such as corrmon-cause failure events, Human-Action related basic events and fire and flooding scenarios. Further in20l6, a model update included new basic events to represent the diverse and flexible mitigation strategies (FLEX). The added SSCs were included in the BVPS-2 SEL.

4.1.1.2 Additional SSCs Included in the SEL Consistent with the ASME Standard (Reference 4), the BVPS-2 IPEEE documentation (Reference 28) and Updated Final Safety Analysis Report (UFSAR) (Reference29, Tahle 8.1 l) were first reviewed to identiff plant structures that should be added to the BVPS-2 SPRA SEL.

Such passive SSCs were not included in the internal events PRA models but are of special interest for SPRA. A total of 13 Seismic Category 1 structures were added. Seismic Categary 2 and non-seismic structures (5 in all) were added if they housed SSCs already on the list. The following structures are included in the BVPS-2 SEL:

. Auxiliary Building (AXLB) r Reactor Containment Building (RCBX)

. Diesel Generator Building (DGBX)

Fuel Handling and Decontamination Building (FULB)

Service Building (SRVB)

. Main Steam and Cable Vault (MSCV) o Control Building (CNTB) r Safeguards Building (SFGB)

. Intake Structure (INTS)

. Primary Plant Demineralized Water Storage Tank Pad and Enclosure (PPDWST)

Refueling Water Storage Tank/Chemical Addition Tank Pad (RWST/CAT)

Service Water Valve Pits (VLVP)

Pipe Trenches (PIPETRENCHES)

. Alternate Intake Structure (AISX)

Emergency Response Facility Substation (ERFS)

I Emergency Response Facility (ERF) Diesel Generator Building (RSGB)

. Turbine Building (TRBB)

Cooling Tower Pump House (CTPH)

These Category 2 and non-seismic structures were considered further for BVPS-2 only when the fragility analysts determine whether they are likely to survive earthquakes that contribute to risk.

No SSCs within the Unit 2 turbine building are credited in the SPRA and therefore not included in the SEL walkdown. The ERF RSGB fragility was evaluated in the earlier SPRA for BVPS-2 and so, notwithstanding the aforementioned, was retained on the SEL for potential walkdown.

ReferenceS2 outlines other passive SSCs added to the BVPS-2 SEL such as nuclear steam supply system (NSSS) components, block walls, polar crane, and piping segments, among many others. The basis for including these additional passive SSCs is also provided in Reference 32.

In addition to adding passive equipment and structures, alternative lists of SSCs for the SEL were considered. These included SSCs such as those associated with the occurrence of a very lESGonsulting

()Rtzzo

2734294-R-036 Reaision 0 May 1L,2017 Page 46 of 1.AS small LOCA as well as those associated to LERF. A review of the SEL for both Diablo Canyon and Surry was performed to identiff potential additions of non-passive SSCs into the BVPS-2 SEL. The complete list of additional non-passive SSCs is provided inReference 32 along with their basis for inclusion into the BVPS-2 SEL.

4.1.2 Relay 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 BVPS-2 SPRA, in accordance with SPID (Reference}), Section 6.4.2 and ASME/ANS PRA Standard (Reference 4), Section 5-2.2. The evaluation resulted in most relay chaffer scenarios screened from further evaluation based on no impact to component function. 167 relays did not screen based on relay chatter evaluation; however after fragility analysis 53 relays have high confidence of a low probability of failure (HCLPF) greater than the screening HCLPF for inclusion into the PRA (i.e., they all screen based on seismic capacity). The remaining relays are included in the PRA model as described in Section 5..I.3. It should be noted that some relays did not screen based on seismic capacity until after the peer review in which the relay fragilities were refined to remove excess conservatisms documented in the Peer Review Report.

For presentation of results circuit breakers and contactors that did not screen are addressed separately from the above relays. All circuit breakers screened from inclusion in the PRA, mainly from high capacity. Contactors identified through circuit analysis were evaluated for the GERS fuirction during failure mode of the MCC that the contactor is housed in. 7 MCC cabinets did not screen from inclusion into the PRA model based on seismic capacity and were included in the SPRA model as described in Section 5.7.3.

The specific SSCs potentially affected by chatter of these relay types and how they are modeled in the PRA are summarized in Section 5.6 of Reference 38.

4.2 WnlxnowN Arpnoncn This sectionprovides a summary of the methodology and scope ofthe seismic walkdowns performed for the SPRA. Walkdowns were performed by persorulel with appropriate qualifications as defined inthe SPID. Walkdowns of those SSCs included onthe seismic equipment list were performed as part of the development of the SEL, and to assess the as-installed condition of these SSCs for use in determining their seismic capacity and performing initial screening.

Walkdowns were performed in accordance with guidance in SPID Section 6.5 (Reference 2) and the associated requirements in the PRA Standard (Reference 4).

lESConsulting rlRtzzo

2734294-R-A36 Rrr:rsion 0 May 11,20L7 Page 47 of 1,45 Several SEL items were previously walked down during the BVPS-2 Seismic IPEEE program.

Those walkdown results were reviewed and the following steps were taken to confirm that the previous walkdown conclusions remained valid.

. A walk-by was performed to confirm that the equipment material condition and configuration is consistent with the walkdown conclusions and that no new significant interactions related to block walls or piping attached to tanks exist.

. If the SEL item was screened out based on the previous walkdown, that screening evaluation was reviewed and reconfirmed for the SPRA.

For some SEL SSCs walkdowns had recently been performed in support of resolution of NTTF 2.3 seismic (Reference 14) and the Expedited Seismic Evaluation Process (ESEP)

(Reference 8), and information from those walkdowns was used where the appropriate level of detail needed for the SPRA was available.

The seismic walkdowns for equipment outside of the RCBX were performed from January 2l through February 1,2013. Seismic walkdowns of equipment in RCBX were performed May I artd2,20l4, during a station refueling outage. A supplemental walkdown was performed on May 30,2014, to further evaluate potential seismic-induced fire and seismic-induced flood. A second round of supplemental walkdowns was performed on February 8,2016, and February 29, 2016, to address findings and obsenrations from the December 2014 SPRA peer review, evaluate recently installed FLEX equipment, and assess the lines connected to the spent fuel pool.

The following paragraphs summarTze the preparation, procedure, and findings of the seismic walkdowns.

Structures, Systems and Components Walkdown The BVPS-2 SEL consisting of approximately 1,950 SSCs was reviewed, analyzed, ffid then reduced to about 750 for walkdown and walk-bys. In addition to selecting representative samples of similar equipment, about 470 check valves and 220 penetrations were excluded as being seismically robust. Approximately 260 SSCs were excluded as being housed within other SSCs that were walked down, and 100 SSCs in the TRBB were excluded since this is a lower capacity structure. An additional 140 components were excluded from walkdowns since they are not curently modeled in the SPRA. These components generally correspond to non-seismic or Seismic Category II systems. 1l SSCs on the SEL correspond to NSSS components. These items were not walked down, but fragility parameters were developed for them based on available drawings and calculations.

The BVPS-2 SEL also includes items needed to maintain containment (CTMT) functions. The RCBX and equipment that support the CTMT functions, and systems required for CTMT performance (e.g., CTMT fan coolers and CTMT isolation valves) were included in the walkdown list, as well as targeted for fragility analysis.

Table 4-2 presents the number of Walkdown components sorted in accordance with the EPRI Equipment Classes. Equipment Class I through Class 2l are assigned consecutively based on the SQUG/Generic Implementation Procedure (GIP) Walkdown Seismic Evaluation Work lESGonsuEing

()R.zzo

2734294-R-036 Reaision 0 May 1.1.,201.7 Pase 48 of145 Sheets (SEWS). Class number (0) is assigned to the remaining components in the Walkdown SEWS as "other" components.

TABLE 4-2 BREAKDOWFI OF EQUIPMENT WALKDOWI\\ LIST BY EQUIPMENT CLASS EPRI Cuss DnscnrprloN CouNr 0

Other 7l I

Motor Control Centers 13 2

Low Voltage Switchgear 7

aJ Medium Voltage, Metal-Clad Switchgear 2

4 Transformers 13 5

Horizontal Pumps 2t 6

Vertical Pumps 15 7

Pneumatic-Operated Valves 84 8A Motor-Operated Valves 13r 8B Solenoid Valves 18 I

Fans l5 r0 Air Handlers 4

ll Chillers 2

t2 Air Compressors J

l3 Motor Generators 0

l4 Distribution Panels 24 l5 Battery Racks 7

16 Battery Chargers And Inverters l3 l7 Engine Generators 6

l8 Instrument (On) Racks 29 19 Temperature Sensors 6

2A Instrument And Control Panels r87 2t Tanks And Heat Exchangers 3t Structures and Distribution Systems 24 SEL Total 726 Walkdown Seismic Review Team The seismic walkdowns were conducted by two Seismic Review Teams (SRT). Each Team was composed of at least two Seismic Capability Engineers (SCE) along with BVPS-2 Station personnel. All of the key individuals performing the walkdowns completed the 1-week walkdown training sponsored by SQUG. In addition, SCEs possess technical degrees with a structural/seismic background and nuclear-related experience. Furthermore, Mr. Farzin Beigi provided continuous support and expert input to each walkdown team throughout the full extent of the station walkdowns, as well as post-walkdown discussions to ensure consistency between walkdown teams.

lESGonsulting

()Rtz71o

273429+R-036 Reaision 0 May 11,2017 Page 49 of 1.45 Seismic Evaluation Walkdown Procedures Prior to the walkdown, the SEL comprising the full scope of the seismic evaluations was reviewed by the SRT and Station Personnel. For the purpose of the equipment walkdown, the SEL was divided into mechanical and electrical (M&E) equipment and distribution systems. The locations of strucfures and components were determined from the station layout drawings. The walkdown sequence, including coordination with station operations, schedule, and route was developed to minimize affecting station operations.

The Walkdown of the SEL items was accomplished in two phases. The first phase was devoted to components that could be examined during normal station operation, while the second phase was planned for the remaining components accessible only during the station outage.

Inaccessible components are addressed by inspection of photographs and existing design analysis documents.

Walkdown of Structures The information required to develop structural fragilities is obtained primarily from design drawings. The seismic walkdown of the structures was limited to verification of the structural location, overall configuration, gross dimensions, and building separation, and any signs of degradation and distress.

Walkdown of Equipment and Distrihution Systems The seismic walkdown of the BVPS-2 M&E equipment was performed in accordance with the methodology of SQUG/GIP and EPRI NP-6041-SL (Reference 7).

The component-specific SQUG/GIP SEWS were utilized to record walkdown observations.

Unlike the SQUG/GIP, the focus here was not to perform screening, but rather to document the specific sets of inclusion/exclusion rules or caveats and cofirmon bases in accordance with prescribed checklists so that the experience-based HCLPF in EPRI NP-6041-SL (Reference 7) can be supported.

The distribution systems comprising of piping, ducting, and cable trays were walked on a sampling basis, reflecting the industry experience that the distribution systems components generally perform well in a seismic event. The sample set of piping, heating, ventilation, and air-conditioning (HVAC) duct and cable trays segments represent the essential distribution systems in the BVPS-2. In general, the observations related to distribution systems focused on seismic vulnerabilities posed by potential excessive differential rnotion between structures and poor design of supports and their anchorage.

The walkdown procedures for different types of components are described in detail in the BVPS-2 Walkdown Report (Reference 40).

tEEConsulting rlRtzzo

273429+R-036 Reaision 0 May 1L,20L7 Pase 5A of H5 Additional Walkdown C onsiderations In support of the plant walkdown, some added lists were developed for inspection by the walkdown team. Three general areas were considered:

Operator Action Locations Fire Ignition Sources Potential Flooding Sources Human failure events (i.e., models of operator actions) were identified inthe BVPS-2 Internal Events PRA. The BVPS-2 SEL Development Report (Reference 32) provides a sunmary of the room locations and path ways needed for recovery action following an earthquake. Excluding the control room and the turbine building, which is assumed failed, a total of nine (9) unique locations were identified where credit for operator actions performed outside the control room is taken. These locations were assessed as part of the human reliability analysis to determine which of these locations are likely to be accessible by the operators following a substantial earthquake.

A listing of the transit routes for actions performed outside the control room is provided in the human reliability analysis notebook (Reference 36). Veriffing that the locations were accessible helped assure that the actions credited in the internal events PRA were still feasible even considering the potential equipment failures that may occur following an earthquake.

Potential f,rre ignition sources were routinely evaluated by the walkdown team. These sources may coincide with SSCs on the seismic list or be in close proximity to SSCs that are on the list.

Only those plant locations evaluated during the walkdown were considered because they contain SSCs on the SEL. However, to provide soms assurance that potential sources were not overlooked, the walkdown team performed two informational searches focused on: (l) potential ignition sources involving flammable liquids and piping containing hydrogen or oil, and (2) electrical equipment that could be the source of a seismic-induced fire but are not already on the SEL.

To assist the plant walkdown, a list of potential flooding sources that should be considered during the walk-by was also developed. This list consisted of fire protection system piping, which is maintained "wet" during plant operation and tanks and coolers represented in the initial BVPS-2 Internal Floods PRA. The list of potential flooding sources is presented inthe BVPS-2 SEL Development Report (Reference 32).

4.2.1 Significant Walkdown Results and Insights Consistent with the guidance fromNP-6041 (Reference 7), no significant findings were noted during the BVPS-2 seismic walkdowrs. Note that previous walkdowns for the NTTF Recommendation 2.3 did identiff adverse conditions that were documented with their dispositions in a separate submittal (Reference 14).

Components on the SEL were evaluated for seismic anchorage and interaction effects in accordance with SPID guidance (ReferenceZ) and ASME/ANS PRA Standard (Reference 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, the potential for seismic-induced fire and flooding scenarios was assessed. Potential lBSConsulting tlRtzzo

273429+R-036 Reuision 0 May 1.1.,201.7 Page 51of 1.45 internal flood scenarios were incorporated into the BVPS-2 SPRA model. The walkdown observations were adequate for use in developing the SSC fragilities for the SPRA.

4.2.2 Seismic Equipment List and Seismic Walkdowns Technical Adequacy The BVPS-2 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 (Reference 4).

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the BVPS-2 SPRA SEL and seismic walkdowns are suitable for this SPRA application.

4.3 Dyu.r.puc Amalysrs oF SrnucruRns This section sunmarizes the dynamic analyses of structures that contain systems and components important to achieving a safe shutdown, using fixed-base an#or Soil Structure Interaction Analyses (as applicable). The section describes the methodologies used, discusses responses at various locations within the structures and relevant outputs, important assumptions and sources of uncertainty.

4.3.1 Fixed-BaseAnalyses No structure at BVPS-2 was analyzed using a fixed based methodology; i.e., SSI was performed for all structures analyzed for the SPRA. Note, however, that fixed-base analyses were performed as verification and validation step in the development of the SSI models, as described in the BVPS-2 Building Seismic Analysis Report (Reference 43).

4.3.2 Soil Structure Interaction (SSI) Analysis The building seismic analysis for BVPS-2 addresses the effects of SSI on the seismic response of the building structures. This analysis accounts for the foundation mat flexibility and its interaction with the flexibility of the supporting geotechnical medium. Both kinematic interaction due to the foundation mat stiffness and inertial interaction due to its mass are accounted for. The seismic incident waves are assumed to propagate vertically in the form of shear waves producing horizontal ground motion and compression waves producing vertical ground motion. Because the solution to the equations of motion is obtained in the frequency domain, the SSI analysis is linear. Strain-compatible soil properties obtained from the site response analysis (Reference23) are used in the analysis without further modification.

The SSI analysis for BVPS-2 structures utilizes RIZZO' s version of the System for Analysis for Soil-Structure-Interaction (SASSI) Program. This version is based on the original SASSI developed in the 1980s at the University of California, Berkeley (Refercnce 42).

The mean (BE) HCSCP are used in the SSI analyses. Although the site response analysis also develops mean-o (lower bound) and mean+o (upper bound) HCSCP, these are not considered in obtaining the seismic response used in the fragility analysis. Rather the effects of the SSI stiffness variation on the seismic demand are incorporated by peak shifting in accordance with the methodology in EPRI 103959 (Reference l l) and EPRI 1019200 (Reference 44). The justification for this approach is discussed in Reference 43, and Calculation l2-4735-F-140 (Reference 45), and summarized as follows:

lESGonsulting riRtzzo

2734294-R-036 Reaision 0 May 1L,20L7 Page 52 of145 r For a given input spectrum shape, the deterministic analysis with conservative structure and soil damping and BE structure and soil stiffiress results in approximately 80 percent non-exceedance probability response which achieves the targeted demand consenratism for conservative deterministic failure margin (CDFM) evaluations.

r The American Society of Civil Engineers (ASCE) 4-98 (Reference 46) procedure of enveloping of lower bound (LB) and upper bound (UB) response and peak shifting provides a consewative design basis response for use in the seismic qualification of multi-mode subsystems. These procedrxes are conservatively biased and are consequently not used for fragility analysis.

o The LB and UB responses do not represent reasonable median-centered values. The use of LB and UB does not result in a CDFM value representative of I percent probability of failure on the composite fragility curve. If LB or UB response is used, then the Bc may need to be re-examined so that the conditional failure probabilities are consistently described in quantification.

The ground motion inputs to the building seismic analysis are represented by a set of time histories (two horizontal and one vertical), each matching the appropriate FIRS (hereafter called the FIRS time histories). The FIRS time histories are based on seed (recorded) time histories selected based on similarity of their response spectral shapes to the spectral shapes of the FIRS.

The seed time histories are conditioned to obtain FIRS time histories whose response spectra closely match the FIRS. This process implements the guidance in Reference 24 and Reference 82.

Selected records are checked to ensure that they meet criteria established by the NRC regarding the adequacy of time histories. Based on the Reference 82 the strong-motion duration is defined as the time required for the Arias Intensity Reference 83 to rise from 5 to 75 percent (D5-75).

The uniformity of the growth of this Arias Intensity is reviewed. The minimum acceptable strong-motion duration should be 6s.

Prior to being used as input to seismic structural analyses, the seed time histories must be conditioned to match the FIRS. Spectral matching analysis is performed to generate spectral-compatible acceleration time histories using the spectral matching computer program, RspMatch09 (References 84 and 85). RspMatch09 uses a time domain spectral matching method, where adjustment of initial time series (seed motions) is made by adding wavelet functions to the initial acceleration time history in the time domain. This adjustment is repeated until its response spectrum becomes comparable to the target spectrum over the desired frequency range.

Spectral matching analysis is performed by running RspMatch09 multiple times, which is specified in the RspMatch09 input file. The output file from the last run is used to conf,rrm that the adjusted time histories meet the criteria stated in Reference 24 artd Reference 82.

To confirm that there is no significant gap in the smoothed power spectral density (PSD) of the matched time histories, the computed PSD are compared to the minimum PSD requirement of lESGonsulting tlRtzzo

273429+R-036 Reaision 0 May 11,,20L7 Pase 53 of 1.45 Reference 82 which refers to the M and R bins from Reference'17. To comply with Reference 82 the minimum PSD are compared to 80Yo of minimum PSD in the frequency range of 0.3-24112.

The full suite of time history information can be found in Reference 23.

The time histories described above and used as input to the building seismic analyses match the FIRS presented in Revision I ofthe BVPS PSHA/FIRS Report. They are not modified to match the FIRS presented inSectbn 3.1.4 fromRevision 4 of the BVPS PSHA/FIRS Report on the basis that the shapes of the FIRS utilized in the building seismic analysis reported in Reference 43 arc very similar to those of the FIRS presentedin Section 3.1.4. Figure 4-1 compares the RCBX horizontal spectra normalized to the RCBX PGA. The comparison illustates that the difference in the horizontal FIRS is relatively insignificant. However, the comparison on Figure y'-2 shows that the vertical specta are diminished in excess of l0Yo in the frequency range of about 8 Hz to 15 Hz because they are now based on the mean of the V/tI ratios where previously, the envelope was used.

1.00 RCBX H-FIRS Damp=5%

/

7

\\

\\

) /

\\ \\ \\

./

7 0.10 1.00 10.00 100.00 Frequency (Hz)

-Rev 1

-Rev 4 Scaled to Rev 1 PGA I.IGURE 4.1 COMPARISON OX'RCBX HORIZONTAL FIRS NORMALIZED TO PGA OX'0.24G fSCffiuttng (IFrzzg u0 Y(,(,

-.E LP(Joc Itl 0.80 0.60 0.40 0.20 0.00

2734294-R-036 Reoision 0 Moy 11,2017 Pase 54 of 145 RCBX V-FIRS Damg = 5%

\\

1.0 0.8

.+

u0 E o.o

-tE L6 0.4 o

CL vl 0.2 0.0 0.1 1.0 10.0 100.0 Frequency (Hz)

ReVl,-

ReV4 X'IGURE 4.2 COMPARISON O['YERTICAL FIRS AT RCBX F'OI]I\\DATION LEVEL Thus, the vertical direction ground motion time histories used in the building seismic analysis are conservatively biased. This conservative bias is justified on the basis that the fragilities of most of the SSCs are controlled by horizontal response, and are therefore not expected to be impacted significantly. However, when controlled by the vertical in-structure response spectra (ISRS),

fragilities could be improved on a selective basis (e.g., relay fragilities). Additionally, the bias is retained to allow for uncertainties in the regulatory acceptability of using mean V/[I ratios instead of the envelope. This is further discussed and justified in the Fragility analysis Report (Reference 4l).

Details of the SSI analyses are provided in the BVPS-2 Building Seismic Analysis Report (Reference 43).

A list of structures and descriptions of dynamic analysis approaches are presentedinTable 4-3 lBSGsrsulting

()Rrzzo

2734294-R-036 Reaision 0 May 11-,201,7 Page 55 of145 TABLE 4-3 DESCRIPTION OF STRUCTURES AND DYNAMIC ANALYSIS METHODS FOR BVPS.2 SPRA 4.3.3 Structure Response Models Details of the structural response models development are provided in the BVPS-2 Building Seismic Analysis Report (Reference 43). The following subsections surrmaize the evaluation of existing lumped-mass stick models, analytical modeling procedure, and structure material properties, stiffness, mass and damping.

4.3.3.L Evaluation of Existing Lumped-Mass Stick Models The design basis seismic analysis of the BVPS-2 structures utilized lumped-mass stick models (LMSM). These models representthe entire mass of afloor slab concentrated at one point. The point masses are then connected with a beam or "stick" representing the respective story stiffness. These models are typical of the prevailing practice when BVPS-2 design was performed.

P.IZZO assessed the acceptability of using stick models in the SPRA project in light of the ASME/ANS requirements (Reference 47). The Report compares ISRS obtained using stick models to the ISRS based on independently developed finite-element models (FEM) for three representative buildings of the Davis-Besse Nuclear Power Station; namely Auxiliary Building Area 7, Reactor Building's Internal Structure, and the Reactor Shield Building.

lESGonsulting

[iRrzzo SrRucruRn Founo,+rron ConurroN Trryn OF Monrl An,+.lysrs MnrHon Couuuurs/Ornnn IivroRuarIoN Auxiliary Building Soil FE Deterministic SSI BE Case, 1 set of T-H in accordance with ASCE 4-98 Reactor Containment Building Soil FE Deterministic SSI BE Case, 1 set of T-H in accordance with ASCE 4-98 Diesel Generator Building Soil FE Deterministic SSI BE Case, I set of T-H in accordance with ASCE 4-98 Fue[ Handling / Decon Buildings Soil FE Detenninistic SSI BE Case, I set of T-H in accordance with ASCE 4-98 Service Building Soil FE Deterministic SSI BE Case, I set of T-H in accordance with ASCE 4-98 Main Steam & Cable Vault Building Soil FE Deterministic SSI BE Case, I set of T-H in accordance with ASCE 4-98 Conffol Building Soil FE Deterministic SSI BE Case, 1 set of T-H in accordance with ASCE 4-98 Safeguards Building Soil FE Deterministic SSI BE Case, I set of T-H in accordance with ASCE 4-98

2734294-R-036 Reaision 0 May L1.,201.7 Paxe 56 of145 Based on the comparisons of the ISRS, the Report concludes that the ISRS from the FEMs are not enveloped by the ISRS from the existing stick models over the entire range of frequencies of interest. However, improvements to the existing stick models to include appropriate representation of flexural stiffness, mass eccentricities, and rigid body rotations may result in acceptable response results.

Because of the significant effort expected to upgrade the existing stick models coupled with the possibility of such models being challenged, the study reported here develops new analytical models based on the FE method. These models represent state of the current practice. However, as a global verification, the total masses used in stick models have been compared to the values represented in the corresponding FE models. The differences are smaller than 10 percent.

4.3.3.2 Development of FE Structure Response Models The building structure FEM are based on geometric information, such as building dimensions, wall and slab thicknesses, structural member locations, and size of openingso etc., taken from building structure layout drawings and details. The parametric information, such as the material properties, live loads, equipment loads, and boundary conditions are obtained on the basis of drawings, existing reports, ffid appropriate codes and standards.

Figure 4-3 presents the generic flow chart describing the procedure utilized to develop and check the FEMs. The structural FEMs are suitably modified for use with the program SASSI in the seismic SSI analysis.

The modeling effort for the building structure starts with the preparation of three dimensional (3-D) drawings representing the building geometry using software with a graphical interface, such as AutoCAD or zuSA. This step develops the geometrical representation of the structural components of the building, such as the foundation and floor slabs, walls and openings, and defines the mid-planes of floors and walls. The geometric model is imported into SAP2000 for FE meshing, assigning element types, ffid material characteristics in support of developing the structural model. Loads, boundary conditions, and any other special analytical requirements are then incorporated to complete the analytical models.

Most of the building structures which house equipment are analyzed using models which represent the building geometry as described above, as well as the dynamic seismic interaction with the supporting geotechnical medium. The models are sufficiently representative to extract seismic forces on the strucfural components and to develop the ISRS at locations of interest for use in the analysis of the equipment supported in the buildings.

lESGonsulting

()Rrzzo

273429LR-036 Reoision 0 May 11,2017 Page 57 of 145 Station Sffuctural Drawings Create 3D Geometry with graphical aids Generate Finite Element Mesh Assign model parameters:

thickness, mass, elastic properties, damping, boundary conditions Finite Element Model Developed NOT OK Verify Finite Element Model: 1-G analyses Mode-Fequency Analysis, Reaction Forces (S4P2000)

FE Model Ready to mnduct Seismic OK FLow cHARr rr r "f,1t#Hrtlr"LopMENr ox' r'EM 4.3.3.3 Material Properties & Structure Stiffness and Mass The building seismic analyses are performed using the best estimate values of structure stiffrress and mass, the BE subsurface Vs profile compatible with the expected seismic shear strains, and "conservative estimates of median damping.' In accordance with ASCE 4-98 @eference 46),

this approach is expected to develop approximately 84tr percentile seismic response suitable for use in the CDFM analysis.

Table 4-4 presents the general material properties of the materials of construction. Information on the structure specific design drawings is also utilized to confirm the material stength.

fSGonsulthu oRrz7[o

273429+R-036 Reaision 0 May 11,20L7 PaRe 58 of145 TABLE 4-4 STRENGTH AND ELASTIC PROPERTIES OF'MATERIALS OF CONSTRUCTION BEAVER VALLEY POWER STATION. UNIT 2 STRUCTURES M,+,rrRr.q,r, CoxsrnucrroN Srnmqcrn Elnsrrc Mouur,us Potsson's R q,rto Concrete Auxiliary Building Reactor Containment Building Diesel Generator Building Fuel Handling and Decontamination Building Service Building Safeguards Building Control Building Main Steam & Cable Vault f*': 3000 psi 3.1x 106 psi 0.25 Rebar ASTM 4615, Gr 60 No. 3 to No. 18 Fy : 60 ksi 29.Ax103 ksi 0.30 ASTM A 36 -

Structural Structural shapes, system supports, component supports Fy: 36 ksi 29.0x103 ksi 0.30

Reference:

BVPS-2 UFSAR (Reference 29)

The values of the Yourg's Modulus in Table 4-4 arc generally in agreement with those based on ACI 349-06 (Reference 48) fornormal weight concrete (8, = 57,000#). The value of the Poisson's ratio is taken to be 0.25 so that the concrete shear modulus Gc : 0.4 Ec, which is consistent with ASCE Standard 43-05 (Reference 49). A unit weight of 150 pounds per cubic foot (pcfl has been adopted for analyses. This value corresponds to normal weight concrete used in the building construction. Consistent with the expected Response (damage) Level, fulI or effective stiffrresses are used for concrete members recommended in ASCE/SEI 43-05 (Reference 49) as shown in Table 4-5.

TABLE 4.5 EFFECTIVE STIF'F'NESS OF REINFORCBD CONCRETE ELEMENTS (REFERENCE 4e) lr{erttlrer Flexural Rigidity Shear Rigidity AxialRigidity Beants-Nonprestressed Beams-Prestressed Colulrrns in courpression Columns in lension

\\Val ls nnd diaphmgrrts-Lhrcrncked 0.5 EJs EJ*

0.7 EJs o.s tr4 E"lo

{"fr <.f.,}

0,5 E+

lfi> I,,l C.An' 6.An' G"An' G"Au' G.Ag,

{H ( lr.}

0.5 G.A$'

(y> y"l Edtg Eltn E.As Ec'{s

\\Ual Is artd diirpltragnrs{racked lESGonrulting (iRtzzo

273429+R-036 Rqtision 0 May 1.1.,201.7 Page 59 of145 The shear stiffrress of walls and diaphragms is represented assuming cracked section properties (Table 4-5) for in-plane shear. Subsequent to the SPRA quantification, a selected sample of the shear walls for the plant structures was assessed to confirm the assumption. This assessment shows that the shear demand corresponding to the median failure capacities of controlling SSCs (HCLPF of about 0.5g PGA) exceeds the concrete shear capacity, which is 2 {f* in accordance with ASCE 43-05. The assessment shows that most walls are cracked.

4.3.3.4 Structural Damping Dynamic analyses of BVPS-2 structures use aconcrete structural damping of 4 percent of sritical for concrete members and 2 percent for steel structural members. This level of damping considers that the buildings will enter only into Response Level I as defined in ASCE/SEI 43-05 (Reference 49). An assessment of damage state in accordance with ASCE/SEI 43-05 (Reference 49) for a selected sample of walls shows that most walls remain in Response Level I.

4.3.4 Seismic Structure Response Analysis Technical Adequacy The BVPS-2 SPRA Seismic SSI Analysis and the Structure Response were subjected to an independent peer review against the pertinent requirements in the PRA Standard (Reference 4).

The peer review assessment, and subsequent disposition of peer review F&Os, is described in Appendix A, arrd establishes that the BVPS-2 SPRA Seismic Structure Response and SSI Analysis are suitable for this SPRA application.

4,4 SSC Fnq,crlrrY ANALYSTS The seismic fragility analysis develops the probability of SSC failure for a given value of the PGA. The fragilities are developed for all of the SSCs that participate in the SPRA accident sequences and included on the SEL. The fragility analysis for the significant risk contributors is particularly based on plant-specific information, and actual current conditions of the SSCs in the plant, as confirmed through the detailed walkdown of the plant, so that the resulting fragility estimates are realistic.

This section summarizes the fragility analysis methodology, presents a tabulation of the fragilities (with appropriate parameters, 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 5.0). Important assumptions and important sources of uncertainty, and any particular fragility-related insights identified, are also discussed.

4.4.L SSC Screening Approach In the context of a SPRA, high capacity components may be screened if their HCLPF capacity is in excess the PGA at very low exceedance frequency (e.g., 2x10-7) on the site-specific hazard curve. The items screened out in this manner require no further fragility analysis as the screening level capacity already contributes negligibly to the CDF. However, the associated screening level at the Beaver Valley site is relatively high, and very few items can be screened out.

A more appropriate screening level is established on a quantitative basis so that the maximum possible increase in CDF/LERF that can be added from accelerations greater than the Screening Threshold does not exceed 2x10-7 to CDF, or 1xl0-8 additionto LERF. This quantitative lEGonsulting rlRtzzo

273429+R-036 Reoision 0 May L1,2017 Page 60 of 1,45 approach uses the CDF/LERF interval success frequency (i.e., the hazard frequency which does not go to core damageflarge early release) and results in a Screening Threshold of 0.7g for excluding SSCs from the Level I PRA model and 2.0g for excluding SSCs from the Level 2 PRA model.

The screening strategy implemented for the BVPS-2 Fragility Analysis is based on the following considerations and is supported by the walkdowns:

The screening is based on the site-specific seismic hazard for the BVPS-2.

r The fragility analysts focus most of their analytical resources on equipment likely to govern the seismic risk, and to minimize their efforts on more robust equipment, or on equipment judged seismically so weak as to not provide any benefit.

r To demonstrate that all seismic risk contributors to CDF and LERF are eventually included in the SPRA, the final screening criterion was adjusted upward based on the fragility estimates for evaluated equipment. The intent is to show that at most, the equipment not evaluated in detail contribute no more than three to four percent to CDF or LERF.

I Sensitivities were performed on fragile components to assess the impact of possible refinement of fragilities. This is documented in the quantification notebook Section6.3.2, Group 6 of sensitivities (Reference 17).

r SPRA is expected to be used in the future for making risk-informed decisions.

For this pulpose, it is useful to keep in the system model all the components whose failure may lead to some important accident sequences. In this w&y, one could judge the impact of upgrading any particular component or even relaxing the test frequency requirements. If the component is screened out and not in the model, the analyst would have to introduce the subject component into the SPRA model for future risk-informed applications.

Where appropriate, the SRT used caveats inthe screening tables in EPRINP-6041-SL (Reference 7) to justify assigning the respective screening level capacities to high seismic capacity components.

The general approach classifies equipment on the SPRA SEL into ranges of HCLPF capacity so as to identifu a set of equipment that are seismically strong enough to mitigate risk, yet not so strong that they do not contribute to seismic CDF and LERF. The approach used is as follows:

1.

Initially screen from fragility analysis all SSCs that are not Seismic Category 1, as being seismically weak.

Screen out all Seismic Category I SSCs that are judged seismically no stronger than the fragility for loss of offsite power, again as being seismically weaki i.e., a HCLPF of 0.1g PGA.

IESGonsulting

{}Rtzzo 2.

273429+R-036 Reaision 0 May 1L,201.7 Page 61 of145 J

4 5

6 7

I Screen out rugged SSCs judged to have a seismic HCLPF greater than the screening level as being seismically robust and; therefore, potentially less likely to contribute to seismic CDF or LERF.

Evaluate the fragilities for the remaining Seismic Category 1 SSCs judged to have a seismic fragility with HCLPF's between 0.1g and the screening level.

Incorporate the evaluated fragilities in Step 4 above into an initial SPRA model to determine the seismic CDF and LERF as a function of seismic hazard level.

Subtract the CDF contribution from each seismic range from the seismic hazard frequency curve to obtain the remaining frequency of seismic events that do not result in core damage as a function of PGA. Identify the seismic magnitude in PGA, at which the adjusted exceedance frequency curve corresponds to 3 percent to 4 percent of the computed CDF. Repeat this step for LERF.

If the PGA values for maximum added seismic CDF and LERF obtained in Step 6 are less than the screening level then no additional SSC fragilities need be evaluated. A11 other unanalyzed SSCs have been shown to have seismic capacities greater than the screening level, or are seismically weak and not credited in the analysis.

If the PGA values corresponding to 3 percent to 4 percent of the computed CDF and LERF as derived in Step 6 are greater than the screening level, then additional SSCs should be evaluated. The choice of which SSCs are to be evaluated next is to be decided by discussions between the fragility analysts and the PRA analysts. Most likely SSCs selected from those initially judged to have HCLPFs greater than the screening level are to be evaluated next. The collaboration between the fragility analysts and PRA modeling team is to also consider how the initial contributors to CDF and LERF can be mitigated by SSCs not yet creditedi e.9., by SSCs screened because they were not Category 1. Afterthe fragility analyses of more SSCs, repeat Step 4 through Step 6 until the CDF and LERF PGA values in Step 6 are less than the screening level, or some higher acceleration level that the fragility analysts can justiff that all other SSCs meet.

The assignment of SSCs to ranges of HCLPFs is supported by EPRINP-6041-SL (ReferenceT).

Therein caveats are provided for equipment to meet in order to assign a generic seismic capacity.

The generic seismic capacity is based on seismic experience as well as results from prior SPRAs.

The screening level to be applied to BVPS-2 components that meet the EPRI caveats is t.8g SA per References 7,44, and 50. This screening level capacity is aHCLPF capacity level and assures the survival of the equipment and function after the earthquake. Anchorage must be verified to also have a HCLPF capacity of at least l.8g SA.

lESGonsulting tlRtzzo

J 4

2734294-R-036 Raision 0 May 1-1-,20L7 Pase 62 of145 Fragilities of components, based onthe screening level HCLPFs, were developed as follows:

1.

The clipped peak of the 84ft percentile non-exceedance probability (NEP) spectra at the equipment location, or the SA at or greater than the lowest estimated/calculated/tested equipment frequency was compared to the 1.8g screening level to determine the ratio of the screening level to the 84ft percentile NEP demand.

2.

The HCLPF of the component was determined as the ratio in Step I times the site-specific Control Point PGA; i.e., 0.249 PGA.

Anchorage HCLPF was determined in accordance with EPRI NP-6041-SL (Reference 7) procedures and using the 84th percentile NEP floor spectra as the demand.

The governing HCLPF was determined as the component screening level, or component's demonstrated test capacity or the anchorage capacity. If the component was subjected to seismic interaction effects, then the resulting HCLPF was the lowest HCLPF, including the HCLPF due to seismic interaction effects.

In accordance with the recommendations in Referenee 2 a generic composite uncertainty, pc, ranging from 0.35 to 0.45 was assumed.

6.

The median ground acceleration capacity of the screened component was calculated from the governing HCLPF as:

Am: HCLPF(*2'33+F.;

Bc was broken down into a Fn of 0.24 to represent randoilrness in the ground motion and response and Bu ranging from 0.26 to 0.38 to represent uncertainty in response and capacity per Reference 2 Table 6-2.

Based on the walkdown observations and past SPRA experience, we conclude the following:

r SEL items deemed to meet the l.8g SA limit can be assigned a generic seismic fragility.

t Manually-operated valves on the SEL are judged to have high seismic capacities. They were removed from the SPRA systems model.

For the SEL items not "screened out" specific seismic fragilities were developed using the design data and walkdown observations.

4.4.2 SSC Fragility Analysis Methodolory For the BVPS-2 SPRA, the following methods were used to determine seismic fragilities for SSCs included in the SPRA. Overall, fragilities of Seismic Category 1 structures were calculated following the separation of variables method whereas the remainder of SSCs not screened out was established using the CDFM method considering betas recommended in Table 6-2 of the SPID (Reference 2). The following subsections describe the implementation of the technical approach in developing the seismic fragilities for the BVPS-2 SSCs.

lESGonsulting tlRtzzo 5

273429+R-036 Rwision 0 May'1,1,20L7 Page 63 of 145 4.4.2.1 F'ragility Evaluation Standards and Guides The standards and guidelines used to develop the fragilities of SSCs are identified below.

l.

EPzu TR-I03959 "Methodology for Developing Seismic Fragilities," (Reference 1l).

2.

EPRI 1002988 "Seismic Fragility Application Guide," (Reference 50).

3.

EPRI 1019200, "Seismic Fragility Applications Guide Update," (Reference 44),

4.

EPRI NP-6041-SL, "Nuclear Plant Seismic Margin," (Reference 7).

5.

ASCE/SEI 43-05, "Seismic Design Criteria for Stnrctures, Systems, and Components in Nuclear Facilities," (Reference 49).

6.

ASCE 4-98, "Seismic Analysis of Safety Related Nuclear Structures," (Reference 46).

7.

EPRI 1025287, o'Screening, Prioritization and Implementation Details (SPID) for the Resolution of Fukushima Near-Term Task Force Recommendation 2.1 : Seismic,"

(Reference 2).

4.4.2.2 CDFM Method The CDFM method is described in detail in Reference 7. The CDFM HCLPF values are determined using the following expression:

HCLPF - Fs. Fp. PGA

where, Fr = Strength factor derived from the comparison between seismic demand (ISRS) and the conservative estimate of seismic capacity (e.9., GERS or design analysis),

Ft, = Inelastic energy absorption factor (taken as 1.0 for briule failure modes),

PGA = Peak ground acceleration of the BVPS-2 control point (i.e., Reactor Containment Building foundation at EL 681 ft) FIRS :0.24 times the acceleration of gravity (g)

The median capacity A* developed in terms of the CDFM approach was estimated by using the following equation:

A*= HCLPF ' rz'tt(filj

where, Fc = Composite logarithmic standard deviation due to randomness and uncertainty, The median capacity estimates, A* are developed using pc values recofirmended in Table 6-2 of the SPID (Reference 2) for various types of SSCs. These values are shown below in Table 4-6 along with the corresponding p, and p,, values.

lESGonsulting rlRtzzo

273429+R-036 Rwision 0 May 1L,20L7 Page 64 of145 Tyrr SSC Conrposrrn Ec Rannour En Uxcr,RrAINTY 0u Csoy"/ Crw Structures & Major Passive Mechanical Components Mounted on Ground or at Low Elevation Within Structures 0.3s 0.24 0.26 2.26 Active Components Mounted at Hieh Elevation in Structures 0.45 0.24 0.38 2.8s Other SSCs 0.40 0.24 0.32 2.s4 TABLE 4.6 RECOMMENDED LOGARITHMIC STANDARD DEYIATIONS FOR SSC (SPID GUIDELINE, TABLE 6-2) 4.4,2,3 Separation of Variables Method The direct method, using separation of variables, develops median capacity on the basis of the median factor of safety (FOS), F1s, which defines the relationship between A* and the value of the ground motion parameter corresponding to the analysis spectra (EPRI TR-l002988 Seismic Fragility Application Guide,2002, EPRI TR-103959 Methodology for Developing Seismic Fragilities):

Am: Frra X Anrs

where, F,u is the seismic safety fastor Anm is the peak ground acceleration (g)

For structures, Fu is defined by:

Fna:FnsxFc Fc is the seismic capacity factor defined as:

Fc:FsxFlr

where, Fs is a factor associated with strength

,F,, is a factor associated with ductility FRs, the structural response factor, was calculated by SOV as a combination of several factors that affect the seismic response:

Fe u = Ground Motion Factor, FD 'Damping Factor, FM - Modeling Factor, Fuc = Modal Combination Factor, Fra : Time History Simulation Factor, Fssl = Soil-Structure Interaction Factor, Fec = Earthquake Component Combination Factor, lEGonsulting rlRtzzo

2734294-R-036 Raision 0 May 1,1,2017 Pase 65 ofL45 Fan = Horizontal Direction Peak Response, Fvc = Vertical Component Response.

Thus, Fsn is defined as:

Ft* = FG* ' Fo' Fu ' Fuc' FrH'Fsst ' FEC' Fno' Fvc Combining the capacity and the response factors the overall median FOS is:

Fu: Fc. Fns F* : ffi2R c+F2R,Rs)l/2 Fu: (F2u,c+p2u,*s)l/2 4.4.2.4 Seismic Demand The FIRS developed in Referenc e 23 are of significantly different shapes than the design basis earthquake (DBE) Safe Shutdown Earthquake (SSE) response spectra. Therefore, scaling of the DBE seismic response and the floor response spectra was not considered adequate to obtain median capacities. Instead, the fragility calculations reported here are based on seismic re-evaluation of facility structures using the new evaluation basis earthquake ground motion.

This re-evaluation also updates the analytical models of the structures as described in Section 4.3.

The seismic demand on the plant SSCs (in terms of forces and moments on building structural components, and in-structure floor response spectra) is obtained on the basis of seismic soil-structure-interaction analysis of selected buildings as reported in the BVPS-2 Building Analysis Report (Reference 43). The seismic SSI analysis is performed following the methodology in ASCE 4-98 (Reference 46), and results in the approximate 84th percentile seismic demand.

For structure fragilities evaluated using the separation of variables approach, the median demand is obtained on the basis of the calculated 84th percentile NEP forces and moments resulting from the SSI analyses, ffid the median demand conservatism ratio factor from the equation in EPzu Report 1019200 (Reference 44). A seismic demand logarithmic standard deviation of 0.2 is used in the equation based on an interpretation of data presented as part of probabilistic SSI studies in literature (References 51 and 52). The resulting median demand conservatism ratio is 1.18.

The seismic demand on equipment is evaluated independently using the 84th percentile NEP floor response spectra at selected points close to the equipment support location. Unlike design analysis, the equipment response used in the CDFM approach is typically based on un-broadened ISRS and frequency shifting. EPRI NP-6041-SL (Reference 7) reconrmends the damping values to calculate the equipment seismic demand for use in the CDFM method. These damping values are presented here in Table 4-7.

lf8Gonsulting

[]Rlzzo

273429+R-036 Reaision 0 May 1.1,2017 Page 66 of 1.A5 TABLE 4-7 RECOMMENDED EQUIPMENT DAMPING FOR ANCHORAGE BASEI} ON EPRI NP.6O4I.SL EeurmENT TYPE Daprprxc Electrical Cabinets Bolted or Welded to Floor 5%

Light, Welded Instrument Racks 3%

Massive, Low-Stressed Components (Pumps, Motors) 3%

Piping s%

Cable Trays t5%

Fluid Containine Tanks - Impulsive Mode s%

Fluid Containing Tanks - Sloshing Mode 0.5%

4.4.2.5 Fragility Evaluation of Seismic Category I Structures The building structures listed below are included in the fragility analysis. The fragilities of these structures are based on new analysis using the separation of variables method previously summarized. The method is described in detail in EPRI TR-l03959 (Reference 1l). Other structures are evaluated on the basis of simplified analysis.

t Auxiliary Building (AXLB) o Reactor Containment Building (RCBX)

Diesel Generator Building (DGBX)

Fuel Handling and Decontamination Building (FULB)

Service Building (SRVB)

Safeguards Building (SFGB)

. Control Building (CNTB)

. Main Steam and Cable Vault (MSCV)

The seismic capacity of a structure is typically controlled by the capacrty of the shear walls, which are the primary lateral load resisting elements. Floor diaphragms are screened on the basis that the seismic margins for these components are generally higher than for the shear walls. The diaphragm shear develops only from the lateral forces on the floor, while the shear walls particularly near the base are subjected to lateral forces accumulated from the stories above.

Based on the typical floor slab thickness (two feet) and span configurations of the floor diaphragms of the BVPS structures, it is judged that their fragilities do not govern over in-plane shear or flexure fragilities of shear walls near the base.

Within each strueture, critical walls are selected for evaluation of fragility. Critical structural members are major walls which failure poses a potential failure of the structure. Yielding of minor walls is not a concern since loads in these walls will be redistributed to the major shear walls. Of these critical walls selected for evaluation, the one calculated to have the lowest safety factor is taken to represent the fragility of the building. Critical walls of a building are generally located at stories which exhibit the most significant inter-story drift based on the displaced shape of the structure under horizontal seismic loads. Typically, two or more floor levels of the building are considered where representative walls are evaluated. One is at the foundation level, lEEGonsulting rlRtzzo

2734294-R-036 Reaision 0 May L1,20L7 Page 67 of745 where the walls are expected to carry the largest shear forces accounting for the total base shear for the structures. A second story level is based on observable inter-story drift. This story is expected to introduce the largest shear deformations in the shear walls.

The fragility of a reinforced concrete wall reflects the strength of the wall accounting for the ultimate strength of the concrete, the yield strength of the reinforcing steel and the energy absorption as the component is cycled in the inelastic range.

The strength capacity calculations follow consensus codes and industry guides such as ACI 3 18 and EPRI 103959 to evaluate potential failure modes, such as diagonal shear cracking, flexure, and shear friction in walls. In general, the critical failure modes of concrete shear walls in Seismic Category I buildings of the BVPS-2 are diagonal shear and flexure. Shear friction is not considered to be a credible failure mode for the BVPS shear walls. This is because there are either no horizontal construction joints, or because the joints are prepared to result in bonding between concrete placed at different times. Similarly, due to heavy reinforcement, the failure mode involving compression failure of the shear wall end sections is not predicted.

The inelastic energy absorption is related to the hysteresis as the structure describes inelastic displacements in sustaining loads up to the ultimate strength of the structural elements. The fragilities of the buildings are evaluated considering two limit states, according to ASCE/SEI 43-05 (Reference 49).

1.

Limit State C (LS-C) defined as limited permanent deformation, and

2.

Limit State A (LS-A) defined as short of collapse, but structurally stable.

ASCE 43 LS-C corresponds to the point where the structure exhibits sufficient strain to induce cracking and cause incipient failure of the anchorage of mounted components. ASCE 43 LS-A corresponds to an advanced limit state allowing pennanent inelastic deformations short of collapse, but structurally stable. This limit state is more representative of gross failure of the structure, whereas LS-C represents a failure of equipment housed within the structure. Inelastic energy absorption factor values presented in Table 5-l of ASCE/SEI 43-05 (Reference 49) consistent with the limit state being evaluated are selected and converted to median level for use in the separation of variables fragility evaluation of the walls.

With the exception of structural damping, all other variables in the building seismic analysis are median values. A conservative value of structural damping (4 percent of critical) is used to develop the ISRS for use in the CDFM calculations. However, a higher damping is used in the fragility analysis of the structure itself withthe value depending onthe limit state being evaluated. For LS-C, 7 percent of critical damping is considered as median. A higher damping of 10 percent of critical is selected for LS-A consistent with the advanced degree of damage.

lffiGontulting rlRrzzo

273429+R-036 Reaision 0 May LL,201.7 Pase 68 of145 4.4.2.6 Grouping of Equipment for Seismic Evaluation The equipment screened in for evaluation are grouped to condense the equipment list into a reasonable number of groups containing similar equipment based on several attributes, including the following:

Equipment types (SQUG GIP Classes), such as horizontal and vertical pumps.

Associated systems.

Potential concerns encountered which could impact the seismic capacity.

. Location, such as building and floor elevation.

Size.

o Manufacturer.

Observations made during walkdowns are also utilizedto assess if components included in a group need a specific evaluation (as opposed to generic approaches) to establish a capacity. For example:

Component does not meet all caveats of respective GIP class; e.g. valves with excessively cantilevered actuators.

Supplemental supports, such as snubbers, rigid struts, or hangers for valve yokes.

Potential of seismic interactions.

Where differences in physical characteristics, such as the dimensions, weight, manufacturer, etc.,

are observed for components included in EPRI equipment classes, additional sub-groups were created so that representative HCLPF values could be developed. Finally, components within groups are subdivided based on building and elevations to address the differences in floor response spectra.

In some instances, a relatively large number of components were grouped together and represented by a component that reasonably bounds the seismic capacity of other somponents in the group. The inherent conservatism in this approach is justified on the basis that the bounding capacity exceeds the risk significance level. Therefore, the seismic fragilities of all of the components bounded by this representative component also have a negligible quantitative impact on the PRA results.

4.4.2,7 Fragility Evaluation of Mechanical and Electrical Equipment In general, fragilities are evaluated for the equipment functional and structural/anchorage capacities, as well as relay and potential interactions where applicable. Functional fragility is typically established by comparing the ISRS near the equipment, clipped according to EPRI 6041, to a capacity spectrum in a frequency range of interest. Most equipment functional capacities are established on the basis of experience data, generic equipment ruggedness spectra (GERS), or qualification test data. These capacities do not represent the anchorage capacity of the equipment and accordingly anchorage fragility evaluation is also necessary where these approaches are used. Anchorage fragility is typically calculated by scaling design basis analyses or by new analysis. For passive equipment such as tanks, only a structural/anchorage fragility is evaluated.

lBSGonsulting ilRtzz-o

273429+R-036 Reaision 0 May 1.1",2017 Pase 69 of145 4.4.2.8 Fragility Evaluation Based on Experience Data A HCLPF capacity based on earthquake experience data, when used is justified by documenting thatthe associated caveats are satisfied. EPRINP-7149-D (Reference 56) and its supplement EPRI NP-7149-D-S1 (Reference 57) document the development of these caveats based on extensive surveys and cataloging of the effects of strong ground motion earthquakes on various classes of equipment mounted in conventional power stations and other industrial facilities. The seismic experience database presented in these reports reflects detailed investigations of some 120 sites withinthe strong-motionregions of some 23 earthquakes from l97l to 1993 by SQUG, EPR[, and EQE International.

EPRI 1019200 (Reference 44) presents further analysis forthe earthquake experience database.

It concludes that components satisffing the requirements to assign al.2g capacity in terms of PGA will exhibit HCLPF capacities developed as follows:

r For ground-mounted items, the mounting level capacity is 1.32g for comparison to either free-field demand or clipped in-structure demand spectra.

. For structure-mounted items, the mounting level capacity is 1.80g for comparison to clipped ISRS demands.

I These experience-based capacities of 1.80g and 1.32g can be used to develop a component functional HCLPF capacity in a manner similar to a capacity response spectrum developed by testing, such as a GERS.

The ISRS, which reflect the calculated floor response spectra, often exhibit highly amplified naffow frequency content. These narrow peaks are not well correlated with potential structural or functional failure. Therefore, when comparing peaked floor response spectra with an experience-based capacity, the peaks in the floor response spectra are "clipped" as described in Appendix Q of EPRI NP-6041-SL (Reference 7).

4.4.2.9 Fragility Evaluation Based on Test Data The seismic capacity of components qualified onthe basis of tests (e.g., electrical cabinets) may utilize either specific qualification testing or generic test data. The seismic capacity is determined as the ratio of the TRS to the required response spectra (RRS) associated with the evaluation basis earthquake. In order to bias the capacity to CDFM level of conservatism, the selected TRS is associated with a 99 percent exceedance probability. Depending upon whether the testing is assembly based or device-based, local amplification may be incorporated to obtain device-based capacities (using, for example, in-cabinet response spectra).

Several reference documents, such as EPRI NP-6041-SL (Reference 7), EPzu TR-103959 (Reference 11), EPRI NP-5223-SL (Reference 58), and SQUG/GIP (Reference 18), present the methodology to develop CDFM level capacities based on Test Response Data for specific classes of M&E equipment. These documents specify the conditions (caveats), under which the GERS may be used. Available TRS for specific equipment are also considered to develop seismic capacities. However, the TRS are taken to represent a LB of the capacity of the respective equipment.

lBSConsulting tlRtz?(o

273429+R-036 Reuision 0 May 1.L,2017 Page 70 of 7.45 Where CDFM level capacities were assigned based on generic test data, the walkdowlr observations provided the basis for considering that the associated caveats are satisfied.

The ISRS, which reflect the calculated floor response specta, often exhibit highly amplified narrow frequency content. These nilrow peaks are not well correlated with potential strustural or functional failure. Therefore, when comparing peaked floor response spectra with a TRS capacity, the peaks in the floor response spectra are clipped as described in Appendix Q of EPzu NP-6041-SL (Reference 7).

4,4.2.10 Fragitity Evaluation Based on New Analysis or Scaling of Existing Analysis Typical codes and standards used in the qualification of equipment by analysis include those published by ASME, American Institute of Steel Construction (AISC), ACI and Institute of Electrical and Electronics Engineers (IEEE) Standard. Additionally, EPRI NP-6041-SL (Reference 7) identifies load combinations and stress limits for pressure retaining components, supports, ffid anchorage.

When equipment is qualified based on design analysis, it was recognized that the component design capacity is determined by code specified stress and design displacement limits. The CDFM capacrty, on the other hand, is obtained for as-built conditions using stress limits corresponding to the code specified minimum stress or the material yield strength with a 95 percent exceedance probability. However, for non-ductile materials EPRI NP-6041-SL (Reference 7) suggests using 70 percent of the material yield as the stress limit.

The evaluation of M&E components based on generic and seismic experience capacities are supplemented with the verification of the equipment anchorage. For anchorage fragility evaluation, approaches include scaling of existing analysis or new analysis. Scaling of existing analyses is performed considering the guidance of EPRI 6041 (ReferenceT). New analysis is performed in accordance with the procedure outlined in the SQUG/GIP (Reference l8). This procedure follows a static equivalent approach, where the inertial load of the equipment is applied at the equipment center of gravity. The inertial load in each direction is equal to the product of the Sa, an equivalent static coefficient, and the mass of the equipment. An equivalent static coefficient of 1.0 is used forthe anchorage analysis of M&E equipment.

The seismic demand on the equipment anchorage in terms of tension and shear is developed consistent with the following equipment characteristics :

Mass of the Equipment o Location of the Center of Gravity o Natural Frequency o Equipment Damping t

Equipment Base Center of Rotation The equipment mass defines the inertial loads, while the location of the center of gravity determines the overturning moment caused by the inertial loads. The seismic anchorage demand of the equipment is determined by shifting the appropriate floor response spectrum to account for the effects of the uncertainties in the structural frequencies, according to EPRI NP-6041-SL (ReferenceT). Then, the lowest natural frequency of the equipment is used to determine the amplified acceleration of the equipment from the shifted ISRS.

lESConsulting rlRrzzo

273429+R-036 Rezision 0 May LL,201"7 Pase 71. of 1.45 4.4.2.11 Fragility Evaluation of Ilistribution Systems Components Distribution systems, piping, cable trays and supports, and HVAC are typically treated on a sampling basis and are evaluated using generic data and earthquake experience data. A conservative 0.509 PGA HCLPF value is assigned to distribution systems in the BVPS-2; i.e., piping, HVAC ducts, and cable trays and conduits, onthe basis of earthquake experience data.

Experience from past strong-motion earthquakes in industrial facilities throughout the world indicated that, in general, distribution systems are seismically rugged. The seismic experience data shows that most types of piping systems exhibit extremely good performance under strong-motion seismic loading, with the pressure boundary being retained in all but a handful of cases. The BVPS-2 Walkdown Report (Reference 40) presents a summary of walkdown observations, which provide the basis to assign a 0.509 PGA HCLPF value to distribution systems.

4.4.2.12 Fragility Evaluation of Relays 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 BVPS-2 SPRA, in accordance with SPID (Reference 2) and ASME/ANS PRA Standard (Reference 4). The evaluation resulted in most relay chatter scenarios screened from further evaluation based on no impact to component function. The relays that were not screened were addressed in the SPRA with appropriate seismic fragility or operator action.

The seismic fragility for the relay chatter mode is developed based on test reports for specific relay models. For the relay chatter evaluation, the CDFM methodology is followed as described in EPRI 6041 (Reference 7).

Appropriate cabinet amplification factors, AFr, are considered to scale the ISRS to an estimated mounting point spectrum. In general, amplification factors from Table Q-l of EPRI 6041 (Reference 7) are used for the horizontal direction and EPzu 3002004396 (Reference 39) for the vertical direction. The recofirmended factors are shown in Table C-8 below:

TABLE 4-8 RT,COMMENDED CABINET AMPLIFICATION FACTORS (BpRr 6041 (REFERENCE 7), EPRr 3002004396 (REFERENCE 39))

DmBcrrou Cnnmrr Tvrr Annrr,rrICATIoN Facron, AF Horizontal Motor Control Centers 3.6 Low and Medium Voltage Switchgears 7.2 Stiff Panels and Control Boards 4.5 Vertical All 4.7 A knockdown factor, Fr,, has been considered to obtain about a 99 percent exceedance level capacity. Representative knockdown factors are presented in Table Q-2 of EPRI 6041 lESGonsulting

(}Rtzzo

273429+R-036 Ratision 0 May L1., 201"7 Pase 72 of145 (Reference 7) and reproduced in Table 4-9 below. Knockdown factors coffesponding to IEEE C37-98 Relay Fragility Tests, GERS

• Relays, ffid Component-Specific Qualification Test: Function During are used for the BVPS-2 relay evaluation.

TABLE 4-9 RECOMMENDED TRS KNOCKDOWIT FACTORS (EPRr 6041 (REFERENCE 7))

Dnra Souncn KuocxnowN FncroR, Fr HCLPF Capacities 1.0 GERS - Non-Relays t.2 GERS - Relays 1.5 IEEE C37-98 Relay Fragility Test 1.08 Component-Specific Qualification Test:

Function During 1.2 Component-Specifi c Qualification Test :

Function After [No Anomalies) 1.0 Component-Specific Qualification Test :

Function After (Anomalies) 1.1 - 1.6 TRS are all broad banded and are not clipped, but RRS were clipped as appropriate. Therefore, Crfactoris l.0withnouncertainty. PeTEPRI 6041 (ReferenceT),whentheTRSarefor multi-axis excitation, and the RRS is predominantly a single-axis excitation, as is the case for relays and contactors mounted on panels in cabinets, then the TRS should be increased by a multi-axis to single-a:<is correction factor, Fus,to remove the unnecessary conservatism.

EPRI 6041 (Reference 7) suggests Fus:1.20.

4.4,2.13 Fragility Evaluation of NSSS Components Ten NSSS components are included in the SEL: Pressurizer, three Reactor Coolant Pumps, Reactor Internals, Control Rods, Reactor Vessel, and three Steam Generators. The fragilities for these NSSS components are based on new analysis, design basis criteria, scaling available seismic calculations, and earthquake experience data.

4.4.2.14 Fragility Evaluation of Block 'Walls No block walls were identified in the vicinity of evaluated equipment.

4.4.2.15 Fragility Evaluation of Non-Seismic Category I SSCs A 0.10g HCLPF capacity is assigned to all Non-Cat 1 SSCs prior to any fragility calculation unless a higher capacity was requested by the PRA analyst. The basis for this capacrty is that it corresponds to the HCLPF recommended for loss of offsite power (LOOP) per the EPRI SPRAIG Report 3002000709 (Reference l5), NUREG-1738 (Reference 59), and NUREG-CR-3558 (Reference 60). The representative failure mode for LOOP is the brifile failure of the ceramic insulators on transformers per NUREG-CR-4334 (Reference 6l) and I'{UREG-CR-3558 (Reference 60). A key function of non-Cat I equipment relates to bringing offsite power into the Station. The equipment that supports this function is judged to have lESGonsulting riRtzzo

2734294-R-036 Reaision A May 11,2017 PaRe 73 of145 HCLPFs greater than or equal to that of offsite power. Therefore, the seismic capacity of off-site power constitutes the weak link in the system. The equipment that supports systems that bring offsite power into the Station are limited by the seismic capacity of LOOP and accordingly may be assigned the same capacity. Other Non-seismic Category I SSCs not related to LOOP are assigned a conservative low HCLPF capacity of 0.lg on the basis that they have such low impact on thb SPRA results and risk quantification is not sensitive to the conservatism in their fragilities.

4,4.2.16 FragilityRefinementProcess The objective of refining seismic fragilities is to assess unintended conservatism in the fragility parameters to subsequently achieve an acceptable risk level quantified in terms of CDF or LERF.

The refinement of seismic fragilities for SSCs constitutes an iterative process between the fragility analyst and PRA systems modeler. This iterative process can be sunrmarized as follows:

1.

The fragility analyst develops seismic fragilities based on generic methods (i.e., earthquake experience or GERS) and scaling of existing anchorage analysis.

2.

This initial set of seismic fragilities is provided to the PRA systems modeler in the form of HCLPF capacities, logarithmic standard deviations, median capacities, and controlling failure modes.

3.

By performing initial risk quantification, the PRA modelerrecords the CDF and LERF values achieved with this initial set of fragilities.

4.

The PRA modeler will then proceed to evaluate the risk level and determine if the resulting CDF and LERF fall within an acceptable risk level.

5.

In case the resulting CDF and LERF does not represent an acceptable risk level, say greater than 10-6, the PRA modeler will identiff and rank the SSCs with the highest contribution to CDF and/or LERF.

6.

This list of top contributors is then provided to the fragility analyst with the intent to refine the SSCs seismic fragilities. In order to refine or provide more representative fragilities, the fragility analyst will recur to several methods including:

t Creating new groups and selecting new representative components.

. Refining of seismic demand through the development of computer models.

. Inclusion of a higher ductility factor.

r Performing a new fragility calculation following the separation of variable approach.

7.

After refinement of seismic fragilities, the fragility analyst will convey the newly refined fragilities to the PRA systems modeler for new risk quantification.

8.

This process is repeated until an acceptable CDF and LERF risk level has been achieved.

The refinement of seismic fragilities for several SSCs in the BVPS-2 PRA model was performed by following this process until an acceptable CDF or LERF was achieved.

lESGonsulting (lRtzz.o

273429+R-436 Reuision 0 May LL, 201.7 Page 74 of 1"45 4.4.3 SSC Fragility Results and Insights Table 5-10 andTable 5-ll inSection 5.0 provide lists of fragilities for SSCs at BVPS-2 determined to be top contributors to risk, based on Fussell-Vesely importance (FVI) from the final SPRA quantifications of CDF and LERF. The Median acceleration capacity Am and associated variabilities pr and Bu are provided for each SSC along with their calculation method, and failure mode addressed in the PRA plant model.

4.4.4 Fragitity Analysis Technical Adequacy The BVPS-2 SPRA SSC Fragility Analysis was subjected to an independent peer review against the pertinent requirements in the PRA Standard (Reference 4).

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, ard establishes that the BVPS-2 SPRA SSC Fragility Analysis is suitahle for this SPRA application.

ABSGonsulting

{iR}zzo

2734294-R-036 Reaision 0 May 1L,20L7 Pase 75 of145 5.0 PLANT SEISMIC 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 identiff 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 Ilnvnr,opMENT oF THE SPRA Pr,,+.nr Snrsprrc Loctc Monrl The BVPS-2 seismic response model was developed by starting with the BVPS-2 internal events at-power PRA working model as of July 2014, and adapting the model in accordance with guidance in the SPID (Reference 2) and PRA Standard (Reference 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. In early 2016, the seismic model was incorporated into the latest at-power PRA model of record-that was effective on December 3 1, 201 5-and the seismic model was further refined to resolve findings and observations from a December 2014 seismic PRA peer review (see Reference 6).

For the BVPS-2 SPRA, the following methods were used to develop the seismic plant response model:

The BVPS-2 PRA is comprised of two major areas of analysis: (1) the identification of seismically-induced sequences of events that could lead to core damage and the estimation of their frequencies of occuffence (the front-end analysis); and (2) the evaluation of the potential response of containment to these sequences, with emphasis on the possible modes of containment failure and the corresponding radionuclide source terms (the back-end analysis).

The overall methodology for both the front-end and back-end analysis can be characterized as the "linked-event tree" approach. Under this approach, a set of linked-event trees was developed for the plant responses needed to model the impacts from seismic initiating events. The model for these plant responses was developed starting from the General Transient event tree set developed for internal events (see Reference 62). This event tree set also considers transient-induced small LOCA. An updated seismic pre-tree was developed to replace the one previously linked to the General Transient event tree set. These event trees allow the safety functions that must be achieved to keep the core cooled to be organized in a way that defines accident sequences that lead to core damage. The potential for failure of each of the safety functions is defined through the construction of a fault tree. The fault trees carry the modeling from the level of safety functions down to the basic hardware failures and human actions (or inactions) that can contribute to a core damage sequence. Using reliability data assembled from a review of operating experience at BVPS-Z and on an industry-wide basis, the integrated models can be evaluated to yield estimates of the frequencies of the core damage accidents of concern.

fESGonsulting

[]Rtzzo

273429+R-036 Ranision 0 May 11,20L7 Page 76 of 1,45 As described in Reference 38, the SPRA model builds upon the intemal events PRA accident sequence models documented in Reference 62. A cross-reference between the top events considered in that model and the system notehooks where the analysis for top events is documented is provided in Table B-1 of Reference 63. The internal events PRA consists of many notebooks listed in Table 3. I I of Reference 63 document the models used as a starting point. The portion of the internal events PRA sequence models used in the SPRA and the changes made to incorporate seismic failure events are documented in Reference 2.

The back-end analysis is essentially the same as that performed for the internal events PRA, as documented in Reference 64. The back-end analysis performed for the internal events PRA employed both deterministic and probabilistic analysis tools to follow the progression of the core damage accidents. Computer codes were used to simulate the meltdown of the core, the failure of the reactor vessel due to contact with molten core materials, and the transport and interactions of core debris in the containment. Because of the large rxrcertainties associated with the progression of a core damage accident, these deterministic calculations were supplemented with assessments that considered the potential for phenomena different from or more severe than those treated in the computer codes (see Reference 64). The results of that analysis included an assessment of the potential for a variety of containment failure modes for each type of core-damage sequence, and an estimate of the magnitude of the radionuclide release that would be associated with each.

The seismic hazard curve for BVPS-2 is shown on Figure 5-l below, taken from Figure 6-7 of Reference2S. The 100 Hz spectral acceleration is selected to represent the zero period PGA at the analysis Reactor Containment Building control point. All SSC fragilities are also developed with referenced to this same control point. The BVPS-2 SSE at 0.1259 has a mean hazard exceedance frequency of 3E-04 per year. The hazard exceedance frequency of 1E-05 is at 0.5S and the exceedance frequency is about 1E-06 per year at 1.0g.

lSSGonsulting tlRtzzo

2734294-R-036 Reoision 0 May 11,20L7 77 1,45 1.ffi,-02 1.0E-03 1.0E-04 1.0E-05 l.ffi -O7 t.(E -08 100.0 Hz I a a j a

\\

ir

\\

\\

-06 1,OE ot,ttlEoo(,x UJ Ett,tro I

dtr I,L Ur I il--trE4

\\

\\

o_10 Acceleration [g]

a

t o.o1 1.tlo 10-oo T.IGURE 5.1 SEISMIC HAZARD EXCEEDANCE CURVES F'OR BEAVER VALLEY SITE AT THE REACTOR CONTAINMENT BUILDING F'OUNDATION, INCLUDING UNCERTAINTIES The BVPS-2 seismic exceedance curves shown on Figure 5-1 are in units of per calendar year.

The SPRA model is to assess the risk of at-power plant operation. Therefore, the exceedance curves are scaled by the Unit2 specific availability factor of 0.936, to obtain the mean exceedance frequency curve for at-power conditions; i.e., the rest of the time the plant is not at-power and the SPRA model does not apply. Table 5-1 lists the scaled and unscaled mean seismic hazard exceedance frequencies at the accelerations provided from Reference23.

fBtGonsulting

()Rrzzo Mean

......r Sth*fdn

{; r *. 16*th*fdn

,r. r S0th*fdn

- 84th_f dn

. 95_fdn t

t a

a a

a a

o a

atrt

2734294-R-036 Rwision 0 May 11,201.7 Page 78 of 1.45 TABLE 5.1 MEAN SEISMIC HAZARI} EXCEEDANCE FREQUENCIES SCALEI} BY PLANT AVAILABILITY AccrlrRATroN G)

ExcrnnANCE FnneunNCY Sc^r,r,rn ny Pr,nnr UxavuLABrLrrY (0.936)

Mral* Cunvr Mn.r,x Cunvr 0.01 1.19E-02 I.l I E-02 0.02 4.33E-03 4.0sE-03 0.03 2.23E-03 2.09E-03 0.04 1.44E-03 1.35E-03 0.05 1.03E-03 9.64E-04 0.06 7.698-04 7.208-04 0.07 6.02E-04 5.63E-04 0.08 4.86E-04 4.55E-04 0.09 4.028-04 3.76E-04 0.10 3.38E-04

3. 16E-04 0.20 8.92E-05 8.3sE-05 0.25 5.54E-05 5.19E-05 0.30 3.77E-0s 3.s3E-05 0.40 1.95E-05 1.83E-05 0.50 1.09E-05 t.02E-05 0.60 6.s8E-06 6.16E-06 0.70 4.21E-06 3.94E-06 0.80 2.788-06 2.60E-06 0.90 1.87E-06 1.75E-06 1.00 1.26E-06 1.l8E-06 2.00 9.70E-08 9.08E-08 3.00 2.44E-08 2.28E-08 5.00 3.6sE-09 3.428-09 The seismic initiating event frequencies and their associated acceleration intervals are found in Table 5-2. The analysis acceleration for computing SSC fragilities is also listed. Finally, the fourhuman reliability analysis (HRA) analysis intervals, are also associated withthe l0 seismic analysis intervals chosen. The basis for this assignment is provided in Reference 38.

The lowest acceleration for the SPRA (0.06S) was selected so that the geometric mean of the acceleration interval would be roughly 0.1g; i.e., the HCLPF value forthe off-site power fragility. This same selection has been made for the SPRA models for other FirstEnergy Nuclear Operating Company (FENOC) plants. Relatively narrow acceleration intervals were selected for those ranges of acceleration where the conditional core-damage probability was expected to change most quickly, and to aid in the demonstration that adding new SSC fragilities with higher capacity would not significantly impact the computed CDF. Therefore, constant interval widths lESGonrulting

[]Rtzzo

273429+R-036 Reaision 0 May 11,2017 Pnop 79 nf145 of 0.1g were selected for the range between 0.4g to 0.8g. Above 0.8g, the acceleration widths of the remaining seismic initiating event intervals were broadened. The higher range of the acceleration intervals is retained to evaluate LERF. Withthe exceptionof G08, the uneven acceleration interval widths still result in the initiating event frequencies decreasing for each interval.

TABLE 5.2 SEISMIC INITIATING EVENT INTERVALS IE NAME PGA Lownn PGA HrcnEn IE FREQ G01 0.06

0. 15 s.38E-04 G02 0.15 0.2s I.t0E-04 G03 0.25 0.4 3.35E-05 G04 0.4 0.5 7.99E-06 G0s 0.5 0.6 4.03E-06 G06 0.6 0.7 2.21E-06 G07 0.7 0.8 1.33E-06 G08 0.8 1.0 1.42E-06 G09 1.0 2.0 1.08E-06 Gl0 2.0 4.99 8.67E-08 Freq. Sum:

6.99E-04 5.1.1 Seismic Initiating Event Impacts The purpose of this section is to document the potential initiating event impacts that may be caused by seismic events so that a suitable plant response model to respond to each of the impacts is accounted for in the SPRA. Fortunately, the BVPS-2 Internal Events PRA includes a long list of initiating event impacts and a number of unique plant response models. These plant response models take the form of linked-event tree sets wherein each set contains a seismic pre-tree, a fire analysis tree, a support tree, one or more frontline trees and a containment tree.

The event tree sets are best distinguished by their frontline tree names since the other event trees mentioned previously are cofirmon to each event tree set, resulting in the following event tee sets:

1.

Excessive (e.9., Reactor Vessel Rupture) LOCAs

2.

Large LOCAs

3.

Medium LOCAs

4.

General Transient/ Small LOCAs

5.

Steam Generator Tube Ruptures

6.

Anticipated Transient without Scram (ATWS), for Transients Involving Failure to Trip

7.

Interfacing Systems LOCAs The sequences for these plant response models are created hy linking the frontline tree to the other trees in the set, including the containment event tree so that Level 1 and Level 2 end states may be calculated; i.e., where the CDF from seismic events is a sequence group (SEIS) defined as the sum of all release categories at the end of the containment event tree. The sequence group lEEConsulting rlRtzzo

2734294-R-036 Reaision 0 May 1.1.,2017 Pase 80 0f145 (LERFS) for large early release from contributed by seismic events is the sum ofjust those release categories acknowledged to result in a large early release; i.e., release categories 8V01, BV01S,8V02, BV02S,8V03, BV03S,8V04, BV04S,8V18, and BVlg. The trailing "S" in these release categories indicates that a small containment penetration fails to isolate. Large, early releases result from containment bypasses (8V18 or BVl9), or from large, early containment failures (8V01, 8V02, 8V03, or BV04) with or without an accompanying small containment isolation failure (i.e. as represented by bin nirme suffix "S").

The basic events included in the internal events PRA models were used in large part to develop the BVPS-2 SEL. These events form a large portion of the SEL. Therefore, the seismic impacts of most SSCs are already accounted for in the internal events PRA models. What has been added to the SEL, are the passive SSC failures and potential relay chatter effects. These passive failures need only be added to the list of seismic impacts affecting aplant response if they are new, cannot be modeled by an existing plant response model, and if the seismic SSC failure probabilities fall below the screening criterion for inclusion in the SPRA model. We have adopted an SSC screening criterion of 0.7e for the SSC HCLPF. SSCs with HCLPFs higher than A.lg may still be added to the model so long as the required plant response model is available.

For the BVPS-2 seismic PRA, we settled on including only the General Transient/ Small LOCAs event tree set. The reasons are seen n Table 4-1 of Section 4.1.1where a review of the fulI list of internal events initiating events is documented for applicability to seismic events.

In summary, the following assumptions and bases are used in the development of the BVPS-2 systems model:

l.

The Internal Events PRA was last formally documented in 2015 (see Reference 89). This BV2REV6 model served as the foundation for the latest version of the seismic PRA presented in the interim revision BV2REV6A model. While the seismic PRA evaluated in the December 2014 Seismic Peer Review used an earlier effective Internal Events PRA model as its foundation, the same methodologies were used when incorporating into the latest model.

2.

The Internal Events PRA is used as the technical basis for both CDF and LERF. All assumptions and success criteria in the Internal Events PRA are retained in the SPRA for the portions of the sequence models that apply (see Reference 89). This assumption provides continuity between the Internal Events PRA and the SPRA. Any future changes to the Internal Events PRA success criteria would be addressed as part of the maintenance and update process of the integrated PRA.

3.

An SSC HCLPF of 0.359 is used as the screening criterion for excluding potential seismic-induced fires. Please see Section 5.5.2 of Reference 38.

4.

The portions of the internal events PRA model that apply to seismic events are:

Transients (which include small and very small loss of coolant accidents [SLOCA and VSLOCAI and losses of offsite power) and seismic events assumed to lead directly to core damage and/or large early release.

5.

ATWS sequences are excluded from the SPRA model on the basis of low frequency; based on multiple redundant trip signals resulting from ground acceleration, as well as AESGonsulting rlRrzzo

6 7

273429+R-036 Reaision 0 May 1.1.,201.7 Pase 81of 1.45 highly reliable operation action to trip the plant, it is assumed thaf the reactor would successfully trip. However, seismic capacity of the control rod drive mechanism is evaluated, and seismic damage to this component is assumed to lead to core damage.

The random catastrophic reactor vessel rupture event sequence model (Excessive LOCA [ELOCA]) is screened from the SPRA on the basis of low frequency. However, seismic capacity of the reactor vessel itself is evaluated, and seismic damage to this component is assumed to lead to core damage.

Sequences involving seismic SSCs failures judged to lead directly to core damage (e.9., polar crane in the Reactor Containment Building falling onto the reactor vessel) are guaranteed to be binned to core damage through inclusion of certain event tree rules.

These SSCs are represented by Top Event ZLI (see Section 4.5.1 of Reference 38).

However, systems that may have an impact on radiological release categories (e.g., containment spray systems) are still evaluated probabilistically; i.e., not guaranteed failed.

Seismic SSCs failures judged to lead directly to core damage plus a large early release (e.9., selected building failures) bypass the usual General Transient initiator event trees, and through the inclusion of certain event tree rules, these sequences are guaranteed to lead to core damage and to a large early failure of the containment, which is always mapped to a large early release category. These SSCs are represented by Top Event ZLz (see Section 4.5.1 of Reference 38).

Sequences involving steam generator tube rupture as a direct result of the seismic motion were not included in the SPRA because no seismic failures that cause a steam generator tube rupture (SGTR) without otherwise failing the steam generator were identified.

Pressure-and temperature-induced SGTR following core damage are still evaluated in the containment event tree, and may have an impact on radiological release.

The Interfacing Systems LOCA (ISLOCA) initiating events model from the Internal Events PRA was reviewed for applicable SSCs, but none were found applicable to seismic failure modes and so the associated sequence model was not used in the SPRA model.

The CDF model screening criterion used for excluding SSCs from the SPRA logic models is an SSC HCLPF value of higher than 0.79. See Section 5.1 of Reference 38 for a further explanation.

The LERF model screening criterion used for excluding SSCs from the SPRA logic models is an SSC HCLPF value of 2.0g or higher. See Section 5.1of Reference 38 for a funher explanation.

Large loss of coolant accidents (LLOCA) are screened from the final SPRA (see Section 5.1 of Reference 38 for screening discussion). All Beaver Valley Unit 2 specific NSSS components large enough to result in these larger breaks were found seismically robust enough to be excluded (see Reference 30). The generic fragility for large breaks suggested by EPRI (Reference 15) has a HCLPF above the 0.7g screening criterion.

lESGonsulting

{}Hrzzo I

I 10.

11.

t2.

13.

t4.

15.

16.

t7.

18.

19.

20 21.

2734294-R-036 Reaision 0 May 1-L,201"7 Pase 82 of 1,45 SSCs located in the twbine building are not credited in the SPRA sequence models. The turbine building is a non-seismic design and so is not resistant to extreme shaking.

Further, it contains many SSCs that are also susceptible to seismic shaking. Therefore, while it is expected that the turbine building and SSCs have some seismic capacity to respond to low accelerations, no credit was assumed for the turbine building for low acceleration ranges.

Although components in the turbine building are assumed failed for all seismic initiators, there are also cables that pass through the turbine building, but these SSCs are not assumed to fail. See Section 4.5.3 of Reference 38 for further discussion on this topic.

No credit is taken for the ERF station blackout diesel generator or any SSCs within the ERF structures. The ERF structures are of non-seismic design and so are not resistant to extreme shaking. Further, they may supply many SSCs that are also susceptible to seismic shaking and failure.

The seismic failure of offsite power is assumed to also impact the normal switchgear which would otherwise bring offsite power to the emergency 4 kV buses. This assumption effectively precludes credit for the cross-tie of power from Unit 1 to the Unit 2 emergency buses. The assumption is conservative. The degree of conservatism depends on the seismic capacities of the normal switchgear buses at both Unit 1 and Unit 2 which are used to align for cross-tie of the emergency buses. As a result, cross-tie capabilities (Top Event XT) are assumed failed for all seismic events for the Unit 2 model.

Seismic SSC failures are assumed to be complete failures, in that the SSC fails to perform its function, ornot. Degraded states of equipment (e.9., where onlythe equipment failure rates differ from the internal events model) for the period following the seismic initiator are not represented.

The assumed SSC seismic failure mode depends on the SSC type and whether the fragility applies to functional failure, structural/anchorage, or interaction faih.re. See Section 5.2 ofReference 38 for further explanation. Relay chatter failure modes are a function of the specific relay and SSC control circuit itself, See Reference 37 for more discussion of relay chatter.

Inadvertent actuation of the Safety Injection (SI) signal or other Engineered Safety Feature Actuation System functions may occur as a result of seismic failures in the actuation logic, or functional failure of the associated cabinets. However, the primary and secondary process racks and reactor protection racks all screen at high seismic capacity; i.e., greater HCLPF than 0.79.

The standby service water system is in a Category 2 building (AISX) and so preliminarily assigned a low seismic capacity, and thus the alternate service water pumps have a high probahility of failure for even the lower seismic events. This conservatism is not expected to be significant because of the redundancy offered by the Category I service water system and the similarity of support systems both systems require for success.

fESGonsulting tlRlzzo

2734294-R-036 Reoision 0 May 11,2017 Paxe 83 of145 2s 26 27 22 23 24 28.

29.

No credit is given to several systems located in the Unit 2 non-seismic turbine building, which is assumed to fail for even lower-magnitude seismic events. See Section 4.5.3 of Reference 38. No credit is given forthe six portable generators during a seismically-caused station blackout because they are located in the non-seismic Unit 1 turbine building, which is also assumed to fail for even lower-magnitude seismic events.

The steam generator atmospheric relief valves and safety valves are highly redundant for steaming the steam generators. We conservatively assume that if they fail seismically, they would all fail to open; i.e., that the sfrong motion occurs before they are called on to open. This assumption is conservative because it would fail all steam generator cooling.

Seismic failures of buildings that are adjacent to the Reactor Containment Building were assumed to fail in a way that opened flow paths around the containment penetrations into each building. The flow area was assumed large enough to lead to a large early release should a core damage sequence also occur. The buildings applicable to this scenario are represented in Top Event ZLZ (see Section 4.5.1 of Reference 38).

Seismic failures of the containment spray nozzles or discharge headers were assumed not to affect the transfer of water from the RWST into the containment. Such failures would affect the spray function but this function is not required to protect the containment.

Many other SSCs are seismically rugged, and therefore their seismic failure probabilities are unchanged from the intemal events PRA; e.g., check valves, manual valves, cable trays, conduits, junction boxes, and local starter boards.

Test and maintenance (T&M) basic events are not affected by seismic events and so were left unchanged. T&M frequencies input to the model determine what components will be out of service for the beginning of the initiating event. Seismic events cannot be predicted and prepared for, in the same way that plants can prepare for a hurricane making landfall by restoring all possible components to senrice, for example, so the T&M frequencies are kept the same as in the latest data update.

Common-cause basic events are not affected by seismic events, ffid so were left unchanged. In the Seismic Event Tree, a component group either survives the earthquake or it fails, which is probabilistically based onthe components' HCLPF values. Thus, if a component group probabilistically fails, all of the correlated components that would have normally been a part of a common-cause group also fail. If a component group succeeds, then all of those components are simply not failed directly by the seismic event, and still have the opportunity to fail in a common-cause manner throughout the mission time of the sequence. No partial failures are assigned, ffid thus the common-cause basic events are not affected. See Section 5.4 of Reference 38 for a deeper discussion of SSC correlation.

The impacts of several Internal Events initiating events are conservatively assumed to occur simultaneously during a seismic initiating event. See Table 4.1-1 of Reference 38 for more details.

The existing Internal Events PRA meets the Capability Category II requirements of the ASME PRA Standard for PRA applications. Table 2-1 of Reference 38 lists the upgrade lESConsulting

()Rrzzo 30.

)734294-R-036 Reoision 0 May L1,201.7 Page 84 af145 and update history of the Beaver Valley Unit 2 PRA through the years since it was first issued as an Individual Plant Examination PRA model in March of 1992. Equipment failure data for random failures, T&M unavailabilities, and plant configuration data are unchanged from the internal events PRA model. All seismic correlation sets and seismic initiating events are stored in the RISKMANTM software (Reference 69) model data. The increasing SSC seismic failure probabilities with acceleration interval are computed from the fragility cunres reported in Reference 41 within the Fragility Module of RISKMAN.

The Am, pr, and pu parameters of the SSC seismic fragility curves are used to compute the acceleration interval dependent failure probabilities and then combined with other fragility curves which lead to the same plant impacts to generate the seismic pre-tree top event failure prohabilities as appropriate. The seismic pre-tree accounts for the seismic SSC failures while the existing event trees account for the random SSC failures.

5.1.2 Seismic Event Trees for Large Early Release The Level 2 PRA Notebook (Reference 64) documents the containment event trees used, the mapping of sequences from the Level 1 plant response into plant damage state bins, the assignment of sequences into release categories, and their categorization into large/small and early/late release states. The same containment event tree (CET) which models the containment response is used here for the SPRA. The LERF sequences are one such categorization of releases and are used for the SPRA calculation of LERF due to seismic events.

During SEL development SSCs related to LERF were identified to prevent inadvertent screening due to the large HCLPF screening cutoff for LERF. These SSCs include but not limited to the containment structure and any SSC that could affect the function of the containment pressure boundary, as well as SSCs that have a role in containment isolation failures.

The release categories assigned to LERF in the LERF analysis for internal events are presented in the PRA Notebook (Reference 64).

A discussion of seismic containment failures resulting in flow paths large enough, should core damage occur, to potentially lead to a large early release is provided in Section 4.5.1 of Reference 38. Seismically-induced large holes in the Reactor Containment Building are represented by Top Event L1 inthe CET. Failure of Top Event Ll represents alarge hole inthe Reactor Containment Building prior to or at the time of Vessel Breach.

Regarding containment isolation failures on smaller lines, caused by seismic accelerations, see also the discussion of containment isolation failures in Section 4.5.2 and Table 4.5-1 of Reference 38" Seismic fragility assessments were performed on the containment isolation valves of the normally open lines of interest. Relay chatter analysis was also performed for the potential opening of isolation valves. These normally open lines, if failed, are modeled in GTRECIRC Top Event CI. CI failure represents openings too small to lead to a large early release and so do not impact the calculation of LERF at BVPS-2.

5.1.3 Relay Chatter Modeling The investigation into SSCs susceptible to relay chatter during a seismic event is documented in Reference 37. Circuit analysis was performed for identified SSCs (MOVs, Pumps, PORV, EDG Loading Circuits etc.). The evaluation of relay chatter considers chatter of not only relays, but lffiGonsulting tlHtzzo

2734294-R-036 Reaision 0 May LL,20L7 85 145 also other non-relay contact devices as electro-mechanical contactors, and motor starters (main and auxiliary contacts); circuit breakers (main and auxiliary contacts); manually-operated control switches, limit torque, and position switches; and mechanical sensor switches including pressure, level, flow, and temperature switches, etc. This includes all the devices identified to be susceptible to high-frequency motion identified in EPRI Phase 2 testing (Reference 90). The circuit analysis evaluated the impact of the contact device (relay) on the SSC and screened out devices based on the following:

l.

Relays that were located in non-seismically designed buildings were screened out as long as the components they wers associated with were also located in a non-seismic building.

The assumption is that both the component and relay fail when the building fails.

2.

The relay impacts indication or annunciation only. Such relays will not physically alter the state of the SSCs. This also includes relays for post-accident monitoring.

3.

The relay is not a lock-out relay and does not impact a seal-in or lock-out. Impacts to seal-in and lock-out relays are the principal concern in this study as these relays are the most likely to have an impact on PRA-related SSCs.

4.

Due to the lack of mechanically moving parts, solid state relays are not prone to chatter.

5.

Timing relays with settings greater than one second are not affected by chatter of upstream relays because they will be de-energized for sufficient time to reset the timing mechanism. However, a timing relay's output contacts may still chatter in response to seismic input.

Those relays that could not be screened had fragilities developed as dessribedinSection 4.7.2 of this submiual. The seismic failure of the relays that did not screen based on capacity was included in the SPRA. Each relay equipment group in the table below represents a correlation group of relays or contact devices that if chatter occrurod (based on calculated fragility) would fail the top event(s) presented in the table below. The following Tahle 5-.1 lists the relays or contact devices that were modeled and their effect on the model if chatter were to occur.

AEEGonsulting rlRtzzo

273429+R-036 Reaision 0 May 11,2017 Page 86 of745 TABLE 5.3

SUMMARY

OF RELAY CHATTER CONSEQUBNCES Top Event Top Event Name Equipment Groups Effect on Model if Top Event Failed ZP-z Relay EQl02, EQl10 (AR440AR & master)

EQr02 EQI r0 Failure of ZR2 leads to guaranteed failure of Top Events zsz, zsw, zsM, zHH, zDP, and RR ZF-3 Relay EQl03 (R223068-AP)

EQr03 Failure of ZR3 leads to guaranteed failure of Top Events zsz, zsw,zLP, zsM, zHH, zQS, and RR ZR4A Relays for EDG 2-l (DGF and D3)

EQI l3A EQl l4A Failure of ZR4A represents the failure of EDG 2-l by failing Top Event AO; if ZOG also failed, then ZM5 is guaranteed failed as well due to lack of Orange AC power ZR4B Relays for EDG 2-2 (DGF and D3)

EQI l38 EQI l4B Failure of ZR4B represents the failure of EDG 2-2by failing Top Event BP; if ZOG also failed, then ZM6 is guaranteed failed as well due to lack of Purple AC power ZI/{z MCC.2.El2 EQl26 Failure of ZM2leads to guaranteed failure of Top Event M2; macro QSPU ffue due to high frequency contactor chatter ZId3 MCC.2-E03, EO4 EQr 16 Failure of ZIn{3 leads to guaranteed failure of Top Events M3, M4; fails WA, WB due to contactor chatter ZM5 MCC.2.EO5, El3 EQl rs Guaranteed failure of Top Event M5; fails SE due to contactor chatter; fails HH due to contactor chatter if there is no SI signal ZMi6 MCC-2-E06, E14 EQl l7 Failure of ZM6 leads to guaranteed failure of Top Events M6 fails HH due to contactor chatter if there is no SI signal 5.1.4 Correlation of Fragilities SSCs not screened by potential impact on the plant were then assigned to correlation sets in part by their seismic capacities. It is important to account for dependencies between the probabilities of seismic failure modes, as appropriate. Past SPRA's have assumed that all identical and redundant equipment located in the same or at least seismically similar response locations, are 100 percent correlated, while assuming that equipment which is identical, but not redundant, (i.e., perform their functions in series) are uncoffelated. Here, by 100 percent correlated we mean that if one equipment item in the redundant set fails seismically, all others in that redundant set are also assumed to fail and via the same failure mode. This is a much stronger linkage than simply saying their failure prohabilities are the same yet the failure probabilities themselves are independent. This 100 percent correlation approach conservatively minimizes the advantages of redundancy; partial correlation is not modeled.

The approach to defining correlation groups in this study is explained below.

All SSCs on the SEL have been screened as seismically rugged, are judged not to impact the PRA model, or have had their seismic capacities assessed. Those assessed have been assigned to one of the EPRI seismic analysis categories as an initial step in computing seismic equipment lESGonsulting TlRtzzo

273429+R-036 Reaision 0 May LL, 201,7 Pase 87 ofL45 fragilities. These categories were further broken down into analysis groups which contain the SSCs sufficiently similar in anchorages as to be expected to all be evaluated in roughly the same way. For example, for BVPS-2 the equipment assigned to the EPRI Category 2l for tanks and heat exchangers was further divided into nine analysis groups due to perceived differences in the analysis needed to assess their seismic capacities.

A further consideration is in the final assessment of equipment capacities. In this study the equipment's HCLPF is used as ameasure of equipment capacity, although it is recognizedthat the capacity is defined by the entire fragility curve, including its parameters for median capacity and variability assigued. The HCLPF assigned is a function of many things, including the equipment type, seismic design classification and the exact SSC location within the building.

The general approach to correlating SSCs into correlation groups was to group those SSCs that of the same equipment types, have roughly the same seismic capacity, ffid subject to the same seismic accelerations. Reasons for not grouping such SSCs are as follows:

1.

SSCs in different EPRI categories are assigned to different correlation groups.

2.

SSCs in different buildings are assigned to different correlation groups.

3.

SSCs on different floors of the sirme building are assigned to different correlation groups.

4.

SSCs which seismic capacities are evaluated differently according to their different analysis groups are assigned to different correlation groups, though sometimes the analysis groups are sufficiently similar that they still should be grouped.

The approach to correlation was first to divide the fulI list of equipment into partial lists of nearly identical equipment. The lists of all equipment in the same EPRI category were separated out, one category atatime. If multiple types of equipment are assignedto the same EPRI category (for example air-operated valves (AOV) and relief valves are assigned to the same EPRI Category 7), then the list reviewed was further broken up by types of equipment within a given EPRI Category.

The next step was to sort the list of equipment within the EPRI category by capacrty as measured by their assessed HCLPFs.

Correlation groups were then assigned based primarily on similarity of the assessed HCLPFs.

While they need not be identical, the grouping into correlation sets was only performed for those SSCs with nearlythe same HCLPF; i.e., within say 0.059 of each other. Grouping equipment with substantially different HCLPFs can be problematical, because then it is unclear which HCLPF to assign to all the SSCs within the correlation set. For this study, the lowest HCLPF within the correlation set was assigned to all SSCs within the set, though most often equipment assignedto the same correlation group had identical HCLPFs. SSCs of the same equipmenttype with HCLPFs that differed by more than 0.059 were generally found to be designed to different seismic design classifications, located in different buildings, were in notably different elevations within the same building, or belonged to a different analysis group of the same EPRI category, indicating that their anchorage design maybe different.

Exceptions to the above rules for assigning SSCs to correlation groups were made for this study and are documented in Reference 38. Generally these assumptions reflect differences in the lESGonsulting

{}Rtzz-o

273429+R-036 Ratision 0 May 11,20L7 Pase 88 of145 depth for fragility analysis for each SSC and the relative importance of the SSCs. The correlation groups defined arepresented in Table 5.4-l of Reference 38 along withthe SSCs assigned to each. Nearly 205 SSCs ars explicitly grouped into 76 correlation groups. Since the SSCs may have a slightly different capacity than that assigned to the entire correlation set, the individual SSC HCLPFs are also listed in the table. Note that these HCLPFs are for the minimum HCLPF values for the different failure modes of the same SSC; i.e., from among the failure modes of functionality, structural/anchorage, relay chatter, or interaction failures.

5.1.5 Human Reliability Analysis The list of post-initiator human actions for the internal events model was analyzed for modification due to seismic affects. Some human failure events (HFE) were excluded from the analysis due to not being associated with the sequence models used to represent seismic initiatorst e.8., HFEs for SGTR initiators.

Every post-initiator HFE retained in the SPRA sequence models was evaluated for the impacts of seismic events. The degree of impact was assumed dependent on the seismic acceleration level.

At very high accelerations, the human error probabilities (HEP) were set to 1.0. The seismic impacts on every post-initiator HFE in the SPRA sequence models is accounted for by the HFE specific, performance shaping factors and selected minimal values that increase with acceleration as a function of plant damage state. The adjusted HFEs use the internal events name with the suffix of "Sn" where n ranges from 1 to 4; i.e., four separate seismic acceleration ranges were evaluated for varying seismic impacts, but in SEIS4, all post-trip actions are set to failed.

Further discussion of the modeling changes made to account for acceleration dependent HEPs is provided in Section 6.0 of Reference 38. A summary of the SPRA HRA HFE HEP Evaluation Process is provided in Table 5-4 below.

AESContulting rlRtzzo

TABLE 5.4 SPRA HFE IIEP EYALUATION PROCBSS

SUMMARY

Snmurc GRoup Aprnoxruarn G Lnvnl Snrsrurc Inrrmrrnc Evsur(s)

ConnnsroNDsTo FoR BVl Arrncrs CR HFE Arrncrs FtrlnHFE Copruprurs SEISl 0-0. l5 GI SSE (and slightly over) r Add 2 min to Tdelay-no other affect e

Add 2 min to Tdelay-r no other affect Plant is designed for SSE-should be little effect; 2 minutes to account for initial shock. Note, that if adding time delays for SPRA also increases the EPRI recommended floor values of dependency, this updated floor value for dependency is applied in the cognitive and execution recoYery portions of the HEP evaluation (this is applied in all cases where the EPRI recommended dependency level has changed, including for SEISI, SEISZ, and SEIS3 evaluations)

SEIS2 0.15-0.8 G2. G7 (ZO3:S and ZO4:S)

Accelerations greater than the SSE in which control room indication is not lost and the conffol room ceiling is still intact.

r Add 2 minute to Tdelay and r

increase cognitive workload and o

execution stress level to high o

If HCR/ORE Cognition, increase CP level to UB r

Add 2 minute to Tdelay and increase cognitive workload and execution stress level to high and r

increase Texe to 2x r

If HCR/ORE Cognition, increase CP level to UB Although control indication is still available seismic events greater than the SSE are likely to cause additional failures that would increase cognitive workload and stress as well as execution time

,iH NO NE od tr It

='

E F>, N G$

H EEHF q-L H.F t*ro\\

o tF, l+ -

UJ (nrtOs'

TABLE 5.4 SPRA HFE HEP EYALUATION PROCESS

SUMMARY

(coNTINUED)

Arrrcrs Frpr,o HFE Courrannrs Susprrc Ixrrmrrxc Evsnr(s)

ConnnspoNDs To FoR BVI ArnnCrS CRHFE Snrsurc Gnoup AppRoxrruarn G Lnvrl r

Add 15 minute to Tdelay and r

increase cognitive workload and execution stress level to high; r

use "monitored, not alarmed" for pcb, r

ro ERF recovery credit and o

increase Texe to 4x If HCR/ORE Cognition, increase CP level to UB When controls are being lost in the control room; there should be a step change. Difiicult to navigate to work site; many components already failed.

USE FLOOR OF 1E.02 FOR INDIVIDUAL HFEs 0.15-0.8 G2. G7 (ZO3:F and ZO4:S)

Accelerations greater than the SSE in which control room indication is lost but the conffol room ceiling is still intact.

r Add 15 minute to Tdelay and r

increase cognitive workload and r

execution stress level to high use "monitored, not alarmed" for pcb, o

no ERF recovery credit If HCR/ORE Cognition, increase CP level to UB SEIS3 High Accelerations Fail 1.0 Fail 1.0 Most CAT 1 buildings fail above l.0g Greater than 0.8 G8, G9, and Gl0 CR ceiling fails at about 0.70569 Catastrophic; failure of the contol room ceiling, failures of SSCs leading to direct core damage, or toxic failure of the propane tank farm.

Fail 1.0 Fail 1.0 SEIS4 Alt Gol - G10

{zo4:F, PT:TOX, or ZLI:F)

N)

RN slr C( >U IP t-r$ N

)-r=

I LH: F o:i o

Fl-CI)

\\osr F

Oa(s (oo l T

)fh UE r., s HO N9 od tr-ttE

2734294-R-036 Reaision 0 May 11,20L7 Page 91, of 145 Human Failure Events were also developed for the FLEX mitigation actions. These are not specific to the seismic PRA as they are designed for an extended LOOP scenario, and specifically accounts for high levels off plant damage and operator stress. The FLEX operator actions were developed using the same methodology as other internal events HFEs. Execution step durations were obtained from the timing validation study performed by BVPS. These actions are failed in the seismic model for the "SEIS4" or high acceleration group identified above.

The use of the sirme method from the internal events model for HRA dependency analysis is valid for the SPRA HRA. The SPRA HRA Notebook (Reference 36) discusses the method used to assess HFE dependency. The SPRA Quantification Notebook also has details of how the HRA dependency analysis was performed for the SPRA (Reference 17). The FENOC HRA Dependency Database (Reference 70) is used to determine the level of dependency between HFE Pairs assigned to the silme HRA seismic interval since only such pairs can appear in the same accident sequence; i.e., SEISl, SEIS2, ffid SEIS3. These pairs with other than zero dependence are then examined individually to see if the dependence need be included in the accident sequence model. Section 4.2 of Reference 17 discusses the HRA dependency analysis further.

Pre-initiator actions are not affected by seismic events and so were not changed from the internal events PRA model.

5.1.6 Seismic-Induced Floods The evaluation of seismic-induced floods was a compilation ofthree activities. First, the internal flooding PRA, Referenc e 27, was utilized to provide a risk-based screening of flood-significant scenarios. The second activrty was the use of the walkdown team to identiff flood sources in and around components that were on the SEL; this is documented in the Seismic Walkdown of BVPS-2, Reference 40. The third activity was to review the tanks not on the SEL and the " uu et" fire suppression system and do awalk-by ofthe components to determine if the assets would screen; this is documented in the SEL Notebook, Reference 32.

As discussed in Section 3.3.6 of Reference 40, the piping evaluationwas risk informed. The systems of interest and flood areas selected were those that had the greatest risk contributions as evaluated in the Internal Flood PRA. Table 3-4 of Reference 40 identifies six flood areas that merited specific walk downs. This table is repeated here as Tahle 5-5.

In addition to those pipe segments identified in Section 3.3.6 of Reference I (Seismic Walkdown), Table 3-4 of that document identifies six additional pipe segments that merited additional specific walk downs. This list was derived from a list of important flood scenarios minus those pipe segments that had previously been walked down. The list of top flood contributors was filtered according to flood location and the top flood areas were CB-5 and floods that could propagate to flood area CB-IA (both located in the Control Building, CNTB) and PA-5 which propagates to PA-3H (both located in the auxiliary building, AXLB).

During the plant walk downs, piping in general, and non-seismic piping in particular were examined to see if there were any unique vulnerabilities in proximity to any of the SSCs examined; see Referenc e 27. A summary of specific seismic-induced flooding interactions is provided in Section 3.3.6 of Reference 40. Appendix B of Reference 40 presents pictures and lEEGonsulting tlRtzzo

2734294-R-036 Reuision 0 May 11,2017 Pase 92 ofL45 the walkdown team's conclusions for the piping segments called out as having the highest conditional probability of core damage given a pipe break occurs.

TABLE 5.5 AREAS EVALUATED FOR RISK SIGNIFICANT FLOOD SCENARIOS Floon Anrcn Svsrnu Buu,nrnc Er-nvauoru CB.5 SERVICE WATER CB 735 PA.5 FIRE PROTECTION AXLB 773 SG.lN ALL SFGB 718 SG-T S ALL SFGB 718 SG.lNA ALL SFGB 737 SG.lSA ALL SFGB 737 5.1.7 Risk Significant Flood Scenarios As a supplement to the SSCs in the internal events PRA, a list of all tanks and coolers at the plant was obtained for review for potential seismic-induced flood sources. This list was reduced by excluding those tanks in plant rooms that contain no SSCs on the SPRA SEL, and to eliminate duplicates that are already on the SPRA SEL. The reduced list of potential flood sources is also shown as Table 3-6 in Reference 32.

The reduced list of potential sources was then filtered by building and those located in the turbine building were also then excluded. For the SPRA, no credit is taken for any equipment in the turbine building and failures do not propagate to adjoining buildings.

To ensure that no important tanks were missed, the SPRA SEL list of tanks, coolerslheat exchangers, and pumps (which have coolers) was reviewed. Those not already on the list were added if the tanks and coolers were not located in the yard or containment, and contained liquids rather than air.

The walkdowns performed by the Seismic Review Team screened these from further consideration either due to their seismic ruggedness, presence of dikes around the tanks, or lack of proximify to SEL components. All tanks were screened based on either: information provided in the internal events flooding analysis, or based on no impact to PRA equipment in the flood area, or too small of a flood source to cause an impact. The small coolers also were screened from either of these screening criteria.

The flood sources from tanks and heat exchangers, although technically screened, were sampled and walked down to validate the assumptions made for their screening. These include the fire protection engine cooler on the diesel driven pump and the spent fuel pool heat exchangers as examples.

No potential flooding sources have been identified for inclusion in the BVPS-2 seismic model.

5.1.8 Seismic-Induced F ires Appendix A in Reference 38 contains a white paper on the subject of seismic-induced fires. The presentation describes ways that seismic-induced fires may be screened, both qualitatively and lESGonsulting rlRtzT-o

2734294-R-036 Reaision 0 May 1"1,20L7 Page 93 of145 quantitatively from further consideration. The flow chart presented at the end of Appendix A in Reference 38, summarizes the variety of ways that screening can be performed on a fire compartment by compartment basis.

The following are some key conclusions from the suggested approach in Appendix A in Reference 38:

1.

The list of equipment of interest as potential flue sources caused by seismic events are:

a.

Tanks, Bottles, and Piping (including turbine-generator, auxiliary boiler) That Contain Hydrogen, Propane, ffid any Other Flammable Gases

b.

Above-Ground Tanks and Piping That Contain Diesel Fuel Oil

c.

Tanks, Equipment, and Piping That Contain Lubricating Oil Turbine-Generator Turbine Lube Oil Storage Tank r Oil-Filled Transfonners Pumps (especially large pumps)

Compressors e Piping

d.

Equipment with Electrical Wire or Bus Bar Connections at 480V and Above Pumps r Oil-Filled Transfonners o Compressors Switchgears/Buses/Jv1CCs r Others (e.9., other applicable NUREG/CR-6850 fire source bins from Fire PRA that are unique and significant for specific plants)

2.

Seismic-induced fires are believed possible only if structural failure of the SSC occurs; i.e., we neglect the functional failure limit if it is lower.

3.

Based on data from other industries, the conditional probability of fire ignition given seismic failure of a potential seismic-induced fire source is bounded by 0.1. An individual seismic-induced fire frequency leading to core damage for a single SSC of 1E-7 per year is assumed as sufficiently small as to be neglected. Due to frequency overlap between the potential seismic-induced fire and other contributions to core damage, a single, SSC seismic-induced fire frequency of 5E-7 per year is sufficiently small as to be negligible.

For this study of Beaver Valley, we adopt the above methodology conclusions and apply the qualitative and quantitative screening of potential seismic-induced fire sources, including the use of walkdown observations to eliminate seismic-induced fires from inclusion in the SPRA logic models. The case for this screening is provided below.

AESGsrsulting

()Rrzzo

2734294-R-036 Reaision 0 Moy 1.1,241.7 Pase 94 of745 Table 5.5-Z of Reference 38 presents the failure frequencies of SSCs with typical HCLPFs ranging from 0.1g to 2.0g. The total frequency column was obtained by summing the convolution of the Beaver Valley mean seismic hazard curve over all seismic intenrals. The frequency of seismic failure of 1E-7 per year corresponds to an SSC HCLPF ofjust greater than 1.0g.

However, this acceleration level has not yet accounted for the conditional probability of ignition given the SSC fails, or of the potential overlap of seismic-induced fires with other contributors to core damage. At Beaver Valley, the conditional core damage probability for accelerations of 0.7g and higher is close to 1.0. Therefore, seismic-induced fires at frequencies greater than 0.7g cannot add significantly to the CDF total. The HCLPF acceleration corresponding to a failure frequency of lE-7 per year, only from accelerations less than 0.7g is then between 0.559 and 0.69. This is an approximate approach, as other contributors to core damage at accelerations less than 0.7g do occur and so there is some potential overlap at lower accelerations that is not credited.

An ignitionprobability of 0.1 reducesthe frequency of SSC failures to justthose that also ignite, resulting in a fire. A corresponding HCLPF value just more than 0.359 would result in a potential fire source adding approximately lE-7 per year to the existing seismic CDF. We observe that this acceleration is selected conservatively both because of the potential for frequency overlap at accelerations less than 0.19, and because it is implicitly assumed by this screening calculation that any seismic-induced fire leads to core damage. Further, the results for the unconditional seismic-induced fire frequencies presented in Reference 3 8 do not yet include a scaling factor on the hazard exceedance curves to account for the plant availability factor. To do so would provide us additional margin. We therefore use 0.359 for an SSC HCLPF as the quantitative screening criterion for excluding potential seismic-induced fires.

Table 5-6 (reproduced from Reference 71) provides a list of the top 25 fire scenarios from the BVPS-2 fire PRA. Out of these 25 scenarios, CB-l and CV-1 fire areas were the dominant contributors and those areas were chosen to have a specific seismically-induced fire walkdown.

IESGonsulting rlRtzzo

273429+R-436 Reaision 0 May L1.,20L7 Page 95 of 1.45 TABLE 5.6 RISK CONTRIBUTING PLANT LOCATIONS FROM THE BEAVER VALLBY TINIT 2 FIRE PRA (REFERENCB 71)

InmraroR ScrxlRro DrscRrpuox Expaunrn DrscnrPTroN FCRIOI FIRE - Benchboard Section C 3-CR-1 Main Control Room, fire in BB section C3 FCRI 02 FIRE - Benchboard Sections A&B 3-CR-1, fire in BB sections A & B common raceway FTBl03 FIRE - T/G Fires Excitor, Hydrogen, Oil, & Catastrophic: FDSI/3i5 2-TB-I Turbine Building, turbine-generator-exciter fires FCB IH8 FIRE - Bin 5 z-CB-l Process lnstrument Rm & Cable Tunnel; worst single tray assumed ignited by cutting/weldine FCVI l5 FIRE - TS#5: FDS0 z-CV-l West Cable Vault & Rod Control Area; transient scenario 5 (see fire modeling)

FTBIO5 FIRE - TS#28 2-TB-1, transient scenario 28 (see fire modeling)

FCVI13 FIFE. TS#4A 2-CV-1, transient scenario 44. (see fre modeling)

FSB303 FIRE - TS#29 BINT : FDS3 2-SB-3 Service Building Cable Tray Area, Bin 7 transient scenario 29 (see fire modeling)

FCVI I I FIRE - TS#3: FDS0 2-CV-1, transient scenario 3 (see fire modeling)

FCV IO4 FIRE - 2DGP-3: FDSI/3 z-CV-1, source panel2DGP-3; fire contained within panel FSB302 FIRE - TS#29 BINT : FDS2 2-SB-3, Bin 7 transient scenario 29 (see fire modeling)

FSB446 FIRE - 4KVS-2C, Section 9-12 HEAF: FDS3/5 2-SB-4 Normal Switchgear, source bus 4KVS-2C breakers 9-12 high energy arcing fault; fire affects external tarqets FMCA.42 FIRE - Multi-Compartment:

Exposing 2-DG-l to 2-PT-l Multi-compartment fre scenario; fire engulfs EDG 2-1 compartment then spreads to engulf Pipe Tunnel FMS IO I FIRE. FULL COMPARTMENT:

Main Steam Valve Area 2-MS-1 Main Steam Valve Area; fire starts at any defined ignition source and assumed to burn whole room FCR103 FIRE - Benchboard Section C 3-CR-1, fire in BB section C common raceway FCB 1 F5 FIRE - RK-2RC-PRT-B: FDS2:

1. Incipient Detection Factor 2-CB-1, source RK-2RC-PRT-B; fire grows outside cabinet FRHIO I FIRE. FULL COMPARTMENT Switchyard Relay House 3-RH-l Relay House in the Switchyard; whole compartment assumed burned from any of the defined sources FRC I 08 FIRE. 767SE z-RC-1 Reactor Containment, Southeast section of 767'elevation FCB1M6 FIRE - TS#38: FDSI z-CB-1, ffansient scenario 38 (see fire modeling)

FCB 1A'6 FIRE - RK*2RC-PRT-A: FDS2:

Wlncipient Detection Factor 2-CB-1, source RK-2RC-PRT-A; fire grows outside cabinet FCB 1 K2 FIRE. TS#17 2-CB-1, transient scenario l7 (see fre modeling)

AESGonsulting rlRrzzo

2734294-R-036 Reaision A May 1"1.,201.7 Page 96 of 1.45 TABLE 5-6 RISK CONTRIBUTING PLANT LOCATIONS FROM THE BEAYER VALLEY T]NIT 2 FIRE PRA (REFERENCE 71)

(coNTINUED)

Iurrraron Scnnanro DrscRmrroF{

Exrawnun Dnscmrrron FCB 185 FIRE - RK*2PRI-PROC-I: FDS2:

w/Incipient Detection Factor 2-CB-1, source RK-2PRI-PROC-1; fire grows outside cabinet FCV105 FIRE - ZIHA*OCABCVI: FDSI/3 2-CV-1, source 2IHA-OCABCVI; fire contained within cabinet FCB 1P7 FIRE - TS#44: FDS1 2-CB-1, transient scenario 44 (see fire modeling)

FCV102 FIRE - PNL*DC2-15: FDS2 2-CV-1, source PNL-DC2-15; fire grows outside cabinet With the quantitative screening criterion established, the potential fire sources previously screened in qualitatively for assessment, according to the arguments of Appendix A in Reference 38, were addressed.

1.

Tanks, bottles, and piping (including turbine-generator, auxiliary boiler) that contain hydrogen, propane, and any other flammable gases.

The flammable gases in the nuclear plant consists basically of hydrogen. It is used as a cover gas on the generator. The gas for the generator is in the yard well away from the plant structure itself and the generator is in the turbine building. We screen potential sources in the turbine building because no credit is taken for SSCs within the turbine building for seismic events.

Hydrogen used for chemistry analysis was screened based on the small quantity involved and the lack of risk significant equipment in the vicinity.

Similarly, we screened potential sources in the yard, since even if they seismically fail, they will not impact other SSCs that are credited.

Above-Ground tanks and piping that contain diesel fuel oil.

Tanks, equipment, and piping that contain lubricating oil.

r Turbine-Generator e Turbine Lube Oil Storage Tank

. Oil-Filled Transfoffners Pumps (Especially Large Pumps)

Compressors

. Piping Table 5-7 below lists potential fire ignition sources at BVPS-2 not included in the SEL. These items were all part of the walkdown and evaluated for their potential to become a seismically-induced fire. The oil and grease sources on the list were part of the larger component and all screened with a HCLPF of greater than 0.3.

fEtGonsulting rlR.zzo 2

J

TABLE 5.7 HYDROGEN AND FLAMMABLB LIQUID IGNITION SOURCES tt jgo NO NE od E

It 5'

E IcrurrroN Souncp ID Icruruorv Souncr Burr,onvc ErnvlrroN Annd Roou Hvunocnu oR Fr,nupranls Lroum Lo.Lnmrcs Fmr Courpanruuxr 2-GMH.

TK21A.B.C.D.E,F,G,H Unit 2 Yard 730 8 cylinders Bulk Hydrogen Storage Tanks Large 2-H-l Misc. Hydrogen Piping Unit 2 Auxiliary Building 735'.-755',

Yard, Auxiliary Building, and Turbine Buildine Large 2-PA-4 TR.2A Unit 2 Yard 730 Transformer 2A Laree 2-TR-5 TR-MT.2 Unit 2 Yard 730 Unit 2 Main Transformer Large 2-TR-1 TR.2C Unit 2 Yard 730 Transformer 2C Laree 2-TR-2 TR.2D Unit 2 Yard 730 Transformer 2D Laree 2-TR-3 TR.2B Unit 2 Yard 730 Transformer 28 Large 2-TR-4 TR.2A Unit 2 Yard 730 Transformer 2A Large 2-TR-5 TR.MT.2 Unit 2 Yard 730 Unit 2 Main Transformer Large 2-TR-t TR.2C Unit 2 Yard 730 Transformer 2C Large 2-TR-2 TR.2D Unit 2 Yard 730 Transformer 2D Large 2-TR-3 TR.2B Unit 2 Yard 730 Transformer 28 Large 2-TR-4 TR.2A Unit 2 Yard 730 Transformer 24.

Large 2-TR-s 2ABM.B14.

Unit 2 Auxiliarv Boilers Buildine 735'.

South -Lower Drum Manway Large 2-SOB-l 2ABM.B 18 Unit 2 Auxiliarv Boilers Buildins 73s'.

Large 2-SOB-1 T/G Excitor Unit 2 Turbine Buildins 730' Turbine-Generator Excitor Large 2-TB-1 T/G Hydrogen Unit 2 Turbine Buildins 730' Turbine-Generator Hydrogen Large 2-TB-t T/G Oir Unit 2 Turbine Buildine 730'.

Turbine-Generator Oil Large 2-TB-1 2TMB-P2O9A Unit 2 Turbine Building 73O',

E.H. Fluid Pump 40 hp -

Mezzazine Large 2-TB-1 2TMB.P2O9B Unit 2 Turbine Building 730',

E.H. Fluid Pump 40 hp Mezz.azine Large 2-TB-l 2EGS+EGz.1 Unit 2 Diesel Generator Building 732 Emergency Diesel Generator 2-l Fuel and Luhe Oil 1200 eat 2-DG-l 2EGS*EG2.2 Unit 2 Diesel Generator Building 732 Emergency Diesel Generator 2-2 Fuel and Lube Oil 1200 eal 2-DG-2 lRG.EG-I Common Emergency Response Facility 735 Ensine ERFS Diesel Ensine Larse 3-ER-z FP.P.2 Common Intake Structure 705' Diesel Ensine Driven Fire Pump Large 3-IS-4 Pumps; e.g., 2SWS-P2IA"B.C Unit 2 Intake Structure 705' Service Ater Pump Cubicles 450 gal. Fuel Oil, 27 Lube Oil 2-IS-4 Pumps and oxy-acetylene Common Intake Structure 705'.

Intake Structure General Areq 705' 165 gal. Fuel Oil, l6 Lube Oil, Oxy-Acetylene Weldine Cart 3-rs-6 F>, N N.\\W qeS.

ts SEgF qLF.ry HOT

(}

rf lr-(JJ (Jr\\Oo\\

2734294-R-036 Rrrt,sion 0 May 1L,201.7 Page 98 of 1,45 Piping containing lubricating oil and hydraulic oil are mostly in the turbine building. The SSCs within the turbine building are not credited in the SPRA and so such pipes in the turbine building are screened. All pipes examined in the SPRA were found to have high capacity, ffid so were screened from further consideration of seismic-induced fires.

4.

Equipment with electrical wire or bus bar connections at 480V and above.

Regarding switchgetr, buses, and MCCs, a walkdown was performed to examine these equipment items focusing on the potential for their structural failures leading to a signifi cant seismic-induced fire.

Both the Division I and Division 2 switchgear rooms were walked down due to these zones being significant contributors to CDF in the Fire PRA and because they could possibly have a high energy arcing fault.

Seismic-induced fire would require both overturning of switchgear and severing of top lines. Top conduits are rigidly braced to the wall. No potential interactions were observed that would puncture/sever top conduits, so the most likely failure mode is judged to be structural/anchorage failure resulting in switchgear overturning and severing of conduits. Preliminary calculations determined a HCLPF >0.309 for structural (anchorage) failure that would be required to initiate overturning. Those preliminary calculations conservatively do not credit the restraint added by the top conduit bracing to prevent overturning, so the actual structural capacity of the component is higher. The transformers in the area are dry type.

The high voltage switchgear in both rooms were all well supported and the potential for any differential movement between the switchgear and the conduits that enter and exit appeared to be minimal thus reducing any potential high energy arcing fault.

480V transformers are used throughout the plant to step down power to a l20vac lighting panel. These were determined to be seismically robust.

No potential seismically-induced fire sources were identified for inclusion in the SPRA.

This conclusion is further supported by the review documented in Reference 72.

5.2 SPRA Plnur Sntsnatc LocIc MonBr, TacururcAt, Aureu,q,cv The BVPS-2 SPRA seismic plant response methodology and analysis were subjected to an independent peer review against the pertinent requirements in the PRA Standard (Reference 4).

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the BVPS-2 SPRA seismic plant response analysis is suitable for this SPRA application.

5.3 Srtsmrc Rrsr< QUaFITIFICITIoN 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 radioactivrty to the environment. This section describes the SPRA quantification methodology and important modeling assumptions.

lE$Gonsulting riRtTzo

273U9+R-036 Reaision 0 May 11,20L7 Page 99 of 745 5.3.1 SPRAQuantificationMethodolory For the BVPS-2 SPRA, the follo*ing approach was used to quantify the seismic plant response model and determine seismic CDF and LERF.

The computer codes used by the BVPS-2 PRA are available from ABSG Consulting Inc.

(ABS Consulting) which is the developer of the RISKMAN software. Technical support and quality assurance are provided by ABS Consulting. The software is classified as Category B software per the FENOC Adminishative Program for Computer Related Activities (Reference 73), and has been site accepted per that program.

5.3.1.1 RISKMAIYTM Software RISKMAN 14.3 was used in the creation and maintenance of both the internal events PRA and in this SPRA, Version 14.3 was also used in the development or the Interval Events PRA.

Version 14.3 was used forthe SPRA and is also now usedto maintainthe internal events PRA models. The features and code limitations of RISKMAN are described in Reference 69 and its companion manuals for each of the main modules.

5.3.2 SPRA Model and Quantilication Assumptions The following assumptions were made as part of the seismic PRA quantification:

1.

The quantification of CDF and LERF sequences is performed by a large, linked-event tree model in which the seismic acceleration intervals are evaluated successively and then the computed frequencies added.

2.

The seismic impacts on types of SSCs represented in the SPRA model are limited to those identified in Tables 5.2 1 of Reference 38.

3.

Screening criteria for the need to include SSCs within the SPRA model were set at 0.lg HCLPF for all SSCs and up to 2.0g for SSCs related to LERF.

4.

In the base-case SPRA model, the assignment of human error probabilities for each HFE is dependent on the associated acceleration range from one of four HRA seismic intervals for which the human effor probability (HEP) is being evaluated (see Reference23).

5.

The base-case accident sequence quantification cutoff used was lxl0-14 per year, for both CDF and LERF. The sensitivity analyses were performed using a sequence frequency cutoff of lxl0-12 per year. See Section 4.3 of Reference 38 for a discussion of CDF and LERF convergence.

5.4 SCIIF Rnsulrs The seismic PRA performed for BVPS-2 shows that the point estimate mean seismic CDF is 8.78x10-06. A discussion of the mean SCDF with uncertainty distribution reflecting the uncertainties in the hazard, fragilities, and model data is presentedin Section 5.6. Important contributors are discussed in the following paragraphs.

The top SCDF accident sequences are documented in the SPRA quantification (Reference 17).

These are briefly sunmarized in Table 5-8.

lBEGonsuEing

[]R]zzo

2734294-R-036 Rwision 0 May LL,20L7 Page 100 of 1.45 TABLE 5-8

SUMMARY

OF TOP SCDF ACCIDENT SEQUENCES Raltr lxrrr.+,ro R

IE FRseurxcv CDF/Yn,m Yo Or SCDF Sneurxcp PRocRrssroN Dnscnrprrou I

G0s 0.5-0.69 4.0316E-06 2.08E-07 2.37%

This sequence is initiated by an earthquake between 0.5g and 0.6g. It causes failure of power from the offsite grid, the failure of all EDGs and a very small LOCA to occru. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage.

2 G06 0.6-0.7g 2.2086E-06 1.92E-07 2.18%

This sequence is initiated by an earthquake between 0.69 and 0.7g. It causes failure of power from the offsite grid, the failure of all EDGs and a very small LOCA to occur. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage. This sequence is essentially the same as that for sequence l.

J G04 0.4-0.5g 7.9900E-06 1.20E-07 137%

This sequence is initiated by an earthquake between 0.4g and 0.5g. It causes failure of power from the offsite grid, the failure of all EDGs and a very small LOCA to occur. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage.

4 G07 0.7-0.8g 1.3307E-06 1.06E-07 l.2lo/o This sequence is initiated by an earthquake between 0.7g and 0.8g. It causes failure of power from the offsite grid, the failure of all EDGs and a very small LOCA to occur. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage.

5 G08 0.8-1.0g t.4t62E-46 8.14E-08 0.93%

This sequence is initiated by an earthquake between 0.8g and 1.0g. It causes failure of power from the offsite grid, the failure of all EDGs and a very small LOCA to occur. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage.

6 G06 0.6-0.79 2.2086E-06 6.32E-08 0.72%

This sequence is initiated by an earthquake between 0.69 and 0.7g. It causes failure of power from the offsite grid and relay chatter resulting in loss ofboth EDG 2-l and 2-2. Avery small LOCA also occurs. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage. This sequence is similar to that for sequence l.

7 G05 0.5-0.69 4.0316E-06 5.62E-08 0.64Vo This sequence is initiated by an earthquake between 0.5g and 0.69. It causes failure of power from the offsite grid and relay chatter resulting in loss of both EDG 2-l and 2-2. Avery small LOCA also occurs. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage. This sequence is similar to that for sequence l.

lESGonsulting rlRtzzo

273429+R-036 Rwision 0 May 1L,201.7 Pase 101 of145 Raur Iurrraro R

IE FnnounNcv CDF/YNAN 7o Of SCDF SneurucB PnocnnssroN DsscRrprrox I

G08 0.8-1.0g 1.4162E-06 5.55E-08 O.63Vo This sequence is initiated by an earthquake between 0.8g and 1.0g. It causes failure of power from the offsite grid, the failure of all EDGs and a sma(( LOCA to occur. The FLEX portable generator is also failed. The failure of all AC power with a LOCA precludes recovery resulting in core damage.

This sequence is similar to sequence 5 except that here the LOCA is small rather than very small.

I G09 1.0-2.0g 1.0829E-06 5.07E-08 0.s8%

This sequence is initiated by a very strong earthquake between l.0g and 2.0g. It causes direct core damage and LERF by either failure of the Reactor Containment Building, MSCV, or gross failure of the steam generators. There are no recovery and core damage results.

10 Gr0 2.0-4.999 8.6694E-08 4.88E-08 0.56%

This sequence is initiated by a very strong earthquake between 2.0g and 4.0g. It causes direct core damage and LERF by either failure of the Reactor Containment Building, MSCV, or gross failure ofthe steam generators. There are no recovery and core damage results.

TABLE 5-8

SUMMARY

OF TOP SCDF ACCIDENT SEQUENCES (coNTTNUED)

SSCs with the most significant seismic failure contributions to SCDF are listed in Table 5-9, sorted by FVI. The seismic fragilities for each of the significant contributors are also provided in Table 5-9, along with the corresponding limiting seismic failure mode and method offragility calculqtion. FVI values for seismic equipment groups were calculated using RISKIvIAN's "Fragile Component Importance Report," for Sequence Group SEIS and Master Frequency File R6AIMP. Table 5-9 shows the top 27 seismic equipment groups, sorted by FV. It was revealed that setting various operator actions to guaranteed failure, with a value of 1.0 (common in the SPRA), was not allowing success sequences to be quantified, and thus there were FV values that were not being calculated appropriately. Sensitivity Case 38 was devised, in which the human actions in the model that had been set to 1.0 were reset to 0.999, ffid the model was quantified. The importance displayed in the following tables use the results from Sensitivity Case 38.

The fragilities reflect the outcome of the refinement process outlined in Section 4.4.2.16.

Among the top SCDF contributors are very small LOCA (VSLOCA), the containment sump screens, transformers for offsite grid connection, recirculation spray coolers, and relay chatter of the AR440AR type relay.

lESGonsulting

(}Rtzz.o

2734294-R-036 Reoision 0 May 1.1.,201-7 Page 1.02 of 1.45 Loss of offsite grid is associated with brittle failure of the cerirmic insulators on transformers.

This is assigned a 0.1g HCLPF which is conservative, but is the recommended HCLPF based on EPRI SPRAIG Report 3002000709 (Reference 15), NUREG-1738 (Reference 59), and NUREG-CR-3558 (Reference 60) reports. The Seismic-Induced Very Small LOCA is associated with the failure of NSSS Piping assumed to occur at the bottom of the reactor pressure vessel. This failure mode is assigned a 0.1259 HCLPF based on the BVPS-2 SSE PGA, which is aligned with Option 3 of Section 5.4.4.2 in the EPRI SPRAIG Report 3002000709 (Reference 15). These two contributors are important to CDF because together they provide a challenge to the plant of providing makeup to the reactor after a LOCA occurs, but both are identified as using industry accepted methodology to obtain the HCLPF values.

The seismic failure containment sump screen or the correlated failure of the recirculation spray heat exchangers fail the ability to recirculate water from the containment sump and back into the reactor coolant system, given a LOCA. The containment sump has a calculated HCLPF of 0.299 with the failure mode of base metal shear stress. The HCLPF of the recirculation spray coolers is conseryatively assumed to be 0.309, failing by the structural or anchorage failure mode, based on other Seismic Category I tanks. Although this conservatism plays a role in the calculated failure of the recirculation ability, the calculated lower HCLPF of the containment sump screens accounts for the majority of this failure.

The AR440AR relay has a calculated HCLPF of Q.429, with a failure mode of chatter. The relays in question are located in reactor protection system panels BV-RK-2RC-PRT-A and -8, and are used to actuate multiple important pumps and motor-operated valves across the plant.

Correlated chatter of these relays is modeled as failing all service water pumps, all high-head injection pumps, valves for the ability to depressurize the RCS for RHR entry, and all recirculation spray pumps.

lESGonsulting

{}R.zzo

2734294-R-036 Reuision 0 May 1.1.,201.7 Page 103 of 1.45 TABLE 5.9 SCI}F IMPORTANCE MEASURES RANKED BY FV RAFrx Gnour Top Evnxr DescRrprrou FV HCLPF (s)

Au (c) pr

$u FnrluRn Moor FRlcrlrrv MrcrHon I

EQ55 ZVS VSLOCA 1.38E-01 0.125 0.32 0.24 0.32 See Note (l)

See Note fl) 2 EQl1 ZSM 2RSS.SSCIOI (CNMT Sump Screens) 5.328-02 0.29 0.72 0.24 0.32 Base metal shear stress CDFM J

EQs3 ZOG TR-2A;B;C;D &

TRF-2-5J;K 4.878-02

0. r0 0.25 0.24 0.32 Conservative low HCLPF of 0.10g assigned based on seismic category Assigned 4

EQTl ZSM 2RSS.

E2lA;B;C;D (Recirc Spray Coolers) 4.14E-02 0.30 0.76 0.24 0.32 Stnrctural Assigned 5

EQl02 ZP-z Pump and MOV Relays AR440AR in Panels BV-RK-2RC.PRT-AIB 3.57E-02 0.42 1.06 0.24 0.32 Relay chatter CDFM 6

EQs6 ZLK SLOCA 3.47E.02 0.32 1.00 0.30 0.40 See Note (2)

See Note Q\\

7 EQl r4B ZR4B D3 Relay for EDG 2.2 3.268-02 0.37 0.93 0.24 0.32 Relay chatter CDFM I

EQ3e ZRW 2QSS-TK2r (RWST) 3.05E-02 0.4s 1.02 0.24 0.26 Overtuming of tank CDFM 9

EQ42 ZDG Diesel Gen Bldg Supply Fans (2HVD-FN270A"B) 2.t8E-02 0.46 1.27 0.24 0.38 Functional failure due to shaft binding and failure of attached duct work CDFM r0 EQs2 ZSW Standby Service Water Pumps &

A]SX 2.t6E.02

0. l0 0.25 0.24 0.32 Conservative low HCLPF of 0.10g assigned based on seismic catesory Assigned l1 EQ76 ZDG 2EGF-TK2lA;B (EDG Fuel Oil Storage Tanks) 1.63E-02 0.50 l.l3 0.24 0.26 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (3) t2 EQl10 ZP(z Master relay (MidTe#AEMCO tvpe-1 56) 1.62E-02 0.48 t.2t 0.24 0.32 Relay chatter CDFM l3 EQ77 ZDG Dampers for EDG Support 1.45E-02 0.51 1.27 0.24 0.32 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (3) t4 EQe0 ZDG 2EGF-LI5203A;B

& 204A;B (EDG Day Tank Level Switches) 1.45E-02 0.51 1.27 0.24 0.32 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (3) 15 EQI l4A ZP.4A D3 Relay for EDG 2-I 1.42E-02 0.29 0.73 0.24 0.32 Relay chatter CDFM IEEGoneuEing tlRrzzo

273429+R-036 Rasision 0 May 11,201-7 Pase 1.04 of 145 TABLE 5.9 SCDF IMPORTANCE MEASTJRES RANKED BY FV (coNTTNUBD)

Raux Gnoup Tor EvrNr DnscRrprroFr FV HCLPF (s)

Atrl

{c) pr pu Farlunr Monn Fmcu,rrv MrrHon 16 EQe4 ZSz Valve Pit &

2SWS-PTl l3A;B (SWS Pump Dschg Press Transmitter) l.3lE-02 0.50 l.l3 0.24 0.26 Pragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

Sse Note (3) t7 EQl 17 ZIs{.6 MCC.2.EO6, E14 t.298-02 0.38 0.96 0.24 0.32 Auxiliary contact chatter, main contact chatter, control relay chatter, and contactor change-of-state CDFM l8 EQ120A ZS2 SWS Underground Piping l.l7E-02 0.51 1.27 0.24 0.32 Structural Assigned Screening Threshold -

See Note (3) l9 EQse ZS2 SWS Metal Expansion Joint Headers t.t7E-02 0.5r 1.27 0.24 o.32 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (3) 20 EQ78 ZS2 Dampers for Service Water Support t.t7E-02 0.51 t.27 0.24 0.32 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (3) 2l EQl08 ZLI MLOCA 1.138-02 0.54 2.00 0.35 0.45 See Note (2)

See Note (2\\

22 EQr3 ZMA 2FWE.LCVIO4A (Level Control for DWSr) 7.r7E-03 0.10 0.25 0.24 0.32 Conservative low HCLPF of 0.10g assigned based on seismic category Assigned 23 EQ60 ZMA Floor-Mounted Instrument Racks 7.178-03 0.10 0.25 0.24 0.32 Conservative low HCLPF of 0.10g assigned based on seismic category Assigned 24 EQr03 ZR3 Pump Relay RK223068-AP in SWGR BV.

4KVS.2AEIDF 6.46E-03 0.55 1.39 0.24 0.32 Relay chatter CDFM 25 EQI 12 ZLI Polar Crane in CTMT s.36E-03 0.60 1.90 0.30 0.40 Failure of brake system Assigned based on similarity to BVPS-I polar crane 26 EQ88 Z,\\F 2FWE.

FEl0lA;B;C (300 GPM Cavitating Venturi) 5.19E-03 0.51 1.27 0.24 0.32 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (3)

ABSGonsulting rlRtzzo

2734294-R 036 Rmision 0 May'1,'1,,20L7 Pase L05 of145 TABLE 5.9 SCDF'IMPORTANCE MEASURES RANKEI} BY FV (coNTTNUED)

Notes:

(1) The fragility for VSLOCA is assumed to have a HCLPF equal to the BV2 Site SSE based on Section 5.4.4 of the EPRI SPRA Implementation Guide.

(2) The fragility for SLOCA and MLOCA is assigned based on following Table H-2 of the EPRI SPRAIG (EPRI 3002000709). The fragilities in Table H-2 are considered to be representative fragilities based on a survey of available industry information. The failure mode specified is the RCS boundary failure.

(3) Assigned Screening Threshold means that the SSCs were determined to be sufficiently seismically rugged as determined from plant walkdown to conservatively assign a screening level HCLPF which initially was 0.5g The most significant non-seismic SSC failures (e.g., random failures of modeled components during the SPRA mission time) are listed rn Table 5-10.

Reference 17 contains the FV and RAW values for each component modeled in the SPRA, for both CDF and LERF sequences. Components were determined to be significant if the component's RAW is greater than 2 or its FV is greater than 0.005 for either CDF or LERF sequences, per the definition from the PRA Standard (Reference 4). RISKMAN report "Component Importance, With Common Cause and Maximum BE RAW" was used for FV, and "Component Importance, Without Common Cause and Maximum BE RAW" was used for RAW, created using the SEISLI sequence group for CDF data. Judging against the above criteria, there were nine risk significant components for CDF sequences. Note that only two components are important based on FV criteria. These were failures of the BV-2EGS-EG2-2 diesel and of the BV-FLEX-GEN-002 generator. The other risk significant components exceeded only the RAW criterion of 2.0 for risk significance. These risk significant components included the station batteries (i.e. BV-BAT-2-l and BV-BAT-2-2) and components associatedwith DC power; e.g., BV-DC-SWBD2-1, BV-DC-SWBD2-2,BAT-BKR2-I-S\\VGR, and BAT-BKR2-2. The importances presented in Tahle 5-1A also use the results from Sensitivity Case 38.

lE$Cotrsulting rlRtzzo RaFrx Gnour Top Ewru IlBscnrprrou FV HCLPF (s)

Awr (c) pr pu F,lrlunr Monr FRecrlrrv Mnruon 27 EQl26 ZMz MCC.2.EI2 5.13E-03 0.46 0.93 0.24 0.18 Auxiliary contact chatter, main contact chatter, control relay chatter, and contactor change-of-state CDFM

2734294-R-036 Rwision 0 May'1,1,201.7 Page 1.06 of 1"45 TABLE 5.10 NON-SBISMIC SIGNIFICANT COMPONENT LIST (SORTEI) BY SCDF FVI)

The contribution of each category of initiating events to the total CDF wi$ calculated and is summarized in Table 5-11 below. The table is sorted by the hazard range of the initiators.

Initiating event category contribution was determined by using RISKMAN's "Contribution of Initiating Events to One Sequence Group" report, using the Master Frequency File R6AMFF with Sequence Group SEISLI, at a report cutoff of lE-14, after quantification truncation of 1E-14.

TABLE 5.17 INITIATING EVENT CONTRIBUTION TO SCDF The major initiating events contributing to core damage from seismic are G04 through G09.

This range of hazards accounts for about 93% of CDF. The DBE for BVPS is 0.1259; which is within the G01 initiator range of accelerations. By contrast, such seismic events contribute much less than 1 percent of the total.

fESG+nsulting 7\\RIZZ'O

\\"J Courounrur ConaroFIrITr I}TSCRIPTION SCI}F FV SCDF RAW BV.2EGS.EG2.2 Emergency Diesel Generator 1.93E-02 l.7lE+00 BV.FLEX.GEN.OO2 FLEX 480V Generator 5.23E-03 I.l4E+00 BV-8AT.2.2 Battery 1.208-03 3.59E+00 BAT.BKR2-2-SWGR 125 VDC Battery Breaker Switchgear 2.58E-05 2.76E+00 BV.DC.SWBD2.2 125VDC SWBD 2.58E-05 2.76E+00 BAT-BKR2-I.SWGR 125 VDC BATTERY BREAKER SWITCHGEAR l.4lE-0s 2.09E+00 BV-DC-SWBDz-I 125VDC SWBD I.41E-05 2.09E+00 BV.BAT.BKR2.2 BAT-2-2 Output Isolation Breaker 2.90E-06 2.768+04 BV-2FWE-TK2IO PPDWST 6.798-07 2.49E+00 Inrunron Haznno Rlucn (s)

FnneurNCY CDF o/o ConrmnurroN CumuutrrvE CI}F G01 0.06-0.15 5.38E-04 l.s2E-08 0.t7%

l.s2E-08 G02 0.15-0.25 1.10E-04 9.20E-08 l.0s%

1.07E-07 G03 0.25-0.4 3.35E-05 4.20E-07 4.78%

5.278-07 G04 0.4-0.5 7.998-46 1.05E-06 1r.96%

1.58E-06 G0s 0.5-0.6 4.03E-06 1.70E-06 19.38%

3.28E-06 G06 0.6-0.7 2.2t8-06 1.68E-06 19.16%

4.96E-06 G07 0.7-0.8 1.33E-06 1.25E-06 14.16%

6.21E-06 G08 0.8-1.0 1.42F.-06 l.4lE-06 16.03%

7.62E.06 G09 1.0-2.0 1.08E-06 1.08E-06 12.32%

8.70E-06 Gl0 2.0-4.99 8.67E-08 8.67E-08 0.99%

8.78E-08 Total 0.06-4.99 6.99E-04 8.78E-06 100%

2734294-R-036 Rsuision 0 May LL,2017 Pase 1.07 of 745 In addition to examining the sequences that contribute to CDF, it can be useful to identiff the systems that are most important. One measure of importance can be determined by evaluating the effect on CDF if the system is assumed to have perfect reliability. This allows the systems to be ranked according to their contributions to overall CDF; i.e., the larger the impact on CDF if the system were perfect, the larger the contribution to the base-case CDF due to the failure of that system. This is a common importance measure, ffid is referred to as FV Importance (FVI).

System FV values were calculated using the data from RISKMAI\\I'S "Component lmportance, With Common Cause and Maximum BE RAW" report, created using the SEISLI sequence group and Master Frequency File R6AIMP. Each component is then grouped into its Maintenance Rule system, and the component FVs for each separate system are added together to determine overall system FV values. The systems modeled in the PRA with a FV greater than or equal to 1E-05 are listed inTable 5-12, sorted by largest FV value. The importances displayedin Table 5-12 use the results from Sensitivity Case 38.

TABLE 5.12 SYSTEM IMPORTANCE BY FUSSELL-\\TESELY Rnxx Svsrpnn #

DnscrurrroN FV I

36A Emergency Diesel Generators & Support Systems 2.40E-02 2

JI 480 Volt Station Service System r.328-02 J

24F, Auxiliary Feedwater System 8.23E-03 4

368 4KV Station Service System 6.428-03 5

30 Service Water System 6.04E-03 6

39 125 VDC Distribution System 4.19E-03 7

13 Containment Depressurization System 3.97E-03 I

44F Area Ventilation Systems - Miscellaneous 3.48E-03 I

l1 Safety Injection System 1.90E,03 l0 06 Reactor Coolant System 9.91E-04 The most important system is the EDGs and Support System. The EDGs would be called upon following a LOOP which is probable after a seismic event.

Reference l7 sunmarizes the contribution to seismic CDF from the most significant post-initiator human actions. Per Reference 4, significant post-initiator operator actions are defined as those operator action basic events that have a FVI value greater than 0.005 or a RAW greater than 2. The importance measures were calculated in RISKMAN and generated through the Basic Event Importance Report for Sequence Group Report in the Event Tree Module.

Reports were generated for the Sequence Group SEISL1 (seismic CDF) and the Operator Action Events were pulled out to make the summary tahle in Appendix J of Reference 17. Appendix J of Reference 17 also uses importances from sensitivity Case 38, however operator actions that are guaranteed failed in the seismic model are excluded. Judging against the above criteria, only one operator action was found to be risk significant, and this action exceeded the criterion only for FVI to CDF. The top action is for the operators failing to align the service water system emergency flow path, given a seismic event greater than the plant SSE in which control room AESGonsuEing

()Rtzzo

273429+R-036 Reuision 0 May 11,20L7 Pase 1.08 of 145 indication is not lost and the control ceiling is intact. This action is to provide makeup to the PPDWST which is the source of auxiliary feedwater.

5.5 SLERF Rrsur,rs The seismic PRA performed for BVPS-2 shows that the point estimate mear seismic LERF is 2.668-7. A discussion of the mean SLERF with uncertainty distribution reflecting the uncertainties in the hazxd, fragilities, and model data is presented in Section 5,6. tmportant contributors are discussed in the following paragraphs.

The top SLERF accident sequences are documented in the SPRA quantification report (Reference 17). These are briefly sunmarized in Table 5-13.

lESGonsulting rlRlzza

2734294-R-036 Reuisian 0 May LL,20L7 Pase 1.09 of 145 TABLE 5-13

SUMMARY

OF TOP SLERF ACCIDENT SEQUENCES Raxx Iurrmrruc Evrxr IE Fnueunxcv SLERF/vn PnRCrxr OF SLERF Sreusncv PRocnassroN DpscmrrroFr 1

G09 1.0-2.09 r.08E-06 s.07E-08 19.05V" This sequence is initiated by an earthquake between l-0g and 2.0g. Seismic failures lead to both core damage and a large early release directly. The most limiting seismic failure is the seismic failure of the SGs.

A large release path is assumed to be caused by the failure of the SGs, either by a direct opening or by overpressure of the containment caused by rapid discharge of both the primary and secondarv coolant.

2 Gl0 2.0-4.99g 8.67E-08 4.88E-08 18.34%

This sequence is initiated by an earthquake between 2.0g and 5.0g. Seismic failures lead to both core damage and a large early release directly. The most limiting seismic failure is the seismic failure of the SGs.

A large release path is assumed to be caused by the failure of the SGs, either by a direct opening or by overpressure of the containment caused by rapid discharge of both the primary and secondary coolant.

This sequence is similar to sequence l.

J Gr0 2.0-4.99s 8.67E-08 1.29E-08 4.85Vo This sequence is initiated by an earthquake between 2.0g and 5.0g. Direct core damage occurs by seismic failure of the control rods to insert. Power from offsite is failed seismically. The EDGs and the service water intake structure also fail seismically causing a station blackout. A small LOCA and failure of the RWST and all 3 AFW pumps are also failed caused by seismic motion. Core damage occurs. One of the large containment peneffations (e.g. personnel airlock) also fail. The large containment penetration failure provides a relsase path resultinE in a large, early release.

4 G08 0.8-1.0g 1.42E-06 4.10E-09 1.54o/o This sequence is initiated by an earthquake between 0.8g and 1.0S. Seismic failures lead to both core damage and a large eady release directly. The most limiting seismic failure is the seismic failure of the SGs.

A large release path is assumed to be caused by the failure of the SGs, either by a direct opening or by overpressure of the containment caused by rapid discharge of both the primary and secondary coolant.

This sequence is similar to sequence l.

fSSGonsulting tlHlz71e^

2734294-R-036 Reuision 0 May 1.1.,201.7 Page 110 of145 TABLE 5-13

SUMMARY

OF TOP SLERF ACCIDENT SEQUENCES (coNTTNUED)

Ratvr INIrrnrNc EvBNT IE FnneunNcv SLERF/vn Pnncnur OF SLERF Sneunucv PRocnrssroN DrscRrrrrox 5

Gl0 2.0-4.99g 8.67E-08 3.68E-09 1.387o This sequence is initiated by an earthquake hetween 2.0g and 5.0g. Seismic failures lead to both core damage and a large early release directly. The most limiting seismic failure is the seismic failure of the SGs.

Additional seismic failures also result. Namely the offsite propane tank farm fails releasing a toxic gas towards the plant. All operator actions are conservatively assumed failed. Other seismic failures include; failure of the EDGs, failure of the service water intake structure, failure of the RWST, failure of all 3 AFW Pumps, and the occurrence of a small LOCA. A large release path is assumed to be caused by the failure of the SGs, either by a direct opening or by overpressure of the containment caused by rapid discharge of both the primary and secondary coolant.

Several other seismic failures also occur (e.g. failure of large penetrations such as the personnel airlock). This sequence is similar to sequence l.

6 G09 1.0-2.0g 1.08E-06 2.60E-09 0.98%

This sequence is initiated by an earthquake between l.0g and 2.0g. Seismic failures of the offlsite tank farm leads to a vapor cloud explosion which is assumed to cause an extemal overpressure of the containment and core damage directly. A large, early release path results. The model does not consider other seismic failures once it is determined that there is a direct core damase and laree. earlv release event.

7 Gl0 2.0-4.99g 8.67E-08 2.34E-09 0.88%

This sequence is initiated by an earthquake between 2.0g and 5.0g. Direct core damage occurs by seismic failure of the control rods to insert. Power from offsite is failed seismically. The EDGs and the service water intake structure also fail seismically causing a station blackout. A small LOCA and failure of the RWST and all 3 AFW Pumps are also failed caused by seismic motion. Core damage occurs. One of the large containment penetrations (e.g. personnel airlock) also fail. The large containment penetration failure provides a release path resulting in a large, early release.

lEEGonrulting tlRtzzo

273429+R-436 Reaision 0 May L1.,201.7 Page'1,11 of 145 TABLE 5-13

SUMMARY

OF TOP SLERF'ACCIDENT SEQUENCES (coNTTNUED)

SSCs with the most significant seismic failure contribution to SLERF are listed in Table 5-14, sorted by FVI. The seismic fragilities for each of the significant contributors is also provided in Tahle 5-14, along with the coruesponding limiting seismic failure mode and method offragility calculation.

Among the top SLERF contributors are the steam generators, containment isolation solenoid-operated valves at MSCV EL 718 ft containment isolation diaphragm operated vales at RCBX EL 718-121 ft and major containment penetrations.

lEGonsulting (lRtz7-o Raxx Iurrrnrnic Evnur IE FnneunF{cv SLERF./vn PsncnNT OF SLERF Sreunxcy PRocRnssroN DnscRrrrrox I

Gt0 2.0-499g 8.678-08 2.34E-09 0.88%

This sequence is initiated by an earthquake between 2.0g and 5.0g. Direct core damage occurs by seismic failure of the control rods to insert. Power from offsite is failed seismically. The EDGs and the service water intake structure also fail seismically causing a station blackout. A small LOCA and failure of the RWST and all 3 AFW pumps, are also caused by seismic motion.

Core damage occurs. One of the large containment penetrations (e.g.personnel airlock) also fail. The large containment penetration failure provides a release path resulting in a large, early release. This sequence is similar to the previous sequence. It only differs in the seismic failures associated with HPI. Since there is a station blackout any, the sequence progressions are the same.

9 Gr0 2.0-4.99g 8.67E-08 2.33E-09 4.87o/o This sequence is initiated by an earthquake between 2.0g and 5.0g. Direct core damage occurs by seismic failure of the control rods to insert. Power from offsite is failed seismically. The EDGs and the service water intake structure also fail seismically causing a station blackout. A small LOCA and failure of the RWST and all 3 AFW PUMPS are also caused by seismic motion.

Core damage occurs. Here a smaller containment isolation line fails to isolate due to seismic motion (i.e.,

EQl25). The release path is still large enough to cause a largs. early release.

l0 G09 1.0-2.09 1.08E-06 2.17E-09 0.82%

This sequence is initiated by an earthquake between l.0g and 2.0g. Direct core damage occurs by seismic failure of the control rods to insert. Power from offsite is failed seismically. The EDGs and the service water intake structure also fail seismically causing a station blackout. A small LOCA and failure of the RWST are also caused by seismic motion. All 3 AFW pumps also fail seismically. Core damage occurs. One of the large containment penetrations (e.9. personnel airlock) also fails seismically. The release path is still large enough to cause a large, early release.

2734294-R-036 Reaision 0 May 11,20L7 Page 112 of 1-45 TABLE 5.14 IMPORTANCE MEASURES FOR SEISMIC COMPONENT FAILURES TO SLERF RANKED BY F'USSEL.VESBLY IMPORTAFICB ABSGonculting rlRlzzfr Rq.xx Gnoup Tor Evnxr Coprpounr,m DrscRrprrox FVI HCLPF (e)

Anr pr Su F,nrlunn Moun Frucu-Irv Mrruoo I

EQOl ZLz STEAM GENERATORS 1.85E-0I 1.08 2.71 0.24 0.32 EXCEEDING ALLOWABLE STRESS IN SUPPORT FRAMING BRACE CDFM 2

EQl25 ZSO CT. ISOL.. OUTBRD SOV MSCV 7I8 1.70E-01 1.09 2.74 0.24 0.32 FUNCTIONAL FAILURE OF SOLENOID CDFM 3

EQl2l ZDI CT. ISOL..INBRD DIAPHRAGM RCBX 718-721 6.09E-02 0.84 2.tt 0.24 0.32 SHAFT BINDING CDFM 4

EQ58A ZCP ZCP: PERSONNEL AIRLOCK 5.45E-02 r.34 3.37 0.24 0.32 MEMBRANE STRESS CDFM 5

EQ58B ZCP ZCP:EMERGENCY AIRLOCK EQ.

HATCH 5.45E-02 1.34 3.37 0.24 0.32 MEIIdBRANE STRESS CDFM 6

EQ58C ZCP ZCP:

CONTAINMENT EQUIPMENT HATCH 5.45E-02 1.34 J.J /

0.24 0.32 MEMBRANE STRESS CDFM 7

EQ58D ZCP ZCP:480V ELECTRICAL PENETRATIONS 5.45E-02 1.34 3.37 0.24 0.32 MEMBRA}iE STRESS CDFM 8

EQ53 ZOG IR-2A;B;C;D & TRF-2-5J;K 3.33E-02 0.1 0.25 0.24 0.32 CONSERVATIVE LOW HCLPF OF O.IOG ASSIGNED BASED ON SEISMIC CATEGORY ASSIGNED I

EQl22 ZDO CT. ISOL.. OUTBRD DIAPHRAGM MSCV 718 3.3rE-02 0.97 2.44 0.24 0.32 SHAFT BINDING CDFM l0 EQr l1 ZPT PROPANE TANK FARM 2.9rE-02 0.45 1.03 0.24 0.26 PIER FLEXURE CDFM ll EQ123 ZPO CT.ISOL. - OUTBRD PRSSR RLF MSCV 718-725 2.268-02 1.05 2.65 4.24 0.32 RELIEF VALVE FTINCTIONAL FAILURE CDFM t2 EQ67 ZLz MSCV BUILDING 1.698-02 1.68 3.64

0. r6 0.31 STRUCTURAL SOV 13 EQ55 ZVS VSLOCA 1.54E-02 0.125 0.32 0.24 0.32 sEE NOrE (l) sEE NOTE (1) t4 EQr r4A ZR4A D3 RELAY FOR EDG 2-l 6.82E-03 0.29 0.73 0.24 0.32 RELAY CHATTER CDFM t5 E,Q52 ZSW STANDBY SERVICE WATER PUMPS &

AISX 5.87E-03 0.1 0.25 0.24 0.32 CONSERVATIVE LOW HCLPF OF O.lOG ASSIGNED BASED ON SEISMIC CATEGORY ASSIGNED l6 EQl l4B ZR4B D3 RELAY FOR EDG 1n 4.51E-03 0.37 0.93 0.24 0.t2 RELAY CHATTER CDFM

273429+R-036 Reaision 0 May 1.1.,201.7 Page 1-L3 of 145 TABLE 5-T4 IMPORTANCE MEASURES F'OR SEISMIC COMPONENT FAILURES TO SLERF RANKED BY FUSSEL.VESELY IMPORTANCE (coNTTNUED) lESGonsuiling

[]Rtzzo Raxx Gnoup Top Evnxr ConpolrcNr DnscRrpuon FVI HCLPF (s)

Au pr pu FnrLURr Monn Fru.ctlIrY Mnruon t7 EQ76 ZDG 2EGF-TK21A;B (EDG FUEL OIL STORAGE TANKS) 2.99E-03 0.5 l.l3 0.24 0.26 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (2) l8 EQI 18 zo3 CONTROL ROOM TNDICATION PANELS 2.298-03 0.52 t.43 0.24 0.38 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (2) l9 EQ77 ZDG DAMPERS FOR EDG SUPPORT 1.87E-03 0.5 1.27 0.24 432 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (2) 20 EQe0 ZDG 2EGF-LIS2O3A;B &

204A;B (EDG DAY TANK LEVEL SWTTCHES) 1.87E-03 0.5 1.27 0.24 0.32 Fragility assigned based on inherent seismic ruggedness Assigned Screening Threshold -

See Note (2) 2l EQ42 ZDG DIESEL GEN BLDG SUPPLY FANS (2HVD-FN270A.,8) 1.87E-03 0.46 r.27 0.24 0.38 FI.JNCTIONAL FAILURE DUE TO S}IAFT BINDING AND FAILIJRE OF ATTACHED DUCT WORK CDFM 22 EQl 19 zo4 CONTROL ROOM CEILING 1.66E-03 0.66 1.00 0.24 0.01 IMPACT WITH VERTICAL BOARDS See Note (3) 23 EQl 16 ZM3 MCC.2-E03, EO4 1.65E-03 0.45 0.91 0.24 0.18 AUXILIARY CONTACT CHATTER, MAIN CONTACT CHATTER" CONTROL RELAY CHATTER, AND CONTACTOR CHANGE-OF-STATE CDFM 24 EQl l ZSM zRSS.SSCIOI (CNMT SUMP SCREENS}

1.47E-03 0.29 0.72 0.24 0.32 BASE METAL SHEAR STRESS CDFM 25 EQ68 ZLz REACTOR CONTAINMENT BUILDING t.29E-03 2.39

\\?1 0.16 0.32 STRUCTURAL SOV 26 EQTl ZSM 2RSS-E2lA;B;C;D (RECIRC SPRAY cooLERS) r.16E-03 0.30 0.76 0.24 0.32 Structural Assigned

273429+R-036 Reaision 0 May 11,20L7 Page LL4 of145 TABLE 5.14 IMPORTANCE MEASURES FOR SEISMIC COMPONENT FAILURES TO SLERF RANKED BY FUSSEL.VESELY IMPORTAI{CE (coNTTNUED)

Notes:

(1) The fragilrty for VSLOCA is assumed to have a HCLPF equal to the BV2 Site SSE based on Section 5.4.4 of the EPRI SPRA Implementation Guide.

(2) Assigned Sueening Threshold means that the SSCs were determined to be sufficiently seismically rugged as determined from plant walkdown to conservatively assign a screening Level HCLPF which initially was 0.5g.

(3) The closure of the gap calculation is carried out as a median-centered analysis which directly provides A*. Generic betas are then adopted to calculate a HCLPF.

The most significant non-seismic SSC SLERF contributors (e.9., random failures of modeled components during the SPRA mission time) are listed in Tahle 5-15.

Reference l7 contains the FV and RAW values for each component modeled in the SPRA, for both CDF and LERF sequences. Components were determined to be significant if the component's RAW is greater than 2 or its FV is greater than 0.005 for either CDF or LERF sequences, per the definition from the PRA Standard (Reference 4). RISKMAN report "Component Importance, With Common Cause and Maximum BE RAW" was used for FV, and "Component Importance, Without Common Cause and Maximum BE RAW'was used for RAW, created using the LERFS sequence group for LERF data. Judging against the above criteria, there were no risk significant components for LERF sequences, however the top 10 components by FV for seismic LERF is presented below. Note that the top five components are related to the Emergency Diesel Generators. The importances presentedin Table 5-/5 also use the results from Sensitivity Case 38.

AEBGonsulting

[]Rtzzo Rar*tx Gnoup Tor Evnrur CorupoxBxr Drscnrruox FVI HCLPF (e)

Au pr pu FerluRu Mour FRncrlrrv Mnrsoo 27 EQ3e ZRQ REFUELING WATER STORAGE TANK (RWST) 1.03E-03 0.45 1.02 0.24 0.32 OVERTURNING OF TANK CDFM 28 EQl30 zo3 CONTROL ROOM BOARDS 1.01E-03 0.58 1.46 0.24 0.32 FUNCTIONAL FAILURE DUE TO SLIDING OF INTERNALS, CURRENT SURGES AND/OR SPTIRIOUS ACTUATION OF RELAYS CDFM

273429+R-036 Rwision 0 May 11,20L7 Page 115 of 1.45 TABLE 5-I5 NON.SEISMIC SIGNIFICANT COMPONENT LIST (SORTED BY SLERF FVI)

Conapourur COIVTPONENT DESCRIPTION SLERF FV BV-FLEX.GEN-OO2 FLEX 480v Generator 2.04E-03 BV.2EGS.EG2-2 Emergency Diesel Generator 1.76E-03 BV-2EGS.EG2-1 Diesel Generator I.10E-03 BV-480VUS-2-8-4C 480v Breaker For MCC 2-803 l.0sE-03 BV.2FWE.P22 Aux Feed Turbine Driven 5.54E-04 BV-FLEX.MU-PP.OO2 FLEX Make-Up Pump 4.88E-04 BV.BAT-2.2 Battery 4.45E-04 BV.4KVS-2DF.2FI O 4160 Volt Breaker For Diesel Gen 2-Z 3.828-04 BV.4KVS-2AE-2EI O 4160 Volt Breaker For Diesel Generator 2.45E-04 BV-480VUS-2-9-5C 480v BKR For MCC-2-E08 2.398-04 A summary of the SLERF results for each seismic hazard interval is presented in Table S-td.

The table is sorted by the hazard range of the initiators. Initiating event category contribution was determined by using RISKMAN's "Contribution of Initiating Events to One Sequence Group" report, using the Master Frequency File R6AMFF with Sequence Group LERFS, at a report cutoff of lE-14, after quantification truncation of lE-14 TABLE 5-16 INITIATING EVENT CONTRIBUTIONS TO LERF As shown inTable 5-16, seismic LERF is dominated by acceleration intervals G09 through Glg which account for 92o/o of the LERF contribution. At these accelerations, the seismic collapse of various buildings causes large openings in the containment through penetrations or failure of the containment itself.

lESGolrsulting rlRtzzo IrurrraroR H.tzmn R.q.Ncr (g)

Iurrcnvnr, FnneunNCy INrnRvar.

LERF otto ConrnrBUTroN Cuprur,,+.TIvE LERJ G01 0.06-0.1s 5.38E-04 1.58E-10 0.06%

1,58E-10 G02 0.1s-0.25 1.10E-04 s.67E-10 0.2r%

7.25F-10 G03 0.2s-0.4 3.35E-0s 6.39E-10 0.24%

1.36E-09 G04 0.4-0.5 7.998-06 1.08E-09 0.41%

2.448-09 G0s 0.5-0.6 4.03E-06 2.08E-09 0.78%

4.528-09 G06 0.6-0.7 2.21E.06 2.578-09 0.96%

7.09E-09 G07 0.7-0.8 1.33E-06 3.04E-09 t.t4%

l.0lE-08 G08 0.8-1.0 t.428-06 I.l9E-08 4.46%

2.20E-08 G09 1.0-2.0 1.08E-06 1.61E-07 60.41%

1.83E-07 Gl0 2.0-4.99 8.67E-08 8.3 5E-08 31.33%

2.668-07 Total 0.06-4.99 6.998-04 2.668-07 100.00%

273429+R-036 Reaision 0 May 11",20L7 Page 1.L6 of 745 Appendix J in Reference 17 surnmarizes the contribution to seismic LERF from the most significant post-initiator human actions. Per Reference 4, significant post-initiator operator actions are defined as those operator action basic events that have a FV Importance value greater than 0.005 or a RAW greater than2, The importance measures were calculated in RISKMAN and generated through the Basic Event Importance Report for Sequence Group Report in the Event Tree Module. Reports were generated for the Sequence Group LERFS (seismic LEFS) and the Operator Action Events were pulled out to make the table in Appendix J in Reference 17.

Appendix J in Reference 17 also uses importances from Sensitivity Case 38. Operator Actions that had a FVI of 0 and RAW of I for both CDF and LERF were excluded from the table as they are not important to the seismic CDF or LERF. Also operator actions that are guaranteed failed for seismic events are excluded.

Although no operator actions meet the risk significant criteria listed above, the top operator action to LERFS (seismic LERF) by FV Importance is the same as the most important actions to seismic CDF.

5.6 SPRAQumvrtFICATIoNUNCERTAINTvAn.q,lvsrs Parameter uncertainty relates to the uncertainty in the computation of the parameter values for initiating event frequencies, component failure probabilities, and HEP that are used in the quantification process of the PRA model. These uncertainties carr be characterized by probability distributions that relate the analysts' degree of belief in the values that these parameters could take. To make a risk-informed decision, the numerical results of the PRA, including their associated uncertainty, must be compared with the appropriate decision criteria.

The RISKMAN software has the capability to correlate selected input distributions, propagate these uncertainties in input parameter distributions via a Monte Carlo quantification, and calculate the probability distributions for the risk metrics of the SPRA. These distributions and main uncertainty parameters (Mean, 5th Percentile, 50th Percentile, and 95th Percentile) are provided below for the seismically initiated CDF and LERF.

The parametric uncertainty results present an estimation of the uncertainty introduced by the data used to quantiff the PRA model. Such data uncertainty typically shows a relatively tight distribution for internal events in a commercial nuclear plant PRA as a result of the types of distributions used (largely lognormal) and the relatively large amount of operational experience for most modeled components. For seismically initiated accident sequences this is not the case.

The uncertainties in the family of seismic hazard exceedance curves, and the SSC fragility curves can be large, and with a much large impact than the data distributions applicable to internal events.

For the propagation of parameter uncertainties to seismic CDF and LERF the Uncertainty Analysis feature of RISKMAN was used. This feature requantifies the sequences using distributions for the input variables (initiators and split fractions) utilizing a Monte Carlo simulation. This method accounts for the uncertainty from all the input data parameters.

This parameter uncertainty estimation does not, however, reflect possible effects on the results from other sources of uncertainty. Such sources may include such things as: optimism or pessimism in definitions of sequence, component, or Human-Action success criteria; limitations lESGonsulting rlRtzz.o

2734294-R-036 Ratision 0 May 1,1,20L7 Page 11-7 of 1.45 in sequence models due to simplifications (for example, not modeling available systems or equipment) made to facilitate quantification; uncertainty in defining human response within the emergency procedures (for example, if there are choices that can be made); degree of completeness in selection of initiating events; assumptions regarding phenomenology or SSCs behavior under accident conditions (for example, RCP seal LOCA modeling assumptions).

While it is difficult to quantiff the effects of such sources of uncertainty, it is important to recognize and evaluate them because there may be specific PRA applications where their effects may have a significant influence on the results.

The results of the base-case seismic model parameter uncertainty analysis are shown in Table 5-17 and on Figure 5-2 and Figure 5-3.

TABLE 5.17 PARAMETER T]NCBRTAINTY ANALYSIS RESULTS MrnN 5o/o s0%

950h Seismic CDF (/Year), 10,000 Samples 8.78E-06 7.80E-07 5.08E-06 2.98E-0s Seismic LERF (/Year), 10,000 Samples 2.668-07 8.43E-09 I.1 5E-07 9.97E-07 lESGonsulting rlRlzz.o

273429+R-036 Raision 0 May 1,1,2017 Page 11.8 of 1,t15 Seismic CDF I

a a

o

\\

/

\\

a a

a I

\\

a a

I

\\

a a

a

/

a

\\

a

\\

Seismic Core Damage Frequency Distribution 1.00E-08 1.00E-07 1.00E-06 1.00E-05 1.00E-04 CDF FIGT]RE 5.2 spRA cDF UNCERTAINTY DTSTRTBUTION (10,000 SAMPLES) lICGoneul$lr0

(}FIzzp

273429+R-036 Reoision 0 May L1,2017 Page 119 of l,aS LERFS Seismic Large Early Release Frequenry Distribution 1.00E-11 1.00E-10 1.00E-09 1.00E-08 1.00E-07 1.00E-06 1.00E-05 1.00E-04 LERF f,.IGT]RE 5.3 SPRA LERF UNCERTAINTY DTSTRTBUTION (10,000 SAMPLES) 5.6.1 ModelUncertainty Model uncertainty arises because different approaches exist to represent plant response. A source of model uncertainty is one related to an issue in which no consensus approach or model exists, and where the choice of approach or model is known to have an ef[ect on the SPRA.

These uncertainties are typically dealt with by making assumptionSi o.9., the approach to address common-cause failure, how a RCP would fail following a loss of seal cooling, the approach to identifr and quantiff HFEs. In general, model uncertainties are addressed through sensitivity studies using different models or assumptions.

The guidance provided in EPRI l0l6737,Treatment of Parameter and Model Uncertaintyfor Probabilistic Risk Assessments (Reference 74),was used to address sources of model uncertainty and related assumptions. It provides a framework for the pragmatic treahent of uncertainty chaructenzation to support risk-informed applications and decision-making. The process includes identification and characterization of sources of model uncertainty and related assumptions; the following sections summarize the sources of uncertainty found in the Level I SPRA.

fSGonsul$ry

()Rrzzo

273429+R-036 Raision 0 May L1-,20L7 Page L20 af745 5.6,2 Understood and Accepted Generic Uncertainties Three issues that are generally understood and accepted as potential generic sources of model uncertainty are:

I.

Treatment of Pre-Initiator and Post-Initiator Human Errors (i.e., screening human error probabilities, realistic HEPs for significant HFEs, realistic HEPs for all HFEs)

2.

Treatment of Potentially Dependent Post-Initiator Human Erors (i.e., no HFE dependence, some dependent HFEs, all HFEs assessed for dependence)

3.

Intra-System Common Cause Events (i.e., generic corrmon cause failure [CCF],

plant-specific CCF)

Based on lessons learned, a standard set of sensitivity cases was recommended to envelope these understood and accepted generic sources of uncertainty at a high level (Reference 16).

The four sensitivity cases are:

1.

All HEPs set to their 5th percentile value

2.

All HEPs set to their 95th percentile value

3.

All CCF probabilities set to their 5th percentile value

4.

All CCF probabilities set to their 95th percentile value The quantitative results of these sensitivities are presented in ^Section 5.7.

5.6.3 Generic Sources of Model Uncertainty A generic list of additional sources of model uncertainty for internal events PRA was identified hased on Reference 16. This list includes those having the highest potential to change risk metrics and decisions, and includes: phenomena or nature of the event or failure mode not completely understood; models based on significant interpretations; and issues with general agreement. Table I-1 in Appendix I of Reference l7 includes the list of generic pressurized water reactor (PWR) sources of model uncertainty and a characterization assessment for the BVPS-2 Level I SPRA.

5.6.4 Plant-Specific Sources of Model Uncertainty An examination of plant-specific features and modeling approaches was also performed to identify any uncertainties not identified on the generic list. This assessment focused on identiffing plant-specific features, modeling approaches and assumptions that were not included inthe generic uncertainties. Table I-ZinAppendix I of ReferencelT includes the list of plant-specific sources of model uncertainty and a SPRA characterization assessment for the BVPS-2 Level 1 PRA; exceptions include generic sources of model uncertainty, alignments, and boundary systems that are not modeled because they have no impact on the PRA function system modeled.

Table I-3 of Reference 17 identifies sources of uncertainties from the assumptions listed in Section 2 of Reference 38. These assumptions are specifically related to the plant-specific SPRA for BVPS-2. The table descrihes the impact of the assumption on the SPRA modeling and then lESGonsulting rlRlzzo

2734294-R-036 Reaision 0 Moy L1,2017 Page 121 of 1"45 characterizes whether the uncertainty in the current assessment could potentially impact plant risk-based applications.

5.6.5 Completeness Uncertainty Completeness uncertainty relates to risk contributors that are not in the SPRA model, nor were they considered inthe development of the model. These include knowntypes such as the scope of the PRA, which does not include some classes of initiating events, hazards, and operating modes; and the level of analysis, which may have omitted phenomena, failure mechanisms, or other factors because their relative contribution is believed to be negligible. They also include ones that are not known such as the effects on risk from aging or organizational changes; and omitted phenomena and failure mechanisms that are unknown. Both can have a significant impact on risk.

No completeness uncertainties were identified for the BVPS-2 Level 1 SPRA, based on the ASME/ANS PRA Standard (Reference 4).

5.7 SPRAQu,uvrtFICATIoNSENSITrvITvAu,r.r,vsrs As presented in Section 5.7.1, four standard sensitivity sfudies were selected for analysis:

r A11 HEP probabilities set to their 5th percentile value.

. All HEP probabilities set to their 95th percentile value.

. All CCF probabilities set to their 5th percentile value.

r All CCF probabilities set to their 95th percentile value, The HEPs and CCF probabilities were changed to the 5th or 95tr percentiles by importing distributions in the datamodule using the import distribution parameters function. The import file was created by exporting the parameters using the export distribution parameters function and the mean values were adjusted to the 5tr or 95th percentile. The percentile values were taken from the RISKMAN titles listing report in the data module. The distributions affected were all Human-Action and beta, gamma, ffid delta factors used in the Multiple Greek Letter corrmon-cause method in the model. Both CDF and LERF were requantified at the 5th or 95ft percentiles for HEPs and CCF probabilities in separate cases.

The resulting 5th and 95th percentile values represent the CCF sensitivity cases listed above. The results of these sensitivity cases are discussed here and compared to the RG 1.174 CDF limit of lx1O4/year for CDF and lxlO-s/year for LERF to obtain insights into the sensitivity of the base PRA model results to these generic high level sources of modeling uncertainty. This approach is followed rather than trying to identify all potential sources of model uncertainty associated with these issues since they are generally understood and accepted as areas of uncertainty that can be significant contributors to CDF. The results of the studies are shown inTable 5-18.

The results indicate that CDF is more sensitive to these uncertainties than LERF, and each of the models are more sensitive to operator action uncertainty than they are to common-cause uncertainty. However, overall the model does not produce drastic changes for these sensitivity studies.

lEEGonsulting

[]Rtzz.o

2734294-R-036 Reaision 0 May LL,201.7 Page 1.22 of 145 TABLE 5.T8 CCF AND HEP SENSITIVITY CASBS Casn iYo 5V" L' FROM BasrLrFrn 95%

95Y" L' FROM Bnsnr,nvn tmP-CDF (/year) 8.61E-06

-1.920/0 9.12E-06 3.90%

Fffi,P-LERF (/year) 2.668-07

-0.r7%

2.578-07 0.36%

CCF-CDF (/year) 8.77E-06

-0.12%

8.80E-06 0.20%

CCF-LERF (/year) 2.668-07

-0.04%

2.67F,-07 0.07%

5.7.L Seismic-Related Sensitivity Cases This section presents the sensitivity results for selected cases defined specifically for the modeling of seismic events.

The uncertainties in the assessment of the seismic hazard curve, and of SSC fragilities are captured in the parameters that define these intermediate results; i.e., by the family of seismic hazard exceedance curves, andthe parameters for each of the SSC fragilitiesl Am, B', and po.

The results of the uncertainty analysis presented inSection 5.6 illustrate the impact of uncertainties in the hazard exceedance curves and fragility curves on CDF and LERF. Therefore no funher sensitivities were performed to assess these parameter uncertainties.

Sensitivity studies described below are used to investigate other sources of uncertainty which impact the modeling of seismic impacts and the quantification methods used.

Each of the assumptions listed previously in Section 2 of the Quantification Notebook (Reference 17) and in other notebooks was examined to determine if a sensitivity case was feasible and instructive. The following areas were investigated:

l.

Modeling of Seismic Impacts

2.

Correlation of Fragilities

3.

Relay Chatter

4.

Human Reliability Analysis

5.

Quantification Methods

6.

Fragility Refinement Impacts The results for each of the seismic-related sensitivity cases are provided in Table 5-19. All sensitivities were performed using the Level 2 model, which can calculate Level I results, but is slightly lower than the actual Level 1 results because sequences that are close to the lE-14 cutoff for core damage will drop below the lE-14 cutoff after progressing through the CET tree for LERF. This is deemed acceptable for these sensitivities because the insights will be the same.

lEBGonrulting

[]Rtzzo

2734294-R-036 Reaision 0 May 11,2017 Page 1.23 of 145 TABLE 5.19 SEISMIC-RELATED SENSITIVITY RESULTS Gnoup Casn Snnsrrrvrry Sruny (Srn Norrs BELow As Wnu)

CDF Vo Cg,tNcn Ng CDF LERF o/"

Cuaucrc rrq LERF 0

0 Base Case 8.71E-06 2.66E-07 I

I LOOP Always TRUE 8.93E-06 254%

2.69E-07 0.89%

I

)

No Turbine Building Impacts 8.64E-06

-0.87%

2.648-07

-0.820/o I

aJ Credit ERFS Black Diesel Generator N/A for Unit 2 I

4 DG 48 Hours (change @T24 to

@T48 local variables for all System Tops) 8.87E-06 1.86%

2.6',1E-07 0.42o/o I

5 Extend LERF Evacuation Time to 48 hrs (Rebin of LATE to LERF due to extended time) 8.7rE-06 0.00%

2.66E-07 0.00%

I 6

No Very Small LOCA 7,64E-06

-12.33Yo 2.70E.07 1.42%

I 7

Eliminate impacts of Block Wall failures N/A for Unit 2 I

r3 Changing Fragil ity Values/Curves 8.71E-06 0.00%

2.628-07

-1.69%

I l5h Eliminate Seismic Failure of SEISMIC Top Event ZRW 8.46E-06

-2.92o/o 2.6',1E-07 0.1 5%

I lsi Remove Failure of Propane Tank 8.70E-06

-0.12%

2.598-07

-232%

I 15i.

Remove FLEX 1.02E-05 16.70%

2.85E-07 7.10%

I 22 Guarantee fail Cross-tie; No Correlation of Unit I and 2 DGs N/A; cross-tie not credited for any seismic 2

l4a Correlate the Seismic Failure of Buildings Directly Causing Core Damage and Large, Early Release 8.71E-06 0.00%

2.668-47

-0.13%

2 r4b Correlate the Large Containment Failures 8.71E-06 0.00%

2.21E'07

-l7.llo/o 2

l4c Un-Correlate Service Water Trains 8.31E-06

-4.s8%

2.678-07 0.lzYo J

9a Remove All Relay Chatter Fragilities 7.89E-06

-9.44o/o 2.668-07

-0.31%

5 9b Remove EDG Relay Chatter Impacts Only 8.43E-06

-5.59o/o 2.658-07

-l.34Vo 4

17.

HEP 5th %

8.618-06

-t.92%

2.668-07

-0,t1yo 4

18.

HEP 95th %

9.12E-06 3.90%

2.67E-07 0.36%

4 t9 SEIS3 Timing Sensitivity 1 (Sens I Tdelay *30 min, Texe xl C& Texe x4 outside MCR) 8.71E-06 0.02%

2.66F-07 0.00%

4 20 SEIS3 Timing Sensitivity 2 (Sens 2 Tdelay +30 rnin, Texe x2 CR, Texe x4 outside MCR)

Same as Case 19 4

21 SEIS3 Timing Sensitivity 3 (Sens 3 Tdelay +15 min, Texe xl CR, Texe x4 outside MCR (max 30 minutes))

8.71E-06

-0.01%

2.668-07 0.00%

4 23 0.1 Minimum SEIS3 IIEP 8.71E-06

-0.02%

2.66E-07 0.00%

4 24 Remove CR Panels/Indications and Ceiling Impacts and Adjust SEIS2 and SEIS3 Acceleration Interval Assignment l.0lE-05 t5.78%

2.88E-07 8.24%

4 25 Remove Toxic Gas Impact 8.68E-06

-0.47o/o 2.668-07

-0.l5%

fESGonsuEing TlHtz71o

2734294-R-036 Reaision 0 May 11,20L7 PaXe 1.24 of 145 TABLE 5-19 SEISMIC-RELATED SENSITIVITY RESULTS (coNTrNUEr))

Gnour Casn Srnsrrrvrry Sruuy (Snu Norrs Bulow As WnI,r)

CDF o/o Cnaxcn rx CDF LERF o/o Cnar.lcn N\\T LERF 4

26 Remove Impacts from CR Panels/Indications, CR Ceiling, and Toxic Gas and Adjust SEISZ and SEIS3 Acceleration lnterval AssiErment 1.01E-0s 15.57o/o 2.88E-07

8. r9%

4 39 Change HEPs Affected by CR Ceiling Tiles Failure from 1.0 to 0.99 8.70E-06

-0.lzYo 2.668-47

-0.03%

5 ll ADDED INTERVALs - expand to model.O5g delta in range of change; 0.259 to 0.5g This case was not performed in Revision l. Note Revision 0 results for this case document that for the interval between.l5g and.25g, the Case 11 results are about TVolower, and for the.25g to.4g range and the.4g to 59 range the results are 7o/o higher. These differences cancel out and the overall results are essentially the same indicating that the base-case discretization involving 10 acceleration intervals is sufficient.

5 27*

CCF 5th %

8.77E-06

-0.12%

2.668-07

-0.04%

5 29.

ccF 95th %

8.80E-06 0.20%

2.67E.07 0.47%

5 30*

Truncation Sensitivity (TRUNC: 1E-09) 5.00E-06

-43.r0%

l.4lE-07 A',t.tgYo 5

31*

Truncation Sensitivity (TRUNC: 1E-I0) 6.s9E-06

-24.92o/o 2.068-07

-22.79%

5 32*

Truncation Sensitivity (TRtrNC : 1E-11) 7.71E-06

-t2.14%

2.44E-07

-8.29%

5 JJ Truncation Sensitivity (TRUNC: 1E-12) 8.36E-06

-4.74%

2.58E-07

-3.20o/o 5

34*

Truncation S ens itivity (TRUNC: lE-13) 8.66E-06

-1.34%

2.63E-07

-l.l3o/o 5

35*

Truncation Sensitivity (TRUNC: lE-14) 8.78E-06 0.00%

2.668-07 0.00%

5 JI Truncation Sensitivity (TRUNC: lE-15) 8.82E-06 0.460/o 2.68E.07 0.61%

5 36.

Zero Maintenance 8.72E-06

-0.69%

2.668-07

-0.lgYo 6

EO5s 2*AMED - Very Small LOCA 8.05E-06

-7.63%

2.68E-47 0.67%

6 EQr l 2*AMED - 2RSS-SSC101 (CNMT Sump Screens) 8.26E-06

-5.17%

2.668-07

-0.020/o 6

EQ71 2*AMED - 2RSS-E2 IA;B;C;D (Recirc Spray) 8.36E-06

-4.04%

2.66E-07 0.01%

6 EQl02 Z+AMED - Pump and MOV Relays AR440AR in 8.40E-06

-3.52%

2.66E.01 0.t4%

6 EQ56 2*AMED - Small LOCA 8.43E-06

-3.19%

2.678-07 0.41%

6 EQl14B 2*AMED - D3 Relay for EDG 2-2 8.44E-06

-3.12%

2.66F-01

-0.03%

6 EO39 2*AMED - 2QSS-TK2I (RWST) 8.46E-06

-2s0%

2.67E.07 0.24o/o 6

EQO I 2*AMED - Steam Generators 8.71E-06 0.02%

2.18E-07

-l 8.10%

lEECqrsulting rlRtzz-o

273429+R-036 Reaision 0 May 1L,201-7 Page 1-25 of 1,45 GRour CasB SnNsrrrvrrv Srunv (Snr Norrs Bsrow As Wnu)

CDF o/o Cu,rncn rru CDF LERF Vo CnaNcn rn LERF 6

EQl2s 2*AMED - Ct. Iso[. - Outbrd SOV MSCV 7I 8.71E-06 0.03%

2.22E-07

-16.65%

6 EQr2r 2*AMED - Ct. Isol. - Inbrd Diaphraem RC 8.72E-06 0.07%

2.51E.07

-s.62%

6 EQ58A 2*AMED - ZCP: Personnel Airlock 8.7IE-06 0.02o/o 2.52E-07

-5.30%

6 EQssB 2*AMED - ZCP: Emergency Airlock 8.71E-06 0.02%

2.528-07

-5.30%

6 EQssC 2*AMED - ZCP: Containment Eq.

hatch 8.71E-06 0.02o/o 2.528-07

-s.30%

6 EQ58D 2*AMED - ZCP: 480V Elec Pen.

8.71E-06 0.OZYy 2.528-07

-5.30o/o 6

EQl22 2*AMED - Ct. Isol - Outboard Diaphram MSCV 718 8.71E-06 0.05%

2.58E-07

-3.08%

TABLE 5.19 SEISMIC-RELATED SENSITIYITY RESULTS (coNTTNUBD)

Note:

• These cases were quantified with the Level I and Level 2 models separately and the CDF results are compared with the 8.78E-6 seismic CDF instead of the CDF bin in the Level2 model which truncates some CDF sequences and has a value of 8.71e-06 5,7,1.1 Group 6: Fragility Refinement Impacts The preceding seismic sensitivity cases reflect those sensitivities defined to determine the impacts of selected modeling assumptions on the CDF and LERF calculations. The cases described below are defined to exafirine the sensitivity of CDF and LERF to assumed improvements in the seismic capacities of the most important equipment fragility groups. One can use the FVI rankings directly for this putpose, but the FVI measure is a bounding measure assuming the SSCs in the equipment fragility groups are made perfect. For these added cases a seismic capacity improvement equal to twice the base-case evaluated capacities is assumed, one equipment group at a time. Further fragility analysis is unlikely to achieve such an asssssed improvement because much effort has already been dedicated to making the SSC seismic capacity assessments as realistic as possible. These cases are incorporated into the model by replacing the base median acceleration capaclty, Am, by twice the Am. The Beta-r and Beta-u values are held the same so that the HCLPF accelerations are also twice the base-case values.

The FVI measures computed from Sensitivity Case 38 were used to identifu fragility groups for these sensitivities as results from this case give more accurate importances, as identified earlier in this submittal. The fragility component groups with FVI less than 0.03 were deleted from further consideration. They were deleted because even if they could be made perfect, the maximum reduction in CDF or LERF wouldbe 0.03. Also deleted from furtherconsideration was the fragility groups for failures of the offsite grid (EQ53). This fragility group was assessed using generic data that is not specific to BV Unit 2 and is an industry accepted value and should not change in the near future.

IE$Gonsulting rlRtzzo

2734294-R-036 Rtuision 0 Moy L1-,20L7 Pase 126 of145 All sensitivities were performed using the Level 2 model, which can calculate Level 1 results, but is slightly lower than the actual Level 1 results because sequences that are close to the 1E-14 cutoff for core damage will drop below the 1E-14 cutoffafter progressing through the CET tree for LERF. This is deemed acceptable for these sensitivities because the insights will be the same.

Table 5-20 below identifies the fragility groups evaluated for the twice Am sensitivities. The CDF and LERF changes are presented for all cases. The top half of Table 5-20 is for CDF contributors and the bottom of the table for LERF contributors. The CDF and LERF changes are nevertheless presented for all cases. The FVI measures from Sensitivity Case 38 are presented in the table as well as the revised CDF and LERF and changes in CDF and LERF are presented.

All CDF frequency changes were less than 1E-6 per year. All LERF frequency changes were less than lE-7 per year. The percent changes in CDF or LERF were, as expected, found to be less than the FVI of the fragility group to that risk measure and in some cases the change in CDF and LERF were negligibly.

The largest potential decrease in CDF would come from refining the VSLOCA fragility. This fragility is based off of industry accepted methodology and although conservative is an accepted value. The same is true for the SLOCA fragility. The remaining fragility groups identified for CDF have all been refined to achieve a realistic fragility. For these groups plant modification would be the only way to achieve the risk reduction presented, in Tahle 5-20. The low seismic CDF of 8.71E-06 justifies the acceptance of the conservativisms in the VSLOCA and SLOCA fragilities as well as eliminates the need for any modifications. Additionally the delta CDFs in the mid to low E-7 range is further justification for accepting the conservativisms in the VSLOCA and SLOCA fragilities and furtherjustifies the basis forno plant modifications.

Similar to the identified CDF components the identified LERF components have also been refined to remove conservatisms. It is judged to achieve the risk reductions identified in the table below a plant modification would be needed for the identified components. The low seismic LERF of 2.66E-07 eliminates the need for any modifications. Additionally the delta LERFs in the mid to low E-8 range is further justification for no plant modifications.

It is concluded that all other fragility groups, not evaluated here, if evaluated with twice the current capacities would lead to a reduction in CDF or IERF of less than 3%. These SSCs are not important enough to justiff refining the fragility because possible conservatisms in the fragility calculations are not driving the model results or masking insights.

lE$Gonsutting (iRtzzo

TABLE 5-20 SENSITIVITY OF MOST IMPORTANT EQUIPMENT GROUPS (2*THEIR MEDIAN CAPACITY)

CasB ID DrscRmuou Basn-Ca,sE, HCLPF cDF rvr @

1E-14 cDF @

lE-14 CDF Dmrnnrxcn FROM PrRcrxr CHaucs tr{

CDF LERF FVr @

1E-14 LERF@

lE-14 LERF DrrrnRrncn FRoM*

PnncnNr Cnaxcr rx LERF' Sensitivities at lE-l4lyex; Set Am :) 2*Am; all SSCs with FVI >3E-Z to CDF (exclude LOOP) 8.71E-06 2.66E'07 EQs5 Very Small LOCA 0.125g 1.38E-01 8.05E-06

-6.63E-07

-'1.61%

r.54E-02 2.688-07 1.77E-09 0.67Vo EQI l 2RSS-SSCl0l (CNMT Sump Screens 0.29s 5.32F.02 8.26E-06

-4.51E-07

-5.17%

1.47E-03 2.658-07

-5.00E-1t

-0.02o/o 2RSS-Ez IA;B;C;D (Recirc Spray EQ71 0.30g 4.14E-02 8.36E-06

-3.528-07 4.04%

l.t6E-03 2.668-07 2.00E-l l 0.01%

EQ102 Pump and MOV Relays AR440AR in 0.42g 3.57F-02 8.40E-06

-3.06E-07

-352%

3.378-04 2.668-07 3.80E-10 0.14%

EQ56 Small LOCA 0.32g 3.47E-42 8.43E-06

-2.78E-07

-3.19o/o

-2.17E-03 2.678-07 r.l0E-09 0.41%

EQl14B D3 Relay for EDG 2-2 0.37e 3.268-02 8.44E-06

-2.728-07

-3.12o/o 4.5IE-03 2.66E.-07

-7.00E-11

-0.03%

EQ3e 2QSS-rKzl (RWSr) 0.45g 3.05E-02 8.46E-06

-2.s3E.07

-2.900/o 1.03E-03 2.678-07 6.40E-10 0.24%

Sensitivities at lE-l4/year; Set Am =) 2+Am, for all SSCs with FVI >3E-2 to LERF (exclude LOOP and VSLOCA)

EQOr Steam Generators 1.08g 4.42E-0s 8.7tE-06 1.50E-09 0.02Yo 1.85E-01 2.18E-07

-4.82E-08

-r 8. l0%

EQl25 Ct. Isol. - Outbrd SOV MSCV 718 1.09g

-0.00E+00 8.71E-06 2.70E-09 0.03%

1.70E-01 2.22E-07

-4.43E-08

-t6.6svo EQl2l Ct. Isol. - Inbrd Diaphragm RC 0.84g

-0.00E+00 8.728-06 6.00E,09 0.07Vo 6.09E-02 2.51E-07

-1.50E-08

-5.620/o EQssA ZCP: PersonnelAirlock 1.34g

-0.00E+00 8.71E-06 2.00E-09 0.02%

5.45E-02 2.528-47

-l.4lE-08

-5.30o/o EQ58B ZCP: Emergency Airlock Eq. Hat 1.34g

• 0.00E+00 8.71E-06 2.00E-09 0.02%

5.45E-02 2.528-07

-l.4lE-08

-s.30%

EQ58C ZCP: Containment Equipment Hat 1.349

• 0.00E+00 8.71E-06 2.00E-09 0.02%

5.45E-02 2.528-07

-l.4lE-08

-5.30%

EQs8D ZCP; 480V Electrical Penetrati 1.349

-0.00E+00 8.71E-06 2.00E-09 0.02%

s.45E-02 2.528-07

-l.4lE-08

-530%

EQl22 Ct. Isol-Outboard Diaphragm MSCV 718 0.979

-0.00E+00 8.71E-06 4.40E-09 0.05%

3.31E-02 2.58E-07

-8.20E-09

-3.08%

F lr:r(< id r{*

-F.l s-

-tDN) x'

\\o F$\\

'N tJt TH NO Ng od E-*

='

E

273429+R-036 Reaision 0 May 11,2017 Pase L28 of145 5.8 SPRA LocIc Monrl AND QuaxuFrcATroN TECHNTcAL Anreu^tcy The BVPS-2 SPRA risk quantification and results interpretation methodology were subjected to an independent peer review against the pertinent requirements in the ASME/ANS PRA Standard (Reference 4).

The peer review assessment, and subsequent disposition of peer review findings, is described in Appendix A, and establishes that the BVPS-2 SPRA seismic plant response analysis is suitable for this SPRA application.

lESGonsulting tlRtzz.o

2734294-R-036 Reaision 0 May 1.1.,201.7 Page 129 of 1.45

6.0 CONCLUSION

S A seismic PRA has been performed for BVPS-2 in accordance with the guidance in the SPID.

The BVPS-2 SPRA shows that the seismic CDF is 8.78x1046 and the seismis LERF is 2.66x10-07.

Further, no seismic hazard vulnerabilities were identified.

The updated BVPS-2 PRA model, which includes the seismic PRA reflects the as-built, as-operated plant as of the freeze date of October 25,2016 and includes the FLEX mitigation strategies equipment and procedure changes. The PRA model provides insights and identifies the most important equipment to responding to a seismic event, but no seismic hazard vulnerabilities were identified. The seismic CDF and LERF are sufficiently low such that possible improvements or modifications to the plant are not considered necessary. In addition, the Group 6 sensitivities inSection 5.7.1.6 ofthis submittal (i.e., Table 5-20) showthat postulated improvements that would increase the seismic capacity of the important seismic failures would not provide a significant reduction in risk.

IESGonsulting tlRtzzo

2734294-R-036 Reaision 0 May 11,2017 Page L30 ofL45 7,0 REFERENCES The dates and revisions of the reference documents in this section correspond to the PRA freeze date of June 2012 unless there was reason to use a more recent version of the document.

1.

NRC (E Leeds and M Johnson) Letter to All Power Reactor Licensees et al., "Request for Information Pursuant to Title l0 of the Code of Federal Regulations 50.54(0 Regarding Recommendations 2.1,2.3 and 9.3 of the Near-Term Task Force Review of Insights from the Fukushima Dai-Ichi Accident," March 12,2012.

2.

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

February 2013.

3.

NTTF 2.1 Seismic Hazard and Screening Report Beaver Valley Power Station Unit 2, Beaver County, Pennsylvania, March 20,2014.

4.

ASME/ANS RA-S-2008, Standardfor Level l/Large Early Release Frequency Probabiltstic Risk Assessment for Nuclear Power Plant Applications, including Addenda 8,2013, American Society of Mechanical Engineers, New York, September 30,2013.

5.

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

6.

PWROG-15008-P, Peer Review of the Beaver Valley Power Station Seismic Probabilistic Risk Assessment, Revision 0, March 2015.

7.

EPRI NP 6041-SL, A Methodolog,t for r4ssessment of Nuclear Power Plant Seismic Margin, Rev. 1,, Electric Power Research Institute, Palo Alto, CA, August 1991.

8.

Expedited Seismic Evaluation Process (ESEP) Report Beaver Valley Power Station Unit 2, Report No. 2734294-R-020, Rev. 0, November 3, 2014.

9.

NUREG-1407, Procedural and Submittal Guidance for the Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabiltties, U.S. Nuclear Regulatory Commission, June 1991.

10. Generic Letter No. 88-20 Supplement 4, Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilittes - llCFR 50.54(f), Nuclear Regulatory Commission, June 1991.

I 1.

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

12. NUREG/CR-0098, Development of Criteriafor Seismic Review of Selected Nuclear Power Plants, Nuclear Regulatory Commission, May 1978.

13. NRC (E Leeds) Letter to All Power Reactor Licensees et al., "Screening and Prioritization Results Regarding Information Pursuant to Title 10 of the Code of Federal Regulations 50.54(F) Regarding Seismic Hazard Re-Evaluations for lSSGolrsulting riRlz7.o

14.

15.

16.

17, 18.

19.

21.

,/,J.

25.

27.

28.

20 11

./J./J.

24 26 2734294-R-036 Reuision 0 May LL,20L7 Page L3L of145 Recommendation 2.1 of the Near-Term Task Force Review of Insights From the Fukushima Dai-Ichi Accident," May 9,2014.

Beaver Valley Power Station Unit 2 Near-Term Task Force Recommendation 2.3 Seismic Walkdown Report, Rev. 0, October 3 1,2012.

EPRI 3002000709, Seismic PRA Implementation Guide, Electric Power Research Institute, Palo Alto, CA, December 2013.

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.

FirstEnergy Nuclear Operating Company, Beaver Valley Unit 2 PRA Notebook, PRA-BV2-AL-R06a-SQU, Seismic Probabilistic Risk Assessment Quantification, Uncertainty, and Sensitivity.

SQUG 2001, "Generic Implementation Procedure (GIP) for Seismic Verification of Nuclear Plant Equipment," SQUG/GIP, Seismic Quality Utility Group, Revision 3A',

2001.

Configuration Baseline Form of RIZZO-HAZARD, V&V Revision 0, Paul C. Rizzo Associates, Inc., Pittsburgh, Pennsylvania, February 201 4.

Engineering Seismic Risk Analysis, Bulletin of the Seismological Society of America, Vol. 58, No. 5, pp. 1583-1606, 1968.

Central and Eastern United States Seismic Source Characterization for Nuclear Facilities, Vols. 1-6, NUREG-zLls, USNRC, Washington, D.C., February 2012.

EPRI (2004, 2006) Ground-Motion Model (GMM) Review Project, Report 3002000717,"

Electric Power Research Institute, Palo Alto, California, June 2013.

Probabilistic Seismic Hazard Analysis and Foundation Input Response Spectra, Beaver Valley Nuclear Power Station Seismic Probabilistic Risk Assessment Project, Report 2734294-R-003, Revision 4, ABS Consulting and P-IZZO Associates, 2016.

NRC, 2007, "A Perfonnance-Based Approach to Define the Site-Specific Earthquake Ground Motion," Regulatory Guide 1.208, USNRC, Washington, D.C., March 2007.

NRC, 2010, "Interim Staff Guidance on Ensuring Hazard-Consistent Seismic Input for Site Response and Soil Structure Interaction Analyses," DC/COL-ISG-017, U.S. Nuclear Regulatory Commission, Washington, D.C., March 2010.

Beaver Valley Power Station Unit 2 Probabilistic Risk Assessment (QU) PRA Quantification, Issue 5A," April 26,2012.

Beaver Valley Power Station Unit 2 PRA Notebook, PRA-BVI-AL-ROSa, (IF) Intemal Flooding Analysis, August 28, 2012.

Duquesne Light Company, "Beaver Valley Unit 2 Probabilistic Risk Assessment, Individual Plants Examination of External Events," Submiued in response to U.S. Nuclear Regulatory Commission Generic Letter 88-20 Supplement 4.

lESGonsulting

$Rtzzo

2734294-R-035 Reaision 0 May L1.,201.7 Page 132 of145 29

'*Beaver Valley Power Station Unit 2, Updated Final Safety Analysis Report,"

Revision 19.

Pacific Gas & Electric Company, "Final Report of the Diablo Canyon Long Term Seismic Program," Diablo Canyon Power Plant Docket Nos. 50-275 and 50-323, July 1988.

Kassawara, R.P., et al., "Surry Seismic Probabilistic Risk Assessment Pilot Plant Review," 1020756, EPRI, Palo Alto, CA 2010.

FirstEnergy Nuclear Operating Company, Beaver Valley Unit 2 PRA Notebook, PRA-BV2-AL-R06a-SEL. Development of the Beaver Valley Unit 2 Seismic Equipment List.

E-Mail from Sum Leung dated 19 July 2012, "Appendix G BV-l Fire PRA Component Selection and Screening," FENOC.

Wright, M.D., "Beaver Valley Power Station Unit I NFPA 805 Fire PRA, Task I Ib, Main Control Room Detailed Fire Modeling, Revision 8", July 12,2011, Scientech Calculation 17756-05.

EPRI 1025286, "Seismic Walkdown Guidance for Resolution of Fukushima Near-Term Task Force Recommendation 2.3: Seismic," Final, June 20L2.

FirstEnergy Nuclear Operating Company, Beaver Valley Unit 2 PRA Notebook, PRA-BV2-AL-R06a-SHR, Seismic PRA Human Reliability Analysis.

FirstEnergy Nuclear Operating Company, Beaver Valley Unit 2 PRA }rlotebook, PRA-BV2-AL-R06a-SRE, Relay Chatter Analysi s.

FirstEnergy Nuclear Operating Company, Beaver Valley Unit 2 PRA Notebook, PRA-BV2-AL-R06a-SMO, Seismic Probabilistic Risk Assessment lnputs and Model.

High-Frequency Program - Application Guidance for Functional Confirmation and Fragility Evaluation Report 3002004396, Electric Power Research Institute, Palo Alto, California, USA, July 2015.

Seismic Walkdown of Beaver Valley Nuclear Power Station, Unit 2 Seismic Probabilistic Risk Assessment Project, Report2734294-R-01l, RevisionZ, ABS Consulting and P.IZZO Associates, 20 1 6.

Fragility Analysis Report Beaver Valley Power Station, Unit 2 Seismic Probabilistic Risk Assessment Project, Report2734294-R-013, Revision 2, ABS Consulting and P.IZZO Associates, 2016.

Computer Program SASSI - User's Manual, prepared by the SASSI Development Team:

John Lysmer, et al., Geotechnical Engineering Division, Civil Engineering Department, University of California, Berkeley, California; and Bechtel Power Corporation, San Francisco, California, 1998.

Building Seismic Analysis Beaver Valley Power Station, Unit 2 Seismic Probabilistic Risk Assessment Project, Report2734294-R-012, Revision 2, ABS Consulting and P-IZZO Associates, 20 1 6.

lESGonsulting (IRtzzo 30.

31.

i/..

36.

JJ.

34.

35.

37.

38.

39.

4r.

42.

40 43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

273429+R-036 Reuision 0 May LL,201.7 Paxe 1"33 of 145 Seismic Fragility Application Guide Update, EPRI 1019200, Electric Power Research Institute, Palo Alto, CA, USA, 2009.

BVPS Soil-Structure Interaction Sensitivity Analyses Using Lower Bound and Upper Bound Soil Profiles, CalculationNo. 12-4735-F-140, Revision O,F-IZZO Associates, Pittsburgh, Pennsylvania, 20 I 6.

Seismic Analysis of Safety-Related Nuclear Structures, ASCE 4-98, American Society of Civil Engineers, 1998.

Assessment of Existing Stick Models for the Auxiliary Building (Area 7), Containment Internal Structures, and the Shield Building, Davis-Besse NPP, Report No. R4 12-4737 20121026, Paul C. Rizzo Associates, Inc., Pittsburgh, Pennsylvania, October 2012.

Code Requirements for Nuclear Safety Related Concrete Structures and Commentary, ACI 349-06 American Concrete Institute, Farmington Hills, Michigan, 2006.

Seismic Design Criteria for Structures, Systems and Components in Nuclear Facilities, ASCE/SEI Standard 43-05, American Society of Civil Engineers, 2005.

Seismic Fragility Application Guide, EPRI-I002988, Electric Power Research Institute, Palo Alto, California, USA, December 2002.

Overview of Probabilistic Seismic Response Evaluation, Early SPRA Workshop, April 7-9, 201 5.

Elkhoraibi, T. et al. 2013, "Probabilistic and Deterministic Soil Structure Interaction furalysis Including Ground Motion Incoherency Effects," Nucl. Eng. Des., 2013.

Not Used.

Effect of Torsional Moments on Walls for BVPS, Calculation No. 12-4735-F-148, Revision 0, P.IZZO Associates, Pittsburgh, Pennsylvania, 201 6.

Building Code Requirements for Structural Concrete and Commentary, ACI 318-11, American Concrete Institute, Farmington Hills, Michigan, 201 I.

Summary of the Seismic Adequacy of Twenty Classes of Equipment Required for the Safe Shutdown of Nuclear Plants," NP-71 49-D, Electric Power Research Institute, Palo Alto, California, USA, March 1991.

Summary of the Seismic Adequacy of Twenty Classes of Equipment Required for the Safe Shutdown of Nuclear Plantsn'o NP-7149-D, Supplement l, Electric Power Research Institute, Palo Alto, California, USA, January 1996.

Generic Seismic Ruggedness of Power Plant Equipment, EPRI NP-5223-SL, Rev. l, Electric Power Research Institute, Palo Alto, California, USA, August 1991.

Technical Study of Spent Fuel Pool Accident Risk at Decommissioning Nuclear Power Plants, NUREG-I738, U.S. Nuclear Regulatory Commission, Washington, D.C.,

February, 2001.

Handbook of Nuclear Power Plant Seismic Fragilities, NUREG/CR-3558, U.S. Nuclear Regulatory Commission, Washington, D.C., I 985.

lB$Gonsulting tlRtTzo 53.

54.

55.

s6.

57.

58.

59.

60

273429+R-036 Reaision 0 May 1.1.,201.7 Page 134 of 145 61.

62.

63.

64.

65.

66.

An Approach to the Quantification of Seismic Margins in Nuclear Power Plants, NUREG/CR-4334, U.S.Nuclear Regulatory Commission, Washington, D.C., August, 1985.

"Beaver Valley Power Station Unit 2, PRA Notebook, PRA-BV2-AL-R06, (AS) Level I Accident Sequence Analysis," First Energy Nuclear Operating Company, November 12, 2015.

"Beaver Valley Power Station Unit 2, PRA Notebook, PRA-BV2-AL-R06, (SO) Systems Analysis Overview and Guidance,n' First Energy Nuclear Operating Company, October 7, 201 5.

First Energy Nuclear Operating Company, "Beaver Valley Power Station Unit 2 PRA Notebook," PRA-BV2-AL-R06, (LE) Level 2 LERF Analysis, October 13, 2015.

"Beaver Valley Unit 2 Nuclear Power Station, 2002 WOG PRA Peer Review,"

January 2002.

Westinghouse Electric Company, "Fire PRA Peer Review against the Fire PRA Standard Supporting Requirements from Section 4 of the ASME/ANS Standard for Level llLarge Early Release Frequency Probabilistic Risk Assessments for Nuclear Power Plant Applications for the Beaver Valley Fire Probabilistic Risk Assessment", Letter, prepared for First Energy Nuclear Company, attachment to Letter, LTR-RAM-II-12-015, April 2012.

Letter, "Follow-on Fire PRA Peer Review Against the Fire PRA Standard Supporting Requirements From Section 4 of the ASME/ANS Standard For The Beaver Valley Unit I Fire Probabilistic Risk Assessment," Westinghouse Electric Company, prepared for First Energy Nuclear Company, attachment to LTR-RAM-II-I 1-008, April 201 1.

Westinghouse Electric Company, "RG 1.200 PRA Focused Peer Review against the ASME PRA Standard Requirements for the Beaver Valley Internal Flooding Probabilistic Risk Assessment," Letter, prepared for First Energy Nuclear Company, attachment to LTR-RAM-II-I 1 -093, September 8, 201 1.

RISKMANTM for Windows, Version 14.3, "IJser's Manual, I Overview Analysis,"

prepared by ABSG Consulting Inc. April 2015.

FENOC HRA Dependency Database v1.0.0 Help Guide, October 15,2013.

Beaver Valley Units I &2, Fire PRA Task I - Plant Boundary Definition and Partitioning", CalculationNO. 10080-Dec-3560, Rev. 1; May 16, 2011.

NFPA 805 Fire PRA Task 5.13 Seismic Fire Interactions," Document No. 8700-01.062-0035, Revision A, November 30, 2010, Scientech Calculation 1 77 56-04..

NOP-SS-1001 FENOC Administrative Program for Computer Related Activities.

EPzu Report 1016737, 'oTreatment of Parameter and Model Uncertainty for Probabilistic Risk Assessments," December 2008.

lBSGonsulting tiRlzz.o 67.

68.

69.

74.

71.

72 73.

74.

2734294-R-036 Reuision 0 May 1-1.,201.7 Pase 1.35 of 145 79 75.

76.

77.

78.

80.

81.

83.

85.

Stokoe, K. H., W. K. Choi, and F-Y Menq, 2003, "Summary Report: Dynamic Laboratory Tests: Unweathered and Weathered Shale Proposed Site of Building 9720-82 Y-12 National Security Complex, Oak Ridge, Tennessee," Department of Civil Engineering, The University of Texas at Austin, Austin, Texas, 2003.

Not Used.

McGuire, R.K, Silva, W.J., and Costantino, C.J.,2001, "Technical Basis for Revision of Regulatory Guidance on Design Ground Motions: Hazard-and Risk-Consistent Ground Motion Spectra Guidelines," NUREG/CR-6728, U.S. Nuclear Regulatory Commission, October.

Toro, 1996, "Probabilistic Models of Site Velocity Profiles for Generic and Site-Specific Ground Motion Amplification Studies, Description and Validation of the Stochastic Ground Motion Model," Report submitted to Brookhaven National Laboratory, Associated Universities, Inc. Upton, New York 1L973, Contract No. 770573, Published as Appendix D in W.J. Silva, N. Abrahamson, G. Toro and Costantino, 1996.

EPRI, 2015, "High Frequency Program Application Guidance for Functional Confirmation and Fragility Evaluation," EPRI Technical Report 3002004396.

Bozorgnia, Y., and Campbell, K.W., 2004, o'The Vertical to Horizontal Response Spectral Ratio and Tentative Procedures for Developing Simplified V/FI and Vertical Design Spectra," Journal of Earthquake Engineedrg, Vol. 8, No.2, 175-207.

Gulerce, 2., and Abratramson, N.A,2011, "Site-Specific Design Spectra for Vertical Ground Motion, Earthquake Spectrq" Vol. 27, No. 4, pp. 1023-1047.

NRC, z}ll,"Standard Review Plan: Section 3.7.1, Seismic Design Parameters, Revision 0, and Section 3,7.2, Seismic System Analysis, Revisior 3," NUREG-0800, U.S. Nuclear Regulatory Commission, Washington, D.C.

Arias, A., 1970, "A Measure of Earthquake Intensity,'o In Seismic Design for Nuclear Power Plants, Ed. R. J. Hansen, MIT Press, Cambridge, Massachusetts, 1970.

P.IZZO,20I l, "Spectral Matching Computer Program: RspMatch0g, Version l.l, User Manual," Paul C. Rizzo Associates, Inc., Pittsburgh, Pennsylvania, Revision 0, April 2011.

RIZZO,2012, "V&V for Spectral Matching Computer Program RspMatch09,"

Revision l, Paul C. Rizzo Associates, Ine., Pittsburgh, Pennsylvania, March 20L2.

IESGonsulting tlRtzzo 82.

84

2734294-R-036 Reaision 0 May LL,201.7 Page 1.36 af 145 86.

87.

88.

First Energy Nuclear Operating Company Letter L-l 6-282, Spent Fuel Pool Evaluation Supplemental Report, Response to NRC Request for Information Pursuant to 10 CFR 50.54(f) Regarding Recommendation 2.1 of the Near-Term Task Force Review of Insights from the Fukushima Dai-ichi Accident, Beaver Valley Power Station Units I and 2, dated Novemb er 7, 2016, ADAMS Accession Number ML 163 124"3 I I.

Not Used USNRC NUREG-1855, "Guidance on the Treatment of Uncertainties Associated with PRAs in Risk-Informed Decision-making," Revision l, 2013.

FirstEnergy Nuclear Operating Company, "Beaver Valley Power Station Unit 2 Probabilistic Risk Assessment Update Report," lssue 6 (PRA-BV2-AL-R06), Effective Date: December 31,2A15.

Electric Power Research Institute,"High Frequency Prograru, " EPR[-3002002997, September 2014.

EPRI Technical Report No. 3002004396, "High Frequency Program - Application Guidance for Functional Confirmation and Fragility Evaluation, Final Report, July 2015.

FirstEnergy Nuclear Operating Company, '.BVPS-2 Cumulative Risk of Screened SSCs for Seismic Initiators", PRA-BV2-17-007-R00, May 2017.

lEEGonsulting tlRtzzo 89.

90.

91.

92.

2734294-R-036 Reuision 0 May 11,2017 Pa*e 1.37 of L45 ABS AC ACI AF AISC AISX ANS AOV ASCE ASME ATWS AXLB BE BVPS BVPS-I BVPS.2 CABX CCF CDF CDFM CET CEUS CEUS.SSC CMU CNTB COV CP CRDM CTMT DAFW DBE DG DGBX DOE DWST EDG EL EPRI ERF ERFS ESEL 8.0 LIST OF ACRONYMS AND ABBREVIATIONS ABSG CONSULTING INC.

AIR CONDITIONING AMERICAN CONCRETE INSTITUTE AMPLIFICATION FACTOR AMERICAN INSTITUTE OF STEEL CONSTRUCTION ALTERNATE INTAKE STRUCTURE AMERICAN NUCLEAR SOCIETY AIR-OPERATED VALVE AMERICAN SOCIETY OF CIVIL ENGINEERS AMERICAN SOCIETY OF MECHANICAL ENGINEERS ANTICIPATED TRANSIENT WITHOUT SCRAM (ALSO ATWT, ANTICIPATED TRANSIENT WITHOUT TRIP)

AIIHLIARY BUILDING BEST ESTIMATE BEAVER VALLEY POWER STATION BEAVER VALLEY POWER STATION, UNIT 1 BEAVER VALLEY POWER STATION, UNIT 2 CHEMICAL ADDITION BUILDING COMMON.CAUSE FAILURE CORE DAMAGE FREQUENCY CONSERVATIVE DETERMINISTIC FAILURE MARGIN CONTAINMENT EVENT TREE CENTRAL AND EASTERN I-INITED STATES CENTRAL AND EASTERN UNITED STATES SEISMIC SOURCE CHARACTERIZATION CONCRETE MASONRY UNIT CONTROL BUILDING COEFFICIENT OF VARIATION COGNITIVE PROBABILITY CONTROL ROD MECHANISM CONTAINMENT DEDICATED AUXILIARY FEEDWATER DESIGN BASIS EARTHQUAKE DIESEL GENERATOR DIESEL GENERATOR BUILDING DEPARTMENT OF ENERGY DEMINERALIZED WATER STORAGE TANK EMERGENCY DIESEL GENERATOR ELEVATION ELECTzuC POWER RESEARCH INSTITUTE EMERGENCY RESPONSE FACILITY EMERGENCY RESPONSE FACILITY SUBSTATION EXPEDITED SEISMIC EQUIPMENT LIST lESGonsulting tlRtz7-o

273429+R-036 Reaision 0 May 1-L,20L7 Page L38 of L45 ESEP ESFAS F&O FE FEM FENOC FIRS FLEX FULB FV FVI FT FWS GERS GIP GMM GMPE GMRS HCLPF HCSCP HEP HF HFE HHSI HID HRA IF INTS IPEEE ISLOCA ISRS LB LERF LHSI LMSM LOCA LOOP LR LS.A LS.C M&E HVAC HX HZ IEEE EXPEDITED SEISMIC EVALUATION PROCESS ENGINEERED SAFETY FEATURE ACTUATION SYSTEM FACTS AND OBSERVATIONS FINITE ELEMENT FINITE.ELEMENT MODEL FIRSTENERGY NUCLEAR OPERATING COMPANY FOUNDATION INPUT RESPONSE SPECTRA DIVERSE AND FLEXIBLE MITIGATION STRATEGIES FUEL HANDLING AND DECONTAMINATION BUILDING FUSSELL.VESELY FUS SELL.VESELY IMPORTANCE FEET FEEDWATER SYSTEM GENERIC EQUIPMENT RUGGEDNESS SPECTRA GENERIC IMPLEMENTATION PROCEDURE GROUND MOTION MODEL GROUND MOTION PREDICTTON EQUATION GROUND MOTION RESPONSE SPECTRA HIGH CONFIDENCE OF A LOW PROBABILITY OF FAILURE HAZARD-C ONS I STENT STRAIN.COMPATIBLE PROPERTIE S HUMAN ERROR PROBABILITIES HIGH FREQUENCY HUMAN FAILURE EVENTS HIGH.HEAD SAFETY INJECTION HAZARD INPUTS DOCUMENT HUMAN RELIABILITY ANALYSIS HEATING, VENTILATION, AND AIR CONDITIONING HEAT EXCHANGER HERTZ INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS INTERVAL FREQUENCY INTAKE STRUCTURE INDIVIDUAL PLANT EXAMINATION FOR EXTERNAL EVENTS INTERFACING SYSTEMS LOCA IN-STRUCTURE RESPONSE SPECTRA LOWER BOI-IND LARGE EARLY RELEASE FREQUENCY LOW-HEAD SAFETY INJECTION LUMPED.MASS STICK MODELS LOSS OF COOLANT ACCIDENT LOSS OF OFFSITE POWER LOWER RANGE LIMIT STATE A LIMIT STATE C MECHANICAL AND ELECTRICAL lESGonsulting TlRIZZO r.I

273429+R-036 Reaision 0 May 11,201,7 Page 1-39 of 145 MAFE MCC MLOCA MOV MSCV NEI NEP NFPA NPTX NRC NSSS NTTF NUREG PDWS PGA PIPETT]NNEL PORV PSD PRA PSHA PWR PZR RAW RCBX RCP RCS REJ RTZZO RLYB RPS RRS RSGB RVT RW RWST SAP SASSI SBO SCDF SCE SEL SEWS SFGB SFP MEAN ANNUAL FREQUENCY OF EXCEEDANCE MOTOR CONTROL CENTER MEDIUM LOCA MOTOR.OPERATED VALVE MAIN STEAM CABLE VAULT NUCLEAR ENERGY INSTITUTE NON.EXCEEDANCE PROBABILITY NATIONAL FIRE PROTECTION AS SOCIATION NORTH PIPE TRENCH NUCLEAR REGULATORY COMMISSION NUCLEAR STEAM SUPPLY SYSTEM NEAR-TERM TASK FORCE US NUCLEAR REGULATORY COMMISSION REGULATION PRIMARY PLANT DEMINERALIZED WATER STORAGE PAD AND ENCLOSURE PEAK GROUND ACCELERATION PIPE TUNNELS PRES SURE.OPERATED RELTEF VALVE POWER SPECTRAL DENSITY PROBABILISTIC RISK ASSESSMENT PROBABILISTIC SEISMIC HAZARD ANALYSIS PRESSURIZED WATER REACTOR PRESSURIZER RISK ACHIEVEMENT WORTH REACTOR CONTAINMENT REACTOR COOLANT PUMP REACTOR COOLANT SYSTEM RUBBER EXPANSION JOINT RTZZO ASSOCIATES SWITCHYARD RELAY HOUSE REACTOR PROTECTION SYSTEM REQUIRED RESPONSE SPECTRA ERF DIESEL GENERATOR BUILDTNG RANDOM VIBRATION THEORY RIVER WATER REFUELING WATER STORAGE TANK PLANT DATABASE SYSTEM FOR ANALYSIS FOR SOIL-STRUCTURE-INTERACTION STATION BLACKOUT SEISMIC CDF SEISMIC CAPABILITY ENGINEER SEISMIC EQUIPMENT LIST SEISMIC EVALUATION WORK SHEETS SAFEGUARDS BUILDING SPENT FUEL POOL AESGortsultlng rlRtzzo qJ

2734294-R-036 Reoision 0 May 1,1, 20L7 Page L40 of 1.45 SPRA SPRAIG SPTX SRT STOR sQSS-TK21 SFR SGTR SHA SHS SI SLERF SLOCA SMA SOV SPID SPR SQUG SRT SRVB SSC SSE SSEL SSI SWBX SWGR TK TRBB TRS TSCR UB UFSAR UHRS UHS UR USI VAC VCT VDC V/H VPA VPB SEISMIC FRAGILITY ELEMENT WITHTN ASME/ANS PRA STANDARD STEAM GENERATOR TUBE RUPTURE SEISMIC HAZARD ANALYSIS ELEMENT WITHIN ASME/ANS PRA STANDARD SEISMI C HAZARD SUBMITTAL SAFETY INJECTION SEISMIC LARGE EARLY RELEASE FREQUENCY SMALL LOSS OF COOLANT ACCIDENTS SEISMIC MARGIN ASSESSMENT SOLENOID.OPERATED VALVE SCREENING, PRIORITIZATION, AND IMPLEMENTATION DETAIL S SEISMIC PRA MODELING ELEMENT WITHTN ASME/ANS PRA STANDARD SEISMIC PROBABILISTIC RISK ASSESSMENT SEISMIC PROBABILISTIC RISK ASSESSMENT IMPLEMENTATION GUIDANCE SOUTH PIPE TRENCH SEISMIC REVIEW TEAM STOREROOM SURROUNDTNG SHIELD WALL FOR REFUELING WATER STORAGE TANK SEISMIC QUALIFICATION UTILITIES GROUP SEISMIC REVIEW TEAM SERVICE BUILDING STRUCTURES, SYSTEMS, AND COMPONENTS SAFE SHUTDOWN EARTHQUAKE SAFE SHUTDOWN EQUIPMENT LIST SOIL STRUCTURE INTERACTION SOLID WASTE BUILDING SWITCHGEAR TANK TURBINE BUILDING TEST RESPONSE SPECTRA TRUNCATED SOIL COLUMN RESPONSE UPPER BOUND UPDATED FINAL SAFETY ANALYSIS REPORT T.]NIFORM HAZARD RESPONSE SPECTRA ULTIMATE HEAT SINK UPPER RANGE I.INRESOLVED SAFETY IS SUE VOLTS (ALTERNATING CURRENT)

VOLUME CONTROL TANK VOLTS (DIRECT CURRENT)

VERTICAL.TO-HORIZONTAL RIVER WATER VALVE, PIT TRAIN lESGotrsulting r\\Rtzzo r.I

273429+R-036 Reaision 0 May 11,2017 Pase L4L of 1.45 VSLOCA WTBX WUS VERY SMALL LOSS OF COOLANT ACCIDENTS WATER TREATMENT BUILDTNG WESTERN UNITED STATES AI A2 A3 AA AF AG AL AM AO AP AS ASP AT AW AX BI BK BL BP BX BY C1 C2 C3 C4 CC CD CE CG CI CP CS D3 D4 D5 D6 DC DO DP PRA Model Top Event Descriptions:

AUXILIARY FEEDWATER. SBO 1 HR MT AUXILIARY FEEDWATER - SBO 24 HR MT AUXTLTARY FEEDWATER (SBO)

FLEX ALTERNATE AFW PUMP AUXILIARY FEEDWATER - NON SBO SI ACCUMULATORS - GENTRA}.IS, SLOCA ST ACCUMULATORS. LLOCA SI ACCUMULATORS - MLOCA AC ELECTRIC POWER ORAI{GE TRAIN ALPHA MODE FAILURE AMSAC SIGNAL ALTERNATE SHUTDOWN PA}IEL AUXILIARY FEEDV/ATER - SGTR AUXILIARY FEEDWATER - ATWS AC POWER CROSS.TIE DUMMY TOP BASEMAT PENETRATION ERF (BLACK) DIESEL POWER LARGE CONTAINMENT BYPASS AC ELECTRIC POWER PURPLE TRAIN AC ELECTRIC POWER TRAINS ORAI{GE & PURPLE - DUMMY TOP EVENT CONTAINMENT BYPASSED CNMT FAILS PRIOR TO VESSEL BREACH CNMT FAILS AT VESSEL BREACH LATE CONTAINMENT FAILURE DUE TO BURN LONG TERM CNMT OVERPRESSURIZATION PRIMARY COMPONENT COOLING WATER SYSTEM OPERATOR COOLDOWN AND DEPRESSURIZE CNMT FAILS DUE TO EARLY H2 BURN LEVEL 1 OR LEVEL 2 SEQUENCE GROUP CONTATNMENT ISOLATION FAILURE TO COOL DEBRIS IN VESSEL TURBINE PLANT COMPONENT COOLING WATER I25V DC 2-3 SUPPLY I25V DC 2.4 SUPPLY 125V DC BATTERY 2-5 SUPPLY I25V DC BATTERY 2-6 SUPPLY FAILURE TO COOL DEBRIS EX.VESSEL I25V DC 2-I SUPPLY I25V DC 2-2 SUPPLY lESConsulting

[]Rtzzo

273429+R-036 Reaision 0 May 1.'1,,201.7 Pase 142 of745 DX DY DZ GE GL GP H3 HC HE HH HL HM HR IA IB IC IP IR IS IW ry LI L2 L3 L4 LC LD LE LH LL LM LS M1 M2 M3 M4 M5 M6 MA ME MF MS MU NA ND I25V DC 2-1 AND 2-2 SUPPLY. DUMMY TOP 125V DC BATTERY 2-5 A}ID 2-6 SUPPLY - DUMMY TOP 125V DC 2-3 AI{D 2.4 SUPPLY. DUMMY TOP FLEX 48OV GENERATOR PORTABLE AC GENERATOR FOR SG LEVEL INSTR ALTERNATE HIGH PRESSURE SG FEEDWATER PUMP LATE BURN OF COMBUSTIBLE GASES HHSI PATH TO COLD LEGS - GENERAL TRAhISIENT HYDROGEN BURN WITHIN 4 HRS OF VB HHSI/CHARGING PUMP TRATNS HHSI PATH TO COLD LEGS - LLOCA HHSI PATH TO COLD LEGS - MLOCA HIGH HEAD COLD LEG RECIRCULATION INSTRUMENT AIR SUPPLY VITAL BUS CHANNEL III (BLUE)

CONTAINMENT INSTRUMENT AIR SYSTEM INDUCED RCS HOT LEG OR SURGE LINE RUPTURE VITAL BUS CHANNEL I (RED)

TEMPERATURE INDUCED SG TUBE RUPTURE VITAL BUS CHANNEL TI (WHITE)

VITAL BUS CHANNEL IV (YELLOW)

LARGE CONTAINMENT FAILURE PRIOR TO VB LARGE CONTAINMENT FAILURE @ VB LARGE LATE CONTAINMENT FAILURE LARGE LONG TERM CNMT OVERPRESSURIZATION FAILURE LHSI PATH TO COLD LEGS. GENERAL TRANSIENT FLEX LOAD SHED LARGE CNMT FAILURE FROM EARLY H2 BURN LHSI PUMP TRAINS LHSI PATH TO COLD LEGS - LLOCA LHSI PATH TO COLD LEGS - MLOCA INDUCED PORV LOCA 480V MCCS (ORANGE) - E07 AhrD El I 480V MCCS (PURPLE) - E08 AND E12 480V MCC (ORANGE) - E03 480V MCC (PURPLE) - E04 480V MCCS (ORANGE) - Eos AND El3 480V MCCS (PURPLE) - E06 AND Er4 MAKEUP TO PPDWST AND AFW PUMPS HIGH PRESSURE MELT EJECTION MATN FEEDWATER FOR SECONDARY HEAT REMOVAL MAIN STEAM ISOLATION MAKEUP TO RV/ST - RECIRC. FAILURE NORMAL BUS 2.4 AC POWER NORMAL BUS 2D AC POWER lEtGonsulting rlHlzzo LJ

273429+R-036 Rutision 0 May LL,20L7 D^-^ 1 na ^l 7 itr NM NR NX OA OB OC OCL OD OF OG OL OP OR OS OT OV PA PI PK PL PR PT QS R1 R2 R3 RC RD RE RI RL RP RR RS RT RW RX SA SB SD SE SL SM SP SS NO MELT CONDITION FROM INJECTION PHASE RECIRCULATION FROM SUMP NOT REQUIRED NORMAL BUS 2A AND 2D AC POWER. DUMMY TOP OPERATOR INITIATES EMERGENCY BORATION - ATWS FEED & BLEED COOLING OPERATOR TRIPS RCPS DURTNG LOSS OF SEAL COOLING OPERATOR TRIPS RCPS DURTNG SEAL LOCA (30)

DEPRESSURIZATION OF RCS FOR RHR ENTRY OPERATOR RESTORES h{ATN FEEDWATER OFFSITE GRID OPERATOR RESTORES COOLING TO SCRUB FAULTED SGTR OPERATOR PREMATURELY TERMINATES SI ALIGNMENT FOR RECIRCULATION OPERATOR INITIATES SAFETY INJECTION OPERATOR TRIPS REACTOR I\\{ANUALLY - SHORT TIME OPERATOR OPENS EDG BLDG DOORS & TEMP FAhIS RCS PRESSURE RELIEF - ATWS PORV ISOLATION PORVS AhID SAFETY VALVES RECLOSE - ATWS POWER LEVEL BELOW 40% - ATWS RCS PRESSURE RELIEF PROPANE TANK FARM DURTNG EARTHQUAKE QUENCH SPRAY SYSTEM SERVICE WATER TRAIN A TO RSS SERVICE WATER TRATN B TO RSS SERVICE U/ATER TRATN A & B TO RSS - DUMMY TOP RECIRCULATION SPRAY TRAIN C RECIRCULATION SPRAY TRAIN D ELECTRIC POWER RECOVERY OPERATOR MANUALLY INSERTS CONTROL RODS. ATWS RCP SEAL LOCA RCS PRESSURE AT VESSEL BREACH RESIDUAL HEAT REMOVAL TRAINS RECIRCULATION SPRAY TRAINS A & B REACTOR TRIP RWST SUPPLY RECIRCULATION SPRAY TRAINS C & D. DUMMY TOP EVENT SOLID STATE PROTECTION SYSTEM TRATN A SOLID STATE PROTECTION SYSTEM TRAIN B SHUTDOWN SEAL ACTUATES RCP SEAL COOLING SECONDARY LEAKAGE TO ATMOSPHERE SUCTION FROM CONTAINMENT SUMP REACTOR COOLANT PUMP SEAL LOCA NO MELT FROM LEVEL 1 lESGonsutting F\\RIZZO

\\.I

273429+R-036 Reoision 0 May 1L,201.7 Dnno 1 tl tl nT 1tC, SW SX TB TR TT VI WA WB WC WM XC XL XM XT xx ZAF ZAI ZAT ZCC ZCD ZCI ZCP ZD5 ZDC ZDG ZDI ZDO ZDP ZGL ZHz ZHC ZHH ZHR ZHV ZIC ZIL ZIN ztv ZLL ZLz ZLI(

ZLO ZLP ZMz FLEX DC TRAIN SWAPOVER SOLID STATE PROTECTION SYSTEM TRAIN A & B RCP THERMAL BARRIER COOLING PRESSURE INDUCED SG TUBE RUPTURE TURBINE TRIP REACTOR VESSEL INTEGRITY - ATWS SERVICE/STAI{DBY SW TRAIN-A PUMPS & FLOW PATH SERVICE/STAI{DBY SW TRATN-B PUMPS & FLOW PATH SERVICE WATER/STANDBY SW SYSTEM TRAINS A & B MAKEUP TO RWST - SGTR, SECONDARY LEAK HHSI & LHSI COLD LEG PATHS. GT, DUMMY TOP HHSI & LHSI COLD LEG PATHS - LLOCA, DUMMY TOP HHSI & LHSI COLD LEG PATHS. MLOCA, DUMMY TOP STATION AC POWER CROSS TIE USE FOR SENSITIVITIES AFW - PPDWST OR ALL 3 PUMPS CT. ISOL..INBRD AOV RCBX 692 REACTOR INTERNALS FOR ATWS PCCW. MEJ REJ PUMPS & HXS; SURGE TANK COOLDOWN A}ID DEPRESSURIZE CONTAINMENT ISOLATION VALVES LARGE CONTAINMENT PENETRATIONS DC TRAIN I-5. BATTERY & CHARGER& SWBD EMERG. DC. SWBD BATTERIES CHARGER DIESEL GENERATORS & SUPPORT CT. ISOL. - INBRD DIAPHRAGM RCBX 718.721 CT. ISOL. - OUTBRD DIAPHRAGM MSCV 7I8 DEPRESSURIZE RCS FOR RHR PORTABLE GENERATOR HHSI PUMP SUCTION VLVS HHSI PATH TO COLD LEG HIGH HEAD SAFETY INJECTION HIGH PRESSURE RECIRC - MOVS 2CHS-TK22 (VOLUME CONTROL TANK)

CONTAINMENT INSTRUMENT AIR CNMT ISOL LETDOWN VLVS; INBOARD AOVS CNMT ISOL LETDOWN VLVS; INBOARD LCVS CNMT ISOL VENTS&DRAINS; INBOARD DIRECT CORE DAMAGE DIRECT CD ANID LERF - RCBX; MSCV; SGS SMALL RCS LOCAS CNMT ISOL LETDOWN VLVS; OUTBOARD LHSI TRAINS MCC.z-EI2 lESGonsultirtg

()Rtzzo

2734294-R-036 Rutision 0 May 11,201.7 Pase L45 of145 ZMl ZMs ZM6 ZMA ZMS ZNIU zo3 zo4 ZOB ZOG ZOR zov ZPO ZPT ZQS ZRIA ZRIB ZP-z ZR3 ZP(4A ZR4B ZRR ZRS ZRW ZS2 ZSM ZSO ZSV ZSW ZTB ZT){

ZVS ZX ZY ZZ MCC.2-E03, EO4 MCC-2-E05, E13 MCC.2-E06, EI4 MAKEUP TO PPDWST A}.ID AFW MATN STEAM ISOLATION MAKEUP TO RWST - PUMPS MOVS CONTROL ROOM TNDICATION PANELS CONTROL ROOM CEILING PZR PORVS& PSVS&BLOCK VALVES AS.IS OFFSITE POWER. NON-SEISMIC SWGR ALIGNMENT FOR RECIRCULATION CNMT ISOL VENTS&DRAINS; OUTBOARD CT. ISOL.. OUTBRD PRSSR RLF MSCV 718.725 PROPA},IE TANK FARM QSS -NOZZLES& PUMPS EDG 2-1 RELAY (EQIooA)

EDG 2-2 RELAY (EQI00B)

RELAY FRAGTLTTY GROUP EQl02 (AR440AR & MASTER)

RELAY FRAGTLTTY GROUP EQl03 (RK223068-AP)

RELAYS FOR EDG 2-l (DGF AND D3)

RELAYS FOR EDG 2-2 (DGF AND D3)

RHR. MOVS& PUMPS&ruffi RECICULATION SPRAY. PUMPS&HXS&HEADER REFUELING WATER STORAGE TANK SERVICE WATER SYSTEM FAILURES INTS CTMT SUMP SUCTION OR 2RSS-P21A; B; C; D CT. ISOL. - OUTBRD SOV MSCV 718 SG SAFETY RELIEF VALVES STANDBY SERVICE WATER PUMPS&AISX FAILURES RCP THERMAL BARRIER COOLING TURBTNE BUILDING VERY SMALL LOCA NON.SETSMIC INITTATING EVENT (SEISMIC TREE)

NO SEISMIC FAILURES (SUPPORT TREE)

NO SETSMTC FAILURES (GENTRAI{S TREE) lESGonsultlng

{}R}zzo

2734294-R-036 Ratision 0 May L1.,201.7 Page A1 of Aa9 APPENDIX A

SUMMARY

OF SPRA PEER REVIEW AND ASSESSMENT OF PRA TECHNICAL ADEQUACY FOR RESPONSE TO NTTF 2.1 SEISMIC 50.54(F) LETTER A.L Overview of Peer Review The Beaver Valley Power Station (BVPS)-2 PRA was subjected to an independent peer review against the pertinent requirements in Part 5 of the ASME/ANS PRA Standard (Reference 4).

The peer review assessment, and subsequent disposition of peer review findings, is summarized here (for the final report, see Reference 6). The scope of the review encompassed the set of technical elements and supporting requirements (SR) for the Seismic Harard Analysis (SHA),

seismic fragilities (SFR), and seismic PRA modeling (SPR) elements for seismic core damage frequency (CDF) and large-early release frequency (LERF). The peer review therefore addressedthe set of SRs identified inTables 6-4 through 6-6 ofthe Screening, Prioritization, ffid Implementation Details (SPID) (Referenc e 2).

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 Revision 2 (Reference l6) and the requirements in Section 1-6 of the ASME/ANS PRA Standard (Reference 4), and presents the significant results of the peer review.

The BVPS Units I and 2 SPRA peer review was conducted during the week of December 1, 2014, at the F'irstEnergy Nuclear Operating Company (FENOC) offices in Aluon, Ohio. As part of the peerreview, awalkdown of portions of BVPS Units I and 2 was performed on December 1,2014, by two members of the peff review team who have the appropriate Seismic Qualification Utilities Group ( SQUG) training.

4.2 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 (Reference 4), using the peer review process defined in NEI L2-13 (Reference 5).

The review was conducted over a four-day period, with a summary and exit meeting on the evening of the fourth day.

The SPRA peer review process defined in (Reference 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 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.

lESGonsulting tlR.zzo

2734294-R-036 Rwision 0 May 1L,20L7 Page A2 of A48 For each area, i.e., SHA, SFR, SPR, a team of two to three peer reviewers were assigned, one having lead responsibihty 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 fuIl 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 (Reference 5): Findings, which identiff issues that must be addressed in order for an SR (or multiple SRs) to meet Capability Category II; Suggestions, which identiff 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.

Section 5 of the ASME/ANS PRA Standard contains a total of 77 SRs under three technical elements. Three (3) of the supporting requirements were judged to be not applicable, and therefore the remainingT4 SRs were reviewed.

A,.3 Peer Review Team Qualifications The review was conducted by Dr. Andrea Maioli and Mr. Kenneth Kiper of Westinghouse, Dr. Martin McCann of Jack R. Benjamin & Associates, Dr. Bob Youngs of AMEC, Mr. Steve Eder of Facility Risk Consultants, Mr. Nathan Barber of Pacific Gas & Electric, Mr. Deepak Rao of Entergy, Dr. Se-Kwon Jung of Duke Energy and Mr. Don Moore of Southern Company. Appendix D of the peer review report (Reference 6) contains the resumes for the reviewers. Reference 6 Table 2-2 shows the review assignments for each reviewer.

Dr. AndreaMaioli, the team lead, has over l0-years' experience at Westinghouse inthe nuclear safety area generally and seismic PRA specifically. He has served as lead engineer for a number of seismic PRA and seismic margin studies for existing and new nuclear power plants.

Dr. Martin McCann was the lead for the SHA technical element. He has 30 years' experience in engineering seismology including site response analysis, specification of ground motion. He was assisted in the hazard review by Dr. Bob Youngs, an internationally-recognized expert in seismology and earthquake hazard assessment.

Mr. Stephen 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 Dr. Se-Kwon Jung and Mr. Donald Moore. Mr. Moore has over 45 years of experience in specialized technical positions and supervisory positions in the field of structural engineering with specific emphasis on seismic analysis and design, seismic risk assessments, and seismic qualification of equipment and subsystems. Dr. Jung has over 10 years' experience in the field of civil and structural engineering with focus on fragility evaluation in support to seismic PRAs.

lEEGonrulting

[]Rtzzo

2734294-R-036 Raision 0 May 11.,201-7 Page A3 of A48 Mr. Ken Kiper was the lead forthe System Response (SPR) technical element. Mr. Kiper joined Westinghouse as a Technical Manager after a 3 l-year career in Seabrook Station. He has experience in virtually every aspect of PRA modeling and applications, including upgrading and maintaining the RISKMAN Seabrook seismic PRA. He was assisted by Mr. Nathan Barber and Mr. Deepak Rao. h{r. Barber has more than 12 years' experience in multiple aspects of PRAs; he is the lead for the Diablo Canyon seismic PRA RISKMAN model update and maintenance.

Mr. Rao has 31 years' experience in essentially every aspects of PRA.

Two working observers (Boback Torkian, Enercon and Tommy John, Dominion) supported the review of the SPR and SFR technical elements. Any observations and findings these working observers generated were given to the peer review team for their review and "ownership." As such, Mr. Torkian and Mr. John assisted with the review but were not formal members of the peer review team.

None of the peer review team members had any involvement in the development of the BVPS-2 SPRA. The peer review team members met the peer reviewer independence criteria in NEI 12-13 (Reference 5).

4.4 Summary of the Peer Review Conclusions The review team's assessment of the SPRA elements is summarized 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 earthquakes 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 seismic source characterization (SSC) and the ground motion modeling (GMM) inputs to the analysis. The SSC ingutl to the PSHA are based on the recently completed Central and Eastern U.S. (CEUS) seismrc 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 satisff the requirements of the standard for data collection and use of a structured expert elicitation process. The SSHAC process describes a process 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 a sufficient depth and detail necessary to satisff scientific and regulatory needs. The level of study is not mandated in the Standard; however, boththe SSC and the GMM parts of the PSHA were developed as aresult of SSHAC Level 3 analyses. Inthe case of the GMM, a SSHAC Level 2 analysis was carried out to update a prior Level 3 study. These Level 3 studies satisfi, 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 paleoseismic information within 320 km of the site. This data collection effort was carried out as part of the CEUS and GMM projects that were the basis lESConsulting tiRtzz.o

2734294-R-036 Ratision 0 May LL,20L7 Page A4 of A48 for the inputs to the Beaver Valley PSHA. To ensure that the database of information that is the hasis for the PSHA is up-to-date, the PSHA analysts did not systematically conduct a review to identiff and gather new geological, seismological, or geophysical data available since the completion of the CEUS-SSC study or information at alevel of detail thatwas notconsidered in the CEUS-SSC regional study that would indicate there should be new seismic sources added to the SSC model or changes to existing sources.

While a systematic review and update effort was not carried out, the PSHA analysts did gather data to update the earthquake catalog to assess whether there was new information since the completion of the CEUS-SSC Project that should be used to update the seismicity parameters. A subjective review of the updated catalog was conducted to conclude that an update to the seismicity parameters was not required.

As part of the CEUS-SSC model sources potentially damaging earthquakes that could occur in the CEUS were modeled. This includes all distributed seismic sources within 640 km and all Repeated Large Magnitude Earthquake (RLME) sources within I 000 km of the BV site. In the implementation of the CEUS model for the Beaver Valley site, all seismic sources in the CEUS model were included in the PSHA. By including all the CEUS seismic sources in the analysis, the contribution of "near-" and "far-field" earthquake sources to ground motions at Beaver Valley were considered.

The Davis-Besse peer review identified the fact that the PSHA software that was used to perform the probabilistic hazard quantification did not perform the uncertainty analysis correctly. This error was not corrected for the Beaver Valley PSHA; therefore, the uncertainty results are not correct in this analysis as well. This error does not impact the estimate of the mean hazard, but it does affect the estimate of the uncertainty in the PSHA results. Consequently, the PSHA inputs to the SPRA uncertainty quantification are incorrect.

The SHA for the Beaver Valley site took into account the effects of local site response.

However, the review team did not find adequate documentation to support the site-specific velocity profile used inthe analysis. Also, because of the limited site-specific data, the study could not properly account for velocity uncertainties as required by the standard. The review also noted that aleatory and epistemic uncertainties in the site response were not separately combined with the uncertainty in the rock seismic hazard results. As a result, the uncertainty in the soil site hazard results is likely underestimated.

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 PSHA. The PSHA fully accounted for the "near-" and "far-field" 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. The PSHA documentation does provide certain information such as magnitude-distance deaggregation plots that provide insight into contributors to the site hazard. However, the PSHA documentation does not provide the results of a systematic sensitivity analysis that evaluates the importance and sensitivity of key parameters to the results. As a result this requirement was not met.

lESGontulting

{}Rtzzo

2734294-R-036 Rwision 0 May L1.,20L7 Page A5 of AaB As required by the Standard, a screening analysis was performed to assess whether in addition to the vibratory ground motion, other seismic hazards, such as fault displacement, landslide, soil liquefaction, or soil settlement, need to be included in the seismic PRA. The review identified a number of areas where funher information should be provided to support the conclusion that other seismic hazards can be screened out. Because of the limitations in the review and screening of other hazards, SHA-I-2 is at this time identified as not MET, pending the resolution of the issues identified in SHA-I-1. This SR can be non-applicable if all the other hazards are indeed confirmed as screened out, or not met if some hazard needs to be retained.

Both the aleatory and epistemic uncertainties have been addressed in characterizing the seismis sources. In addition, uncertainties in each step of the hazard analysis were propagated and displayed in the final quantification of hazard estimates for the Beaver Valley site. As noted above, the PSHA software that was used to perform the hazard calculations implements an approach for the propagation of the uncertainties in the analysis that is not correct. As a result, the uncertainty in the seismic hazard is not properly quantified.

ln summary, the PSHA performed for the BVPS is based on the CEUS and GMM regional studies which are SSHAC Level 3 efforts. There are a couple of instances where the standard is not met, including a computational issue with the PSHA software that impacts the uncertainty analysis. The PSHA is well documented which supports the review process and its future use by FENOC.

SFR The Standard requires that all the structures, systems, and components (SSC) that play a role in the seismic PRA be identified as candidates for subsequent seismic-fragility evaluation. This was performed through the development of the Walkdown Seismic Equipment List (SEL). As permitted by the Standard, extremely seismically rugged and seismically insensitive items in the list were screened out - i.e., no seismic-fragility evaluation is required for these items.

Additional high seismic capacity screening was performed for systems and components using the Electric Power Research Institute (EPRI) seismic margins screening tables. As required by the Standard, anchorage adequacy was verified when generic functional capacity was used. Some of the items with 0.509 based generic capacity ended up being top contributors to CDF. For these cases, no additional justification for use of the generic fragilities was provided as required by the Standard.

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 and then subsequently utilized in the evaluation of seismic fragilities. New 3-D building models were developed for all structures and used for this purpose. However, the review team noted that the modeling methods and the performance objective for the building response analysis were suitable for the calculation of fragilities for equipment and relays (based on the Conservative Deterministic Failure Margin [CDFMJ approach), but not realistic for the calculation of fragilities for the building structures (based on the separation of variables approach). The review teams also noted that simpliffing assumptions used in the soil-structure interaction analyses of buildings were not fully justified and that sensitivity studies or other more detailed evaluation may be warranted.

AESGonsulting rlRrzzo

2734294-R-036 Ratision 0 Moy 1,L, 20L7 Page A6 of 448 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 evaluated the potential for seismic-induced fires and floods, and found no hazard sources. The walkdown observations were subsequently incorporated in the seismic-fragility evaluations. However the review team noted some inconsistencies between the configurations assumed for the anchorage fragility calculations and actual field conditions, which resulted in excess conservatism.

The SPRA identifies the relevant failure modes for the SSCs through a review of plant design documents, earthquake experience data, and walkdowns. Subsequently, seismic-fragility evaluations were performed for the critical failure modes of the SSCs. The review team noted however that the failure modes, analytical assumptions, and associated capacities assigned to certain SSCs including relay fragilities have conservative bias and are thus not realistic.

The Standard requires that the seismic-fragility parameters be based on plant-specific data supplemented as needed by earthquake experience data, fragility test data, ffid generic qualification test data. The review team found that this requirement was satisfied, but noted as described above that certain fragilities are not realistic and that the basis for use of generic lower bound fragilities should be revisited in certain cases.

In conclusion, seismic fragilities were developed for strucfures, systems, and components associated with the SEL. This included development of new building models and performance of site-specific response analyses for generation of in-sfructure response spectra. Component screening was performed using available industry guidance at 0.509. [Note that although the peer review report says 0.5g, the final screening value for BVPS-2 was increased to 0.7g during the process of model refinements.] Thorough walkdowns were performed and documented.

Many detailed calculations were performed to assess SSC fragility, and the documentation was comprehensive. However unrealistic assumptions were noted in different steps of the evaluation process, resulting in fragilities with conservative bias.

SPR The plant response model developed for the BVPS-2 SPRA represents a state-of-the-art model and documentation that fully meet the requirements of the Standard. The model, as reviewed, represents a final-draft version, which will need to be finalized along with the standard quantifi cation steps and revised documentation.

The SPRA model was developed by modifuine 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 effors, The HRA modeling and documentation was recognized as a best practice. This HRA used the EPRI HRA Calculator and adjusts performance shaping factors (PSF) to account for four levels of earthquake intensity. Specific adjustments were made to the delay time and execution time, to stress, and to cognitive work load. These adjustments were implemented through the HRA Calculator for each action modeled in the SPRA.

lESGonsulting tlR.zzo

2734294-R-436 RerJision 0 May 1.1,2017 Page A7 of A48 The use of RISKMAN in the seismic model development and quantification fully met the challenges of integrating a seismic risk model. A significant number of sensitivities were performed to understand the impact of the various modeling and screening assumptions. In these aspects, the quantification of the BVPS-2 SPRA is judged to meet the PRA Standard.

It is apparent that the quantification process was used to inform as appropriate the fragility aspects; e.9., selection of the screening values and of the specific fragility items to be refined.

The peer review team concluded that the BVPS-2 SPRA has an appropriate level of resolution for CDF evaluation, but that conservative fragilities may be masking some of the LERF contributors.

The FENOC PRA team went beyond the current state-of-practice in addressing seismic-induced fires ffid, especially, seismic-induced floods, leveraging the existing fire and floods PRA for a more systematic assessment of these scenarios. This was recognized as a best practice by the peer review team.

In conclusion, the seismic PRA model integrates the seismic hazard, the seismic fragilities, and the systems-analysis aspects appropriately to quanti+/ CDF and LERF. The seismic PRA analysis was extensively documented in a manner that facilitates applying and updating the SPRA model.

4.5 Summary of the Assessment of Supporting Requirements and Findings Table,4-f presents a sunmary of the SRs graded as not met or not Capability Category II, and the disposition for each. Section A,10 presents summary of the Finding F&Os and the disposition for each.

AESGonsulting tiRtzz.o

2734294-R-036 Ratision 0 May 1.L,2017 Page AB of A48 TABLE A.1

SUMMARY

OF'SRS GRADED AS NOT MET OR CAPABILITY CATEGORY I FOR SUPPORTING REQUIREMENTS COYERED BY THE BVPS-2 SPRA PEER REVIEW SR Assnssnn C.+,panrlrrv Canrconv AssocrnrED F'nrunrc F&Os DrsrosrrroN ro Acurnvn MET on Cnranrlrrv C.r.rncony II SHA SHA.F1 CC-I

?-1 1-)

L L)

L L

Associated F&Os have heen resolved. SR is iudged to now achieve CC-II.

SHA.F2 Not MET 2-3 Associated F&Os have been resolved. SR is iudeed to be met.

SHA-I2 Not MET 2-26,2-27,2-28,2-29,2-3r Associated F&Os have been resolved. SR is iudged to be met.

SHA.J3 Not MET 2-30 Associated F&Os have been resolved. SR is iudeed to be met.

SFR SFR-A'2 CC-I 4-6,4-13, 4-t6 F&Os 4-13 and 4-16 have been resolved as prescribed by the peer review team.

For F&O 4-6, further justification has been provided as to why the generic fragilities described in the F&O are acceptable for use, per HLR-SFR-F, as directed in SFR-A2. This is demonstrated through the use of sensitivrty studies. See the "Plant Response or Disposition" section of this F&O in Section 4.10.

This SR is iudeed to now achieve CC-[.

SPR

[None]

N/A N/A N/A A.6 Summary of Technical Adequacy of the SPRA for the 50.54(f) Response The set of SR from the ASME/ANS PRA Standard (Reference 4) that is identified in Tables 6-4 through 6-6 ofthe SPID (Reference2) define the technical attributes of aPRA 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 BVPS-2 SPRA model meets the expectations for PRA scope and technical adequacy as presented in RG 1.200, Revision 2 (Reference l6) as clarified in the SPID (Reference 2).

lESConsulting

(]Rtzzo

2734294-R-036 Raision 0 May 1,L,2017 A9 48 The main body of this report provides a description of the SPRA methodology, including:

Summary of the SHA (Section 3)

Summary of the structures and fragilities analysis (Section 4)

Summary of the seismic walkdowns performed (Section 4) r Summary of the internal events at power PRA model on which the SPRA is based, for CDF and LERF (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 Revision 2 is available if required to facilitate the Nuclear Regulatory Commission (NRC) staff s review of this submittal.

The BVPS-2 SPRA reflects the as-built and as-operated plant as of the cutoff date for the SPRA, October 25,2016. This includes outage modifications, non-outage modifications, and other configuration control items through September 30, 2016. There are no permanent plant changes that have not been reflected in the SPRA model.

L.7 Summary of SPRA Capability Relative to SPID Table 6-4 through Tahle 6-6 The Owners Group performed a full scope peer review of the BVPS-2 internal events PRA and internal flooding PRA that forms the basis for the SPRA to determine compliance with ASME PRA Standard, RA-S-2008, including the 2009 Addenda A (Reference 4) and RG 1.200 (Reference 16) during the week of June 6, 201 I. This review documented findings for all SRs which failed to meet at least Capability Category II. AII of the internal events and internal flooding PRA peer review findings that may affect the SPRA model have been addressed.

The Owners Group performed a peer review of the BVPS-2 SPRA in December 2014. The results of this peer review are discussed above, including resolution of SRs not assessed by the peer review as meeting Capability Category II, and resolution of peer review findings pertinent to this submittal. The peer review team expressed the opinion that the BVPS-2 seismic PRA model is of good quality and integrates the seismic hazard, the seismic fragilities, and the systems-analysis aspects appropriately to quantit/ CDF and LERF. The general conclusion of the peer review was that the BVPS-2 SPRA is judged to be suitable for use for risk-informed applications, lEEGonsulting

[]Rtzz.$

2734294-R-036 Rsoision 0 May 11,20L7 Page A10 of AaB I

Table A-1 in Section A.5 provides a summary of the disposition of SRs judged by the peer review to be not met, or not meeting Capability Category II.

. Section A.I0 provides a sunmary of the disposition of the open SPRA peer review findings.

Table A-2 provides an assessment of the expected impact onthe results of the BVPS-2 SPRA of those SRs and peer review Findings that have not been fully addressed.

TABLE A-2

SUMMARY

OF IMPACT OF NOT MET SRS AND OPEN PEER REVIEW FINDINGS A.8 ldentification of Key Assumptions and Uncertainties Relevant to the SPRA Results The PRA Standard (Reference 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 (Reference 88) and EPRI 1016737 (ReferenceT4) provide guidance on assessment of uncertainty for applications of a PRA. As described in NUREG-1855 (Reference 88), sources of uncertainty include "parametric" uncertainties, oomodeling" uncertainties, and "completeness" (or scope and level of detail) uncertainties.

Parametric uncertainty was addressed as part of the BVPS-2 SPRA model quantification (see Section 5 of this submittal).

r 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 BVPS-2 SPRA technical elements are noted in the SPRA documentation that was subject to peer review, and a sunmary of important modeling assumptions is included in Section 5. These important modeling assumptions were considered when identifying sensitivity cases for quantification.

lESGon*ulting

[]Rtzz.o F&O SunaunRY oF Issuu Nor Fur,r,v Rnsor,vnn Iprp.q.cr oN SPRA Rpsulrs N/A N/A This table is not applicable, as all F&Os listed in Table A-1 have been fully dispositioned in Section A.10. It is judged by the utility that the associated SRs now achieve at least CC-[ (or MET, for SRs in which no capability category is assigned), ffid that no further action is needed to address any SPRA F&Os. This table is retained to maintain the numbering order from the template.

2734294-R-036 Ratision 0 Moy 11.,20L7 Page 41.1. of 448 r 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. For example, the current seismic PRA only considers scenarios initiated from power operation, not from shutdown conditions. A few specific issues of PRA completeness were identified in the SPRA peer review and the associated F&Os were addressed and resolved.

A summary ofpotentially important sources of uncertainty inthe BVPS-2 SPRA is listed in Table A-3.

TABLE A-3

SUMMARY

OF POTENTIALLY IMPORTANT SOURCBS OF UNCERTAINTY PRA Elrnarcxr SuprmaRy oF TREATMENT oF Souncrs oF UucnnrArNTY PER Pnrcn Rrvrrcw PornxuAI, IMPACT oN SPRA REsuLrs Seismic Hazard The BVPS-2 SPRA peer review team noted that 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 Beaver Valley site. As noted above, the PSHA software that was used to perform the hazard calculations implements an approach for the propagation of the uncertainties in the analysis that is not correct.

The BVPS-2 SPRA peer review team noted that the uncertainty in the seismic hazard is not properly quantified. In response, associated F&Os were addressed and resolved. The seismic hazard reasonably reflects sources of uncertainty.

Seismic Fragilities Section 5.7.1.6 of the main report presents sensitivities performed which adjust the high confidence of a low probability of failures (HCLFP) of the top seismic SSC failures to assess the impact of assumptions and uncertainties in the fraeility calculations.

Seismic PRA Model Section 5.7.l of the main report presents sensitivities performed that assesses the impact of assumptions and sources of uncertainties in the SPRA model.

lESConeulting tlRrzzo

2734294-R-036 Rmision 0 May L1.,20L7 Page A12 of AaB A.9 Identification of Plant Changes Not Reflected in the SPRA The BVPS-2 SPRA reflects the as-built and as-operated plant as of the cutoff date for the SPRA, which was October 25,2016. This includes outage modifications, non-outage modifications, and other configuration control items. Tahle,4-4 lists significant plant changes subsequent to this date and provides a qualitative assessment of the likely impact of those changes on the SPRA results and insights.

TABLE A-4

SUMMARY

OF SIGNIF'ICANT PLANT CHANGES SINCE SPRA CUTOFF DATB DrscmrrloN oF PLAr\\rr Cnnucn Iurncr ox SPRA Rnsulrs N/A This table is not applicable, as there have been no significant plant changes from the date of SPRA modeling cutoff. This table is retained to maintain the numbering order from the template.

4.10 Summary of Finding F&Os and Disposition Status Note that some findings only pertain to Unit 1 and are noted that way in the details of the finding. The dispositions of these findings are judged to have resolved the issues identified and thus the seismic PRA meets Capability Category II or higher for all supporting requirements in Section 5 of the ASME/ANS PRA standard (Reference 4). It is believedthatthe standard bounds the SPID, however it has been identified that the SPID contains specific guidance that differs from the Standard or expands it in 16 different areas. These 16 topics are specifically addressed below. Based on this and the results of the peer review along with the resolutions to the findings the SPRA is judged to meet or exceed the SPID (ReferenceZ).

Topic 1: Seismic Hazard (SPD Sections 2.112.2, and 2.3)

The PSHA submitted to the NRC in response to the NTTF 2.1 50.54(f) letter in March of 2014 has been updated following the peer review for use in the final SPRA model. The guidance presented in the SPID (Reference 2) was followed for developing the PSHA update. The PSHA update is described in Section 3.1. I of this report.

Topic 2: Site Seismic Response (SPID Section 2.4)

The site response analysis submiued to the NRC in response to the NTTF 2.1 50.54(f) letter in March of 2014 has been updated following the peer review for use in the final SPRA model.

The guidance presented in the SPID (Reference 2) was followed for developing the site response analysis update. The site geotechnical model used for the site response analysis is described in Section 3.1.1.2 while the site response analysis results are described in Section 3.1.1.3 ofthis report.

lESConsulting rlR.zzo

2734294-R-036 Ratision 0 Moy 1L,2017 Page 41.3 of A48 Topic 3: Definition of the Control Point for the SSE-to-GMRS-Comparison Aspeet of the Site Analysis (SPID Section 2.4.21 The PSHA and site response analysis are used to derive Foundation Input Response Spectra (FIRS) at several foundation elevations for critical structures to support the development of fragilities. Section 3.1.1.2 summarizes the elevations for the FIRS. The SPRA does not explicitly derive a ground motion response spectra (GMRS). The GMRS for the site is consistent with the SSE control point is defined in the Updated Final Safety Analysis Report (UFSAR) (Reference29). Section 3.1.2 of this report compares the GMRS submitted to the NRC in response to the NTTF 2.1 50.54(f) letter in March of 2014 with the FIRS for the Reactor Containment Building foundation elevation. The FIRS are derived consistent with NRC Interim Staff Guidance as described in Section 3. I. 1.2 of this report.

Topic 4: Adequacy of the Structural Model (SPID Section 6.3.1)

Entirely new finite element structural models were developed for the SPRA which meet the intents of Criteria 1 through 7 inthe SPID (Reference 2) Section 6.3.1. Details onthe structural models can be found in Section 4.3 of this submiual.

Topic 5: Use of Fixed-Based Dynamic Seismic Analysis of Structures for Sites Previously Defined as "Rock" (SPID Section 6.3.3)

Fixed-base dynamic seismic analysis of structures was not used for the SPRA since BVPS is characterized as a soil site.

Topic 6: Use of Seismic Response Scaling (SPID Section 6.3.2)

Seismic response scaling was not used for the SPRA.

Topic 7: Use of New Response Analysis for Building Response, ISRS, and Fragilities New response analysis is not specifically addressed in the SPID for use in developing In-Structure Response Spectrum (ISRS) and fragilities. The requirements for new analysis are found in the standard under supporting requirements SFR-C2, C4, C5, and C6. The peer review team reported all four of these requirements are either met for Capability Category II or are not applicable for the BVPS-I SPRA. Furthermore the FIRS site response is developed with appropriate huilding specific soil velocity profiles and captures the uncertainty and variability in material dynamic properties as described in Section 3.1.1.2 of this submittal.

Topic 8: Screening by Capacity to Select SSCs for Seismic Fragility Analysis (SPID Section 6.4.3)

The screening approach is documented in Section4,4.1 of this document. The selection of SSCs for seismic fragility analysis used a capacity-based screening approach. This approach meets the recorrmendations in Section 6.4.3 of the SPID (Reference 2). All screened SSCs are retained in the PRA model. Note that analysis assessment PRA-BV2-L7-007-R00 (Reference 92) documents the cumulative impact of all screened SSCs at <SYo and further shows that no screened SSCs are significant based on importance measures.

lESGonsulting fiRtzz.o

2734294-R-036 Ratision 0 May LL,201-7 Page A1-4 of A48 Topic 9: Use of the CIlFIWIIybrid Methodolory for Fragility Analysis (SPD Section 6.4.1)

The CDFM methodology used for fragility analysis is documented in Section 4.4.2.2 of this submittal andmeets the recommendations in section 6.4.1 of the SPID (Reference 2).

Recommended variabilities in Table 6-2 of the SPID were used to develop full seismic fragility curoes.

Topic 10: Capacities of SSCs Sensitive to High-Frequencies (SPD Section 6.4.2)

Contact devices identified in EPRI Phase 2 testing (Reference 90) as being sensitive to high-frequency seismic motion were included in the relay chatter evaluation documented in Section 5.1.3 of this submittal. The flow chart on Figure 6-7 of the SPID (ReferenceZ) can be applied to the high-frequency analysis because all high-frequency susceptible components of interest were identified through circuit analysis and if not screened from the circuit analysis had a high-frequency capacity calculated. The High Frequency Fragility Calculations were performed in accordance with EPRI's High Frequency Program - Application Guidance for Functional Confirmation and Fragility Evaluation (Reference 91). During the high-frequency fragility calculation a capacity versus demand evaluation is performed, and in all cases the capacity was greater than the demand, and therefore no components required replacement.

Topic 11: Capacities of Relays Sensitive to High-Frequencies (SPD Section 6.4.21 The standard is acceptable for the fragility analysis, but additional guidance is presented in the SPID for circuit analysis and operator actions analysis. The BVPS-I SPRA does not credit any specific operator action in response to any seismic-induced relay chatter. Circuit analysis was performed to identifu relays that can potentially impact plant SSCs if chatter were to occur, ffid screen out the relay devices that do not pose a safety concern. The circuit analysis was performed in accordance with the Standard and also meets the SPID (Reference 2) and is documented in Section 5.1.3 of this submittal.

Topic 12: Selection of Dominant Risk Contributors that Require Fragility Analysis Using the Separation of Variahles Methodology (SPID Section 6.4.1)

The SPRA uses the CDFM methodology for the bulk of SSCs requiring seismic fragility analysis.

Separation of Variables was not required. This is supported by the sensitivities presented in Section 5.7.1.6 of this submittal combined with a sufficiently low seismic CDF (SCDF) or 8.78E-06 and seismic large-early release frequency (SLERF) of 2.66E-07. The sensitivities argue that even if the high confidence of a low probability of failure (HCLPF) of the top contributors were improved the reduction in risk is not worth the additional analysis. Furthermore with the low SCDF/SLERF values any potential conservatisms/uncertainties in the CDFM methodology are deemed acceptable.

Topic 13: Evaluation of LERF (SPID Section 6.5.1)

The evaluationLERF is judgedto meet each of the elements of section 6.5.1 of the SPID (Reference 2) including Table 6-3. Section 5.1.2 of this submiual details the evaluation of LERF inthe SPRA. In addition Sensitivity Case 5 in Section 5.7.1 addresses the potential impact of a seismic event extending the evacuation time.

lESComsulting

[iRtzzlo

2734294-R-036 Ratision 0 May LL,20L7 Page A1.5 of A48 Topic 14: Peer Review of the Seismic PRA, Accounting for NEI 12-13 (SPID Section 6.7)

The peer review of the seismic PRA performed meets the elements in Section 6.7 of the SPID (ReferenceZ), An in-process peer review was not performed for the SPRA. Although it is not specifically stated in the peer review report (Reference 6), the lead fragility peer reviewer and one of the two supporting fragility peer reviewer has successfully completed the SQUG training course. Additionally the fragility peer review team lead wrote most of the training course and conducted most of the original classroom lectures.

Topic 15: I)ocumentation of the Seismic PRA (SPID Section 6.8)

This submittal is judged to meet the documentation requirements of section 6.8 of the SPID.

Additionally, all documentation supporting requirements were judged met by the peer review team with the exception of SHA-J3 which is judged to be met by the response to finding 2-30.

Topic 16: Review of Plant Modifications and Licensee Actions, If Any There are no modifications necessary to achieve the appropriate risk profile.

ABSGonsulting tlRtzzo

2734294-R-036 Raision 0 May 11-,2017 Page A1-6 of A48 F&O 2-1 PRA Peer Review Fact & Observation 2-1 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-FI (and other affected Supporting Requirement SHA-F3 ).

DETAILS (Peer Review Team)

As part of the Davis-Besse peer review it was determined that the propagation of the epistemic uncertainty in the ground motion models is not correctly carried out in the estimate of the total seismic hazard at the site. The PSHA report acknowledges this finding and indicates the BVPS PSHA will be updated when an appropriate methodology is implemented in the seismic hazard software.

This issue does not impact the estimate of the mean hazard.

BASIS FOR SIGNIFICANCE (Peer Review Team)

The methodology that is implemented in the P*IZZO Associates (RIZZO) seismic hazard software to propagate the uncertainty in the ground motion models for individual seismic sources to determine the uncertainty in the total seismic hazard at a site does not correctly implement the ground motion logic tree.

POSSIBLE RESOLUTION (Peer Review Team)

The methodology that is used in the RIZZO seismic hazard software to combine the seismic hazard for individual seismic sources to estimate the total seismic hazard (the propagation of epistemic uncertainties in the ground motion model) should be changed to properly implement the ground motion logic tree. The PSHA calculations for the BVPS should be re-run, including the estimate of the rock site hazard results and the incorporation of the uncertainty in the local site response to estimate the FIRS.

The methodology that is used in the F*IZZO seismic hazard software to combine the seismic hazard for individual seismic sources to estimate the total seismic hazard (the propagation of epistemic uncertainties in the ground motion model) should be changed to properly implement the ground motion logic tree. The PSHA calculations for the BVNS should be re-run, including the estimate of the rock site hazard results and the incorporation of the uncertainty in the local site response to estimate the FIRS.

PLANT RESPONSE OR RESOLUTION (ABS Consulting, RIZZO Associates, and FENOC)

The method for combining seismic hazard curves from individual sources is revised such that when combining hazard curves for one seismic source (consistent with CEUS-SSC logic tree structure) each ground motion prediction equation (GMPE) is considered separately (consistent with the EPRI-GMM logic tree). Accordingly, the post-processing scripts that implement the combination method are revised 1) to retain intermediate seismic hazard results (for each source and for each GMPE), and 2) to combine the full set of seismic hazard curves to correctly derive the total mean and fractiles. Documentation of the revised combination method is provided in more detail below.

lBSGonsulting

(]Rlzza^

2734294-R-036 Ratision 0 May 11.,201.7 Page A17 af AaB Enhancement of the method to propagate uncertainty in local site response is described in the Disposition to F&O 2-2.

The revised control point hazard (reactor building tRB] foundation level) due to the above revision in the hazard combination method, and incorporating enhancements to better propagate uncertainty in site response to address F&O 2-2, exhibits insignificant changes to the mean hazard and the mean uniform hazard response spectra used to determine the FIRS, while the low and high fractiles show small differences. Therefore, as discussed further below, the fragility analyses of plant SSCs, which are based onthe reported FIRS, are unafflected.

Note that RIZZO-HAZARD software that calculates the hazard for each branch of the PSHA logic tree is fully verified and validated and produces correct results. The issue identified in this F&O is related to post-processing scripts that combine outputs from RIZZO-HAZARD, and not with the basic hazard computation.

Revision of Method for Ha4ar4 Corubination F.IZZO Calculation No. l2-4735-F-120, Revision 2, develops the seismic hazard for hard-rock conditions. [t describes the post-processing scripts that incorporate the GMPE correlation model, and provides details of the methodology implemented to derive the hard-rock total seismic hazard curves as follows:

Section 5.2.3: Describes the GMPE correlation model used to combine hazard curves from RLME and distributed seismicity sources.

Section 5.4: Describes the RIZZO-HAZARD hazard curve data files per source zones (RLME and distributed seismicity sources), GMPE, ffid magnitude-range weighting cases used in the recurrence relationship (Cases A, B and E).

r Section 5.9.5: Describes the combination of the hazard curves from Section 5.4 to obtain total rock hazard curves. The scripts described in this section perform the following steps:

Uploading the hazard curves per GMPE and the three magnitude-range weighting cases used in the recurrence relationship for the distributed seismicity sources, and only by GMPE in the case of RLME source zones (Files described in Section 5.4).

Combining hazard curves from source zones (Section 5.4) considering correlations among the magnitude-range weighting in the recurrence relationships and among GMPEs when two distributed seismicity sources are combined, and the GMPE correlation model described in Section 5.2.3 when an RLME and distributed seismicrty source are combined.

RIZZO Calculation I2-4735-F-120, Revision 2 and Calculation l2-4735-F-I21, Revision 2 document the resulting mean hazard and the hazard fractile curves for hard rock at the BVPS Site implementing the above revisions; and Section 4.3 of ABS/RIZZO Report 2734294-R-003, Revision 4 (updated PSHA Report) summarizes the revised hard rock hazard.

lE$Gonsulting tlR,zzo

2734294-R-036 Ratision 0 May 1L,2017 41.8 448 Revision of Methoj fo.r-Prqpaeation of Uncertainty in Local Site Response The "...incorporation of the uncertainty in the local site response to estimate the FIRS." inthe Peer Review Suggested Resolution for Finding F&O 2-1 is addressed in the response to Finding F&O 2-2.

A$$essment of Effect of Revised GMRS and FIRS on Fragilitv-Analvs,ep The FIRS reported in Section 6.4 of ABS/RIZZO Report 2734294-R-0A3, Revision 4 showminor differences as compared to the FIRS reported previously in ABS/RIZZO 2734294-R-003, Revision I. However, the differences in the spectral shapes are insignificant. Based on a comparison of the spectral shapes of the FIRS the impacts on the fragilities reported in ABS/RIZZO Report 2734294-R-A06, Revision 0, are also insignificant. Therefore, the ground motion time histories, the building analysis, ffid the fragility analysis remain unaffected. This is further discussed and justified in the Section 5.5 of the revised Fragility Analysis Reports (ABS/RIZZO Report 2734294-R-006, Revision I andABS/RIZZO Report 2734294-R-013, Revision I).

IESGonsulting

$Rtzzo

2734294-R-036 Rutision 0 May 1,1,20L7 Page A1-9 of A48 F&O 2-2 PRA Peer Review Fact & Observation 2-2 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-F I (and other affected Supporting Requirement SHA-J I ).

DETAILS (Peer Review Team)

To estimate the seismic hazard at the top of the soil column (e.9., at the RB base elevation) the aleatory and epistemic uncertainties in the rock PSHA results and the site amplification factors are not combined to estimate the total epistemic uncertainty in the soil harard, This issue does not impact the estimate of the mean hazard.

BASIS FOR SIGNIFICANCE (Peer Review Team)

To estimate the seismic hazard at the top of the soil column, the rock PSHA results are combined with the probabilistic characterization of the site amplification factors. The site amplification is represented by the mean and standard deviation for the total uncertainty (combined aleatory and epistemic uncertainty) and the assumption that the amplification factors are lognormally distributed. Thus, the epistemic uncertainty in the rock site haeard is probabilistically combined with the site amplification aleatory and epistemic uncertainty. As a result, the epistemic uncertainty in the site amplification is not combined with the rock hazard uncertainty to estimate the uncertainty in the soil hazard, leading to the rxrcertainty in the soil hazard being underestimated.

Since the aleatory and epistemic uncertainty in the site amplification are considered, the estimate of the mean soil hazard should not be efflected.

The approach that is used under-estimates the epistemic uncertainty in the soil hazard and is therefore unconservative. As a result the uncertainty in the seismic risk (CDF and LERI') will be underestimated.

As currently implemented the process for generating the input to the SPRA quantification (a series of 100 hazard curves) also does not combine the rock site hazard and the site amplification uncertainties.

POSSIBLE RESOLUTION (Peer Review Team)

As part of the site response analysis, maintain the segregation of aleatory and epistemic uncertainties and propagate these properly when combined with the rock hazard results to estimate the seismic hazard and the top of the soil column.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and RIZZO Associates)

RIZZO Calculation l2-4735-F-117 is revised (Revision 2)to appropriately segregate the aleatory and epistemic uncertainties in site response such that they can be properly propagated when combining the site response with the rock hazard results to obtain control point (i.e., "soil")

hazard. RIZZO Calculation 12-4735-F-l18, Revision 2 (Reactor Building foundation),

Calculation 12-4735-F-123, Revision I (AUX, DGB, FDB, MSVCV, SFGB, SRV, and CB foundation), and Calculation l2-4735-F-125, Revision I (Intake Structure foundation) illustrate that the revised treatment of the uncertainties in the site response analysis, along with other lESGonsulting

(}Rtzza

2734294-R-036 Rwision 0 Moy L1.,20L7 Page A20 of AaB changes to address F&O's, result in insignificant changes in mean horizontal control point hazafi atthe top of the soil column and corresponding UHRS usedto develop FIRS, while the low and high fractiles show small differences. As discussed firther below, the fragility analyses of plant SSCs, which are based on the reported FIRS, are unaffected.

Revised Treatment gf Aleatorv and Euistemic Uncertaintv in Site Responqq Analvsis The logic tree of input parameters for site response analysis, shown on Figure 5-1 (Section of 5.2 of ABS/RIZZO Report 2734294-R-003, Revision 4), has 20 branches accounting for various combinations of input parameters reflecting epistemic uncertainties in the site response analysis.

The aleatory variability is represented by 30 combinations of randomized Vs profiles (from hard rock to the control point elevation at the top of the soil column)n irnd corresponding randomized G/Gmax and damping curves. The end branches of the logic tree reflect epistemic uncertainty in the various site response inputs and take into account guidance on characterizing uncertainty provided in Screening, Prioritization, and Implementation Details (SPID) for the Resolutton of Fukushima Near-Term Task Force Recommendation 2.1 : Seismic (EPRI, 2013b). The calculation results in the mean and standard deviation of the site amplification functions (SAF) for each branch of the logic tree for each of 1l hard-rock hazard levels.

In RIZZO Calculation l2-4735-F-117, Revision I, which is summarized in ABS/RIZZO Report 2734294-R-003, Reviston 1, the approach described in EPRI (2013b, Section B-6) was followed to develop probabilistic hazard curves. Site amplification functions were determined for each combination of response frequency and hard-rock ground motion amplitude weighted sums over the 20 site response models. This effectively transfers the epistemic uncertainty in site response into aleatory uncertainty.

In RIZZO Calculation I2-4735-F-l17, Revision 2, which is summarized in ABS/RIZZO Report 2734294-R-003, Revtston 4, andrelated RIZZO Calculations l2-4735-F-122, Revision 2 and l2-4735-F-124, Revision 2,the site response results are surlmarized to maintain the general characteristics of site amplification uncertainty related to epistemic uncertainty in site response analysis inputs. Epistemic uncertainty in site response analysis inputs that does not translate into significant epistemic uncertainty in SAFs is averaged (i.e., transferred to aleatory uncertainty). Epistemic uncertainty in site response analysis that leads to relatively significant uncertainty in SAFs is retained and carried into the control point (soil) hazard calculation.

More specifically, the control point (RB foundation level) hazwd is obtained by the convolution of hard-rock hazard with the SAF, as described in Section 6.1 of the ABS/RIZZO Report 2734294-R-003, Revision 4. Although in principle this process is able to segregate and propagate the aleatory and epistemic uncertainty in the site response, the previous analysis (ABS/RIZZO Report 2734294-R-003, Revision /) treats epistemic uncertainty as aleatory variability, consistent with the SPID guidance (EPRI Technical Report #1025257,2013b).

However, we concurwiththis F&O thatthe propagation of epistemic uncertainty in site response into the PSHA more accurately determines the control point (soil) hazard fractiles. In response to the F&O, RIZZO Calculation I2-4735-F-117, Revision 2 describes the method used to properly segregate and propagate the aleatory and epistemic uncertainties in the convolution of the SAFs with the hard-rock hazard.

AESGonsutting

[iRtzz.o

2734294-R-036 Reuision 0 May 1-1.,201-7 Page A21 of A48 Because it is computationally prohibitive to incorporate the full set of epistemic simulations (20 branches x 36 spectral frequencies x I I HR hazard levels) into the hazard analysis, a simplified approach is utilized. This approach examines the SAFs at each end branch of the site response logic tree for all levels of input motions, ffid bins them into three groups of epistemic branches based on which inputs dominate the epistemic uncertainty in site response and on the similarity of the SAF. Section 5.10 of the ABS/RIZZO Report 2734294-R-003, Revision 4 describes the grouping and develops representative SAFs for each group. The respective group SAFs are used to convolve with the hard-rock hazard and propagate the epistemic uncertainties in developing the control point hazard.

Calculation l2-4735-F-117, Revision 2 describes the basis for the SAF grouping (three groups),

and presents Tables and Figures displaying the SAF for each group and at each of the seven spectral frequencies. This calculation is also expanded to document additional details on the derivation of the inputs used in the site response analysis. Much of this material was previously included in Section 5.0 of ABS/RIZZO Report 2734294-R-003, Revision l. Further, Calculation I2-4735-F-I 18, Revision 2 provides details about how the GMPE correlation model and the epistemic uncertainty in SAF are incorporated in the process, as follows:

Section 5.1 describes how the three groups of SAF are applied to the hard-rock hazard curve for each branch of the logic tree to obtain a new population of hazard curves at the RB foundation elevation.

Section 5.2 describes how the scripts from Calculation F-120 (hard-rock hazard curves) are modifiedto apply one of the three SAF groups andperform the full combination of the hazard curves considering the CEUS-SSC and EPRI-GMM model logic trees. The modification to the script saves the hazard curves at the RB foundation calculated with each of the three SAF groups. Section 5.2 also describes how the three sets of hazard curves at the RB foundation obtained from the three SAF groups are combined to obtain the total RB foundation hazard curves.

r Calculation 12-4735-F-143, Revision 2 describes how the control point (soil) hazard distribution for the RB foundation, which is determined by appropriately segregating epistemic uncertainty and aleatory variability in site response analysis and then propagating them properly when combining them with rock hazard results, is used to provide the 100 hazard curves used as input to SPRA quantification.

Note that, other than the guidance in the SPID, no other guidance is available on how site response epistemic uncertainty should be assessed as part of deriving seismic hazard curves, particularly hazard curve fractiles, while maintaining reasonable computational efforts. Given that site response epistemic uncertainty essentially impacts each GMPE used in the hazard computation, the grouping approach focuses on the critical site response epistemic uncertainty while maintaining computational viability in developing accurate mean hazard curves at the elevations where FIRS are needed for fragility calculation, and hazard fractiles at the RB foundation elevation to which the fragilities are referenced.

AESGonsulting rlRtzzo

2734294-R-036 Rmision 0 Moy 1,L,201"7 Page A22 of A48 Assessment of Effect of Rqyise.d,_GMRS and FIRS on Fragilitv Analv$e$

The FIRS reported in Section 6.4 of ABS/NZZO Report 27i4294-R-0A3, Revision 4 show minor differences as compared to the FIRS reported previously in ABS/RIZZO 2734294-R-003, Revtsion I. However, the differences in the spectral shapes are insignificant. Based on a comparison of the spectral shapes of the FIRS the impacts onthe fragilities reported in ABS/RIZZO Report 2734294-R-006, Revision 0, are also insignificant. Therefore, the ground motion time histories, the building analysis, ffid the fragility analysis remain unafflected. This is further discussed andjustified inthe Section 5.5 of the revised Fragility Analysis Reports (ABS/RIZZO Report 2734294-R-006, Revision 1 for BVPS Unit I and ABS/RIZZO Report 2734294-R-013, Revision 1"fo, BVPS Llnit 2).

lESGsrsulting (IRt74zlo

2734294-R-036 Rwision 0 May 1.1,201,7 Page 423 of AaB F&O 2-3 PRA Peer Review Fact & Observation 2-3 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-F2.

IIETAILS (Peer Review Team)

Sensitivity studies and intermediate results have not been systematically carried out and reported in the PSHA documentation. While some deaggregation results are reported (which can be interpreted as intermediate and sensitivity calculations), a systematic demonstration of sensitivity of the results to key parameters is not presented.

BASIS FOR SIGNIFICANCE (Peer Review Team)

The PSHA report does not present a comprehensive assessment of the sensitivity of the seismic hazardresults to the different elements of the analysisi e.9., seismic source model uncertainty, ground motion model uncertainty, etc.

POSSIBLE RESOLUTION (Peer Review Team)

Perform and present sensitivity calculations that demonstrate the sensitivity of the hazardresults to elements of the PSHA; ground motion attenuation models; estimates of site amplification; alternative soil profiles, estimates of kappa, etc. The sensitivity of the hazard to different factors in the PSHA could be demonstrated by adding "tornado plots" at different ground motion levels to the various branches in the logic tree. These plots show which sources of epistemic uncertainty are most important. It should include the source model uncertainty, ground motion model uncertainty, and site response uncertainty. Currently, the total uncertainty is shown by the hazard fractiles, but it is not broken down to provide understanding as to what is most important.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and HIZZ,O Associates)

The response to this F&O improves the documentation andpresentation of some ofthe intermediate hazard results and provides additional sensitivity calculations to provide insight into what inputs more strongly contribute to the overall distribution of hazard results. It does not change the hazard or the seismic demand on which the fragilities are based.

RIZZO Calculqtion F-144, Revision 1 develops the total variance deaggregation for 100 Hz surface hazard for all the logic tree branches and for different ground motion levels represented by mean annual frequency of exceedances (MAFE). The total hazard variance is deaggregated in terms of the following PSHA elements:

Seismic source model uncertainty Alternative source model approach Mmax Recurrence rates Magnitude weighting case used to determine recurrence rates Thickness of the seismogenic layer lESGonsulting tiRtzzo

273429+R-036 Rwision 0 May 1L,201,7 Page AZa of A48 r Ground motion model uncertainty Site response uncertainty Alternative SAF groupings The deaggregated variance is a measure of relative contribution of epistemic uncertainty in each element to the total variance. These relative contributions are response frequency and annual frequency of exceedance (AFE) dependent.

Additionally, RIZZO Calculation F-l17, Revision 2 develops median and standard deviations of SAFs for the 20 epistemic branches of the site response inputs logic tree. The logic tree represents the assessed uncertainty in geologic profile, seismic source spectra model, profile damping, and site kappa. These intermediate results are documented in Section 5.8.8 of ABS/NZZO Report 2734294-R-003, Revision 4 (the updated PSHA Report).

RIZZO Calculation F-144 Revision / shows that the dominant contributor to the total variance is the epistemic uncertainty in the ground motion model; i.e., GMPEs. As the MAFE gets lower, the epistemic uncertainty in maximum magnitude, the three magnitude-range cases used for deriving recurrence rate, and the eight recrurence rate realizations become more significant.

Similarly, RIZZO Calculation F-I17, Revision 2 shows that the most significant factor impacting the SAFs is the uncertainty in geologic profile definition.

The above sensitivity studies were performed for additional insight of the epistemic uncertainty only, and do not affect or change any inputs to the PRA model.

Section 5.9 of ABS/RIZZO Report 2734294-R-003, Revision 4 documents the contribution of different sources of uncertainties modeled in the PSHA. It describes the wide range of sensitivity calculations and also presents an assessment of the variance contribution to the hazard for all PSHA inputs (seismic source, ground motion, and site response). The variance assessment is accomplished for a wide range of ground motion levels represented by the annual frequencies of exceedance. Figure 5-37 displays the variance contribution for each PSHA input.

This is effectively similar to "tornado plots," and provides an understanding of which PSHA inputs are more significant from an epistemic uncertainty perspective.

AESGonsulting (iRtzzo

2734294-R-036 Ratision 0 May 1.L, 201"7 Page A25 of A48 F&O 2-26 PRA Peer Review Fact & Observation 2-26 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-II (and other affected Supporting Requirement SHA-I2).

DETAILS (Peer Review Team)

A screening assessment of other seismic hazards was performed. There are a number of technical questions associated with elements of the analysis for some of the other seismic hazards.

BASIS FOR SIGNIFICANCE (Peer Review Team)

The NRC has identified two dams that are upstream of the BVPS that may pose a flood hazard.

In fact there are multiple dams upstream of the plant.

The PSHA report does not address the potential for seismically-initiated dam failure that could impact the dams. A large seismic event in the region could potentially simultaneously cause high ground motions at the BVPS and at the upstream dams leading to dam failure and damage to the BVPS.

POSSIBLE RESOLUTION (Peer Review Team)

The potential seismically-initiated failure of upstream dams and their flooding consequence should be addressed as part of the seismic screening analysis.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and RIZZO Associates)

Section 7.3.5 ("Seismically Induced Dam Failures") of has been added to ABS/RIZZO Report 2734294-R-003, Revision 4 to include an assessment for the potential seismically-induced failure of upstream dams and their flooding consequences. The analysis reported in BVPS-2 UFSAR (Appendix2.4A) concludes that the failure of the upstream Conemaugh Dam, which is the most critical with respect to flooding, raises the flood stage to EL 725.2 ft. This is less than design basis flood level of EL 730.0 ft. Therefore, this seismic-related hazard is screened out from further analysis.

lESGonsulting

(]Rtzzo

2734294-R-036 Rwision 0 May L1,2017 Page 426 of A48 F&O 2-27 PRA Peer Review Fact & Observation 2-27 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-II (and other affected Supporting Requirement SHA-I2).

DETAILS (Peer Review Team)

A screening assessment of other seismic hazards was performed. There are a number of technical questions associated with elements of the analysis for some of the other seismic hazards.

BASIS FOR SIGNIF'ICANCE (Peer Review Team)

It is argued that the consequence of slope failure that is based on the minimum FOS slip surfaces is negligible because they do not intersect critical structures. However, analyses are not presented of slip surfaces that would have safety consequences to plant structures in order to show margins against these slope failures.

Impact of failure of slope in Cross Section 2-Z on the Intake Structure itself has not been clearly assessed. It is stated on Page 410 3rd paragraph: "In the event of a failure in Section}-2,the material of the lower slope is expected to displace less than one-half of a foot. The upper slope in Section}-Z is expected to be retained by the retaining structure. These displacements are relatively small and do not affect the function of the Intake Structure." It is not clear that this has been clearly analyzed in the context that a HCLPF for displacements has been analyzed.

A generic procedure has been used to estimate the HCLPF for soil structures. It is not clear that the generic procedure that includes (at least implicitly) estimates of aleatory and epistemic uncertainty in soil properties, stability analyses, etc. is an appropriate basis to estimate the HCLPF and serve as a basis for screening.

POSSIBLE RESOLUTION (Peer Review Team)

The analysis should evaluate potential slope failure modes that would impact critical structures and components.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and RJIZZ,O Associates)

Section 7.3.3 of ABS/RIZZO Report 2734294-R-003, Revision 4 evaluates three permanent slopes whose failure could affect safety-related functions, including:

r Slope north of the Unit 1 (Figure 2.6-3 of BVPS-1 UFSAR) t Riverward slopes involving Service Water Piping (Figure 2,5,4-57 of BVPS-2 UFSAR)

Intake Channel Slopes (Figures 2.5.4-37 and 6l of BVPS-2 UFSAR)

As reported in Section 7.3.3, the slope stability analyses for the above permanent slopes are performed using Version 7.23 of the SLOPE/W Stability Analysis Program (Geo-Slope, 2007; RIZZO,2012b). The HCLPFs are obtained using site-specific geotechnical characteristics obtained from the FSAR. As described below, the HCLPFs for slope failures are smaller than 0.5g. However, the slope failure is screened on the basis of the consequence to the affected lESGonsulting (lRtzza

273429+R-036 Reoision 0 May 1.1,2017 Page A27 of A48 SSCs. The consequences are assessed based on the expected post-failure displacements, which are significantly smaller than the distance to the affected structures.

The slope north of Unit t has a HCLPF value of 0.5g PGA. It is noted, however, that this failure mode does not affect the Turbine Building (TRBB) because the failure circle is expected to daylight about 150 ftfrom the Turbine Building foundation. The HCLPF value ofthe analyzed failure circle is takento be a conservative lower bound affecting the TB. This is in excess of the assumed HCLPF of the TB structure. Potential failure surfaces involvingthe TB footprintwould be characterized by larger margins and are not controlling failure modes assosiated with slope failure affecting the TB.

The minimum slope stability factor of safety for the Riverward Slope is 1.54. The corresponding HCLPF value is 0.339 PGA. In the event of a slope stability failure, a maximum displacement of I inch is predicted. Based on the acceleration required to cause 1 inch of displacement, the HCLPF capacity associated with slope displacement is 0.389. This analysis also shows that the critical slip surface outcrops approximately 150 ft from the Intake Structure. Therefore, possible displacements due to the slope failure caused by an earthquake are not expected to affect the structural integrity of the Intake Structure. Shallower failure surfaces extending to the Intake Structure are expected to have larger factors of safety than the critical slip surface, and therefore do not represent controlling failure modes for slope failure.

The factors of safety for the upper and lower slopes at the intake are calculated to be 1.66 and 1.43., and the corresponding HCLPF values for slope failure are 0.369 and 0.3 1g. In the event of slope failure, the upper slope is expected to be retained by the retaining stucture. The unrestrained displacements of the lower slope are less than one foot. Therefore, it will not affect the function of the Intake Structure, which is more than 90 ft from the toe of the slope.

The analyses presented conclude that potential failure of the intake slopes and the resulting displacement profiles do not affect the structural integrity of the structures or the function of the Intake Channel.

lESGonsulting (iRtzz.o

2734294-R-036 Rwision 0 May 1-1.,201.7 Page 428 of Aal F&O 2-28 PRA Peer Review Fact & Observation 2-28 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-I I (and other affected Supporting Requirement SHA-I2).

DETAILS (Peer Review Team)

A screening assessment of other seismic hazards was performed. There are a number of technical questions associated with elements of the analysis for some of the other seismic hazards.

BASIS FOR SIGNIFICANCE (Peer Review Team)

Text in the PSHA report at the bottom of Page 405 indicates a minimum HCLPF for Unit 2 bearing capacity of 0.459. The minimum value in Table 7-1 for Unit 2 appears to be 0.509.

POSSIBLE RESOLUTION (Peer Review Team)

Modify the text to be consistent with the analysis results.

PLAFIT RESPONSE OR RBSOLUTION (ABS Consulting and RIZZO Associates)

Section 7.3.2 of ABS/RIZZO Report 2734294-R-003, Revision 4 has been revised to be consistent with the minimum HCLPF presented in Table 7-l. This is a documentation change and does not affect PRA inputs.

IESGonsulting tiR.zzo

2734294-R-036 Raision 0 May 11,201.7 Page 429 of AaB F'&O 2-29 PRA Peer Review Fact & Observation 2-29 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-II (and other affected Supporting Requirement SHA-I2).

DETAILS (Peer Review Team)

A screening assessment of other seismic hazards was performed. There are a number of technical questions associated with elements of the analysis for some of the other seismic hazards.

BASIS FOR SIGNIF'ICANCE (Peer Review Team)

There is no indication that lateral spreading of the ground in the vicinity of the Intake Strusture or other critical structures has been assessed.

A generic procedure has been used to estimate the HCLPF for soil structures. It is not clear that the generic procedure includes (at least implicitly) estimates of aleatory and epistemic uncertainty in soil properties, stability analyses, etc. is an appropriate basis to estimate the HCLPF and serve as a basis for screening.

POSSIBLE RESOLUTION (Peer Review Team)

The analysis should evaluate the potential for liquefaction and lateral spreading that could impact critical structures and components.

PLANT RESPONSE OR RBSOLUTION (ABS Consulting and RIZZO Associates)

As described in Section 7.3.4 of ABS/RIZZO Report 2734294-R-003, Revision 4,the foundations for all power block structures are supported on either in-situ competent soil in the higher terrace or on engineered structural backfill. NUREG/CR-5741 concludes that the liquefaction susceptibility of terrace soils fromthe Pleistocene period is'very low'. Additionally, the liquefaction potential is also overy low' when depth of the groundwater is greater than about 50 ft (NUREG/CR-5741; NRC, 2000). All of the power block structures satisff both conditions, and are therefore not affected by liquefaction, and this failure mode is screened out for the power block SSCs.

Section 7,3.4 presents the detailed liquefaction analysis of the yard areabetween the plant and the intake. The reported liquefaction analysis is based on conservative design parameters in the FSAR such as recorded SPT blow counts, the particle size distribution and fines content, and the water table elevation. These are taken to be the 84th percentile values. The calculated HCLPFs for liquefaction and its effects on affected SSCs (buried pipes) thus represent CDFMs.

Based on the calculated settlements due to liquefaction, and assuming an allowable seismic-induced settlement associated with the buried lines of 3 inches, the HCLPF value associated with seismic-induced settlement is 0.399. Allowing for a nominal ductility (Fp:2.0), the HCLPF associated with structural integrity of the buriedpipes is about 0.8g. This is significantly in excess ofthe CDFM HCLPF values of equipment inthe Intake Structure.

Therefore, the liquefaction failure mode affecting the plant SSCs is screened out. Additionally, due to the generally flat topography lateral spreading is not an issue.

lEEConsulting

{}Rrzzo

2734294-R-036 Rwision 0 May 11,2017 Page A30 of Aa$

F&O 2-30 PRA Peer Review Fact & Observation 2-30 was identified in the Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-J3.

DETAILS (Peer Review Team)

A foundational element of PSHA as it has evolved over the past 30 years is the development and implementation of methods to identiff, evaluate, and model sources of epistemic (model and parametric) uncertainty in the estimate of ground motion haards. As such fairly rigorous analyses are carried out (SSHAC studies) to quantitatively address model uncertainties.

At the same time there is within any analysis sources of uncertainty that are not directly modeled and assumptions that are made for pragmatic or other reasons. There are also sources of model uncertainty that are embedded in the context of current practice that are 'accepted' and typically not subject to critical review. For instance, in the PSHA it is standard practice to assume that the temporal occurrence of earthquakes is defined by a Poisson process. This assumption is well accepted despite the fact that it violates certain fundamentally understanding of tectonic processes (shain accumulation). A second practice is the fact that earthquake aftershocks are not modeled in the PSHA, even though they may be significant events (depending on the size of the main event).

In the spirit of the standard it seems appropriate that sources of model uncertainty that are modeled as well as sources of uncertainty and associated assumptions as they relate to the site-specific analysis should be identified/discussed and their influence on the results discussed.

As SPRA reviews and the use of the standard has evolved, it would seem the former interpretation is reasonable, but potentially incomplete. It is reasonable from the perspective that documentation of the sources of model uncertainty and their contribution to the site-specific hazard results is a valuable product that supports the peer review process and assessments in the future as new information becomes available). Similarly, documenting assumptions provides similar support for peer reviews and future 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:

l.

The documentation should present quantitative results and discussion the sources of epistemic uncertainty that are modeled and their contribution to the total uncertainty in the seismic hazard.

2.

The documentation should discuss elements of the PSHA model where these may be latent sources of model uncertainty that are not modeled and assumptions that are made in performing the analysis.

BASIS FOR SIGNIFICAF{CE (Peer Review Team)

The documentation of the sources of model uncertainty analysis and a description of the analysis assumptions is not complete in the PSHA report in its current form such that a clear lESGonsulting tlRtzzo

2734J,94-R-036 Reaision 0 May 1.1,2017 Page A3L of AaB understanding of the contribution of individual sources of uncertainty to the estimate of hazard are understood. Limited information on the contribution of seismic sources to the total mean hazard is presented, but information on the contributors to the uncertainty is not provided.

With respect to addressing model uncertainties and associated assumptions there are some examples that can be identified in the Beaver Valley (BV) PSHA. These are:

1.

In the site response analysis the assumption is made that the lD equivalent-linear model (SHAKE type) to estimate the site amplification and ground motion input to plant structures is appropriate. In addition, an assumption is made that the variation in the rock topography does not significantly influence the ground motion that is input to the plant.

This modeling approach and the potential model uncertainty that it represents relative to the conditions at the BV site should be addressed.

2.

In the estimate of vertical ground motions, an envelope of alternative V/H ratio models was used. This approach is conservative. It is implicitly assumed this approach is reasonable and appropriate as a basis to provide input to the seismic fragility analysis.

This assumption and its potential implications is a topic that should be identified and discussed in the context of addressing this requirement.

POSSIBLE RESOLUTION (Peer Review Team)

The resolution to this finding could involve:

1.

Documentation and discussion of the contribution of different sources of uncertainty that are modeled in the PSHA. The documentation of the confribution of different sources of uncertainty can be shown by means of "tornado plots" that quantiff the sensitivity of the hazard at different ground motion levels to the various branches in the logic tree. These plots show which sources of epistemic uncertainty are most important. It should include the source model uncertainty, ground motion model uncertainty, ffid site response uncertainty. Currently, the total uncertainty is shown by the hazard fractiles, but it is not broken down to provide understanding as to what is most important.

2.

Identification and discussion of model assumptions that are made.

PLANT RESPONSB OR RESOLUTION (ABS Consulting and HIZ,ZO Associates)

This F&O relates to the documentation of the sensitivity analyses addressed in response to F&O 2-3 and documentation of model assumptions. It does not affect the hazard definition or the UHRS.

As stated inthe Disposition of F&O 2-3, Section 5.9 of ABS/RIZZO Report 2734294-R-003, Revision 4 documents the contribution to hazard of different sources of uncertainties modeled in the PSHA. Additionally, Section 5.8.8 presents details of the sensitivity of the site amplification factors to various inputs to the site response analysis such as geologic profile, ground motion amplitude, seismic source spectra, profile damping assumptionso and site kappa.

Section 5.9 concludes that the dominant contributor to the total hazard variance is the epistemic uncertainty in GMPEs. As the MAFE gets lower, the epistemic uncertainty in maximum magnitude, the three magnitude-range cases used for deriving recurrence rates and the eight lESGonsulting

()Rtzz,o

2734294-R-036 Rwision 0 May 11,2017 Page A32 of A48 recturence rate realizations become more significant. Similarly, Section 5.8.8 concludes that the most significant factor impacting the SAFs is the uncertainty in geologic profile definition.

The modeling assumptions for various elements of the PSHA are described in Section 2.0 for the seismic source models and in Section 3.0 forthe ground motionmodels, and inthe references cited therein. The modeling assumptions for the site response analysis are described in Section 5.0 and cited references.

Assumptions used are those associated with current standards of practice. Examples are as follows:

. Ergodic assumption as applied to the estimation of ma:rimum earthquake magnitude for distributed seismicity sources and to ground motion prediction.

r Seismic source characterization model The spatial distribution of seismicity is generally temporally stationary.

The occurrence of independent earthquakes is a stationary Poisson process.

The size distribution of earthquake magnitudes for distributed seismicity sources follows an doubly truncated exponential distribution.

Ground motion characterization Variability in ground motion follows a lognormal distribution.

r Site response analysis Use of equivalent-linear analysis and vertically propagating shear waves adequately represents the important trends in site response for the levels of ground motion considered.

A site geotechnical model consisting of homogeneous, horizontal layers adequately represents the site conditions.

Conditions at the Beaver Valley sites are consistent with the standard practice use of the above assumptions.

lESGonsulting

[]Rrzzo

2734294-R-0s6 Ranision 0 May 11, 20L7 Page A33 of A48 F&O 2-31 PRA Peer Review Fact & Observation 2-31was identified in ttre Probabilistic Seismic Hazards Analysis High Level Requirement, Supporting Requirement SHA-II (and other affected Supporting Requirements SHA-I2, SFR-D I ).

DETAILS (Peer Review Team)

A screening assessment of other seismic hazards was performed. There are a several technical questions associated with elements of the analysis for some of the other seismic haeards.

BASIS FOR SIGNIFICANCE (Peer Review Team)

An analysis was performed to assess potential bearing capacity failures.

Calculation 12-4736-F-033 Rl presents the methodology for calculating the bearing capacity; however it does not discuss how the HCLPF is estimated. As such it is not clear if the HCLPF estimates, which are the basis for screening bearing capacity failures are appropriate.

Discussions with the analyst involved in the analysis suggests that the median capacity for a bearing failure may not be significantly higher than the estimated median capacity. If this is the case, additional support for screening out this failure mode is required.

POSSIBLE RESOLUTION (Peer Review Team)

Provide documentation of the methodology for estimating the bearing capacity HCLPF. If the median seismic capacity is not significantly higher than the estimated HCLPF, then additional basis for screening out this failure mode should be provided PLANT RESPONSE OR RESOLUTION (ABS Consulting and RIZZO Associates)

ABS/RIZZO Report 2734294-R-003, Revision 4 includes revisions to enhance the discussion of bearing capacrty HCLPF values. It provides additional basis to screen out bearing capacity failure. Based on available margins assuming linear behavior, the HCLPF is sufficiently large to accommodate the possibility that, due to inherent nonlinearities, the median capacity is not significantly larger than the HCLPF.

Section 7.3.2 of ABS/RIZZO Report 2734294-R-003, Revision 4 documents the methodology for estimating the HCLPF associated with bearing capacity failure. The factors of safety reported in the FSAR indicate relatively significant margins against bearing capacity failure under SSE conditions. To account for potential uplift at higher ground motion levels, a bounding analysis is performed. This analysis conservatively ignores that uplift reduces the demand overturning moment. On the other hand, it accounts for the fact that uplift reduces the effective bearing area and therefore increases the bearing pressure and reduces the effective bearing capacity.

Table 7-f presents the resulting conservative bounds for the HCLPF values. The minimum bounding HCLPF for the BVPS Unit 1 and Unit 2 structures is 0.539 and 0.5g, respectively.

It is noted that uplift of the foundation mat due to seismic ground motion significantly reduces overturning moments and in turn the bearing pressure. These reductions in demand, along with (1) the calculated bounding HCLPFs in Table 7-1 and (2) the significant margins under SSE conditions, are used as basis to screen out bearing capacity as the controlling failure mode for the BVPS structures.

lEtGonsulting

[]Rtzz.o

2734294-R-036 Ratision 0 Moy 1,L,2017 Page A34 of A48 F&O 4-6 PRA Peer Review Fact & Observation 4-6 was identified in the Seismic Fragility Analysis High Level Requirement, Supporting Requirement SFR-A2 (and other affected Supporting Requirement SFR-F2).

IIETAILS (Peer Review Team)

Excess conservatism and unrealistic assumptions were noted in a number of calculations providing the fragility parameters for components identified as top contributors to CDF.

BASIS FOR SIGNIFICANCE (Peer Review Team)

(sequential letters added by FENOC "for clartty in Plant Response or Resolution section) a)

BVl residual heat removal (RHR) pumps are evaluated in2734294-C-106, Revision 0 BVPS1 Seismic Fragility for Vertical Pumps, Section 5.4. EW and NS seismic accelerations are enveloped. Three percent damping is used but response is dominated by the steel support frame. CDFM capacity is scaled from a conservative design calculation.

The design calculation includes operational considerations and seismic nozzle loads, but it is not checked if these loads are realistic for fragility evaluation purposes. No inelastic energy absorption factor is used. Weld capacity is governed by base metal and this is not a realistic failure modes per American Institute of Steel Construction (AISC).

b)

BVI Pressurizer power operated relief valve 2RCS-PCV455C is evaluated in 2734294-C-208, Revision 0 BVPS2 Seismic Fragility for Motor _ Solenoid Operated Valves, Section 5.2. A conservative lower bound natural frequency estimate is used in the evaluation, and conservative generic capacity is assigned. A value lower than the calculated HCLPF capacity was used in the quantification.

c)

BV I Pressurizer relief valve 2RCS-RV5 5 1A is evaluated in calculati on 2734294-C-207,

Revision 0 BVPS2 Seismic Fragility for Pneumatic Operated Valves, Section 5.27. A conservative lower bound natural frequency estimate is used in the evaluation, and conservative generic capacity is assigned. A value lower than the calculated HCLPF capacity was used in the quantification.

d)

BV2 battery charger 2BAT-CHG}-7 is evaluated in calculation2734294-C-216, Revision 0 BVPS2 Seismic Fragility for Battery Chargers and Inverters, Section 5.2.

Weight is determined by Reference to Generic Implementation Procedure (GIP) and 3x weight of sheet metal is used. However this is for a control cabinet, not a battery charger.

A battery charger weight should be based on 45 lbs/ft3. The resulting weight by generic method is 1485 Ibs, not 1104 lbs as used in the calculation. A conservative 0.60 knock down factor is used in the fragility calculation for anchorage capacity due to unknown anchor type, but the anchor type was clearly identified during the peer review walkdown.

e)

BV2 Motor Control Center (MCC)-2-E06 is evaluated in 2734294-C-201, Revision 0 BVPS2 Seismic Fragility for Motor Control Centers, Section 5.5. Functional capacity of the MCC is based on ratio of generic equipment ruggedness spectra (GERS) to ISRS for 1 I Hz response in the vertical direction. The realistic failure mode of the MCC associated with vertical motion is not described. The anchorage section of the calculation IESComsulting rlRtzzo

2734294-R-036 Raision 0 May 11,2017 435 448 states thatvertical frequency is at least SSHzbut 18HZ is used for functional evaluation.

A plug weld detail is assumed for the base connection. Plug weld capacrty is governed by base metal capacity, although AISC no longer recognizes base metal as a realistic failure mode for filet welds.

D The BV 1 Primary Plant demineralized water storage tank (DWST) is evaluated in Calculation 124736 F-l35. Although it is essentially axisymmetric, loads are increased by 40% based on 100-40-40 considerations which are not applicable, thus introducing excess conservatism. The failure mode of tank wall bending is not applicable since the anchor chairs are encased in concrete.

g)

BVI RHR heat exchangers are evaluated in calculation 2734294-C-121, Revision 0 BVPS1 Seismic Fragility for Tanks and Heat Exchangers, Section 5.9. The 19.8 Hz frequency estimate is conservatively applied in all directions. The same input motion scape factor is used in all directions. CDFM capacity is scaled from a conservative design calculation. The design calculation includes operational considerations and seismic nozzle loads, but it is not checked if these loads are realistic for fragility evaluation purposes. No inelastic energy absorption factor is used.

h) 2FWS-FCV479 is evaluated in calculation 2734294-C-207, Revision 0 BVPS2 Seismic Fragility for Pneumatic Operated Valves, Section 5.13. Lack of meeting SQUG caveats is not described clearly in the calculation. A lower bound frequency estimate is used in the evaluation. A value lower than the calculated HCLPF capacity was used in the quantification.

i)

The functional/anchorage HCLPF capacity for the representative battery charger, BAT-CHGI-5, is conservatively assumedto be 0.1 g. Since this is one of the risk significant items ranked within top ten contributors to the seismie CDF, its fragility needs to be refined to obtain a more realistic estimate of the seismic fragility.

j)

For a group fans on isolators listed in Table 5.3-l of 2734294-C-109, Revision 0, the obtained HCLPF capacrty is calculated as2.29 g on Page 3l of 2734294-C-109, Revision 0. When a review of top contributors to seismic CDF, it is noticed that the fragility capacity for Emergency Switchgear heating, ventilation, and air-conditioning (HVAC) Fans is set to the HCLPF capacity of 0.3 g. Please explain the difference.

POSSIBLE RESOLUTION (Peer Review Team)

The Standard requires that realistic fragilities are used for top contributors to CDF. More realistic fragility analysis is required for these items.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and RIZZO Associates) a)

In response to the F&O, a refined fragility was calculated for the BVl RHR pumps in Revision 1 of BVPS-1 Calculation 2734294-C-106. To this end, existing computer analytical models of the pumps and support frame documented in design calculations were reproduced in a new Calculation l2-4735-F-141. This facilitated the elimination of conservatisms from the design calculation scaling approach in 2734294-C-106, lESGonsulting tlRtzzo

b) 2734294-R-036 Raision 0 May 1L, 2017 Page A36 of A48 Revision 0. Conservative assumptions removed by: 1) performing analysis in the NS and EW directions with their respective seismic accelerations, rather than an envel ope,Z) using specific motorweights for "A" and "B" pumps instead of an envelope,3) retaining plus or minus signs for nozzle loads instead of conservatively assuming absolute maxima,

4) including dead weight of the support frame, 5) transferring calculated pump foot reactions from "A" and o'8" pump models to the support frame model instead of envelope, 6) determining seismic responses of pump/frame based on 7To damping for welded steel structures and of piping nozzle loads based on 57o damping per ASCE 43-05 instead of conservative design damping, 7) using pinned connections at the support frame to reinforced concrete pier anchorage locations instead of conservatively assuming fixed connections, 8) applying the 100-40-40 rule for combining seismic spatial components in the three orthogonal directions instead of an absolute sum, 9) using the governing thermal condition instead of an envelope of potential thermal conditions for RHR pump suction and discharge nozzle loads, and l0) scaling seismi c nozzle loads based on resonant frequencies of piping reported in design evaluations. In Revision 0 of Calculation2734294-C-106, the governing failure mode was of the ductile steel anchorage and an inelastic energy absorption factor of greater than unity could have been warranted. However, Revision 1 of Calculation 2734294-C-106 expanded the structural fragility section for the RHR pumps to evaluation concrete-related failure modes of the anchorage calculated in accordance with ACI 349-06. The governing structural/anchorage failure mode of the pumps is concrete breakout failure ofthe pump support frame to reinforced concrete piers cast-in-place anchor bolts. An inelastic energy absorption factor was not used because this failure mode is briule; i.e., Fp:I. With respect to weld capacity, the capacity used in Revision 0 of Calculation 2734294-C-106 is in accordance with ANSI/AISC 360-10 Section J2.4 which states: "the design strength of welds shall be the lower value of the base material and the weld metal strength." All of these details are addressed in the Revision I of Calculation 2734294-C-106 an#or new Calculation 12-473 5-F-l 41, Revision 0.

It should be noted that the peer review F&O report has a typographical error in the Basis for Significance section of this F&O. The first word in the second paragraph is o'BVl,"

but the rest of the paragraph is about the Unit 2 pressurizer power operated relief valve (PORV), 2RCS-PCV455C. The fragility for this valve was updated after the BV2 model was locked. The fragility report summary table in Revision I of the fragility Report 2734294-R-013 reflects the updated valve HCLPF of 0.549, which has been incorporated into the PRA model (original HCLPF was 0.329). In addition, the Seismic PRA Quantification Notebook now includes a group of sensitivities in Section 6 which address the models sensitivity to refinement of fragilities. These new cases only look at seismic components whose Fussell-Vesely importance (FVI) is greater than 0.03; anything less is considered to not significantly change results even if the HCLPF values were improved. Those SSCs that had a FVI >0.03 had sensitivities performed in which the HCLPF was doubled, in order to bound the small changes in HCLPF values that would more realistically be expected. This was done for both CDF and LERF. In many cases, the sensitivity showed a small change in CDF/LERF, indicating that improving the fragility would have very little effect on the model. In the cases for which there is a lESGonsulting

{}Rtzzo

c) 273429+R-036 Ratision 0 May 1L, 201,7 Page A37 of A48 noticeable change, the fragilities of those SSCs are deemed to be already realistic-ither because they were refined following peer reviewo or the peer review team did not identiff any lack of realism in the fragility calculations-and calculating a more robust fragility is not seen as plausible. Therefore, the PRA team concludes that there are no possible conservatisms in the fragility calculations that are driving the model results or masking insights. The fragility for the PORV identified in this F&O was refined and also has a low Fussell-Vesely (FV), signiffing that any additional refinement would not have a significant impact on the CDF/LERF.

The fragility for this valve (BV2 pressurizer relief valve 2RCS-RV551A) was updated after the BV2 model was locked prior to peer review. The fragility report summary table in Revision I of the fragility Report 2734294-R-013 reflects the updated valve HCLPF of 0.559 which has been incorporated into the PRA model (original HCLPF was 0.329). In addition, the Seismic PRA Quantification Notebook now includes a group of sensitivities in Section 6 which address the models sensitivity to refinement of fragilities. These new cases only look at seismic components whose FVI is greater than 0.03; anything less is considered to not significantly change results even if the HCLPF values were improved.

Those SSCs that had a FVI >0.03 had sensitivities performed in which the HCLPF was doubled, in order to bound the small changes in HCLPF values that would more realistically be expected. This was done for both CDF and LERF. In many cases, the sensitivity showed a small change in CDFILERF, indicating thar improving the fragility would have very little effect on the model. In the cases for which there is a noticeable change, the fragilities of those SSCs are deemed to be already realistic-or because they were refined following the peer review-and calculating a more robust fragility is not seen as plausible. Therefore, the PRA team concludes that there are no possible conservatisms in the fragility calculations that are driving the model results or masking insights. Since the peer review, the fragility for the pressurizer relief valve identified in this F&O was refined, and also has a low FV, signifying that any additional refinement would not have a significant impact on the CDF/LERF.

To resolve this F&O, the fragility for BV2 battery charger 2BAT-CHG}-7 was evaluated with estimated weight of 1485 lbs calculated based on 45 pounds per cubic foot for battery chargers per the SQUG GIP and anchorage capacity based on plant-specific walkdown observations by qualified personnel from ABS Consulting,RlZZo, an#or FENOC that the anchors are shell-type Philips studs. Revision 1 of BVPS-2 Calculat ion 27 3 4294-C-21 6 documents the updated evaluation of 2BAT-CHG}-7.

PeTEPRI TR-102180, minimum frequencies of free standing MCCs are inthe range of 3-10 Hz in the horizontal direction. The minimum horizontal frequency of 7 Hz was appropriately used in this calculation as the lower bound estimate. While the vertical frequency of MCCs were consideredto be at 33 Hz and above for evaluation of anchorage, the minimum frequency considered in functional fragility analysis was limited to 15 Hzto account forpotentially damaging local modes ofthe MCC and intemal components (e.g. breakers, contactors, transformers) with lower resonant frequencies.

For anchorage fragility calculation, these local modes will not result in significant anchor loads and the evaluation is based on only the global resonant frequency which was judged lESGonsulting

[iHtzza*

d) e)

0 s) 2734294-R-036 Ratision 0 Moy 1,1,20L7 Page A38 of A48 above 33 Hz. Details of the connection between the MCC and base channel were not available during preparation of Revision 0 of this fragility analysis and a worst case scenario of plug weld anchorage was assumed for MCCs. Further walkdowns performed by qualified personnel from ABS, RIZZO, and/or FENOC confirmed that MCCs are connected to their base channel sills with 3/8-inch-diameter bolts. Therefore, calculation of HCLPF due to plug weld capacity is removed inRevision I ofthis calculation. In Revision 0 of this calculation, the plug weld capacity considered base metal capacity consistent with requirements in AISC 360-10, which considers base metal shear capacity as a potential failure mode. In Revision 0 of this calculation, plug welds governed the anchorage capacity of MCCs. Welded connections are considered brittle connections per EPRI NP-6041-SL and therefore an inelastic energy absorption factor of 1.0 was assigned. Also, in Revision I of this calculation the anchorage capacity is governed by headed studs in concrete, which are also considered to have brittle failure mode and an inelastic absorption capacity of 1.0 is assigned. Revision 1 of the MCC fragility Calculation2734294-C-201 includes the previously described expanded discussion and the updated anchorage evaluation.

Revision 1 of Calculation 2734294-C-121 was issued to calculate a refined fragility for the BV I Primary Plant DWST. To this end, horizontal loads are no longer combined withthe 100-40-40 rule in consideration ofthe essentially axisymmetric tank shape. The BVl walkdown Report 2734294-R-004, Revision 1 clearly shows the BVI Primary Plant DWST anchor chairs are not encased in concrete and therefore the last part of the peer review comment is not applicable.

The Seismic PRA Quantification Notebook now includes a group of sensitivities in Section 6 which address the models sensitivity to refinement of fragilities. These new cases only look at seismic components whose FVI is greater than 0.03; anything less is considered to not significantly change results even if the HCLPF values were improved.

Those SSCs that had a FVI >0.03 had sensitivities performed in which the HCLPF was doubled, in order to bound the small changes in HCLPF values that would more realistically be expected. This was done for both CDF and LERF. In many cases, the sensitivity showed a small change in CDF/LERF, indicating that improving the fragility would have very little effect on the model. In the cases for which there is a noticeable change, the fragilities of those SSCs are deemed to be already realistic-or because they were refined following the peer review-and calculating a more robust fragility is not seen as plausible. Therefore, the PRA team concludes that there are no possible conservatisms in the fragility calculations that are driving the model results or masking insights. The RHR HXs identified in this F&O has a low FV, signiffing that any additional refinement would not have a significant impact on the CDF/LERF.

The HCLPF for 2FWS-FCV479 was increased from 0.289 to 0.419 after locking the BV2 model. The summary table in Revision 1 of the BVPS-2 Fragility Report 2734294-R-013 was updated to match the fragilrty reported in Revision I of Calculation 2734294-C-207.

Also, the Seismic PRA Quantification Notebook now includes a group of sensitivities in Section 6 which address the models sensitivity to refinement of fragilities. These new cases only look at seismic components whose FVI is greater than 0.03; anything less is lESConsulting

[]Rtzzo h)

i) i)

2734294-R-036 Reaision 0 Moy 1.1,201"7 Page A39 of A48 considered to not significantly change results even if the HCLPF values were improved.

Those SSCs that had a FVI >0.03 had sensitivities performed in which the HCLPF was doubled, in order to bound the small changes in HCTPF values that would more realistically be expected. This was done for both CDF and LERF. In many cases, the sensitivity showed a small change in CDFILERF, indicating that improving the fragility would have very little effect on the model. In the cases for which there is a noticeable change, the fragilities of those SSCs are deemed to be already realistic<r because they were refined following the peer review-and calculating a more robust fragility is not seen as plausible. Therefore, the PRA team concludes that there are no possible conservatisms in the fragility calculations that are driving the model results or masking insights. Since the peer review, the fragility for the Flow Control Valve identified in this F&O was refined, and also has a low FV, signifying that any additional refinement would not have a significant impact on the CDF/LERF.

Revision 1 of Calculation 2734?94-C-116 was issued to calculate arefined fragility for BAT-CHGl-S. To this end, an experience-based approach of 1.5 x Reference Spectrum was used to establish a functional fragility. The anchorage fragility is in excess of the functional fragility based on a review of the seismic characteristics of the component and its anchorage and walkdown photographs and observations documented in the walkdown Report 2734294-R-004, Revision l. The governing HCLPF based on the refined calculation was 0.709.

The correct and final HCLPF value is 2.299. The 0.309 value was originally submitted to the PRA modeler for its initial risk quantification using conservative assumptions. This fan subsequently showed as a top contributor and a more representative fragility of 2.299 was calculated. The 0.309 HCLPF value was incorrectly left in Revision 0 of the Fragility Report (2734294-R-006). This value has been corrected in Revision l.

lESGonsulting (iRtzzo

2734294-R-036 Ratision 0 May 1,L, 2017 Page A40 of AaB F&O 4-13 PRA Peer Review Fact & Observation4-13 was identified inthe Seismic Fragility Analysis High Level Requirement, Supporting Requirement SFR-A2 (and other affected Supporting Requirements SFR-F4, SPR-E6).

IIETAILS (Peer Review Team)

The LERF model appears to be conservative with regard to the structural failures modeled in Top Event ZLZ that are mapped directly to CDF and LERF.

Building structures are important to LERF. Fragilities should be realistic.

BASIS FOR SIGNIFICANCE (Peer Review Team)

Structural failures in Top Event ZLZ are important contributors to LERF. However, it is not clear from the documentation how these failures cause core damage and containment failure.

This is especially true for the MS Cable Vault structure (where it is not clear how core damage is guaranteed) and the Containment (where the dominant failure mode is an internal wall, not a functional failure of containment).

The failure mode of buildings needs to be realistic. There is no explanation of how a failure of a single internal wall leads to gross failure of the Reactor/Containment Bldg.

Report 2734294-C-133, Revision 0, states the lowest HCLPF of the Reactor/Containment building walls is 0.619. This is an internal wall (690-INT-W2). This HCLPF is assigned as the gross failure mode of the Reactor/Containment Bldg.

There is a discrepancy in structural damping. Calculation2734294-C-133 Fragility Analysis Reactor Containment (RCBX) Section7.2 Damping Factor states seismic demand is based on 7% structural damping. But Report 2734294-R-006, Section 7.2.4 Modeling of Structural Parameters states the structural damping of 4%o is assumed based on the expected damage level.

In the typical huilding response analysis, the 4% damping is used to be consistent with the CDFM approach. However, when the building structural responses obtained from the CDFM building analysis are used with the separation of variables approach, it is stated that the converted building responses are equivalent to response analysis results corresponding to 7Yo structural damping. For example, on Page 54 of 2734294-C-128 R0 BVPSl Fragility Analysis Auxiliary Building (AXLB), it is stated that the seismic demand is based onTo/o structural damping. The basis for this when the 4%o structural damping is actually used for the CDFM approach is not described.

Forces and moments for selected major shear walls and columns are provided in Tables A.I-l and A.I-2 of 2734294-R-005, Part A. Then, these appear to be converted to median values for use with separation of variables and presented in Section 5.2 of Calculation 2734294-C-128, Revision 0. It is not clear how this conversion was conducted. Please provide the process for how the CDFM-calculated demands were converted to the corresponding median demands.

lE$Gonsulting tlRtzzo

2734294-R-036 Rwision 0 May 1.1,201.7 Page A41-of A48 A review of building fragility calculations shows that the variabilities associated with the following fragility parameters were not included:

r Horizontal Direction Peak Response

. Vertical Component Response

. Time History No fragilities are calculated for floor diaphragms.

In shear wall fragilities, axial compression forces are neglected.

Forces and moments for selected structural components are provided as follows:

. 2734294-R-005, Revision 1 Part A, Attachment A.I for Auxiliary Building

. 2734294-R-005, Revision I Part B, Attachment B.I for Reactor Building e 2734294-R-005, Revision I Part C, Attachment C.I for Diesel Generating Building

. 2734294-R-005, Revision I Part D, Attachment D.I for Fuel and Decontamination Buildings r 2734294-R-005, Revision I Part E, Attachment E.I for Service Building

. 2734294-R-005, Revision 1 Part F, Attachment F.I for Main Steam Valve and Cable Vault Building

. 2734294-R-005, Revision I Part G, Attachment G.I for Intake Structure t 2734294-R-005, Revision 1 Part H, Attachment H.I for Safeguards Building All these include twisting moments in the summary tables. It is not described how the twisting moments are considered as part of building fragility evaluations.

Section 6.3 in 2734294-R-005, Revision 1 Part H states the following:

"This approach consenratively assumes that all accelerations are co-directional and ignores the effects due to mode shapes. This conservative bias could be as high as about 50% in individual structural components, but it is considered acceptable because the fragilities of the structural components, such as reinforced concrete walls, are generally high and; therefore, will not contribute to the CDF (fragilities of other components will control). If subsequent calculations determine otherwise, the specific strucfural components will be re-evaluated to obtain more accurate estimates of forces and moments. We anticipate that this will be accomplished by integrating stresses from the SASSI analysis."

This statement acknowledges conservatisms embedded in the seismic demands for Safeguards Building and justifies them based on the assumption that the corresponding building fragilities do not play a major role in the plant risk such as CDF. However, a review of top 10 contributors to LERF reveals that Safeguards Building is one of the three buildings that are ranked within the first top three contributors to LERF, along with Main Steam Cable Vault (MSCV) and Reactor Containment Buildings. Therefore, the building fragilities for these three buildings need to be refined by eliminating the aforementioned conservatisms.

ll$Gonsulting

[iRtzzo

2734294-R-036 Rwision 0 May 11,2017 Page A42 of 448 While the documents mentioned in this finding are from BV l, this ohservation also extends to BV2.

POSSIBLE RESOLUTION (Peer Review Team)

Review the dominant contributors to LERF to assure they are assessed as realistically as possible. Document the assumptions that are used to map the structural failures to CDF and LERF.

Provide basis that the lowest fragility of a component of a building represents the gross failure fragility of the building.

Correct the discrepancy in the description of structural damping.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and RIZZO Associates)

In Revision 0 structural fragility calculations which were reviewed by the peer review team, a fragility was calculated for the limit state of structural deformation causing failure of equipment anchorage using ASCE 43-05 inelastic energy absorption factors for Limit State C. The failure of equipment supported within these structures will lead to core damage. The capacity calculated for structural deformation causing failure of equipment anchorage was also conseryatively taken as representative for collapse. Collapse of the CIS or adjacent buildings such as the MS Cable Vault structure can be assumed to guarantee containment failure.

Revision 2 of structural fragility calculations include a calculation of the capacity for the limit state of incipient collapse using ASCE 43-05 inelastic energy absorption factors for Limit State A. Revision 2 of structural fragility Calculations 2734294-C-128 through C-135 and C-228 through C-235 include expanded discussion of limit states and a calculation of collapse capacity.

In Revision 0 of the reactor building structural fragility calculation which was reviewed by the peer review team, a fragility was calculated for the limit state of structural deformation causing failure of equipment anchorage using ASCE 43-05 inelastic energy absorption factors for Limit State C. The failure of equipment supported within the reactor building will lead to core damage. The capacity calculated for structural deformation causing failure of equipment anchorage was also conservatively taken as representative for collapse. Collapse of the CIS can be assumed to guarantee containment failure. Revision 2 of the reactor building structural fragility calculation includes a calculation of capacity for the limit state of incipient collapse using ASCE 43-05 inelastic energy absorption factors for Limit State A. To address this F&O, discussion was added stating that the critical structural members for which fragilities are calculated are major walls and columns for which failure poses a potential gross loss of structural stability that could lead to collapse of the structure. Yielding of minor walls is not a concern since loads in these walls will be redistributed to the major shear walls. Internal wall-690-INT-W2 is categorized as a major shear wall. Failure of internal wall 690-INT-W2 according to the limit state of structural deformation causing failure of equipment anchorage leads to core damage. Failure of internal wall 690-INT-W2 according to the limit state of incipient collapse leads to large-early release. Revision 2 of the reactor building structural fragility Calculations 2734294-C-133 (Unit 1) and 2734294-C-233 (Unit 2) include expanded discussion of failure modes, limit states and a calculation of collapse capacity.

lESGonsulting tlRtzz.o

2734294-R-036 Raision 0 May 11,2017 Page A43 of A48 As pointed out by the peer review team, 4o/o structural damping based on Response Level I was used to obtain the seismic structural response documented in Revision I of the Building Seismic Analysis Reports 2734294-R-005 (Unit l) and 2734294-R-012 (Unit 2) which is appropriate for development of ISRS for use in CDFM equipment HCLPF calculations. However, for evaluating forces and moments in structural members using the separation of variables method, a higher level of structural damping is permissible per ASCE 43-05. To address the finding, structural fragility Calculations 2734294-C-128 through -135 andC-228 through C-235 were revised as follows. For fragility evaluation of the limit state of collapse used for LERF quantification, Response Level 3 structural damping of l0% was used for evaluating seismic-induced forces and moments in structures by elastic analysis as permiued by ASCE 43-05. For fragility evaluation of the limit state of structural deformation causing failure of equipment anchorage used for CDF quantification, structural damping was limited to Response Level 2 of 7od since the structure will be at a less degraded condition at the limit state which will cause incipient failure of wall mounted anchorage. The higher damping levels and associated variabilities were incorporated in to the fragility analysis via the Damping Factor, one of the Separation of Variables Structural Response Factors. This change results inaSZYo higher seismic capacity for the limit state of structural deformation causing failure of equipment anchorage and a 58Yo higher seismic capacity for the limit state of collapse. Revision 2 of structural fragility Calculations 2734294-C-128 through C-135 and C-228 through C-235 include the above updates.

In response to this finding, structural fragility Calculations 2734294-C-128 through C-I35 and C-228 through C-235 following the separation of variables methodology were revised to convert CDFM level demands defined at the 84th percentile NEP to median demand using the following approach. The seismic demand used in the structural fragility calculations reviewed by the peer review team was developed with I time history and BE soil properties in accordance with ASCE 4-98, which resulted in an approximately 84th percentile NEP structural response appropriate for CDFM evaluations. In order to achieve realistic structural fragilities, the 84th percentile NEP seismic forces and moments in the walls and colufirns were reduced by a median demand conservatism ratio factor based on EPRI Report 1019200 in the revised calculations.

The median demand conservatism ratio factor was calculated using a seismic demand logarithmic standard deviation based on probabilistic SSI studies in literature. Structural fragility calculations following the separation of variables methodology were revised to reduce seismic forces and moments in the walls and columns by the median demand conservatism ratio factor to obtain a median response. As a result, structural fragilities increased by approximately 18%. Revision 2 of structural fragility Calculations 2734294-C-128 through C-135 and C-228 through C-235 include the above updates.

A detailed breakdown of the logarithmic standard deviations associated to each of the aforementioned factors is presented in the Revision 2 of the fragility calculations for each of the structures evaluated in the BVPS. It is noted that these calculations assume that variabilities associated with the Horizontal Direction Peak Response, the Vertical Component Response and Time History simulation do not contribute significantly to the log standard deviations in the seismic demand. In response to F&O 4-13, this assumption is re-examined as follows. The variability associated with the horizontal direction peak response accounts for the fact that the PGA in any one horizontal direction may exceed the geo-mean PGA used to base the fragilities.

lESGoneulting

[iRtzz]a

2734294-R-036 Reuision 0 Moy 11-,201,7 Page Aaa of Aat Althoughthe SRSS method is used in calculating the total seismic shear in awall, much of this shear is determined by the input motion parallel to the orientation of the wall. Therefore, the corresponding log standard deviation is taken to be 0.12 consistent with the recommendations in EPRI TR-103959. Because the vertical FIRS is site-specific, the variability associated with the basic variable "vertical component response" is typically represented by pr in the range of 0.22 to 0.28 and Bu less than about 0.2 (EPRI 103959). However, the effect of the vertical load on the wall shear capacity is relatively small (see also response to F&O 4-13). Therefore, the associated pr in seismic margin is relatively small (on the order of 0.01). The time histories used in the analysis closely match the target FIRS at 5o/o damping. The peaks and valleys are less than plus or minus l0% above or below the target FIRS at range of frequencies of 2.5 Hz to I Hz, near the fundamental frequency of the building; i.e., 4Hz. Thus, it is judged that a time history simulation factor is 1.0 and an uncertainty of 0.05 is used consistentwith EPRI TR 103959.

Also, Recent EPRI workshops have recommended that if only one time history is used in obtaining the 84th percentile response a random variability of 0.15 should be assigned to reflect effects of random phasing of the Fourier components on the resulting peak response. Revision 1 of the fragility analysis reports 2734294-R-006 (Unit 1) and 2734294-R-013 (Unit 2) include the discussion of these fragility analysis factors and the updated structural fragility parameters.

Floor diaphragm fragilities were considered to not govern over the in-plane shear and moment capacities of vertical structural members. The primary purpose of floor diaphragms part of the lateral force resisting system is to transfer lateral forces in a given floor into the vertical members of the lateral force resisting system. Typical floor slab thickness of BVPS buildings is 2 ft and longer spans ile supported by beams composite with the slabs. Given the typical thickness and configurations of the floor diaphragms, it is judged their fragilities do not govern over in-plane shear or flexure fragilities of shear walls near the base resisting lateral forces accumulated from the stories above. Revision I of the fragility analysis reports 2734294-R-006 (Unit 1) and 27 34294-R-0 I 3 (Unit 2) include the j ustification for the omission of floor diaphragm fragility evaluation.

In response to this finding, a representative structural fragility calculation was revised to demonstrate the effect of the axial compressive forces on shear wall shear capacity. The effect was found to be insignificant and therefore it was concluded the assumption to omit the effect from calculations remains valid. The other structural fragility calculations were revised to reference the representative calculation for the basis for omission of axial compressive load effect on shear wall shear capacity. Revision? of structural fragility Calculations 2734294-C-128 through C-135 andC-228 through C-235 include the above described updates.

To address this F&O, Calculation l2-4735-F-148, Revision 0, was preparedto elaborate and demonstrate how twisting moments reported in the Building Seismic Analysis Reports 2734294-R-005 (Unit 1) and 2734294-R-012 (Unit 2) affect building seismic fragilities documented in structural fragility Calculations 2734294-C-128 tlrough C-135 and C-228 through C-235. The calculation clarifies that the reported twisting moments are the resultant of the distribution of out-of-plane shear forces on the elements that comprise the section cuts. Also, the calculation estimates the out-of-plane shear strength factor for both with and without the effects of the twisting moment for a representative BVPS structure. The strength factors are IESGonsulting

()Rtzzo

2734294-R-036 Raision 0 May 1L,20L7 Page Aa1 of Aa9 compared to the reported strength factors from the structural fragility calculation, which are based on in-plane shear. Including the effects of the resultant twisting moments, the calculation demonstrates thatthe maximum out-of-plane shear is well withinthe shearcapacity of the wall, and confirms that out-of-plane shear does not govern the wall fragility.

The justification for the approach to obtain forces and moments used as inputs to structural fragility calculations was clarified and augmented. The approach implemented to obtain the response quantities on the structural members uses the maximum absolute accelerations resulting from the SSI analyses in an equipment static analysis of the fixed-base structure. The equivalent static analysis is performed using the program SAP2000. The equivalent static analysis conservatively assumes that all response accelerations are co-directional and ignores the effects due to mode shapes. However, this is justified on the basis that the dominant mode shape is typically characterized by monotonically increasing shear displacements with height. The conservative bias could be as high as 50% for some structural components such as columns and other elements which may be influenced by local modes. The approach is further judged to be acceptable on the following basis. Fragility refinements were performed which increased the HCLPFs of the CDF related failure mode (i.e., building deformation causing equipment failure) by a factor ranging from 1.3 to 1.8. For LERF, a refined fragility was calculated (i.e., building collapse) which increases the HCLPF used in quantification by a factor ranging from 2.21o 2.9.

Considering these increase factors, the fragilities of structural components such as reinforced concrete shear walls are high and therefore are not expected to be significant contributors to CDF or LERF. The above descrihed basis is documented in Revision 2 of structural fragility Calculations 2734294-C-128 through C-135 andC-228 through C-235 and Revision 2 of Building Seismic Analysis Reports 2734294-R-005 (Unit 1) and 2734294-R-012 (Unit 2).

lESGonsulting

()Rtzzo

2734294-R-036 Raision 0 May 1L,2A1.7 Page Aa6 of Aa$

F&O 4-16 PRA Peer Review Fact & Observation 4-16 was identified in the Seismic Fragility Analysis High Level Requirement, Supporting Requirement SFR-A2 (and other affected Supporting Requirement SFhF4).

IIETAILS (Peer Review Team)

Containment building analysis for BVI and BV2 is not realistic.

BASIS FOR SIGNIFICANCE (Peer Review Team)

On Page 18 of 2734294-R-005, Part B, the second paragraph states that the steel liner is anchored to the concrete inside surface at sufficiently close intervals so that the overall deformation of the liner is essentially the same that of the concrete wall; thus, performing as additional reinforcement. Then, on Page 28 of 2734294-R-005, Part B, the third paragraph further states that the steel lining on the internal face of the reinforced walls of the RCS was modeled by defining a concrete equivalent thickness; such that the moment of inertia per unit width results is equal to 0.5 the moment of inertia of concrete (cracked stiffiress) plus the moment of inertia from the transformed steel lining area. The mass and weight densities are modified accordingly, to match the actual values for steel plus concrete.

As stated above, the steel liner is not explicitly treated in the analysis model and converted to the equivalent concrete thickness. Then, 2734294-R-005, Revision l, Part B, Attachment B.I, presents resulting section forces and moments for a section cut located at EL 690 as follows, which is slightly less than the bottom of the steel liner elevation ofEL 690 ft-lI-inches.

It is important to note that the obtained forces and moments in Tables B.I-l and B.I-2 are consistent with the requirements of the CDFM approach. Thus, they need to be adjusted to be median-centered values when the separation of variables approach is used for building fragility evaluations. However, when Section 5.2 of 2734294-C-133, Revision 0, is reviewed, it is found that the values from Tables B.I-1 and B.I-2 are directly copied and used in the fragility evaluation.

POSSIBLE RESOLUTION (Peer Review Team)

Based on these findings and observations, the following should be addressed:

Explain why the CDFM-related section forces and moments from Tables B.I-1 and B.I-2 of 2734294-R-005, Part B, are directly used for the separation of variables fragility evaluation in Section 5.2 of 2734294-C-133, Revision 0.

Explain why the twisting bending moments from Tables B.I-l and B.I-2 of 2734294-R-00s Part B, are completely ignored in Section 5.2 of 2734294-C-133, Revision 0.

It appears that the obtained section cut forces presented in Tables B.I-1 and B.I-2 of 2734294-R-005, Part B, are for the combined section of the concrete and the steel liner. This approach may be reasonable when the overall section capacity of the combined section is evaluated assuming the perfect composite action at the interface between the liner and the concrete section. However, this approach does not consider another potential mode associated lESGonsulting tlRtzzo

2734294-R-036 Ranision 0 May 1,1, 201.7 Page A47 of Aa9 with the liner itself such as rupturing due to excessive strain. This failure mode should be separately evaluated.

PLANT RESPONSE OR RESOLUTION (ABS Consulting and HIZZ;O Associates)

With respect to CDFM forces and moments, in response to this finding, reactor building structural fragility Calculations 2734294-C-133 (Unit 1) and 2734294-C-233 (Unit 2) were revised to convert CDFM level demands defined at the 84tr percentile NEP to median demand using the following approach.

The seismic demand used in the structural fragility calculations reviewed by the peer review team was developed with one time history and BEs soil properties in accordance with ASCE 4-98, which resulted in an approximately 84th percentile NEP structural response appropriate for CDFM evaluations. In order to achieve realistic structural fragilities, the 84th percentile NEP seismic forces and moments in the walls and columns were reduced by a median demand conservatism ratio factor based on EPRI Report 1019200 in the revised calculations.

The median demand conservatism ratio factor was calculated using a seismic demand logarithmic standard deviation based on probabilistic SSI studies in literature. Structural fragility calculations following the separation of variables methodology were revised to reduce seismic forces and moments in the walls and columns by the median demand conservatism ratio factor to obtain a median response. As a result, structural fragilities increased by approximately 18%.

Relatedto twisting moments, to address this F&O, Calculation l2-4735-F-148, Revision 0, was prepared to elaborate and demonstrate how twisting moments reported in the reactor building fragility calculations affect building sei smic fragilitie s.

The calculation clarifies that the reported twisting moments are the resultant of the distribution of out-of-plane shear forces on the elements that comprise the section cuts. Also, the calculation estimates the out-of-plane shear strength factor for both with and without the effects of the twisting moment for a representative BVPS structure. The strength factors are compared to the reported strength factors from the structural fragility calculation which are based on in-plane shear. Including the effects of the resultant twisting moments, the calculation demonstrates that the maximum out-of-plane shear is well withinthe shear capacity of the wall, and confirms that out-of-plane shear does not govern the wall fragility.

Pertaining to the combined concrete and steel liner section, as pointed out by the peer reviewers, the steel liner is not explicitly treated in the analysis model and converted to the equivalent concrete thickness. This approach adequately captures the dynamic response of the steel liner/concrete shield.

For cylindrical shell structures such as the containment building, local shear or bending failures will not govern the capacity under seismic loading. Instead, global failure will govern where the whole cross section is engaged in shear or flexure eliciting a composite response. Thus, local failure of the steel liner is precluded under seismic loading.

AESGortsulting

[]Rtzzo

2734294-R-036 Ranision 0 May 11,2017 Page A48 of A48 Revision I of the Fragility Analysis Reports (2734294-R-006/2734294-013) include the rationale for not evaluating rupture fragility of the containment steel liner.

lEf,Gonsulting tlRtzzo

ABSG COHSULTING INC.

300 Commerce, Suite 200 lrvine, CA 92602 Telephone 714-7344242 Fax 714-7344252 NORTH AMERICA lESGonsulting ABS GROUP OF COMPANIES, INC.

16855 Northchase Drive Houston, TX 77060 Telephone 281-673-2800 Fax 281-673-2801 EUROPE Sofia, Bulgaria Telephone 359-2-9632049 Piraeus, Greece Telephone 30-21 04294046 Genoa, ltaly Telephone 39-01 0-2512090 Hamburg, Gernany Telephone 4940-300-92-22-21 Las Arenas, Spain Telephone 34-944644444 Rotterdam, The Netherlands Telephone 31-10106-0778 Amsterdam, The Netherlands Telephone 31-205-207-947 G0teborg, Sweden Telephone 46-70-283-0234 Bergen, Norway Telephone 47-55-55-10-90 Oslo, Nonruay felephone 4747-57-2740 Stavanger, Noway Telephone 47-51-93-92-20 Trondheim, Nonray Telephone 47-73-900-500 A$IA-PACIFIC Ahmedabad, lndia Telephone 079 4000 9595 NaviMumbai, lndia Telephone 91 757{780 New Delhi, lndia Telephone g1-1 145634738 Yokohama, Japan Telephone 81 45450-1250 Kuala Lumpur, Malaysia Telephone 603-79822455 Kuala Lumpur, Malaysia Telephone 603-2161-5755 Beijing, PR China Telephone 86-10-581 12921 Shanghai, PR China Telephone 86.21{876-9266 Busan, Korea Telephone 82-51-8524661 Seoul, Korea Telephone 82-2-5524661 Alexandna Point, Singapore Telephone 6562704663 Kaohsiung, Taiwan, Republic of China Telephone 886-7-271-3463 Bangkok, Thailand Telephone 662-399-2420 West Perh, WA 6005 Telephone 61{-9486-9909 INTERNET Additional office information can be found at:

www,absgroup.com SOUTH AMERICA 2100 Space Park Drive, Suite 100 Houston, TX 77058 Telephone 713-929-6800 Energy Crossing ll, E. Building 15011 Katy Freeway, Suite 100 Houston, TX 77094 1525 Wilson Boulevard, Suite 625 Arlington, VA 22209 Telephone 703-682-7373 Fax 703-682-7374 10301 Technology Drive Knoxville, TN 37932 Telephone 865-966-5232 Fax 86S966-5287 1745 Shea Center Drive, Suite 400 Highland Ranch, CO 80129 Telephone 303-674-2990 1390 Piccard Drive, Suite 350 Rockville, MD 20850 Telephone 301 -907-9100 Fax 301-990-7185 3115 East Lion Lane, Suite 160 Salt Lake City, UT 84121 Telephone 801 -333-7676 Fax 801-333-7677 140 Heimer Road, Suite 300 San Antonio, TX 78232 Telephone 210-495-51 95 Fax 210495-5134 823 Congress Avenue, Suite 1510 Austin, TX 78701 Te le phone 512-7 32-2223 Fax 512-233-2210 55 Westport Plaza, Suite 700 St. Louis, M0 63146 Telephone 314-81 g-1550 Fax 314{19-1551 One Chelsea Street New London, CT 06320 Telephone 860-7014608 100 Danbury Road, Suite 105 Ridgefield, CT 06877 Telephone 2034314281 Fax 203431-3643 1360 Truxtun Avenue, Suite 103 North Charleston, SC 29405 Telephone 843-297-0690 152 Blades Lane, Suite N Glen Burnie, MD 21060 Telephone 410-5144450 MEXTC0 Maca6, Brazil Telephone 55-22-2763J018 Rio de Janeiro, Brazil Telephone 55-21-3179-31 82 Sao Paulo, Brazil Telephone 55-1 1-3707-1 055 Vifia del Mar, Chile Telephone 56-32-2381 780 Bogota, Colombia Telephone 571-2960718 Chuao, Venezuela Telephone 58-21 2-959-7442 Lima, Peru Telephone 51 -1 437-7430 Manaus, Brazil Telephone 55-92-321351 1 Montevideo, Uruguay Telephone 5982-2-901 33 UHITED KINGDOM EQE House, The Beacons Wanington Road Birchwood, Wanington Cheshire WA3 6WJ Telephone 44-1925-287300 3 Pride Place Pride Park Derby DE24 8QR Telephone 44-0-1332-254-01 0 Unit 3b Damery Works Woodford, Berkley Gloucestershire GL13 9JR Telephone 44{-1 454-269-300 ABS House 1 Frying Pan Alley London E1 7HR Telephone U-207 -377 4422 Aberdeen AB25 1XQ Telephone 444-1224-392100 London W1T 4TQ Telephone 444-203-301 -5900

['llDDLE EAST Ciudad del Carmen, Mexico Telephone 52-938-3824530 Mexico City, Mexico Telephone 52-55-551 14240 Monteney, Mexico Telephone 52-81+319-0290 Reynosa, Mexico Telephone 52-899-920-2642 Veracruz, Mexico Telephone 52-229-980+133 Dhahran, Kingdom of Saudi Arabia Telephone 966-3-868-9999 Ahmadi, Kuwait Telephone 965-3263886 Doha, State of Qatar Telephone 97444-13106 Muscat, $ultanate of Oman Telephone 968-597950 lstanbul, Turkey Telepho ne 90-212-661 4127 Abu Dhabi, United Arab Emirates Telephone 971 -26912000 Dubai, United Arab Emirates Telephone 971 4-33061 16