Latest revision as of 17:01, 19 January 2022

Text

NUREG/BR-0058, Rev. 5, Appendix H Severe Accident Risk Analysis - Dfc
	ML21096A294
Person / Time
Issue date:	04/30/2021
From:	Pamela Noto; Office of Nuclear Material Safety and Safeguards
To:	;
	Malone, Tina
Shared Package
ML21096A269	List: ML21096A292; ML21096A293; ML21096A294; ML21096A295;
References
	NUREG/BR-0058, Rev. 5
	Download: ML21096A294 (150)
	v • d • e

1 APPENDIX H 2 SEVERE ACCIDENT RISK ANALYSIS

1 TABLE OF CONTENTS 2 LIST OF FIGURES...................................................................................................... H-v 3 LIST OF TABLES ..................................................................................................... H-vii 4 ABBREVIATIONS AND ACRONYMS ....................................................................... H-ix 5 H.1 PURPOSE ......................................................................................................... H-1 6 H.2 BACKGROUND................................................................................................. H-2 7 H.3 SEVERE REACTOR ACCIDENT RISK INFORMATION USED IN SAFETY 8 GOAL EVALUATION AND COST-BENEFIT ANALYSIS ................................. H-6 9 H.3.1 Probabilistic Risk Assessment Model Selection Guidance .................................. H-6 10 H.3.1.1 Probabilistic Risk Assessment Model Scope .......................................... H-7 11 H.3.1.2 The Structure of Traditional Nuclear Power Plant Probabilistic Risk 12 Assessment Models............................................................................... H-8 13 H.3.2 Risk Metrics for Evaluating Substantial Safety Benefit ........................................ H-9 14 H.3.3 Common Analysis Elements .............................................................................. H-12 15 H.3.3.1 Accident Sequence Analysis ................................................................ H-12 16 H.3.3.2 Quantification of Change in Accident Frequency .................................. H-14 17 H.3.3.3 Quantification of Change in Consequences ......................................... H-15 18 H.3.3.4 Identification and Estimation of Affected Parameters ........................... H-16 19 H.4 GRADED APPROACH TO ANALYSIS ........................................................... H-18 20 H.4.1. Example of Approach ........................................................................................ H-20 21 H.4.2. Sources of Information ...................................................................................... H-22 22 H.5 MAJOR-EFFORT ANALYSIS ......................................................................... H-28 23 H.5.1 Accident Sequence Analysis .............................................................................. H-28 24 H.5.2 Severe Accident Progression Analysis .............................................................. H-30 25 H.5.2.1 Sources of Information .......................................................................... H-30 26 H.5.2.2 MELCOR Modeling Approach .............................................................. H-31 27 H.5.3 Offsite Consequence Analysis ........................................................................... H-32 28 H.5.3.1 Sources of Information .......................................................................... H-32 29 H.5.3.2 MACCS Modeling Approach ................................................................. H-32 30 H.6 SUPPLEMENTAL ANALYSES ....................................................................... H-36 31 H.6.1 Uncertainty Analyses ......................................................................................... H-36 32 H.6.1.1 Uncertainties in PRA Models ................................................................ H-36 33 H.6.2 Sensitivity Analyses and Plant-to-Plant Variability Analyses ............................. H-39 H-iii NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.6.2.1 Sensitivity Analyses .............................................................................. H-40 2 H.6.2.2 Plant-to-Plant Variability Analyses ........................................................ H-40 3 H.7 PRESENTATION OF RESULTSINPUTS TO REGULATORY 4 ANALYSIS ....................................................................................................... H-42 5 H.7.1 Aggregating Probabilistic Risk Assessment Results from Different Hazards ..... H-42 6 H.7.2 Offsite Consequence Measures ......................................................................... H-42 7 H.7.2.1 Conditional Consequence Measures .................................................... H-42 8 H.7.3 Evaluation of Regulatory Alternatives ................................................................ H-44 9 H.7.3.1 Results from the Core Damage Event Tree Quantification ................... H-44 10 H.7.3.2 Results from the Accident Progression Event Tree Quantification ....... H-44 11 H.7.3.3 Results from MELCOR Analysis ........................................................... H-45 12 H.7.4 Risk Integration Results and Key Insights.......................................................... H-46 13 H.8 REFERENCES ................................................................................................ H-50 14 ENCLOSURE H-1: DESCRIPTION OF ANALYTICAL TOOLS AND 15 CAPABILITIES ......................................................................... H-58 16 ENCLOSURE H-2:

SUMMARY

OF THE STATE-OF-THE-ART REACTOR 17 CONSEQUENCE ANALYSES (SOARCA) PROJECT ............. H-73 18 ENCLOSURE H-3:

SUMMARY

OF DETAILED ANALYSES FOR SECY-12-0157, 19 CONSIDERATION OF ADDITIONAL REQUIREMENTS FOR 20 CONTAINMENT VENTING SYSTEMS FOR BOILING WATER 21 REACTORS WITH MARK I AND MARK II 22 CONTAINMENTS ................................................................... H-77 23 ENCLOSURE H-4:

SUMMARY

OF DETAILED ANALYSES FOR SECY-15-0085, 24 EVALUATION OF THE CONTAINMENT PROTECTION AND 25 RELEASE REDUCTION FOR MARK I AND MARK II 26 BOILING-WATER REACTORS RULEMAKING 27 ACTIVITIES............................................................................. H-92 28 ENCLOSURE H-5:

SUMMARY

OF DETAILED ANALYSES FOR SECY-13-0112 29 AND NUREG-2161, CONSEQUENCE STUDY OF A BEYOND-30 DESIGN-BASIS EARTHQUAKE AFFECTING THE SPENT 31 FUEL POOL FOR A U.S. MARK I BOILING-WATER 32 REACTOR ............................................................................ H-111 33 ENCLOSURE H-6:

SUMMARY

OF DETAILED ANALYSES IN COMSECY-13-0030, 34 ENCLOSURE 1, REGULATORY ANALYSIS FOR JAPAN 35 LESSONS-LEARNED TIER 3 ISSUE ON EXPEDITED 36 TRANSFER OF SPENT FUEL.............................................. H-130 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-iv

1 2 LIST OF FIGURES 3

4 Figure H-1 Overall Logic and Structure of Traditional NPP PRA Models ............................. H-9 5 Figure H-2 Distribution of 2016 Point Estimates for Total CDF, U.S. Plants ...................... H-11 6 Figure H-3 Distribution of 2016 Point Estimates for LERF, U.S. Plants .............................. H-11 7 Figure H-4 Uncertainty in Average Individual Latent Cancer Fatality Risk (0-10 miles) in 8 the 2015 Containment Protection and Release Reduction Regulatory 9 Analysis............................................................................................................. H-22 10 Figure H-5 Modular Approach to Event Tree Development in CPRR Analysis................... H-29 11 Figure H-6 Parametric Uncertainty Analysis Results for Individual Latent Cancer 12 Fatality Risk ...................................................................................................... H-39 13 Figure H-7 Likelihood of a Leak and Magnitude of Releases from Beyond-Design-Basis 14 Earthquake........................................................................................................ H-45 15 Figure H-8 Comparison of Regulatory Analysis Alternatives Using Population Dose 16 Risk (0-50 miles) ............................................................................................... H-48 17 Figure H-9 Reduction in 50-mile Offsite Cost Risk ($/reactor-year) ................................. H-48 18 Figure H-10 Uncertainty in Reduction in 50-mile Offsite Cost Risk ...................................... H-49 19 Figure H-11 Overall Logic and Structure of Traditional NPP PRA Models and Role of 20 SAPHIRE, MELCOR, and MACCS Code Suites .............................................. H-59 21 Figure H-12 Simplified Diagram of Event Tree with Initiating Event (IE) and Two 22 Supporting Fault Trees ..................................................................................... H-62 23 Figure H-13 Simplified Event Tree Structure ........................................................................ H-82 24 Figure H-14 Reduction in 50-mile Population Dose Risk (person-rem/ry) .......................... H-85 25 Figure H-15 Reduction in 50-mile Offsite Cost Risk ($/ry) .................................................. H-85 26 Figure H-16 Uncertainty in Reduction in 50-mile Population Dose Risk ............................... H-87 27 Figure H-17 Uncertainty in Reduction in 50-mile Offsite Cost Risk ...................................... H-87 28 Figure H-18 Uncertainty in Average Individual Latent Cancer Fatality Risk (0-10 miles)..... H-98 29 Figure H-19 Modular Approach to Event Tree Development .............................................. H-100 30 Figure H-20 Conditional Probability of SFP Liner Leakage and SFP Release 31 Magnitude ....................................................................................................... H-126 32 H-v NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 LIST OF TABLES 3

4 Table H-1 Options Defining Scope of Commercial NPP PRAs ............................................. H-8 5 Table H-2 Reactors with Published SAMA Analyses .......................................................... H-23 6 Table H-3 Salem Nuclear Generating Station Core Damage Frequency for Internal 7 Events at Power ................................................................................................. H-26 8 Table H-4 Salem Nuclear Generating Station Breakdown of Population Dose by 9 Containment Release Mode ............................................................................... H-27 10 Table H-5 Ratio of Consequences for 1-Year Intermediate Phase Duration 11 Sensitivity Cases to Baseline Cases in the Containment Protection and 12 Release Reduction Analysis ............................................................................... H-41 13 Table H-6 Severe Accident Consequence Analysis ResultsExample ............................. H-44 14 Table H-7 Risk Estimates by Regulatory Analysis Subalternative ...................................... H-47 15 Table H-8 Conditional Annual Average Individual Latent Cancer Fatality Risk from 16 SOARCA Unmitigated Scenarios within 10 miles of the Plant ........................... H-75 17 Table H-9 Hypothetical Plant Modifications ........................................................................ H-81 18 Table H-10 Mapping of Simplified Event Tree Sequences to Plant Modifications 19 and MELCOR Cases .......................................................................................... H-82 20 Table H-11 Parameter Values Used to Estimate Radiological Release Frequencies........... H-83 21 Table H-12 Mean MACCS Consequence Results for Selected MELCOR Accident 22 Scenarios ........................................................................................................... H-83 23 Table H-13 Point Estimate Risk Values for Each Hypothetical Plant Modification................ H-84 24 Table H-14 Risk Reductions from Severe-Accident-Capable Venting System Plant 25 Modifications ...................................................................................................... H-84 26 Table H-15 Parameter Uncertainty Distributions .................................................................. H-86 27 Table H-16 Summary of Quantitative Cost-Benefit Analysis Results for Filtered 28 Containment Vent System using a $2,000 per Person-Rem Conversion 29 Factor ................................................................................................................. H-89 30 Table H-17 Summary of Adjusted Quantitative Cost-Benefit Analysis Results for 31 Filtered Containment Vent System using a $4,000 per Person-Rem 32 Conversion Factor .............................................................................................. H-89 33 Table H-18 Ratings Assigned to Each Alternative by Qualitative Factor .............................. H-90 34 Table H-19 Summary of Regulatory Subalternatives and Distinguishing Attributes ............. H-96 35 Table H-20 MACCS Results for 18 Mark I Source Term Bins ............................................ H-103 36 Table H-21 MACCS Results for 9 Mark II Source Term Bins ............................................. H-104 37 Table H-22 Risk Estimates by Regulatory Analysis Subalternative .................................... H-106 38 Table H-23 Uncertainty Analysis Inputs .............................................................................. H-107 39 Table H-24 Results for Baseline Cases with Different Site Files ........................................ H-108 40 Table H-25 Operating Cycle Phase Descriptions ............................................................... H-115 41 Table H-26 Scenario Descriptions for a Given Operating Cycle Phase .............................. H-116 42 Table H-27 Summary of Release Results for High-Density Configurations ........................ H-119 43 Table H-28 Summary of Release Results for Low-Density Configurations ........................ H-119 H-vii NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Table H-29 Binning of MELCOR Release Sequences into Release Categories for 2 High-Density Configurations ............................................................................. H-121 3 Table H-30 Binning of MELCOR Release Sequences into Release Categories for 4 Low-Density Configurations ............................................................................. H-121 5 Table H-31 Mean Reduction in Offsite Consequence Results Associated with Option 2 ... H-123 6 Table H-32 Summary of Benefits and Costs within 50 Miles for Option 2 .......................... H-128 7 Table H-33 Combined Effect of $4,000 per Person-Rem Conversion Factor and 8 Consequences Beyond 50 Miles for Option 2 .................................................. H-128 9 Table H-34 SFP Groupings Used for the Staff's Technical and Cost-Benefit Analyses ..... H-134 10 Table H-35 Key Input Parameters Used for Sensitivity Analyses ....................................... H-137 11 Table H-36 Summary of Net Benefits for Each Spent Fuel Pool Group* ............................ H-138 12 Table H-37 Net Benefits for Low-Density SFP Storage for Groups 1-4 from Combined 13 Sensitivity Analyses that Analyzed Consequences Beyond 80 kilometers 14 (50 Miles) and Using an Adjusted Dollar per Person-Rem Conversion 15 Factor ............................................................................................................... H-139 16 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-viii

1 2 ABBREVIATIONS AND ACRONYMS 3

4 delta or incremental 5 $ U.S. dollars 6 ADAMS Agency wide Documents Access and Management System 7 ANS American Nuclear Society 8 AP1000 Advanced Passive 1000 9 APET accident progression event tree 10 ASME American Society of Mechanical Engineers 11 ATD atmospheric transport and dispersion 12 Ba chemical element barium 13 B&W Babcock and Wilcox 14 BWR boiling-water reactor 15 C degrees Celsius 16 Ci consequences for each potential accident i 17 CDET core damage event tree 18 CDF core damage frequency 19 Ce chemical element cerium 20 CE Combustion Engineering 21 CFR Code of Federal Regulations 22 Ci radiation units in Curies 23 CPRR containment protection and release reduction 24 Cs chemical element cesium 25 DF decontamination factor 26 DOE U.S. Department of Energy 27 DW drywell 28 DWF drywell first strategy 29 ELAP extended loss of alternating current power 30 EPA U.S. Environmental Protection Agency 31 EPRI Electrical Power Research Institute 32 EPZ emergency planning zone 33 ESP early site permit 34 ETE evacuation time estimate 35 F degree Fahrenheit 36 FLEX flexible coping strategies 37 FR Federal Register 38 GE General Electric 39 gpm flow rate in gallons per minute 40 I chemical element iodine 41 IE initiating event 42 ILRT integrated leak rate testing 43 IPE individual plant examination 44 IPEEE individual plant examination for external events 45 ISLOCA interfacing systems loss-of-coolant accident 46 K degrees Kelvin 47 Kg/m3 gas density in kilograms per cubic meter 48 Kg/s mass flow rate in kilograms per second 49 KI chemical compound potassium iodide 50 La chemical element lanthanum H-ix NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 LERF large early release frequency 2 LMT liner melt-through 3 LTSBO long-term station blackout 4 LWR light-water reactor 5 MACCS MELCOR Accident Consequence Code System 6 MCi radiation unit in million Curies 7 Mo chemical element molybdenum 8 Mod modification 9 MWt megawatt thermal 10 NEI Nuclear Energy Institute 11 NFPA National Fire Protection Association 12 NPP nuclear power plant 13 NRC U.S. Nuclear Regulatory Commission 14 NSSS nuclear steam supply systems 15 NTTF Near-Term Task Force 16 OCP operating cycle phase 17 OMB Office of Management and Budget 18 OP overpressurization 19 Pi probability or frequency of potential accident i 20 PAG protective action guide 21 PRA probabilistic risk assessment 22 Psi pounds per square inch 23 psig pounds per square inch gauge 24 PWR pressurized-water reactor 25 QHO quantitative health objective 26 R risk 27 RC release category 28 RPV reactor pressure vessel 29 Ru chemical element rubidium 30 RuO2 chemical compound ruthenium oxide 31 Ry reactor-year 32 SAMA severe accident mitigation alternative 33 SAMDA severe accident mitigation design alternative 34 SAPHIRE Systems Analysis Program for Hands-on Integrated Reliability 35 Evaluations 36 SAWA severe accident water addition 37 SAWM severe accident water management 38 SBO station blackout 39 SFP spent fuel pool 40 SGTR steam generator tube rupture 41 SOARCA State-of-the-Art Reactor Consequence Analyses 42 SPAR Standardized Plant Analysis Risk 43 SRM staff requirements memorandum 44 STSBO short-term station blackout 45 Te chemical element tellurium 46 U.S. United States 47 W rate of sensible heat 48 WWF wetwell first strategy 49 Xe chemical element xenon NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-x

1 SEVERE ACCIDENT RISK ANALYSIS 2

3 H.1 PURPOSE 4

5 The purpose of this appendix is to provide guidance and best practices for use at the 6 U.S. Nuclear Regulatory Commission (NRC) when performing probabilistic risk assessments 7 (PRAs) and consequence analyses as part of regulatory, backfit, and environmental analyses 8 for nuclear power reactors.

9 10 Used in conjunction with the discussion in Section 5 of this NUREG, this appendix explains how 11 to perform the safety goal evaluation and the valuation of the public health (accident) and 12 economic consequences (offsite property) attributes for the purposes of cost-benefit analysis. It 13 provides references on sources of information and an overview of the tools and methods used 14 to estimate baselines and changes in core damage frequency (CDF), large early release 15 frequency (LERF), public health risk, and offsite economic consequences risk. Onsite risk 16 attributesoccupational health risk (accident) and onsite property riskare also relevant to 17 nuclear power reactor severe accident risk but are not within the scope of this appendix.

18 Finally, the guidance on performing offsite consequence analyses is useful for reference when 19 conducting the severe accident mitigation alternative (SAMA) and severe accident mitigation 20 design alternative (SAMDA) analyses that are required under the National Environmental Policy 21 Act (see Appendix I, National Environmental Policy Act Cost-Benefit Analysis Guidance, to this 22 NUREG).

23 24 This appendix does not impose new requirements, establish NRC policy, or instruct NRC staff in 25 preparing cost estimates. Rather, it provides information on accepted state-of-practice methods 26 for estimating the frequency and consequence components of the risk from hypothetical 27 accidents at nuclear power plants (NPPs), for the purposes of safety goal evaluations and 28 cost-benefit analyses for regulatory, backfitting, forward fitting, issue finality, and National 29 Environmental Policy Act environmental review analyses.

30 H-1 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.2 BACKGROUND 2

3 The quantification of risks associated with postulated severe accidents is an integral part of the 4 NRCs regulatory policy and practices. A severe accident is an accident that involves 5 extensive core damage and fission product release into the reactor vessel and containment, 6 with potential release to the environment (NRC, 2013f; ASME/ANS, 2009). The NRC uses 7 PRAs for the severe accident risk quantification that is needed in regulatory, backfit, and 8 environmental analyses.

9 10 The NRC has a long history of using PRA techniques to characterize severe accident risks in 11 support of its regulatory processes and decisions. Since the completion of the seminal Reactor 12 Safety Study (WASH-1400, Reactor Safety Study: An Assessment of Accident Risks in 13 U.S. Commercial Nuclear Power Plants, issued October 1975 (NRC, 1975)), PRAs have 14 provided important, actionable safety insights through a number of different studies. In the late 15 1970s, the NRC used insights from PRA in consideration of topics, including the likelihood of 16 loss-of-coolant accidents, the reliability of direct current power supplies, and the effectiveness of 17 alternate containment designs (NRC, 2016c). In the early 1980s, the NRC relied on PRA 18 techniques to address unresolved safety issues involving accidents such as the anticipated 19 transient without scram (NRC, 1978) and station blackout (SBO) rules (NRC, 1988b). The NRC 20 considered risk arguments in support of licensee requests to extend equipment outage times 21 and the Commission used information from licensee-sponsored PRAs to inform its decision in 22 1985 to allow continued operation of the Indian Point power plants (NRC, 2016c).

23 24 In 1985, the Commission issued a policy statement on severe accidents, which recognized that 25 plant-specific PRAs had exposed unique vulnerabilities to severe accidents and were a 26 potential source of significant new safety information to identify instances of undue risk 27 (NRC, 1985). This policy statement led to the issuance of Generic Letter 88-20, Individual 28 Plant Examination for Severe Accident Vulnerabilities10 CFR 50.54(f), dated 29 November 23, 1988 (NRC, 1988a), asking each licensee to conduct an individual plant 30 examination (IPE) to identify plant-specific vulnerabilities to severe accidents and report the 31 results to the Commission, and later to Generic Letter 88-20, Supplement 4, Individual Plant 32 Examination of External Events (IPEEE) for Severe Accident Vulnerabilities10 CFR 50.54(f),

33 dated June 28, 1991 (NRC, 1991), which focused on severe accidents initiated by external 34 events. As a result, 74 PRAs representing 106 U.S. NPPs were completed; the assessments 35 calculated CDF and LERF 1 and gave the utilities a method for tracking improvements made in 36 terms of risk abatement and cost effectiveness (Keller and Modarres, 2005). The NRC 37 documents its staff summary and evaluation of licensee submittals under this program in 38 NUREG-1560, Individual Plant Examination Program: Perspectives on Reactor Safety and 39 Plant Performance, issued December 1997 (NRC, 1997a), and NUREG-1742, Perspectives 40 Gained from the Individual Plant Examination of External Events (IPEEE) ProgramFinal 41 Report, issued April 2002 (NRC, 2002), for the IPEs and IPEEEs, respectively. The NRC had 42 also sponsored an assessment of the risks from severe accidents in five commercial nuclear 43 power plants in the United States which was published in 1990 as NUREG-1150, Severe 44 Accident Risks: An Assessment for Five U.S. Nuclear Power Plants (NRC, 1990b).

45 NUREG-1150 and supplementary studies based on NUREG-1150 were the main sources of 46 information and basis for the NRCs 1997 NUREG/BR-0184, Regulatory Analysis Technical 47 Evaluation Handbook, Final Report (NRC, 1997b); for example, see NUREG/BR-0184, 1 LERF is defined as The frequency of a rapid, unmitigated release of airborne fission products from the containment to the environment that occurs before effective implementation of offsite emergency response, and protective actions, such that there is a potential for early health effects (NRC, 2013f).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-2

1 Table 5.3 and Appendix B.4.

2 3 The Commission formally endorsed the use of PRA methods in nuclear regulatory activities in 4 its 1995 policy statement (NRC, 1995a), which includes the following precepts:

5 6

The use of PRA technology should be increased in all regulatory matters to the extent 7 supported by the state-of-the-art in PRA methods and data and in a manner that 8 complements the NRCs deterministic approach and supports the NRCs traditional 9 defense-in-depth philosophy.

10 11

PRA and associated analyses (e.g., sensitivity studies, uncertainty analyses, and 12 importance measures) should be used in regulatory matters, where practical within the 13 bounds of the state-of-the-art, to reduce unnecessary conservatism associated with 14 current regulatory requirements, regulatory guides, license commitments, and staff 15 practices. Where appropriate, PRA should be used to support the proposal for 16 additional regulatory requirements in accordance with 10 CFR 50.109 (Backfit Rule) 17 [Title 10 of the Code of Federal Regulations (10 CFR) 50.109, Backfitting].

18 19

PRA evaluations in support of regulatory decisions should be as realistic as practicable 20 and appropriate supporting data should be publicly available for review.

21 22 The 1995 policy statement introduced the concept of risk-informed regulation; which solidified 23 the role of PRA methods and results in regulatory decisionmaking. Today, the NRC conducts 24 risk analyses for a wide range of regulatory activities and processes. Examples of activities that 25 rely on PRA include:

26 27

Regulatory analysis and backfit analysis: PRAs are used to determine whether 28 additional new regulatory requirements for licensees could lead to a substantial safety 29 improvement. Potential benefits such as reduced public health risk or reduced risk of 30 offsite economic consequences are quantified as part of the cost-benefit analysis to 31 justify new or amended rules or guidance.

32 33

New reactor certification and licensing: 10 CFR 52.47, Contents of Applications; 34 Technical Information, requires that an application for standard design certification 35 contain a description of the plant-specific PRA and its results. A similar requirement 36 applies to combined license applicants in 10 CFR 52.79, Contents of Applications; 37 Technical Information in Final Safety Analysis Report.

38 39

Risk-informed decisionmaking:

40 41 o Changes in plant licensing basis: Operating reactor licensees may use risk 42 information to support a voluntary change from a plants current licensing basis 43 to a new licensing basis. Regulatory Guide 1.174, An Approach for Using 44 Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific 45 Changes to the Licensing Basis (current version), provides guidance on the use 46 of PRA findings and risk insights to a support licensee request for changes to a 47 plants licensing basis.

48 H-3 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 o Reactor oversight: The NRCs regulatory framework for reactor oversight is 2 risk-informed and performance based. 2 The Reactor Oversight Process uses 3 performance indicators and inspection findings that are color coded according to 4 safety/risk significance. Within the Reactor Oversight Processs strategic 5 performance area of reactor safety, significance determinations of inspection 6 findings and events rely on plant-specific risk information, such as the changes in 7 CDF and LERF.

8 9

Environmental reviews: The licensee prepares an environmental report and submits it 10 to the NRC for independent evaluation as part of an application for license renewal for 11 an existing reactor, a design certification application for a new reactor, and a 12 construction and operating license application for a new reactor. These reports are 13 required to include SAMA or SAMDA evaluations to identify potential features or actions 14 that would prevent or mitigate the consequences of a severe accident. These 15 requirements appear in 10 CFR 51.53(c)(3)(ii)(L) for operating reactor license renewal 16 applicants; 10 CFR 51.55, Environmental Report; Standard Design Certification, for 17 new reactor design certification applicants; and 10 CFR 51.75, Draft Environmental 18 Impact StatementConstruction Permit, Early Site Permit, or Combined License, for 19 new reactor construction permits, early site permits, 3 and combined license 20 environmental impact statements. A PRA and offsite consequence analysis would 21 support whether these SAMA are cost-beneficial.

22 23 In addition, the 2011 accident at the Fukushima Dai-ichi NPP in Japan initiated a large-scale 24 effort by the staff to identify potential modifications to equipment and operational requirements 25 to address the lessons learned from this disaster. The NRC undertook a number of major 26 regulatory analyses to inform Commission decisions. Notable examples are listed below, with 27 additional information available in enclosures to this appendix as indicated. The following 28 analyses are regulatory analyses that supported these NRC decisions. The enclosures to this 29 appendix summarize these analyses and highlight the approaches and evaluation criteria that 30 were used, the information that was provided, the results and insights, and the resulting 31 Commission decision, if applicable. These enclosures are intended to provide useful examples 32 for performing these types of analyses.

33 34

SECY-12-0157, Consideration of Additional Requirements for Containment Venting 35 Systems for Boiling Water Reactors with Mark I and Mark II Containments, dated 36 November 26, 2012 (NRC, 2012h) and SRM-SECY-12-0157, Consideration of 37 Additional Requirements for Containment Venting Systems for Boiling Water Reactors 38 with Mark I and Mark II Containments, dated May 19, 2013 (NRC, 2013h). See also 39 Enclosure H-3.

40 41

SECY-15-0085, Evaluation of the Containment Protection and Release Reduction for 42 Mark I and Mark II Boiling-Water Reactor Rulemaking Activities, dated June 18, 2015 43 (NRC, 2015a) and SRM-SECY-15-0085, Evaluation of the Containment Protection and 44 Release Reduction for Mark I and Mark II Boiling-Water Reactor Rulemaking Activities, 45 dated August 19, 2015 (NRC, 2015c). See also Enclosure H-4.

46 47

The spent fuel pool (SFP) study supporting the evaluation of expedited transfer or spent 48 fuel, SECY-13-0112, Consequence Study of a Beyond-Design-Basis Earthquake 2 https://www.nrc.gov/reactors/operating/oversight/rop-description.html 3 This applies if a design has been chosen at the early site permit stage.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-4

1 Affecting the Spent Fuel Pool for a U.S. Mark I Boiling-Water Reactor, dated 2 October 9, 2013 (NRC, 2013e). See also Enclosure H-5.

3 4

COMSECY-13-0030, Staff Evaluation and Recommendation for Japan 5 Lessons-Learned Tier 3 Issue on Expedited Transfer of Spent Fuel, dated 6 November 12, 2013 (NRC, 2013g) and SRM-COMSECY-13-0030, Staff 7 RequirementsStaff Evaluation and Recommendation for Japan Lessons-Learned Tier 8 3 Issue on Expedited Transfer of Spent Fuel, dated May 23, 2014 (NRC, 2014h). See 9 also Enclosure H-6.

10 11

Mitigation of beyond-design basis events is described in SECY-15-0065, Proposed 12 Rulemaking: Mitigation of Beyond-Design-Basis Events, dated April 30, 2015 13 (NRC, 2015d) and SRM-SECY-15-0065, Staff RequirementsProposed Rulemaking:

14 Mitigation of Beyond-Design-Basis Events, dated August 27, 2015 (NRC, 2015f).

15 16 These activities have resulted in a more consistent and technically justified application of PRA 17 and severe accident consequence analysis in the NRCs regulatory process and serve as the 18 basis for this guidance. The following sections explain the risk information, tools, methods, and 19 approaches that are used to conduct these analyses.

20 H-5 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.3 SEVERE REACTOR ACCIDENT RISK INFORMATION USED IN 2 SAFETY GOAL EVALUATION AND COST-BENEFIT ANALYSIS 3

4 The NRC uses a risk analysis framework to determine when a proposed requirement may meet 5 the substantial additional protection standard and to provide some of the metrics needed to 6 weigh the costs against the benefits of a regulatory action. Evaluating the benefits associated 7 with a regulatory action requires the quantification of both the likelihood and the conditional 8 consequences of fission product release for a spectrum of hypothetical severe accident 9 scenarios. The complexity of the risk analysis depends on the type of analysis to be conducted.

10 This appendix should be used with Section 2.1 of this NUREG to understand the level of effort 11 needed for each type of analysis and the factors that should be used to determine which 12 analysis is appropriate.

13 Staff should consult the most current PRA information available when beginning a new analysis.

14 15 A basic principle of this NUREG is that each analysis should be adequate for its intended 16 application in terms of the type of information supplied, the level of detail provided, the level of 17 uncertainty, and the availability of design margin. In general, the severe accident risk analysis 18 considers plant systems and operator responses to initiating events leading to core damage 19 (Level 1 PRA) and accident progression to the release of fission products to the environment 20 (Level 2 PRA), while combining estimates of radiological release category frequencies and their 21 associated consequences (Level 3 PRA) to produce risk estimates. This section details the 22 technical approach used to complete each portion of the risk evaluation. These discussions 23 assume familiarity with the concepts of risk as related to the nuclear industry, as well as 24 knowledge of event- and fault-tree terminology. The analyst should consult existing PRAs and 25 standard references 4 for further information on these concepts. Sections H.4 through H.6 26 provide specific guidance for performing analyses.

27 28 H.3.1 Probabilistic Risk Assessment Model Selection Guidance 29 30 The purpose of this section is to provide the analyst with guidance on selecting PRA models to 31 perform safety goal screenings and estimate the potential public health benefits (from avoided 32 accidents) associated with a proposed regulatory action. Performing these evaluations requires 33 a PRA model to analyze the effects of the proposed action. The most important considerations 34 for selecting the PRA model are its scope and its level of detail, which together should be 35 sufficient to assess the issues of concern.

36 4 For instance, NUREG/CR-2300, PRA Procedures Guide: A Guide to the Performance of Probabilistic Risk Assessments for Nuclear Power Plants, issued January 1983 (NRC, 1983a), and NUREG-0492, Fault Tree Handbook, issued January 1981 (NRC, 1981).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-6

1 H.3.1.1 Probabilistic Risk Assessment Model Scope 2

3 The NPP PRA models can vary in scope, depending on their intended application or use. As 4 summarized in Table H-1, the scope of a PRA is defined by the extent to which various options 5 for the following five factors are modeled and analyzed:

6 7 (1) Radiological sources: The NPP sites contain multiple sources that could potentially 8 release radioactive material into the environment under accident conditions. Although 9 most PRA models focus on the reactor core, other important sources that could be 10 modeled in the PRA to estimate the public health accident risk from an NPP site include 11 (1) spent nuclear fuel (both wet and dry storage), (2) fresh nuclear fuel, and 12 (3) radiological waste storage tanks.

13 14 (2) Exposed population: In estimating the numbers of radiological health effect cases 15 attributable to a postulated nuclear accident, both onsite and offsite populations may be 16 considered. Typical NPP PRA models estimate the radiological health risk to members 17 of the general public located at various distances from the NPP site. Although these 18 PRA models do not consider the risk to onsite workers and first responders to a nuclear 19 accident, the radiological health risks to these groups typically are considered as part of 20 other attributes included in a regulatory analysis (e.g., occupational health (accident)).

21 22 (3) Initiating event hazard groups: Initiating events cause the plant to deviate from its 23 intended operating state and challenge plant control, safety systems, and operator 24 actions designed to prevent reactor core damage and the release of radioactive material 25 to the environment. These events include failure of equipment from (1) internal causes 26 (e.g., transients, loss-of-coolant accidents, internal floods, internal fires) or (2) external 27 causes (e.g., earthquakes, high winds, tsunamis). In an NPP PRA model, similar 28 causes of initiating events are organized by hazard group and are then assessed using 29 common assumptions, methods, and data to characterize their effects on the plant.

30 31 (4) Plant operating states: In determining the public risk from NPP operations, it is 32 important to consider not only the response of the plant to initiating events occurring 33 during at-power operation but also its response to initiating events occurring while the 34 plant is in other operating states, such as low-power and shutdown. Plant operating 35 states are used to subdivide the plant operating cycle into unique states defined by 36 various characteristics (e.g., reactor power, coolant temperature, coolant pressure, 37 coolant level, equipment configuration) so that the plant response can be assumed to be 38 the same for all initiating events that occur when a plant is assumed to be in a particular 39 plant operating state.

40 41 (5) End state (level of risk characterization): The NPP PRA models can be used to 42 calculate risk metrics at different end states. The text below discusses in more detail the 43 three different end states or levels of risk characterization that traditionally have been 44 used in NPP PRA models.

45 H-7 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Table H-1 Options Defining Scope of Commercial NPP PRAs Factor Scoping Options for Commercial NPP PRAs Reactor core(s)

Radiological sources Spent nuclear fuel (SFP and dry cask storage)

Other radioactive sources (e.g., fresh fuel and radiological wastes)

Exposed population Offsite population Internal hazards

Traditional internal events (transients, loss-of-coolant accidents)

Internal floods Initiating event

Internal fires hazard groups External hazards

Seismic events (earthquakes)

Other site-specific external hazards (e.g., high winds, external flooding)

At-power Plant operating states Low-power Shutdown Level 1 PRA: Initiating event to onset of core damage Level 1 plus LERF: Level 1 plus limited scope Level 2, which is End state/Level of sufficient for the purpose of calculating LERF risk characterization Level 2 PRA: Initiating event to radioactive material release from containment Level 3 PRA: Initiating event to offsite radiological consequences 2

3 The most important aspects to consider when evaluating the scope of a PRA model is to ensure 4 that it includes significant risk contributors that are relevant to the evaluation of a proposed 5 regulatory action and that the level of detail is appropriate with respect to scope, level of detail, 6 and technical acceptability.

7 8 H.3.1.2 The Structure of Traditional Nuclear Power Plant Probabilistic Risk Assessment 9 Models 10 11 Risk can be characterized in many ways, depending on the end states of interest for a decision 12 or application. To provide some overall logic and structure and to facilitate evaluation of 13 intermediate results, PRAs for NPPs have traditionally been organized into three analysis levels.

14 Three sequential adverse end states that can occur in the progression of postulated NPP 15 accident scenarios define these levels (1) onset of damage to the nuclear fuel in the reactor 16 core (termed core damage), (2) release of radioactive materials from the NPP containment 17 structure to the surrounding environment (termed radiological release), and (3) adverse human 18 health, environmental, and economic consequences that occur beyond the boundary of the NPP 19 site (commonly referred to as offsite radiological consequences).

20 21 Figure H-1 illustrates the overall logic and structure of traditional NPP PRA models, including 22 the types of results that are produced at each level.

23 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-8

Increasing PRA model scope and complexity LEVEL 3 PRA MODEL LEVEL 2 PRA MODEL LEVEL 1 PRA MODEL Initiating Events & Mitigating Severe Accident Conditional Probabilistic Systems Response Logic Phenomenological Models Consequence Analysis Models & Containment Systems Models Core Damage Radiological Release Offsite Radiological Consequences

Total core damage frequency

Radiological release category

Population dose

Core damage sequence frequencies

Adverse human health effects information

Representative source term

Contaminated areas

Importance measures information

Economic costs 1

2 Figure H-1 Overall Logic and Structure of Traditional NPP PRA Models 3

4 In NPP Level 3 PRAs, the output of PRA logic models that estimate the frequencies of a 5 representative set of radiological release categories intended to capture a reasonably complete 6 spectrum of possible accident scenarios is combined with the conditional consequence results 7 for each release category. For each outcome of interest, the consequences are then summed 8 across all radiological release categories to estimate the mean annual risk of that outcome.

9 10 The first step in conducting the analysis is to identify the potential source of risk (e.g., reactor 11 core, spent fuel, dry cask storage), reactor operating states (e.g., at-power, low-power, 12 shutdown), and hazards of concern (e.g., internal events, external events, all hazards) for 13 analysis. The potential source of risk will usually be determined by the objective statement 14 described in Sections 2.3.1 and 2.3.2 of this NUREG, which provide guidance on defining the 15 regulatory problem statement and identifying regulatory alternatives. A complete assessment of 16 alternatives that includes all relevant accident scenarios may require the development of plant-17 specific, full-scope Level 3 PRAs for each plant type of interest. However, this may exceed the 18 required level of detail necessary for a regulatory analysis. For most regulatory analyses, the 19 regulatory problem statement will delineate the accident initiators and sequences to be 20 considered.

21 22 H.3.2 Risk Metrics for Evaluating Substantial Safety Benefit 23 24 For potential backfit considerations, it is useful to have an approximation of the range of the 25 CDFs and LERFs for relevant classes of plants. Section 2.4.1 of this NUREG describes the 26 quantitative risk thresholds for substantial safety benefit. The NRC uses LERF instead of the 27 historical conditional containment failure probability (see for example, Regulatory Guide 1.174).

28 The analyst has access to a current body of CDF and LERF information of operating NPPs from 29 a variety of sources. These sources include the NRCs plant-specific Standardized Plant 30 Analysis Risk (SPAR) models, risk information in SAMA analyses supporting license renewal H-9 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 applications, and license amendment requests supporting risk-informed regulatory applications 2 such as those for risk-informed in-service inspection (NRC, 2003).

3 4 Figures H-2 (CDF) and H-3 (LERF) show representative distributions of point estimates for CDF 5 and LERF, published in NUREG-2201, Probabilistic Risk Assessment and Regulatory 6 Decisionmaking: Some Frequently Asked Questions, issued September 2016 (NRC, 2016c).

7 The purpose of these figures in this appendix is to provide a general illustration of the 8 distribution of CDFs and LERFs. These figures depict the CDF and LERF for a subset of the 9 U.S. fleet of operating power reactors, based on information readily available through NRC 10 regulatory applications. As noted in NUREG-2201, the CDFs are based on 2016 estimates for 11 61 units from license amendment requests to change requirements or SAMA analyses as part 12 of the environmental evaluation conducted by license renewal applicants. The earliest result is 13 from a 2002 analysis, but over 80 percent of the results are from 2008 or later. The estimates 14 are based on PRAs with different scopes, for example, the majority included internal plus 15 external event initiators while a minority included internal event initiators only.

16 17 The point estimate for CDFs range from about 4x10-6 per reactor-year to approximately 1x10-4 18 per reactor-year, with a mean and median of about 5x10-5 per reactor-year. The point estimates 19 for LERFs range from about 8x10-8 per reactor-year to approximately 3x10-5 per reactor-year, 20 with a mean of approximately 4x10-6 per reactor-year and a median of about 3x10-6 per 21 reactor-year. The source information for these estimates typically do not include uncertainty 22 estimates. NUREG-2201 also notes that it is important to recognize:

23 24 * [P]ast PRAs have consistently shown that potential vulnerabilities (and 25 therefore plant risk) are highly plant specific.

26

Design and operational changes addressing lessons identified by PRAs can 27 lead to significant changes in CDF...

28

The above estimates for total CDF are developed by adding the CDFs 29 estimated for different accident scenarios.

30

The CDF contributions from accidents caused by internal hazards (e.g.,

31 floods, fires) and external events (e.g., earthquakes, high winds, and external 32 floods) can be significant.

33 (Source: NUREG-2201, p. 36) 34 35 It is important to note that external events are sometimes out-of-scope or handled much less 36 rigorously than internal events (for example, in SAMA analyses for operating reactors). See 37 additional discussion in Section H.4.2, Sources of Information, and table notes under Tables 38 H-3 and H-4. Similar information is available for new and advanced reactors (see Section 39 H.5.2), with the exception that large release frequency is used instead of LERF.

40 41 As noted above, the analyst should access available risk information that is current at the time 42 of a future regulatory or cost-benefit analysis. Figures H-2 and H-3 provide an example based 43 on 2016 data for a subset of operating reactor units, with the aforementioned limitations.

44 45 46 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-10

1 2 Figure H-2 Distribution of 2016 Point Estimates for Total CDF, U.S. Plants 3 (Source: NUREG-2201, Figure 4-3) 4 5

6 Figure H-3 Distribution of 2016 Point Estimates for LERF, U.S. Plants 7 (Source: NUREG-2201, Figure 5-2)

H-11 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 H.3.3 Common Analysis Elements 3

4 Risk (R) is, summed over the spectrum of potential accidents, the product of (1) the probability 5 (or frequency) (Pi) and (2) associated consequences (Ci) for each potential accident (i) in the 6 spectrum, as shown in the equation below:

7 8 =

9 10 Hence, estimating the public health (accident) risk and offsite economic consequences (offsite 11 property damage) risk in a cost-benefit analysis for a proposed action requires the estimation of 12 both (1) the change in probabilities (frequencies) and (2) the change in consequences 13 associated with accidents in the spectrum of relevant accidents. Therefore, the common 14 analysis elements include the following:

15 16

An accident sequence analysis to identify the relevant accidents 17 18

Quantification of frequencies associated with individual accident sequences for the 19 probability/frequency portion of the risk equation 20 21

Quantification of the public health and offsite economic consequence associated with 22 each accident sequence, for the consequence portion of the risk equation 23 24 The following sections discuss these elements in greater detail.

25 26 H.3.3.1 Accident Sequence Analysis 27 28 An accident sequence analysis systematically identifies risk-significant accident sequences and 29 quantifies their frequency. Logic models provide the probabilistic framework for assessing the 30 change in risk associated with a regulatory analysis alternative. These models consist of event 31 trees to identify the set of possible accident sequences that lead to fission product release and 32 rely on accident progression simulations performed for a specific accident sequence to 33 understand how a combination of successes and failures affects the facility. The following 34 examples are for a nuclear power plant, but the principles apply to all NRC-regulated facilities.

35 36 PRA Logic Model Structure 37 38 One PRA modeling approach is to construct logic models using event trees and fault trees. An 39 event tree represents different plant and operator responses in terms of sequences of undesired 40 system states, such as core damage or fission product release, that could occur following an 41 initiating event. The probabilistic (Level 1 and Level 2 frequency) portions of an accident 42 sequence analysis are assessed using Core Damage Event Trees (CDETs) and Accident 43 Progression Event Trees (APETs). A fault tree identifies different combinations of basic events 44 (e.g., initiating events; failures of systems, structures, and components; and human failure 45 events) that could lead to a system failure. Fault tree models are linked to the event tree 46 sequences and allow for the identification and evaluation of minimal cut setsthe minimum 47 combinations of events needed to result in an adverse end state of interest (e.g., core damage).

48 When linked together, these logic structures provide an integrated perspective that can capture 49 major system dependencies.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-12

1 2 Care should be taken to ensure that the modeling is sufficiently detailed and is technically 3 adequate to provide the needed confidence in the resultsfor its use in the regulatory analysis 4 and for its role in the integrated decision process, which is critical for coherent decision-making.

5 Because the standards and industry PRA programs are not prescriptive, there is some freedom 6 on how to model these logic structures. The choice of specific assumptions, a particular 7 approximation, or a modeling choice or simplification may, however, influence the results.

8 These underlying assumptions and approximations made in the development of the PRA model 9 give rise to uncertainty and should be explicitly identified and quantified to aid the 10 decisionmaker in understanding the results and the potential range of costs and benefits. The 11 treatment of uncertainty and sensitivity analysis are further discussed in Section H.6.

12 13 PRA Logic Model Level of Detail 14 15 Much like the scope, the level of detail of an NPP PRA model can vary, depending on its 16 intended application or use. The level of detail is defined by the degree to which (1) the actual 17 plant is modeled and (2) the unlimited range of potential accident scenarios is simplified.

18 Although the goal of a PRA is to represent the NPP as-designed, as-built, and as-operated as 19 realistically as practicable, some compromises are made to keep the PRA model manageable, 20 considering time and resource constraints.

21 22 For each of the technical elements that comprise a PRA model, the level of detail may vary by 23 the extent to which the following is true:

24 25

Plant systems and operator actions are credited in modeling plant-specific design and 26 operation 27 28

Plant-specific operating experience and data for the plants structures, systems, and 29 components are incorporated into the model 30 31

Realism (as opposed to intentional conservatism) is incorporated into analyses that 32 predict the expected plant and operator responses 33 34 Furthermore, the logic structures (e.g., event trees and fault trees) in the model are simplified 35 representations of the complete range of potential accident scenarios. Simplifications are made 36 through underlying assumptions and approximations such as (1) the consolidation into 37 representative hazard groups of initiating event causes and (2) the screening out of certain 38 equipment failure modes.

39 40 Although the level of detail needed for an NPP PRA model is largely dependent upon the 41 requirements associated with its intended use (e.g., a PRA should meet the relevant American 42 Society of Mechanical Engineers [ASME] and American Nuclear Society [ANS] PRA standards 43 for operating reactor licensing changes), at a minimum, it needs to be detailed enough to model 44 the major system dependencies and to capture the significant risk contributors.

45 46 The level of effort required to construct these logic models depends upon the availability of 47 information and preexisting models developed for the specific site of interest and on the amount 48 of information that is obtainable from the licensee. The NRC has developed SPAR models for 49 all NPPs used to support various risk-informed activities. However, depending upon the scope 50 of the regulatory analysis, these models may need to be expanded to address other hazards or 51 plant conditions. To the extent possible, the analyst should use existing information, in addition H-13 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 to related research efforts, 5 to complete the regulatory analysis efficiently. Qualitative insights 2 may be needed to supplement incomplete quantitative modeling.

3 4 Assumptions about which systems will be available (or should be probabilistically considered) 5 are dependent upon the type of initiating event being considered. For example, if the initiating 6 event is seismically induced, consideration should be given to whether a given safety system 7 realistically would be available. The assumptions used in developing the event trees should be 8 clearly delineated for the systems that are probabilistically considered. In constructing the event 9 trees, systems or modes of operations for which reliability data are not available should not be 10 credited or probabilistically considered. The analyst should document for reference these 11 assumptions and all hardware-related failure event probabilities that are incorporated in the 12 CDETs and APETs.

13 14 H.3.3.2 Quantification of Change in Accident Frequency 15 16 The change in accident frequency is a key factor for several of the cost-benefit analysis 17 attributes. Estimates of the change in accident frequencies resulting from a proposed NRC 18 action are based on the effects of the action on appropriate parameters in the accident 19 equation. Examples of these parameters might be system or component failure probabilities, 20 including those for the facilitys containment structure. The estimation process involves two 21 steps(1) identification of the parameters affected by a proposed NRC action, and 22 (2) estimation of the values of these affected parameters before and after the action takes 23 place.

24 25 The parameter values are substituted in the accident equation to yield the base- and 26 adjusted-case accident sequence frequencies. The sum of their differences is the reduction in 27 accident frequency caused by the proposed NRC action. The frequency of accident sequence i 28 initiated by event j is 29 30 =

31 32 where = the frequency, F, of minimal cut set k for accident sequence i initiated by event j 33 Source: (NRC, 1997b).

34 5 For example, related research efforts include SPAR external events modeling (https://saphire.inl.gov/current_models_public.cfm), fire risk research under National Fire Protection Association (NFPA) 805, Performance-Based Standard for Fire Protection for Light-Water Reactor Electric Generating Plants (current version) (https://www.nrc.gov/reactors/operating/ops-experience/fire-protection/protection-rule/protection-rule-overview.html), and generic issue evaluations (https://www.nrc.gov/about-nrc/regulatory/gen-issues/dashboard.html).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-14

1 A minimal cut set represents a unique and minimum combination of occurrences at lower levels 2 in a structural hierarchy (e.g., component failures that are typically represented by basic events 3 in PRA model fault trees) needed to produce an overall occurrence (e.g., facility damage) at a 4 higher level. It takes the form of a product of these lower level occurrences. The affected 5 parameters comprise one or more of the multiplicative terms in the minimal cut sets. Thus, the 6 change in accident sequence frequency i, between the base model and the adjusted model that 7 incorporates the proposed action, is 8

9 = =

10 11 Source: (NRC, 1997b) 12 13 The changes in accident frequency for each affected accident sequence are added. Reduction 14 in accident frequency is algebraically positive; increase is negative. This equation assumes that 15 the model structure remains valid for risk evaluations after a proposed action. It is possible for 16 a proposed action to result in a change to the model structure (e.g., by adding or removing top 17 events in an event tree). Therefore, in addition to potentially changing the values of parameters 18 that comprise a base-case set of minimal cut sets, a proposed action can change the structure 19 of the minimal cut sets and create new minimal cut sets that were not included in the base case.

20 This would require an evaluation beyond quantification of the above equation, which only 21 quantifies the change of frequencies of existing minimal cut sets.

22 23 Each accident sequence that ends in core damage is binned for further analysis into a plant 24 damage state with other core damage sequences having plant conditions that are expected to 25 result in similar accident progression behavior. The frequencies of the sequences with a core 26 damage end state are summed to estimate the CDF for an initiating event. The characteristics 27 that define each plant damage state bin comprise the initial conditions for the APET. Similarly, 28 the APETs evaluate the containment response to those sequences that result in core damage 29 and provide the frequencies of sequences with end states of release to the environment.

30 31 Source terms are binned into release categories based on release characteristics such as 32 magnitude and timing of release. Binning both the plant damage states, and source terms 33 reduces the total number of accident progression and consequence simulations that are 34 required. In summing the CDF and LERF/large release frequency, the analyst should consider 35 all significant accident sequences. Significant accident sequences, as defined in Regulatory 36 Guide 1.200, An Approach for Determining the Technical Adequacy of Probabilistic Risk 37 Assessment Results for Risk-Informed Activities (current version), are those that, when ranked, 38 compose 95 percent of the CDF or LERF, or that individually contribute more than 1 percent to 39 the CDF.

40 41 In practice, the computation of change in the frequency of CDF and release categories for both 42 the standard analysis and the major effort uses PRA software, such as Systems Analysis 43 Program for Hands-On Integrated Reliability Evaluations (SAPHIRE), are discussed in 44 Enclosure H-1, Description of Analytical Tools and Capabilities, to this appendix.

45 46 H.3.3.3 Quantification of Change in Consequences 47 48 Many analyses may assume that new consequence evaluations will not be needed. If the 49 change in risk can be captured through a change in accident sequence frequencies only, then 50 the overall risk equation can use the existing public health and economic consequence H-15 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 assessments associated with those accident sequences. This assumption is embedded when 2 existing population dose and offsite economic consequence multipliers (e.g., population dose 3 factors in Section 5.3.2.1 of this NUREG) are used for severe accident sequences. However, if 4 a proposed action affects an accidents conditional consequences, then the risk quantification 5 approach should explicitly account for the change in conditional consequences, as noted at the 6 end of Section 5.3.2.1.1 of this NUREG. If the existing PRA model does not adequately capture 7 the change in risk associated with the proposed change, then the PRA model should be revised 8 to support the analysis.

9 10 Regulatory analyses involving large light-water reactors historically have been estimated using a 11 50-mile radius from the site (see Section 5.2.1 of this NUREG). The analyst chooses the 12 distance based on the potentially affected area (e.g., where offsite population dose and offsite 13 property damage is incurred). Offsite consequences for other distances 6 have been considered 14 in recent detailed analyses where individual plants with site-specific information were evaluated.

15 Section H.5 and Enclosures H-4 through H-6 to this appendix discuss examples. For small 16 modular reactors and advanced reactors, the radius should be chosen based on design-specific 17 details, site characteristics, and precedents.

18 19 H.3.3.4 Identification and Estimation of Affected Parameters 20 21 An action may affect accident frequencies only, accident consequences only, or both accident 22 frequencies and consequences. Actions that may change existing PRA model structures 23 (e.g., by adding or removing events in an event tree or changing consequences of existing 24 accident sequences) will require additional analysis steps compared to actions that affect only 25 the relative frequencies of existing accident sequences and associated consequences.

26 27 If appropriate PRA models are available, these can be used to identify the affected parameters.

28 For example, all NPP PRA studies include accident sequences involving loss of emergency 29 alternating current power. If the minimal cut sets used in the analytical modeling of these 30 sequences contain parameters appropriate to an action related to loss of emergency alternating 31 current power, then these PRA studies would be appropriate for use in the analysis. In this 32 case, the analyst can readily identify the affected parameters and their estimated values.

33 34 Within the scope of an analysis, the identification of affected parameters may require more than 35 the direct use of existing PRA models. Existing studies may need to be modified. The effort 36 may involve (1) performing an expanded or independent analysis of the accident sequences 37 associated with an action, using previous studies only as a guideline, or (2) combining several 38 existing PRA studies to form a composite study more applicable to a generic action. Care 39 should be taken to ensure that assumptions, modeling, and uncertainty characterization are 40 appropriate and valid to support decisionmaking.

41 42 Assuming the analyst has identified affected parameters, the next step is to estimate the 43 base- and adjusted-case values of the affected parameters, which are then used to estimate the 44 base- and adjusted-case total accident sequence frequencies and associated consequences.

45 The sum of the differences between the base- and adjusted-cases is the change in accident 46 frequency, the consequence resulting from the action, or both.

47 48 In some cases, additional modeling is required, where identification of affected parameters 6 The analyst should also consider the capabilities and range of validity of analytical tools when selecting these distances.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-16

1 requires the type of analysis associated with a much greater level of detail and, most likely, a 2 significantly expanded scope. NRC programs related to unresolved generic safety issues for 3 power reactors offer examples of where major efforts were required in the past. Such programs 4 tend to be multiyear tasks. The expected level of detail and quality of analysis should be 5 consistent with current standard practice and may entail peer review.

H-17 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.4 GRADED APPROACH TO ANALYSIS 2

3 As with most areas of NRCs regulation (e.g., NRCs Strategic Plan: Fiscal Years 2018-2022 4 [NRC, 2018a]), staff are expected to take a risk-informed approach to severe accident risk 5 analyses supporting regulatory analyses. NRCs Office of Nuclear Reactor Regulation Office 6 Instruction LIC-504, Integrated Risk-Informed Decision-Making Process for Emergent Issues 7 (NRC, 2014i) describes different levels of approach, namely a graded approach, to using risk 8 information that, while tailored to decision-making for emergent issues, is conceptually 9 appropriate to the use of risk information in regulatory analyses too. A graded approach is one 10 where the level of rigor applied depends on the importance, e.g., risk significance and 11 applicability (see for example, discussion in Management Directive 6.4, Generic Issues 12 Program [NRC, 2015g]). As noted in LIC-504, Regulatory Guide 1.174, and NUREG-1855, 13 Guidance on the Treatment of Uncertainties Associated with PRAs in Risk-Informed Decision 14 Making (NRC, 2017a), it is particularly important to assess uncertainties in the risk analyses 15 and understand how uncertainties may affect the comparison of risk measures with decision 16 criteria.

17 18 In some cases, an initial screening-type analysis may be sufficient to disposition the evaluation 19 of a potential regulatory action. For example, if it is necessary to show a substantial safety 20 benefit and possibly to get an initial assessment of whether a potential regulatory change may 21 be cost-beneficial, existing compilations of risk information may be sufficient to make the 22 determination (this would be analogous to answering yes to the question in NRCs LIC-504 23 Section 4.2.2, Is the Issue Clearly of Low Safety Significance? [NRC.2014i]). For such an 24 approach, the potential benefits should be maximized, and (if pursuing an initial cost-benefit 25 assessment) the potential costs minimized, to ensure that a potentially warranted action is not 26 unduly screened out. Furthermore, uncertainty in these screening or bounding-type analyses 27 and its potential impact should be considered.

28 29 In the absence of a new major-effort analysis, existing risk information would be used, e.g., by 30 selecting the maximum CDF for the class of affected plants and the highest known conditional 31 consequences within the class of affected plants. Current CDFs at the time of an analysis are 32 available, such as in the information sets used to create Figures H-2 and H-3 above. While the 33 conditional consequences may be harder to find, several sources of information (discussed in 34 Section H.5.2) exist and could provide the needed estimates. The highest conditional 35 consequences for a class of plants typically will be tied to the highest population sites. Both 36 10-mile- and 50-mile-radius populations should be considered for large light-water reactors; for 37 small modular reactors and advanced reactors, the radius could be chosen based on 38 design-specific details and precedence (such as EPZ and Protective Action Guides [PAGs]).

39 The joint consideration of a sites meteorological profile, population distribution, and licensed 40 thermal power (since total radiological releases for a given accident are expected to scale with 41 core power) is important. The offsite populations residing within 50 miles of the operating NPPs 42 in the United States varied from 180,000 to 17 million, according to the 2000 and 2010 43 censuses (NRC, 1996 and supplements). As of 2019, the licensed thermal power for individual 44 large light-water reactors in the United States varied from 1,700 megawatts thermal (MWt) to 45 4,400 MWt (NRC, 2019b).

46 47 As discussed in Section 5.3.2 of this NUREG, the estimation of the avoided public health effects 48 and avoided offsite economic consequences is calculated from current risk information from 49 existing studies. The avoided consequences are computed by multiplying the change in 50 frequency of each significant release category by its consequence metrics and then applying a NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-18

1 summation over all affected release categories. This approach should only be used if the staff 2 deems that existing risk studies adequately capture the accident scenarios, associated 3 frequencies and consequences, for the issue under consideration.

4 5 At the simplest level, the analysis assumes values of affected parameters are readily available 6 or can be derived easily. Sources of data that are readily accessible include existing PRA 7 studies, which provide parameter values in forms appropriate for accident frequency 8 calculations (e.g., frequencies for initiators and unavailability or demand failure probabilities for 9 subsequent failures of systems, structures, and components).

10 11 After identifying base and adjusted-case values for the parameters in the plant-risk equation 12 that are affected by the proposed regulatory action (see Section 5.3.2 of this NUREG), the 13 analyst calculates the change in accident frequency as the sum of the differences between the 14 base- and adjusted-case values for all affected accident sequences.

15 16 Uncertainties are prevalent in any risk assessment and should be addressed (see Section 17 H.6.3.1 for a discussion on different kinds of uncertainties). For example, an error factor on the 18 best estimate of the reduction in total accident frequency may be used to estimate high and low 19 values for the sensitivity calculations in the analysis for power reactor facilities. Past analyses 20 have used error factors of 5-10 or more, depending on the events analyzed 7. Error factors from 21 the specific risk assessment being used, if available, or knowledge of typical error factors from 22 current analogous risk assessments, should be employed.

23 24 An analyst who is unable to identify affected parameters for an action can estimate changes in 25 accident frequency using professional judgment. Expert opinion also plays a prime role in 26 estimating adjusted-case parameter values. Typically, existing data are applied to yield 27 base-case values, leaving only engineering judgment for arriving at adjusted-case values.

28 Reaching consensus among multiple experts can increase confidence, and the magnitudes of 29 parameter values normally encountered in PRA studies can serve as rough guidelines.

30 31 At a more detailed level, but still within the scope of a standard analysis, the analyst may 32 conduct reasonably detailed statistical modeling or extensive data compilation when values of 33 affected parameters are not readily available. While existing PRA studies may provide some 34 data for use in statistical modeling, the level of detail required normally would be greater than 35 they could provide. Statistical modeling may use stochastic simulation methods and involve 36 statistical analysis techniques using extensive data.

37 38 NUREG/CR-2800, Guidelines for Nuclear Power Plant Safety Issue Prioritization Information 39 Development, issued September 1983 (NRC, 1983b), discusses the calculation of change in 40 core melt accident frequency for power reactors, and provides examples. Such calculations are 41 typical for a standard cost-benefit analysis. A useful reference is Nuclear Energy Institute 42 (NEI)-05-01, Revision A, Severe Accident Mitigation Alternatives (SAMA) Analysis Guidance 43 Document, issued November 2005 (NEI, 2005), because SAMA analyses follow a similar 44 process to that of regulatory and cost-benefit analyses. A SAMA analysis includes searches for 45 potential generic industry and plant-specific improvements to address important risk 46 contributors, and cost-benefit analyses to evaluate these potential improvements.

47 7 See for example: https://nrcoe.inl.gov/resultsdb/publicdocs/AvgPerf/ComponentUR2015.pdf H-19 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.4.1. Example of Approach 2

3 The staff analysis summarized in Enclosure H-3, Summary of Detailed Analyses for 4 SECY-12-0157, Consideration of Additional Requirements for Containment Venting Systems 5 for Boiling Water Reactors with Mark I and Mark II Containments, provides an example of a 6 practical modern approach to what was historically called a standard analysis. To evaluate 7 the potential risk reduction benefit of the proposed action, the staff first reviewed insights from 8 available risk studies. These sources of risk information included (1) the IPEs completed in 9 response to Generic Letter 88-20 (NRC, 1988a, NRC, 1997a), (2) applicable risk-informed 10 license amendment requests, which in this case were the requests for integrated leak rate 11 testing (ILRT) (see Table 2 of NRC, 2012h), and (3) SAMA analyses submitted with license 12 renewal applications for operating NPPs (NRC, 1996, and supplements). The ILRT license 13 amendment requests were considered because they estimated post-core-damage containment-14 related risk benefits that informed the evaluation of potential benefits of installing containment 15 venting systems. The staff collected the following information from these sources:

16 17

Identification of the conditional containment failure probabilities from the class of plants 18 under consideration (e.g., boiling-water reactors [BWRs] with Mark I and Mark II 19 containments), for base-case conditions in the IPEs and ILRTs, as well as sensitivity to 20 extended ILRT intervals 21 22

Identification of dominant contributors to early containment failure 23 24

Evaluation of whether past SAMA analyses considered filtered severe accident venting, 25 and if so, whether they found it to be a potentially cost-beneficial plant improvement at 26 the time of the license renewal application 27 28 This evaluation of available risk insights contributed to the technical approach for evaluating 29 potential benefits by helping the staff to develop the branches on the event tree for sequence 30 evaluation and benefit quantification (see Enclosure H-3 to this NUREG for more details of this 31 analysis).

32 33 A safety goal evaluation is required as part of a regulatory analysis in which regulatory 34 alternatives are analyzed to determine whether they are generic safety enhancement backfits 35 subject to the substantial additional protection standard. To perform the safety goal evaluation, 36 the staff should analyze the regulatory alternatives to directly compare the potential safety 37 benefits to the QHOs for average individual early fatality risk and average individual latent 38 cancer fatality risk described in the Commissions Safety Goal Policy Statement 8 (NRC, 1986).

39 To determine the relative costs and benefits, the analyst should compare each of the 40 alternatives to the regulatory baseline.

8 In 1986, the NRC published the Safety Goal Policy Statement, whose objective was to, establish goals that broadly define acceptable level of radiological risk to the public from nuclear power plant operation (NRC, 1986).

This policy stated two qualitative safety goals, supported by two quantitative objectives which are commonly called QHOs: (1) the risk to an average individual in the vicinity (1 mile) of a nuclear power plant of prompt fatalities that might result from reactor accidents should not exceed one-tenth of one percent (0.1 percent) of the sum of prompt fatality risks resulting from other accidents to which members of the U.S. population are generally exposed; and (2) the risk to the population in the area (within 10 miles) near a nuclear power plant of cancer fatalities that might result from nuclear power plant operation should not exceed one-tenth of one percent (0.1 percent) of the sum of cancer fatality risks resulting from all other causes. Since the QHOs are tied to the prompt fatality risks and cancer fatality risks from all other causes in the U.S., the actual QHOs can change over time.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-20

1 2 A successful strategy used in the past for the safety goal evaluation is to employ a high-level 3 and conservatively high estimate to maximize the potential benefit of a regulatory alternative for 4 comparison to the regulatory baseline, to determine whether an alternative may meet the 5 substantial safety benefit threshold. For example, in the Containment Protection and Release 6 Reduction (CPRR) regulatory analysis described in Enclosure H-4 to this appendix, the staff 7 performed a screening analysis for the average individual latent cancer fatality risk QHO for the 8 relevant plantsall U.S. BWRs with Mark I containments (a total of 22 units at 15 sites) and 9 Mark II containments (a total of eight units at five sites). For this screening analysis, the staff 10 developed a conservatively high estimate of the frequency-weighted average of an individual 11 latent cancer fatality risk within 10 miles of the plant using the following parameter values:

12 13

An extended loss of alternating current power (ELAP) 9 frequency value of 7x10-5 per 14 reactor-yearwhich represented the highest value among all BWRs with Mark I and 15 Mark II containments 16 17

A success probability for flexible coping strategies (FLEX) equipment of 0.6 per 18 demandwhich assumed the implementation of FLEX will successfully mitigate an 19 accident involving an ELAP 6 out of 10 times 20 21

A conditional average individual latent cancer fatality risk of 2x10-3 per eventwhich 22 represented the highest value among all BWRs with Mark I and Mark II containments 23 from the detailed analyses 24 25 These assumed parameter values resulted in a conservatively high estimate of a 26 frequency-weighted individual latent cancer fatality risk within 10 miles of approximately 27 7x10-8 per reactor-year (labelled as High-Level Conservative Estimate in Figure H-4), which is 28 greater than an order of magnitude less than the QHO for an average individual latent cancer 29 fatality risk of approximately 2x10-6 per reactor-year. This conservatively high estimate did not 30 take credit for any of the accident strategies and capabilities described in the 20 CPRR 31 alternatives and subalternatives. Figure H-4 shows the incremental benefit (in terms of 32 individual latent cancer fatality risk on the y-axis) for each alternative on the x-axis-33 subalternatives within Alternatives 2 to 4 compared to the status quo, Alternative 1.

34 35 Because the conditional early fatality risk was essentially zero, a comparable analysis for the 36 early fatality QHO was not needed.

37 9 An ELAP is defined as an SBO that lasts longer than the SBO coping duration specified in 10 CFR 50.63, Loss of all alternating current power.

H-21 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-4 Uncertainty in Average Individual Latent Cancer Fatality Risk (0-10 miles) in 3 the 2015 Containment Protection and Release Reduction Regulatory 4 Analysis 5 (Source: SECY-15-0085, Enclosure, Figure 3-3) 6 7 H.4.2. Sources of Information 8

9 As noted in the Background section above, historically, NUREG-1150, Severe Accident Risks:

10 An Assessment for Five U.S. Nuclear Power Plants, issued December 1990 (NRC, 1990b),

11 and supplementary studies based on NUREG-1150, were the main sources of information for 12 the NRCs typical regulatory analyses. The analyst should consult the SPAR Program owner to 13 collect the most current risk information and insights at the time of a new analysis. For 14 example, the NRC maintains SPAR models for use in the Reactor Oversight Process and other 15 risk-informed regulatory activities, as noted in Section H.3.3.1 and discussed further in 16 Enclosure H-1. Risk-informed applications and SAMA analyses are other examples of sources 17 of information, as discussed further below.

18 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-22

1 Risk-informed license amendment requests 10 cover a range of plant and risk scenarios that 2 should be consulted according to the risk scope under consideration. The 10 CFR 50.54(f) 3 letter responses are another source of information for a variety of plant and accident types. For 4 example, in response to the lessons learned from the Fukushima Dai-ichi accident, the NRC 5 issued a 10 CFR 50.54(f) letter (NRC, 2012i) to all operating NPP licensees to reevaluate the 6 seismic and flooding hazards at their sites using updated seismic and flood hazard information 7 and present-day regulatory guidance and methodologies and, if necessary, to request that they 8 perform a risk evaluation. The responses to the letter provide post-2012 seismic CDF and 9 seismic LERF information for operating NPPs. 11 10 11 The SAMA analyses may provide useful information since SAMA analyses (1) cover all nuclear 12 steam supply systems (NSSS) and containment types for the operating fleet of NPPs (see 13 Table H-2), as well as new reactors under construction (e.g., SAMA and SAMDA analyses for 14 the advanced passive 1000 [AP1000]), and (2) have been evaluated for the known risk profile 15 (e.g., different accident initiators and scenarios) for each subject plant at the time of analysis.

16 The SAMA analyses report on the rank of contributors to CDF (see the example in Table H-3),

17 the rank of contributors to LERF (occasionally), the rank of contributors of different release 18 categories or containment release modes to population dose (see example in Table H-4), and 19 the maximum attainable benefit in terms of the offsite dose and offsite economic cost risks 20 (within a 50-mile radius from the plant) that would be saved if all potential accidents could be 21 eliminated at the plant. These analyses 12 are documented in license applications and in the 22 staffs environmental evaluations. 13 As noted in the main body Section 2, the SAMA analyses 23 documented in the NUREG-1437 supplements report quantitative internal events CDFs, and 24 external events multipliers in the range of 1.2 to 12, with an average value of 3.2 (based on 25 51 of the 57 supplements published between 1999 and 2016 that reported external events 26 multipliers for 82 individual reactors). This means that the total CDF was estimated to be 1.2 to 27 12 times the internal events CDF, with an average value of 3.2 times the internal events CDF.

28 Additional SAMA analyses have been performed for design certifications and combined license 29 new reactor reviews. 14 When using data from SAMA analyses, the analyst should be aware 30 that the agency undertakes SAMA analyses to meet NEPAs hard look requirement; as a 31 result, some aspects of SAMA analyses may require further consideration before the agency 32 relies on them to meet its obligations under the Atomic Energy Act of 1954, as amended.

33 34 Table H-2 Reactors with Published SAMA Analyses Licensed NUREG-1437a, b Containment Thermal Supplement Number NSSS Type Plant Name Type Power (unless noted (MWt) otherwise)

Arkansas 1 2568 3 Oconee 1 2568 2 B&W Lowered Loop Oconee 2 2568 2 Dry, Ambient Oconee 3 2568 2 B&W Raised Loop Davis-Besse 2817 52 CE Arkansas 2 3026 19 10 For example, see risk-informed technical specification changes discussed here:

https://www.nrc.gov/reactors/operating/licensing/techspecs/risk-management-tech-specifications.html 11 https://www.nrc.gov/reactors/operating/ops-experience/japan-dashboard/seismic-reevaluations.html 12 https://www.nrc.gov/reactors/operating/licensing/renewal/applications.html contains links to all NPP license renewal applications and the NRCs reviews.

13 https://www.nrc.gov/reading-rm/doc-collections/nuregs/staff/sr1437/

14 https://www.nrc.gov/reactors/new-reactors.html H-23 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

Licensed NUREG-1437a, b Containment Thermal Supplement Number NSSS Type Plant Name Type Power (unless noted (MWt) otherwise)

Calvert Cliffs 1 2737 1 Calvert Cliffs 2 2737 1 Millstone 2 2700 22 Palisades 2565 27 Saint Lucie 1 3020 11 Saint Lucie 2 3020 11 Waterford 3 3716 59 Palo Verde 1 3990 43 Large Dry, CE 80 Palo Verde 2 3990 43 Ambient Palo Verde 3 3990 43 GE 2 Nine Mile Point 1 1850 24 Dresden 2 2957 17 Dresden 3 2957 17 GE 3 Monticello 2004 26 Quad Cities 1 2957 16 Quad Cities 2 2957 16 Browns Ferry 1 3952 21 Browns Ferry 2 3952 21 Browns Ferry 3 3952 21 Brunswick 1 2923 25 Mark I Brunswick 2 2923 25 Cooper 2419 41 Duane Arnold 1912 42 GE 4 Fermi 2 3486 56 FitzPatrick 2536 31 Hatch 1 2804 4 Hatch 2 2804 4 Hope Creek 1 2902 45 Peach Bottom 2 4016 10 Peach Bottom 3 4016 10 Limerick 1 3515 49 Limerick 2 3515 49 GE 4 Susquehanna 1 3952 35 Susquehanna 2 3952 35 Mark II Columbia 3544 47 La Salle 1 3546 57 GE 5 La Salle 2 3546 57 Nine Mile Point 2 3988 24 Grand Gulf 1 4408 50 Mark III GE 6 River Bend 1 3091 58 Ginna 1775 14 Dry, Ambient Westinghouse 2-loop Point Beach 1 1800 23 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-24

Licensed NUREG-1437a, b Containment Thermal Supplement Number NSSS Type Plant Name Type Power (unless noted (MWt) otherwise)

Point Beach 2 1800 23 Prairie Island 1 1677 39 Prairie Island 2 1677 39 Beaver Valley 1 2900 36 Beaver Valley 2 2900 36 Dry, North Anna 1 2940 7 Westinghouse 3-loop Subatmospheric North Anna 2 2940 7 Surry 1 2587 6 Surry 2 2587 6 Farley 1 2775 18 Farley 2 2775 18 Harris 1 2948 33 Dry, Ambient Westinghouse 3-loop Robinson 2 2339 13 Summer 2900 15 Turkey Point 3 2644 5 Turkey Point 4 2644 5 Braidwood 1 3645 55 Braidwood 2 3645 55 Byron 1 3645 54 Byron 2 3645 54 Callaway 3565 51 Indian Point 2 3216 38 Indian Point 3 3216 38 Millstone 3 3650 22 Dry, Ambient Westinghouse 4-Loop Salem 1 3459 45 Salem 2 3459 45 Seabrook 1 3648 46 South Texas 1 3853 48 South Texas 2 3853 48 Vogtle 1 3626 34 Vogtle 2 3626 34 Wolf Creek 1 3565 32 Catawba 1 3469 9 Catawba 2 3411 9 D.C. Cook 1 3304 20 D.C. Cook 2 3468 20 Ice Condenser Westinghouse 4-Loop McGuire 1 3411 8 McGuire 2 3411 8 Sequoyah 1 3455 53 Sequoyah 2 3455 53 Watts Bar 2 3411 NUREG-0498, Supp. 2c H-25 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

Licensed NUREG-1437a, b Containment Thermal Supplement Number NSSS Type Plant Name Type Power (unless noted (MWt) otherwise)

Vogtle 3d NUREG-1872d AP1000 Westinghouse 2-Loop Vogtle 4d NUREG-1872d 1 a Information current as of 2019 2 b NUREG-1437 and supplements are available at: https://www.nrc.gov/reading-rm/doc-3 collections/nuregs/staff/sr1437/

4 c NRC, 2013i.

5 d Under construction; NUREG-1872, Final Environmental Impact Statement for an Early Site Permit (ESP) at the 6 Vogtle ESP Electric Generating Plant Site, issued August 2008 (NRC, 2008).

7 8 Table H-3 Salem Nuclear Generating Station Core Damage Frequency for Internal 9 Events at Power CDF1 % Contribution Initiating Event (per year) to CDF2 Loss of Control Area Ventilation 1.8 x 10-5 37 Loss of Offsite Power (LOOP) 8.1 x 10-6 17 Loss of Service Water 6.6 x 10-6 14 Internal Floods 4.5 x 10-6 9 Transients 4.0 x 10-6 8 Steam Generator Tube Rupture (SGTR) 2.7 x 10-6 6 Loss of Component Cooling Water (CCW) 1.0 x 10-6 2 Anticipated Transient Without Scram (ATWS) 7.4 x 10-7 2 Loss of 125V DC Bus A 6.9 x 10-7 2 Others (less than 1 percent each)3 1.8 x 10-6 4 Total CDF (internal events at power)4 4.8 x 10-5 100 10 1 Calculated from Fussel-Vesely risk reduction worth (RRW) provided in response to NRC staff RAI 1.e 11 (PSEG, 2010a).

12 2 Based on internal events CDF contribution and total internal events CDF.

13 3 CDF value derived as the difference between the total Internal Events CDF and the sum of the individual internal 14 events CDFs calculated from RRW.

15 4 The results only covers a fraction of the total plant risk profile, so their usefulness for regulatory decision-making 16 may be limited for situations where the analysis is evaluating changes involving not at power or external events.

17 (Source: NUREG-1437, Supplement 45, Table F-1) 18 19 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-26

1 Table H-4 Salem Nuclear Generating Station Breakdown of Population Dose by 2 Containment Release Mode Population Dose Percent Containment Release Mode (Person-Rem1 Per Year) Contribution2 Containment overpressure (Late) 42.9 55 Steam generator rupture 31.9 41 Containment isolation failure 2.3 3 Containment intact 0.2 <1 Interfacing system Loss-of-Coolant Accident 0.6 <1 (LOCA)

Catastrophic isolation failure 0.4 <1 Basemat melt-through (late) Negligible Negligible Total3,4 78.2 100 3 1 One person-rem = 0.01 person-Sv 4 2 Derived from Table E.3-7 of the ER (PSEG 2009).

5 3 Column totals may be different due to rounding.

6 4 The results only covers a fraction of the total plant risk profile, so their usefulness for regulatory decision making 7 may be limited for situations where the analysis is evaluating changes involving not at power or external events.

8 (Source: NUREG-1437, Supplement 45, Table F-2) 9 10 The State-of-the-Art Reactor Consequence Analyses (SOARCA), (see Enclosure H-2 to this 11 appendix) is another source of information for potential offsite public health consequences 12 within the scope of the severe accident scenarios studied for three operating reactor types in 13 the United States. 15 SOARCA analyses, including uncertainty analyses, were conducted for 14 short-term and long-term SBO accidents at a BWR with a Mark I containment in Pennsylvania; 15 a three-loop Westinghouse NSSS pressurized-water reactor (PWR) with a subatmospheric 16 large, dry containment in Virginia; and a four-loop Westinghouse NSSS PWR with an ice 17 condenser containment in Tennessee. Deterministic analyses were also conducted for an 18 interfacing systems loss-of-coolant accident at the PWR with a large, dry containment.

19 Consequence results were reported as individual latent cancer risks and individual early fatality 20 risks for different radial rings out to 50 miles from the site. The SOARCA studies focused on 21 accident progression, source term, and conditional consequences should the postulated 22 accidents occur. The project did not include within its scope new work to calculate the 23 frequencies associated with the postulated severe accidents. And just like information from 24 modern plant-specific risk-informed license amendment requests, or plant-specific SAMA 25 analyses, the SOARCA studies were conducted for specific reactor types and sites.

26 15 The SOARCA analyses are documented in a series of NUREG and NUREG/CR reports (NRC, 2012a; NRC, 2012j; NRC, 2013a; NRC, 2013b; NRC, 2014a; NRC, 2014b; NRC, 2016b; NRC, 2019a; NRC, 2020).

H-27 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.5 MAJOR-EFFORT ANALYSIS 2

3 When additional rigor is required, a major-effort analysis is performed. Enclosures H-4 4 through H-6 to this appendix summarize the major-effort regulatory analyses that the staff 5 completed in the 2013 to 2015 timeframe. This section summarizes approaches and 6 considerations for the common technical elements in a major-effort regulatory analysis: accident 7 sequence analysis, accident progression (Level 2 PRA) analysis, and offsite consequence 8 (Level 3 PRA) analysis.

9 10 H.5.1 Accident Sequence Analysis 11 12 A major-effort analysis should begin with an accident sequence analysis. The analyst should 13 consider the following factors during the development of the technical approach for selecting the 14 relevant set of accident sequences:

15 16

The risk evaluation should provide risk metrics for all regulatory analysis subalternatives 17 and do so according to the approved scope, schedule, and allocated resources.

18 19

Consistent with the NRCs regulatory analysis guidelines, the risk evaluation should 20 provide fleet-average risk estimates. Therefore, the technical approach should consider 21 the impacts of plant-to-plant variability (for example, see Section H.6.2.2).

22 23

The staff should leverage existing relevant sources of accident sequence information 24 and develop new information where required.

25 26

The analyst should develop CDETs to (1) model the impact of equipment failures and 27 operator actions occurring before core damage that affect severe accident progression 28 and the probability that regulatory alternatives are successfully implemented, (2) match 29 the initial and boundary conditions used in the thermal-hydraulic simulation of severe 30 accidents in MELCOR, and (3) consider mitigating strategies for beyond-design-basis 31 external events, as applicable.

32 33

The analyst should develop APETs to model regulatory alternatives.

34 35 Enclosures H-3 through H-6 to this appendix include discussions of the accident sequence 36 analyses for three detailed regulatory analyses. As discussed in Enclosure H-4 to this 37 appendix, analysts successfully used a modular approach to develop the CDETs and APETs, 38 as shown in Figure H-5. This modeling approach streamlined the development of risk estimates 39 for the CPRR technical basis rulemaking and provides a good example for future detailed 40 analyses. Enclosure H-1 to this appendix describes the NRC-sponsored software, SAPHIRE.

41 SAPHIRE can be used for accident sequence modeling with CDETs and APETs and frequency 42 analysis.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-28

1 2 Figure H-5 Modular Approach to Event Tree Development in CPRR Analysis 3 (Source: NUREG-2206, issued March 2018, Figure 2-1) 4 H-29 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.5.2 Severe Accident Progression Analysis 2

3 The next step of a major-effort analysis is to complete a severe accident progression and 4 source term analysis, analogous to a Level 2 PRA. The objective of the severe accident 5 progression analysis is to generate a technical basis quantitatively characterizing thermal and 6 mechanical challenges to engineered barriers to fission product release to the environment.

7 This analysis provides a chronology of postulated accidents resulting in significant damage to 8 reactor fuel and generates quantitative estimates of a radioactive material release to the 9 environment. The staff has used the MELCOR code 16 (Humphries et al., 2015), described in 10 Enclosure H-1 to this appendix, to model accident progression and fission product release 11 estimates for each of the selected accident scenarios in the detailed analyses.

12 13 The two broad purposes for conducting MELCOR calculations are: (1) to evaluate reactor 14 systems and containment thermal-hydraulics under severe accident conditions, and (2) to 15 assess the timing and magnitude of fission products released to the environment. Three 16 outputsthe containment temperature and pressure signatures, along with hydrogen 17 distribution through the containment and reactor buildingprovide information to assess the 18 status of reactor plant and containment integrity under varying postulated conditions. This 19 information may provide the basis for investigating other regulatory subalternatives. Analysts 20 use the timing and magnitude of fission product release information to characterize the source 21 terms in the consequence analysis described in Section H.5.3.

22 23 The MELCOR calculations are deterministic in nature and simulate different possible outcomes 24 or plant damage states, given the initial conditions that are specified in the accident sequence 25 analysis. The analyst should choose representative plant models based on the requirements of 26 the regulatory analysis (e.g., reflective of the relevant class(es) of NSSSs, containments, and 27 operational features). For efficiency, the representative MELCOR plant models can use existing 28 input decks developed for recent studies when available and relevant. For example, the 29 regulatory analyses discussed in Enclosures H-3 and H-4 to this appendix started with the 30 SOARCA Peach Bottom Atomic Power Station input deck for Mark I containments.

31 32 H.5.2.1 Sources of Information 33 34 NUREG/CR-7008, MELCOR Best Practices as Applied in the SOARCA Project, issued 35 August 2014 (NRC, 2014a), describes the best practices in modeling approach and parameter 36 selections that support the best estimate analyses in the 2012 SOARCA project, for a General 37 Electric BWR with a Mark I containment and a Westinghouse 3-loop PWR with a large, dry, 38 subatmospheric containment. The input models should follow the guidance of 39 NUREG/CR-7008, supplemented with updates and insights from the most recent MELCOR 40 analyses available (e.g., later SOARCA studies, such as NUREG/CR-7245, State-of-the-Art 41 Reactor Consequence Analyses (SOARCA) Project: Sequoyah Integrated Deterministic and 42 Uncertainty Analyses, issued 2019 (NRC, 2019a), for a Westinghouse 4-loop PWR with an ice 43 condenser containment, and NUREG/CR-7155, State-of-the-Art Reactor Consequence 44 Analyses Project: Uncertainty Analysis of the Unmitigated Long-Term Station Blackout of the 45 Peach Bottom Atomic Power Station, issued May 2016 (NRC, 2016b), and future studies, such 46 as the NRCs Site Level 3 PRA, 17 for a Westinghouse 4-loop PWR with a large, dry 47 containment).

48 16 http://melcor.sandia.gov/

17 https://www.nrc.gov/about-nrc/regulatory/research/level3-pra-project.html NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-30

1 Each operating NPP has an updated final safety analysis report that describes the facilitys 2 design bases and technical specifications and provides a safety analysis of each plant system 3 (10 CFR 50.34(b)). The updated final safety analysis report describes plant components and 4 containment features. The analyst can use this information to construct the MELCOR model.

5 6 IPEs provide information on the types of accidents that have a potential for occurring and the 7 location of failures. As previously discussed, each operating plant has one of these risk 8 analyses for internal events and many have IPEEEs.

9 10 Severe accident management guidelines are a source of information for characterizing operator 11 and plant response to severe accidents. These guidelines are developed by the utility and 12 provide guidance for operator actions in the event of a severe accident. These guidelines 13 contain strategies to stop or slow the progression of fuel damage, maintain containment, and 14 mitigate radiological releases.

15 16 H.5.2.2 MELCOR Modeling Approach 17 18 An accident progression analysis should be a collection of simulations of specific accident 19 sequences that is conducted to understand how a regulatory alternative affects the plant and 20 estimate the fission product release (source term) resulting from the accident sequence.

21 22 A MELCOR calculation matrix is developed to delineate runs evaluating each regulatory 23 analysis alternative, the various potential plant lineups, and the sensitivity analyses performed 24 for pre- and post-core damage mitigation measures. The calculations should clearly state the 25 initial and boundary conditions for the analysis and base the model nodalization on the specific 26 events that are being examined. The calculations should line up with APET and CDET 27 sequences in the accident sequence analysis.

28 29 Each accident sequence is binned into a release category that is represented by a MELCOR 30 source term. MelMACCS, which provides an interface between MELCOR and MACCS, can 31 read a MELCOR source term and provide the following data for each source term:

32 33

Time-dependent release fraction of chemical groups 18 34 35

Time-independent distribution by particle size diameter for 10 aerosol size bins 36 characterized by geometric mean diameters 37 38

Height of each MELCOR release pathway 39 40

Time-dependent data needed to estimate buoyant plume rise, including rate of release 41 of sensible heat (W), mass flow (kg/s), and gas density (kg/m3) 42 43 The MELCOR source terms become input for the next step of the analysis, which are used to 44 estimate the offsite consequences using the MELCOR Accident Consequence Code System 45 (MACCS) suite of codes.

46 18 For example, Noble Gases (Xe), Alkali Metals (Cs), Alkali Earths (Ba), Halogens (I), Chalcogens (Te), Platinoids (Ru), Early Transition Elements (Mo), Tetravalents (Ce), and Trivalents (La)) for each MELCOR release pathway H-31 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.5.3 Offsite Consequence Analysis 2

3 Similar to the MELCOR analysis, the consequences discussed here are conditional and do not 4 factor in the probability of release. The MACCS suite of codes 19 is the NRCs code system for 5 performing offsite consequence analyses for severe accident risk assessments. The NRC uses 6 MACCS to analyze hypothetical accident scenarios, and almost all parameters in the code may 7 be modified. This functionality provides substantial flexibility and allows for the characterization 8 of uncertainties. Enclosure H-1 to this appendix provides more details on the MACCS code and 9 its capabilities.

10 11 H.5.3.1 Sources of Information 12 13 Similar to the MELCOR SOARCA best practices, NUREG/CR-7009, MACCS Best Practices as 14 Applied in the State-of-the-Art Reactor Consequence Analyses (SOARCA) Project, issued 15 August 2014 (NRC, 2014b), describes the parameter selections that supported the 16 best-estimate MACCS analyses in the 2012 SOARCA study. The MACCS input models should 17 follow the guidance of NUREG/CR-7009, supplemented with updates and insights from the 18 most recent MACCS analyses (e.g., later SOARCA studies, such as NUREG/CR-7245 and 19 NUREG/CR-7155) and guidance. NUREG/CR-4551, Volume 2, Revision 1, Part 7, Evaluation 20 of Severe Accident Risks: Quantification of Major Input Parameters: MACCS Input, issued 21 December 1990 (NRC, 1990c), describes the development of shielding parameters for 22 NUREG-1150 is in greater detail.

23 24 H.5.3.2 MACCS Modeling Approach 25 26 There is considerable variation in site characteristics, such as population size and distribution, 27 land use, economic values, weather, and emergency response characteristics (e.g., road 28 networks, use of potassium iodide). Site-specific models historically have been developed for 29 plant and containment types and then adapted using a series of sensitivity calculations to 30 assess the potential impact of the site-specific parameters on the results. For efficiency, the 31 analyst can use existing MACCS input decks developed for recent studies when available and 32 relevant. For example, the regulatory analyses discussed in Enclosures H-3 and H-4 to this 33 appendix started with the SOARCA Peach Bottom MACCS input deck.

34 35 Source Term Characterization 36 37 The source terms developed from the severe accident progression analysis with similar release 38 fractions and release timing characteristics may be binned to reduce the number of MACCS 39 cases that must be run. The binning should be based, at a minimum, on cumulative cesium and 40 iodine release fractions and the warning times associated with each source term. Historically, 41 the cesium group has been the most important for long-term offsite consequences (e.g., latent 42 cancer fatality risk), and the iodine group has been the most important for early offsite 43 consequences (e.g., early fatality risk).

44 45 The MelMACCS code 20 in the MACCS suite of codes is used to create a MACCS input file to 46 represent the radiological source term developed using MELCOR. MelMACCS allows the user 47 19 https://maccs.sandia.gov/

20 https://maccs.sandia.gov/melmaccs.aspx NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-32

1 to associate the MELCOR mass values with an ORIGEN output to convert masses of chemical 2 classes to activities of individual radionuclides. In addition, the code needs the following data to 3 characterize each source term:

4 5

Radionuclide releases are divided into hourly segments to be consistent with the hourly 6 meteorological observations. If meteorological sampling is being used, the most 7 risk-significant plume should be identified to align the release with the weather data for 8 each weather bin. This is often taken to be the plume segment with the highest iodine 9 chemical group release fraction.

10 11

Building height and width are used to estimate the initial horizontal and vertical plume 12 dispersion caused by building wake effects.

13 14

Ground height in the MELCOR reference frame is used to adjust the MELCOR release 15 heights relative to grade.

16 17

Reference time, which is the difference between accident initiation time in MELCOR and 18 scram time. This value, which is used to properly account for decay and ingrowth of 19 radioactivity within MACCS, is usually zero but may be non-zero for some MELCOR 20 simulations.

21 22 Site and Meteorological Data 23 24 MACCS uses a polar grid to model the exposures to people, land contamination, and protective 25 actions of people and land. MACCS allows the user to choose 16, 32, 48, or 64 angular sectors 26 for grid division. The analyst should choose 64 angular sectors to provide the greatest 27 resolution. MACCS allows the user to divide the grid into a maximum of 35 radial rings, at 28 specified radii from the plant. The boundaries are selected to be consistent with certain areas of 29 interest. For example, for large LWR accidents, a radial boundary should be set at roughly 30 1 mile from the approximated site boundary to evaluate individual early fatalities for which the 31 NRCs early fatality QHO applies (NRC, 1986). This boundary is set at 10 miles to approximate 32 the plume exposure EPZ and latent fatality QHO, and at 50 miles to capture the majority of 33 radiological and economic consequences.

34 35 The SecPop preprocessor code in the MACCS suite of codes is typically used to generate 36 site-specific population and the economic data required for consequence calculations.

37 Population data should be scaled forward to the year of interest from the year of the census 38 data contained in SecPop using population growth data from the U.S. Census Bureau.

39 Additionally, the economic values contained in SecPop are from the U.S. Department of 40 Agriculture and U.S. Department of Commerce and should be scaled forward from the base 41 year data to the year of interest, using the consumer price index for all urban consumers.

42 43 The analyst should obtain raw weather data for the representative site from the site 44 meteorological towers for at least 2 full calendar years. Even though only 1 year of weather 45 data is necessary to complete the calculation, multiple years are beneficial for comparison to 46 ensure that the year selected is not anomalous (e.g., an abnormally dry or rainy year). The 47 inherent assumption in using historical data to quantify the consequences of a future event is 48 that future weather data will be statistically similar to historical data. The most complete year of 49 data should be chosen, and any missing data filled in by NRC meteorologists in accordance 50 with the U.S. Environmental Protection Agencys (EPAs) EPA-454/R-99-005, Meteorological H-33 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Monitoring Guidance for Regulatory Applications, issued February 2000 (EPA, 2000). The 2 methodology described in NUREG-0917, Nuclear Regulatory Commission Staff Computer 3 Programs for Use with Meteorological Data, issued July 1982 (NRC, 1982b), is used to perform 4 quality assurance evaluations of all meteorological data. In accordance with Regulatory 5 Guide 1.23, Meteorological Monitoring Programs for Nuclear Power Plants (current version),

6 the completeness of the recorded data (the data recovery rate) should be greater than 7 90 percent for the wind speed, wind direction, and atmospheric stability parameters. The 8 nonuniform bin sampling approach may be used to capture the effects of variable weather, 9 consistent with modeling best practices and recent consequence analyses.

10 11 Protective Action Modeling 12 13 EPA-400/R-17/001, PAG Manual: Protective Action Guides and Planning Guidance for 14 Radiological Incidents, issued January 2017, describes the emergency phase as the beginning 15 of a radiological incident when immediate decisions for effective use of protective actions are 16 required and must therefore be based primarily on the status of the radiological incident and the 17 prognosis for worsening conditions (EPA, 2017). Offsite response organization emergency 18 plans are required to include detailed evacuation plans for the plume exposure EPZ (NRC and 19 FEMA, 1980). Site-specific information should be obtained from offsite response organization 20 emergency response plans and the licensees evacuation time estimate (ETE) reports to 21 support the development of timelines for protective action implementation. The protective action 22 modeling assumptions have an important impact on offsite consequences.

23 24 MACCS input parameters related to evacuation modeling are taken primarily from the 25 site-specific ETE reports, which the licensee develops and updates under 10 CFR 50.47 (b)(10).

26 ETEs provide the time required to evacuate various sectors and distances within the EPZ for 27 transient and permanent residents, and these times are used to develop response timing and 28 travel speeds for evacuating cohorts 21 in MACCS.

29 30 Important information in an ETE report includes demographic and response data for four 31 population segments, which may be readily converted into cohorts, if appropriate. These 32 population segments are (1) permanent residents and transient population, 33 (2) transit-dependent permanent residents (e.g., people who do not have access to a vehicle or 34 are dependent upon help from outside the home to evacuate), (3) special facility residents 35 (e.g., people in nursing homes, assisted living centers, hospitals, jails, prisons), and (4) schools, 36 including all public and private educational facilities within the EPZ. In general, delineating the 37 population into more cohorts (beyond these four) allows greater fidelity in modeling the 38 emergency response of the public. In recent practice, the staff has further divided the ETE 39 cohorts into additional groups (e.g., in order to separate the 10 percent of the permanent 40 general population who may evacuate later than the other 90 percent of the general population).

41 42 The licensees ETE report typically includes about 10 scenarios that vary by season, day of the 43 week, time of day, and weather conditions, as well as other EPZ-specific situations such as 44 special events. The ETEs do not consider most external events and their impact on road 45 infrastructure, and it is important for the analyst to account for these impacts in the model. The 46 Sequoyah SOARCA analysis provides an example of how the impact of seismic events may be 47 considered in MACCS modeling (NRC, 2019a), if seismic events are important for the scope of 48 accidents under consideration.

21 As explained in more detail in Enclosure H-1 to this appendix, a cohort in MACCS is a group that is modeled as behaving similarly (e.g., evacuating at the same time and speed).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-34

1 2 In modeling the early phase relocation actions, the dose criteria to trigger the actions should be 3 consistent with the current EPA PAGs. In MACCS, emergency phase relocation is modeled 4 with two user-specified dose criteria to trigger the action and a relocation time for the population 5 affected by each dose. This modeling should consider site-specific features such as source 6 term, site information, and local demographics.

7 8 Although decisions about cleanup and reoccupation of affected areas would involve both 9 radiological and non-radiological considerations, it is customary in MACCS to use the dose 10 criteria for intermediate phase relocation as a surrogate for decisions about long-term 11 habitability. In determining the relocation and habitability dose criteria for the intermediate and 12 long-term phases, state-specific guidance for relocation following the early phase (as a 13 surrogate for decisions regarding habitability) should be followed when available. Absent 14 state-specific guidance, the analyst should use the EPA relocation PAGs.

15 H-35 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.6 SUPPLEMENTAL ANALYSES 2

3 Much like other parts of the regulatory analysis, the extent of supplemental analyses should be 4 commensurate with the complexity of the problem and associated uncertainties. At a minimum, 5 the analyst should identify important sources of uncertainty and influential assumptions and 6 evaluate their impacts on analysis outcomes. The results of these investigations should be 7 summarized in the report provided to decision makers, as discussed in Section 7.4, Risk 8 Integration Results and Key Insights.

9 10 H.6.1 Uncertainty Analyses 11 12 Appendix C, Treatment of Uncertainty, to this NUREG contains a general discussion of 13 uncertainties. The discussion below focuses on PRA uncertainties relevant to major-effort 14 analyses.

15 16 H.6.1.1 Uncertainties in PRA Models 17 18 When using PRA results as part of any regulatory decisionmaking process, it is important to 19 understand the types, sources, and potential impact of uncertainties associated with PRA 20 models and how to treat them in the decisionmaking process. Using PRA for regulatory 21 decisionmaking requires that the associated uncertainties and their implications be 22 characterized. For a major-effort analysis, the models and available information for projecting 23 severe accident consequences contain large uncertainties. The explicit identification and 24 quantification of sources of uncertainty of a consequence analysis are necessary to aid the 25 decisionmaker in understanding the results and the potential range of costs and benefits.

26 27 Although PRA models have several different sources of uncertainty, there are two principal 28 categories: aleatory and epistemic. Aleatory uncertainty arises from the random nature of the 29 basic events and phenomena (e.g., weather) modeled in PRAs. Because PRAs use 30 probabilistic distributions to estimate the frequencies or probabilities of these basic events, the 31 PRA model itself is an explicit model of the aleatory uncertainty. Similarly, the explicit modeling 32 of different weather conditions in the Level 3 portion of a PRA is a treatment of aleatory 33 uncertainty.

34 35 Epistemic uncertainties arise from incompleteness in the collective state of knowledge about 36 how to represent plant behavior in PRA models. These uncertainties relate to how well the PRA 37 model reflects the as-designed, as-built, as-operated plant and to how well it predicts the 38 response of the plant to various scenarios. Since these uncertainties can have a significant 39 impact on the interpretation and use of PRA results, it is important that they be appropriately 40 identified and characterized and that the analysis address important uncertainties. The 41 following three types of epistemic uncertainty are associated with PRA models:

42 43

Parameter Uncertainty: Parameter uncertainty relates to the uncertainty of input 44 parameter values. Probability distributions for the input parameters quantify the 45 frequencies or probabilities of basic events in the PRA logic model. Importantly, this 46 assumes that the selection of the probability distribution to model the likelihood of the 47 basic event is agreed upon; if uncertainty exists about this selection, it is more 48 appropriately considered model uncertainty.

49 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-36

1

Model Uncertainty: Model uncertainty arises from a lack of knowledge of physical 2 phenomena; failure modes related to the behavior of systems, structures, and 3 components under various conditions; or other phenomena modeled in a PRA (e.g., the 4 location and habits of members of the public in different exposure scenarios). This can 5 result in the use of different approaches to modeling certain aspects of the plant and 6 public response that can significantly impact the overall PRA model. Since uncertainty 7 exists about which approach is most appropriate, this leads to uncertainty in the PRA 8 results. Model uncertainty can also arise from uncertainty in the logic structure of the 9 PRA model or in the selection of the probability distribution used to model the likelihood 10 of the basic events in the PRA model. Sensitivity analyses typically address model 11 uncertainties to determine the sensitivity of the PRA results to alternative modeling 12 approaches. The ASME/ANS PRA standards (ASME/ANS, 2009, 2014, 2017) treat 13 Level 2 and Level 3 deterministic analysis uncertainties as model uncertainty, even 14 those that relate to input parameters in the MELCOR and MACCS consequence models.

15 16

Completeness Uncertainty: Completeness uncertainty arises from limitations in the 17 scope and completeness of the PRA model. These uncertainties can be addressed by 18 supplementing the PRA with additional analyses to demonstrate their impact is not 19 significant. The PRA model may have additional uncertainties from unknown risk 20 contributors, and defense-in-depth principles typically address them. See for example, 21 the discussion in NUREG/KM-0009, Historical Review and Observations of 22 Defense-in-Depth (NRC, 2016d). Section 3.1 of NUREG/KM-0009 notes the role of 23 defense-in-depth in a risk-informed regulatory framework to compensate for 24 uncertainties, in particular unquantified and unquantifiable uncertainties. Similar to the 25 framework laid out in Regulatory Guide 1.174 for risk-informed plant-specific changes to 26 licensing bases, consideration of completeness uncertainty means that a regulatory 27 analysis should not be overly reliant on precise risk quantification alone.

28 29 Although PRA cannot account for the unknown and identify all unexpected event scenarios, it 30 can (1) identify some originally unforeseen scenarios, (2) identify where some of the 31 uncertainties exist in plant design and operation, and (3) for some uncertainties, quantify the 32 extent of the uncertainty.

33 34 NUREG-1855, Guidance on the Treatment of Uncertainties Associated with PRAs in 35 Risk-Informed Decision Making, issued March 2017 (NRC, 2017a), contains useful general 36 guidance. NUREG-1855 focuses on sources of uncertainty associated with PRAs used to 37 estimate CDF and LERF, since these are the metrics for current risk-informed regulatory 38 decisions, such as risk-informed changes in the licensing basis. However, the principles and 39 broad guidance are more generally applicable to analyses that encompass additional Level 2 40 (accident progression and source terms) and Level 3 PRA (offsite consequences) information.

41 42 Several reference documents contain useful compendiums of sources of uncertainties in Level 2 43 and Level 3 PRA analyses. An Electrical Power Research Institute (EPRI) companion 44 document to NUREG-1855 lists sources of Level 2 analysis uncertainties identified at a 45 workshop of practitioners (EPRI, 2012). A joint Commission of European Communities expert 46 elicitation conducted in the 1990s identified sources of Level 3 analysis uncertainties (NRC and 47 Commission of European Communities, 1995). The uncertainties for non-site-specific 48 parameters from this expert elicitation were further mapped on to MACCS code input 49 parameters and documented for use in MACCS analyses in NUREG/CR-7161, Synthesis of 50 Distributions Representing Important Non-Site-Specific Parameters in Off-Site Consequence 51 Analyses, issued April 2013 (NRC, 2013c). The NRCs Site Level 3 PRA will have companion H-37 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 uncertainty documents for the Level 2 and Level 3 analyses. SOARCA uncertainty analyses are 2 documented for specific SBO scenarios at three NPPs (NRC, 2016b; NRC, 2019a; NRC, 2020).

3 The SOARCA analyses identified and propagated input parameter uncertainties through the 4 MELCOR and MACCS analyses and showed the effects of MELCOR uncertainties on accident 5 progression and radionuclide release metrics, as well as the combined effects of MELCOR and 6 MACCS uncertainties on offsite consequence metrics.

7 8 As noted above, NUREG-1855 and the ASME/ANS PRA standard categorize most uncertainties 9 embodied in the Level 2 and Level 3 portions of the PRA as model uncertainties. For the 10 purposes of consequence analyses supporting regulatory analysis, the outputs from MELCOR 11 and MACCS analyses become inputs to the regulatory and cost-benefit analyses as, for 12 example, individual early and latent cancer fatality risk (for QHO comparisons) and averted 13 population dose and offsite economic cost risks (for quantification of benefits to be compared 14 against implementation costs).

15 16 It is practical to treat the relevant PRA outputs as parameter uncertainties for cost-benefit 17 analysis. The regulatory bases documents for CPRR (NRC, 2018b) and filtered vents 18 (NRC, 2012h) contain examples of how to characterize and propagate uncertainties. Table 12 of 19 Enclosure 5 to the filtered vents analysis (NRC, 2012h) shows how the uncertainty was 20 described for all relevant inputs to the offsite risk analysis. The point estimates of the base-case 21 inputs such as CDF and MACCS consequences were specified to be the arithmetic means of 22 their respective distributions, and the distribution type and shape factors (such as the and 23 parameters for the beta distribution, or the error factor for the lognormal distribution), were 24 specified as well. The staff used a Monte Carlo process to propagate the uncertainty in each of 25 these inputs, as well as the uncertainty in the onsite cost elements. The results are shown for 26 each proposed modification and are presented as the distributions of averted cost (benefit) 27 elements for (1) public dose risk, (2) offsite economic cost risk, (3) onsite worker dose risk, and 28 (4) onsite cost risk. The CPRR risk analysis similarly assigned uncertainty distributions to the 29 following important inputs: the frequency of extended loss of alternating current power events, 30 the seismic hazard curves, the seismic fragility curves, random equipment failures, operator 31 actions, and consequences. The staff used a Monte Carlo process to propagate these 32 uncertainties and show the resulting distribution of individual latent cancer risk for the different 33 regulatory alternatives under consideration (NRC, 2015a, Figure 4-5), which is reproduced as 34 Figure H-6 as an illustrative example.

35 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-38

1 2 Figure H-6 Parametric Uncertainty Analysis Results for Individual Latent Cancer Fatality 3 Risk 4

5 H.6.2 Sensitivity Analyses and Plant-to-Plant Variability Analyses 6

7 Sensitivity analysis refers to studying the impact of one uncertain input on the analysis output, 8 without regard to relative probabilities. Uncertainty analysis typically evaluates the integrated 9 impact on the output of a collection of uncertain inputs that are assigned distributions of values, 10 resulting in a distribution of output results. In contrast, sensitivity analysis typically evaluates the 11 impact of one input on the output, and without consideration of the probability of different 12 outcomes. Two-way or joint sensitivity analyses similarly can study the impact of two or more 13 uncertain inputs on the outputs of interest.

14 15 Sensitivity analyses are typically used for particular categories of inputs. It is more appropriate 16 to use sensitivity, rather than uncertainty, analysis for input values subject to the 17 decisionmakers value choices; the dollar per person-rem conversion factor used in cost-benefit 18 analysis is one example. Inputs that depend on variability within the population of affected 19 plants is another example where sensitivity analysis is more appropriate.

20 H-39 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.6.2.1 Sensitivity Analyses 2

3 The regulatory analyses discussed in Enclosures H-3 through H-6 of this appendix used 4 sensitivity analyses to address the impact of different values for various inputs. For example, at 5 the time of the filtered vents analysis (Enclosure H-3), CPRR analysis (Enclosure H-4), and 6 expedited spent fuel transfer analysis (Enclosure H-6), the staff was in the process of updating 7 the dollar per person-rem conversion factor. The staff thus performed sensitivity analyses to 8 evaluate the impact on the results of increasing the dollar per person-rem conversion factor 9 from the 1995 value of $2,000 per person-rem to $4,000 per person-rem.

10 11 H.6.2.2 Plant-to-Plant Variability Analyses 12 13 Variability refers to the inherent heterogeneity of data in an assessment because of the diversity 14 of the regulated facilities. When conducting an analysis for a generic requirement that would 15 apply to a number of different plants, the staff usually chooses a representative plant and site 16 for the base-case analysis. To assess the potential difference in analysis outcomes for the 17 affected variable population of sites and facilities, the staff should complete a plant-to-plant 18 variability analysis. For example, the expedited spent fuel transfer regulatory analysis 19 (NRC, 2013g) and technical basis (NRC, 2014d), as well as the CPRR analysis (NRC, 2015a; 20 NRC, 2018b), included sensitivity analyses that showed the effect of the same accident 21 occurring at different sites.

22 23 For the CPRR analysis, the staff performed MACCS sensitivity calculations to analyze the 24 influence of site-to-site variations and protective action variations on the offsite consequences.

25 The staff conducted the following sensitivity calculations:

26 27

population (low, medium, high) 28

evacuation delay (1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months , 3 hour3.472222e-5 days 8.333333e-4 hours 4.960317e-6 weeks 1.1415e-6 months , 6 hour6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months , no evacuation) 29

nonevacuating cohort size (5 percent of EPZ population) 30

intermediate phase duration (0, 3 months, and 1 year) 31

long-term habitability criterion (500 millirem per year and 2 rem per year), which can vary 32 among states in the United States 33 34 Table H-5 shows one example of results from this set of sensitivity calculations. This table 35 shows the ratio of results if the intermediate phase duration were 1 year instead of the baseline 36 duration of 3 months. The color coding visually shows the significance to various metrics.

37 Yellow indicates a ratio of near 1, meaning there was no significant difference, while colors 38 closer to red or green indicate a larger influence on results. Results are reported for three sites 39 with representative low, medium, and high populations, coupled with low, medium, and high 40 source terms for Mark I and Mark II containments. Table H-5 shows that the conditional offsite 41 costs for the high source terms at all six sites evaluated are approximately 1.6 times higher 42 when the intermediate phase is assumed to last for 1 year versus 3 months.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-40

1 Table H-5 Ratio of Consequences for 1-Year Intermediate Phase Duration Sensitivity 2 Cases to Baseline Cases in the Containment Protection and Release 3 Reduction Analysis Individual Land (sq mi) Population Base Model Early Individual Latent Cancer Population Dose Offsite Cost Exceeding Long- Subject to Long-Fatality Fatality Risk (person-rem) ($ 2013) Term Habitability Term Protective Site Source Term Risk Criterion Actions 0-1.3 mi 0-100 0-10 mi 0-50 mi 0-100 mi 0-50 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi and beyond mi Mark I - Low (Bin 3) 0.88 0.89 0.88 0.98 0.99 0.98 0.98 1.00 1.00 0.00 0.00 Mark I - Peach Bottom Low - Hatch Mark I - Med (Bin 10) 1.07 0.93 0.91 0.97 0.97 1.38 1.18 0.86 0.92 0.48 0.48 Mark I - High (Bin 17) 1.04 0.98 0.93 0.98 0.96 1.61 1.39 0.80 0.87 0.60 0.53 Medium - Mark I - Low (Bin 3) 0.88 0.88 0.88 0.94 0.95 0.96 0.96 1.00 1.00 0.14 0.14 Vermont Mark I - Med (Bin 10) 1.06 0.92 0.89 0.93 0.92 1.39 1.04 0.73 0.86 0.57 0.57 Yankee Mark I - High (Bin 17) Individual 1.02 0.97 0.91 0.97 0.92 1.58 1.33 0.71 0.82 0.59 0.46 Mark I - Low (Bin 3) early fatality 0.88 0.89 0.88 0.95 0.95 0.97 0.97 1.00 1.00 0.16 0.16 High - Peach Mark I - Med (Bin 10) risk is zero 1.07 0.92 0.90 0.93 0.92 1.31 1.16 0.91 0.94 0.39 0.39 Bottom Mark I - High (Bin 17) for all 1.04 0.97 0.93 0.97 0.94 1.60 1.46 0.86 0.89 0.55 0.51 Mark II - Low (Bin 2) baseline 0.90 0.93 0.93 0.99 0.99 1.00 1.00 1.00 1.00 *

Low -

Mark II - Med (Bin 5) and 0.96 0.92 0.92 0.98 0.98 1.00 1.00 0.99 1.00 0.29 0.29 Mark II - Limerick Columbia Mark II - High (Bin 8) sensitivity 1.18 0.98 0.98 0.98 0.98 1.50 1.49 0.86 0.90 0.20 0.19 Mark II - Low (Bin 2) cases. 0.90 0.93 0.93 0.96 0.96 1.00 1.00 1.00 1.00 *

Medium -

Mark II - Med (Bin 5) 0.98 0.93 0.90 0.95 0.93 1.18 1.11 0.94 0.97 0.44 0.44 Susquehanna Mark II - High (Bin 8) 1.18 0.98 0.98 0.97 0.97 1.63 1.49 0.62 0.81 0.26 0.21 Mark II - Low (Bin 2) 0.90 0.93 0.93 0.93 0.94 1.00 1.00 1.00 1.00 *

High -

Mark II - Med (Bin 5) 1.00 0.92 0.91 0.94 0.93 1.08 1.06 0.96 0.97 0.45 0.45 Limerick 4 Mark II - High (Bin 8) 1.17 0.97 0.98 0.95 0.96 1.57 1.48 0.68 0.81 0.21 0.20 5

An asterisk indicates that the values of both the numerator and denominator in the ratio are zero.

6 (Source: NUREG-2206, Table 4-33) 7 H-41 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.7 PRESENTATION OF RESULTSINPUTS TO REGULATORY 2 ANALYSIS 3

4 H.7.1 Aggregating Probabilistic Risk Assessment Results from Different Hazards 5

6 For many regulatory applications, it is necessary to consider the contributions from several 7 hazards to a specific risk metric. When considering multiple hazards, a PRA model can be a 8 fully integrated model in which all hazards are combined into a single logic structure, a set of 9 individual PRA models for each hazard, or a mixture of the two. When combining the results of 10 PRA models for several hazards, the levels of detail and approximation included in the PRA 11 model may differ from one hazard to the next. Because of the methods and data used, a high 12 level of uncertainty can exist in PRAs for internal fires, external events (seismic, high wind, and 13 others), and low-power/shutdown conditions. In principle, this uncertainty could be reduced by 14 developing models to the same level of detail and rigor associated with internal events, at-power 15 PRAs. A larger uncertainty in a subset of the total PRA analyses can result in greater 16 uncertainty. The analyst needs to understand the main sources of conservatism in the PRA 17 associated with any of the hazards that can potentially impact the regulatory application. When 18 interpreting the results of the comparison of risk metrics to acceptance criteria or guidelines, it is 19 important to focus not only on the aggregated numerical result but also on the relative 20 importance and uncertainty of the main contributors to the risk metric.

21 22 H.7.2 Offsite Consequence Measures 23 24 An analyst uses several offsite consequence measures to characterize the impacts resulting 25 from a severe accident. For the purposes of a regulatory analysis, the individual early fatality 26 risk, latent cancer fatality risk, population dose, and offsite economic costs should all be 27 presented. The first two enable comparisons with the NRCs QHOs, and the latter two are 28 needed to quantify the affected parameters (accident offsite consequences) in the cost-benefit 29 equation.

30 31 H.7.2.1 Conditional Consequence Measures 32 33 Conditional offsite consequence results should be presented, first, for each source term bin. In 34 other words, given that an accident occurs and results in a particular source term bin, the offsite 35 consequences should be presented. The next step is to map the source term bins onto the 36 release categories developed in the accident sequence analysis, for the purposes of risk 37 integration.

38 39 Early Fatality Risk 40 41 Individual early fatality risk for the area within approximately 1 mile of the site boundary is 42 provided as an input for the evaluation of the NRCs early fatality QHO (NRC, 2015a). 22 43 44 Latent Cancer Fatality Risk 45 46 The individual latent cancer fatality risk is the risk of an average individual within the specified 47 spatial element contracting a fatal cancer caused by early, intermediate, and long-term radiation 22 If no one resides within 1 mile of the site boundary an individual should be assumed to reside within 1 mile for evaluation purposes.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-42

1 exposures. The analyst calculates this population-weighted metric by dividing the expected 2 number of fatal cancers in a spatial element by the population residing in that element. The 3 analysis should show the individual latent cancer fatality risk for the areas within 10- and 4 50-miles from the reactor site. The 10-mile area corresponds to the QHO for cancer fatality risk 5 (NRC, 2015a) and to the plume exposure EPZ. The analysis also should display the results for 6 the 50-mile area, as the NRCs regulatory analyses use this distance (other distances may be 7 appropriate, depending on facility type, as discussed in Section H.3.3.3).

8 9 Population Dose Risk 10 11 The offsite population dose, measured in person-rem, represents the sum of the doses from all 12 exposure pathways multiplied by the size of the population within a specified area. This metric 13 quantifies the public health (accident) attribute, as discussed in Sections 5.2.1 and 5.3.2.1 of 14 this NUREG. The dose to the population within a 50-mile radius (or other appropriate distance, 15 as discussed in Section H.3.3.3) from the reactor facility is reported for each source term bin.

16 MACCS reports the population dose per event (i.e., the conditional dose, given a particular 17 accident), and this value needs to be converted to the population dose per reactor-year by 18 multiplying by the event frequency.

19 20 Offsite Economic Cost Risk 21 22 The offsite economic costs resulting from an accident scenario correspond to the economic 23 consequences (offsite property) attribute described in Sections 5.2.5 and 5.3.2.5 of this 24 NUREG. This metric sums the costs of the protective actions taken to reduce offsite exposure 25 and restore land to usability and habitability. The offsite economic costs are computed directly 26 by MACCS and should be reported for the area within a 50-mile radius (or other appropriate 27 distance, as discussed in Section H.3.3.3) of the reactor facility for each source term bin.

28 29 Other Results 30 31 In addition to risk estimates, other consequence results provide risk insights about the various 32 alternatives. Some examples include the number of displaced individuals, land contamination, 33 and the extent over which protective actions may be needed. Discussion of these other results 34 may provide a better understanding of the extent and severity of the accident scenarios.

35 36 Table H-6 gives one example of how this information might be tabulated. This table is taken 37 from the CPRR analysis (NRC, 2015a; NRC, 2018b) and shows each of these consequence 38 results and their corresponding source term bins. This CPRR analysis (similar to the SFP study 39 [NRC, 2014d]) reported other results, such as land contamination and size of the population 40 affected by long-term protective actions, at radii of 50 miles and 100 miles from the reactor site.

41 H-43 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Table H-6 Severe Accident Consequence Analysis ResultsExample 2

3 (Source: SECY-15-0085, Enclosure, Table 4-22) 4 5 The consequence results presented in Table H-6 do not account for the event frequency, 6 (e.g., they are conditional on the occurrence of the postulated accident). Also, it is important to 7 note that these results are strongly dependent on the assumed (modeled) protective actions.

8 9 H.7.3 Evaluation of Regulatory Alternatives 10 11 H.7.3.1 Results from the Core Damage Event Tree Quantification 12 13 The analysis should tabulate the point estimates for relevant initiating event frequency, CDF, 14 and conditional core damage probability by site for each regulatory alternative. These tables 15 provide insight into the efficacy of the different strategies and present fleet averages for CDF 16 and conditional core damage probability for comparison.

17 18 Basic events, such as equipment and human failure events, should be tabulated with 19 importance measures (Risk Achievement Worth and Fussel-Vesely) with respect to CDF. A 20 table should show plant damage state frequencies for each regulatory alternative.

21 22 H.7.3.2 Results from the Accident Progression Event Tree Quantification 23 24 The analysis should tabulate the conditional containment failure probability for each APET to 25 demonstrate the efficacy of different mitigation alternatives. It should also tabulate the 26 frequencies of significant release categories for each APET.

27 28 The accident sequence analysis results show the CDF frequency from the initiating event and 29 provide insights into the relative contributions of various factors (e.g., external hazards, 30 equipment failures, human errors) to overall CDF. Figure H-7 shows an example of accident 31 sequence analysis and radioactive release summary results from the SFP study (NRC, 2014d).

32 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-44

1 2 Figure H-7 Likelihood of a Leak and Magnitude of Releases from Beyond-Design-Basis 3 Earthquake 4 (Source: NUREG-2161, Figure ES-1) 5 6 H.7.3.3 Results from MELCOR Analysis 7

8 The MELCOR results are classified into two broad categories: (1) thermal-hydraulic output and 9 (2) source term output. The timing of key events for the accident progression should be 10 presented and discussed for select MELCOR cases. In addition, time plots should be provided 11 for some important thermal-hydraulic outputs. Some examples include the following:

12 13

Reactor pressure vessel pressure, temperature, and water level 14 15

Containment pressure and temperature, to determine the likelihood of failure of 16 containment and various components by overpressure, overtemperature, or both 17 18

Hydrogen and other noncondensable gas generation and migration, to contribute to 19 containment overpressurization; also, to determine the potential for combustion in, for 20 example, the reactor building or the vent line 21 H-45 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 These discussions assist the analyst in assessing how each regulatory alternative would impact 2 the accident progression and the state of containment vulnerability under severe accident 3 conditions. They also provide the decisionmaker with qualitative information and a technical 4 basis for developing potential staff guidance for implementing a regulatory alternative.

5 6 H.7.4 Risk Integration Results and Key Insights 7

8 The final step is to present the results as integrated risk measures, which multiplies the 9 frequencies of different accident sequences with their conditional consequences. For example, 10 for each regulatory alternative (or subalternative), the population dose risk and offsite economic 11 cost risks should be presented on a per-reactor-year basis. Table H-7 and Figure H-8 show 12 example presentations of results, taken from the CPRR analysis (NRC, 2015a; NRC, 2018b).

13 The affected parameters that are quantified in the cost-benefit equation, population dose risk, 14 and economic cost risk, associated with each regulatory analysis subalternative are presented 15 for 50-mile and 100-mile radial distances. Additional measures are also presented, such as 16 land exceeding habitability criterion. Figures H-9 and H-10 show another example, taken from 17 the filtered vents analysis (NRC, 2012h), which presents the change (compared to the status 18 quo) in offsite economic cost risk per year for each regulatory alternative, called a Mod 19 (Figure H-9). Furthermore, the results of the uncertainty quantification are shown for those 20 alternatives (Figure H-10) with a positive change.

21 22 In addition to quantitative risk results, important qualitative insights and assumptions should also 23 be presented, on the most important contributors to risk and uncertainty. The supplementary 24 analyses discussed in Section H.6 make an essential contribution to this summary discussion 25 for decision makers, since those investigations help identify the impact of uncertainties and the 26 sensitivity of results to different assumptions. For example, the Technical Evaluation Summary 27 of the CPRR analysis (NRC, 2015a, Section 4.6 of Enclosure) presented the key insights from 28 the risk evaluation, MELCOR analysis, and MACCS analysis. These insights included the 29 following:

30 31

A discussion of the most important contributors to accident frequency (e.g., the major 32 contribution to seismically induced ELAP is from earthquakes that cause site peak 33 ground accelerations in the range of 0.3 to 0.75g) 34 35

A discussion of important assumptions (e.g., the evaluation assumed that 60 percent of 36 the time, the pre-core-damage water addition [FLEX] will be successful in preventing 37 core damage) 38 39

A discussion of accident progression and source term insights (e.g., the highest 40 calculated release to the environment results from a main steam line creep rupture 41 scenario, which is one of the least likely scenarios) 42 43

A discussion of offsite consequence insights (e.g., that, for all Mark I and Mark II source 44 terms, there is zero early fatality risk because the source terms are not large enough to 45 exceed the threshold for the acute dose to the red bone marrow, which is typically the 46 most sensitive tissue for early fatalities) 47 48

A discussion of important uncertainties and their key drivers (e.g., that the 49 5 percent/95 percent parametric uncertainty interval of the estimated risks is more than 50 1 order of magnitude and is largely driven by uncertainty in the seismic hazard curves)

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-46

2 3 1 Individual Land Exceeding Population Subject Fraction of Early Individual Latent Cancer Population Dose Offsite Cost Long-Term to Long-Term Core-Damage Fatality Fatality Risk (/y) (person-rem/y) ($ 2013/y) Habitability Criterion Protective Actions Frequency Risk (/y) (square miles/y) (persons/y)

Regulatory Analysis Index Uncontrolled 0-1.3 mi Sub-Alternative Vented 0-10 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi Release and beyond 1 1 0% 100% 0.0E+00 3.0E-09 8.6E-10 4.2E-10 1.3E+01 2.3E+01 9.9E+04 1.3E+05 4.4E-03 7.6E-03 5.1E-01 5.8E-01 2 2A 0% 100% 0.0E+00 3.0E-09 8.6E-10 4.2E-10 1.3E+01 2.3E+01 9.9E+04 1.3E+05 4.4E-03 7.6E-03 5.1E-01 5.8E-01 (Source: NUREG-2206, Table 5-1) 3 3A 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 4 3B 42% 58% 0.0E+00 2.1E-09 6.7E-10 3.4E-10 1.1E+01 1.9E+01 7.4E+04 1.0E+05 3.4E-03 6.4E-03 4.1E-01 4.9E-01 5 4Ai(1) 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 6 4Ai(2) 42% 58% 0.0E+00 2.1E-09 6.1E-10 3.1E-10 9.5E+00 1.7E+01 6.8E+04 9.0E+04 3.2E-03 5.8E-03 3.6E-01 4.1E-01 7 4Aii(1) 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 8 4Aii(2) 42% 58% 0.0E+00 2.4E-09 7.7E-10 3.9E-10 1.2E+01 2.2E+01 8.9E+04 1.2E+05 3.9E-03 7.3E-03 4.8E-01 5.8E-01 9 4Aiii(1) 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 H-47 10 4Aiii(2) 42% 58% 0.0E+00 2.0E-09 5.6E-10 2.7E-10 8.7E+00 1.5E+01 6.2E+04 7.9E+04 3.0E-03 5.1E-03 3.1E-01 3.4E-01 11 4Bi(1) 58% 42% 0.0E+00 1.3E-09 3.1E-10 1.5E-10 4.5E+00 7.8E+00 2.9E+04 3.7E+04 1.6E-03 2.5E-03 1.5E-01 1.6E-01 12 4Bi(2) 42% 58% 0.0E+00 1.4E-09 3.3E-10 1.5E-10 4.8E+00 8.2E+00 3.1E+04 3.8E+04 1.8E-03 2.7E-03 1.6E-01 1.6E-01 13 4Bii 42% 58% 0.0E+00 1.4E-09 3.2E-10 1.5E-10 4.6E+00 7.9E+00 3.0E+04 3.7E+04 1.7E-03 2.6E-03 1.5E-01 1.5E-01 14 4Biii 42% 58% 0.0E+00 1.4E-09 3.2E-10 1.5E-10 4.7E+00 8.1E+00 3.1E+04 3.7E+04 1.7E-03 2.6E-03 1.5E-01 1.6E-01 15 4Biv 40% 60% 0.0E+00 1.3E-09 3.1E-10 1.5E-10 4.6E+00 7.8E+00 3.0E+04 3.6E+04 1.7E-03 2.6E-03 1.5E-01 1.5E-01 16 4Ci(1) 58% 42% 0.0E+00 1.3E-09 3.1E-10 1.5E-10 4.5E+00 7.8E+00 2.9E+04 3.7E+04 1.6E-03 2.5E-03 1.5E-01 1.6E-01 Table H-7 Risk Estimates by Regulatory Analysis Subalternative 17 4Ci(2) 42% 58% 0.0E+00 1.3E-09 3.1E-10 1.4E-10 4.5E+00 7.6E+00 3.0E+04 3.7E+04 1.6E-03 2.4E-03 1.5E-01 1.6E-01 18 4Cii 42% 58% 0.0E+00 1.3E-09 3.0E-10 1.4E-10 4.4E+00 7.4E+00 2.9E+04 3.6E+04 1.5E-03 2.3E-03 1.5E-01 1.5E-01 19 4Ciii 42% 58% 0.0E+00 1.3E-09 3.1E-10 1.4E-10 4.4E+00 7.6E+00 3.0E+04 3.7E+04 1.6E-03 2.4E-03 1.5E-01 1.6E-01 20 4Civ 40% 60% 0.0E+00 1.3E-09 3.0E-10 1.4E-10 4.3E+00 7.4E+00 2.9E+04 3.6E+04 1.5E-03 2.3E-03 1.5E-01 1.5E-01 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-8 Comparison of Regulatory Analysis Alternatives Using Population Dose Risk 3 (0-50 miles) 4 (Source: NUREG-2206, Figure 5-2) 5 6

7 Figure H-9 Reduction in 50-mile Offsite Cost Risk ($/reactor-year) 8 (Source: SECY-12-0157, Enclosure 5c, Figure 5)

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-48

1 2

3 Figure H-10 Uncertainty in Reduction in 50-mile Offsite Cost Risk 4 (Source: SECY-12-0157, Enclosure 5c, Figure 10) 5 H-49 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 H.8 REFERENCES 2

3 42 USC 2011 et seq. Atomic Energy Act of 1954, as amended.

4 5 American Society of Mechanical Engineers/American Nuclear Society (ASME/ANS), Standard 6 for Probabilistic Risk Assessment for Nuclear Power Plant Applications, RA S-2002, 2002.

7 ASME/ANS, Standard for Level 1/Large Early Release Frequency Probabilistic Risk 8 Assessment for Nuclear Power Plant Applications, RA-Sa-2009, Addendum A to RA-S-2008, 9 2009.

10 ASME/ANS, Severe Accident Progression and Radiological Release (Level 2) PRA Standard 11 for Nuclear Power Plant Applications for Light Water Reactors (LWRs), RA-S-1.2-2014, 2014.

12 ASME/ANS, Standard for Radiological Accident Offsite Consequence Analysis (Level 3 PRA) 13 to Support Nuclear Installation Applications, RA-S-1.3-2017, 2017.

14 Code of Federal Regulations (CFR), Title 10, Energy, Part 50, Domestic Licensing of 15 Production and Utilization Facilities. Available at http://www.nrc.gov/reading-rm/doc-16 collections/cfr/part050/.

17 18 CFR, Title 10, Energy, Part 51, Environmental Protection Regulations for Domestic Licensing 19 and Related Regulatory Functions. Available at https://www.nrc.gov/reading-rm/doc-20 collections/cfr/part051/.

21 22 CFR, Title 10, Energy, Part 52, Licenses, Certifications, and Approvals for Nuclear Power 23 Plants. Available at https://www.nrc.gov/reading-rm/doc-collections/cfr/part052/.

24 25 CFR, Title 10, Energy, Part 72, Licensing Requirements for the Independent Storage of Spent 26 Nuclear Fuel, High-Level Radioactive Waste, and Reactor-Related Greater than Class C 27 Waste. Available at https://www.nrc.gov/reading-rm/doc-collections/cfr/part072/.

28 29 Electrical Power Research Institute (EPRI), Practical Guidance on the Use of Probabilistic Risk 30 Assessment in Risk-Informed Applications with a Focus on the Treatment of Uncertainty, 31 Appendix E, Report 1026511, 2012.

32 Gregory, J.J., et al., Task 5 Letter Report: MACCS2 Uncertainty Analysis of EARLY Exposure 33 Results, NRC-JCN-W6352, Sandia National Laboratories, 2000.

34 35 Helton, J.C. Uncertainty and Sensitivity Analysis in the Presence of Stochastic and Subjective 36 Uncertainty, Journal of Statistical Computation and Simulation 197; 57 (1-4): 3-76, 1995.

37 38 Humphries, L.L., et al., MELCOR Computer Code Manuals: Primer and Users Guide, Version 39 2.1.6840 2015, SAND2015-6691 R, Volume 1, Sandia National Laboratories, 2015.

40 Agencywide Documents Access and Management System (ADAMS) Accession 41 No. ML15300A479.

42 43 Keller, W., and Modarres, M., A historical overview of probabilistic risk assessment 44 development and its use in the nuclear power industry: a tribute to the late Professor Norman 45 Carl Rasmussen, Reliability Engineering & System Safety, Volume 89, Issue 3, 2005, 46 Pages 271-285, ISSN 0951-8320. Available at http://dx.doi.org/10.1016/j.ress.2004.08.022.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-50

1 2 National Fire Protection Association (NFPA) Standards Council, Performance-Based Standard 3 for Fire Protection for Light-Water Reactor Electric Generating Plants, NFPA Standard 805, 4 current version.

5 6 Nuclear Energy Institute (NEI), Severe Accident Mitigation Alternatives (SAMA) Analysis, 7 Guidance Document, NEI-05-01, Revision A, 2005. ADAMS Accession No. ML060530203.

8 9 NEI, Industry Guidance for Compliance with Order EA-13-109: BWR Mark I & II Reliable 10 Hardened Containment Vents Capable of Operation Under Severe Accident Conditions, 11 NEI 13-02, Revision 0C3, 2014. ADAMS Accession No. ML14294A789.

12 13 Office of Management and Budget (OMB), Regulatory Analysis, Circular A-4, September 17, 14 2003. Available at https://www.whitehouse.gov/sites/whitehouse.gov/files/omb/circulars/A4/a-15 4.pdf.

16 17 Stein, A.F., Draxler, R.R, Rolph, G.D., Stunder, B.J.B., Cohen, M.D., and Ngan, F., NOAAs 18 HYSPLIT Atmospheric Transport and Dispersion Modeling System, 2015. Available at 19 https://doi.org/10.1175/BAMS-D-14-00110.1.

20 21 U.S. Department of Energy (DOE), MACCS2 Computer Code Application Guidance for 22 Documented Safety Analysis, Final Report, DOE-EH-4.2.1.4 MACCS2-Code Guidance, 2004.

23 Available at 24 https://www.energy.gov/sites/prod/files/2013/07/f2/Final_MACCS2_Guidance_Report_June_1_

25 2004.pdf.

26 27 U.S. Environmental Protection Agency (EPA), Meteorological Monitoring Guidance for 28 Regulatory Modeling Applications, EPA-454/R-99-005, 2000. Available at 29 https://www3.epa.gov/scram001/guidance/met/mmgrma.pdf.

30 31 EPA, PAG Manual: Protective Action Guides and Planning Guidance for Radiological 32 Incidents, EPA-400/R-17/001, 2017. Available at https://www.epa.gov/radiation/pag-manuals-33 and-resources.

34 35 U.S. Nuclear Regulatory Commission (NRC), Reactor Safety Study: An Assessment of 36 Accident Risks in U.S. Commercial Nuclear Power Plants, WASH-1400 (NUREG-75/014),

37 1975. Available at https://www.nrc.gov/reading-rm/doc-collections/nuregs/staff/sr75-014/.

38 NRC Anticipated Transients Without Scram for Light Water Reactors, NUREG-0460, 1978.

39 ADAMS Accession No. ML15134A025.

40 NRC, Fault Tree Handbook, NUREG-0492, 1981. ADAMS Accession No. ML100780465.

41 42 NRC, Nuclear Regulatory Commission Staff Computer Programs for Use with Meteorological 43 Data, NUREG-0917, 1982b. ADAMS Accession No. ML12061A136.

44 NRC, PRA Procedures Guide: A Guide to the Performance of Probabilistic Risk Assessments 45 for Nuclear Power Plants, NUREG/CR-2300, 1983a. Available at https://www.nrc.gov/reading-46 rm/doc-collections/nuregs/contract/cr2300/.

47 H-51 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 NRC, Guidelines for Nuclear Power Plant Safety Issue Prioritization Information Development, 2 NUREG/CR-2800, 1983b.

3 4 NRC, Safety Goals for Nuclear Power Plant Operation, NUREG-0880, Revision 1, 1983c.

5 ADAMS Accession No. ML071770230.

6 7 NRC, Policy Statement on Severe Reactor Accidents Regarding Future Designs and Existing 8 Plants, Federal Register, 1985 (50 FR 32138). ADAMS Accession No. ML003711521.

9 10 NRC, Safety Goals for the Operations of Nuclear Power Plants; Policy Statement; 11 Republication, Federal Register, 1986 (51 FR 30028). Available at http://www.nrc.gov/reading-12 rm/doc-collections/commission/policy/51fr30028.pdf.

13 14 NRC, Individual Plant Examination for Severe Accident Vulnerabilities10 CFR 50.54(f),

15 Generic Letter 88-20, 1988a. Available at https://www.nrc.gov/reading-rm/doc-collections/gen-16 comm/gen-letters/1988/gl88020.html.

17 NRC, Evaluation of Station Blackout Accidents at Nuclear Power Plants, Technical Findings 18 Related to Unresolved Safety Issue A-44, Final Report, NUREG-1032, 1988b.

19 NRC, MELCOR Accident Consequence Code System (MACCS), NUREG/CR-4691, Sandia 20 National Laboratories, 1990a. ADAMS Accession No. ML063560409.

21 NRC, Severe Accident Risks: An Assessment for Five U.S. Nuclear Power PlantsFinal 22 Summary Report, NUREG-1150, 1990b. Available at https://www.nrc.gov/reading-rm/doc-23 collections/nuregs/staff/sr1150/v1/.

24 25 NRC, Evaluation of Severe Accident Risks: Quantification of Major Input Parameters: MACCS 26 (MELCOR Accident Consequence Code System) Input, NUREG/CR-4551, Volume 2, Revision 27 1, Part 7, 1990c. Available at https://www.osti.gov/servlets/purl/6360728.

28 NRC, Individual Plant Examination of External Events (IPEEE) for Severe Accident 29 Vulnerabilities10 CFR 50.54(f), Generic Letter 88-20, Supplement 4, 1991.

30 31 NRC, Software Quality Assurance Program and Guidelines, NUREG/BR-0167, 1993. ADAMS 32 Accession No. ML012750471.

33 NRC, Use of Probabilistic Risk Assessment Methods in Nuclear Regulatory Activities; Final 34 Policy Statement, Federal Register, August 16, 1995a (60 FR 42622). Available at 35 https://www.nrc.gov/reading-rm/doc-collections/commission/policy/.

36 NRC, Reassessment of NRCs Dollar Per Person-Rem Conversion Factor Policy, 37 NUREG-1530, 1995b. ADAMS Accession No. ML063470485.

38 NRC, Generic Environmental Impact Statement for License Renewal of Nuclear Plants: Main 39 Report, NUREG-1437, Volume 1, 1996. Available at https://www.nrc.gov/reading-rm/doc-40 collections/nuregs/staff/sr1437/v1/.

41 NRC, Individual Plant Examination Program: Perspectives on Reactor Safety and Plant 42 Performance, NUREG-1560, 1997a. ADAMS Accession No. ML063550244.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-52

1 NRC, Regulatory Analysis Technical Evaluation Handbook, Final Report, NUREG/BR-0184, 2 1997b. ADAMS Accession No. ML050190193.

3 NRC, Code Manual for MACCS2: Users Guide, NUREG/CR-6613, Volume 1, Sandia National 4 Laboratories, 1998. ADAMS Accession No. ML110030976.

5 NRC, Technical Study of Spent Fuel Pool Accident Risk at Decommissioning Nuclear Power 6 Plants, NUREG-1738, 2001. ADAMS Accession No. ML010430066.

7 NRC, Perspectives Gained from the Individual Plant Examination of External Events (IPEEE) 8 ProgramFinal Report, NUREG-1742, 2002 Available at https://www.nrc.gov/reading-rm/doc-9 collections/nuregs/staff/sr1742/.

10 NRC, An Approach for Plant-Specific Risk-Informed Decisionmaking for Inservice Inspection of 11 Piping, Regulatory Guide 1.178, Revision 1, 2003. ADAMS Accession No. ML032510128.

12 NRC, Reevaluation of Station Blackout Risk at Nuclear Power Plants, NUREG/CR-6890, 13 2005. Available at https://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr6890/.

14 NRC, Meteorological Monitoring Programs for Nuclear Power Plants, Regulatory Guide 1.23, 15 current version.

16 NRC, Final Environmental Impact Statement for an Early Site Permit (ESP) at the Vogtle ESP 17 Electric Generating Plant Site, NUREG-1872, 2008. Available at https://www.nrc.gov/reading-18 rm/doc-collections/nuregs/staff/sr1872/.

19 NRC, An Approach for Determining the Technical Adequacy of Probabilistic Risk Assessment 20 Results for Risk-Informed Activities, Regulatory Guide 1.200, current version.

21 NRC, Systems Analysis Programs for Hands-on Integrated Reliability Evaluations (SAPHIRE) 22 Version 8, NUREG/CR-7039, Idaho National Laboratory, 2011a. Available at 23 https://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr7039/.

24 NRC, Near-Term Report and Recommendations for Agency Actions Following the Events in 25 Japan, SECY-11-0093, 2011b. ADAMS Accession No. ML11186A950 (package).

26 NRC, Prioritization of Recommended Actions to be Taken in Response to Fukushima Lessons 27 Learned, SECY-11-0137, 2011c. ADAMS Accession No. ML11269A204.

28 NRC, Generic Environmental Impact Statement for License Renewal of Nuclear Plants:

29 Regarding Hope Creek Generating Station and Salem Nuclear Generating Station, Units 1 and 30 2, NUREG-1437, Supplement 45, Final Report, 2011d. Available at 31 https://www.nrc.gov/reading-rm/doc-collections/nuregs/staff/sr1437/supplement45/.

32 33 NRC, State-of-the-Art Reactor Consequence Analyses (SOARCA) Report, NUREG-1935, 34 2012a. ADAMS Accession No. ML12332A057.

35 36 NRC, Issuance of Order to Modify Licenses with Regard to Requirements for Mitigation 37 Strategies for Beyond-Design-Basis External Events, Order EA-12-049, 2012b. ADAMS 38 Accession No. ML12054A735.

H-53 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 NRC, Issuance of Order to Modify Licenses with Regard to Reliable Hardened Containment 2 Vents, Order EA-12-050, 2012c. ADAMS Accession No. ML12054A694.

3 NRC, Order Modifying Licenses with Regard to Reliable Spent Fuel Pool Instrumentation, 4 Order EA-12-051, 2012d. ADAMS Accession No. ML12056A044.

5 NRC, Staff RequirementsSECY-12-0025Proposed Orders and Requests for Information in 6 Response to Lessons Learned from Japan's March 11, 2011, Great Tohoku Earthquake and 7 Tsunami, SRM-SECY-12-0025, 2012e. ADAMS Accession No. ML120690347.

8 NRC, Tier 3 Program Plans and 6-Month Status Update in Response to Lessons Learned from 9 Japans March 11, 2011, Great Tohoku Earthquake and Subsequent Tsunami, SECY-12-0095 10 (redacted), 2012f. ADAMS Accession No. ML12208A208.

11 NRC, Consideration of Economic Consequences within the U.S. Nuclear Regulatory 12 Commissions Regulatory Framework, SECY-12-0110, Enclosure 9, MELCOR Accident 13 Consequence Code System, Version 2 (MACCS2), 2012g. ADAMS Accession 14 No. ML12173A509.

15 NRC, Consideration of Additional Requirements for Containment Venting Systems for Boiling 16 Water Reactors with Mark I and Mark II Containments, SECY-12-0157 (redacted), 2012h.

17 ADAMS Accession No. ML12325A704 18 19 NRC, Letter with Subject Request for Information Pursuant to Title 10 of the Code of Federal 20 Regulations 50.54(f) Regarding Recommendations 2.1,2.3, and 9.3, of the Near-term Task 21 Force Review of Insights from the Fukushima Dai-ichi Accident, 2012i. ADAMS Accession 22 No. ML12053A340.

23 24 NRC, Modeling Potential Reactor Accident Consequences (NUREG/BR-0359, Revision 1),

25 NUREG/BR-0359, Revision 1, 2012j. ADAMS Accession No. ML12347A049.

26 27 NRC, State-of-the-Art Reactor Consequence Analyses Project: Peach Bottom Integrated 28 Analysis, NUREG/CR-7110, Volume 1, 2013a. ADAMS Accession No. ML13150A053.

29 30 NRC, State-of-the-Art Reactor Consequence Analyses Project: Surry Integrated Analysis, 31 NUREG/CR-7110, Volume 2, Revision 1, 2013b. ADAMS Accession No. ML13240A242.

32 33 NRC, Synthesis of Distributions Representing Important Non-Site-Specific Parameters in 34 Off-Site Consequence Analyses, NUREG/CR-7161, Sandia National Laboratories, 2013c.

35 ADAMS Accession No. ML13114A322.

36 37 NRC, Issuance of Order Modifying Licenses with Regard to Reliable Hardened Containment 38 Vents Capable of Operation Under Severe Accident Conditions, Order EA-13-109, 2013d.

39 ADAMS Accession No. ML13143A321.

40 41 NRC, Consequence Study of a Beyond-Design-Basis Earthquake Affecting the Spent Fuel Pool 42 for a U.S. Mark I Boiling-Water Reactor, SECY-13-0112, 2013e. ADAMS Accession 43 No. ML13256A334 (package).

44 45 NRC, Glossary of Risk-Related Terms in Support of Risk-Informed Decisionmaking, 46 NUREG-2122, 2013f. ADAMS Accession No. ML13311A353.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-54

1 2 NRC, Staff Evaluation and Recommendation for Japan Lessons-Learned Tier 3 Issue on 3 Expedited Transfer of Spent Fuel, COMSECY-13-0030, 2013g. ADAMS Accession 4 No. ML13329A918 (package).

5 6 NRC, Staff RequirementsSECY-12-0157Consideration of Additional Requirements for 7 Containment Venting Systems for Boiling Water Reactors with Mark I and Mark II 8 Containments, SRM-SECY-12-0157, 2013h.

9 10 NRC, Final Environmental Statement: Related to the Operation of Watts Bar Nuclear Plant, 11 Unit 2Final Report, NUREG-0498, Supplement 2, 2013i. Available at 12 https://www.nrc.gov/reading-rm/doc-collections/nuregs/staff/sr0498/.

13 14 NRC, Updated Schedule and Plans for Japan Lessons-Learned Tier 3 Issue on Expedited 15 Transfer of Spent Fuel, May 7, 2013j. ADAMS Accession No. ML13105A122.

16 17 NRC, MELCOR Best Practices as Applied in the State-of-the-Art Reactor Consequence 18 Analyses (SOARCA) Project, NUREG/CR-7008, Sandia National Laboratories, 2014a.

19 ADAMS Accession No. ML14234A136.

20 NRC, MACCS Best Practices as Applied in the State-of-the-Art Reactor Consequence 21 Analyses (SOARCA) Project, NUREG/CR-7009, Sandia National Laboratories, 2014b.

22 ADAMS Accession No. ML14234A148.

23 NRC, Compendium of Analyses to Investigate Select Level 1 Probabilistic Risk Assessment 24 End-State Definition and Success Criteria Modeling Issues, Energy Research, Inc.,

25 NUREG/CR-7177, 2014c. ADAMS Accession No. ML14148A126.

26 NRC, Consequence Study of a Beyond-Design-Basis Earthquake Affecting the Spent Fuel Pool 27 for a U.S. Mark I Boiling-Water Reactor, NUREG-2161, 2014d. ADAMS Accession 28 No. ML14255A365.

29 NRC, NRC Incident Investigation Program, Management Directive 8.3, 2014e. ADAMS 30 Accession No. ML13175A294.

31 NRC, Qualitative Consideration of Factors in the Development of Regulatory Analyses and 32 Backfit Analyses, SECY-14-0087, 2014f. ADAMS Accession No. ML14127A458 (package).

33 34 NRC, NRC Non-Concurrence Process, Management Directive 10.158, 2014g. ADAMS 35 Accession No. ML13176A371.

36 NRC, Staff RequirementsCOMSECY-13-0030Staff Evaluation and Recommendation for 37 Japan Lessons-Learned Tier 3 Issue on Expedited Transfer of Spent Fuel, 38 SRM-COMSECY-13-0030, 2014h. ADAMS Accession No. ML14143A360.

39 NRC, Integrated Risk-Informed Decision-Making Process for Emergent Issues, LIC-504, 40 Revision 4, 2014i. ADAMS Accession No. ML14035A143.

41 NRC, Evaluation of Containment Protection and Release Reduction for Mark I and Mark II 42 Boiling-Water Reactors Rulemaking Activities (10 CFR Part 50), SECY-15-0085, 2015a.

43 ADAMS Accession No. ML15005A079.

H-55 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 NRC, Commission Voting RecordEvaluation of the Containment Protection and Release 2 Reduction for Mark I and Mark II Boiling-Water Reactors Rulemaking Activities (10 CFR Part 50) 3 (RIN-3150-AJ26), CVR SECY-15-0085, 2015b. ADAMS Accession No. ML15231A524.

4 NRC, Staff RequirementsEvaluation of Containment Protection and Release Reduction for 5 Mark I and Mark II Boiling-Water Reactors Rulemaking Activities (10 CFR Part 50),

6 SRM-SECY-15-0085, 2015c. ADAMS Accession No. ML15231A471.

7 NRC, Proposed Rulemaking: Mitigation of Beyond-Design-Basis Events, SECY-15-0065, 8 2015d. ADAMS Accession No. ML15049A213.

9 NRC, Seventh 6-Month Status Update on Response to Lessons Learned from Japan's March 10 11, 2011, Great Tohoku Earthquake and Subsequent Tsunami, SECY-15-0059, 2015e.

11 ADAMS Accession No. ML15069A444.

12 NRC, Staff RequirementsProposed Rulemaking: Mitigation of Beyond-Design-Basis Events, 13 SECY-15-0065, 2015f. ADAMS Accession No. ML15231A471.

14 NRC, Generic Issues Program, Management Directive 6.4, 2015g. ADAMS Accession No.

15 ML18073A162.

16 NRC, Closure of Fukushima Tier 3 Recommendations Related to Containment Vents, 17 Hydrogen Control, and Enhanced Instrumentation, SECY-16-0041, 2016a. ADAMS Accession 18 No. ML16049A079 (package).

19 NRC, State-of-the-Art Reactor Consequence Analyses Project: Uncertainty Analysis of the 20 Unmitigated Long-Term Station Blackout of the Peach Bottom Atomic Power Station, 21 NUREG/CR-7155, SAND2012-10702P, 2016b. ADAMS Accession No. ML16133A461.

22 NRC, Probabilistic Risk Assessment and Regulatory Decisionmaking: Some Frequently Asked 23 Questions, NUREG-2201, 2016c. ADAMS Accession No. ML16245A032.

24 NRC, Historical Review and Observations of Defense-in-Depth, NUREG/KM-0009, 2016d.

25 ADAMS Accession No. ML16104A071.

26 NRC, Guidance on the Treatment of Uncertainties Associated with PRAs in Risk-Informed 27 Decisionmaking, Final Report, NUREG-1855, Revision 1, 2017a. ADAMS Accession 28 No. ML17062A466.

29 NRC, Proposed Revision to NUREG-1530 Reassessment of NRCs Dollar Per Person-Rem 30 Conversion Factor Policy, SECY-17-0017, 2017b. ADAMS Accession No. ML16147A293 31 (package).

32 NRC, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on 33 Plant-Specific Changes to the Licensing Basis, Regulatory Guide 1.174, current version.

34 NRC, Strategic Plan: Fiscal Years 2018-2022, NUREG-1614, Volume 7, 2018a. ADAMS 35 Accession No. ML18032A561.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-56

1 NRC, Technical Basis for the Containment Protection and Release Reduction Rulemaking for 2 Boiling-Water Reactors with Mark I and Mark II Containments, NUREG-2206, 2018b. ADAMS 3 Accession No. ML18065A048.

4 NRC, State-of-the-Art Reactor Consequence Analyses (SOARCA) Project: Sequoyah 5 Integrated Deterministic and Uncertainty Analyses, NUREG/CR-7245, Sandia National 6 Laboratories, 2019a.

7 NRC, 2019-2020 Information Digest, NUREG-1350, Volume 31, 2019b. ADAMS Accession 8 No. ML19242D326.

9 NRC, Benefits and Uses of the State-of-the-Art Reactor Consequence Analyses (SOARCA) 10 Project, Research Information Letter 19-01, 2019c.

11 NRC, SecPop Version 4: Sector Population, Land Fraction, and Economic Estimation 12 Program, NUREG/CR-6525, Revision 2, Sandia National Laboratories, 2019d. ADAMS 13 Accession No. ML19182A284.

14 NRC, State-of-the-Art Reactor Consequence Analyses (SOARCA) Project: Uncertainty 15 Analysis of the Unmitigated Short-Term Station Blackout of Surry Power Station, 16 NUREG/CR-7262, Sandia National Laboratories, 2020.

17 NRC and Commission of European Communities, Probabilistic Accident Consequence 18 Uncertainty Analysis, NUREG/CR-6244 Report Series, 1995.

19 NRC and FEMA, Criteria for Preparation and Evaluation of Radiological Emergency Response 20 Plans and Preparedness in Support of Nuclear Power Plants, NUREG-0654 and FEMA-REP-1, 21 Revision 1, 1980. ADAMS Accession No. ML040420012.

22 H-57 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 ENCLOSURE H-1: DESCRIPTION OF ANALYTICAL TOOLS AND 2 CAPABILITIES 3

4 Risk can be characterized in many ways, depending on the end states of interest for a decision 5 or application. To provide some overall logic and structure and to facilitate evaluation of 6 intermediate results, probabilistic risk assessments (PRAs) for nuclear power plants (NPPs) 7 have traditionally been organized into three analysis levels, with the scope and level of 8 complexity of the PRA model increasing with each level. These levels are defined by three 9 sequential adverse end states that can occur in the progression of postulated NPP accident 10 scenarios: (1) core damage, (2) radiological release, and (3) offsite radiological consequences.

11 12 Several computer codes exist for performing PRA and severe accident consequence analysis.

13 For regulatory analyses that require detailed analyses of offsite consequences, most recent 14 light-water reactor applications have used the U.S. Nuclear Regulatory Commission 15 (NRC)-sponsored MELCOR and MELCOR Accident Consequence Code System (MACCS) 16 code suites. These codes include state-of-the-art integrated modeling of severe accident 17 behavior that incorporates insights from decades of research into severe accident 18 phenomenology and radiation health effects. The NRC-sponsored Systems Analysis Programs 19 for Hands-on Integrated Reliability Evaluations (SAPHIRE) code is also available for performing 20 PRAs using event trees and fault trees. Figure H-11 notes the role of these three code suites in 21 NPP PRAs. The sections below describe these code suites, their capabilities, and their typical 22 uses.

23 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-58

SAPHIRE 1

2 MELCOR MACCS 3

4 Figure H-11 Overall Logic and Structure of Traditional NPP PRA Models and Role of 5 SAPHIRE, MELCOR, and MACCS Code Suites 6

7 Severe Accident Scenario Modeling and Frequency Analysis 8

9 Systems Analysis Programs for Hands-on Integrated Reliability Evaluations 10 (SAPHIRE) 11 12 SAPHIRE is an NRC-sponsored software application that the Idaho National Laboratory 13 developed and maintains for performing PRAs of complex engineered facilities, systems, or 14 processes.

15 16 The NRC uses SAPHIRE to develop Level 1 and Level 2 PRA logic models for NPPs. The end 17 state of interest for a Level 1 PRA is core damage. SAPHIRE can (1) model plant and operator 18 responses to initiating events to identify sequences (combinations of system and operator action 19 successes and failures) that result in either the achievement of a safe state or the onset of core 20 damage, (2) quantify the frequencies of sequences that result in core damage and total core 21 damage frequency (CDF) for the NPP, and (3) identify important contributors to CDF. The end 22 state of interest for a Level 2 PRA is radiological release. SAPHIRE can also be used to 23 expand upon a Level 1 PRA model to (1) model containment systems and operator responses 24 to severe accident conditions, (2) quantify radiological release category frequenciesincluding 25 a large early release frequency (LERF), and (3) identify important contributors to radiological 26 release category frequencies. A Level 3 PRA combines the results of the SAPHIRE radiological 27 release category frequencies (from the Level 2 PRA) with the results from the corresponding 28 MACCS offsite radiological consequence model to provide an overall characterization of the risk 29 to the offsite public from a broad spectrum of postulated accidents involving a modeled NPP 30 site.

31 H-59 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 SAPHIRE contains graphical editors for creating, viewing, and modifying fault tree and event 2 tree models that serve as logical representations of accident sequences that can occur at an 3 NPP. SAPHIRE uses event tree and fault tree models, coupled with accident sequence linkage 4 rules and postprocessing rules, to generate unique combinations of individual failures 5 (i.e., minimal cut sets) that can result in an undesired end state. SAPHIRE quantifies the 6 frequencies and probabilities associated with the minimal cut sets to estimate the frequencies of 7 selected undesired end states. In addition, SAPHIRE includes many useful features to support 8 the frequency quantification of PRA models and identification of significant contributors to risk 9 (e.g., calculation of traditional PRA importance measures described below). Finally, SAPHIRE 10 can perform an uncertainty analysis using either Monte Carlo or Latin Hypercube sampling 11 methods to estimate the uncertainty in calculated results (e.g., CDF, LERF, or importance 12 measures) caused by epistemic 23 uncertainties in input parameters for basic events in the 13 Level 1 and Level 2 PRA logic models.

14 15 NUREG/CR-7039, Systems Analysis Programs for Hands-on Integrated Reliability Evaluations 16 Version 8, issued June 2011, contains detailed information about the features and capabilities 17 of SAPHIRE Version 8. Some basic features and capabilities in SAPHIRE include the following:

18 19

Basic events: Basic events typically represent events involving failures of structures, 20 systems, or components; adverse environmental or phenomenological conditions that 21 could lead to failures; or human failure events for operator actions. Basic events are 22 logically linked together in fault trees and provide SAPHIRE with the probabilistic 23 information (e.g., failure data input and type of probability calculation) needed to quantify 24 the PRA model. Basic events appear as circles at the bottom of the example in 25 Figure H-12 (feeding System A and System B fault trees).

26 27

Fault trees: A fault tree generally represents a failure model. A fault tree model consists 28 of a top event (e.g., failure of System A in the example in Figure H-12), usually defined 29 by a heading in an event tree (e.g., System A appears as a heading in the example 30 event tree in Figure H-12, for the initiating event IE). A combination of basic events 31 must occur to result in the undesired top event, using a logic structure as a model for the 32 basic events.

33 34

Event trees: An event tree is a logic structure that chains sequential events together to 35 model the likelihood of the potential outcome(s) of those events. The simple example in 36 Figure H-12 contains a chain of three events: initiating event IE, System A (success or 37 failure), and System B (success or failure). The analyst defines accident sequences 38 using an event tree to indicate the failure or success of top events. Each heading in the 39 event tree is associated with a system fault tree. Event trees are constructed and 40 modified using a graphical editor that allows the linkage of multiple event trees and the 41 creation of very large event trees.

42 43

Rule-based fault tree linking: In generating accident sequences, the analyst uses a set 44 of defined rules to reduce the complexity of the overall logic structure.

45 46

Cut sets: A cut set is a combination of faults that must occur together to result in the 47 failure of a top event. To solve an accident sequence, SAPHIRE constructs a fault tree 23 Epistemic uncertainty is the uncertainty related to the lack of knowledge or confidence about the system or model and is also known as state-of-knowledge uncertainty (NUREG-2122, Glossary of Risk-Related Terms in Support of Risk-Informed Decision Making, issued November 2013).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-60

1 for those systems that are defined to be failed in the sequence logic by creating a 2 temporary AND gate with these systems as inputs. SAPHIRE then solves this fault 3 tree using specified cut set probability truncation values. This process results in a list of 4 cut sets for the failed systems in the accident sequence. SAPHIRE then uses Boolean 5 reduction techniques to further reduce this list of cut sets to the set of minimal cut sets 6 for the accident sequence. The analyst can specify one of three main cut set 7 quantification techniques, depending on the desired tradeoff between accuracy and 8 computation time.

9 10

Uncertainty analysis: Both Monte Carlo and Latin Hypercube sampling methods are 11 available for performing an uncertainty analysis. The uncertainty analysis functions in 12 SAPHIRE estimate the uncertainty in calculated output quantities caused by epistemic 13 uncertainties in the basic event frequencies or probabilities. These output quantities 14 include (1) fault tree top event probabilities, (2) event tree sequence frequencies, (3) end 15 state frequencies, or (4) importance measures. In an uncertainty analysis, SAPHIRE 16 samples analyst-specified distributions for each basic event in a group of cut sets and 17 then quantifies these cut sets using the sampled values.

18 19

Importance measures: SAPHIRE can quantify a range of traditional importance 20 measures that are used to measure the absolute or relative importance of basic events 21 in the PRA model to specified end-state frequencies. As previously stated, uncertainty 22 analyses on these measures can use Monte Carlo or Latin Hypercube sampling 23 techniques.

24 25 The NRC designed its SAPHIRE software development and maintenance program to provide an 26 analytical tool that performs risk calculations accurately and efficiently and reports the results in 27 a clear and concise manner to support risk-informed decisionmaking. Idaho National 28 Laboratory has created a software quality assurance program to ensure SAPHIRE continues to 29 meet its requirements as new features and changes are implemented.

30 H-61 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-12 Simplified Diagram of Event Tree with Initiating Event (IE) and Two 3 Supporting Fault Trees 4

5 Standardized Plant Analysis Risk Models 6

7 The NRC established the Standardized Plant Analysis Risk (SPAR) model program to support 8 regulatory reviews and independent evaluations of risk-related issues. The SPAR models are 9 plant-specific NRC-developed PRA models using standardized modeling conventions and data.

10 This standardization allows agency risk analysts to efficiently use SPAR models for diverse 11 plant designs in support of various regulatory activities. The regulatory uses of SPAR models 12 include the following:

13 14

Inspection Program (e.g., Significance Determination Process Phase 3): Determine the 15 risk significance (with respect to CDF and LERF) of inspection findings or of events to 16 decide (1) the allocation and characterization of inspection resources, (2) the initiation of 17 an inspection team, or (3) the need for further analysis or action by other agency 18 organizations.

19 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-62

1

Management Directive 8.3, NRC Incident Investigation Program: Estimate the risk 2 significance of events or conditions at operating NPPs so the agency can analyze and 3 evaluate the implications of plant operating experience to (1) compare the operating 4 experience with the results of licensee PRAs, (2) identify risk-significant conditions that 5 need additional regulatory attention, (3) identify conditions that need less regulatory 6 attention, and (4) evaluate the risk impact of regulatory or licensee programs.

7 8

Accident Sequence Precursor Program: Screen and analyze operating experience data 9 using a systematic approach to identify those events or conditions that are precursors to 10 severe accident sequences (core damage events).

11 12

Generic Issues Program: Provide the capability to resolve generic safety issues, both 13 for screening (or prioritization) and conducting a more rigorous analysis to (1) determine 14 if licensees should be required to make a change to their plants or (2) assess if the 15 agency should modify or eliminate one or more existing regulatory requirements.

16 17

License Amendment Reviews: Enable the NRC staff to make risk-informed decisions on 18 plant-specific changes to the licensing basis as proposed by licensees and provide risk 19 perspectives in support of agency reviews of licensee submittals.

20 21

Verification of Performance Indicators: Assist in (1) identifying threshold values for 22 risk-based performance indicators and (2) developing integrated or aggregate 23 performance indicators.

24 25

Special Studies: Undertake various studies in support of risk-informed regulatory 26 decisions (e.g., regulatory analysis and backfit analysis).

27 28

Operating Experience: Support and provide rigorous and peer reviewed evaluations of 29 operating experience, thereby demonstrating the agencys ability to analyze operating 30 experience independently of licensee PRAs and thus enhancing the technical credibility 31 of the agency.

32 33 The SPAR models allow agency risk analysts to perform independent evaluations of regulatory 34 issues without reliance on licensee-developed PRA models and analyses. The SPAR models 35 integrate systems analysis, accident scenarios, component failure likelihoods, and human 36 reliability analysis into a coherent model that reflects the design and operation of a specific 37 plant. These models give agency risk analysts the capability to (1) quantify the expected risk of 38 an NPP in terms of CDF or LERF, (2) identify and understand the attributes that significantly 39 contribute to risk, and (3) develop insights on how to manage that risk.

40 41 The SPAR models use an NRC-developed standard set of event trees and standardized input 42 data for initiating event frequencies, equipment performance, and human performance.

43 However, these input data may be modified to be more plant- or event-specific, when needed.

44 The system fault trees contained in the SPAR models are generally not as detailed as those 45 contained in licensee PRA models. However, SPAR models may need to be more advanced in 46 some areas than licensee PRA models (e.g., modeling of support system initiating events and 47 electrical power recovery). The staff has performed detailed cut set reviews for all SPAR 48 models to (1) more accurately model plant operation and configuration and (2) identify 49 significant differences between licensee PRAs and the corresponding SPAR models.

50 H-63 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 In addition to internal events, at-power models, the staff has developed the following models for 2 a subset of units: (1) external event models based on the licensee responses to Generic 3 Letter 88-20, Supplement 4, Individual Plant Examination of External Events for Severe 4 Accident Vulnerabilities, dated June 28, 1991, (2) low-power/shutdown models, and 5 (3) extended Level 1 PRA models supporting limited Level 2 PRA modeling and quantification of 6 LERF. SPAR model development work in these areas is ongoing. The staff has updated all 7 internal events models to include FLEX modeling. Additionally, the staff has developed 8 design-specific internal events SPAR models for new reactor designs and is developing a plant 9 specific new reactor SPAR model.

10 11 The staff has developed a formal SPAR model quality assurance plan and the Risk Assessment 12 Standardization Project Handbook. The SPAR model quality assurance plan provides 13 reasonable assurance that the SPAR models used by agency risk analysts represent the 14 as-built, as-operated plants to the extent intended within the scope of the SPAR models. As 15 part of this plan, the staff periodically updates the SPAR models for operating NPPs to reflect 16 the most recent operating experience and reliability data, performing routine updates to 17 approximately 6 SPAR models per year. The Risk Assessment Standardization Project 18 Handbook implements a formal, written process for maintaining SPAR models that are 19 sufficiently representative of the as-built, as-operated plants to support model uses. The staff 20 and Idaho National Laboratory also developed a SAPHIRE quality assurance program that is 21 compliant with NUREG/BR-0167, Software Quality Assurance Program and Guidelines, and 22 developed and released SAPHIRE Version 8, issued February 1993, which was independently 23 verified and validated.

24 25 American Society of Mechanical Engineers and American Nuclear Society PRA 26 Standard 27 28 In 2009, the staff, along with peer review teams comprised of industry experts, performed a peer 29 review of a representative boiling-water reactor SPAR model and a representative 30 pressurized-water reactor SPAR model in accordance with the American Society of Mechanical 31 Engineers (ASME) and American Nuclear Society (ANS) PRA Standard, ASME RA-S-2002, 32 Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications, and 33 Regulatory Guide 1.200, An Approach for Determining the Technical Adequacy of Probabilistic 34 Risk Assessment Results for Risk-Informed Activities. The peer review teams concluded 35 thatwithin constraints on access to licensee data and resourcesthe SPAR models are an 36 appropriate tool to provide a check and to prompt questions on the licensee-maintained and 37 peer reviewed PRA. The staff therefore concluded that SPAR models are an efficient tool for 38 obtaining qualitative and quantitative insights for agency risk-informed applications.

39 40 Severe Accident Progression and Source Term Analysis 41 42 The MELCOR Code 43 44 The MELCOR code is a fully integrated, engineering-level computer code designed to model the 45 progression of a broad spectrum of postulated severe accidents in light-water reactors and in 46 nonreactor systems (e.g., spent fuel pool and dry cask). MELCOR has been under continuous 47 development by the NRC and Sandia National Laboratories. Current activities involve the 48 development and implementation of new and improved models to predict the severe accident 49 behavior of various reactor (both light water and nonlight water) and spent fuel pool designs and 50 to reduce modeling uncertainties. In addition, enhancements and more flexibility are being NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-64

1 added to the code to evaluate the safety of accident-tolerant fuel designs. MELCOR represents 2 the current state-of-the-art in accident progression analysis, which has developed from domestic 3 and international research. The MELCOR code development meets the following criteria:

4 5

The prediction of phenomena is in qualitative agreement with the current 6 understanding of physics, and uncertainties are in quantitative agreement with 7 experiments.

8 9

The focus is on mechanistic models, where feasible, with adequate flexibility for 10 parametric models.

11 12

The code is portable, robust, and relatively fast running, and its maintenance 13 follows established Software Quality Assurance standards.

14 15

Detailed code documentation (including user guide, model reference, and 16 assessment) is available.

17 18 The NRC uses MELCOR to model severe accident progression and to compute the resulting 19 source terms for use in plant-specific PRAs and regulatory and backfit analyses. Recent 20 examples include the technical bases for the following NRC studies:

21 22

Enclosure H-3, Summary of Detailed Analyses for SECY-12-0157, of this appendix 23 summarizes the detailed analyses supporting SECY-12-0157, Consideration of 24 Additional Requirements for Containment Venting Systems for Boiling Water Reactors 25 with Mark I and Mark II Containments, dated November 26, 2012.

26 27

Enclosure H-4, Summary of Detailed Analyses for SECY-15-0085, of this appendix 28 summarizes the detailed analyses supporting SECY-15-0085, Evaluation of the 29 Containment Protection and Release Reduction for Mark I and Mark II Boiling-Water 30 Reactors Rulemaking Activities, dated June 18, 2015; the NRC subsequently published 31 the detailed analyses as NUREG-2206, Technical Basis for the Containment Protection 32 and Release Reduction Rulemaking for Boiling-Water Reactors with Mark I and Mark II 33 Containments, issued March 2018.

34 35

Enclosure H-5, Summary of Detailed Analyses for SECY-13-0112 and NUREG-2161, 36 of this appendix summarizes the detailed analyses supporting SECY-13-0112, 37 Consequence Study of a Beyond-Design-Basis Earthquake Affecting the Spent Fuel 38 Pool for a U.S. Mark I Boiling-Water Reactor, dated October 9, 2013, which was 39 documented in NUREG-2161, Consequence Study of a Beyond-Design-Basis 40 Earthquake Affecting the Spent Fuel Pool for a U.S. Mark I Boiling-Water Reactor, 41 issued September 2014.

42 43

Enclosure H-6, Summary of Detailed Analyses in COMSECY-13-0030, Enclosure 1, of 44 this appendix summarizes the detailed analyses supporting COMSECY-13-0030, Staff 45 Evaluation and Recommendation for Japan Lessons-Learned Tier 3 Issue on Expedited 46 Transfer of Spent Fuel, dated November 12, 2013.

47 48 Level 1 success criteria analyses have used MELCOR, as noted in Figure H-11 (see, for 49 example, NUREG/CR-7177, Compendium of Analyses to Investigate Select Level 1 H-65 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Probabilistic Risk Assessment End-State Definition and Success Criteria Modeling Issues, 2 issued May 2014). The discussion of the MACCS code below notes a variety of NRC research 3 studies that have used MELCOR. Additionally, some international organizations have used the 4 code to assess severe accident management strategies.

5 6 MELCOR Code Structure 7

8 MELCOR is a modular code consisting of three general types of packages: (1) basic physical 9 phenomena (i.e., hydrodynamicscontrol volume and flowpaths, heat and mass transfer to 10 structures, gas combustion, and aerosol and vapor physics), (2) reactor-specific phenomena 11 (i.e., decay heat generation, core degradation and relocation, ex-vessel [outside the reactor 12 vessel] phenomena, and engineering safety systems), and (3) support functions 13 (i.e., thermodynamics, equations of state, material properties, data-handling utilities, and 14 equation solvers). These packages model the major systems of an NPP and their associated 15 interactions. The various code packages have been written with well-defined interfaces 16 between them. This allows the exchange of complete and consistent information among them 17 so that all phenomena are coupled at every step.

18 19 MELCOR modeling makes use of a control volume approach in describing the plant system. No 20 specific nodalization (how the control volumes are defined) of a system is forced on the user, 21 which allows a choice of the degree of detail appropriate to the task at hand. Reactor-specific 22 geometry is imposed only in modeling the reactor core. Even here, one basic model suffices for 23 representing various core and fuel assembly designs, and a wide range of levels of modeling 24 detail is possible.

25 26 MELCOR Source Term 27 28 The MELCOR output binary plot file contains the time-dependent variables of interest as a 29 function of time at a frequency specified by the user. Of interest in Level 2 and Level 3 30 consequence analyses, MELCOR provides data on fluid flows and radionuclide transport to the 31 environment through flowpaths identified as release paths. This information constitutes the 32 source term and defines the magnitude and timing of the release of radionuclides. It is 33 characterized by the following MELCOR plot variables:

34 35

nominal aerosol density 36 37

fluid temperature 38 39

enthalpy 40 41

cumulative fluid mass flow 42 43

released radioactive mass for each radionuclide class 44 45

aerosol size distribution 46 47 This information can be converted into a MACCS input file by the MelMACCS preprocessor 48 code. The sections below describe MelMACCS, along with other associated codes.

49 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-66

1 ASME/ANS Level 2 PRA Standard 2

3 In January 2015, ASME/ANS issued for trial use ASME/ANS RA-S-1.2-2014: Severe Accident 4 Progression and Radiological Release (Level 2) PRA Standard for Nuclear Power Plant 5 Applications for LWRs. The NRCs Site Level 3 PRA Level 2 analysis team used a 6 prepublication draft of this trial use Level 2 PRA standard in a pilot application to perform a 7 self-assessment of its draft internal events and floods Level 2 PRA.

8 9 Severe Accident Consequence Analysis 10 11 The MELCOR Accident Consequence Code System (MACCS) 12 13 MACCS is the NRC code used to estimate the offsite consequences associated with a 14 hypothetical release of radioactive material into the atmosphere from a severe accident at an 15 NPP. The code models atmospheric transport and dispersion (ATD); mitigative actions based 16 on dose projections; dose accumulation by several pathways, including food and water 17 ingestion; early and latent health effects; and economic costs. MACCS is currently the only 18 code used in the United States for the offsite consequence analyses portion of NPP Level 3 19 PRAs.

20 21 As indicated in the main body of this NUREG, the NRC uses MACCS to estimate the averted 22 offsite property damage cost and the averted offsite dose cost elements in the performance of 23 cost-benefit analyses as part of backfit and regulatory analyses. The NRC has also used 24 MACCS to support calculations of individual latent cancer fatality and prompt fatality risks for 25 comparison to quantitative health objectives. As with the previous discussion on MELCOR, 26 recent examples in which the NRC used MACCS in regulatory analyses include SECY-12-0157, 27 SECY-15-0085, SECY-13-0112, and COMSECY-13-0030. The U.S. NPP license renewal 28 applicants use MACCS to support the plant-specific evaluation of severe accident mitigation 29 alternatives (SAMAs) that may be required as part of the applicants environmental report for 30 license renewal. Additionally, MACCS is used in severe accident analyses and severe accident 31 mitigation design alternative (SAMDA) assessments for environmental analyses supporting 32 design certification, early site permit, and combined construction and operating license reviews 33 for new reactors.

34 35 A variety of NRC research studies also used MACCS. The State-of-the-Art Reactor 36 Consequence Analyses (SOARCA) project used MELCOR and MACCS to develop best 37 estimates of the offsite radiological health consequences for potential severe reactor accidents 38 at Peach Bottom Atomic Power Station (Peach Bottom), the Surry Power Station, and the 39 Sequoyah Nuclear Plant. The MELCOR and MACCS best practices as applied in the 2012 40 SOARCA project were respectively documented in NUREG/CR-7008, MELCOR Best Practices 41 as Applied in the State-of-the-Art Reactor Consequence Analyses Project, and 42 NUREG/CR-7009, MACCS Best Practices as Applied in the State-of-the-Art Reactor 43 Consequence Analyses Project, both issued August 2014. Three SOARCA uncertainty 44 analyses have also been completed, including one for the Peach Bottom unmitigated long-term 45 station blackout, documented in NUREG/CR-7155, State-of-the-Art Reactor Consequence 46 Analyses Project: Uncertainty Analysis of the Unmitigated Long-Term Station Blackout of the 47 Peach Bottom Atomic Power Station, issued May 2016. These studies propagated uncertainty 48 for a variety of key uncertain MELCOR and MACCS parameters to develop insights into the 49 overall sensitivity of SOARCA results and conclusions to input uncertainty and to identify the 50 most influential input parameters for accident progression and offsite consequences. MACCS 51 was also used in a consequence study of a beyond-design-basis earthquake affecting the spent H-67 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 fuel pool for a U.S. Mark I boiling-water reactor and is documented in NUREG-2161. In 2 addition, the NRCs Full-Scope Site Level 3 PRA for a reference NPP site uses MACCS to 3 support the offsite consequence analyses.

4 5 MACCS Code Structure 6

7 The MACCS code is subdivided into three modules that handle the various components of the 8 consequence analysis calculation: ATMOS, EARLY, and CHRONC. These modules estimate 9 consequences in sequential steps:

10 11 1. ATMOS models atmospheric transport and deposition of radioactive materials onto land 12 and water bodies.

13 14 2. EARLY calculates the acute and lifetime doses, along with the associated health effects, 15 during the emergency phase simulation.

16 17 3. CHRONC calculates the estimated exposures and health effects during an intermediate 18 period of up to 1-year (intermediate phase) and computes the long-term (e.g., 50 years) 19 exposures and health effects (late-phase model). CHRONC also calculates the 20 economic costs of the intermediate and long-term protective actions, as well as the cost 21 of the emergency response actions in the EARLY module.

22 23 The following sections summarize the MACCS code models. More detailed descriptions appear 24 in the MACCS Code User Guide and Model Description, which includes NUREG/CR-4691, 25 MELCOR Accident Consequence Code System, issued February 1990 (NRC, 1990a) and 26 NUREG/CR-6613, Code Manual for MACCS2, issued May 1998 (NRC, 1998).

27 28 Atmospheric Transport and Dispersion 29 30 ATMOS models the dispersion of radioactive materials released into the atmosphere using the 31 straight-line Gaussian plume segment model with provisions for meander and surface 32 roughness effects. The ATD model treats buoyant plume rise, initial plume size caused by 33 building wake effects, release of up to 500 plume segments, dispersion under given 34 meteorological conditions, deposition under given dry and wet (precipitation) conditions, and 35 decay and ingrowths of up to 150 radionuclides and a maximum of six generations.

36 37 The analyst has the option of using a single weather sequence. Sampling among multiple 38 weather sequences is used in probabilistic consequence analysis studies to evaluate the 39 variability in consequences that can result from uncertain weather conditions at the time of a 40 future, hypothetical release of radioactive material. The results generated by the ATD model 41 include radionuclide concentrations in air, on land, and as a function of time and distance from 42 the release source; these results are subsequently used to model early, intermediate, and 43 long-term phase radiological exposure, as discussed below.

44 45 Early (Emergency) Phase Protective Actions and Exposure Pathways 46 47 The EARLY module in MACCS assesses the time period immediately following a radioactive 48 release while releases are ongoing. This is analogous to the emergency phase of a severe 49 accident. Early phase exposure calculations account for reductions in dose from the use of 50 emergency response measures such as sheltering, evacuation, and relocation of the population.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-68

1 MACCS models sheltering and evacuation for user-specified population cohorts. 24 Different 2 shielding factors for the different exposure pathways (i.e., cloudshine, groundshine, inhalation, 3 and deposition on the skin) are associated with three types of activities: (1) normal activity, 4 (2) sheltering, and (3) evacuation.

5 6 Intermediate Phase Protective Actions and Exposure Pathways 7

8 MACCS can model an intermediate phase following the end of the early phase. The only 9 protective action modeled in this phase is relocation. If the projected dose to a population 10 exceeds a user-specified threshold over a user-specified time duration, the population is 11 assumed to be relocated to an uncontaminated area for the entire duration of this phase. The 12 user defines a corresponding per-capita per diem economic cost. If the projected dose does not 13 reach the user-specified threshold, MACCS models exposure pathways for groundshine and 14 inhalation of resuspended material.

15 16 Long-Term Phase Protective Actions and Exposure Pathways 17 18 In the long-term phase, which follows the intermediate phase and can last, from months to 19 years, protective actions are defined to keep the dose to an individual below specified limits.

20 Protective actions in this phase include dose reduction measures, such as decontamination and 21 interdiction of contaminated areas. Decisions on protective actions are based on two sets of 22 independent criteria relating to whether land, at a specific location and time, is suitable for 23 human habitation (habitability) or agricultural production (farmability). Habitability and 24 farmability are defined by a set of user-specified maximum doses and a user-specified exposure 25 period to receive those doses. The long-term phase includes both direct exposure pathways 26 (i.e., groundshine, resuspension inhalation) and indirect exposure pathways through ingestion 27 (i.e., food and water consumption).

28 29 Health Effects Modeling 30 31 MACCS employs a user-specified dose conversion factor file based on the most recent 32 U.S. Environmental Protection Agency (EPA) guidance, currently, EPAs Federal Guidance 33 Report No. 13, Cancer Risk Coefficients for Environmental Exposure to Radionuclides, issued 34 September 1999. Federal Guidance Report No. 13 converts the integrated air concentration 35 and ground deposition of 825 radionuclides to a whole-body effective dose and individual organ 36 doses for 26 tissues and organs and for four exposure pathways. In general, the radiological 37 dose to a receptor (i.e., person) in each spatial element (i.e., an area of land) is the product of 38 the radionuclide concentration or quantity, the exposure duration, the shielding factor, the dose 39 conversion factor, and the usage factor (e.g., breathing rate). The total dose to an organ or the 40 whole body is then obtained by summation across the relevant exposure pathways and 41 radionuclides.

42 43 Offsite Consequence Measures 44 45 The results of a MACCS analysis can be reported in terms of population dose, health risks to 46 the public, land contamination, population subject to long-term protective actions, and economic 47 costs. Consequence results discussed in this section are conditional consequences 48 (i.e., assuming the accident occurs). Therefore, this section does not consider the different 24 Cohorts are subsets of the population with similar characteristics (e.g., school children in school at the time of the accident).

H-69 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 probabilities or frequencies of the different accident progression scenarios. Typical cost-benefit 2 analyses and SAMDA/SAMA analyses generally report the individual risks, population dose, 3 and economic costs as mean values (i.e., expected values). The values are averaged over 4 sampled weather conditions representing a year of meteorological data and over the entire 5 residential population within a circular or annular region. Past PRA applications have also 6 shown complementary cumulative distribution functions of these consequence measures (the 7 outputs of analysis), illustrating variability across weather conditions (inputs to the analysis).

8 9 Population Dose 10 11 As noted above, in general, the radiological dose to a receptor in each spatial element is the 12 product of the radionuclide concentration or quantity, the exposure duration, the shielding factor, 13 the dose conversion factor, and the usage factor (e.g., breathing rate). The total dose to an 14 organ or the whole body is then obtained by summation across the relevant exposure pathways 15 and radionuclides. Long-term population dose results are summed over the user-specified 16 areas of interest and reported in person-Sieverts.

17 18 Individual (Population-Weighted) Latent Cancer Fatality Risk and Early Fatality Risk 19 20 The individual, population-weighted, latent cancer fatality 25 risk calculations include only the 21 direct exposure pathways (i.e., groundshine, cloudshine, cloud inhalation, and resuspension 22 inhalation) and exclude the ingestion (i.e., consumption of food and water) pathways. The 23 MACCS early fatality model provides a pooled risk estimate of death from any of a number of 24 competing causes of early death, such as hematopoietic, gastrointestinal, and pulmonary 25 syndromes. Only the early phase exposure pathways are considered in the calculation of 26 individual early fatality risk. The individual latent cancer fatality and early fatality risks are 27 computed over user-specified regions. For example, for a large light-water reactor, a 10-mile 28 radius circular region centered on the plant is used, for purposes of comparison to the latent 29 cancer fatality risk quantitative health objective, and within 1 mile of the site boundary is used, 30 for purposes of comparison to the prompt fatality risk quantitative health objective (NRC, 1986).

31 32 Economic Consequences 33 34 The offsite economic consequences model in MACCS estimates the direct offsite costs that 35 result from protective actions modeled to reduce radiation exposures to the public. The current 36 cost-based economic model treats the following costs:

37 38

Evacuation costs: The daily cost of compensation for evacuees could include food, 39 housing, transportation, and lost income.

40 41

Relocation costs: The costs associated with relocating individuals during the 42 intermediate and long-term phases.

43 44

Decontamination of property: Costs are to decontaminate inhabited areas and farmland.

45 46

Loss of use: Economic losses from loss of return on investment and depreciation of 47 property value are incurred while property is temporarily interdicted. The depreciation of 48 value of the buildings and other structures results from lack of habitation and 49 maintenance.

25 This is a fatal cancer incurred from radiological exposure.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-70

1 2

Condemnation of property: Economic losses result from the permanent interdiction of 3 property.

4 5

Disposal of contaminated farm products and interdiction of farming: The economic cost 6 is from the loss of sales of farm products.

7 8 To obtain the total offsite economic costs, all the costs for the six cost categories are summed 9 over the entire region of interest affected by the atmospheric release. Many of the values 10 affecting the economic cost model are user inputs and thus can account for a variety of costs 11 and can be adjusted for inflation, new technology, or changes in policy or practices.

12 13 Ongoing Updates 14 15 Work is ongoing to update the MACCS code to include additional state-of-practice modeling 16 approaches (SECY-12-0110, Consideration of Economic Consequences within the 17 U.S. Nuclear Regulatory Commissions Regulatory Framework, Enclosure 9, MELCOR 18 Accident Consequence Code System, Version 2 (MACCS2), dated August 14, 2012).

19 Alternate ATD models are being implemented within MACCS by adding the capability to use 20 results from the National Oceanic and Atmospheric Administrations HYbrid Single-Particle 21 Lagrangian Integrated Trajectory (HYSPLIT) code (Stein et al., 2015). This will allow the use of 22 models that may provide a better representation of atmospheric transport, dispersion, and 23 deposition at longer ranges or in complex windfields. In addition, an alternative economic model 24 will use regional gross domestic product-based input-output models to capture the upstream 25 supply chain impacts of affected industries outside areas directly affected by radiological 26 releases.

27 28 Associated Codes 29 30 WinMACCS 31 32 WinMACCS is a graphical user interface that assists the user in constructing and executing 33 MACCS input files. The graphical user interface acts as a wizard that identifies what input is 34 necessary for a particular calculation. WinMACCS allows the user to interact with graphical 35 tools to aid in user input by visualization, such as defining an evacuation network using a map 36 with the polar grid superimposed.

37 38 MelMACCS 39 40 MelMACCS is a graphical user interface that converts source term information from the severe 41 accident analysis code MELCOR into a form suitable for use in the consequence analysis code 42 MACCS. MelMACCS processes MELCOR information for use in the ATMOS package of 43 MACCS for atmospheric transport and dispersion. Not all MACCS variables for source term 44 input are directly obtained from a MELCOR plot file. The variables not provided are either 45 calculated from other values in the plot file or are requested in the MelMACCS interface.

46 47 SecPop 48 49 SecPop is a preprocessor code for MACCS that enables the use of site-specific population, 50 economic, and land use data in the calculation of offsite consequences. SecPop uses a 51 block-level database of the U.S. population based on the U.S. Census and county-level data for H-71 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 economic information from the U.S. Department of Agriculture Census of Agriculture and 2 Bureau of Economic Analysis. SecPop allows the user to scale population and economic data 3 from the database years to a target year based on a user-specified growth rate. The output of 4 SecPop is a site file that is input into MACCS. NUREG/CR-6525, Revision 2, SecPop Version 5 4: Sector Population, Land Fraction, and Economic Estimation Program, issued June 2019, 6 provides more information.

7 8 COMIDA2 9

10 COMIDA2 is a preprocessor code that models the food-chain dose pathway. COMIDA2 can 11 calculate estimates of radionuclide concentrations in agricultural products after a radioactive 12 release following a hypothetical severe accident. This code calculates the uptake of 13 radioisotopes into the edible portions of plants as a function of the development of the plant. It 14 also considers the decay chains of nuclides, up to four daughters, and can, therefore, consider 15 the loss and ingrowth of radioisotopes in the plant.

16 17 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-72

1 ENCLOSURE H-2:

SUMMARY

OF THE STATE-OF-THE-ART REACTOR 2 CONSEQUENCE ANALYSES (SOARCA) PROJECT 3

4 Project Overview 5

6 The U.S. Nuclear Energy Commission (NRC) initiated the State-of-the-Art Reactor 7 Consequence Analyses (SOARCA) project to further its understanding of the realistic 8 consequences of severe reactor accidents. SOARCA addresses the consequences of rare but 9 severe accidents at commercial reactors in the United States. The SOARCA analysts focused 10 on accident progression, source term, and conditional consequences should the postulated 11 accidents occur. The project did not include within its scope new work to calculate the 12 frequencies associated with the postulated severe accidents.

13 14 The project, which began in 2006, combined information available at the time about the pilot 15 plants layout and operations, local population and site data, and emergency preparedness 16 plans. The NRC analyzed information using the MELCOR and MELCOR Accident 17 Consequence Code System (MACCS) suite of computer codes for integrated severe accident 18 progression and offsite consequence modeling. The modeling incorporated insights from 19 decades of research into severe reactor accidents.

20 21 Plants and Accident Scenarios Studied 22 23 The NRC staff initially evaluated potential consequences of select, important severe accidents 24 at the Peach Bottom Atomic Power Station (Peach Bottom) and Surry Power Station (Surry) 25 (NRC, 2012a). Selected accidents included station blackout scenarios for both plants and 26 bypass scenarios for Surry. Peach Bottom is a General Electric boiling-water reactor with a 27 Mark I containment, located in Pennsylvania; Surry is a Westinghouse 3-loop pressurized-water 28 reactor (PWR) with a subatmospheric large, dry containment, located in Virginia. The staff 29 subsequently evaluated a more limited set of scenarios at a third plant, the Sequoyah Nuclear 30 Plant (Sequoyah), a Westinghouse 4-loop PWR with an ice condenser containment, located in 31 Tennessee (NRC, 2019a). The Sequoyah study focused on issues unique to the ice condenser 32 containment design because of its lower design pressure and smaller volume. For this third 33 study, the staff also conducted an uncertainty analysis for one of the scenarios concurrently with 34 the deterministic calculations, in which it conducted uncertainty analyses for one scenario each 35 at the Peach Bottom and Surry plants after the initial deterministic SOARCA calculations 36 (NRC, 2016b and NRC, 2015a, a draft that will be updated for the Surry uncertainty analysis).

37 38 The SOARCA projects main findings fall into three basic areas: how a reactor accident 39 progresses, how existing systems and emergency measures can affect an accidents outcome, 40 and how an accident would affect public health. The 2012 project findings, corroborated by 41 subsequent uncertainty analyses and the Sequoyah analyses, include the following:

42 43

Existing resources and procedures can stop an accident, slow it down, or reduce its 44 impact before it can affect public health, if successfully implemented.

45 46

Even if accidents proceed without successful intervention, they generally take longer to 47 happen and release less radioactive material within the simulation time than earlier 48 analyses suggested. Hence, some accidents that may have been traditionally classified 49 as large-early release scenarios (e.g., interfacing systems loss-of-coolant accident for H-73 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Surry) may no longer contribute to large early release frequency because release is 2 delayed beyond the time assumed to successfully evacuate the close-in population.

3 4

The analyzed accidents pose essentially zero risk of early death (from radiological 5 consequences) and only a negligible increase in the risk of a long-term cancer death, to 6 a member of the public.

7 8

The small risk for the calculated individual cancer fatalities is dominated by the long-term 9 accumulation of very small doses (below allowable habitability criteria) to the public in 10 the affected area.

11 12 The NRC makes supporting technical information available on the deterministic Peach Bottom 13 analysis and Surry analysis in NUREG/CR-7110, State-of-the-Art Reactor Consequence 14 Analyses Project: Peach Bottom Integrated Analysis, Volume 1, issued May 2013 15 (NRC, 2013a), and NUREG/CR-7110, State-of-the-Art Reactor Consequence Analyses Project:

16 Surry Integrated Analysis, Volume 2, issued August 2013 (NRC, 2013b). NUREG/BR-0359, 17 Modeling Potential Reactor Accident Consequences, issued December 2012, describes this 18 Peach Bottom and Surry research for a general audience. The Peach Bottom uncertainty 19 analysis of the unmitigated long-term station blackout (LTSBO) scenario is available in 20 NUREG/CR-7155, State-of-the-Art Reactor Consequence Analyses Project: Uncertainty 21 Analysis of the Unmitigated Long-Term Station Blackout of the Peach Bottom Atomic Power 22 Station, issued May 2016. The Sequoyah integrated deterministic and uncertainty analyses 23 are available in NUREG/CR-7245, State-of-the-Art Reactor Consequence Analyses (SOARCA) 24 Project: Sequoyah Integrated Deterministic and Uncertainty Analyses, issued October 2019 25 (NRC, 2019a). The Surry uncertainty analysis of the unmitigated short-term station blackout 26 (STSBO), including a potential induced steam generator tube rupture, is available in 27 NUREG/CR-7262, State-of-the-Art Reactor Consequence Analyses (SOARCA) Project:

28 Uncertainty Analysis of the Unmitigated Short-Term Station Blackout of Surry Power Station, 29 issued in 2020 (NRC, 2020).

30 31 Results of the Mitigated Scenarios 32 33 One of the goals of the original Peach Bottom and Surry SOARCA analyses was to study the 34 benefits of the then-recently established mitigation measures in Title 10 of the Code of Federal 35 Regulations (10 CFR) 50.54(hh) (formerly B.5.b) for the accidents analyzed. All mitigated cases 36 of SOARCA scenarios, except for one, result in prevention of core damage or no offsite release 37 of radioactive material. The only mitigated case still leading to an offsite release was the Surry 38 STSBO-induced steam generator tube rupture. In this case, mitigation is still beneficial in that it 39 keeps most radioactive material inside containment and delays the onset of containment failure 40 by about 2 days (NRC, 2012a). The NRC made no attempt to quantify the likelihood that 41 mitigation would be successful and conducted no human reliability analysis. Instead, the 42 scenarios were analyzed twiceone case assuming that mitigation was successful and an 43 unmitigated case assuming successful mitigation did not occur.

44 45 The mitigated scenarios show zero individual early fatality risk from radiation exposure and zero 46 risk or a very small risk of long-term cancer fatalities, depending on the specific scenario. The 47 SOARCA results demonstrate the potential benefits of the mitigation measures analyzed in this 48 project. SOARCA shows that successful mitigation either prevents core damage or prevents, 49 delays, or reduces offsite health consequences.

50 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-74

1 The NRC was nearing completion of the SOARCA analyses when the accident at the 2 Fukushima Dai-ichi plants in Japan occurred in 2011. The NRC did not redefine or reanalyze 3 the scenarios following the Fukushima accident. It included a brief comparison to the 4 Fukushima Dai-ichi nuclear power plant accident in the Peach Bottom uncertainty analysis 5 technical report (NRC, 2016b). None of the SOARCA analyses included the use of flexible 6 coping strategies (FLEX) because FLEX was still under development at the time of the analysis.

7 8 Results of Unmitigated Scenarios 9

10 Even the unmitigated scenarios result in essentially zero individual early fatality risk from 11 radiation exposure. Although these unmitigated scenarios result in core damage and release of 12 radioactive material to the environment, the release is delayed, which allows the population to 13 take protective actions (including evacuation and sheltering). The individual risk of long-term 14 cancer fatality is calculated to be very small. Table H-8 shows the point estimates 15 (NRC, 2012a; NRC, 2019a), as well as uncertainty analysis bands where available 16 (NRC, 2016b; NRC, 2019a; NRC, 2020), for the conditional risk (assuming that the accident 17 occurs) to the public living between 0 and 10 miles from the plants, assuming the linear no-18 threshold dose response model. The SOARCA analyses calculated risk to individuals out to 50 19 miles from the plants. For some scenarios, the risks to the 10- to 30-mile population (outside 20 the plume exposure pathway emergency planning zone) are slightly higher than the risk to the 21 0- to 10-mile population. Considering that the frequencies estimated for these scenarios are in 22 the range of one per 100,000 to one per 30 million reactor-years, the absolute risk of long-term 23 cancer fatality from the analyzed SOARCA scenarios is projected to be negligible.

24 25 Table H-8 Conditional Annual Average Individual Latent Cancer Fatality Risk from 26 SOARCA Unmitigated Scenarios within 10 miles of the Plant Peach Bottom Surry Sequoyah Scenario Induced LTSBO STSBO LTSBO STSBO ISLOCA STSBO SGTR Point estimatea 9x10-5 2x10-4 5x10-5 9x10-5 3x10-4 3x10-4 8x10-5 5th percentileb 3x10-5 3x10-7 1x10-8 N/A N/A N/A N/A 95th percentileb 4x10-4 2x10-4 2x10-4 27 a The Peach Bottom and Surry accident simulations were carried out to 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months ; the Sequoyah accident was 28 simulated out to 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months .

29 b The Peach Bottom uncertainty analysis simulation was carried out to 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months ; the Surry and Sequoyah uncertainty 30 analysis simulations were carried out to 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months . The Surry STSBO 5th and 95th percentiles include induced steam 31 generator tube rupture (SGTR).

32 33 Notable Assumptions 34 35 The SOARCA models assume that 99.5 percent of the population residing in the 10-mile 36 emergency planning zone will evacuate as ordered. Shadow evacuationsthe voluntary 37 evacuation of members of the public who have not been ordered to evacuateare also 38 modeled for 10- to 15-mile or 10- to 20-mile radius annular rings around the plants. The 39 Sequoyah analysis explicitly considered the potential impact of the seismic initiating event on 40 emergency response and included sensitivity calculations for extended sheltering-in-place with 41 and without degraded shielding caused due to structural damage, in case evacuation is delayed 42 (NRC, 2019a). The Peach Bottom and Surry calculations assume the unmitigated accident 43 releases can be terminated within 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months . The Sequoyah calculation assumes releases can 44 be terminated within 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months .

H-75 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Uses of SOARCA Models and Insights 3

4 SOARCA models and insights were subsequently leveraged in a variety of projects, including 5 the analyses summarized in Enclosures H-3 through H-6 to this appendix. The NRC also 6 published research Information Letter 19-01, Benefits and Uses of the State-of-the-Art Reactor 7 Consequence Analyses (SOARCA) Project, issued 2019 (NRC, 2019c), which summarizes 8 many of the uses of the SOARCA body of work.

9 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-76

1 ENCLOSURE H-3:

SUMMARY

OF DETAILED ANALYSES FOR 2 SECY-12-0157, CONSIDERATION OF ADDITIONAL REQUIREMENTS 3 FOR CONTAINMENT VENTING SYSTEMS FOR BOILING WATER 4 REACTORS WITH MARK I AND MARK II CONTAINMENTS 5

6 This enclosure summarizes the 2012 analyses supporting the consideration of additional 7 requirements for containment venting systems for boiling-water reactors (BWRs) with Mark I 8 and Mark II containments, following the 2011 accident at the Fukushima Dai-ichi nuclear power 9 plant in Japan. The contents of this enclosure should be considered with the Commission 10 direction in its staff requirements memorandum (SRM)-SECY-12-0157, Consideration of 11 Additional Requirements for Containment Venting Systems for Boiling Water Reactors with 12 Mark I and Mark II Containments, dated March 19, 2013, and the subsequent analysis 13 described in Enclosure H-4, Summary of Detailed Analyses for SECY-15-0085, Evaluation of 14 the Containment Protection and Release Reduction for Mark I and Mark II Boiling-Water 15 Reactors Rulemaking Activities to this appendix. A summary of SRM-SECY-12-0157 is 16 provided at the end of this enclosure.

17 18 Problem Statement and Regulatory Objectives 19 20 The accident that occurred on March 11, 2011, at the Fukushima Dai-ichi nuclear power plant in 21 Japan underscored the potential need for nuclear power plant safety improvements related to 22 beyond-design-basis events involving natural hazards and their causal effects on plant systems 23 and barriers from an extended loss of electrical power and access to heat removal systems. As 24 part of its response to lessons learned from this accident, the U.S. Nuclear Regulatory 25 Commission (NRC) staff issued Order EA-12-050, Issuance of Order to Modify Licenses with 26 Regard to Reliable Hardened Containment Vents, dated March 12, 2012. This order required 27 licensees that use the boiling-water reactor (BWR) with Mark I and Mark II containment designs 28 to install hardened containment vents. These hardened containment vents would address 29 problems encountered during the Fukushima accident by providing plant operators with 30 improved methods for venting containment during accident conditions and thereby preventing 31 containment overpressurization and subsequent failure.

32 33 While developing the requirements for Order EA-12-050, the staff acknowledged that questions 34 remained about maintaining containment integrity and limiting the release of radiological 35 materials if licensees used the venting systems during severe accident conditions. In 36 SECY-11-0137, Prioritization of Recommended Actions to be Taken in Response to Fukushima 37 Lessons Learned, dated October 3, 2011, the staff also identified the addition of an engineered 38 filtered vent system to improve reliability and limit the release of radiological materials should 39 the venting systems be used after significant core damage had occurred.

40 41 Regulatory Alternatives 42 43 The NRC considered four regulatory alternatives that address containment venting systems for 44 BWRs with Mark I and Mark II containments in the regulatory analysis performed in support of 45 SECY-12-0157:

46 47

Option 1: Reliable Hardened Vents (Status Quo). Continue to implement 48 Order EA-12-050 and install reliable hardened vents to reduce the probability of failure of 49 BWR Mark I and Mark II containments and take no additional action to improve their H-77 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 ability to operate under severe accident conditions or to require the installation of an 2 engineered filtered vent system. This alternative represented the status quo and served 3 as the regulatory baseline against which the costs and benefits of other alternatives 4 were measured.

5 6

Option 2: Severe-Accident-Capable Venting System Order (without Filter). Upgrade or 7 replace the reliable hardened vents required by Order EA-12-050 with a containment 8 venting system designed and installed to remain functional during severe accident 9 conditions. This alternative would increase confidence in maintaining containment 10 functionality following core damage events. Although venting containment during severe 11 accident conditions may result in significant radiological releases, it would prevent 12 overpressurization and reduce the probability of gross containment failures that could 13 hamper accident management and result in larger radiological releases.

14 15

Option 3: Filtered Severe Accident Venting System Order. Design and install an 16 engineered filtered containment venting system that is intended to prevent the release of 17 significant amounts of radiological materials for dominant severe accident scenarios at 18 BWRs with Mark I and Mark II containments. The engineered filtering system would 19 need to operate under severe accident conditions to reduce the amount of radiological 20 material released to the environment from venting containment to prevent 21 overpressurization.

22 23

Option 4: Severe Accident Confinement Strategies. Pursue development of 24 requirements and technical acceptance criteria for confinement strategies and require 25 licensees to justify operator actions and systems or combinations of systems 26 (e.g., suppression pools, containment sprays, and engineered filters) to accomplish the 27 function and meet the requirements. For this option, the staff did not evaluate a specific 28 filtering system; instead, it drew on insights from various sensitivity studies to define a 29 possible approach.

30 31 Safety Goal Evaluation 32 33 This regulatory analysis required a safety goal evaluation because each of the alternatives was 34 considered a generic safety enhancement backfit subject to the substantial additional protection 35 standard in Title 10 of the Code of Federal Regulations (10 CFR) 50.109 (a)(3). Each 36 alternative, if implemented, would improve containment performance by reducing the probability 37 of containment failure, given the assumed occurrence of a severe accident scenario, or the 38 amount of radiological material released to the environment from a severe accident scenario, or 39 both. However, since none of the alternatives would impact the frequency of core damage 40 accidents (i.e., the change in core damage frequency (CDF) for each alternative relative to the 41 regulatory baseline was zero), the safety goal screening criteria in the regulatory analysis 42 guidelines could not be used to determine whether each alternative could result in a substantial 43 increase in overall protection of public health and safety.

44 45 Therefore, the Japan Lessons-Learned Steering Committee (NRC, 2011c) evaluated whether 46 imposition of requirements for severe-accident-capable or filtered venting systems would satisfy 47 the substantial additional protection standard. The Japan Lessons-Learned Steering Committee 48 decided that the staff should take the next step within the regulatory analysis process by 49 estimating and evaluating the costs and benefits.

50 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-78

1 Technical Evaluation 2

3 To support the assessment of the quantitative costs and benefits of severe-accident-capable 4 vents (Option 2) and filtered containment venting (Option 3), the staff (with support from Sandia 5 National Laboratories) analyzed selected accident scenarios for a BWR plant with a Mark I 6 containment. The staff used the NRCs severe accident analysis code, MELCOR, and the 7 MELCOR Accident Consequence Code System (MACCS) to perform the analysis. The staff 8 used the MELCOR code to calculate fission product release estimates for each of the selected 9 accident scenarios, and this information was used in MACCS to calculate the offsite radiological 10 consequences for each of the selected accident scenarios. Enclosure H-1, Description of 11 Analytical Tools and Capabilities, to this appendix describes these codes and their capabilities 12 in more detail.

13 14 Accident Scenario Selection 15 16 The selection of accident scenarios considered for MELCOR and MACCS analyses was 17 informed by both the State-of-the-Art Reactor Consequence Analyses (SOARCA) studies and a 18 study of the Fukushima accident that Sandia National Laboratories was performing at the time.

19 Two of the accident scenarios from the SOARCA study for Peach Bottom Atomic Power Station 20 (Peach Bottom) selected for MELCOR and MACCS analyses were (1) the long-term station 21 blackout (LTSBO) and (2) the short-term station blackout (STSBO).

22 23 MELCOR Severe Accident Progression and Source Term Analyses 24 25 Thirty MELCOR cases were run, simulating accident scenarios with different possible outcomes.

26 Cases 2, 3, 6, 7, 12, 13, 14, and 15 became MELCOR base cases, with the results used for 27 MACCS consequence calculations and for the regulatory analysis. The remaining cases were 28 run as variations of the base cases for sensitivity analyses. The base cases represented the 29 following accident scenarios:

30 31

Case 2: No venting or spray 32 33

Case 3: Wetwell venting but no spray 34 35

Case 6: Core spray only 36 37

Case 7: Core spray with wetwell venting 38 39

Case 12: Drywell venting 40 41

Case 13: Drywell venting and drywell spray 42 43

Case 14: Drywell spray only 44 45

Case 15: Drywell spray with wetwell venting 46 47 Collectively, the base cases encompassed all representative combinations of prevention and 48 mitigation measures considered in the description of alternatives used in the regulatory analysis.

49 Case 2 with no venting or spray mapped to Option 1 (status quo). Likewise, all venting cases 50 (Cases 3, 7, 12, 13, and 15) mapped to Option 2 (severe-accident-capable vent) andwhen H-79 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 considered in combination with an external filterto Option 3 (filtered vent). Case 6 and 2 Case 14 (both without venting but with sprays) were considered variations of Option 1.

3 4 The selected MELCOR accident scenarios were organized into four groups to compare the 5 effect of venting and additional mitigation actions:

6 7

Base case: Case 2 and Case 3 8

9

Core spray after reactor pressure vessel failure: Case 6 and Case 7 10 11

Main steamline failure with drywell venting at 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months s: Case 12 and Case 13 12 13

Drywell spray at 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months s: Case 14 and Case 15 14 15 MACCS Consequence Analyses 16 17 The analysts used MACCS to perform consequence analyses for selected accident scenarios to 18 calculate offsite doses and land contamination and their effect on members of the public with 19 respect to individual prompt and latent cancer fatality risk, land contamination areas, population 20 dose, and economic costs. They used the Peach Bottom unmitigated LTSBO MACCS input 21 deck from the SOARCA study, with two key modifications. One modification was the modeling 22 of the ingestion pathway, which was excluded in the SOARCA analyses. Another modification 23 was the use of revised source terms calculated from the MELCOR analyses for this study to 24 account for variation in the LTSBO scenario and the effect of adding an external filter to the vent 25 paths.

26 27 Risk Evaluation 28 29 The analysts constructed a simplified event tree to estimate the radiological release frequencies 30 of the MELCOR accident scenarios. Coupled with the MACCS consequence results developed 31 for each MELCOR scenario, this simplified event tree provided the information needed to 32 assess the reduction in risk resulting from the installation of a severe-accident-capable venting 33 system. The simplified event tree structure used to estimate radiological release frequencies 34 was designed to allow assessment of a wide range of severe-accident-capable vent system 35 designs that varied depending on (1) where the vent is attached (wetwell or drywell), (2) how the 36 vent is actuated (manually by the operator or passively using a rupture disk), and (3) whether 37 the severe-accident-capable venting system has a filter. Table H-9 identifies the nine 38 hypothetical plant modifications (Mods) that were assessed using the simplified event tree 39 structure.

40 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-80

1 Table H-9 Hypothetical Plant Modifications Severe-Accident- Severe-Accident- Severe-Accident-Identifier Capable Vent Filter Capable Vent Location Capable Vent Actuation Mod 0 NA None NA (current situation)

Mod 1 No Wetwell Manual Mod 2 No Wetwell Passive Mod 3 No Drywell Manual Mod 4 No Drywell Passive Mod 5 Yes Wetwell Manual Mod 6 Yes Wetwell Passive Mod 7 Yes Drywell Manual Mod 8 Yes Drywell Passive 2

3 The simplified event tree shown in Figure H-13 traced the accident progression starting from the 4 onset of core damage. The first two event tree headings parsed the total CDF according to the 5 type of hazard that initiated the accident (internal or external) and the type of core damage 6 sequence (station blackout [SBO] sequences, bypass sequences in which venting containment 7 has little or no impact because the containment is bypassed, fast sequences that evolve rapidly 8 and reduce the available time for the operator to manually open the severe-accident-capable 9 vent, and other sequences). Subsequent event tree headings consider (1) operation of the 10 severe-accident-capable vent, (2) offsite power recovery (which is influenced by the type of 11 hazard that initiated the accident), and (3) the availability of a water supply (portable pump) to 12 the drywell. Each sequence was assigned to one of four possible containment status end 13 states:

14 15

Vented: The severe-accident-capable vent is opened, preventing containment 16 overpressurization failure. A source of water to the drywell exists, preventing liner 17 melt-through.

18 19

Liner Melt-through (LMT): The severe-accident-capable vent is opened, preventing 20 containment overpressurization failure. No source of water to the drywell exists, and 21 liner melt-through occurs.

22 23

Overpressurization (OP): The severe-accident-capable vent is closed, resulting in 24 containment overpressurization failure. A source of water to the drywell exists, 25 preventing liner melt-through.

26 27

OP + LMT: The severe-accident-capable vent is closed, resulting in containment 28 overpressurization failure. No source of water to the drywell exists, and liner 29 melt-through occurs.

30 H-81 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-13 Simplified Event Tree Structure 3

4 This simplified event tree delineates 16 post-core-damage accident sequences. Each sequence 5 in the simplified event tree was assigned to a unique containment status. This mapping, 6 together with the definitions of the hypothetical plant modifications shown in Table H-10, 7 determined the specific MELCOR case and MACCS calculation that applies to each sequence, 8 as shown in Table H-11.

9 10 Table H-10 Mapping of Simplified Event Tree Sequences to Plant Modifications and 11 MELCOR Cases Modification Description Release Sequence Containment Status End State Vented OP + LMT LMT Sequence: 1, OP Sequence: 3, Mod Filter Location Actuation Sequence: 2, 4, 5, 10, and Sequence: 7 8, 9, 12, 15, 6, 11, and 14 13 and 16 0 NA NA None NA NA Case 6 Case 2 1 No Wetwell Manual Case 7 or 15 Case 3 Case 6 Case 2 2 No Wetwell Passive (no filter) (no filter) 3 No Drywell Manual Case 13 Case 12 Case 14 Case 2 4 No Drywell Passive (no filter) (no filter) 5 Yes Wetwell Manual Case 7 or 15 Case 3 Case 6 Case 2 6 Yes Wetwell Passive (filter) (filter) 7 Yes Drywell Manual Case 13 Case 12 Case 14 Case 2 8 Yes Drywell Passive (filter) (filter) 12 13 Analysts developed parameter values based on information from a variety of sources to 14 estimate the radiological release frequencies for each sequence in the simplified event tree.

15 Table H-11 summarizes this information.

16 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-82

1 Table H-11 Parameter Values Used to Estimate Radiological Release Frequencies Parameter Value Basis Standardized Plant Analysis Risk CDF 2.0x10-5 per reactor-year (ry)

(SPAR) external hazard models SPAR external hazard models; Fraction of total CDF due to external 0.8 review of previous probabilistic hazards risk assessments (PRAs)

Other 0.83 Breakdown of sequence types for SBO 0.12 SPAR internal hazard models internal hazardsa Bypass 0.05 Fast 0.01 Breakdown of sequence types for Other 0.95 Review of previous PRAs; external hazardsa Bypass 0.05 engineering judgment Mod 0 1 Vent not installed Mods 1, 3, 5, 7other SPAR-H method (manual vent; 0.3 or SBO longer available time)

Probability that severe-accident-capable SPAR-H method (manual vent; vent fails to open Mods 1, 3, 5, 7fast 0.5 shorter available time)

Engineering judgment (passive Mods 2, 4, 6, 8 0.001 vent mechanical failure)

Conditional probability that offsite power is not recovered by the time of lower Historical data 0.38 head failure given not recovered at the (NUREG/CR-6890) time of core damage (internal hazards)

SPAR-H; consistent with SPAR Probability that portable pump for core 0.3 B.5.b study by Idaho National spray or drywell spray fails Laboratory 2 a The values may not total to one due to rounding.

3 4 MACCS is used to calculate the mean conditional offsite radiological consequences per release, 5 conditioned on the assumed occurrence of the accident scenario that each MELCOR case 6 represented. Table H-12 provides the mean results for the 50-mile population dose and 50-mile 7 offsite cost consequence metrics.

8 9 Table H-12 Mean MACCS Consequence Results for Selected MELCOR Accident 10 Scenarios Core Drywell 50-mile Population Dose 50-mile Offsite Cost Casea,b Venting Location Spray Spray (person-rem/event) (million $/event) 2 no no no NA 514,000 1,910 3F no no yes wetwell 183,000 274 3NF no no yes wetwell 397,000 1,730 6 yes no no NA 305,000 847 7F yes no yes wetwell 37,300 18 7NF yes no yes wetwell 235,000 484 12F no no yes drywell 232,000 391 12NF no no yes drywell 3,810,000 33,300 13F no yes yes drywell 59,990 38 13NF no yes yes drywell 3,860,000 33,000 14 no yes no NA 86,100 116 15F no yes yes wetwell 43,300 20 15NF no yes yes wetwell 280,000 588 11 a F: filtered case 12 b NF: not filtered case 13 H-83 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 The analysts calculated risk by combining the frequencies of radiological releases with their 2 conditional offsite radiological consequences. Table H-13 provides the point estimate values for 3 the 50-mile population dose risk and the 50-mile offsite cost risk for each of the nine 4 hypothetical plant modifications.

5 6 Table H-13 Point Estimate Risk Values for Each Hypothetical Plant Modification Vent Vent Vent 50-mile Population Dose Risk 50-mile Offsite Cost Mod Filtered Location Actuation (person-rem/reactor-year [ry]) Risk ($/ry) 0 NA None NA 10.2 $37,884 1 No Wetwell Manual 7.2 $24,041 2 No Wetwell Passive 5.9 $18,117 3 No Drywell Manual 54.5 $452,466 4 No Drywell Passive 73.5 $630,000 5 Yes Wetwell Manual 4.5 $13,958 6 Yes Wetwell Passive 2.0 $3,717 7 Yes Drywell Manual 4.9 $14,540 8 Yes Drywell Passive 2.6 $4,642 7

8 9 Table H-14 provides the risk reductions (relative to Mod 0, the current situation) associated with 10 implementing plant modifications for the severe-accident-capable venting system (Mods 1 11 through 8). Figures H-14 and H-15 graphically illustrate this information.

12 13 Table H-14 Risk Reductions from Severe-Accident-Capable Venting System Plant 14 Modifications Reduction in 50-mile Vent Vent Vent Reduction in 50-mile Mod Population Dose Risk Filtered Location Actuation Offsite Cost Risk ($/ry)

(person-rem/ry) 1 No Wetwell Manual 3.0 $13,842 2 No Wetwell Passive 4.3 $19,767 3 No Drywell Manual (44.3)a ($414,582) 4 No Drywell Passive (63.3) ($592,117) 5 Yes Wetwell Manual 5.7 $23,926 6 Yes Wetwell Passive 8.2 $34,166 7 Yes Drywell Manual 5.3 $23,344 8 Yes Drywell Passive 7.6 $33,242 15 a Negative values are shown using parentheses (e.g., negative 44.3 is displayed as (44.3)).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-84

1 2 Figure H-14 Reduction in 50-mile Population Dose Risk (person-rem/ry) 3 4

5 Figure H-15 Reduction in 50-mile Offsite Cost Risk ($/ry) 6 H-85 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 To gain further insight into the risk reductions afforded by the hypothetical plant modifications, 2 analysts performed a simple parametric Monte Carlo uncertainty analysis. They assigned an 3 uncertainty distribution to each of the parameters used to quantify the radiological release 4 frequencies and to each of the consequences. Table H-15 shows parameters that specify the 5 uncertainty distribution.

6 7 Table H-15 Parameter Uncertainty Distributions Parameter Mean Distribution CDF 2.0x10-05/ry Lognormal; error factor = 10 Fraction of total CDF due to external 0.8 Beta; = 0.5, = 0.125 hazards Other 0.83 Dirichleta SBO 0.12 1 (other) = 41 Breakdown of sequence types for 2 (SBO) = 6 internal hazards Bypass 0.05 3 (bypass) = 2.5 Fast 0.01 4 (fast) = 0.5 Breakdown of sequence types for Other 0.95 Beta; (bypass) = 0.5, external hazards Bypass 0.05 (bypass) = 9.5 Mod 0 1 Held constant Probability that Mods 1, 3, 5, 7 0.3 Beta; = 0.5, = 1.167 severe-accident-capable vent fails to other or SBO open Mods 1, 3, 5, 7fast 0.5 Beta; = 0.5, = 0.5 Mods 2, 4, 6, 8 0.001 Beta; = 0.5, = 499.5 Conditional probability that offsite power is not recovered by the time of lower head failure given not 0.38 Beta; = 0.5, = 0.816 recovered at the time of core damage (internal hazards)

Probability that portable pump for 0.3 Beta; = 0.5, = 1.167 core spray or drywell spray fails Lognormal; error factor = 10 Within a given consequence Consequences Per Table H-6 category, consequences were assumed to be totally dependent.

8 a The Dirichlet distribution is a family of continuous multivariate probability distributions parameterized by a vector of 9 positive reals. It is a multivariate generalization of the Beta distribution. Dirichlet distributions are commonly used as 10 prior distributions in Bayesian statistics.

11 12 Figures H-16 and H-17 show the results 26 of the parametric uncertainty analysis. These figures 13 show that, although somewhat higher, the mean values are very close to the corresponding 14 point estimates. In general, the ratio of the 95th percentile to the point estimate varies from 15 3.5 to 4.0 depending on the consequence category. The major contributors to uncertainty in the 16 risk reduction results were uncertainty in both the CDF and the conditional consequences.

17 26 These figures do not show the results of Mods 3 and 4 because the results are negative (i.e., detrimental compared to the status quo), as shown in Figures H-16 and H-17.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-86

1 2 Figure H-16 Uncertainty in Reduction in 50-mile Population Dose Risk 3

4 5 Figure H-17 Uncertainty in Reduction in 50-mile Offsite Cost Risk 6

7 H-87 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 These risk results that incorporated insights from the MELCOR and MACCS analyses led to the 2 following specific conclusions about severe-accident-capable venting:

3 4

The installation of an unfiltered wetwell severe-accident-capable venting system would 5 reduce public health risk and offsite economic cost risk. By contrast, the installation of 6 an unfiltered drywell severe-accident-capable venting system would increase public 7 health risk and offsite economic cost risk.

8 9

The installation of a filtered severe-accident-capable venting system (attached to either 10 the wetwell or the drywell) would reduce public health risk and offsite economic cost risk.

11 The installation of an external filter into the severe-accident-capable venting system is 12 preferable.

13 14

By preventing containment overpressurization failure, the successful operation of a 15 severe-accident-capable venting system promotes access to plant areas where portable 16 pumps could be installed to provide core debris cooling.

17 18

Passive actuation (via a rupture disk) is preferred to manual actuation because it is more 19 reliable and thus results in larger risk reductions.

20 21

The uncertainty in the amount of risk reduction achieved by the installation of a 22 severe-accident-capable venting system comes mainly from uncertainty both in the CDF 23 and in the consequences resulting from radiological releases.

24 25 Cost-Benefit Analysis Results 26 27 The reductions in 50-mile population dose risk and 50-mile offsite cost risk (relative to Mod 0, 28 the current situation) associated with implementation of the severe-accident-capable venting 29 system plant modifications (Mods 1 through 8) were respectively used to calculate the values of 30 the public health and offsite property attributes for Options 2 and 3 in a cost-benefit analysis.

31 For the purposes of this analysis, Option 2 used the results for Mod 2 and Option 3 used the 32 results for Mod 6. These results corresponded to the plant design modifications that achieved 33 the largest risk reduction for each alternative.

34 35 Table H-16 summarizes the results of the quantitative cost-benefit analysis of a 36 severe-accident-capable (Option 2) and filtered vent system (Option 3) that used the regulatory 37 analysis guidelines that were in effect at the time. This table includes results for both the 38 base-case analysis that used the best estimate CDF value of 2.0x10-5 per reactor-year and a 39 one-way sensitivity analysis in which a CDF value of 2.0x10-4 per reactor-year was used to 40 evaluate the impact on the results of varying this important uncertain parameter.

41 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-88

1 Table H-16 Summary of Quantitative Cost-Benefit Analysis Results for Filtered 2 Containment Vent System using a $2,000 per Person-Rem Conversion 3 Factor Severe-Accident-Capable Engineered Filtered Venting Systems Venting Systems Attribute Base Casea Sensitivity Casea Base Casea Sensitivity Casea CDF=2.0x10 /ry -5 CDF=2.0x10-4/ry CDF=2.0x10 /ry -5 CDF=2.0x10-4/ry Public Health 150 1,500 290 2,900 Occupational Health 11 110 19 190 Offsite Property 348 3,480 600 6,000 Onsite Property 268 2,680 430 4,300 Industry (2,000)b (2,000) (15,000) (15,000)

Implementation Industry Operation n/a n/a (1,100) (1,100)

NRC Implementation (27) (27) (27) (27)

Net Benefit (1,250) 5,743 (14,778) (2,737) 4 a Values are in thousand dollars per unit.

5 b Negative values are shown using parentheses (e.g., negative 2,000 is displayed as (2,000)).

6 (Source: SECY-12-0157, Enclosure 1, Table 1) 7 8 At the time of the analysis, the staff was updating the dollar per person-rem conversion factor 9 policy and performed sensitivity analyses to evaluate the impact on results of increasing the 10 dollar per person-rem conversion factor from $2,000 per person-rem to $4,000 per person-rem.

11 Table H-17 summarizes the results of these sensitivity analyses.

12 13 Table H-17 Summary of Adjusted Quantitative Cost-Benefit Analysis Results for 14 Filtered Containment Vent System using a $4,000 per Person-Rem 15 Conversion Factor Severe-Accident-Capable Engineered Filtered Venting Systems Venting Systems Attribute Base Casea Sensitivity Casea Base Casea Sensitivity Casea CDF=2.0x10-5/ry CDF=2.0x10-4/ry CDF=2.0x10-5/ry CDF=2.0x10-4/ry Public Health 300 3,000 580 5,800 Occupational Health 22 220 38 380 Offsite Property 348 3,480 600 6,000 Onsite Property 268 2,680 430 4,300 Industry (2,000)b (2,000) (15,000) (15,000)

Implementation Industry Operation n/a n/a (1,100) (1,100)

NRC Implementation (27) (27) (27) (27)

Net Benefit (1,089) 7,353 (14,479) 353 16 a Values are in thousand dollars per unit.

17 b Negative values are shown using parentheses (e.g., negative 2,000 is displayed as (2,000)).

18 (Source: SECY-12-0157, Enclosure 1, Table 3) 19 20 Qualitative Factors 21 22 Because the net benefits for both Option 2 and Option 3 were negative for the base case, the 23 quantitative cost-benefit analysis did not appear to justify the imposition of additional 24 requirements on the venting systems for BWR Mark I and Mark II containments under 25 base-case assumptions. However, a one-way sensitivity analysis using a CDF value in the 26 upper range of its uncertainty band resulted in a positive net benefit for Option 2, indicating it 27 may be cost-beneficial. Moreover, a two-way sensitivity analysis within which the higher CDF H-89 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 value and a $4,000 per person-rem conversion factor was used resulted in a positive net benefit 2 for both Option 2 and Option 3, indicating that both options may be cost-beneficial, with Option 2 3 being the preferred alternative because of its greater net benefit.

4 5 However, in addition to performing these quantitative cost-benefit analyses, the staff considered 6 several qualitative factors in its regulatory analysis. For each qualitative factor, the staff 7 assigned a qualitative rating to each alternative. This qualitative rating used the number of 8 up-arrows to indicate the impact of considering that qualitative factor on the relative desirability 9 of the alternative. Table H-18 shows these qualitative ratings.

10 11 Table H-18 Ratings Assigned to Each Alternative by Qualitative Factor Qualitative Factor Option 1 Option 2 Option 3 Option 4 Defense-in-depth Uncertainties Severe accident management Hydrogen control External events Multiunit events Independence of barriers Emergency planning Consistency between reactor technologies Severe accident policy statement International practices 12 Source: Summarized from SECY-12-0157, Enclosure 1 13 14 Note: The analyst should refer to the Commissions response and direction on qualitative factors in 15 SRM-SECY-12-0157 and Appendix A, Qualitative Factors Assessment Tools, to this NUREG before 16 presenting qualitative factors in this manner.

17 18 Summary and Conclusion 19 20 The staff determined that many of the qualitative factors supported the following:

21 22

Pursuing an improved venting system for BWRs with Mark I and Mark II containments to 23 address specific design concerns (e.g., high conditional containment failure probability 24 given core melt) 25 26

Providing additional support for severe accident management functions by preventing 27 radiological releases, hydrogen, and steam from entering the reactor building or other 28 locations on the site 29 30

Minimizing the contamination of the site environment 31 32

Reducing the reliance on emergency planning for the protection of public health and 33 safety 34 35 Considering both the quantitative cost-benefit analysis results and the qualitative factors, the 36 staff further determined that Options 2 and 3, and most likely Option 4, were cost-justified, 37 based on the substantial increase in overall protection of public health and safety that would be 38 provided by addressing severe accident conditions for BWRs with Mark I and Mark II 39 containments.

40 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-90

1 Based on its regulatory analysis, the staff concluded that Option 3 (installation of engineered 2 filtered venting systems for Mark I and Mark II containments) was the alternative that would 3 provide the most regulatory certainty and the most timely implementation.

4 5 Commissions Response to the Staffs Analysis and Recommendations 6

7 The Commission approved Option 2 and directed the staff to further evaluate Options 3 and 4.

8 Enclosure H-4 to this appendix summarizes the staffs further evaluation of Options 3 and 4.

9 The Commission also directed the staff to seek detailed Commission guidance on the use of 10 qualitative factors in a future notation vote paper. In response, the staff submitted 11 SECY-14-0087, Qualitative Consideration of Factors in the Development of Regulatory 12 Analyses and Backfit Analyses, dated August 14, 2014, and developed Appendix A to this 13 NUREG.

14 H-91 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 ENCLOSURE H-4:

SUMMARY

OF DETAILED ANALYSES FOR 2 SECY-15-0085, EVALUATION OF THE CONTAINMENT PROTECTION 3 AND RELEASE REDUCTION FOR MARK I AND MARK II BOILING-4 WATER REACTORS RULEMAKING ACTIVITIES 5

6 This enclosure summarizes the detailed analyses supporting the evaluation of containment 7 protection and release reduction strategies for boiling-water reactor (BWR) plants with Mark I 8 and Mark II containments, as documented in SECY-15-0085, Evaluation of the Containment 9 Protection and Release Reduction for Mark I and Mark II Boiling-Water Reactors Rulemaking 10 Activities, dated June 18, 2015, as well as in NUREG-2206, Technical Basis for the 11 Containment Protection and Release Reduction Rulemaking for Boiling-Water Reactors with 12 Mark I and Mark II Containments, issued March 2018. The contents of this enclosure should 13 be considered with the previous detailed analyses supporting SECY-12-0157, Consideration of 14 Additional Requirements for Containment Venting Systems for Boiling Water Reactors with 15 Mark I and Mark II Containments, dated November 26, 2012. Enclosure H-3, Summary of 16 Detailed Analyses for SECY-12-0157, Consideration of Additional Requirements for 17 Containment Venting Systems for Boiling Water Reactors with Mark I and Mark II 18 Containments, to this appendix summarizes the detailed analyses for SECY-12-0157.

19 20 Problem Statement and Regulatory Objectives 21 22 The accident that occurred on March 11, 2011, at the Fukushima Dai-ichi nuclear power plant in 23 Japan underscored the importance of reliable operation of containment vents for BWR plants 24 with Mark I and Mark II containments. As part of its response to the lessons learned from this 25 accident, the staff of the U.S. Nuclear Regulatory Commission (NRC) issued Order EA-12-050, 26 Issuance of Order to Modify Licenses with Regard to Reliable Hardened Containment Vents, 27 dated March 12, 2012. This Order required licensees that operate BWRs with Mark I and 28 Mark II containment designs to install hardened containment vents. These vents would address 29 problems encountered during the Fukushima accident by providing plant operators with 30 improved methods for venting containment during accident conditions and thereby preventing 31 containment overpressurization and subsequent failure. In SECY-11-0137, Prioritization of 32 Recommended Actions to be Taken in Response to Fukushima Lessons Learned, dated 33 October 3, 2011, the staff also identified an issue involving containment vent filtration and 34 included a recommendation for the addition of an engineered filtered vent system to improve 35 reliability and limit the release of radiological materials if the venting systems are used in a 36 severe accident after the occurrence of significant core damage.

37 38 In SECY-12-0157, the staff analyzed whether additional requirements might be warranted to 39 address venting from BWRs with Mark I and Mark II containments after core damage and 40 whether filtering of radiological materials that may be released from the vents would be 41 necessary. The staff evaluated four regulatory options, including (1) the status quowhich 42 served as the regulatory baseline and assumed the staff would continue to implement 43 Order EA-12-050 and install reliable hardened vents to reduce the probability of failure of BWR 44 Mark I and Mark II containments but would take no additional action, (2) upgrade or replace the 45 reliable hardened vents required by Order EA-12-050 with a containment venting system 46 designed and installed to remain functional during severe accident conditions, (3) design and 47 install an engineered filtered containment venting system intended to prevent the release of 48 significant amounts of radioactive material following the dominant severe accident sequences at 49 BWRs with Mark I and Mark II containments, and (4) pursue development of requirements and 50 technical acceptance criteria for performance-based severe accident confinement strategies.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-92

1 The NRC staff provided an evaluation that considered both results from quantitative cost-benefit 2 analyses and qualitative factors related to the four options and recommended that the 3 Commission approve Option 3 to require the installation of an engineered filtering system.

4 While acknowledging that the quantitative analyses indicated the costs of the proposed actions 5 outweighed the benefits, the staff recommended in SECY-12-0157 that the Commission 6 consider both the quantitative and qualitative factors and concluded the proposed additional 7 regulatory actions associated with Option 3 were cost-justified.

8 9 In its staff requirements memorandum (SRM) for SECY-12-0157, dated March 19, 2013, the 10 Commission directed the staff to (1) issue a modification to Order EA-12-050 to require BWR 11 licensees with Mark I and Mark II containments to upgrade or replace the reliable hardened 12 vents required by Order EA-12-050 with a containment venting system designed and installed to 13 remain functional during severe accident conditions, and (2) develop technical bases and 14 pursue rulemaking for filtering strategies with drywell filtration and severe accident management 15 of BWR Mark I and Mark II containments. The Commission further ordered that the technical 16 bases should (1) assume that severe-accident-capable vents had been ordered and, as a 17 consequence of that action, should assume that the benefits of these vents accrue equally to 18 engineered filters and to filtration strategies, (2) explore requirements associated with measures 19 to enhance the capability to maintain confinement integrity and to cool core debris, and 20 (3) evaluate multiple performance criteria, including a required decontamination factor and 21 equipment and procedure availability like those required to implement Title 10 of the Code of 22 Federal Regulations (10 CFR) 50.54 (hh). 27 23 24 In response to SRM-SECY-12-0157, the staff issued Order EA-13-109, Issuance of Order To 25 Modify Licenses with Regard to Reliable Hardened Containment Vents Capable of Operation 26 Under Severe Accident Conditions, dated June 6, 2013, which rescinded certain requirements 27 imposed in Order EA-12-050 and required BWR licensees with Mark I and Mark II containments 28 to upgrade or replace their vents with a containment venting system designed and installed to 29 remain functional during severe accident conditions. Order EA-13-109 had two primary 30 requirements that would be implemented sequentially in two phases:

31 32 1. Phase 1: Upgrade the venting capabilities from the containment wetwell to provide 33 reliable, severe-accident-capable hardened vents to assist in preventing core damage 34 and, if necessary, to provide venting capability during severe accident conditions.

35 36 2. Phase 2: Either install a reliable severe-accident-capable drywell venting system or 37 develop and implement a reliable containment venting strategy that makes it unlikely that 38 a licensee would need to vent from the containment drywell during severe accident 39 conditions.

40 41 In response to Order EA-13-109, the severe accident water addition (SAWA) approach required 42 licensees to use water addition in combination with one of two strategies(1) a 43 severe-accident-capable drywell vent designed to lower temperature limits, or (2) severe 44 accident water management (SAWM) to control the water levels in the suppression pool such 45 that it would be unlikely that a licensee would need to vent from the containment drywell during 46 severe accident conditions (Nuclear Energy Institute (NEI), 2014).

47 27 The SRM for SECY-12-0157 provided additional directions which are addressed in SECY-15-0085.

H-93 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 With the issuance of Order EA-13-109, the staff also began developing the regulatory basis for 2 the containment protection and release reduction (CPRR) 28 rulemaking for BWRs with Mark I 3 and Mark II containments. The objective of the CPRR regulatory basis was to determine what, 4 if any, additional requirements were warranted on filtering strategies and severe accident 5 management for BWRs with Mark I and Mark II containments, assuming the installation of 6 severe-accident-capable hardened vents per Order EA-13-109.

7 8 Regulatory Alternatives 9

10 The staff interacted with industry and members of the public and identified four major regulatory 11 alternatives comprising numerous subalternatives for choices on filtering strategies and severe 12 accident management for BWRs with Mark I and Mark II containment designs. The four main 13 CPRR regulatory alternatives considered in the regulatory analysis performed in support of 14 SECY-15-0085 were the following:

15 16

Alternative 1: Severe-Accident-Capable Vents (Status Quo). Continue with the 17 implementation of Order EA-13-109 and installation of severe-accident-capable vents, 18 without taking additional regulatory actions related to BWR Mark I and Mark II 19 containments. This alternative represented the status quo and served as the regulatory 20 baseline against which the benefits and costs of other alternatives were measured.

21 22

Alternative 2: Rulemaking to Make Order EA-13-109 Generically Applicable. Pursue 23 rulemaking to make Order EA-13-109 generically applicable to protect BWR Mark I and 24 Mark II containments against overpressurization. The potential benefits associated with 25 this option resulted from making generically applicable the requirements in 26 Order EA-13-109 related to improved reporting, change control, and other aspects of 27 controlling licensing basis information.

28 29

Alternative 3: Rulemaking to Make Order EA-13-109 Generically Applicable and 30 Additional Requirements for SAWA to Address Uncontrolled Releases from Major 31 Containment Failure Modes. Pursue rulemaking to address overall BWR Mark I and 32 Mark II containment protection against multiple failure modes by making 33 Order EA-13-109 generically applicable and requiring external water addition points that 34 would allow water to be added into the reactor pressure vessel (RPV) or drywell to 35 prevent containment failure from both overpressurization and liner melt-through.

36 37

Alternative 4: Rulemaking to Reduce Releases during Controlled Venting (Filtering 38 Strategies, Engineered Filters). Pursue rulemaking to address both containment 39 protection against multiple failure modes and release reduction measures for controlling 40 releases through the containment venting systems. This alternative would make 41 Order EA-13-109 generically applicable and require external water addition into the RPV 42 or drywell. In addition, licensees would be required to reduce the fission products 43 released from containment by (1) implementing strategies to maximize the availability 44 and efficiency of the wetwell in scrubbing or filtering fission products before venting from 45 containment or (2) installing an engineered filter in the containment vent paths (or both).

46 28 As the rulemaking progressed, the staff determined that the original rulemaking name (filtering strategies) no longer matched the purpose of the activity. The staff believed it was more logical to have the rulemaking reflect the two issues being analyzedenhanced containment protection and release reduction.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-94

1 A CPRR strategy is an action taken before or during a severe accident to protect the 2 containments structural integrity or to reduce the amount of radiological material released to the 3 environment. Examples include containment venting following core damage (a containment 4 protection strategy) and the installation of engineered filters on the containment vent lines (a 5 release reduction strategy). Such high-level strategies can be divided into more specific 6 categories according to how they are implemented. From the four main regulatory alternatives 7 defined above, 20 regulatory subalternatives were defined by specific combinations of CPRR 8 strategies. These combinations of CPRR strategies considered many factors, including the 9 following:

10 11

Wetwell and drywell venting priority (before and after core damage) 12 13

Venting actuation (before and after core damage) 14 15

Venting operation mode (before and after core damage) 16 17

Vent reclosure if core damage is imminent 18 19

Postaccident water injection location and operating mode 20 21

Filter size and decontamination factor 22 23 Table 19 summarizes the 20 regulatory subalternatives, how each subalternative maps to the 24 options defined in SECY-12-0157 and the alternatives defined in SECY-15-0085, and the 25 combinations of CPRR strategies used to distinguish among them.

26 27 Safety Goal Evaluation 28 29 A safety goal evaluation for Alternative 3 and Alternative 4 was performed in this regulatory 30 analysis because these two main regulatory alternatives were considered generic safety 31 enhancement backfits subject to the substantial additional protection standard at 32 10 CFR 50.109(a)(3). Each alternative, if implemented, would improve containment 33 performance by reducing (1) the probability of containment failure, given the assumed 34 occurrence of a severe accident scenario, and/or (2) the amount of radiological material 35 released to the environment from a severe accident scenario. However, since none of the 36 alternatives would impact the frequency of core damage accidents (i.e., the change in core 37 damage frequency (CDF) for each alternative relative to the regulatory baseline was zero), the 38 safety goal screening criteria in the regulatory analysis guidelines could not be used to 39 determine whether each alternative could result in a substantial increase in overall protection of 40 public health and safety.

41 42 To perform the safety goal evaluation, the staff analyzed numerous regulatory alternatives to 43 directly compare their potential safety benefits to the quantitative health objectives (QHOs) for 44 average individual early fatality risk and average individual latent cancer fatality risk described in 45 the Commissions Safety Goal Policy Statement (NRC, 1986). Each of the alternatives was 46 compared to Alternative 1 (status quo and regulatory baseline) to determine the relative benefits 47 and costs of the alternative.

48 49 The staff determined there was zero average individual early fatality risk, conditioned on the 50 assumed occurrence of the modeled severe accident scenarios. In part this resulted from the H-95 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 fact that the modeled accident progression resulted in releases that begin late when compared 2 to the time needed to evacuate members of the public living near the modeled nuclear power 3 plant site.

4 5 Table H-19 Summary of Regulatory Subalternatives and Distinguishing Attributes Before Core Damage After Core Damage Postaccident Water SECY-12-0157 Option Venting Operation Mode Venting Operation Mode Reclose Valve if Core Postaccident Water Venting Actuation Venting Actuation Filter Size and DF Regulatory SECY-15-0085 Venting Priority Damage is Imminent Injection Operating Mode Venting Priority Subalternative Alternative Injection Location Index 1 1 2 NA WWF M AV Yes NA NA WWF M OLO NA 2 2A 2 NA WWF M AV Yes NA NA WWF M OLO NA 3 3A 2 1,2,3 WWF M AV Yes RPV SAWA WWF M OLO NA 4 3B 2 1,2,3 WWF M AV Yes DW SAWA WWF M OLO NA 5 4Ai(1) 4 4 WWF M AV Yes RPV SAWA WWF M VC NA 6 4Ai(2) 4 4 WWF M AV Yes DW SAWA WWF M VC NA 7 4Aii(1) 4 4 WWF M AV Yes RPV SAWM WWF M OLO NA 8 4Aii(2) 4 4 WWF M AV Yes DW SAWM WWF M OLO NA 9 4Aiii(1) 4 4 WWF M AV Yes RPV SAWM WWF M VC NA 10 4Aiii(2) 4 4 WWF M AV Yes DW SAWM WWF M VC NA 11 4Bi(1) 3 4 WWF M AV Yes RPV SAWA WWF M OLO S 12 4Bi(2) 3 4 WWF M AV Yes DW SAWA WWF M OLO S 13 4Bii 3 4 WWF M AV Yes DW SAWA DWF M OLO S 14 4Biii 3 4 WWF M AV Yes DW SAWA DWF P OLO S 15 4Biv 3 4 DWF P OLO No DW SAWA DWF P OLO S 16 4Ci(1) 3 4 WWF M AV Yes RPV SAWA WWF M OLO L 17 4Ci(2) 3 4 WWF M AV Yes DW SAWA WWF M OLO L 18 4Cii 3 4 WWF M AV Yes DW SAWA DWF M OLO L 19 4Ciii 3 4 WWF M AV Yes DW SAWA DWF P OLO L 20 4Civ 3 4 DWF P OLO No DW SAWA DWF P OLO L Venting Priority Postaccident Water Injection Location DWF: drywell first strategy DW: drywell via external connection WWF: wetwell first strategy RPV: reactor pressure vessel via external connection Venting Actuation Postaccident Water Injection Operating Mode M: manual SAWA severe accident water addition P: passive (rupture disc) SAWM severe accident water management Venting Operation Mode Filter Size and Decontamination Factor (DF)

AV: anticipatory venting L: large with DF of 1000 OLO: open at 15 psig and leave open S: small with DF of 10 VC: venting cycling at primary containment pressure limit with 10 psi band 6 (Source: NUREG-2206, Table 2-2) 7 8 The staff then performed a screening analysis for the average individual latent cancer fatality 9 risk QHO by evaluating all United States (U.S.) BWRs with Mark I containments (a total of 10 22 units at 15 sites) and Mark II containments (a total of eight units at five sites). For this 11 screening analysis, the staff developed a conservative high estimate of frequency-weighted NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-96

1 average individual latent cancer fatality risk within 10 miles using the following parameter 2 values:

3 4

An extended loss of alternating current power (ELAP) 29 frequency value of 7x10-5 per 5 reactor-yearwhich represented the highest value among all BWRs with Mark I and 6 Mark II containments 7

8

A success probability for flexible coping strategies (FLEX) equipment of 0.6 per 9 demandwhich assumed implementation of FLEX will successfully mitigate an accident 10 involving an ELAP 6 out of 10 times 11 12

A conditional average individual latent cancer fatality risk of 2x10-3 per eventwhich 13 represented the highest value among all BWRs with Mark I and Mark II containments 14 15 These assumed parameter values resulted in a conservative high estimate of 16 frequency-weighted individual latent cancer fatality risk within 10 miles of approximately 17 7x10-8 per reactor-year, which is greater than an order of magnitude less than the QHO for an 18 average individual latent cancer fatality risk of approximately 2x10-6 per reactor-year. This 19 conservative high estimate did not take credit for any of the accident strategies and capabilities 20 described in the 20 CPRR alternatives and subalternatives. Figure H-19 shows the incremental 21 benefit for each alternative and subalternative, compared to the status quo and Order 22 EA-13-109. If licensees were to choose to implement SAWA/SAWM as part of compliance with 23 EA-13-109, the uncertainty band for Alternative 3 would apply. However, since EA-13-109 did 24 not specifically require SAWA/SAWM, it was not credited in Figure H-18 for Alternative 1 or 25 Alternative 2.

26 27 If an ELAP occurs and results in core damage, an engineered filtered containment venting 28 system would reduce offsite consequences. However, because the average individual latent 29 cancer fatality risk within 10 miles for the status quo alternative (Alternative 1) was already well 30 below the associated QHO, the staff concluded that the design and installation of an engineered 31 filtered containment venting system or a performance-based confinement strategy for BWRs 32 with Mark I and Mark II containments would not meet the threshold for a substantial safety 33 enhancement. Moreover, although this analysis did not include all accident scenarios that a 34 full-scope Level 3 PRA would need to consider, the staff concluded that none of the alternatives 35 could result in a substantial increase in overall protection of public health and safety. Therefore, 36 the staff recommended that rulemaking not be pursued for SECY-12-0157 Option 3 or Option 4.

37 Furthermore, the staff concluded that a detailed regulatory analysis of the various alternatives 38 was not warranted and would provide little additional insight into the regulatory decision 39 because the margin to the QHOs did not support a substantial safety benefit.

40 29 An ELAP is defined as a station blackout (SBO) that lasts longer than the SBO coping duration specified in 10 CFR 50.63, Loss of all alternating current power.

H-97 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-18 Uncertainty in Average Individual Latent Cancer Fatality Risk (0-10 miles) 3 (Source: SECY-15-0085, Enclosure, Figure 3-3) 4 5 Technical Evaluation 6

7 Accident Scenario Selection 8

9 The staff considered the following factors during the development of the technical approach for 10 the accident sequence analysis performed for SECY-15-0085:

11 12

The risk evaluation should provide risk metrics for each of the 20 CPRR regulatory 13 analysis subalternatives, according to the schedule established by the Commission and 14 the resources allotted by NRC management.

15 16

Consistent with the NRCs regulatory analysis guidelines, the risk evaluation should 17 provide fleet-average risk estimates. Therefore, the technical approach should consider 18 the impacts of plant-to-plant variability.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-98

1 2

Consistent with Recommendation 5.1 in the Fukushima Near-Term Task Force (NTTF) 3 report, the accident sequence analysis should focus on accidents initiated by ELAP 4 events.

5 6

The generic estimates of release sequence frequencies and conditional consequences 7 in NUREG/BR-0184, Regulatory Analysis Technical Evaluation Handbook, issued 8 January 1997, were developed from previous probabilistic risk assessments (PRAs) that 9 did not consider CPRR strategies and therefore cannot be used to provide an adequate 10 technical basis for the CPRR risk evaluation.

11 12

Core damage event trees (CDETs) should be developed to (1) model the impact of 13 equipment failures and operator actions occurring before core damage that affect severe 14 accident progression and the probability that CPRR strategies are successfully 15 implemented, (2) match the initial and boundary conditions used in the thermal-hydraulic 16 simulation of severe accidents in MELCOR, and (3) probabilistically consider mitigating 17 strategies for beyond-design-basis external events required by Order EA-12-049, 18 Issuance of Order to Modify Licenses with Regard to Requirements for Mitigation 19 Strategies for Beyond-Design-Basis External Events, dated March 12, 2012.

20 21

The CPRR strategies addressed in the set of 20 regulatory analysis subalternatives are 22 specified at a conceptual level. Therefore, it is acceptable to develop high-level generic 23 accident progression event trees (APETs) to model the CPRR strategies because no 24 information is available about their specific design details.

25 26 Analysts used a modular approach to develop the CDETs and APETs, as shown in Figure H-19.

27 This modeling approach streamlined the development of risk estimates.

28 H-99 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-19 Modular Approach to Event Tree Development 3 (Source: NUREG-2206, Figure 2-1) 4 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-100

1 MELCOR Severe Accident Progression and Source Term Analyses 2

3 The MELCOR analyses addressed two main categories: (1) reactor systems and containment 4 thermal-hydraulics under severe accident conditions and (2) assessment of source termsthe 5 timing, magnitude, and other characteristics of fission product releases to the environment. The 6 first category provided insight into the state of containment vulnerability under severe accident 7 conditions and information needed to assess containment integrity. The second category 8 provided information needed to assess the offsite radiological consequences associated with 9 releases of radioactive materials to the environment.

10 11 The NRC based the development of the MELCOR calculation matrices (see Table 3-2 and 12 Table 3-3, NRC, 2018b) on the CPRR alternatives defined by the accident sequence analysis.

13 The MELCOR analyses investigated detailed accident progression, containment response, and 14 source terms for representative Mark I and Mark II containment designs following an ELAP.

15 The selection of accident scenarios considered for MELCOR analyses was informed by the 16 State-of-the-Art Reactor Consequence Analyses (SOARCA) Project (see Enclosure H-2, 17 Summary of the State-of-the-Art Reactor Consequence Analyses (SOARCA) Project, to this 18 appendix), the Fukushima Dai-ichi nuclear power plant accident reconstruction study (Sandia 19 National Laboratories, 2012), and the detailed analyses in SECY-12-0157. The representative 20 Mark I containment selected was similar in configuration to Peach Bottom Atomic Power Station 21 (Peach Bottom), Unit 2, and the representative Mark II containment was similar in configuration 22 to LaSalle County Station (LaSalle). The Mark I MELCOR calculation matrix included sensitivity 23 cases to evaluate the impact on results of using plausible alternative assumptions about 24 multiple factors, including (1) mode of venting, (2) status of RPV depressurization, (3) mode of 25 FLEX water injection, and (4) water management. The Mark II MELCOR calculation matrix 26 included a subset of the Mark I matrix, based on the insights from the Mark I MELCOR 27 calculations, and included sensitivity cases to evaluate the impact of the pedestal and lower 28 cavity designs among the fleet by modifying the base model.

29 30 The scope and technical approach for the MELCOR analyses performed in support of 31 SECY-15-0085 were similar to those of SECY-12-0157. In both cases, the technical approach 32 considered best estimate modeling of accident progression and incorporated both preventive 33 and mitigative accident management measures, including (1) venting, (2) water addition, water 34 management, or both, and (3) installation of engineered filters. However, an important 35 distinction between the technical approaches is that, in SECY-12-0157, water addition was 36 considered in a generic way because the industrys post-Fukushima Dai-ichi severe accident 37 management strategies were still evolving and the concepts of SAWA and SAWM had not yet 38 emerged. Moreover, the industry was formulating its FLEX strategy for severe accident 39 mitigation applications at the time. By contrast, these various concepts and severe accident 40 management measures were more mature by the time detailed analyses were performed for 41 SECY-15-0085 and were, therefore, considered in developing the technical approach for these 42 analyses.

43 44 MACCS Consequence Analyses 45 46 Like the MELCOR analyses, the scope and technical approach for the MACCS analyses 47 performed in support of SECY-15-0085 were similar to those of SECY-12-0157. The NRC used 48 MACCS to calculate offsite radiological consequences with site-specific population, economic, 49 land use, weather, and evacuation data for reference Mark I and Mark II sites. The agency 50 selected Peach Bottom and the Limerick Generating Station (Limerick) as the site-specific 51 reference models for the offsite consequence analyses to enable greater modeling fidelity for H-101 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 sites with relatively high population densities (Peach Bottom had the second highest population 2 within a 50-mile radius among the 15 Mark I sites and Limerick had the highest population within 3 a 50-mile radius among the five Mark II sites).

4 5 The staff performed offsite consequence analyses for the source terms generated by MELCOR 6 corresponding to different CPRR accident management strategies following an ELAP event. It 7 assessed the relative public health risk reduction associated with various containment protection 8 and release reduction measures with respect to various offsite radiological consequence 9 measures, including (1) average individual early fatality risk and average individual latent cancer 10 fatality risk, (2) population dose, (3) land contamination, (4) economic costs, and (5) displaced 11 population. Land contamination areas and displaced populations represented additional 12 consequence metrics that the staff reported for consideration by decisionmakers, although they 13 are not required as inputs to safety goal evaluations or regulatory analyses. The calculated 14 offsite radiological consequences were weighted by accident frequency to assess relative public 15 health risk reduction.

16 17 Tables H-20 and H-21 show the summary MACCS results respectively for the 18 Mark I and the 18 9 Mark II source term bins. As shown on the tables, the staff reported some consequence 19 metrics out to a 100-mile radius from the plant.

20 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-102

1 Table H-20 MACCS Results for 18 Mark I Source Term Bins

Hrs with Individiual Early Population Dose Start Individual Latent Cancer Fatality Risk Rep Case Rep Case Significant Fatality Risk (person-rem)

Bin Rep Case Time Cs (%) I (%) Cs (hrs)

Release* 0-1.3 mi and beyond 0-10 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 1 28DF1000 0.0006% 0.006% 14.9 7 0 4.65E-07 4.57E-08 2.06E-08 1,620 2,380 2 48DF100 0.002% 0.02% 11.4 8 0 1.90E-06 1.90E-07 8.69E-08 5,480 8,260 3 10DF100 0.01% 0.08% 16.3 6 0 6.25E-06 7.16E-07 3.21E-07 16,500 27,300 4 7DF1000 0.02% 0.26% 14.9 20 0 1.72E-05 2.35E-06 1.01E-06 48,400 77,600 5 11DF10 0.06% 0.78% 14.4 4 0 2.03E-05 3.36E-06 1.62E-06 71,200 127,000 6 48 0.23% 1.69% 11.4 8 0 7.95E-05 1.61E-05 7.79E-06 253,000 450,000 7 15 0.60% 5.85% 14.9 7 0 1.21E-04 3.28E-05 1.64E-05 524,000 932,000 8 46 0.98% 11.01% 14.8 17 0 1.53E-04 4.59E-05 2.34E-05 790,000 1,410,000 9 5DF10 1.05% 2.89% 24.2 34 0 3.55E-04 7.50E-05 3.35E-05 1,040,000 1,720,000 10 5 1.39% 6.46% 24.2 41 0 4.06E-04 9.78E-05 4.51E-05 1,360,000 2,290,000 11 8 1.49% 19.25% 14.9 5 0 1.35E-04 6.41E-05 3.43E-05 1,110,000 2,030,000 12 1 1.93% 22.68% 14.9 22 0 2.91E-04 1.01E-04 5.23E-05 1,720,000 3,090,000 13 41DF1000 3.40% 7.65% 9.8 17 0 5.22E-04 1.49E-04 7.89E-05 1,900,000 3,610,000 14 22dw 2.82% 18.64% 14.9 27 0 4.27E-04 1.28E-04 6.57E-05 1,830,000 3,320,000 15 53 2.79% 29.05% 17.4 13 0 2.59E-04 1.19E-04 6.96E-05 1,740,000 3,520,000 16 41 4.54% 14.10% 9.8 16 0 5.57E-04 1.75E-04 9.82E-05 2,300,000 4,520,000 17 3DF10 8.85% 24.65% 9.8 63 0 7.10E-04 2.95E-04 1.68E-04 3,830,000 7,720,000 18 52 15.90% 34.32% 17.4 11 0 5.39E-04 2.23E-04 1.50E-04 3,080,000 6,870,000

Hrs with Land (sq mi) Exceeding Population Subject to Start Offsite Cost ($ 2013) Long-Term Habitability Long-Term Protective Rep Case Rep Case Significant Bin Rep Case Time Criterion Actions Cs (%) I (%) Cs (hrs)

Release*

0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 1 28DF1000 0.0006% 0.006% 14.9 7 78,900,000 78,900,000 0 0 - -

2 48DF100 0.002% 0.02% 11.4 8 79,700,000 79,700,000 1 1 0 0 3 10DF100 0.01% 0.08% 16.3 6 98,100,000 98,700,000 10 11 1 1 4 7DF1000 0.02% 0.26% 14.9 20 141,000,000 141,000,000 23 23 7 7 5 11DF10 0.06% 0.78% 14.4 4 220,000,000 240,000,000 41 65 118 118 6 48 0.23% 1.69% 11.4 8 1,150,000,000 1,390,000,000 116 175 3,440 3,440 7 15 0.60% 5.85% 14.9 7 2,740,000,000 3,690,000,000 190 361 15,000 16,600 8 46 0.98% 11.01% 14.8 17 3,760,000,000 5,220,000,000 242 506 20,700 27,400 9 5DF10 1.05% 2.89% 24.2 34 7,290,000,000 8,600,000,000 351 429 35,200 35,200 10 5 1.39% 6.46% 24.2 41 9,900,000,000 12,000,000,000 479 715 51,400 51,500 11 8 1.49% 19.25% 14.9 5 5,960,000,000 9,720,000,000 286 673 40,500 55,800 12 1 1.93% 22.68% 14.9 22 13,000,000,000 17,400,000,000 549 1,040 64,500 79,700 13 41DF1000 3.40% 7.65% 9.8 17 19,400,000,000 24,700,000,000 783 1,170 168,000 190,000 14 22dw 2.82% 18.64% 14.9 27 12,900,000,000 18,300,000,000 544 1,010 93,700 114,000 15 53 2.79% 29.05% 17.4 13 15,700,000,000 26,500,000,000 573 1,290 111,000 142,000 16 41 4.54% 14.10% 9.8 16 25,500,000,000 35,400,000,000 904 1,500 235,000 281,000 17 3DF10 8.85% 24.65% 9.8 63 47,000,000,000 68,100,000,000 1,360 2,470 417,000 504,000 2 18 52 15.90% 34.32% 17.4 11 46,500,000,000 87,700,000,000 987 2,170 467,000 873,000 3 Note: To quantify the time signature of a source term release, an hourly plume segment is 4 considered significant if it contributes at least 0.5 percent of that source terms total cumulative 5 cesium release to the environment. Cesium, rather than iodine, was selected here because all of the 6 resulting offsite consequences are driven by long-term phase exposures.

7 (Source: NUREG-2206, Table 4-22) 8 H-103 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Table H-21 MACCS Results for 9 Mark II Source Term Bins

Hrs with Individual Early Population Dose Start Individual Latent Cancer Fatality Risk Rep Case Rep Case Significant Fatality Risk (person-rem)

Bin Rep Case Time Cs (%) I (%) Cs (hrs) 0-1.3 mi and beyond 0-10 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi Release*

1 11DF1000 0.00004% 0.0005% 20.3 20 0 9.72E-08 1.03E-08 3.45E-09 282 345 2 5DF1000 0.0006% 0.005% 32.2 20 0 1.15E-06 1.81E-07 6.35E-08 4,340 5,440 3 42DF100 0.0043% 0.037% 14.3 13 0 6.58E-06 8.67E-07 3.02E-07 20,700 26,700 4 11 0.042% 0.45% 20.3 20 0 7.90E-05 9.68E-06 3.27E-06 202,000 261,000 5 51DF10 0.23% 2.01% 16.6 9 0 1.35E-04 3.39E-05 1.21E-05 689,000 888,000 6 5 0.55% 4.94% 32.2 20 0 2.29E-04 1.05E-04 4.01E-05 2,160,000 2,900,000 7 3 1.09% 10.26% 14.3 20 0 3.08E-04 1.88E-04 7.43E-05 4,140,000 5,580,000 8 1 2.46% 19.81% 22.8 25 0 4.70E-04 3.17E-04 1.25E-04 6,110,000 8,260,000 9 52 3.57% 28.67% 16.6 10 0 4.03E-04 2.46E-04 1.01E-04 5,430,000 7,440,000

Hrs with Land (sq mi) Exceeding Start Population Subject to Long-Rep Case Rep Case Significant Offsite Cost ($ 2013) Long-Term Habitability Bin Rep Case Time Term Protective Actions Cs (%) I (%) Cs Criterion (hrs)

Release* 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 1 11DF1000 0.00004% 0.0005% 20.3 20 381,000,000 381,000,000 - - - -

2 5DF1000 0.0006% 0.005% 32.2 20 381,000,000 381,000,000 0 0 - -

3 42DF100 0.0043% 0.037% 14.3 13 393,000,000 393,000,000 2 2 0 0 4 11 0.042% 0.45% 20.3 20 844,000,000 846,000,000 44 47 1,030 1,030 5 51DF10 0.23% 2.01% 16.6 9 4,250,000,000 4,380,000,000 130 221 15,400 15,400 6 5 0.55% 4.94% 32.2 20 24,000,000,000 28,000,000,000 303 551 62,400 62,400 7 3 1.09% 10.26% 14.3 20 80,800,000,000 105,400,000,000 698 1,200 619,000 649,000 8 1 2.46% 19.81% 22.8 25 85,500,000,000 109,300,000,000 854 1,680 721,000 741,000 2 9 52 3.57% 28.67% 16.6 10 53,600,000,000 63,800,000,000 618 1,400 414,000 449,000 3

Note: To quantify the time signature of a source term release, an hourly plume segment is 4 considered significant if it contributes at least 0.5 percent of that source terms total cumulative 5 cesium release to the environment. Cesium, rather than iodine, was selected here because all the 6 resulting offsite consequences are driven by long-term phase exposures.

7 (Source: NUREG-2206, Table 4-23) 8 9 The offsite radiological consequence estimates for SECY-15-0085 were like those of 10 SECY-12-0157. However, an important distinction between the detailed analyses for 11 SECY-15-0085 and SECY-12-0157 is the use of different performance criteria to evaluate the 12 offsite radiological consequence results. Although not explicitly stated, the detailed analyses for 13 SECY-12-0157 implicitly assumed decontamination factor (DF) as a performance criterion.

14 Specifically, consistent with international nuclear safety practices and guidelines, a DF value of 15 1,000 was established as a performance target. This is equivalent to one-tenth of one percent 16 of cesium release to the environment and serves as an indirect measure of latent cancer fatality 17 risk and land contamination risk. By contrast, SECY-15-0085 defined six performance criteria 18 related to the attributes of (1) conditional containment failure probability, (2) DF, (3) equipment 19 and procedure availability, (4) total population dose, (5) margin to the QHOs, and (6) long-term 20 relocation. Ultimately, the detailed analyses for SECY-15-0085 used the margin to the safety 21 goal QHOs for average individual early fatality risk within 1 mile and average individual latent 22 cancer fatality risk within 10 miles as the performance criteria to determine whether each 23 alternative could result in a substantial increase in the overall protection of public health and 24 safety.

25 26 Risk Evaluation 27 28 The staff expanded the scope and level of detail of the PRA model developed for 29 SECY-12-0157 for the detailed analyses for SECY-15-0085. The PRA model used in 30 SECY-12-0157 did not delineate core damage accident sequences. Instead, it relied on a 31 generic estimate of CDF developed from previous NRC staff and licensee PRAs. To provide a NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-104

1 quantitative basis for regulatory decisionmaking, the PRA performed in support of 2 SECY-15-0085 included the following features:

3 4

Models to estimate the frequency of ELAP events resulting from internal events and 5 earthquakes, based on industry-developed re-evaluations of seismic hazard estimates.

6 7

CDETs that delineate accident sequences from the occurrence of an ELAP event to the 8 onset of core damage. The CDETs reflect SBO mitigation strategies using installed 9 plant and portable equipment.

10 11

APETs that delineate accident sequences from the onset of core damage to the release 12 of radioactive materials to the environment. The APETs reflect CPRR strategies such as 13 post-core-damage containment venting and water addition.

14 15

Models that include random and seismically-induced equipment failures.

16 17

In-control room and local manual operator actions consistent with emergency operating 18 procedures and severe accident management guidelines.

19 20

Models that identify important contributors to CDF.

21 22

Sensitivity analyses to gain insight into how plausible alternative assumptions about 23 human error probability estimates impact the quantitative results.

24 25 These revisions to the PRA model resulted in a lower value for conditional CDF, conditioned on 26 the assumed occurrence of an ELAP, than was reported in SECY-12-0157. The model 27 calculated the CDF caused by ELAPs to be 8.9x10-6 per reactor-year, which was about two 28 times lower than the value of 1.6x10-5 that SECY-12-0157 estimated. The CDF calculation 29 averaged together the CDF for each BWR plant that was included in the scope of the accident 30 sequence analysis.

31 32 Table H-22 summarizes the risk estimates of each regulatory analysis subalternative. These 33 risk estimates represent the point estimate, baseline-case results.

34 H-105 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

2 3 1 Individual Land Exceeding Population Subject to Fraction of Early Individual Latent Cancer Population Dose Offsite Cost Long-Term Long-Term Core-Damage Fatality Fatality Risk (/y) (person-rem/y) ($ 2013/y) Habitability Criterion Protective Actions Frequency Risk (/y) (square miles/y) (persons/y)

Regulatory Analysis Index Uncontrolled 0-1.3 mi Sub-Alternative Vented 0-10 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi Release and beyond 1 1 0% 100% 0.0E+00 3.0E-09 8.6E-10 4.2E-10 1.3E+01 2.3E+01 9.9E+04 1.3E+05 4.4E-03 7.6E-03 5.1E-01 5.8E-01 2 2A 0% 100% 0.0E+00 3.0E-09 8.6E-10 4.2E-10 1.3E+01 2.3E+01 9.9E+04 1.3E+05 4.4E-03 7.6E-03 5.1E-01 5.8E-01 (Source: NUREG-2206, Table 5-1) 3 3A 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 4 3B 42% 58% 0.0E+00 2.1E-09 6.7E-10 3.4E-10 1.1E+01 1.9E+01 7.4E+04 1.0E+05 3.4E-03 6.4E-03 4.1E-01 4.9E-01 5 4Ai(1) 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 6 4Ai(2) 42% 58% 0.0E+00 2.1E-09 6.1E-10 3.1E-10 9.5E+00 1.7E+01 6.8E+04 9.0E+04 3.2E-03 5.8E-03 3.6E-01 4.1E-01 7 4Aii(1) 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 8 4Aii(2) 42% 58% 0.0E+00 2.4E-09 7.7E-10 3.9E-10 1.2E+01 2.2E+01 8.9E+04 1.2E+05 3.9E-03 7.3E-03 4.8E-01 5.8E-01 9 4Aiii(1) 58% 42% 0.0E+00 1.8E-09 5.5E-10 2.7E-10 8.6E+00 1.5E+01 6.5E+04 8.5E+04 2.9E-03 5.0E-03 3.3E-01 3.9E-01 H-106 10 4Aiii(2) 42% 58% 0.0E+00 2.0E-09 5.6E-10 2.7E-10 8.7E+00 1.5E+01 6.2E+04 7.9E+04 3.0E-03 5.1E-03 3.1E-01 3.4E-01 11 4Bi(1) 58% 42% 0.0E+00 1.3E-09 3.1E-10 1.5E-10 4.5E+00 7.8E+00 2.9E+04 3.7E+04 1.6E-03 2.5E-03 1.5E-01 1.6E-01 12 4Bi(2) 42% 58% 0.0E+00 1.4E-09 3.3E-10 1.5E-10 4.8E+00 8.2E+00 3.1E+04 3.8E+04 1.8E-03 2.7E-03 1.6E-01 1.6E-01 13 4Bii 42% 58% 0.0E+00 1.4E-09 3.2E-10 1.5E-10 4.6E+00 7.9E+00 3.0E+04 3.7E+04 1.7E-03 2.6E-03 1.5E-01 1.5E-01 14 4Biii 42% 58% 0.0E+00 1.4E-09 3.2E-10 1.5E-10 4.7E+00 8.1E+00 3.1E+04 3.7E+04 1.7E-03 2.6E-03 1.5E-01 1.6E-01 15 4Biv 40% 60% 0.0E+00 1.3E-09 3.1E-10 1.5E-10 4.6E+00 7.8E+00 3.0E+04 3.6E+04 1.7E-03 2.6E-03 1.5E-01 1.5E-01 16 4Ci(1) 58% 42% 0.0E+00 1.3E-09 3.1E-10 1.5E-10 4.5E+00 7.8E+00 2.9E+04 3.7E+04 1.6E-03 2.5E-03 1.5E-01 1.6E-01 17 4Ci(2) 42% 58% 0.0E+00 1.3E-09 3.1E-10 1.4E-10 4.5E+00 7.6E+00 3.0E+04 3.7E+04 1.6E-03 2.4E-03 1.5E-01 1.6E-01 Table H-22 Risk Estimates by Regulatory Analysis Subalternative 18 4Cii 42% 58% 0.0E+00 1.3E-09 3.0E-10 1.4E-10 4.4E+00 7.4E+00 2.9E+04 3.6E+04 1.5E-03 2.3E-03 1.5E-01 1.5E-01 19 4Ciii 42% 58% 0.0E+00 1.3E-09 3.1E-10 1.4E-10 4.4E+00 7.6E+00 3.0E+04 3.7E+04 1.6E-03 2.4E-03 1.5E-01 1.6E-01 20 4Civ 40% 60% 0.0E+00 1.3E-09 3.0E-10 1.4E-10 4.3E+00 7.4E+00 2.9E+04 3.6E+04 1.5E-03 2.3E-03 1.5E-01 1.5E-01

1 2 In addition to these point estimate baseline-case results, the staff conducted uncertainty and 3 sensitivity analyses. The staff performed a parametric Monte Carlo uncertainty analysis to gain 4 additional perspective into the uncertainty of the point estimate risk evaluation results. The 5 uncertainty analysis considered seismic hazard curves, seismic fragility curves, random 6 equipment failures, operator actions, and consequences. Table H-23 summarizes information 7 used to perform the parametric uncertainty analysis. Figure H-19 shows the results of the 8 uncertainty analysis.

9 10 Table H-23 Uncertainty Analysis Inputs Events Distribution Remarks An error factor of 15 maximizes the ratio of the 95th percentile to the mean value. This approach does not Frequency of Lognormal explicitly consider the uncertainty in the offsite power ELAPs due to Mean = point estimate recovery curves or the uncertainty in the EPS reliability internal events Error factor =15 parameters (failure rate and failure-on-demand probability).

Normal parameters were developed for each point on the seismic hazard curve using the fractile information Seismic hazard Lognormal provided by licensees in their responses to the 10 CFR curves 50.54(f) information request concerning NTTF Recommendation 2.1.

Double lognormal, using Traditional approach to modeling uncertainty in seismic Seismic fragilities the developed values of fragility.

C50, R, and U Lognormal Hardware-related An error factor of 15 maximizes the ratio of the 95th Mean = point estimate failures percentile to the mean value.

Error factor = 15 Human failure Constrained A constrained non-informative prior distribution is a beta events non-informative prior distribution with mean = point estimate and = 0.5.

Lognormal Conditional Informed by preliminary results of the SOARCA Mean = point estimate consequences uncertainty analysis project.a Error factor = 10 11 a NUREG/CR-7155 (draft), State-of-the-Art Reactor Consequence Analyses Project, Uncertainty Analysis of the 12 Unmitigated Long-Term Station Blackout of the Peach Bottom Atomic Power Station.

13 (Source: NUREG-2206, Table 5-2) 14 15 Staff also performed MACCS sensitivity calculations to analyze the influence of site to site 16 variation. The following sensitivities were conducted:

17 18

Population (low, medium, high) 19 20

Evacuation delay (1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months ) 21 22

Nonevacuating cohort size (5 percent of emergency planning zone population) 23 24

Intermediate phase duration (0, 3 months, and 1 year) 25 26

Long-term habitability criterion (500 mrem per year and 2 rem per year), which can vary 27 among states in the U.S.

28 H-107 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 A final sensitivity calculation examined evacuation delays on the risk to determine the influence 2 of the plume arrival time on the evacuating population (base case, 3 hour3.472222e-5 days 8.333333e-4 hours 4.960317e-6 weeks 1.1415e-6 months delay, 6 hour6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months delay, 3 no evacuation).

4 5 The results of these sensitivity analyses appear in a series of tables in Chapter 4 of 6 NUREG-2206, which report the ratio of the consequences for the sensitivity cases compared to 7 the baseline cases. Table H-24 below shows an example of these sensitivity results tables, 8 analyzing the effect of different site files (different populations) on the baseline-case results.

9 The results show that individual latent cancer fatality risk is relatively insensitive to site file data 10 (variations are within 60 percent). Population dose is directly related to population size, so the 11 sensitivity cases show a strong increase in population dose for larger population site files. For 12 example, for the Mark II high source term, the high site file case has a population dose about 13 11 times higher than the low site file case. For a given source term, the total offsite cost also 14 increases with higher population site files.

15 16 Table H-24 Results for Baseline Cases with Different Site Files Land (sq mi)

Individual Population Subject Base Model Individual Latent Cancer Population Dose Offsite Cost Exceeding Long-Early Fatality to Long-Term Fatality Risk (person-rem) ($ 2013) Term Habitability Source Term Site File Risk Protective Actions Criterion 0-1.3 mi and 0-10 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi 0-50 mi 0-100 mi beyond Med (VT Yankee) / Low (Hatch) 1.52 0.98 0.90 0.92 1.19 2.79 2.75 0.39 0.43 6.20 6.20 Mark I - Low (Bin 3)

Mark I - Peach High (Peach Bottom) / Low (Hatch) 0.94 0.74 0.96 2.82 2.07 4.65 4.57 1.53 1.45 2.07 2.07 Med (VT Yankee) / Low (Hatch) 1.25 0.98 0.97 1.88 2.37 3.08 3.60 0.67 0.72 2.91 2.92 Mark I - Med (Bin 10) Individual Bottom High (Peach Bottom) / Low (Hatch) 1.02 0.83 1.02 5.83 4.00 8.84 8.22 1.28 1.08 7.15 7.15 early fatality Med (VT Yankee) / Low (Hatch) 1.23 1.05 1.08 2.26 3.33 3.58 4.95 0.82 0.82 3.11 4.16 Mark I - High (Bin 17) risk is zero High (Peach Bottom) / Low (Hatch) 1.00 0.89 1.00 6.78 5.04 11.11 9.33 1.11 0.98 9.96 9.59 for all Med (Susquehanna) / Low (Columbia) *

Mark II - Limerick 1.20 0.93 0.49 0.70 1.00 4.90 4.90 3.93 3.93 Mark II - Low (Bin 2) baseline and High (Limerick) / Low (Columbia) 1.63 1.10 0.69 2.33 2.25 20.48 20.48 12.79 12.79 *

sensitivity Med (Susquehanna) / Low (Columbia) cases. 0.94 0.86 0.49 1.38 1.96 2.32 2.33 0.40 0.56 6.35 6.35 Mark II - Med (Bin 5)

High (Limerick) / Low (Columbia) 1.17 1.03 0.65 6.53 4.82 11.71 10.63 0.52 0.61 28.96 28.96 Med (Susquehanna) / Low (Columbia) 0.89 0.85 0.59 2.06 3.71 3.07 6.60 0.61 0.76 3.00 3.42 Mark II - High (Bin 8) 17 High (Limerick) / Low (Columbia) 1.07 1.04 0.68 10.82 9.32 18.49 17.97 0.69 0.75 17.87 17.09 18

Indicates that both the numerator and denominator in the ratio are zero 19 (Source: NUREG-2206, Table 4-36) 20 21 Cost-Benefit Analysis Results 22 23 Although the potential benefits from possible measures to limit releases through the 24 containment venting systems during severe accidents were well below the NRCs threshold for 25 developing regulatory requirements, the staff reported updated industry cost estimates for 26 implementing the CPRR alternatives in SECY-15-0085. However, these updated cost estimates 27 did not change the staffs conclusion from SECY-12-0157 that none of the proposed regulatory 28 alternatives would satisfy the substantial additional protection standard at 10 CFR 50.109 (a)(3).

29 30 Summary and Conclusion 31 32 The staff developed a risk evaluation and evaluated alternative courses of action related to 33 filtering strategies and severe accident management of BWRs with Mark I and Mark II 34 containments relative to the safety goal QHOs. The staff determined that the possible plant 35 modifications (e.g., engineered filters) to enhance containment protection and release reduction 36 capability beyond those imposed by Order EA-13-109 could result in reductions in offsite 37 consequences. However, these reductions would not meet the quantitative threshold for a 38 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-108

1 substantial safety enhancement because the average individual early fatality risk and average 2 individual latent cancer fatality risk are well below the QHOs without additional plant 3 modifications.

4 5 Based on the results of the detailed analyses for SECY-15-0085, the staff planned to proceed 6 with Alternative 3: Rulemaking to Make Order EA-13-109 Generically Applicable and Additional 7 Requirements for SAWA to Address Uncontrolled Releases from Major Containment Failure 8 Modes. The rulemaking would include the planned implementation of Phase 2 of the order to 9 require licensees of BWRs with Mark I and Mark II containments to have the capability to add 10 water from external sources and control the flow to cool core debris during severe accident 11 conditions. The staff concluded that the ability to provide post-core-damage water addition 12 results in worthwhile additional protection for public health and safety by: (1) protecting the 13 integrity of the containment; (2) reducing the release of radioactive materials in some severe 14 accident scenarios; and (3) contributing to the balance between accident prevention and 15 mitigation.

16 17 The staffs plan to proceed with Alternative 3 for the CPRR rulemaking differed from the staffs 18 recommendation in SECY-12-0157 to require the installation of an engineered filtering system.

19 More detailed analyses resulted in the following findings:

20 21

The CDF from an ELAP event was lower than estimated in SECY-12-0157.

22 23

The identification of important contributors to CDF and sensitivity analyses enhanced the 24 staffs confidence in its quantitative analyses and therefore reduced the importance of 25 remaining uncertainties.

26 27

External water addition was shown to avert containment failure and achieve benefits in 28 terms of averted health risks in a wider range of scenarios than an engineering filtering 29 system (e.g., in scenarios where the release pathway bypasses the filtering system).

30 31 Therefore, the staff recommended proceeding with a proposed rulemaking to address the 32 containment protection improvements related to venting and water addition without including 33 requirements for installing engineered filtering systems.

34 35 Commissions Response to the Staffs Analysis and Recommendations 36 37 The Commission disapproved the staff's plan to proceed with Alternative 3. Instead, the 38 Commission approved Alternative 1, which was to continue with the implementation of Order 39 EA-13-109 and installation of severe-accident-capable vents (including SAWA/SAWM as part of 40 Phase 2 compliance with the Order), without taking additional regulatory actions related to BWR 41 Mark I and Mark II containments. The reasoning for this decision was articulated in the 42 Chairmans comments in the Commission Voting Record. The Chairman noted that there is no 43 practical difference in safety outcomes between Alternatives 1 and 3Order EA-13-109, which 44 was imposed on all BWRs with Mark I and II containments in 2013, already serves as a legally 45 binding mechanism that effectively achieves the results the staff is seeking[Furthermore]

46 there are no expectations that a BWR with a Mark I or II containment will ever be licensed to 47 operate in the United States again, which obviated the need to expend agency resources to 48 make Order EA-13-109 generically applicable through rulemaking (NRC, 2015b).

49 H-109 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 The Commission further directed the staff to leverage the draft regulatory basis to the extent 2 applicable to support resolution of the post-Fukushima Dai-ichi Tier 3 item related to 3 containments of other designs (NTTF Recommendation 5.2). The NTTF Recommendation 5.2 4 was subsequently closed by SECY-16-0041, Closure of Fukushima Tier 3 Recommendations 5 Related to Containment Vents, Hydrogen Control, and Enhanced Instrumentation, dated 6 March 31, 2016, with no further regulatory action.

7 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-110

1 ENCLOSURE H-5:

SUMMARY

OF DETAILED ANALYSES FOR 2 SECY-13-0112 AND NUREG-2161, CONSEQUENCE STUDY OF A 3 BEYOND-DESIGN-BASIS EARTHQUAKE AFFECTING THE SPENT 4 FUEL POOL FOR A U.S. MARK I BOILING-WATER REACTOR 5

6 This enclosure summarizes the detailed analyses supporting the evaluation of expedited spent 7 fuel transfer from the spent fuel pool (SFP) to dry cask storage for a reference plant, as 8 documented in SECY-13-0112, Consequence Study of a Beyond-Design-Basis Earthquake 9 Affecting the Spent Fuel Pool for a U.S. Mark I Boiling-Water Reactor, dated October 9, 2013, 10 and in NUREG-2161, Consequence Study of a Beyond-Design-Basis Earthquake Affecting the 11 Spent Fuel Pool for a U.S. Mark I Boiling-Water Reactor. The contents of this enclosure should 12 be considered with the subsequent detailed analyses supporting COMSECY-13-0030, Staff 13 Evaluation and Recommendation for Japan Lessons-Learned Tier 3 Issue on Expedited 14 Transfer of Spent Fuel. Enclosure H-6, Summary of Detailed Analyses in 15 COMESECY-13-0030, Enclosure H-1, Regulatory Analysis for Japan Lessons-Learned Tier 3 16 Issue on Expedited Transfer of Spent Fuel, to this appendix summarizes the detailed analyses 17 for COMSECY-13-0030.

18 19 Problem Statement and Regulatory Objectives 20 21 Previous risk studies have shown that storage of spent fuel in a high-density configuration in 22 SFPs is safe and that the risk is appropriately low (see for example, NUREG-1738, Technical 23 Study of Spent Fuel Pool Accident Risk at Decommissioning Nuclear Power Plants). These 24 studies used simplified and sometimes bounding assumptions and models to characterize the 25 likelihood and consequences of beyond-design-basis accidents involving SFPs. As part of the 26 Nuclear Regulatory Commissions (NRCs) post-9/11 security assessments, detailed 27 thermal-hydraulic and severe accident progression models for SFPs were developed and 28 applied to assess the realistic heatup of spent fuel under various pool draining conditions. In 29 2009, together with these post-9/11 security assessments, the NRC issued additional regulatory 30 requirements codified in Title 10 of the Code of Federal Regulations (10 CFR) Part 50, 31 Section 54, Conditions of licenses. In particular, 10 CFR 50.54(hh)(2) requires that each 32 reactor licensee develop and implement guidance and strategies intended to maintain or restore 33 core cooling, containment, and SFP cooling capabilities under conditions associated with certain 34 beyond-design-basis events.

35 36 Following the 2011 accident at the Fukushima Dai-ichi nuclear power plant in Japan that 37 resulted from the Tohoku earthquake and tsunami, several stakeholders submitted comments to 38 the NRC Commission and staff requesting that regulatory action be taken to require the 39 expedited transfer of spent fuel stored in SFPs to dry casks. The basis for these requests was 40 that expediting the transfer of spent fuel in SFPs to dry casks would reduce the potential 41 consequences associated with a loss of SFP coolant inventory by decreasing the amount of 42 spent fuel stored in affected SFPs, thereby decreasing the heat generation rate and 43 radionuclide source term associated with affected spent fuel. In response to Commission 44 direction in staff requirements memorandum (SRM)-SECY-12-0025, Staff Requirements 45 SECY-12-0025Proposed Orders and Requests for Information in Response to Lessons 46 Learned from Japans March 11, 2011, Great Tohoku Earthquake and Tsunami, dated 47 March 9, 2012, the staff implemented regulatory actions that originated from the Near-Term 48 H-111 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Task Force (NTTF) recommendations to enhance reactor and SFP safety. The staff issued two 2 orders requiring enhancements to SFP safety:

3 4 1. Order EA-12-049, Issuance of Order to Modify Licenses with Regard to Requirements 5 for Mitigation Strategies for Beyond-Design-Basis External Events, dated 6 March 12, 2012, which requires that licensees develop, implement, and maintain 7 guidance and strategies to maintain or restore core cooling, containment, and SFP 8 cooling capabilities following a beyond-design-basis external event.

9 10 2. Order EA-12-051, Order Modifying Licenses with Regard to Reliable Spent Fuel Pool 11 Instrumentation, dated March 12, 2012, which requires that licensees install reliable 12 means of remotely monitoring wide-range SFP levels to support effective prioritization of 13 event mitigation and recovery actions in the event of a beyond-design-basis external 14 event.

15 16 The results are based on previous risk studies without these enhancements, in which the staff 17 had concluded that existing requirements for both SFPs and dry casks provide adequate 18 protection of public health and safety. However, in response to events following the accident at 19 Fukushima, the staff determined that it should (1) confirm that high-density SFP configurations 20 continue to provide adequate protection of public health and safety; and (2) assess potential 21 safety benefits (or detriments) and costs associated with expediting the transfer of spent fuel 22 from the SFP to dry casks at a reference plant with a boiling-water reactor (BWR) and Mark I 23 containment design (the same type of reactor involved in the Fukushima Dai-ichi nuclear power 24 plant accident).

25 26 Regulatory Alternatives 27 28 The regulatory analyses performed in support of SECY-13-0112 and NUREG-2161 considered 29 the following two regulatory alternatives that address spent fuel storage requirements:

30 31 1. Option 1: Maintain Existing Spent Fuel Storage Requirements (Status Quo). This 32 alternative reflected the Commission decision not to expedite the storage of spent fuel 33 from SFPs to dry casks but to continue with the NRCs existing regulatory requirements 34 for spent fuel storage. Under this alternative, spent fuel is moved into dry storage only 35 as necessary to accommodate fuel assemblies being removed from the core during 36 refueling operations. It also assumed that all applicable requirements and guidance to 37 date had been implemented, but no implementation was assumed for related generic 38 issues or other staff requirements or guidance that were unresolved or still under review 39 at the time of the analysis. This alternative assumed (1) continued storage of spent fuel 40 in high-density racks within a relatively full SFP, and (2) compliance with all current 41 regulatory requirements, including those described above for 10 CFR 50.54(hh)(2),

42 Order EA-12-049, and Order EA-12-051. 30 Furthermore, because SFPs have a limited 43 amount of available storageeven after licensees expanded their storage capacity 44 using high-density storage racksthe alternative assumed that the existing practice of 45 transferring spent fuel from SFPs to casks in accordance with 10 CFR Part 72, 46 Licensing Requirements for the Independent Storage of Spent Nuclear Fuel and 30 Although Option 1 assumed compliance with the post-Fukushima mitigation strategies required under Order EA-12-049 and the reliable SFP instrumentation required under Order EA-12-051, this was not explicitly modeled as part of the study. Instead, compliance with these requirements was treated as a qualitative factor that would significantly enhance the likelihood of successful mitigation, and thereby reduce the conditional probability of radiological release under Option 1.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-112

1 High-Level Radioactive Waste, and Reactor-Related Greater than Class C Waste, 2 would continue. This alternative represented the status quo and served as the 3 regulatory baseline against which the costs and benefits of Option 2 were measured.

4 5 2. Option 2: Expedited Spent Fuel Transfer to Achieve Low-density SFP Storage. This 6 alternative assumed that older spent fuel assemblies would be expeditiously moved from 7 SFP storage to dry cask storage beginning in 2014 to achieve and maintain a 8 low-density loading of spent fuel in existing high-density racks within 5 years. It did not 9 evaluate re-racking of the SFP to a low-density rack configuration because such a 10 situation was judged to be inefficient in terms of regulatory benefit, given that much of 11 the benefit could be achieved by storing less fuel in the existing high-density racks.

12 Because of the low-density SFP loading, this alternative had a smaller long-lived 13 radionuclide inventory in the SFP, a lower overall heat load in the SFP, and a slight 14 increase in the initial water inventory that displaced the removed spent fuel assemblies.

15 16 The staff recognized potential cost and risk impacts associated with the transfer of spent fuel 17 from SFPs to dry casks after 5 years of cooling and during long-term dry cask storage. If 18 included, these cost and risk impacts would have reduced the overall net benefit of Option 2 19 relative to Option 1. However, these effects were conservatively ignored to calculate the 20 potential benefit per reactor-year by comparing only the safety of high-density SFP storage to 21 low-density SFP storage and its implementation costs.

22 23 Safety Goal Evaluation 24 25 To perform the safety goal evaluation, the staff analyzed the regulatory alternatives to directly 26 compare their potential safety benefits to the quantitative health objectives (QHOs) for average 27 individual early fatality risk and average individual latent cancer fatality risk described in the 28 Commissions Safety Goal Policy Statement (NRC, 1986).

29 30 Since the reactor building that houses the SFP does not provide a containment barrier like the 31 containment structure surrounding the reactor coreespecially under conditions postulated to 32 dominate the release of radioactive materials from spent fuelthe staff assumed the frequency 33 of a release of radioactive material to the environment would be the same as the frequency of 34 spent fuel damage. Under this assumption, the radiological release frequency was estimated to 35 range from 7x10-7 to 5x10-6 per reactor-year, when considering all initiators that could challenge 36 SFP cooling or integrity.

37 38 Despite the large releases for certain predicted accident progressions, the staff determined 39 there was zero average individual early fatality risk, conditioned on the assumed occurrence of 40 the modeled severe accident scenarios. In part, this was because the modeled accident 41 progressions resulted in releases that begin late relative to the time needed to evacuate 42 members of the public living near the modeled nuclear power plant site.

43 44 Using the upper limit of the spent fuel damage and radiological release frequency of 5x10-6 per 45 reactor-year combined with a conditional average individual latent cancer fatality risk within 46 10 miles of 4x10-4 resulted in a bounding average individual latent cancer fatality risk of 47 2x10-9 per reactor-year. This calculated value was about 3 orders of magnitude below the QHO 48 of 2x10-6 per reactor-year for an average individual latent cancer fatality risk within 10 miles.

49 The staff therefore concluded that Option 2 could not result in a substantial increase in overall 50 protection of public health and safety.

51 H-113 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Technical Evaluation 2

3 The staff performed detailed analyses using state-of-the-art, validated, deterministic methods 4 and assumptions, supplemented with probabilistic insights where practical.

5 6 The study considered two SFP configurations:

7 8 1. High-density Loading Configuration: A relatively full SFP in which the hottest spent fuel 9 assemblies are surrounded by four cooler fuel assemblies in a 1x4 loading pattern 10 throughout the pool 31 11 12 2. Low-density Loading Configuration: A minimally loaded pool in which all spent fuel with 13 at least 5 years of pool cooling has been removed to ensure the hottest fuel assemblies 14 are surrounded by additional water 15 16 To evaluate the potential benefits of mitigation strategies required in 10 CFR 50.54 (hh)(2), the 17 study analyzed each loading configuration for two different cases(1) the mitigated case, in 18 which 10 CFR 50.54 (hh)(2) mitigation strategies were assumed to be successful and (2) the 19 unmitigated case, in which these mitigation strategies were assumed to be unsuccessful.

20 Following the evaluation of these cases, the staff performed a limited scope human reliability 21 analysis to estimate the likelihood of successful operator actions implementing 22 10 CFR 50.54(hh)(2) mitigation measures to prevent fuel damage. Key assumptions made in 23 this limited scope human reliability analysis are that (1) post-earthquake onsite portable 24 mitigation equipment required by 10 CFR 50.54(hh)(2) was available, (2) minimum plant staffing 25 was available for implementing SFP mitigation, and (3) operators had access to areas needed 26 to implement mitigation measures. The study considered scenarios in which some preplanned 27 and improvised mitigating actions were either unsuccessful or not implemented before the 28 analysis was terminated at 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months . For example, in addition to the 10 CFR 50.54(hh)(2) 29 mitigation strategies, the site emergency response organization would request support from 30 offsite response organizations to implement additional mitigating actions that are improvised, 31 such as pumping water into the SFP using a fire truck. However, these additional mitigating 32 actions were determined to be beyond the scope of the study.

33 34 Accident Scenario Selection 35 36 Previous risk studies had shown that earthquakes represent the dominant risk contributor for 37 SFPs. Therefore, to deliberately challenge the integrity of the SFP, the accident initiator for this 38 study was a beyond-design-basis earthquake with ground motion (0.7g peak ground 39 acceleration) stronger than the maximum earthquake reasonably expected to occur for the 40 reference plant. An earthquake of this severity was estimated to occur about once every 41 60,000 years.

42 43 The SFP accident scenarios evaluated in this study were developed for a single operating cycle.

44 However, the conditions of the SFP change throughout an operating cycle. For example, the 45 SFP can change from being an isolated pool to being hydraulically connected to the reactor 46 vessel (e.g., during refueling operations), or spent fuel can be moved around within the SFP 47 during a cycle to satisfy regulatory requirements with respect to criticality or heat distribution.

48 Such changes affect the consequences of a postulated accident. Therefore, for this study, the 31 A limited sensitivity analysis of a 1x8 spent fuel configuration and a uniform configuration was also performed to better understand the potential effects of plausible alternative SFP configurations on results and insights.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-114

1 continual changes that occur during a single operating cycle were discretized into discrete 2 quasi-steady snapshots referred to as operating cycle phases (OCPs). Since the number of 3 OCPs has a roughly linear scaling effect on the number of MELCOR analyses required, the 4 study defined in terms of the minimum number that most accurately represented pool-reactor 5 configurations (i.e., whether the SFP is connected to the reactor), spent fuel loading 6 configurations, and decay heat levels. Five OCPs were identified based on the timing of fuel 7 movement, key changes in pool-reactor configuration, and peak assembly and whole pool 8 decay heat curves, as listed in Table H-25. Note that, while the beyond-design-basis 9 earthquake described above is equally likely to happen throughout an entire operating cycle, the 10 conditional probability of it occurring during a given OCP is the length of time in an OCP divided 11 by the duration of the entire operating cycle (i.e., fraction of time in each OCP).

12 13 Table H-25 Operating Cycle Phase Descriptions OCP Time % of Total OCP Pool-Reactor OCP Description Duration Operating No. Configuration*

(days) Cycle 1 Defueling of reactor core (~1/3 core) 2-8 0.9 Refueling Reactor testing, maintenance, 2 8-25 2.4 Refueling inspection and refueling Highest decay power portion of 3 25-60 5 Unconnected non-outage period Next highest decay power portion of 4 60-240 25.7 Unconnected non-outage period 240-700; 5 Remainder of operating cycle 66 Unconnected 0-2 14 *Note: The refueling pool-reactor configuration refers to the configuration in which the SFP and the reactor are 15 hydraulically connected. During other stages of the operating cycle, the SFP and reactor are not connected.

16 17 As part of scenario development, the study also considered onsite mitigation and offsite support.

18 It treated onsite mitigation by modeling two cases, successful and unsuccessful mitigation, for 19 each scenario. Successful mitigation occurred when mitigative actions required by 20 10 CFR 50.54(hh)(2) were successfully deployed, additional onsite capabilities were used to 21 extend the use of the mitigation equipment, and arrival of offsite resources allowed the 22 mitigative equipment to be used until onsite capabilities could be recovered. Unsuccessful 23 mitigation occurred when none of the onsite mitigative actions were successful for an extended 24 period. Offsite support was treated using the following assumptions:

25 26

Offsite support arrives within 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months .

27 28

Actions are planned, and equipment is staged within 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months .

29 30

The accident progression analysis is truncated if the fuel is not uncovered and the pool 31 can be refilled by 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months with an injection rate of 500 gallons per minute.

32 33

If the above mitigation actions are unsuccessful, the sequence is run to 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months .

34 35 To develop accident scenarios, the NRC made several key assumptions based on structural 36 analyses, including (1) all offsite and onsite alternating current power is lost as a result of the 37 seismic event, (2) direct current power may be lost, (3) 10 CFR 50.54(hh)(2) equipment, when 38 credited, is available for the duration of the event, (4) tearing of the SFP liner is possible, and 39 (5) there is no failure of penetrations. Based on these and other assumptions, the NRC 40 developed six accident cases for each OCP using a combination of zero, small, and moderate H-115 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 leakage damage states with successful and unsuccessful mitigation actions taken for each 2 leakage scenario. The staff used these accident cases for both high- and low-density loading 3 configurations, as summarized in Table H-26.

4 5 Table H-26 Scenario Descriptions for a Given Operating Cycle Phase Scenario Characteristics Case No.

SFP Leakage Rate Mitigation?

1 Yes None 2 No 3 Yes Small 4 No 5 Yes Moderate 6 No 6

7 8 MELCOR Severe Accident Progression and Source Term Analyses 9

10 Analysts used the MELCOR code (Version 1.8.6) to model severe accident progression for the 11 scenarios described in the previous section. Enclosure H-1, Description of Analytical Tools and 12 Capabilities, to this appendix describes the MELCOR code. The code was ideal for modeling 13 accident progression for SFPs because SFP models had already been developed and 14 validated, and it was also capable of modeling in-building transport/retention and radionuclide 15 release, the latter of which was a key input for subsequent accident consequence analysis 16 modeling using the MELCOR Accident Consequence Code System (MACCS).

17 18 To facilitate modeling of the SFP for BWR fuel assemblies, the staff used a recently developed 19 rack component for improved spent fuel rack modeling and an oxidation kinetics model. These 20 two additions to MELCOR enabled the evaluation of two types of SFP accidents: a partial 21 loss-of-coolant inventory or boiloff accident, and a complete loss-of-coolant inventory accident.

22 A partial loss-of-coolant inventory or boiloff accident could involve no or late uncovery of the 23 bottom of the racks, and boiloff of the coolant could ultimately lead to hydrogen combustion. A 24 complete loss-of-coolant accident occurs when the bottom of the racks is uncovered, leading to 25 air oxidation of the cladding and enhanced ruthenium release.

26 27 The staff used the radionuclide package in MELCOR to model the release and transport of 28 fission product vapors and aerosols. It tracks radionuclides by combining them into material 29 classes, which are groups of elements with similar chemical and transport behavior. The SFP 30 MELCOR model includes 15 default material classes and 2 user-defined classes that can model 31 cesium iodide and cesium molybdate behavior. This study modified the default cesium, iodine, 32 and molybdenum radionuclide classes to accommodate new insights obtained from the Phebus 33 experimental program. 32 In addition, the staff developed a new ruthenium release model in 34 which it adjusted the default vapor pressure parameters for the ruthenium material class to 35 match the ruthenium dioxide vapor pressure at 2,200 K. However, it only used this latter model 36 in scenarios involving rapid draindown (i.e., moderate leak rates) in the SFP. All scenarios 37 applied a 5 percent gap release criterion.

38 32 The PHEBUS Fission Products international research program took place between 1988 and 2010. Its purpose was to improve the understanding of the phenomena occurring during a core meltdown accident in a light-water reactor and to reduce uncertainties in calculated radionuclide releases for reactor safety evaluations that model core meltdown accidents.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-116

1 The decay heat and radionuclide packages were used to calculate the fission product inventory 2 and specific decay power for 29 elemental groups; the specific elemental decay power is 3 compiled as a function of time after shutdown. Because these packages were originally 4 designed for reactor accident progression analyses, the shutdown time for each assembly is the 5 same. Unlike the case for reactor accidents, SFP accidents involve fuel assemblies with 6 multiple shutdown times. To address this discrepancy, a scaling procedure in MELCOR 7 enabled the use of batch-average decay heat results. Each batch also used a post-processing 8 routine with MELCOR-predicted release fractions and actual inventories. Lastly, to map the 9 calculated releases from MELCOR to the MACCS 33 code for accident consequence analyses, 10 the MELCOR input file was modified to enable tracking of fission product releases from each 11 ring, or collection of assemblies in the MELCOR radial nodalization, as well as the subsequent 12 releases to the environment.

13 14 To calculate the above mentioned radionuclides and decay heats, the reference plants utility 15 provided information for all assemblies that had been discharged from the reference plant to the 16 SFP over 18 cycles. From this information, the actual analysis basis for the high-density SFP 17 inventory was 3,055 assemblies, based on the SFP capacity of 3,819 assemblies minus 18 764 assemblies to accommodate a full core offload capability. Although the utility provided data 19 for 18 discharge cycles, this study only included cycles 7-18, since these cycles provided the 20 requisite target inventory (3,055 assemblies). For the burnup analysis, the ORIGEN code 21 simulated the irradiation and decay history for each of the 3,055 assemblies. In this case, the 22 assemblies were each decayed to a reference date, which was the end of the last cycle (18),

23 and the resulting inventories were combined into groups for analysis. These analysis groups 24 were additionally decayed to determine assembly activities and decay heat power to simulate 25 cooling of the discharged fuel after reactor shutdown. The assemblies were then placed into six 26 groups according to the cycle in which they were discharged. The benefit of grouping these 27 assemblies in this manner is that it facilitated the use of the data for analyses of low-density 28 SFP configurations in which assemblies that had been cooled for more than 5 years were 29 removed.

30 31 Description of SFP MELCOR Models 32 33 The SFP for the reference plant is located on the refueling floor of the reactor building. In one 34 corner of the SFP is a cask area. At the bottom of the SFP, high-density SFP racks are located 35 to store the SFP. During operation, these racks are covered with approximately 23 feet of water 36 to provide radiation shielding. Each rack is rectilinear in shape and comes in nine different 37 sizes, and a total of 3,819 storage locations are located in the pool. Each stainless-steel rack 38 includes cell assemblies, a baseplate with flow-through holes, and base support assemblies.

39 40 For the entire SFP model, MELCOR used a series of control volumes for regions at the top and 41 bottom of the SFP (see Figures 39 and 40 in NUREG-2161). The region at the bottom of the 42 SFP containing the empty and loaded spent fuel storage racks was more finely divided into 43 several control volumes to enable detailed analyses of all 3,819 storage locations for high- and 44 low-density configurations. The BWR assembly canisters were modeled using the MELCOR 45 canister component. In addition to the detailed SFP model, the staff used a simplified reactor 46 building model consisting solely of the refueling room, since the bulk of the reactor building 33 At the time of this analysis, the MACCS code was called the MACCS2 code, a leftover notation from the time that the original MACCS code was substantially upgraded to Version 2. Since then, the staff has referred to the code as the MACCS code and notes the version number of the code used in a particular analysis, since code development and maintenance continues.

H-117 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 components do not play a significant role in SFP accidents. The refueling room was modeled 2 using a single control volume in MELCOR, which accounted for nominal reactor building 3 leakage and simulated overpressure failure flowpaths.

4 5 To model reactor outages in which the SFP and the reactor are hydraulically connected 6 (i.e., OCP1 and OCP2), a single control volume represented the reactor well and 7 separator/dryer pool. This control volume was then connected to the spent fuel model 8 described above for the analyses. For each OCP, the assembly layout was also modified to 9 account for assembly offloads for both the high- and low-density loadings.

10 11 MELCOR Accident Progression Analysis Results and Source Terms 12 13 The MELCOR analyses of the six cases per OCP and illustrated in Table H-26 revealed that 14 four classes of scenarios did not lead to a release:

15 16

boiloff scenarios with no SFP leaks 17 18

mitigated scenarios for small leaks 19 20

unmitigated scenarios in late phases (OCP4, OCP5) 21 22

mitigated moderate leak scenarios in OCP2, OCP3, OCP4, and OCP5 23 24 For the boiloff scenarios, a simplified MELCOR model in which all assemblies are combined in 25 only two rings (collections of assemblies) that represent the fuel and empty cells was used to 26 estimate the pool heatup and water level drop. The study used the thermal-hydraulic models in 27 MELCOR, and the simplified model for boiloff, to evaluate sets of both low-density and 28 high-density cases. For both sets, no release occurred because the water level never dropped 29 below the top of the SFP racks. If boiloff of the coolant below the top of the SFP racks had 30 occurred, it could have led to steam generation, oxidation of the cladding, hydrogen production, 31 and possibly hydrogen combustion and release of radionuclides. Similarly, none of the 32 mitigated scenarios for small leaks led to release during any OCP because the rate of water 33 injection (500 gallons per minute) as a mitigative action ensured that the fuel never became 34 uncovered or overheated.

35 36 The results of MELCOR analyses of the unmitigated scenarios in OCP4 and OCP5 indicated 37 that, although there was fuel heatup in both high- and low-density configurations after the rack 38 baseplate was uncovered, there was no release because the total decay heat of the assemblies 39 in these stages was at least 37 to 48 percent lower than the total decay heat of assemblies in 40 OCP3, and natural circulation was sufficient to slow down the rate of fuel heatup to the point at 41 which the fuel failure could occur.

42 43 For moderate leaks, mitigation involved spray activation for outage phases OCP1 and OCP2, 44 and direct injection for post-outage phases OCP3, OCP4, and OCP5. The results of analyses 45 of moderate leaks during phase OCP2 indicated that no releases occurred from various heat 46 transfer mechanisms. Since the unmitigated scenarios for phases OCP3, OCP4, and OCP5 led 47 to no release, the study only evaluated the results of the high-density moderate leak scenario 48 for phase OCP3 (with and without spray flow turned on). The staff determined that modeling the 49 mitigation of moderate leak scenarios with and without the spray mechanism activated led to no 50 release of radionuclides because the fuel clad temperature never surpassed 900 degrees NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-118

1 Celsius (C) (1,652 degrees Fahrenheit (F)), at which point gap release would begin to occur. A 2 key observation was that these results underscored the importance of natural circulation of air 3 through the racks for heat removal to help keep the fuel clad temperatures below the gap 4 release temperature. The study also modeled the moderate leak scenario for OCP3, assuming 5 an additional 3-hour delayed activation of the spray for a spray activation time of 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months after 6 the leak occurs. In this case, it was shown that the maximum clad temperature reached just 7 under 627 degrees C (1,160 degrees F) after 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months , at which point the activated spray was 8 sufficient to keep the fuel clad well below the gap release temperature of 900 degrees C 9 (1,652 degrees F).

10 11 The 14 scenarios that led to release of radionuclides can be categorized as follows:

12 13

unmitigated small leaks in OCP1, OCP2, and OCP3, in both high- and low-density 14 configurations 15 16

unmitigated moderate leaks in OCP1, OCP2, and OCP3, in both high- and low-density 17 configurations 18 19

mitigated moderate leak in OCP1 in both high- and low-density configurations 20 21 Tables H-27 and H-28 summarize the release characteristics for the 14 scenarios that led to a 22 release of radionuclides.

23 24 Table H-27 Summary of Release Results for High-Density Configurations Scenario Characteristics Release Characteristics High-Density Cesium Cs-137 Iodine I-131 SFP 50.54(hh)(2)

Case No. Release at Released Release at Released Leakage Equipment?

72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months (MCi) 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months (MCi)

Small No 0.6% 0.33 3.5% 0.27 OCP1 Moderate Yes 0.5% 0.26 5.0% 0.39 Moderate No 1.5% 0.8 2.1% 0.16 Small No 17.1% 7.90 17.1% 1.91 OCP2 Moderate No 1.6% 0.73 2.0% 0.22 Small No 42.0% 24.20 51.2% 0.73 OCP3 Moderate No 0.7% 0.39 0.7% 0.01 25 26 27 Table H-28 Summary of Release Results for Low-Density Configurations Scenario Characteristics Release Characteristics Low-Density Cesium Cs-137 Iodine I-131 SFP 50.54(hh)(2)

Case No. Release at Released Release at Released Leakage Equipment?

72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months (MCi) 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months (MCi)

Small No 3.1% 0.33 4.6% 0.36 OCP1 Moderate Yes 1.8% 0.19 7.0% 0.55 Moderate No 0.5% 0.05 1.7% 0.13 Small No 1.7% 0.28 3.3% 0.37 OCP2 Moderate No 0.4% 0.07 0.7% 0.08 Small No 0.6% 0.10 1.2% 0.02 OCP3 Moderate No 0.1% 0.02 0.2% 0.00 H-119 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Unmitigated moderate leaks for high-density configurations in OCP1, OCP2, and OCP3 did not 2 lead to hydrogen deflagration, and the releases were relatively low since oxygen depletion 3 limited clad oxidation and fuel heatup. Similarly, none of the scenarios for the low-density 4 configurations led to hydrogen deflagration, and the release fractions were typically low and 5 comparable to the analogous scenario for the high-density loading configuration. One exception 6 to this trend is the low-density OCP1 scenario for mitigated moderate leaks. In this case, the 7 low-density case has slightly higher releases than the high-density cases because there was 8 higher and faster heatup of the most recently discharged assemblies in the low-density cases.

9 The higher initial fuel temperatures in the low-density case led to slightly higher releases.

10 Notably, the highest release fractions for cesium and iodine were observed for scenarios that 11 led to hydrogen combustion; namely, unmitigated small leaks for high-density configurations in 12 OCP2 and OCP3.

13 14 The release data in the tables above were used as input for the accident consequence 15 analyses, as described in the following section.

16 17 MACCS Consequence Analyses 18 19 Based on results from the MELCOR modeling of SFP accident progression scenarios, the staff 20 used Version 2 of the MELCOR Accident Consequence Code System (MACCS, Revision 3.7.0) 21 to model offsite consequence analyses. MACCS can evaluate the impacts of atmospheric 22 releases of radioactive aerosols and vapors on human health and on the environment by using 23 site-specific weather conditions, population data, and evacuation plans. Quantification of the 24 effects of offsite radioactive releases on human health is accomplished by modeling and 25 evaluating the relevant dose pathways; namely, cloudshine, inhalation, groundshine, and 26 ingestion. Enclosure H-1 to this appendix describes the MACCS suite of codes.

27 28 A source term definition was created for each accident consequence evaluation as described 29 below. The ORIGEN code calculated the activity levels of the different radionuclides of the fuel 30 in the SFP, while the plume characteristicsincluding chemical group release rates, aerosol 31 size distributions, density, and mass flow rateswere obtained from the MELCOR analyses 32 described in the previous section. The 14 MELCOR sequences that led to release (see 33 Tables H-27 and H-28 above) were binned by their cesium (Cs)-137 and iodine (I)-131 release 34 activities to lessen the computational cost of the MACCS calculations. Sequences were first 35 grouped into three bins based on their Cs-137 release activities (i.e., 0-0.25, 0.25-0.55, and 36 greater than 0.55 megacuries (MCi) of Cs-137 released) because Cs-137 is the most significant 37 contributor to long-term consequences and groundshine dose. The sequences were then 38 binned based on I-131 release (i.e., 0-0.5, 0.5-5, and greater than 5 MCi of I-131 released) 39 because I-131 is a good indicator for short-lived radionuclides that may be released from 40 recently discharged fuel. In this manner, the 14 release sequences were ultimately binned into 41 nine radiological release categories (RCs), with only four RCs containing at least two release 42 sequences. The staff chose one sequence from each of the four RCs to represent the entire 43 RC except for RC33. The study analyzed both release sequences in RC3 because these 44 release sequences had the highest releases of all sequences. The binning of the 14 MELCOR 45 sequences that led to release is illustrated in Tables H-29 and H-30 for high-density and 46 low-density loading cases with and without mitigation. The sequences that were selected for 47 further analysis are indicated in Tables H-29 and H-30 with bold text for emphasis.

48 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-120

1 Table H-29 Binning of MELCOR Release Sequences into Release Categories for 2 High-Density Configurations Scenario Characteristics Release Characteristics High-Density 50.54(hh)(2) Cs-137 I-131 Sequence SFP Release Case No. Equipment Released Released Analyzed in Leakage Category Deployed (MCi) (MCi) MACCS Small** No 0.33 0.27 RC12 Yes OCP1 Moderate Yes 0.26 0.39 RC12 No Moderate No 0.8 0.16 RC21 No Small No 7.90 1.91 RC33 Yes*

OCP2 Moderate No 0.73 0.22 RC21 Yes Small No 24.20 0.73 RC33 Yes*

OCP3 Moderate No 0.39 0.01 RC11 No 3 *The release scenarios for both sequences in RC33 were evaluated in MACCS because of the comparatively higher 4 releases compared to other scenarios.

5 **The sequences that were selected for further analysis are indicated with bold font.

6 7 Table H-30 Binning of MELCOR Release Sequences into Release Categories for 8 Low-Density Configurations Scenario Characteristics Release Characteristics Low-Density 50.54(hh)(2) Cs-137 I-131 Sequence SFP Release Case No. Equipment Released Released Analyzed in Leakage Category Deployed (MCi) (MCi) MACCS Small No 0.33 0.36 RC12 No OCP1 Moderate Yes 0.19 0.55 RC12 No Moderate No 0.05 0.13 RC11 No Small No 0.28 0.37 RC12 No OCP2 Moderate No 0.07 0.08 RC11 No Small No 0.10 0.02 RC11 Yes OCP3 Moderate No 0.02 0.00 RC11 No 9

10 *The sequence that was selected for further analysis is indicated with bold font.

11 12 The release data described above were used in MACCS for subsequent atmospheric transport 13 and dispersion modeling; exposure, dosimetry, and health effects modeling; emergency 14 response modeling; and long-term protective action modeling, as described in the next section.

15 16 MACCS Model Descriptions 17 18 Atmospheric Transport and Dispersion Modeling 19 20 The MACCS straight-line Gaussian plume segment dispersion model was used to model the 21 atmospheric transport and dispersion of radionuclides released for a given accident scenario.

22 The study divided radionuclides released into the atmosphere into plume segments that are 23 1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months or less to match the resolution of the dispersion models to that of the weather data. In 24 addition, the aerosol size distributions obtained from MELCOR, combined with the aerosol 25 velocity data obtained from NUREG/CR-7161, Synthesis of Distributions Representing 26 Important Non-Site-Specific Parameters in Off-Site Consequence Analyses, issued April 2013, 27 were used to model deposition rates of aerosols from the plume to the ground.

28 H-121 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 One year of hourly meteorological data from onsite meteorological tower observations 2 documented in NUREG-1935, State-of-the-Art Reactor Consequence Analyses (SOARCA) 3 Report, was used for atmospheric modeling in this study. Specifically, the study used 4 meteorological data from the year 2006 at the reference plant site was used. Since the exact 5 weather conditions for a potential future accident are unknown, MACCS accounts for weather 6 variability by analyzing a statistically significant set of weather trials. In this way, the modeled 7 results are an ensemble that represents the full spectrum of meteorological conditions. The 8 nonuniform weather binning strategy used to sample sets of weather data is based on the 9 approach used in NUREG/CR-7009, MACCS Best Practices as Applied in the State-of-the-Art 10 Reactor Consequence Analyses (SOARCA) Project, issued August 2014.

11 12 Exposure, Dosimetry, and Health Effects Modeling 13 14 Groundshine, cloudshine, inhalation, and ingestion are exposure pathways considered in 15 MACCS to calculate population dose and health effects. In general, food ingestion parameters 16 in NUREG/CR-6613, Volume 1, Code Manual for MACCS2: Users Guide, issued May 1998, 17 were used to calculate ingestion dose. Shielding factors applied to evacuation, normal activity, 18 and sheltering for each dose pathway were obtained from NUREG/CR-7009.

19 20 The Federal Guidance Report 13, Cancer Risk Coefficients for Environmental Exposures to 21 Radionuclides, issued September 1999, provided the dose coefficients, risk factors, and 22 relative biological effectiveness. As implemented in MACCS, the Federal Guidance Report 23 13 dose coefficients along with the dose and dose rate effectiveness factors were incorporated 24 in the dose response modeling for the early phase for doses less than 20 rem and in the 25 long-term phase of the offsite consequences. The risk factors were implemented in MACCS for 26 seven organ-specific cancers, as well as residual cancers that were not accounted for directly.

27 NUREG/CR-7161 provided parameters related to health effects, as well as other 28 non-site-specific data used for consequence analysis.

29 30 The NRC used SECPOP2000 to create a MACCS site file containing population and economic 31 data for 16 compass sectors. The site file was then interpolated onto a 64-compass sector grid 32 to improve spatial resolution for the consequence analysis. Site population data were 33 extrapolated to the year 2011 using census data from the year 2000 and a multiplier of 1.1051 34 from the U.S. Census Bureau to account for the average population growth in the United States 35 between 2000 and 2011. Similarly, economic values from the SECPOP2000 database, whose 36 values are based on year 2002 economic data, were scaled by 1.250 derived, based on the 37 consumer price index to account for price escalation (i.e., increasing value of the dollar) 38 between 2002 and 2011.

39 40 Emergency Response Modeling 41 42 The MACCS models for the emergency phase, which is the 7-day period following the start of a 43 release, calculated the dose and associated health effects to the public as well as the effects of 44 emergency preparedness actions that protect the public. To model emergency response the 45 staff developed three evacuation models based on whether 4-day dose projections were 46 expected to exceed 1 rem for a member of the public, at which point the protective action 47 guideline (PAG) was considered to be exceeded(1) a small projected dose that does not 48 exceed the PAG at the emergency planning zone (EPZ), (2) a large projected dose (within 49 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months ) that exceeds the PAG at the EPZ, and (3) a large projected dose (within 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ) 50 that exceeds the PAG at the EPZ. For each model, specific protective actions (e.g., general 51 public evacuation, hotspot relocation, shadow relocation) were included for populations within NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-122

1 and beyond the EPZ. To model population evacuation in these models, the population was 2 divided into cohorts, which are population groups that move differently from other groups. The 3 cohorts were loaded onto the roadway network at a specified time, and a set of speed values 4 were applied per cohort for the early, middle, and late periods of the evacuation. The last 5 10 percent of the population to evacuate (i.e., the evacuation tail) was modeled as a separate 6 cohort. For residents within the EPZ, the MACCS potassium iodide model used in the analysis 7 assumes that potassium iodide would only be distributed within the EPZ, and 50 percent of the 8 population within the EPZ would have access to and take it as directed.

9 10 Long-term Protective Action Modeling 11 12 MACCS was also used to model the long-term protective action phase (i.e., the 50-year period 13 following the 7-day emergency phase). Three protective actions were modeled for 14 contaminated land during the long-term phase: interdiction, decontamination, and 15 condemnation. In the MACCS model, interdiction and condemnation are defined in terms of 16 habitability. Interdiction is a temporary relocation during which land contamination levels are 17 reduced by decontamination, natural weathering, and radioactive decay to restore habitability. If 18 contamination levels cannot be adequately reduced to restore habitability within 30 years, the 19 land is considered condemned, and the population is modeled not to return during the long-term 20 phase (i.e., permanently relocated). Based on the location of the reference plant in 21 Pennsylvania, this study used a habitability criterion of 500 millirem (mrem) per year beginning 22 in the first year. Two levels of decontamination with decontamination factors of 3 and 15 were 23 modeled for a 1-year timespan. The cost of decontamination during this period was determined 24 using values in NUREG/CR-7009.

25 26 This study also considered land suitable for farming (farmability). Values used to define 27 farmability were taken from NUREG-1150, Severe Accident Risks: An Assessment for Five 28 U.S. Nuclear Power Plants, issued December 1990. Agricultural land with contamination 29 levels in excess of the farmability criteria was considered unfarmable, and no farming was 30 allowed until the farmability criteria were satisfied.

31 32 MACCS Consequence Analysis Results 33 34 Table H-31 summarizes the mean reduction in offsite consequence results in terms of averted 35 population dose (person-rem) and averted economic costs (2012 dollars) associated with 36 implementing Option 2 (expedited spent fuel transfer to achieve low-density SFP storage). The 37 reported consequence metrics represent averted consequences that were calculated by taking 38 the difference between consequences for Option 1 (regulatory baseline) and consequences for 39 Option 2.

40 41 Table H-31 Mean Reduction in Offsite Consequence Results Associated with Option 2 Consequence Metrica Best Estimate Low Estimate High Estimate Averted 50-mile Population Dose (person-rem) 124 60 1260 Averted 50-mile Economic Costs (2012 dollars) $723,300 $1,073,300 $4,587,800 42 a The reported consequence metrics represent averted consequences that were calculated by taking the difference 43 between consequences for Option 1 (regulatory baseline) and consequences for Option 2 (expedited spent fuel 44 transfer to achieve low-density SFP storage).

45 46 The consequence metrics for population dose and economic costs can vary significantly with 47 the criterion used to measure or estimate the level of land contamination and to inform decisions 48 about when to allow relocated populations to return to contaminated land areas. The offsite H-123 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 consequence analysis performed in support of SECY-13-0112 and NUREG-2161 used three 2 PAG levels based on annual dose to calculate the estimates of averted population dose and 3 averted economic costs within 50 miles: (1) the U.S. Environmental Protection Agency (EPA) 4 intermediate phase PAG level of 2 rem in the first year, and 500 mrem annually thereafter, was 5 used to calculate the best estimate, (2) the more stringent Pennsylvania PAG level of 500 mrem 6 annually starting with the first year was used to calculate the low estimate, and (3) the less 7 stringent 2 rem annually was used to calculate the high estimate. The analysis calculated all 8 estimates assuming a remaining licensed term of 22 years (until 2034) for the reference plant 9 and using the reference sites offsite population density within a 50-mile radius from the site 10 (approximately 722 people per square mile).

11 12 The study included a limited treatment of uncertainty by describing results for a range of 13 sensitivity analyses performed to evaluate the effect of certain assumptions on results and 14 insights. Factors addressed in these sensitivity analyses included the following:

15 16

using a more favorable 1x8 fuel assembly pattern 17 18

using an unfavorable uniform fuel assembly pattern 19 20

radiative heat transfer 21 22

hydrogen combustion ignition criterion 23 24

occurrence of concurrent events involving the reactor or multiunit events 25 26

molten core-concrete interaction 27 28

alternative accident scenario truncation times 29 30

effects of reactor building leakage on hydrogen combustion and accident progression 31 32 Risk Evaluation 33 34 This study was a limited scope consequence analysis supplemented with probabilistic insights 35 to provide additional context and perspectives about the relative likelihood of events and 36 consequences. This analysis considered the following as examples of probabilistic insights:

37 38

risk information from past studies for accident scenario selection 39 40

initiating event frequency information 41 42

initiating event timing effects (e.g., the relative likelihood of an event occurring during 43 each OCP and the likely configurations incurred) 44 45

relative likelihoods of damage state characteristics 46 47

probabilistic consequence analysis to account for effects of statistical variability in offsite 48 weather conditions on offsite radiological consequences 49 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-124

1 While these elements provided some of the benefits of PRA, this study did not perform several 2 elements of a traditional PRA. The following are examples of traditional PRA elements that 3 were excluded from this study:

4 5

failure modes and effects analysis (except for certain structures, systems, or 6 components specifically identified in the study) 7 8

data analysis and component reliability estimation 9

10

dependency analysis 11 12

human reliability analysis as part of the accident progression and recovery (except the 13 limited scope human reliability analysis that was performed as described above) 14 15

system fault tree and accident sequence event tree development and quantification 16 17 Figure H-20 illustrates the conditional probability of SFP liner leakage and magnitude of release 18 from the SFPconditioned on the assumed occurrence of the beyond-design-basis earthquake 19 considered in the studyfor postulated accident scenarios that occur in different phases of the 20 operating cycle. The figure shows the results for both the high-density and low-density loading 21 configurations, as well as for the mitigated and unmitigated cases.

22 23 The inclusion of probabilistic insights allowed analysts to consider some aspects of likelihood 24 but could not support making definitive statements about SFP risk. This study focused on a 25 specific portion of the overall risk profileSFP accidents caused by large seismic events 26 between 0.5g and 1g. This study can therefore be used to corroborate or challenge the 27 continued applicability of estimates for this part of the risk profile based on previous studies. In 28 addition, since large seismic events have been shown in the past to be a dominant contributor 29 to SFP risk, this comparison helps to predict whether a full-scope PRA would be expected to 30 result in an overall decrease or increase in estimated risk. Therefore, the results of this study 31 can draw supportable, but not definitive, conclusions about overall SFP risk.

32 H-125 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 2 Figure H-20 Conditional Probability of SFP Liner Leakage and SFP Release Magnitude 3

4 Cost-Benefit Analysis Results 5

6 Table H-32 summarizes the results of the quantitative cost-benefit analysis for the best estimate 7 and low- and high-estimate cases for Option 2, documented in NUREG-2161, Appendix D. At 8 the time this regulatory analysis was prepared, returns on investments were well below the 9 3 percent and 7 percent discount rates described in the Office of Management and Budget 10 (OMB) Circular No. A-4, Regulatory Analysis, dated September 17, 2003. A sensitivity 11 analysis was performed using a 0 percent discount rate that produced undiscounted values in 12 constant dollars. Although it was common practice to provide undiscounted values for costs 13 and benefits for information purposes within regulatory analyses, it was not common practice to 14 report such results as part of a sensitivity analysis. However, the staff chose to report the 15 undiscounted costs and benefits as part of a sensitivity analysis for this regulatory analysis to 16 account for current market trends and future predictions. Note that this enclosure 34 only 17 discusses the calculation of public health and offsite property attributes, which is based on the 18 detailed severe accident analysis using MELCOR and MACCS.

19 34 Methods for calculating occupational health, onsite property, and implementation costs are discussed elsewhere in NUREG-2161.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-126

1 In addition to the sensitivity analysis described above to evaluate the effect on results of using a 2 0 percent discount rate, the staff performed sensitivity analyses to account for the effect on the 3 results of (1) using an alternative dollar per person-rem conversion factor ($4,000 per 4 person-rem instead of $2,000 per person-rem), (2) extending the analysis of consequences 5 beyond a 50-mile circular radius around the site, and (3) combining the effects of using the 6 $4,000 per person-rem conversion factor and extending the analysis of consequences beyond 7 50 miles from the site. Tables H-32 and H-33 summarize the results of these sensitivity 8 analyses.

9 10 As shown in Table H-33, requiring the expedited transfer of spent fuel from the SFP to dry cask 11 storage to achieve low-density SFP storage at the reference plant did not achieve a positive net 12 benefit for eight of the nine cases presented. The undiscounted high-estimate casewhich 13 reflects the costs and benefits at the time in which they are incurred with no present worth 14 conversion and which assumes the least stringent habitability criterionresulted in a positive 15 net benefit of about $27.1 million. However, the other high-estimate cases resulted in negative 16 net benefits of about ($10.6 million) and ($25.1 million), which differed from this case by 17 adjusting future costs and benefits into 2012 dollars using 3 percent and 7 percent discount 18 rates, respectively.

H-127 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Table H-32 Summary of Benefits and Costs within 50 Miles for Option 2 Best Estimatea Low Estimatea High Estimatea Attribute Undiscounted 3% NPV 7% NPV Undiscounted 3% NPV 7% NPV Undiscounted 3% NPV 7% NPV Public Health

$247,700 $179,500 $124,600 $119,700 $86,700 $60,200 $2,520,000 $1,825,500 $1,267,000 (Accident)

Occupational Health

$1,300 $900 $700 $700 $500 $300 $21,300 $15,400 $10,700 (Accident)

Offsite Property $723,300 $524,000 $363,700 $1,073,300 $777,500 $539,700 $4,587,800 $3,323,400 $2,306,700 Onsite Property $10,400 $6,900 $4,300 $4,480 $2,950 $1,830 $378,600 $249,600 $155,800 Total Benefits $982,700 $711,300 $493,300 $1,198,200 $867,700 $602,000 $7,507,700 $5,413,900 $3,740,200 Occupational Health

($9,000)c ($24,000) ($27,000) ($9,000) ($24,000) ($27,000) ($9,000) ($24,000) ($27,000)

(Routine)

Industry

($15,660,000) ($41,820,000) ($46,770,000) ($15,660,000) ($41,820,000) ($46,770,000) ($15,660,000) ($41,820,000) ($46,770,000)

Implementation Industry Operation ($730,000) ($252,000) ($64,000) ($730,000) ($252,000) ($64,000) ($730,000) ($252,000) ($64,000)

NRC Implementation NCb NCb NCb NCb NCb NCb NCb NCb NCb NRC Operation NCb NCb NCb NCb NCb NCb NCb NCb NCb NUREG/BR-0058, Rev. 5, App. H, Rev. 0 Total Costs ($16,399,000) ($42,096,000) ($46,861,000) ($16,399,000) ($42,096,000) ($46,861,000) ($16,399,000) ($42,096,000) ($46,861,000)

Net Benefit ($15,416,000) ($41,385,000) ($46,368,000) ($15,200,800) ($41,228,300) ($46,259,000) ($8,891,300) ($36,682,100) ($43,120,800) a 2 Discounted net present value (NPV) results are expressed in 2012 dollars. Undiscounted results are expressed in constant dollars.

b 3 NC: Not calculated c Negative values are shown using parentheses (e.g., negative $9,000 is displayed as ($9,000)).

4 H-128 5 6 Table H-33 Combined Effect of $4,000 per Person-Rem Conversion Factor and Consequences Beyond 50 Miles for 7 Option 2 Best Estimatea Low Estimatea High Estimatea Attribute Undiscounted 3% NPV 7% NPV Undiscounted 3% NPV 7% NPV Undiscounted 3% NPV 7% NPV Public Health

$3,566,900 $2,583,800 $1,793,400 $2,162,500 $1,566,500 $1,087,300 $31,471,600 $22,798,200 $15,823,400 (Accident)

Occupational Health

$2,500 $1,900 $1,400 $1,300 $1,000 $700 $42,700 $30,900 $21,400 (Accident)

Offsite Property $2,139,300 $1,549,700 $1,075,600 $4,968,300 $3,599,100 $2,498,000 $11,586,600 $8,393,400 $5,825,500 Onsite Property $10,400 $6,900 $4,300 $4,680 $3,150 $2,030 $378,600 $249,600 $155,800 Total Benefits $5,719,100 $4,142,300 $2,874,700 $7,136,800 $5,169,800 $3,588,000 $43,479,500 $31,472,100 $21,826,100 Occupational Health

($18,000) c ($49,000) ($54,000) ($18,000) ($49,000) ($54,000) ($18,000) ($49,000) ($54,000)

(Routine)

Industry

($15,660,000) ($41,820,000) ($46,770,000) ($15,660,000) ($41,820,000) ($46,770,000) ($15,660,000) ($41,820,000) ($46,770,000)

Implementation Industry Operation ($730,000) ($252,000) ($64,000) ($730,000) ($252,000) ($64,000) ($730,000) ($252,000) ($64,000)

NRC Implementation NCb NC NC NC NC NC NC NC NC NRC Operation NC NC NC NC NC NC NC NC NC Total Costs ($16,408,000) ($42,121,000) ($46,888,000) ($16,408,000) ($42,121,000) ($46,888,000) ($16,408,000) ($42,121,000) ($46,888,000)

Net Benefit ($10,689,000) ($37,979,000) ($44,013,000) ($9,271,200) ($36,951,200) ($43,300,000) $27,071,500 ($10,648,900) ($25,061,900) a 8 Discounted net present value (NPV) results are expressed in 2012 dollars. Undiscounted results are expressed in constant dollars.

b 9 NC: Not calculated c Negative values are shown using parentheses (e.g., negative $18,000 is displayed as ($18,000)).

10

1 Summary and Conclusion 2

3 Table H-32 shows that requiring the expedited transfer of spent fuel from the SFP to dry cask 4 storage to achieve low-density SFP storage does not achieve a cost-beneficial increase in 5 public health and safety for the reference plant using the current regulatory framework. In 6 addition, three sensitivity analyses (Table H-33) also showed that the regulatory alternative 7 represented by Option 2 was not cost-beneficial for any cases in which costs and benefits 8 incurred in the future were discounted to their present worth using 3 percent and 7 percent 9 discount rates consistent with OMB guidance. Moreover, the staff identified other 10 considerations that would further reduce the quantified benefits, thereby making Option 2 even 11 less justifiable. These other considerations included (1) the costs and risks associated with the 12 handling and movement of spent fuel casks in the reactor building, (2) the post-Fukushima 13 mitigation strategies required under Order EA-12-049 and the reliable SFP instrumentation 14 required under Order EA-12-051, which significantly enhance the likelihood of successful 15 mitigation, and thereby reduce the conditional probability of radiological release, and (3) the 16 possibility of other favorable SFP loading configurations.

17 18 Based on its quantitative cost-benefit analysis, the staff concluded that the added costs involved 19 in expediting the transfer of spent fuel from the SFP to dry cask storage to achieve low-density 20 SFP storage at the reference plant were not warranted. In addition, based on the results of its 21 safety goal evaluation, the staff concluded that this regulatory alternative could not result in a 22 substantial increase in overall protection of public health and safety. Together, these analyses 23 indicated thatfor the reference plantrequiring the expedited transfer of spent fuel from the 24 SFP to dry cask storage to achieve low-density SFP storage was not justified.

25 26 However through its analysis, the staff discovered that an alternative 1x8 high-density fuel 27 configuration may have significantly lower implementation costs and potentially similar benefits 28 to the low-density configuration. The staff therefore recommended that this alternativein 29 addition to other possible SFP loading configurationsbe evaluated further as part of a 30 subsequent regulatory analysis that would be performed to more broadly assess whether any 31 significant safety benefits (or detriments) would occur from requiring expedited spent fuel 32 transfer from SFPs to dry storage casks for the range of SFP designs at existing and new 33 (future) nuclear power plants. In SECY-12-0095, Tier 3 Program Plans and 6-Month Status 34 Update in Response to Lessons Learned from Japans March 11, 2011, Great Tohoku 35 Earthquake and Subsequent Tsunami, dated July 13, 2012, the staff provided a five-step plan 36 to evaluate whether regulatory action is warranted for the expedited transfer of spent fuel from 37 SFPs into dry cask storage. Enclosure H-6 to this appendix summarizes the subsequent 38 regulatory analysis that addresses this issue and that is documented in COMSECY-13-0030.

39 40 Commissions Response to the Staffs Analysis and Recommendations 41 42 The staff provided SECY-13-0112 to the Commission as an information paper instead of a 43 notation vote paper. Therefore, the Commission did not vote on the staffs analysis and its 44 recommendations provided therein. However, after receiving the Tier 3 program plan 45 documented in SECY-12-0095, the Commission directed the staff in several related SRMs.

46 Enclosure H-6 to this appendix summarizes this Commission direction.

47 H-129 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 ENCLOSURE H-6:

SUMMARY

OF DETAILED ANALYSES IN 2 COMSECY-13-0030, ENCLOSURE 1, REGULATORY ANALYSIS FOR 3 JAPAN LESSONS-LEARNED TIER 3 ISSUE ON EXPEDITED 4 TRANSFER OF SPENT FUEL 5

6 This enclosure summarizes the U.S. Nuclear Regulatory Commission (NRC) staffs regulatory 7 analyses of whether expedited transfer of spent fuel to dry cask storage is warranted, as 8 documented in COMSECY-13-0030, Staff Evaluation and Recommendation, Enclosure 1, 9 Regulatory Analysis for Japan Lessons-Learned Tier 3 Issue on Expedited Transfer of Spent 10 Fuel, dated November 12, 2013. These analyses used insights from and expanded upon the 11 staffs previous evaluations described in NUREG-2161, Consequence Study of a 12 Beyond-Design-Basis Earthquake Affecting the Spent Fuel Pool for a U.S. Mark I Boiling-Water 13 Reactor, issued September 2014, and SECY-13-0112, Enclosure 1, Consequence Study of a 14 Beyond-Design-Basis Earthquake Affecting the Spent Fuel Pool for a U.S. Mark I Boiling-Water 15 Reactor, dated October 9, 2014, and summarized in Enclosure H-5, Summary of Detailed 16 Analyses for SECY-13-0112 and NUREG-2161, Consequence Study of a Beyond-Design-Basis 17 Earthquake Affecting the Spent Fuel Pool for a U.S. Mark I Boiling-Water Reactor, of this 18 appendix. As such, this enclosure should be considered with the content of Enclosure H-5.

19 Problem Statement and Regulatory Objectives 20 21 The March 11, 2011, Great Thoku earthquake and subsequent tsunami in Japan caused 22 extensive damage to the nuclear reactors at the Fukushima Dai-ichi nuclear power plant.

23 Although the spent fuel pools (SFPs) and spent fuel assemblies remained intact, the event led 24 to questions about the safe storage of spent fuel in SFPs and whether expedited transfer of 25 spent fuel to dry cask storage was necessary. The event also generated increased interest in 26 understanding the consequences of SFP accidents. On March 23, 2011, the NRC, in response 27 to the accident at Fukushima Dai-ichi, on March 23, 2011, the NRC established a Near-Term 28 Task Force (NTTF) to determine whether the NRC should make any near- or long-term 29 improvements to its regulatory system, based on insights obtained from the Fukushima Dai-ichi 30 accident. Nearly 4 months later, the NTTF provided its recommendations for regulatory 31 improvements, including those to enhance SFP safety, in a Task Force Report to the 32 Commission (NRC, 2011b). Around the same time, the NRC Office of Nuclear Regulatory 33 Research initiated a project evaluating the consequences of a beyond-design-basis earthquake 34 affecting an SFP at a Mark I boiling-water reactor in the United States. The results of this study, 35 hereafter referred to as the Spent Fuel Pool Study (SFP study), were later documented in 36 NUREG-2161 and SECY-13-0112, Enclosure 1, and are summarized in Enclosure H-5 of this 37 appendix.

38 39 In accordance with Commission direction, the staff prioritized its recommendations in 40 SECY-11-0137, Prioritization of Recommended Actions to Be Taken in Response to 41 Fukushima Lessons Learned. The staff identified expedited transfer of spent fuel to dry cask 42 storage as an additional issue that was not identified in the Task Force Report but may warrant 43 further consideration. In SECY-12-0025, Proposed Orders and Requests for Information in 44 Response to Lessons Learned from Japans March 11, 2011, Great Thoku Earthquake and 45 Tsunami, dated March 9, 2012, the staff prioritized this issue in the Tier 3 category, since it 46 required further staff study to determine whether it warranted regulatory action. The staff also 47 proposed two orders to the Commission that would increase SFP safety by (1) requiring 48 installation of enhanced SFP instrumentation and (2) developing additional strategies and NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-130

1 guidance to mitigate beyond-design-basis phenomena by maintaining or restoring SFP cooling, 2 core cooling, and containment capabilities.

3 4 The Commission approved these orders aimed at improving spent fuel safety:

5 6 1) Order EA-12-049, Order Modifying Licenses with Regard to Requirements for Mitigation 7 Strategies for Beyond-Design-Basis External Events, dated March 12, 2012 8

9 This Order requires licensees to develop, implement, and maintain guidance and 10 strategies to maintain or restore SFP cooling capabilities, independent of alternating 11 current power, following a beyond-design-basis external event.

12 13 2) Order EA-12-051, Issuance of Order to Modify Licenses with Regard to Reliable Spent 14 Fuel Pool Instrumentation, dated March 12, 2012 15 16 This Order requires licensees to install reliable means of remotely monitoring wide-range 17 SFP levels to support effective prioritization of event mitigation and recovery actions in 18 the event of a beyond-design-basis external event.

19 20 In SECY-12-0095, Tier 3 Program Plans and 6-Month Status Update in Response to Lessons 21 Learned from Japans March 11, 2011, Great Tohoku Earthquake and Subsequent Tsunami, 22 dated July 13, 2012, the staff outlined a five-step plan to evaluate the Tier 3 issue of whether 23 regulatory action to expedite the transfer of spent fuel to dry cask storage was needed.

24 25 In a memorandum to the Commission entitled, Updated Schedule and Plans for Japan 26 Lessons-Learned Tier 3 Issue on Expedited Transfer of Spent Fuel, dated May 7, 2017, the 27 staff provided a shortened three-phase plan for resolving the Tier 3 Issue on expedited transfer 28 of spent fuel. The first phase of the plan was to conduct a regulatory analysis, leveraging 29 results and insights from the ongoing SFP study, to determine whether a substantial increase in 30 public health and safety can be achieved through an expedited transfer to dry storage casks.

31 Then, if the results of the regulatory analysis indicated that it warranted additional study, the 32 staff would proceed to Phases 2 and 3 of the plan and perform more detailed analyses using 33 refined assumptions to confirm the need for regulatory action. The staff provided its findings 34 from the Phase 1 study to the Commission in COMSECY-13-0030, which are summarized 35 below.

36 37 Regulatory Alternatives 38 39 The staff considered two regulatory alternatives in its analysis:

40 41

Option 1: Maintain the existing spent fuel storage requirements (regulatory baseline).

42 This option, hereafter referred to as the regulatory baseline, refers to the case in which 43 the Commission opts to continue with the existing licensing requirements for spent fuel 44 storage rather than require the expedited transfer of spent fuel from SFPs to dry storage.

45 The existing regulations require that spent fuel, which is stored in SFPs in high-density 46 racks, be moved from SFPs into dry cask storage only when necessary to accommodate 47 spent fuel being offloaded from the core. In addition, the SFP must always allocate 48 enough space to accommodate at least one full core of reactor fuel in case of 49 emergencies or other operational contingencies. The regulatory baseline assumed that 50 all applicable requirements and guidance to date have been implemented, but it 51 assumed no implementation for any related generic issues or other staff requirements or H-131 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 guidance that were unresolved or still under review. For the regulatory analysis, the 2 baseline condition assumed that spent fuel was stored in high-density racks in a 3 relatively full SFP, and that there was full compliance with all regulatory requirements, 4 including those outlined in Title 10 of the Code of Federal Regulations (10 CFR) 5 50.54(hh)(2) with respect to spent fuel configuration and SFP preventive and mitigative 6 capabilities. To increase conservatism in the analysis, for the regulatory baseline it was 7 assumed that there was no successful mitigation of the SFP accident. In addition, 8 because SFPs are relatively full even after using high-density storage racks, the current 9 practice of transferring spent fuel to dry storage in accordance with 10 CFR Part 72, 10 Licensing Requirements for the Independent Storage of Spent Nuclear Fuel, High-Level 11 Radioactive Waste, and Reactor-Related Greater Than Class C Waste, is assumed to 12 continue. Lastly, although the it was assumed that licensees had implemented the 13 requirements of Order EA-12-049 and Order EA-12-051 to enhance their ability to 14 respond to beyond-design-basis events, the staffs evaluation did not quantitatively 15 consider the capabilities implemented to satisfy these requirements. The regulatory 16 baseline represents the status quo against which the second alternative is compared.

17 18

Option 2: Expedite the transfer of spent fuel from SFPs to dry cask storage (low-density 19 SFP). For this alternative, spent fuel assemblies that have been cooled in the SFP for at 20 least 5 years after discharge would be expeditiously moved from the SFP to dry cask 21 storage beginning in 2014 to achieve and maintain low-density loading of spent fuel in 22 the existing high-density racks. For this option, the SFP would have a lower long-lived 23 radionuclide inventory, a lower overall heat load, and a slightly higher water inventory 24 because of the removed spent fuel assemblies. On the other hand, loading, handling, 25 and moving casks to achieve this configuration increase the cost and risk impacts 26 associated with this alternative. Therefore, to maximize the delta benefit of this 27 alternative relative to the status quo (i.e., Option 1), the staffs analysis conservatively 28 did not include these additional costs and risks associated with transferring and handling 29 casks in their analyses. The staff also assumed that mitigative actions in accordance 30 with 10 CFR 50.54(hh)(2) were successful to further increase the regulatory benefit of 31 this alternative, and, similar to Option 1, did not quantitatively consider the requirements 32 of Order EA-12-049 and Order EA-12-051 in the evaluation.

33 34 Safety Goal Evaluation 35 36 As part of its two-part regulatory analysis, the staff performed a safety goal screening evaluation 37 to determine whether requiring the expedited transfer of spent fuel to dry cask storage would 38 provide a significant safety benefit compared to the regulatory baseline, regardless of whether 39 the action would be cost-beneficial. The staff performed the safety goal screening evaluation by 40 comparing the calculated risks to the public from the severe accidents at the plants considered 41 in this study to the two quantitative health objectives (QHOs) for average individual prompt 42 fatalities and average individual latent cancer fatalities, as outlined in the NRCs Safety Goals 43 Policy Statement (NRC, 1986). These QHOs, which are subsequently used to determine 44 whether the NRCs safety goals are met, are as follows:

45 46 (1) The risk to an average individual near a nuclear power plant of prompt fatalities that 47 might result from reactor accidents should not exceed 1/10 of 1 percent (0.1 percent) of 48 the sum of prompt fatality risks resulting from other accidents to which members of the 49 U.S. population are generally exposed.

50 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-132

1 (2) The risk to the population in the area near a nuclear power plant of cancer fatalities that 2 might result from nuclear power plant operation should not exceed 1/10 of 1 percent 3 (0.1 percent) of the sum of cancer fatality risks resulting from all other causes.

4 5 For an average individual within 1.6 kilometers (1 mile), the prompt fatality QHO is 6 5x10-7 per year as estimated in NUREG-0880, Revision 1, Safety Goals for Nuclear Power 7 Plant Operation, issued May 1983. The staffs analysis for expedited transfer of spent fuel 8 showed that there are no offsite early fatalities from acute radiation effects, despite the large 9 releases for some low-probability accident progressions analyzed.

10 11 The cancer fatality QHO listed in NUREG-0880, Revision 1, is 2x10-6 per year for an average 12 individual living within 16 kilometers (10 miles) of a nuclear power plant site. The staff 13 calculated an updated QHO value for comparison, using the most up-to-date estimate of the 14 number of cancer fatalities and the total U.S. population at the time, which yielded a risk of 15 1.84x10-3 per year. One-tenth of 1 percent of this value results in a QHO of 1.84x10-6 per year, 16 which is lower than the value listed in NUREG-0880.

17 18 The staff determined the risk of latent cancer fatalities to a population living near a nuclear 19 power plant by multiplying the bounding frequency of damage to spent fuel (3.46x10-5 per year) 20 with the estimate from the SFP study for conditional individual latent cancer fatality risk within a 21 16-kilometer (10-mile) radius (4.4x10-4 per year). This yielded a conservative high estimate of 22 individual latent cancer fatality risk of 1.52x10-8 cancer fatalities per year for an SFP accident, 23 which is less than one percent of the 1.84x10-6 per year QHO calculated above.

24 25 The staff noted three important limitations to the above evaluation:

26 27 (1) The safety goals outlined in the Safety Goal Policy Statement are intended to 28 encompass all accident scenarios at a nuclear power plant site, while this analysis only 29 considered initiating events that challenge the integrity or cooling of the SFP, which are 30 the most important contributors to SFP risk.

31 32 (2) Although an SFP accident might affect larger areas and more people than a reactor 33 accident, protective actions, such as relocation of the public, would result in the risks to 34 individuals beyond 16 kilometers (10 miles) being similar to the risk to individuals located 35 closer to the plant.

36 37 (3) The total or cumulative radiation dose to the population might be higher for an SFP 38 accident than for a reactor accident, even though the risk to individuals living near or far 39 from the plant remains below the QHOs.

40 41 Based on these results, the staff concluded that the continued use of high-density loadings in 42 SFPs at nuclear power plants does not challenge the NRCs safety goals. Expediting transfer of 43 spent fuel into dry cask storage would provide no more than a minor safety improvement.

44 H-133 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 Technical Evaluation 2

3 Description of Representative Plants 4

5 The staff organized U.S. SFPs into seven groups based on spent fuel configuration, rack 6 designs, and SFP capacities, as shown in Table H-34.

7 8 Table H-34 SFP Groupings Used for the Staff's Technical and Cost-Benefit Analyses SFP No. of Average Year When No. of Group Description Reactor Reactor Operating SFPs No. Units License Expires Boiling-water reactors (BWRs) with Mark I and Mark 1 31 31 2037 II containments and with nonshared SFPs Pressurized-water reactors (PWRs) and BWRs with 2 49 49 2040 Mark III containments with nonshared SFPs 3 AP1000 SFPs 4 4 2078 4 Reactor units with shared SFPs 20 10 2038 5 SFPs located below grade1 (these are included in group 2)

Decommissioned plants with spent fuel stored in 6 7 6 N/A pool2,3 Decommissioned plants with fuel stored in an ISFSI 7 21 N/A N/A using dry casks

1. Group 5 is a special set of currently operating PWRs for which damage to the pool structure would not result in a rapid loss of water inventory.

2 The Zion 1 and 2 decommissioned reactor units share a single SFP.

3 Group 6 includes the GE-Hitachi Morris wet independent spent fuel storage installation (ISFSI) site.

9 10 The technical evaluations discussed in this section and the cost-benefit analyses focused on 11 Group 1 through Group 4 in Table H-34; the analyses excluded Group 5 through Group 7 for the 12 following reasons:

13 14

Group 5 SFPs are less susceptible to the formation of small or medium leaks because 15 there is no open space around the pool liner and concrete structure.

16 17

Group 6 SFPs are no longer receiving spent fuel discharged from the reactor following 18 decommissioning, and several plants had extended plant outages before announcing 19 cessation of plant operation.

20 21

The spent fuel in Group 7 is already in dry cask storage.

22 23 The analyses also included operational strategies such as those used to expand onsite storage.

24 25 Spent Fuel Pool Accident Modeling 26 27 The analyses described relied heavily on the models and data used in the SFP study.

28 NUREG-2161 and SECY-13-0112, Enclosure 1, provide more detailed information about the 29 models developed for the SFP study. This subsection focuses on the most relevant technical 30 information that will enable comprehension of the cost-benefit analyses described in the next 31 section.

32 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-134

1 Seismic Hazard Model and Characterization of Seismic Event Likelihood 2

3 The analyses used the 2008 U.S. Geological Survey seismic hazard model that was available at 4 the time (and used for the SFP study) to evaluate seismic hazards at central and eastern 5 U.S. nuclear plants. Although this model considered hazards at western U.S. sites (e.g., Diablo 6 Canyon), the accident analyses did not include western sites because they were not addressed 7 in Generic Issue 199, 35 which only focused on central and eastern U.S. sites. Using peak 8 ground acceleration and hazard exceedance frequency data from the U.S. Geological Survey, 9 the staff determined that the hazard exceedance frequency curves of the Peach Bottom Atomic 10 Power Station (Peach Bottom), the reference plant used for the SFP study, bound those of 11 reactors in SFP Group 1 through Group 4 over a wide peak ground acceleration range.

12 13 To translate hazard exceedance frequencies into seismic initiating event frequencies, the staff 14 also partitioned the peak ground acceleration ranges for Peach Bottom and for sites in SFP 15 Group 1 through Group 4 into four discrete bins. Since the SFP study demonstrated that 16 damage to the SFP and other related structures was not credible for seismic bins 1 and 2, the 17 staff only used seismic initiator event frequencies from bins 3 and 4 of each SFP group (and 18 Peach Bottom). Specifically, the analyses used seismic initiating event frequencies from bins 3 19 (1.7x10-5 per year) and 4 (4.9x10-6 per year) for Peach Bottom for both the low- and base-case 20 analyses because these hazard exceedance frequencies bound most of the other reactor sites.

21 To account for some reactor site hazard exceedance frequencies exceeded those of Peach 22 Bottom for bins 3 and 4, for each SFP group, the analyses used the site with the largest plant 23 exceedance frequencies in bins 3 and 4 to generate high-estimate seismic initiating event 24 frequencies for subsequent sensitivity analyses (see Table H-35).

25 26 Consequence Analyses 27 28 The MELCOR Accident Consequence Code System (MACCS 36) code was used to model 29 atmospheric transport and dispersion, emergency response, and long-term consequences. The 30 atmospheric transport and dispersion model used for these analyses was based on the Peach 31 Bottom MACCS results described in the SFP study. The MACCS model for Peach Bottom used 32 a straight-line Gaussian plume segment model. For both the SFP study and this study, the 33 atmospheric release of radionuclides was discretized into up to 1-hour plume segments to 34 account for variations in the release rate and the changes in wind direction. Meteorological data 35 used for the MACCS analyses consisted of 1 year of hourly meteorological data (i.e., 8,760 data 36 points for each meteorological parameter) for Peach Bottom evaluated in the SFP study. The 37 specific year of meteorological data chosen for Peach Bottom was 2006, and stability class data 38 were derived from temperature measurements at two elevations on the site meteorological 39 towers.

40 41 The study used population densities and site distribution characteristics for SFPs in the United 42 States to generate the site population and economic data required for MACCS and cost-benefit 43 analyses. The SFP sites were binned based on average population densities within 44 80 kilometers (50 miles) of the sites, and representative sites were selected to represent various 45 population densities. Peach Bottom, Surry Power Station, Palisades Nuclear Plant, and Point 46 Beach Nuclear Plant represented population densities in the 90th percentile, the mean, the 35 https://www.nrc.gov/about-nrc/regulatory/gen-issues/dashboard.html#genericIssue/genericIssueDetails/3 36 At the time of this analysis, the MACCS code was called the MACCS2 code, a leftover notation from the time that the original MACCS code was substantially upgraded to Version 2. Since then, the staff has referred to the code as the MACCS code and notes the version number of the code used in a particular analysis since code development and maintenance continues.

H-135 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 median, and the 20th percentiles, respectively. For each representative site, site population and 2 economic data were created for 16 compass sectors and interpolated onto a 64-compass sector 3 grid for better spatial resolution for consequence analyses. The staff escalated 2000 census 4 data and 2002 economic data to 2011 values.

5 6 Population densities and distributions near SFP locations representing the 90th, mean, median, 7 and 20th percentiles were used for respective high-, base-, median-, and low-estimate 8 sensitivity studies of site population demographics. The study used these data as additional 9 inputs into MACCS calculations to assess the effect of population density on the averted public 10 health (accident) attribute. Since an SFP fire could affect public health consequences beyond 11 80 kilometers (50 miles), sensitivity analyses were also conducted using base-case 12 assumptions and the standard value ($2,000 per person-rem), along with a sensitivity value 13 ($4,000 per person-rem) for the person-rem conversion factor. The study used the $4,000 per 14 person-rem sensitivity value because the staff was reassessing the dollar per person-rem factor 15 at the time as part of its efforts to update NUREG-1530, Reassessment of NRCs Dollar Per 16 Person-Rem Conversion Factor Policy, issued December 1995, and Revision 1, issued 17 August 2015 (NRC, 1995b; NRC, 2017b).

18 19 The study evaluated the relationship between population densities, distribution characteristics, 20 and offsite property values near SFP sites by conducting sensitivity analyses in which the site 21 population densities and distributions were varied. The site populations, distributions, and 22 economic data for the high-, base-, median-, and low-estimate cases described above served 23 as additional input into the MACCS calculations that otherwise used values specific to the 24 reference plant. The staff also evaluated the impact on offsite property costs as a result of 25 extending offsite consequences beyond 80 kilometers (50 miles). In this case, the base-case 26 assumptions and the intermediate protective action guidelines criterion were used, as explained 27 below.

28 29 The SFP study used the emergency response model in MACCS to model doses, health effects, 30 and emergency response during the 7-day period following the start of a release during a 31 severe accident. The long-term phase, which is the period following the 7-day emergency 32 phase, was modeled for 50 years to calculate consequences from exposure of an average 33 person. The habitability criterion used in MACCS, to determine whether land is inhabitable after 34 decontamination, was 2 rem in the first year and 500 millirem (mrem) each year thereafter for 35 the base-case evaluations. This criterion was based on the U.S. Environmental Protection 36 Agencys protective action guidelines as outlined in EPA-400/R-17/001, PAG Manual:

37 Protective Action Guides and Planning Guidance for Radiological Incidents, issued 38 January 2017 (EPA, 2017). However, for habitability, some States (e.g., Pennsylvania) have 39 adopted a habitability criterion of 500 mrem annually. To account for the uncertainties in the 40 way in which States define their habitability criteria, the staff also performed sensitivity studies in 41 which the low estimate case used 500 mrem per year, while the high-estimate case used a 42 conservative 2 rem per year.

43 44 Cost-Benefit Analysis 45 46 A cost-benefit analysis informed the Commissions decision whether to expedite spent fuel 47 transfer to dry cask storage. This analysis was more expansive than that performed for the SFP 48 study, as it evaluated SFP configurations at all U.S. nuclear power plants and it incorporated 49 insights from the SFP study and other previous studies, where possible.

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-136

1 Methodology 2

3 The staff first identified the attributes that would be impacted by expedited fuel transfer and 4 performed quantitative and qualitative analyses on those attributes, including public health 5 (accident) and occupational health (routine and accident), onsite property, offsite property, 6 industry implementation and operational activities, and NRC implementation and operational 7 activities. The analysis did not include the NRCs implementation and operational activity costs; 8 this simplification is acceptable because it is consistent with the approach to maximize the 9 benefit of the alternative.

10 11 The staff determined the costs and benefits associated with each attribute for each alternative, 12 converting them into monetary values where practicable and discounting them to a net present 13 value. Specifically, the staff used a constant 7 percent discount rate as a base-case value and 14 used 3 percent as a sensitivity value to approximate the real rate of return on long-term 15 government debt, which is a proxy for the real rate of return on savings. In addition, the Office 16 of Management and Budget (OMB) Circular No. A-4, Regulatory Analysis, dated 17 September 17, 2003, suggests using a lower but positive discount rate, in addition to the 18 discount rates of 3 percent and 7 percent, if the decisionmaking will have important 19 intergenerational benefits. Therefore, for this study, the staff included a 2 percent discount rate 20 to represent the lower bound for the certainty-equivalency rate in 100 years. The staff analyzed 21 the total discounted quantitative costs and benefits for each alternative to determine whether 22 there was a positive benefit for expedited transfer. The staff also considered qualitative costs 23 and benefits in assessing whether there was a positive benefit.

24 25 The staff performed a sensitivity analysis to identify key input parameters that have the greatest 26 impact on the results. Starting with the parameters for the base case, it varied the input 27 parameters to generate low- and high-estimates that it compared with the base-case results to 28 determine the sensitivity of the results to the input parameter. The results of these analyses 29 indicated that, in addition to discount values used for present value calculations, dollar 30 per person-rem conversion factors, calculated consequences from the site, habitability criteria, 31 and seismic initiator frequency were also key input parameters that strongly affected the net 32 results. Table H-35 summarizes the base-case and sensitivity values used for the key input 33 parameters.

34 35 Table H-35 Key Input Parameters Used for Sensitivity Analyses Methodology Input Parameter Base Case Value Sensitivity Value(s)

Net Present Value (NPV) 7% NPV 2 and 3% NPV Dollar per person-rem

$2,000 $4,000 Conversion Factor Calculated Consequences from 50 miles Beyond 50 miles Site 2 rem in the first year and 500 mrem per year and 2 rem per Habitability Criteria 500 mrem each year year thereafter Bin 3: 1.65x10-5 per year Bin 3: 2.24x10-5-5.64x10-5 per year Seismic Initiator Frequencya Bin 4: 4.90x10 per year

-6 Bin 4: 7.09x10-6-2.00x10-5 per year 36 a As discussed in the SFP study, damage to the SFP and other relevant structures, systems, and components is 37 not credible for events in bins 1 and 2.

38 H-137 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 The staff made its recommendation on the implementation of each alternative based on 2 qualitative attributes, uncertainties, sensitivities, and the quantified costs and benefits taken 3 from quantitative attributes. If the quantified and qualified benefits were greater than the 4 quantified and qualified costs, then the staff recommended the alternative be implemented.

5 Otherwise, the staff recommended that the alternative not be implemented.

6 7 Cost-Benefit Analysis Results 8

9 Table H-36 summarizes the net benefits (i.e., the sum of total benefits and total costs) for each 10 SFP group. The table includes the corresponding values obtained from additional sensitivity 11 analyses in which the discount rate of 7 percent, which the NRC uses for regulatory 12 decisionmaking, was varied to 2 percent and 3 percent in accordance with the 13 recommendations in OMB Circular A-4. In addition to the conservative assumptions used to 14 generate the base-case values, low- and high-estimates are provided that combine the range of 15 expected SFP attributes to model the range of pool accidents postulated.

16 17 Table H-36 Summary of Net Benefits for Each Spent Fuel Pool Group*

SFP Low Estimate Base Case High Estimate Group (2012 million dollars) (2012 million dollars) (2012 million dollars)

No. 2% NPV 3% NPV 7% NPV 7% NPV 2% NPV 3% NPV 7% NPV 1 ($53)** ($55) ($52) ($45) $70 $54 $21 2 ($51) ($54) ($51) ($45) $86 $67 $26 3 ($42) ($36) ($17) ($12) $66 $45 $17 4 ($49) ($50) ($49) ($39) $160 $130 $74 18

Note: The values listed in COMSECY-13-0030, Enclosure 1, have been rounded to two significant figures here.

19 ** Negative values are shown using parentheses (e.g., negative $53 is displayed as ($53)).

20 21 Attributes that led to net costs for SFP Group 1 through Group 4 are industry implementation 22 and occupational health (routine) costs, with implementation costs far surpassing routine 23 occupational health costs. For Group 1, Group 2, and Group 4, these costs are dominated by 24 the additional capital costs for the dry storage containers (DSCs) and loading costs for the 25 storage systems to achieve low-density storage in the SFP above that required for the 26 regulatory baseline. Since the spent fuel stored in Group 3 SFPs is not expected to require dry 27 storage until 2038, additional costs beyond the DSC capital costs and loading costs include 28 ISFSI annual operation and maintenance costs required to establish the ISFSI and store spent 29 fuel there 15 years earlier than in the regulatory baseline.

30 31 Positive attributes (i.e., benefits and cost offsets) that offset the net costs described above are 32 public health (accident), occupational health (accident), offsite property, and onsite property.

33 For all groups, the offsite property cost offset is the largest contributor to the benefits, the 34 majority of which occur during the long-term phase. However, as Table H-37 illustrates, these 35 benefits and cost offsets do not create a positive net benefit for low-, high-, or 36 base-case-estimates with any of the discount rates applied.

37 38 The staff performed sensitivity analyses to provide additional consideration for the safety goal 39 screening evaluation. Table H-37 summarizes the results of the sensitivity analyses considering 40 the combined effects of adjusting the dollar per person-rem conversion factor from $2,000 to 41 $4,000 and of extending consequence analyses beyond 80 kilometers (50 miles) from the site.

42 NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-138

1 Table H-37 Net Benefits for Low-Density SFP Storage for Groups 1-4 from Combined 2 Sensitivity Analyses that Analyzed Consequences Beyond 80 kilometers (50 3 Miles) and Using an Adjusted Dollar per Person-Rem Conversion Factor SFP Low Estimate Base Case High Estimate Group (2012 million dollars)* (2012 million dollars)* (2012 million dollars)*

No. 2% NPV 3% NPV 7% NPV 2% NPV 3% NPV 7% NPV 2% NPV 3% NPV 7% NPV 1 ($51)** ($54) ($51) $9.5 $0.17 ($15) $880 $779 $506 2 ($48) ($51) ($49) $19 $7.7 ($12) $1,100 $916 $569 3 ($39) ($33) ($16) $32 $21 $6.8 $749 $563 $233 4 ($45) ($47) ($44) $40 $28 $5.8 $1,900 $1,600 $1,100 4

Note: the original values for this analysis listed in COMSECY-13-0030, Enclosure 1, have been rounded to two 5 significant figures.

6 ** Negative values are shown using parentheses (e.g., negative $51 is displayed as ($51)).

7 8 The sensitivity results provided in Table H-37 show that there are cases using conservative 9 assumptions for each SFP group in which the low-density spent fuel storage alternative was 10 cost-justified. However, after considering the analysis results, operating history, and limited 11 safety benefits of possible plant changes, the staff concluded that further study would be 12 unlikely to support future actions requiring expedited transfer.

13 14 Summary and Conclusion 15 16 The staff performed a regulatory analysis that included all U.S. SFPs to determine whether 17 expedited transfer of spent fuel from SFPs to dry cask storage was warranted. As part of the 18 regulatory analysis, the staff conducted a technical evaluation using insights from recently 19 completed SFPs, a safety goal screening evaluation, and a cost-benefit analysis. The results of 20 the technical evaluation of the consequences of seismic events impacting four different 21 categories of SFPs indicated that no offsite fatalities were expected to occur, similar to the 22 results obtained from the SFP study and other studies, and that the predicted long-term 23 exposure of the population, which could result in latent cancer fatalities, was low.

24 25 The safety goal screening evaluation revealed that SFP accidents are a small contributor to the 26 overall risks for public health and safety (less than 1 percent of the QHOs), and therefore any 27 reductions in risk associated with expedited transfer of spent fuel only would have a marginal 28 safety benefit. In addition, the cost-benefit analysis demonstrated that the added costs of 29 expediting transfer of spent fuel to dry cask storage were not warranted considering the 30 marginal safety benefits that would result. As part of the analysis, the staff identified attributes 31 affected by expedited transfer and analyzed them quantitatively and qualitatively, where 32 possible. When considering the discount rates combined with very conservative SFP 33 assumptions, the costs of implementing expedited transfer greatly outweighed the benefits of 34 doing so. However, the combination of high estimates for important parameters used in 35 subsequent sensitivity analyses resulted in large economic consequences, such that the 36 calculated benefits from expedited transfer of spent fuel to dry cask storage for those cases 37 outweighed the associated costs. For those cases, the staff concluded that there was only a 38 marginal safety improvement in terms of public health and safety, asserting that the 39 assumptions made in the analyses were selected in a generally conservative manner such that 40 the base case is the primary basis for the staffs recommendation.

41 42 Based on the analyses presented in COMSECY-13-0030, the staff concluded that additional 43 studies were not needed to reasonably conclude that the expedited transfer of spent fuel to dry H-139 NUREG/BR-0058, Rev. 5, App. H, Rev. 0

1 cask storage would provide only a marginal increase in the overall protection of public health 2 and safety. The staff also informed the Commission that it recommended no further regulatory 3 action for the resolution of this Tier 3 issue.

4 5 Staff Non-concurrence 6

7 In accordance with Management Directive 10.158, NRC Non-Concurrence Process, dated 8 March 14, 2014, a member of the NRC technical staff submitted a non-concurrence on 9 COMSECY-13-0030. Enclosure 2 to COMSECY-13-0030 provides documentation associated 10 with this non-concurrence.

11 12 The non-concurrence raised several issues with the detailed analyses performed in support of 13 COMSECY-13-0030, including (1) other potentially cost-beneficial approaches to improving the 14 safety of SFPs should have been evaluated, in addition to Option 2, (2) the base case analysis 15 should have used different assumptions for factors that were ultimately evaluated only as 16 sensitivity analyses (e.g., the dollar per person-rem conversation factor, the region over which 17 offsite radiological consequences are aggregated), (3) the staff should acknowledge the 18 limitations of using safety goals and QHOs that were developed for reactor accidents to 19 determine whether a proposed regulatory action pertaining to SFP safety would constitute a 20 substantial safety enhancement, and (4) the presentation of results should have provided a 21 more balanced and neutral view of the range of findings that were obtained by using the 22 high-estimate cases and sensitivity analyses.

23 24 The staff made several improvements to COMSECY-13-0030 in response to the concerns 25 raised in the non-concurrence. However, after considering the analysis results, operating 26 history, and limited safety benefits of possible plant changes, the staff ultimately concluded that 27 additional studies would be unlikely to support a requirement to expedite transfer of spent fuel 28 from SFP storage to dry cask storage to achieve a low-density SFP loading configuration.

29 30 Commissions Response to the Staffs Analysis and Recommendations 31 32 In the staff requirements memorandum for Staff Requirements Memoranda 33 (SRM)-COMSECY-13-0030, dated May 23, 2014, Staff Evaluation and Recommendation for 34 Japan Lessons-Learned Tier 3 Issue on Expedited Transfer of Spent Fuel, the Commission 35 approved the staff's recommendation that the Tier 3 Japan lessons-learned activities for 36 expedited transfer be closed, and that no further generic assessments be conducted. The 37 Commission also directed the staff to perform several other related activities for completeness 38 and closure of the Tier 3 issue, including modifying the regulatory analysis provided in 39 COMSECY-13-0030 to explain why the 1x8 configuration would not provide a substantial 40 increase in safety. The staff addressed the above issues in SECY-15-0059, Seventh 6-Month 41 Status Update on Response to Lessons Learned from Japans March 11, 2011, Great Tohoku 42 Earthquake and Subsequent Tsunami, Enclosure 3, dated April 9, 2015 (NRC, 2015e).

NUREG/BR-0058, Rev. 5, App. H, Rev. 0 H-140

ML21096A294: Difference between revisions

Latest revision as of 17:01, 19 January 2022

Contents

Text

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

Navigation menu

ML21096A294: Difference between revisions

Latest revision as of 17:01, 19 January 2022

Text

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

Navigation menu

Search