ML23263A639

DPO Case File DPO-2021-001 -- Redacted-Public

ML23263A639
Person / Time
Issue date:	09/20/2023
From:	Jennivine Rankin Office of Nuclear Reactor Regulation
To:	Office of Nuclear Reactor Regulation
References
DPO Case File DPO-2021-001
Download: ML23263A639 (1)

Text

DPO Case File for DPO-2021-001 The following pdf represents a collection of documents associated with the submittal and disposition of a differing professional opinion (DPO) from an NRC employee involving the FitzPatrick Amendment Concerning an Alternate Source Term for Calculating LOCA Accident Dose Consequences.

Management Directive (MD) 10.159, The NRC Differing Professional Opinions Program, describes the DPO Program. https://www.nrc.gov/docs/ML2312/ML23123A099.pdf The DPO Program is a formal process that allows employees and NRC contractors to have their differing views on established, mission-related issues considered by the highest level managers in their organizations, i.e., Office Directors and Regional Administrators. The process also provides managers with an independent, three-person review of the issue (one person chosen by the employee). After a decision is issued to an employee, they may appeal the decision to the Executive Director for Operations (or the Commission, for those offices that report to the Commission).

Because the disposition of a DPO represents a multi-step process, readers should view the records as a collection. In other words, reading a document in isolation will not provide the correct context for how this issue was reviewed and considered by the NRC.

It is important to note that the DPO submittal includes the personal opinions, views, and concerns by NRC employees. The NRCs evaluation of the concerns and the NRCs final position are included in the DPO Decision.

The records in this collection have been reviewed and approved for public dissemination.

Document 1: DPO Submittal Document 2: Memo Establishing DPO Panel Document 3: DPO Panel Report Document 4: DPO Decision Document 5: DPO Appeal Submittal Document 6: Statement of Views Document 7: DPO Appeal Decision

Document 1: DPO Submittal

Differing Professional Opinion Related to Approved License Amendment No. 338 for the FitzPatrick Nuclear Power Plant

1.0 BACKGROUND

By letter dated August 8, 2019 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML19220A043), as supplemented by letters dated August 27, 2019; January 16, 2020; and March 30, 2020 (ADAMS Accession Nos. ML19261A168, ML20017A052, and ML20090E279, respectively), Exelon Generation Company, LLC (Exelon or the licensee), submitted a license amendment request (LAR) for the James A. FitzPatrick Nuclear Power Plant (FitzPatrick). In part, this LAR asked to adopt the alternative source term (AST) in accordance with Title 10 of the Code of Federal Regulations (10 CFR) 50.67, Accident source term, and to revise its technical specifications (TS).

On July 21, 2020, the U.S. Nuclear Regulatory Commission (NRC) staff approved the requested changes to the license in Amendment 338 (ADAMS Accession No. ML20140A070). A copy of the NRC staffs related safety evaluation (SE) is attached to the July 21, 2020, letter notifying FitzPatrick of the Issuance of Amendment 338.

Please note that nonpublic information is contained in double brackets (( )).

2.0 DIFFERING PROFESSIONAL OPINION ISSUES This differing professional opinion (DPO) concerns two changes to the TS that adversely impact the performance of fission product barriers as described in the design and licensing basis for FitzPatrick. These barriers protect workers, public health and safety, and the environment in the event of a radiological accident at FitzPatrick. This DPO asserts that the NRC improperly approved an expansion of operating authority that allows the following:

(1) main steamline isolation valves (MSIVs) to leak at a rate approximately four times higher than was previously authorized (2) removal of requirements for the main steam leakage collection system (MSLCS),

which supplemented the isolation function of the MSIVs by stopping fission products that leak through the MSIVs after an accident from making their way to the environment 1

2.1 BACKGROUND

ON MAIN STEAMLINE ISOLATION VALVES AND THE MAIN STEAM LEAKAGE COLLECTION SYSTEM FitzPatrick is a boiling-water reactor, which operates by boiling water that is in direct contact with the reactor fuel rods in the reactor core. The steam produced is transported directly to the power turbines by means of large main steamlines. Because these steamlines bypass the reactor containment and could carry radioactivity from the reactor core to the environment, two quick-closing safety-related MSIVs were included in the original design. The MSIVs on each steamline partially isolate the containment boundary from the environment in a core damage accident. This isolation accomplishes the critical safety function of mitigating the release of fission products if an accident were to occur and radiation was released into the boiling-water reactor and its containment.

Because MSIVs (stem packing, valve bonnets, and seats) and downstream piping and components (including the power conversion system 1) cannot be made leaktight and are allowed to leak during operations (and thus during accidents), the original design approved by the NRC included an MSLCS. Its purpose was to treat allowed MSIV leakage by creating a negative pressure that collects and filters this leakage before releasing it to the environment. 2 Per Regulatory Guide (RG) 1.96, Design of Main Steam Isolation Valve Leakage Control Systems for Boiling Water Reactor Nuclear Power Plants, a design objective of these systems was to reduce and control of stem packing leakage or other direct leakage to the steam tunnel, because this leakage would escape to the turbine building and the environment via the steam tunnel.

Because MSIVs are not leaktight, acceptable leakage limits were established and incorporated into the plant design and TS in accordance with 10 CFR 50.36, Technical specifications. Licensees periodically test MSIV leakage in accordance with established surveillance requirements to ensure the leakage remains below these limits. Licensees also periodically perform TS surveillances to verify that the MSLCS is functional.

Per 10 CFR 50.36, the TS are derived from the analyses and evaluations included in the updated final safety analysis report (UFSAR). To justify the change to the FitzPatrick TS, FitzPatrick provided a revised UFSAR radiological assessment for a design-basis 1 The FitzPatrick UFSAR Section 1.6.1.4, Power Conversion System, states the following: The unit utilizes a Power Conversion System which includes a turbine-generator, a main condenser, condensate pumps, a steam jet air ejector, Turbine Sealing System, Turbine Bypass System, condensate demineralizers, Reboiler System, condensate booster pumps, reactor feed pumps, feedwater heaters, drain coolers and Condensate Storage System to produce electrical power from the steam coming from the reactor, condense the steam into water, and return the heated feedwater to the reactor. The Circulating Water System removes the heat rejected to the main condenser.

2 UFSAR, Section 9.19.2, Safety Design Bases, for the MSLCS states, in part, the following: The MSLCS is designed to collect and process leakage across the seats of the MSIVs and to collect and process stem packing leakage from the outboard containment MSIVs following a design basis LOCA.

The effluent of the MSLCS is processed by the Standby Gas Treatment System (SGTS) and is exhausted through the Stack. The negative pressure in the SGTS is sufficient to provide the required flow through the MSLCS to collect all postulated leakage. The MSLCS is designed to meet the requirements of a Seismic Class I system, as defined in Section 12.2.

2

loss-of-coolant accident (LOCA). 3 The NRC staffs approval of these changes is stated to be based primarily on 10 CFR 50.67.

The MSIVs and the MSLCS were credited in the original plant accident design to protect the health and safety of the public. In particular, the original FitzPatrick safety evaluation report (SER) (ADAMS Accession No. ML19182A200) states the following in Section 5:

Leakage through the closed main steam line isolation valves following a postulated LOCA presently relies on the low leakage characteristic of the valve [at the time, 11.5 standard cubic feet per hour (scfh)]. The acceptability of the present leakage limits and the need for an auxiliary sealing system are under study by the staff. There is nothing in the existing design which would preclude incorporation of an additional sealing feature if such is determined necessary. The applicant will continue to study developments in this area.

FitzPatricks SER, Section 5.2.3, Containment Isolation, of Supplement 1 (ADAMS Accession No. ML15205A100) states the following:

On the basis of Staff calculations of the effects of main steam line isolation valve leakage and a continuing concern by the ACRS [Advisory Committee on Reactor Safeguards] 4, PASNY [Power Authority of the State of New York] was requested to submit plans for installation of a supplementary sealing system [MSLCS]. PASNY has committed, in a letter dated January 11, 1973, to demonstrate and install an acceptable sealing system at the time of the plants first refueling outage. 5 Contrary to the above licensing basis that required an MSLCS and limited the MSIV leakage to low values (11.5 scfh), Amendment 338 fails to meet the regulatory requirements in 10 CFR 50.67 and 10 CFR 50.36 necessary to remove the MSLCS from the TS and increase the allowed MSIV leakage.

This DPO documents how the NRCs FitzPatrick SE does not comply with NRC regulations. It also documents how the NRC did not require FitzPatrick to comply with NRC regulations and its implementing guidance when it issued Amendment 338. The NRC did not follow its own internal processes used to ensure the regulations are met.

3 UFSAR, Section 9.19.1, Safety Objective, for the MSLCS states, in part, the following: The safety objective of the Main Steam Leakage Collection System (MSLCS) is to collect and process leakage past the main steam isolation valves (MSIVs) following a Loss-Of-Coolant Accident (LOCA) so that resultant exposures are maintained below the values specified by 10 CFR 100.

4 In its letter on the construction permit review of the FitzPatrick plant (January 27, 1970), the Advisory Committee on Reactor Safeguards noted that additional features to control MSIV leakage should be considered.

5 It is important to note that when these NRC staff calculations were performed, the regulation used to evaluate the FitzPatricks MSIVs and the MSLCS design was 10 CFR Part 100, Reactor site criteria.

Unlike 10 CFR 50.67, 10 CFR Part 100 does not specify that the approval of the design must be based only on the applicants analyses.

3

Summaries of specific issues are provided below, and the basis and details supporting these issues are provided in Enclosure 1.

Summary of Issue 1 FitzPatricks LAR, dated August 8, 2019, requested NRC approval for adopting an AST in accordance with 10 CFR 50.67. FitzPatrick stated that its LOCA analysis for showing compliance with 10 CFR 50.67 followed the guidance in RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Revision 0, issued July 2000 (ADAMS Accession No. ML003716792), and NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants LWR [Light-Water Reactor] Edition, and (SRP) Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Revision 0, issued July 2000 (ADAMS Accession No. ML003734190).

In 10 CFR 50.67, the NRC requires the staff to issue the requested amendment only if the applicants analysis demonstrates with reasonable assurance that the dose criteria in 10 CFR 50.67 are met. Contrary to the 10 CFR 50.67 regulation and the guidance the licensee used to demonstrate compliance with that regulation, the NRC staff issued the amendment based on the staffs independent analysis, rather than using the licensees analysis. The NRCs analysis used risk and engineering insights, such as credit for transport of MSIV leakage to the condenser, and holdup and deposition of this leakage in the condenser. The FitzPatrick analysis approved by the NRC in Amendment 338 (which became the analysis of record) does not credit transport of the MSIV leakage to the condenser or credit holdup and deposition of the MSIV leakage in the condenser.

Summary of Issue 2 The LOCA analysis for Amendment 338 proposed by FitzPatrick and now accepted by the NRC is fundamentally flawed. It is inconsistent with the RG 1.183 and RG 1.194 methods that the licensee stated it used to show compliance with 10 CFR 50.67. For example, the FitzPatrick LOCA analysis grossly overestimated the amount of the removal of radioactivity due to sprays and natural deposition in the steamlines. Therefore, the analysis did not demonstrate compliance with 10 CFR 50.67. When these fundamental flaws are addressed, the FitzPatrick LOCA doses do not meet the 10 CFR 50.67 dose criterion necessary for approving the TS changes approved by the NRC staff.

Summary of Issue 3 The NRCs SE for Amendment 338 is fundamentally flawed and inconsistent with regulatory requirements in 10 CFR 50.36. In direct conflict with 10 CFR 50.36(c)(2)(ii),

Criterion 3; and RG 1.183, the NRC staff confirmatory analysis credits the plants power conversion system piping and valves when these structures, systems, and components (SSCs) do not have established TS limiting conditions for operation (LCOs). 10 CFR 50.36 requires the incorporation of these LCOs to ensure that the facility SSCs have and will maintain the assumed integrity in the LOCA analysis.

4

Summary of Issue 4 The NRCs approval of the FitzPatricks LOCA analysis, based upon an NRC staff analysis that credited portions of the power conversion system not credited in the FitzPatrick analysis, does not ensure that a clear, logical, and consistent design basis exists to support evaluations of future modifications, SEs and NRC inspections for FitzPatrick.

Summary of Issue 5 Even if it was appropriate for the staff to consider their risk and engineering insights to determine the licensees compliance with 10 CFR 50.67, those risk and engineering insights would not compensate for the uncertainties (and errors) in the FitzPatrick LOCA analysis and provide reasonable assurance that 10 CFR 50.67 is met.

5

Enclosure 1 Summary of Differing Professional Opinion Issues and Details Supporting These Issues Summary of Issue 1 FitzPatricks LAR, dated August 8, 2019, requested NRC approval for adopting an AST in accordance with 10 CFR 50.67. FitzPatrick stated that its LOCA analysis for showing compliance with 10 CFR 50.67 followed the guidance in RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Revision 0, issued July 2000 (ADAMS Accession No. ML003716792), and NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants LWR [Light-Water Reactor] Edition, and (SRP) Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Revision 0, issued July 2000 (ADAMS Accession No. ML003734190).

In 10 CFR 50.67, the NRC requires the staff to issue the requested amendment only if the applicants analysis demonstrates with reasonable assurance that the dose criteria in 10 CFR 50.67 are met. Contrary to the 10 CFR 50.67 regulation and the guidance the licensee used to demonstrate compliance with that regulation, the NRC staff issued the amendment based on the staffs independent analysis, rather than using the licensees analysis. The NRCs analysis used risk and engineering insights, such as credit for transport of MSIV leakage to the condenser, and holdup and deposition of this leakage in the condenser. The FitzPatrick analysis approved by the NRC in Amendment 338 (which became the analysis of record) does not credit transport of the MSIV leakage to the condenser or credit holdup and deposition of the MSIV leakage in the condenser.

Supporting Details for Issue 1 These details are broken into four major sections. Section 1 provides the applicable regulations, guidance, and background pertaining to Issue 1. Section 2 discusses FitzPatricks loss-of-coolant accident (LOCA) analysis provided in the LAR to demonstrate compliance with 10 CFR 50.67. Section 3 provides details on the NRC staff risk and engineering insights. Section 4 shows how these insights are not part of the licensees analysis required by the 10 CFR 50.67 regulation in order to grant the FitzPatrick LAR.

1

Section 1 Regulatory Background on the 10 CFR 50.67 Regulation and the Methods Used by FitzPatrick To Demonstrate Compliance with 10 CFR 50.67 (Regulatory Guide 1.183 and Standard Review Plan Section 15.0.1)

The NRC, through the Atomic Energy Act, ensures that the primary responsibility for the safety of a nuclear installation rests with the licensee. The NRCs regulatory programs are based on the premise that the safety of commercial nuclear power reactor operations is the responsibility of NRC licensees. The Commission has noted that the purpose of having regulations is to flesh out the adequate protection standard and that compliance with such regulations and guidance may be presumed to assure adequate protection at a minimum. 1 Licensees proposing to make changes to nuclear facilities must demonstrate that adequate protection of public health and safety is maintained after making those changes.

Licensees who propose to incorporate 10 CFR 50.67 into their license must demonstrate that their facility, with any proposed modifications, will not result in accident conditions exceeding the criteria in 10 CFR 50.67. In addition, they must continue to comply all applicable regulations in their licensing and design bases.

Per the August 18, 2019, LAR, FitzPatrick created a LOCA analysis to show compliance with 10 CFR 50.67 following the guidance in RG 1.183, Revision 0, issued July 2000, and SRP Section 15.0.1, Revision 0, issued July 2000.

Below is a brief discussion of 10 CFR 50.67, RG 1.183, and SRP 15.0.1, as they pertain to DPO Issue 1.

In 10 CFR 50.67, the NRC requires the following:

(2) The NRC may issue an amendment only if the applicants analysis 2 demonstrates with reasonable assurance that:

(i) An individual located at any point on the boundary of the exclusion area for any 2-hour period following the onset of the postulated fission product release, would not receive a radiation dose in excess of 0.25 Sv (25 rem) total effective dose equivalent (TEDE),

1 See, for example, Revision of Backfitting Process for Power Reactors, Federal Register, Volume 53, June 6, 1988, pp. 20603, 20606.

2.

It is typically an unstated requirement that staff approval is based on an applicants analysis, which becomes their licensing basis, and that any staff analyses are performed solely for the purpose of confirming the applicants analyses. However, 10 CFR 50.67 explicitly requires the use of the applicants analysis as the demonstration of reasonable assurance and hence the basis of the license. Therefore, the staff review must rely on the applicants analysis as the basis for approval; not the staff analyses.

2

(ii) An individual located at any point on the outer boundary of the low population zone, who is exposed to the radioactive cloud resulting from the postulated fission product release (during the entire period of its passage),

would not receive a radiation dose in excess of 0.25 Sv (25 rem) total effective dose equivalent (TEDE), and (iii) Adequate radiation protection is provided to permit access to and occupancy of the control room under accident conditions without personnel receiving radiation exposures in excess of 0.05 Sv (5 rem) total effective dose equivalent (TEDE) for the duration of the accident.

RG 1.183 provides an approved methodology for analyzing the radiological consequences of several design-basis accidents (DBAs) (including the DBA LOCA) to show compliance with 10 CFR 50.67. RG 1.183 provides guidance to licensees on acceptable assumptions for modeling MSIV leakage in boiling-water reactors (BWRs).

RG 1.183, Regulatory Position 1.5, in part, states the following:

The NRC staffs finding that the amendment may be approved must be based on the licensees analyses, since it is these analyses that will become part of the design basis of the facility.

SRP 15.0.1 provides guidance to the staff for the review of AST amendment requests. It states that the NRC reviewer should evaluate the proposed change against the guidance in RG 1.183 and the following:

Independent calculations should be performed as necessary to conclude, with reasonable assurance, that the applicants analyses are acceptable.

The staffs approval of the application is to be based on the licensees docketed information.

It also states the following:

The reviewer should evaluate the AST 3 proposed by the licensee against the guidance in RG-1.183. Differences between the licensees proposal and the guidance should be resolved with the licensee.

3 Per SRP Section 15.0.1, implementation of an AST within the SRP includes any associated plant modifications. These modifications may be to systems or procedures identified in the final safety analysis report or changes to the technical specifications (TS), as is the case for FitzPatrick.

3

Section 2 FitzPatricks Analysis To Show Compliance with 10 CFR 50.67 In the LAR, FitzPatrick provided a revised LOCA analysis to justify their assertion that they comply with 10 CFR 50.67. This LOCA analysis is provided in calculation JAF-CALC 00005. 4 In its March 30, 2020, supplement, FitzPatrick acknowledges that calculation JAF-CALC-19-00005 does not model or credit radioactivity holdup or deposition in the condenser. The supplement, in part, states, A further conservatism that is not currently modeled in JAF-CALC-19-00005 is the holdup and aerosol deposition provided by the condenser.

JAF-CALC-19-00005, Revision 0, states the following:

Since MSLCS [main steam leakage collection system] is no longer credited, no ESFs [engineered safety features] are assumed to be available to collect or treat MSIV leakage. Releases are assumed to be from the Seismic Class I Turbine Stop Valves (TSVs) without credit for holdup or dilution in the condenser or turbine building. The release is treated as a ground level release for offsite dose assessment.

The NRC staff acknowledges that JAF-CALC-19-005 does not credit the condenser in Section 3.1.1.4.3.3 of the NRC staffs SE for FitzPatrick dated July 21, 2020. Section 3.1.1.4.3.3, in part, states the following:

In its letter dated March 30, 2020, the licensee stated that aerosol holdup and deposition provided by the condenser is not modeled in JAF-CALC-19-00005 for FitzPatrick.

Note that the March 30, 2020, supplement does provide a sensitivity analysis showing FitzPatricks estimates of the radiological consequences if the condenser were to be credited, but this analysis was not provided as the proposed analysis of record nor did the NRC staff use it in the SE as discussed below.

Under the heading Condenser Holdup and Deposition Sensitivity, the SE states the following:

In this sensitivity, the leakage is assumed to travel to the condenser through the drain lines from the main steam line piping between the MSIVs.

This conservatively neglects any holdup and deposition in the outboard main steam line piping. Modeling the release to the condenser from the piping between the MSIV is consistent with other plants in the Exelon fleet (e.g. LaSalle and Limerick). Operating experience associated with the North 4 JAF-CALC-19-00005, Revision 0, was transmitted to the NRC as Attachment 6 to the LAR submittal dated August 8, 2019. The March 30, 2020, supplement to the LAR described a revision to the calculation and provided the changes made but did not show the full calculation.

4

Anna earthquake and post-Fukushima evaluations have shown that components and piping systems typically used in this release path are sufficiently rugged to ensure they are capable of performing some level of radioactivity removal during and following a safe shutdown earthquake (SSE). Thus, it is reasonable to assume that the condenser pathway could be made available for mitigating the consequences of MSIV leakage.

However, the NRC staff did not use the above discussed sensitivity analysis to determine reasonable assurance of compliance with 10 CFR 50.67 for the FitzPatrick LAR. Per the March 30, 2020, supplement, the sensitivity analysis that assumed credit for the condenser also assumed a 2-micron aerosol mass median diameter and geometric standard deviation of 2.0 particle size in the 20-group method. The NRC staffs SE clearly states that these sensitivity analyses were not used in the licensees proposed analysis of record nor did the NRC staff use them to show compliance with 10 CFR 50.67. The NRC staffs July 21, 2020, SE states the following:

The licensee assumed a 2-micron aerosol mass median diameter and geometric standard deviation of 2.0 particle size in the 20-group method to recalculate the aerosol removal rates in its sensitivity analysis. The licensee claimed that this modeling assumption is conservative and results in a much smaller gravitational settling and spray removal rate. The NRC staff did not review and evaluate this assumption because: (1) no basis was provided for the assumption, (2) this assumption was not used in the licensees proposed analysis of record, and (3) it was not used by the NRC staff to determine reasonable assurance for complying with 10 CFR 50.67 for this particular LAR.

Section 3.1.1.4.4 of the NRC staffs July 21, 2020, SE also states that the analysis used to show compliance with the regulations is not the sensitivity analysis performed by the licensee. 5 The now approved analysis of record is the JAF-CALC-19-00005 analysis discussed in the response to Request for Additional Information (RAI) ARCB-RAI-1B (ADAMS Accession No. ML20090E279). The JAF-CALC-19-00005 calculation and the RAI response do not credit transport to or holdup and deposition in the condenser.

Therefore, based on the above information, it is clear that the applicants analysis used to show compliance with 10 CFR 50.67 (and which became the now approved analysis of record) does not credit transport to and holdup and deposition in the condenser. Contrary to this fact and the 10 CFR 50.67 requirement that the NRC staff use the applicants analysis to demonstrate compliance with 10 CFR 50.67, the staff used an analysis that included NRC staff risk and engineering insights to demonstrate compliance with 10 CFR 50.67. These insights (which credit transport to and holdup and deposition in the 5

Section 3.1.1.4.4 states: The NRC staff notes that the licensees base case was produced for the purpose of conducting a sensitivity analysis and is not intended to replace the accident analysis of record, which is the revised analysis discussed in the letter dated March 30, 2020, in response to the RAI concerning obstructions (ARCB-RAI-1B).

5

condenser) are discussed below.

Section 3 NRC Staff Risk and Engineering Insights Discussed in the Safety Evaluation To aid the discussion that follows, this section cites parts of the NRCs SE in regard to the NRC staffs consideration of risk and engineering insights. Portions of the quotations have been underlined to highlight areas considered in the Section 4 discussion.

Section 3.1.1.4.4, NRC Staff Evaluation of Licensees Sensitivity Analysis, of the July 21, 2020, SE states the following:

However, the staff also considers it reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage. The NRC staffs consideration of risk and engineering insights is discussed in Section 3.5 of this SE.

Section 3.5, NRC Staff Risk and Engineering Insights, of the July 21, 2020, SE states the following:

The LAR was not submitted as a formal risk-informed submittal with probabilistic risk assessment information in accordance with the guidance of RG 1.174, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis (ADAMS Accession No. ML17317A256). Thus, the NRC staffs findings are primarily based on traditional deterministic review approaches.

In the SRM to SECY-19-0036, the Commission directed the staff to apply risk-informed principles in any licensing review or other regulatory decision when strict, prescriptive application of deterministic criteria is unnecessary to provide for reasonable assurance of adequate protection of public health and safety. Risk-informed principles are consistent with the Commission direction in the SRM for SECY-19-0036. Since the application is not a fully risk-informed submittal (with probabilistic risk information), the staff does not apply risk as the basis for acceptance of a request; however, the following risk and engineering insights inform the technical review by supporting the deterministic safety conclusions and enhance the technical reviewers confidence in their technical evaluations.

As described in Section 3.1.1.4.3.3, the licensee stated that aerosol holdup and deposition provided by the condenser is not modeled in JAF-CALC-19-00005 and that depending on the event scenario, multiple pathways could exist to route activity to the condenser, including the drain lines and the turbine itself. The licensee concluded that it is reasonable to 6

assume that the condenser pathway could be made available for mitigating the consequences of MSIV leakage.

The NRC staff performed an independent assessment evaluating the capability of the (PCS) and main condenser to serve as a holdup volume for MSIV leakage. The staff evaluated the seismic capacity of the SSCs

[structures, systems, and components] in the PCS, including the main steam piping, equalization header, and main condenser, to assess whether they would be available to provide a holdup volume for fission products following an SSE. The staff used engineering information such as operations and design knowledge, as well as probabilistic and risk information to complete the evaluation. The staff also leveraged more recent relevant operating experience such as that obtained from the March 11, 2011, magnitude 9.0 Great East Japan Earthquake that caused the accident at the Fukushima Dai-ichi Nuclear Power Plant and the August 23, 2011, magnitude 5.8 earthquake near the North Anna Power Station. The staffs independent assessment found it is reasonable to conclude that the SSCs in the PCS would be available following an SSE and that the likelihood of them being unavailable to serve as a volume for holdup and retention is very low.

The assessment provides an insight when addressing uncertainties in the calculation of the dose consequences of MSIV leakage. Specifically, the staff recognizes that there is a high probability that doses will be significantly lower than those estimated using deterministic methods that do not credit holdup and retention of the MSIV leakage within the PCS.

Based on the available information and assessments using conservatively biased assumptions about the seismic capacity of the SSCs in the realistic pathway, the staff determined that there is high confidence that the MSLs

[main steam lines] and the PCS will be available for fission product dilution, holdup, and retention, especially at the seismic accelerations at a plants design-basis SSE. Conservatisms and risk insights result in additional safety margin. In addition, as mentioned in the statements of consideration for 10 CFR 50.67, defense in depth is addressed using a DBA in the deterministic dose calculation. Therefore, consistent with the statements of consideration for 10 CFR 50.67, the principles of risk-informed decisionmaking, and the Commission direction staff has determined these risk and engineering insights support its reasonable assurance finding based on its deterministic review.

Section 3.6, NRC Staff Conclusions on Alternative Source Term, of the July 21, 2020, SE states the following:

The NRC staff concludes there is reasonable assurance supported by risk and engineering insights, that the licensees estimates for the EAB 7

[exclusion area boundary], LPZ [low population zone], and CR [control room] doses comply with the cited acceptance criteria.

Section 4 Risk and Engineering Insights Used in the NRC Staffs Safety Evaluation That Are Not in the FitzPatrick LOCA Analysis Neither the LAR nor its supplements explicitly provide an analysis (proposed to become the new analysis of record) that discusses the risk or engineering insights used by the NRC staff in its SE. Support for this conclusion is provided below using text from the NRC staffs SE Sections 3.1.1.4.4, 3.5, and 3.6.

Section 3.1.1.4.4 of the NRC staffs SE states that it is reasonable to include the probability of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell spray in calculating the dose consequences of MSIV leakage and refers the reader to Section 3.5 of the SE. Section 3.5 makes statements that provide insights into the NRC staffs review and approval of the LAR. Section 3.5 discusses how the LAR was not submitted as a risk-informed submittal and that the NRC staffs findings are primarily based on a traditional deterministic review. The choice of the language in the SE that the NRC staff findings are primarily based, rather than are based, is important because it means that something other than traditional methods was used to approve the amendment. ((Also, a review of the RPS-Licensing system shows that the risk reviewers who incorporated these risk and engineering insights billed a similar number of hours as the deterministic reviewer (214 vs. 285 hours0.0033 days <br />0.0792 hours <br />4.712302e-4 weeks <br />1.084425e-4 months <br />), indicating that the risk and engineering insights were important and significant to the approval of this LAR.))

Section 3.5 then describes how the NRC staff performed an independent assessment evaluating the capability of the power conversion system and main condenser to serve as a holdup volume for MSIV leakage. Section 3.5 also states that using the NRC staffs assessment (not the licensees analysis as required by 10 CFR 50.67), the staff found it reasonable to conclude the SSCs in the power conversion system would be available following an SSE. Furthermore, the SE describes how the NRC staffs assessment provided insights into the calculation of dose consequences of MSIV leakage, and the staff believes there is a high probability that the doses will be significantly lower than those estimated using deterministic methods.

Section 3.5 also states that the risk and engineering insights inform the technical review by supporting the deterministic safety conclusions and enhance the technical reviewers confidence in their technical evaluations. Similarly, Section 3.6 confirms that the NRCs conclusion of reasonable assurance is supported by the NRCs risk and engineering insights. Because reasonable assurance was reached by informing and supporting the deterministic methods with risk and engineering insights, it means that the NRC staff needed something other than the FitzPatrick deterministic analysis to determine if there was reasonable assurance that the dose criteria in 10 CFR 50.67 are met and approve the amendment.

8

The observation that reasonable assurance could not be reached without the NRCs risk and engineering insights is supported by the following two facts. First, FitzPatricks LOCA analysis is based on erroneous and nonconservative methods and is inconsistent with regulatory guidance used to show compliance with 10 CFR 50.67 (please see Issue 2 below for specific details that support this statement). Because the LOCA analysis is fundamentally flawed, the NRC staff could not rely on the FitzPatrick analysis alone to make its safety finding. Secondly, if by using a deterministic approach the FitzPatrick analysis could be shown to comply with the 10 CFR 50.67 dose criteria, then the ((214))

hours 6 spent on the risk portion (billed by the LIC-206 team 7 that created these risk and engineering insights) of the safety evaluation would be unnecessary. FitzPatrick would have been inappropriately charged (((estimated to be approximately $60,000, based upon

$279/hr))) for work that was not needed to show compliance with 10 CFR 50.67 and to approve the amendment. 8

((In addition to these facts internal NRC documents further support that the NRC staffs determination of reasonable assurance that 10 CFR 50.67 is met, is based, in part, upon the NRC staffs analysis that included the risk insights (not contained in the now approved FitzPatrick analysis of record). During an Office of Nuclear Regulatory Regulation All Hands Meeting on December 7, 2020, a slide summarizing the staffs review of four MSIV LARs was provided. One of the reviews summarized is the NRCs review for FitzPatricks Amendment 338. The summary slide stated that: Conclusions of reasonable assurance were based on a combination of deterministic methodology and risk insights. Also, an Interim Staff Guidance (ISG) is being developed by three of the authors of the FitzPatrick SE. The ISG is based upon how the NRC staff reviewed four LARs including the FitzPatrick LOCA analysis (as described in an email dated December 16, 2020). A second 6 The ((214)) hours cited is an estimate based upon a review of the RPS-Licensing system. It is likely a gross underestimate of the total effort to create these insights. In addition to the FitzPatrick SE, a technical paper was written that provides more details supporting the staffs insights. This paper is not referenced in the SE because it was not finalized. The paper was authored by two senior level engineers and a manager. A draft of the paper provided in a February 7, 2020 email mentioned notable contributions from five other individuals (see the Acknowledgements on page 17). Many other staff were involved in meetings and reviewed the paper. It was revised several times. ((During a December 3, 2020 internal meeting, an author of the FitzPatrick SE described the time spent on the technical paper as being thousands and thousands of hours. While this is hopefully an exaggeration, the time spent was likely a large number of hours.)) The numerous hours spent in developing the technical paper do not appear to be reflected in the ((214-hour)) value cited.

7 LIC-206, Integrated Risk-Informed Decision-Making for Licensing Reviews, dated June 26, 2020 (ADAMS Accession No. ML19263A645), was used by a team of reviewers that developed the risk and engineering insights.

8 DPO issue 1 does not assert that waste, fraud or abuse occurred. This DPO issue is based, in part, upon the belief that the risk and engineering insights were necessary to making a regulatory finding that 10 CFR 50.67 is met and that the NRC staffs finding of reasonable assurance was based, in part, on these insights. ((If during the process of reviewing this DPO it is determined that the risk and engineering insights were not necessary or that the approval was not based (in any way) upon these insights, then the Inspector General (IG) should be notified to ask them to investigate why the NRC staff charged the licensee for risk and engineering insights and expended extremely large amounts of staff hours on a technical report that are not necessary to make its regulatory finding of compliance with 10 CFR 50.67.))

9

email on December 16, 2020 transmitted the draft ISG. The ISG, based upon what was done for the FitzPatrick SE, proposes SE language for all future MSIV leakage reviews.

Per the draft ISG, these SEs would state that: The NRC staff concludes with reasonable assurance, based in part on the risk and engineering insights to compensate for uncertainties in the evaluation of the dose consequences from the MSIV release pathway, that the licensees dose estimates will comply with the acceptance criteria.))

It is once again important to note that FitzPatrick did not propose these risk and engineering insights in its LAR. They were created by an NRC staff analysis. A search of the original LAR shows that it contains the word risk a few times, but not in the context of showing compliance with 10 CFR 50.67. The LAR supplement does not contain the word risk. The original LAR shows the word insight(s) twice, but not in the context of showing compliance with 10 CFR 50.67. The supplement does not contain the word insights.

The licensee did not provide an analysis that included the probability of a pathway to the condenser to offset uncertainties in aerosol removal. The now approved analysis of record used to show compliance with 10 CFR 50.67 does not credit transport to the condenser or deposition and holdup in the condenser. Therefore, based on the 10 CFR 50.67 requirement to approve the LAR on the basis of the licensees analysis, the NRC staff should not use the probability of a pathway to the condenser to confirm the licensees compliance with 10 CFR 50.67.

However, similar to Section 3.5 of the staffs SE, the licensees March 30, 2020, supplement states the following:

A further conservatism that is not currently modeled in JAF-CALC-19-00005 is the holdup and aerosol deposition provided by the condenser. Depending on the event scenario, multiple pathways could exist to route activity to the condenser including the drain lines and the turbine itself.

In this sensitivity, the leakage is assumed to travel to the condenser through the drain lines from the main steam line piping between the MSIVs.

This conservatively neglects any holdup and deposition in the outboard main steam line piping. Modeling the release to the condenser from the piping between the MSIV is consistent with other plants in the Exelon fleet (e.g. LaSalle and Limerick). Operating experience associated with the North Anna earthquake and post-Fukushima evaluations have shown that components and piping systems typically used in this release path are sufficiently rugged to ensure they are capable of performing some level of radioactivity removal during and following a safe shutdown earthquake (SSE). Thus, it is reasonable to assume that the condenser could be made available for mitigating the consequences of MSIV leakage.

Note that as previously mentioned, the licensees LAR did not credit the holdup and deposition of aerosols in the condenser in its proposed and now accepted design-basis analysis of record (JAF-CALC-19-00005) used to show compliance with 10 CFR 50.67.

10

The above LAR only briefly mentioned the operating experience discussed in the NRC staffs engineering and risk insights, and this information was not credited in the FitzPatrick analysis. The docketed information provided by the licensee concludes that it is reasonable to assume the condenser could be used but not that the licensee will use these pathways to mitigate MSIV leakage after an accident. Nowhere in the docketed information does the licensee provide under oath and affirmation that (1) it will have a pathway available to the condenser, (2) procedures will use this pathway to reduce the main steamline leakage radioactivity, and (3) these flow paths to the condenser will enable any MSIV leakage to flow downstream of the MSIVs so that the deposition credited in the downstream main steamline piping in the FitzPatrick LOCA is consistent with the limiting pathway for MSIV leakage after an accident.

Based upon the above discussion, it is evident that the NRC staff was uncomfortable with the uncertainties (errors) in the licensees analysis and that the NRC staffs risk and engineering insights were important and necessary to the approval of the amendment. By considering credit for transport of MSIV leakage to the condenser, and holdup and deposition of this leakage in the condenser, the NRC staff considered credit for SSCs not credited in the licensees design-basis LOCA analysis and not required to be used in the event of an accident. Therefore, in direct conflict with the requirements of the 10 CFR 50.67 regulation the NRC staff issued the amendment based, in part, on the NRC staffs analyses that considered risk and engineering insights and not by using the applicants analysis to demonstrate with reasonable assurance that the criteria in 10 CFR 50.67 are met.

11

Summary of Issue 2 The LOCA analysis for Amendment 338 proposed by FitzPatrick and now accepted by the NRC is fundamentally flawed. It is inconsistent with the RG 1.183 and RG 1.194 methods that the licensee stated it used to show compliance with 10 CFR 50.67. For example, the FitzPatrick LOCA analysis grossly overestimated the amount of the removal of radioactivity due to sprays and natural deposition in the steamlines. Therefore, the analysis did not demonstrate compliance with 10 CFR 50.67. When these fundamental flaws are addressed, the FitzPatrick LOCA doses do not meet the 10 CFR 50.67 dose criterion necessary for approving the TS changes approved by the NRC staff.

Supporting Details for Issue 2 The details supporting Issue 2 are broken into three major sections. Section 1 provides the applicable regulatory guidance pertaining to Issue 2. Section 2 provides details to show how FitzPatrick did not comply with the regulatory guidance (RG 1.183 and RG 1.194) it used to show compliance with 10 CFR 50.67. Section 3 summarizes the impact on the dose analysis when regulatory guidance is not followed.

Section 1 Regulatory Guidance Applicable to Issue 2RG 1.183 and RG 1.194 In Attachment 2 of the August 8, 2019 LAR is a table entitled Regulatory Guide 1.183 Conformance Matrix. In this table, the licensee stated that its LOCA analysis conforms with RG-1.183, Regulatory Position 5.1.2 (see page 48), and RG-1.183, Appendix A, Regulatory Positions 3.1, 3.3, 6.3, 6.4, and 6.5 (see pages 50, 51, 60, and 61). In addition, Attachment 1 of the August 8, 2019 LAR it states that the new atmospheric dispersion factors (/Q) analysis complies with the methodology of RG 1.194 (see page 30). Below is a summary of the RG 1.183 and RG 1.194 regulatory positions as they pertain to DPO Issue 2.

RG 1.183, Regulatory Position 5.1.2, states:

Credit may be taken for accident mitigation features that are classified as safety-related, are required to be operable by technical specifications, are powered by emergency power sources, and are either automatically actuated or, in limited cases, have actuation requirements explicitly addressed in emergency operating procedures. The single active component failure that results in the most limiting radiological consequences should be assumed. Assumptions regarding the occurrence and timing of a loss of offsite power should be selected with the objective of maximizing the postulated radiological consequences.

12

Appendix A of RG 1.183, Regulatory Position 3.1, in part, states:

The radioactivity released from the fuel should be assumed to mix instantaneously and homogeneously throughout the free air volume of the primary containment in PWRs or the drywell in BWRs as it is released.

Appendix A to RG 1.183, Regulatory Position 3.3, in part, states the following:

Reduction in airborne radioactivity in the containment by containment spray systems that have been designed and are maintained in accordance with Chapter 6.5.2 of the SRP (Ref. A-1) may be credited. Acceptable models for the removal of iodine and aerosols are described in Chapter 6.5.2 of the SRP.

Appendix A to RG 1.183, Regulatory Position 6.3, states the following:

Reduction of the amount of released radioactivity by deposition and plateout on steam system piping upstream of the outboard MSIVs may be credited, but the amount of reduction in concentration allowed will be evaluated on an individual case basis.

Appendix A to RG 1.183, Regulatory Position 6.4, states the following:

In the absence of collection and treatment of releases by ESFs such as the MSIV leakage control system, or as described in paragraph 6.5 below, the MSIV leakage should be assumed to be released to the environment as an unprocessed, ground-level release. Holdup and dilution in the turbine building should not be assumed.

Appendix A to RG 1.183, Regulatory Position 6.5, states the following:

A reduction in MSIV releases that is due to holdup and deposition in main steam piping downstream of the MSIVs and in the main condenser, including the treatment of air ejector effluent by offgas systems, may be credited if the components and piping systems used in the release path are capable of performing their safety function during and following a safe shutdown earthquake (SSE). The amount of reduction allowed will be evaluated on an individual case basis. References A-9 and A-10 provide guidance on acceptable models.

RG 1.194, Regulatory Position 2.0, in part, states the following:

Selection of conservative, bounding source-to-receptor combinations and less detailed site parameters for the /Q evaluation may be sufficient to establish compliance with regulatory guidelines.

13

Section 2 How the FitzPatrick LOCA Does Not Conform to the Regulatory Guidance Used To Meet 10 CFR 50.67 This section and its subsections describe in more detail why the FitzPatrick LOCA analysis does not conform with the regulatory guidance methods FitzPatrick stated they used to show compliance with 10 CFR 50.67. Section 2 is broken into subsections to describe each issue in the FitzPatrick LOCA analysis.

Section 2.1 Lack of Conformance to RG 1.183, Appendix A, Regulatory Position 3.1 and Statements in the LAR Appendix A, RG 1.183, Regulatory Position 3.1, in part, states:

The radioactivity released from the fuel should be assumed to mix instantaneously and homogeneously throughout the free air volume of the primary containment in PWRs or the drywell in BWRs as it is released.

Per page 88 of calculation JAF-CALC-19-00005, Revision 0, provided as Attachment 6 to the August 8, 2019 LAR, it states that the FitzPatrick LOCA calculation conforms to Regulatory Position 3.1. 9 The initiating event for the design basis LOCA radiological analysis is that the reactor coolant system piping breaks. Note that because of the hypothesized break, the drywell free volume now includes the reactor pressure vessel free volume (for example the steam dome) and any modeled main steamline piping up to the MSIV containment isolation valves. 10 So, if Regulatory Positions 3.1 was followed, the radioactivity released from the fuel would instantaneously and homogeneously be mixed through the drywell, steam dome and main steamline piping up to the MSIV containment isolation valves.

Contrary to RG 1.183, Regulatory Position 3.1 and FitzPatricks statement that they conformed to this regulatory position, FitzPatrick did not assume that the radioactive release from the fuel was instantaneously and homogenously mixed throughout all the free air volume in the primary containment (which now includes the main steamline piping up to the MSIV containment isolation valves). The main steamline piping up to the MSIV containment isolation values was excluded as discussed below.

Page 75 of calculation JAF-CALC-19-00005, Revision 0, Figure 3 provides the model used by FitzPatrick to model the MSIV leakage and is provided below. Figure 3 shows the 9

Calculation JAF-CALC-19-00005 was revised as a result of an NRC staff RAI, however, the licensee did not provide information stating that they no longer conformed with Regulatory Position 3.1.

10 Attachment 1 to the August 8, 2019, LAR (Section 3.11.9 on page 20) describes how the MSIVs are functionally part of the primary containment boundary. It also describes how mixing occurs between the reactor pressure vessel (steam dome) and the rest of the containment via a hypothesized piping break.

14

volumes upstream of the MSIV to be modeled as separate volumes (see volumes V1 and V3 on Figure 3) from the assumed source of the radioactivity (labelled as the reactor pressure vessel). Page 28, of calculation JAF-CALC-19-00005, Revision 0 provides the flow rate from the reactor pressure vessel to these volumes. If Regulatory Position 3.1 were followed, the radioactivity would have been instantaneously and homogenously mixed into volumes V1 and V3 rather than slowly released to these volumes. Page 76 of calculation JAF-CALC-19-00005, Revision 0, Figure 4 provides an even more simplistic representation of the model used by FitzPatrick to model the MSIV leakage. This model shows how Volumes 1 and 3 are not included in the drywell/containment volume.

Figure 3 from Calculation JAF-CALC-19-00005, Revision 0 Table 5.2-1, Input Parameters of calculation JAF-CALC-19-00005, Revision 0 (page 18) states that containment elemental iodine and particulate (natural) deposition/plateout is not credited. 11,12 Contrary to these statements that no containment elemental or particulate deposition is credited, the FitzPatrick model does credit deposition in volumes V1 and V3 that are part of the primary containment. Table 3 of calculation JAF-CALC 11 Removal of elemental and particulate radioactivity due to containment sprays is credited, natural deposition is not.

12 Calculation JAF-CALC-19-00005 was revised as a result of an NRC staff RAI, however, the licensee did not provide information stating that they now credited containment elemental and particulate deposition.

15

00005, Revision 0 (page 48) shows the settling surfaces areas and volumes credited for particulate deposition up to the MSIVs and Table 4A-1 of same calculation (page 52) provides the particulate (aerosol) deposition credited in volumes V1 and V3 (noted as the Inboard volumes for main steamlines B and C). Tables 5A and 5B of calculation JAF-CALC-19-00005, Revision 0 (page 56) shows the volumes credited for elemental iodine deposition and Table 5D (page 57) and 5F (page 59) of the same calculation show the elemental iodine deposition credited for volumes V1 and V3.

Section 2.2 Lack of Conformance with RG 1.183, Appendix A, Regulatory Positions 6.4 and 6.5 The FitzPatrick LAR requested, and the NRC approved, removal of the MSLCS from the FitzPatrick TS. This means the system can no longer be relied on and credited in the accident analysis. The purpose of this system was to collect and treat outboard MSIV valve stem, bonnet and seat leakage. Absent the MSLCS being credited, Regulatory Position 6.4 states that the MSIV leakage should be assumed to be released to the environment as an unprocessed ground-level release. Regulatory Position 6.4 also states that holdup and dilution in the turbine building should not be assumed. to the FitzPatrick letter dated August 8, 2019 (page 96 of Calculation JAF-CALC-19-00005, Revision 0), states the following regarding conformance to RG 1.183, Regulatory Position 6.4:

Since [the] MSLC is no longer credited, no ESF[s] are assumed to be available to collect or treat MSIV leakage. Releases are assumed to be from the Seismic Class I Turbine Stop Valves (TSVs) without holdup or dilution in the condenser or turbine building. The release is treated as a ground level release for dose assessment.

Contrary to Appendix A to RG 1.183, Regulatory Positions 6.4 and 6.5, FitzPatrick credited a reduction in MSIV releases in the steamline piping downstream of the MSIVs but did not demonstrate that the components and piping used in this release pathway are capable of performing their safety function. FitzPatricks safety analysis is based on the assumption that the leakage will only go through the MSIV valve seats and down the steamline to the TSVs. It also assumes that the TSVs will be open to the environment in a way that causes the leakage to follow this pathway. However, the pathway assumed in the FitzPatrick analysis is only one of several possible pathways to the environment, and FitzPatrick did not justify that the components and piping in the assumed pathway will cause the leakage only to go down the pathway assumed in FitzPatricks LOCA safety analysis. 13 13 Note that when performing these analyses, the various failure modes of the TSV must be considered to find the limiting release pathway. The FitzPatrick analysis assumed that the TSVs would leak to the environment. A more limiting scenario exists when the TSVs are assumed not to leak. This scenario would inhibit any MSIV seat leakage from flowing through the credited piping downstream of the MSIVs and would consider leakage through more limiting upstream pathways such as through the valve stem packing or bonnet.

16

Based on the above details, FitzPatrick did not comply with Regulatory Position 6.5 or have a leakage control system. Therefore, according to Regulatory Position 6.4, the analysis should have assumed that all MSIV leakage is released directly to the environment as an unprocessed release and without credit for the piping downstream of the MSIVs that could be bypassed. If FitzPatrick complied with Regulatory Position 6.4, the amount of radioactivity holdup and deposition would be significantly less than in the now approved analysis of record, and the design-basis doses would be significantly larger.

Section 2.3 Lack of Conformance to RG 1.183, Appendix A, Regulatory Position 6.4Dilution in the Turbine Building Incorporated into the Atmospheric Dispersion Factors for the Control Room , page 8 of the August 8, 2019, LAR states, The Turbine Stop Valves (TSVs) are assumed to be the release point for MSIV leakage.

Contrary to the guidance in RG 1.183, Appendix A, Regulatory Position 6.4, which states that Holdup and dilution in the turbine building should not be assumed and the quoted text in Section 2.2 above stating that in the FitzPatrick analysis that Releases are assumed without holdup or dilution in the condenser or turbine building, dilution was credited in the turbine building in the FitzPatrick analysis. The credit for dilution is not explicit but is implicitly considered in the method used to calculate the atmospheric dispersion factors. Rather than assuming the release point is on the surface of the turbine building, the FitzPatrick LOCA calculation assumes the release is at the TSVs within the turbine building. The extra distance from the turbine building surface to the TSV results in lower atmospheric dispersion factors, because the release is assumed to travel over a longer distance and thus is diluted within the turbine building.

Table 1 on page 11 of calculation JAF-CALC-19-00004, Revision 0, provided as to the August 8, 2019 LAR, gives the distance from between the column line 12 (the edge of the turbine building) and the TSVs is 32 feet, 9 3/4 inches. 17 This additional distance credited in the turbine building effectively reduced the atmospheric dispersion factors by over a factor of 2. 18 This reduction is equivalent to crediting a factor of two dilution in the turbine building. The FitzPatrick model is, therefore, a 17 This distance is provided in an equation in Table 1. The value of 32-9 3/4 is provided for the value of the distance from Column 12 to TSV. To show that Column 12 is the edge the turbine building please see FitzPatricks UFSAR Figure 12.3-16, Turbine Building Plan- El. 252-0 in.

18 The atmospheric dispersion factors were recalculated without credit for the distance within the turbine building. The direction to the control room was assumed to be the same as used by the licensee. This assumption is appropriate given that the integrity of the turbine building is not assured by a safety related system and the release point at the surface could be at the same direction as previously assumed by the licensee.

18

significant departure from the RG 1.183 methods FitzPatrick stated they used to demonstrate compliance with 10 CFR 50.67. If RG 1.183 was followed and dilution not credited in the turbine building, the FitzPatrick design basis control room LOCA doses would exceed those required to be met in 10 CFR 50.67 (based upon this issue alone). A recalculation of the FitzPatrick control room doses addressing this one issue is provided later in the DPO.

Section 2.4 Lack of Conformance to RG 1.183, Regulatory Position 5.1.2, and RG 1.194, Regulatory Position 2.0 , page 8 to the August 8, 2019, LAR states the following:

Additional release pathways out of the Turbine Building in the event of a LOCA were evaluated in ECR 5000012124 (Ref. 16). The smoke ejector vents on top of the Turbine Building are normally closed and would not open in a post-LOCA scenario. The access door from the Turbine Building 323 elevation to the Administrative Building roof is normally kept closed, locked, and monitored by security. There is also a Turbine Building exhaust duct that is 60 from the Control Room Intake. This was also determined not to be an appropriate release location because the vent is monitored for radiation and would be closed to prevent a radiological release to the environment. In the event of a loss of offsite power (LOOP) in combination with the LOCA, there would be considerable holdup in the Turbine Building.

Therefore, using the TSVs as the release location to the environment is both conservative and appropriate. , page 30 to the August 8, 2019 LAR states that The new /Q analysis complied with the methodology of RG 1.194. In Attachment 2 of the August 8, 2019 LAR is a table entitled Regulatory Guide 1.183 Conformance Matrix. In this table, the licensee stated that its LOCA analysis conforms with RG-1.183, Regulatory Position 5.1.2 (see page 48).

In part, RG 1.194, Regulatory Position 2.0, in part states the following:

Selection of conservative, bounding source-to-receptor combinations and less detailed site parameters for the /Q evaluation may be sufficient to establish compliance with regulatory guidelines.

RG 1.194, Regulatory Position 2.0, allows for the selection of conservative, bounding source-to-receptor combinations rather than calculating the atmospheric dispersion factors (or /Qs) for each source-to-receptor combination. The above paragraph taken from page 9 of the JAF-CALC-19-00004, Revision 0, attached to the August 8, 2019, LAR discusses the other source-to-receptor combinations and why they were not considered.

Contrary to RG 1.183, Regulatory Position 5.1.2, which, in part, states that Credit may be 19

taken for accident mitigation features that are classified as safety-related, [and] are required to be operable by TSs, the above justification credits turbine-building configurations and SSCs (smoke ejector vents, doors, and turbine-building exhausts) that are not safety related and are not required to be operable by the TS. Therefore, the bounding source-to-receptor combination and atmospheric dispersion factors are not assured.

Crediting these turbine building configurations and SSCs is inconsistent with the methodology in RG 1.183 for design calculations, because what is a normal configuration is not a required configuration to ensure safety. Also, nonsafety-related systems are not as reliable as safety-related systems and are not required to operate. If a licensee decides to change these configurations, there are no configuration controls in place to ensure that they are in the configurations assumed in the safety-related calculations, and there is no legal obligation to be in those configurations.

The effect of not complying with the above regulatory positions has not completely been assessed because the necessary turbine building drawings are not currently available. A qualitative example of one transport pathway shows the non-conservative impact of FitzPatrick not complying with Regulatory Position 5.1.2.

The turbine building exhaust is not safety related and should be considered as a potential release point if FitzPatrick had complied with Regulatory Position 5.1.2. If the turbine building exhaust to control room source-to-receptor combination was considered, the source exhaust location is closer to the control room intake than the distance assumed to be traversed in the environment when assuming the release is from the TSVs. Therefore, by not following the Regulatory Guidance methods the licensees analysis underestimated the atmospheric dispersion factor values. If RG 1.194 and RG 1.183 were to be followed, as the licensee claimed, the LOCA doses estimated by FitzPatrick would be larger.

Section 2.5 The FitzPatrick LOCA Now Approved Analysis of Record Does Not Consider the Worst Break Location for the Radiological Consequence Analysis In an e-mail dated February 8, 2020 (ADAMS Accession No. ML20035D576), the NRC staff requested additional information from FitzPatrick to justify that assuming a recirculation line rupture instead of a main steamline rupture is consistent with the guidance from RG 1.183 (see RAI ARCB-RAI-3 and its regulatory basis). In its letter dated March 30, 2020, the licensee responded by first stating that its analysis submitted in the LAR conservatively modeled only main steamlines B and C, which are symmetrical and shorter than lines A and D, and therefore would provide less volume for deposition.

However, to address the NRC staffs concern, the licensee provided the results of a separate analysis in which all four main steamlines were modeled to quantify the dose consequences of assuming a main steamline break in one of the lines. The licensees evaluation modeling the break in the main steamline also incorporated the changes in drywell spray removal but did not consider the lack of conformance to the other RG 1.183 regulatory positions, discussed under DPO Issue 2.

20

The NRCs SE stated the following:

The licensees results demonstrated that the dose consequences based on modeling two steam lines and a recirculation line break as submitted in the LAR resulted in a slightly higher calculated dose for the CR [control room]

than the analysis which modeled an MSL [main steamline] break in one line. The licensee stated that while the offsite doses for the ruptured MSL case were slightly higher than the analysis submitted in the LAR, the offsite doses remain well below the acceptance criteria. The NRC staff concludes that the analysis of the dose consequences resulting from an assumed MSL break provided by the licensee demonstrates that the impact of including an MSL break does not significantly impact the dose consequences from MSIV leakage for FitzPatrick.

The NRC staffs RAI states the following:

Please provide additional information to justify that assuming a recirculation line rupture instead of a main steam line rupture is consistent with the guidance from RG 1.183 that assumptions should be selected with the objective of maximizing the postulated radiological consequences.

Contrary to the stated guidance in this RAI (that assumptions should be selected with the objective of maximizing the postulated consequences) 19, the licensees analysis in its March 30, 2020, response does not adequately demonstrate that the now approved accident analysis of record maximizes the postulated radiological consequences. The licensees failure to demonstrate this conformance is due to two factors.

First, the licensees calculation provided in the response to the RAI is not consistent with the now approved analysis of record submitted as its new design basis. Four steamlines are considered rather than the two in the now approved analysis of record. Therefore, it is invalid for determining the impact of the licensees lack of modeling the break in the piping up to the first MSIV. When checking the impact of a questionable assumption, the impact of that assumption cannot be determined by making multiple changes that are inconsistent with the proposed licensing basis. These other changes can obfuscate the impact of the 19 The arguments made in the RAI are not as supportable as the following reasons to consider the impact of a reactor coolant system break in the steam line piping up to the outboard MSIV.

Appendix A, RG 1.183, Assumptions for Evaluating the Radiological Consequences of the A LWR Loss-of-Coolant Accident, states: With regards to radiological consequences, a large-break LOCA is assumed as the design basis case for evaluating the performance of release mitigation systems and the containment and for evaluating the proposed siting of a facility, and Leaks up to a double-ended rupture of the largest pipe of the reactor coolant system are included. 10 CFR 50.2, Definitions states: For nuclear power reactors of the direct cycle boiling water type, the reactor coolant system extends to and includes the outermost containment isolation valve in the main steam and feedwater piping. Based upon 10 CFR 50.2 and the quoted Appendix A statements it is clear that a reactor coolant system break up to the MSIVs must be considered.

21

issue being studied.

Secondly, while the NRC staffs SE concluded that the licensees analysis demonstrated that the impact of including a main steamline break does not significantly impact the dose consequences, this conclusion is not valid when the many other modeling issues in the LOCA analysis described in Section 2 are considered. These issues and the factor discussed in the paragraph above mask the significance of not considering a main steamline break in the analysis.

Section 2.6 The FitzPatrick LOCA Analysis Grossly Overestimated the Amount of Deposition in the Steamlines Regulatory Positions 6.3 and 6.5 in Appendix A to RG 1.183 allow credit for a reduction of MSIV leakage in the steamline if certain conditions are met and state that they will be evaluated on an individual case basis. The FitzPatrick revised LOCA analysis proposed to use a previously unapproved method for modeling aerosol settling in the main steamlines. As stated in the NRCs SE dated July 21, 2020, the proposed model treated aerosol removal by sprays and aerosol removal in the main steamlines as independent processes. They are not independent. The upstream source term in the containment has a direct impact on the downstream deposition rates in the steamline. In addition, the so-called 20-group probabilistic settling velocity distribution for modeling aerosol gravitational settling in the main steamlines is based on erroneous assumptions from Accident Evaluation Report (AEB) 98-03 ((that have been known to be flawed for over 14 years (since August 23, 2006).)) In addition, the FitzPatrick model does not appropriately model the impact of changing aerosol particle size distributions as the radioactivity moves down the steamline. The result of these issues is that the new model proposed by FitzPatrick and accepted by the NRC grossly overestimated the deposition in the steamlines and, therefore, does not accurately demonstrate compliance with 10 CFR 50.67.

History of Aerosol Deposition Modeling Issues To examine the Section 2.6 issue in more detail, a summary of the history of the aerosol deposition modeling issues is provided below.

Regulatory Information Summary (RIS) 2006-04, Experience with Implementation of Alternative Source Terms, dated March 7, 2006 (ADAMS Accession No. ML053460347),

was written, in part, to inform licensees that if they chose to use AEB 98-03, they needed to provide appropriate justifications that the assumptions used in AEB 98-03 are applicable for their particular design. RIS 2006-04 also noted that the settling velocity (rate at which aerosols are removed) should account for the distribution of particle sizes in that volume being analyzed. These distributions of particles sizes are dependent on containment sprays. So RIS 2006-04 essentially informed licensees that phenomena that impact the distribution of aerosol sizes, such as containment sprays, need to be considered and the assumptions made in these models need to be justified.

22

((An August 23, 2006, letter from the Director of Risk Assessment and Special Projects in the Office of Nuclear Regulatory Research to the Associate Director for Risk Assessment

& New Projects in the Office of Nuclear Reactor Regulation (NRR) (ADAMS Accession No. ML062010059) corrected several issues in the AEB 98-03, Revision 0, assessment.

The document recalculated the aerosol deposition velocities used to determine how much of the radiation in the steamline would make it out of the steamline. The report showed that Revision 0 erroneously overestimated the natural aerosol deposition settling velocity by approximately a factor of 100 (5.34E-6 (AEB 98-03, Revision 1) to 3.4E-4 m/s (AEB 98-03, Revision 0). The August 23, 2006, transmittal letter also noted that AEB 98-03

[Revision 0] was used to eliminate the TS requirements for MSLCS in BWRs. Therefore, the analyses used to justify the removal of this safety-related system to protect the health and safety of the public have been known to be erroneous and nonconservative for over 14 years.))

((In a response to the August 23, 2006, letter, the NRC staff requested the Office of Research to develop a more holistic analysis of the deposition expected in the steamline.))

SAND2008-6601, Analysis of Main Steam Isolation Valve Leakage in Design Basis Accidents Using MELCOR 1.8.6 and RADTRAD, issued October 2008 (ADAMS Accession No. ML083180196), provides the results of this analysis. SAND2008-6601 represents the best available state-of-the-art methods that incorporate and synthesize the uncertainty involved in the determination of aerosol deposition in the steamline and condenser. It has been vetted in independent studies and reviewed by numerous internal experts and internationally recognized experts on accident analysis and aerosol physics.

These reviews found the report to contain sound and defensible science that the NRC to this day has failed to fully address in their reviews of erroneous analyses such as FitzPatricks.

In October 2009, the NRC incorporated the results of SAND2008-6601 in a draft revision of RG 1.183 (Draft Guide (DG)-1199) and provided it to the public for comments. 20 These comments were resolved, and appropriate changes were made to RG 1.183. The changes to RG 1.183 and the resolution of public comments were sent to over 40 internal reviewers at least three times, and all known internal comments were resolved with those reviewers.

Since the draft final Revision 1 of RG 1.183 ((DFRG-1.183) was completed in December 2011, NRR management has not concurred on the review package, essentially blocking its issuance. No documented or verbal reasons for blocking its issuance were provided, until recently. 21 20 Note that the current version of RG 1.183 does not endorse any methods for calculating these aerosol depositions.

21 In June 2019, DFRG-1.183 was updated to the current regulatory guide formatting and requirements.

Also in 2019, Exelon submitted four LARs asking for increases in allowable MSIV leakage. NRR management decided to delay concurrence on the DFRG to inform it with lessons learned from these MSIV leakage reviews that included the FitzPatrick LAR. This delay is currently expected to last until the second quarter of 2022 or about another 3 years from when the June 2019 DFRG 1.183 was ready for concurrence. This decision further delays incorporating the knowledge of errors in the AEB 98-03 method into the NRC staffs regulatory guidance.

23

((In a user need request dated April 20, 2015 (ADAMS Accession No. ML15096A515),

NRR asked the Office of Nuclear Reactor Research to conduct an independent review of DFRG-1.183 for technical adequacy and the resolution of public comments. An internal memorandum dated May 20, 2019 (ADAMS Accession No. ML19094A625) provided a close out out of the research assistance request and transmitted a technical note estimating the removal of aerosols in steam lines.)) In SAND2017-2651, A Note on Aerosol Removal by Gravitational Settling in a Horizontal Steam Pipe, issued February 2017 (ADAMS Accession No. ML19094A465), two experts (on aerosols and nuclear accidents) showed how the AEB 98-03, Revision 0 study, used as the basis for FitzPatricks approved steamline model, grossly overestimates the amount of aerosol deposition (as compared to expert scientific knowledge) and that the total amount of decontamination (removal of aerosols in the steam piping) would be much smaller than if the erroneous AEB 98-03, Revision 0, is used.22 In addition to the above history, the NRC staff has documented issues with the MSIV leakage pathway modeling in numerous RAIs, in SEs for BWRs, and discussed these issues in two public workshops on RG 1.183, and in its periodic review of RG 1.183, entitled Results of Periodic Review of Regulatory Guide 1.183, dated June 25, 2018 (ADAMS Accession No. ML18149A069).

An example of the staffs documentation of these concerns appears in an SE for the Nine Mile Point, Unit 2 (NMP2) LOCA, dated May 29, 2008 (ADAMS Accession No. ML081230439). In this SE, the NRC staff stated that to address historically documented NRC concerns about the use of AEB 98-03 and concerns about crediting drywell sprays, 23 the NMP2 containment bypass model (1) limited the assumed credit for the aerosol settling velocity in the steamline (6.6E-5 meters/second for NMP2 versus 1.17E-3 meters per second for Perry (AEB 98-03, Revision 0)a reduction by a factor of 18), (2) limited the deposition credited to only one pipe segment between the closed isolation valves, and (3) did not credit deposition in a steamline with the assumed stuck-open valve. This is a clear example of how the NRC acknowledged that crediting drywell sprays will impact the aerosol deposition in the steamlines and that additional conservatisms are needed in the NMP2 LAR to show compliance with 10 CFR 50.67.

22 Table 5-1 of the SAND2017-2651 shows that the maximum decontamination factor for the single well mixed volume studied is between 13 and 27 times smaller than the maximum decontamination derived using AEB 98-03.

23 The NRC staff SE for NMP2 stated that the modeling of various mechanisms for radioactive aerosol removal (drywell spray removal and settling credit) should consider the effects of one model on the others and that NMP2 conservatively limited the credited gravitational settling effects on airborne activity, in part, by using the aerosol settling conservatisms.

24

The Treatment of Aerosol Settling in FitzPatricks Main Steamline Compared to Existing Staff Knowledge and Science FitzPatricks LAR proposed a new main steamline deposition model. The now approved model estimated the removal coefficients in both credited steamline volumes to be approximately 22 hr-1 for 30 days. This is over twice the value in the known to be erroneous and nonconservative AEB 98-03, Revision 0, model that estimated these values to be about 9.0 hr-1 for the inboard and outboard steamline volumes. The difference between using the values of 22 hr-1 by FitzPatrick compared to 9 hr-1 (AEB 98-03) is significant. The values FitzPatrick proposed and the NRC approved, are unsupportable.

The FitzPatrick model credits approximately four times more radiation deposition than even the erroneous AEB 98-03, Revision 0.

When compared to the NRCs most state-of-the-art deposition model, the SAND2008-6604 model, the SAND2008-6601 deposition factors are about a factor of 10 to 20 less than the FitzPatrick analysis values for the volumes between the MSIVs and downstream of the outboard MSIV (0.7 hr-1 to 1.3 hr-1). 24 This is roughly the same as the factor of 18 reduction in removal coefficients used by Nine Mile Point described above.

Furthermore, FitzPatrick reassessed the aerosol deposition in the main steamline and provided the results in a March 30, 2020, letter to the NRC staff. The FitzPatrick reanalysis assessed the impact of sprays on the main steamline aerosol deposition. The analysis showed that the removal coefficients (or lambdas) when considering drywell sprays were much less than in the now approved analysis of record LOCA that FitzPatrick used to justify the LAR. Where the now approved analysis of record estimated the removal coefficients in the steamline to be approximately 22 hr-1, the values calculated considering the drywell sprays were calculated to be on the order of 0.06 hr-1 to 0.598 hr-1. 25 However, these revised values were not used by either the licensee in its now approved design-basis analysis of record or the NRC in its review of the FitzPatrick LAR.

A FitzPatrick reviewer estimated the proposed removal coefficients provided in FitzPatricks now approved analysis of record. The results of this analysis were not officially documented but were provided in two e-mails from the NRC reviewer who wrote the SE section on FitzPatricks removal coefficients. The e-mails, dated August 31, 2020, 24 There are some significant differences in the SAND2008-6601 and the FitzPatrick model that would tend to make this comparison more favorable for FitzPatrick. The Sandia study credited more deposition mechanisms than did FitzPatrick.

25 The NRC staff did not perform a complete review of these removal coefficients, because (1) no basis was provided for the assumptions used, (2) these removal coefficients were not used in the licensees proposed and now approved analysis of record, and (3) the NRC staff did not use them for demonstrating compliance with 10 CFR 50.67 for this particular LAR. However, the NRC staff notes that the licensee assumed a 2-micron aerosol mass median diameter and a geometric standard deviation of 2.0 particle size in its reassessment, is not conservative given that the range of particles is expected to be as low as 1 micron and the deposition rate of smaller particles is less than for larger particles. Therefore, even though the recalculated removal rates are much lower than those assumed in the now approved analysis of record, they are likely non-conservative because they do not consider the range of particle sizes expected for this accident.

25

and November 10, 2020 (which provided a draft report with the NRC reanalysis of the FitzPatrick removal coefficients entitled Re-Evaluation of AEB 98-03 Settling Velocity Distribution and the Multi-Group Method), showed that the NRC reevaluation calculated the removal coefficients to be between approximately 9.9 hr-1 and 11.2 hr-1.

Table 1 below summarizes the historical information discussed above and provides the magnitude of main steamline aerosol deposition calculated by FitzPatrick and by the NRC (for this LAR, the AEB 98-03 calculations, and using state-of-the-art methods in SAND2008-6601).

Contrary to the staffs longstanding knowledge of large errors in the estimated removal coefficients in the steamline and the NRCs public and internal documentation of these errors, the NRC staff approved the FitzPatrick erroneous analysis that clearly overestimated the removal of radioactivity in the steamline. The removal coefficients approved were over twice as large as those in the known-to-be-erroneous AEB 98-03, Revision 0, calculation.

Contrary to Regulatory Positions 6.3 and 6.5 in Appendix A to RG 1.183, the NRC staff evaluation of these amounts of the reduction due to deposition in the main steamline piping review of the proposed deposition model did not incorporate longstanding staff knowledge that these models were grossly nonconservative and do not correlate with known science. Also contrary to RG 1.183, which states that The removal rate constants selected for use in design basis calculations should be those that will maximize the dose consequences, the FitzPatrick main steamline removal coefficients are over twice as large as any previously approved using AEB 98-03, Revision 0 and 1 to 2 orders of magnitude greater than state-of-the-art science estimates.

Contrary to the guidance in RIS 2006-04, the FitzPatrick now approved analysis of record does not account for the changes in aerosol sizes due to the drywell sprays. When FitzPatrick accounted for these changes in aerosol sizes, the main steamline aerosol removal factors were between 37 and 367 times smaller than the aerosol removal in the FitzPatrick now approved analysis of record.

Also contrary to the guidance in RIS 2006-04, the FitzPatrick now approved analysis of record does not account for the changes in the aerosol size distributions as the radioactivity moves down the steamline piping. Calculation JAF-CALC-19-00005, Revision 0, states that the NRC staff had the following concern:

Gravitational settling would be expected to be at a lesser rate for the later sections of piping and at later times considering that the larger and heavier aerosols would have already settled out of the main steam lines in upstream sections of piping.

Inspection of the approved removal coefficients based on the unproven, unbenchmarked FitzPatrick main steamline aerosol deposition model, shows that there is hardly any change between the inboard and outboard deposition volumes (22.03 hr-1 to 22.00 hr-1 in 26

one steamline and 22.23 hr-1 to 21.96 hr-1 in the other) and they do not change over time.

In comparison, the state-of-the-art SAND2008-6601 model and FitzPatrick shows large changes due to the changing aerosol sizes over time and space. Despite having this knowledge since 2008, the NRC approved the erroneous FitzPatrick model which does not adequately address the concerns in RIS 2006-04.

Therefore, the FitzPatrick analysis of record now approved cannot legitimately be used to show compliance with 10 CFR 50.67. The NRC knows that the aerosol deposition factors are grossly nonconservative and inconsistent with known science. Even the licensees re-evaluated estimates (contained in the RAI responses) used to assess the impact of the sprays (now credited in the FitzPatrick design), show the nonconservative nature of the approved FitzPatrick model. Despite this knowledge, the NRC approved the erroneous modeling of aerosol deposition in the steamline.

Section 2.7 Lack of Conformance with RG 1.183, Appendix A, Regulatory Position 3.3 RG 1.183, Appendix A, Regulatory Position 3.3, states that SRP Section 6.5.2, Containment Spray as a Fission Product Cleanup System, provides an acceptable model for the removal of iodine and aerosols. The NRCs SE, dated July 21, 2020, states the following:

The NRC staff notes that SRP Section 6.5.2 limits the elemental iodine removal constant to a value of 20 hr-1. The NRC staff performed confirmatory calculations using the RADTRAD program to assess the impact of including this limitation on the licensees dose analysis and concluded that the impact was not significant for this case. Accordingly, the NRC staff finds this deviation from the SRP acceptable for this LAR. As discussed in the SRP, the elemental iodine removal constant should be limited to 20 hr-1. Because the significance of this limitation to the dose calculation is case-specific, the NRC staffs finding in this SE is only applicable to this LAR.

Contrary to the above NRC staffs SE statements regarding SRP 6.5.2, the FitzPatrick LOCA analysis used an elemental removal constant of 26.36 hr-1, which is greater than the 20 hr-1 value noted in the NRC staffs SE. 26 While the NRC staffs SE addressed this 26 Per page 2 of Attachment 1 to the FitzPatrick letter dated March 30, 2020, the revised particulate aerosol removal coefficient used in the now approved LOCA analysis of record (JAF-CALC-19-00005, Revision 1) is provided. While the March 30, 2020 letter does not address the elemental spray coefficient, Attachment 1 to the August 8, 2019 LAR, page 19 states: the elemental iodine removal coefficient is conservatively assumed to be the same as the particulate aerosol removal coefficient. Therefore, for this issue it is assumed that Revision 1 of the now approved analysis of record also changed the assumed elemental spray removal when the particulate spray removal coefficient revised. If this assumption is not true the value would be the value in Revision 0 which is 30 hr-1.

27

issue, this difference in elemental removal is more significant when the other issues in the LOCA analysis, discussed above, are considered. These issues mask the significance of not using a value of 20 hr-1 in the LOCA analysis.

28

Section 3 Summary and Impact on Dose Analysis When the Regulatory Guidance Is Not Followed Each of the above issues is significant to the accident analysis used to show compliance with 10 CFR 50.67. Some issues are more significant than others. While the margin between the now approved analysis of record doses and the 10 CFR 50.67 reference values (or dose limits) is large for the exclusion area boundary and low population zone, the margin to the control room dose limit is much less.

Table 2 shows the now approved analysis of record results. This table is taken from , page 3 of FitzPatricks March 30, 2020 response to the NRC staffs request for additional information. Per Table 2, the control room total dose is estimated to be 4.67 rem TEDE. Since the 10 CFR 50.67 control room dose limit is 5 rem TEDE, the FitzPatrick now approved analysis of record has a dose margin of only 0.33 rem TEDE to the 10 CFR 50.67 limits.

The impact of selected issues 27, discussed in Section 2, on the FitzPatrick LOCA control room doses was evaluated and is provided in Table 3. Case 1 in Table 3 shows the impact of changing the FitzPatrick now approved analysis of record and making it consistent with Appendix A, RG 1.183, Regulatory Position 6.4. This one change would cause the now approved analysis of record control room results to exceed 10 CFR 50.67.

Case 2 shows how changing the now approved analysis of record to be consistent with the RG 1.183 regulatory positions discussed in Sections 2.1 - 2.3 above would result in doses a factor of approximately 100 higher than the now approved analysis of record. It is expected that if the other issues identified above were to be considered the control room doses would be higher than Case 2.

When the fundamental flaws, described in Section 2 are addressed, the FitzPatrick LOCA control room doses do not meet the 10 CFR 50.67 dose criterion necessary for approving the TS changes approved by the NRC staff.

27 Only selected issues were evaluated due to the limited time allotted for writing a DPO.

29

Table 2: Now Approved LOCA Analysis of Record Results 31

Table 3: FitzPatrick LOCA Control Room Dose When Selected Issues Identified in Issue 2, Section 2 Are Addressed Total Case Description Addressed DPO Issues 2 Section and Description 35 Control 2.1 2.2 2.3 Room Dose Containment Mixing & Leakage Pathway to Turbine Building Deposition Environment Dilution (TEDE)

N/A JAF-CALC-19-00005, Rev. 1 36 4.67 1 DPO Estimate 37 X 8 2 DPO Estimate 38 X X X 466 N/A 10 CFR 50.67 Limit 5 35 Note that not all Issue 2 sections are listed here. Only a limited number of evaluations were completed.

36 Revision 1 is the now approved FitzPatrick LOCA analysis of record taken from Table 2 above.

37 The Fitzpatrick model was revised to be consistent with the regulatory guidance as discussed in DPO Issue 2, Sections 2.3. The amount of dilution in the turbine building taken by FitzPatrick reduced the dose by roughly a factor of 2. If no turbine building dilution was taken (consistent with RG 1.183) the MSIV leakage doses would be approximately a factor of 2 higher for the MSIV leakage pathway.

Also, the total control room dose estimates for Case 1 and 2 in this table do not include the impact of the decreased spray removal discussed in the FitzPatrick RAI response dated March 30, 2020 to RAI ARCB-RAI-1B (Attachment 1, page 2). It is expected that using these smaller spray removal coefficients would increase the total control room dose estimates calculated for this DPO.

38 The Fitzpatrick model was revised to be consistent with the regulatory guidance as discussed in DPO Issue 2, Sections 2.1, 2.2 and 2.3.

32

Summary of Issue 3 The NRCs SE for Amendment 338 is fundamentally flawed and inconsistent with regulatory requirements in 10 CFR 50.36. In direct conflict with 10 CFR 50.36(c)(2)(ii),

Criterion 3; and RG 1.183, the NRC staff confirmatory analysis credits the plants power conversion system piping and valves when these structures, systems, and components (SSCs) do not have established TS limiting conditions for operation (LCOs). 10 CFR 50.36 requires the incorporation of these LCOs to ensure that the facility SSCs have and will maintain the assumed integrity in the LOCA analysis.

Supporting Details for Issue 3 The details supporting Issue 3 are broken into five sections. Section 1 provides an introduction to Issue 3. Section 2 provides applicable parts of the 10 CFR 50.36 regulation, regulatory guidance and technical specification bases pertaining to Issue 2.

Section 3 discusses details about Criterion 3 of 10 CFR 50.36(c)(2)(ii). Section 4 provides a discussion of leakage pathways that bypass the condenser credited in the NRCs confirmatory analysis. Section 5 discusses valve stem and other direct leakage pathways that bypass the condenser.

Section 1 Introduction Licenses to operate utilization facilities (nuclear reactors used for power production) must include TS in accordance with Section 182 of the Atomic Energy Act of 1954, as amended. TS provide the specific characteristics of the facility and the conditions which are the lowest functional capability or performance levels of equipment required for safe operation of the facility. The TS are so important that they are a part of the license and cannot be changed without prior Commission approval.

These TS are generally linked to matters that relate to the prevention of accidents or to the mitigation of the consequences of accidents. By using systematic analysis and evaluation 39 of the facility, licensees are expected to identify the SSCs that are related to maintaining the integrity of the physical barriers designed to contain radioactivity. Those SSCs are expected to be the subject of limiting conditions of operation in the TS of the operating license (see 10 CFR 50.36, Criteria 2 and 3). However, contrary to the requirements in regulations, the NRC staffs LOCA analysis credited the integrity of physical barriers used to mitigate the consequences of accidents without requiring that these SSCs be in the TS.

39 Per 10 CFR 50.36(b), those derived from the analyses and evaluations in the safety analysis report and amendments. These are the analyses performed using RG 1.183.

33

Section 2 Regulations, Regulatory Guidance, and Technical Specification Bases Below is a brief discussion of 10 CFR 50.36, RG 1.183, FitzPatricks TS bases, and Criterion 3 in 10 CFR 50.36, as they pertain to DPO Issue 3.

In 10 CFR 50.36(c)(2)(ii), the NRC requires, in part, the following:

A technical specification limiting condition for operation of a nuclear reactor must be established for each item meeting one or more of the following criteria:

Criterion 3. A structure, system, or component that is part of the primary success path and which functions 40 or actuates to mitigate a design basis accident or transient that either assumes the failure of or presents a challenge to the integrity of a fission product barrier.

RG 1.183, Regulatory Position 5.1.2, in part, states the following:

Credit may be taken for accident mitigation features that are classified as safety-related, are required to be operable by technical specifications, are powered by emergency power sources, and are either automatically actuated or, in limited cases, have actuation requirements explicitly addressed in emergency operating procedures.

RG 1.183, Appendix A, Regulatory Positions 6.4 and 6.5, states the following:

In the absence of collection and treatment of releases by ESFs such as the MSIV leakage control system, or as described in paragraph 6.5 below, the MSIV leakage should be assumed to be released to the environment as an unprocessed, ground-level release. Holdup and dilution in the turbine building should not be assumed.

A reduction in MSIV releases that is due to holdup and deposition in main steam piping downstream of the MSIVs and in the main condenser, including the treatment of air ejector effluent by offgas systems, may be credited if the components and piping systems used in the release path are capable of performing their safety function during and following a safe shutdown earthquake (SSE). The amount of reduction allowed will be evaluated on an individual case basis. References A-9 and A-10 provide guidance on acceptable models.

Amendment 338 revised TS 3.6.1.3 and 3.6.1.8. The bases for these TS are discussed below as they pertain to Issue 3.

40 Historically, both passive and active SSCs are used to perform these functions.

34

FitzPatricks TS Bases B 3.6.1.3, Primary Containment Isolation Valves, Applicable Safety Analysis, in part, states the following:

The PCIV [Primary Containment Isolation Valves] LCO was derived from the assumptions related to minimizing the loss of reactor coolant inventory, and establishing the primary containment boundary during major accidents.

As part of the primary containment boundary, PCIV OPERABILITY supports leak tightness of primary containment. Therefore, the safety analysis of any event requiring isolation of primary containment is applicable to this LCO.

PCIVs satisfy Criterion 3 of 10 CFR 50.36(c)(2)(ii) (Ref. 7).

FitzPatricks TS Bases B 3.6.1.1, Primary Containment, Applicable Safety Analysis, in part, states the following:

The DBA that postulates the maximum release of radioactive material within primary containment is a Loss of Coolant Accident (LOCA). In the analysis of this accident, it is assumed that primary containment is OPERABLE such that release of fission products to the environment is controlled by the rate of primary containment leakage.

Analytical methods and assumptions involving the primary containment are presented in References 1 [UFSAR, Section 5.2] and 2 [UFSAR, Section 14.6.1.3]. The safety analyses assume a nonmechanistic fission product release following a DBA, which forms the basis for determination of offsite doses. The fission product release is, in turn, based on an assumed leakage rate from the primary containment. OPERABILITY of the primary containment ensures that the leakage rate assumed in the safety analyses is not exceeded.

Primary containment satisfies Criterion 3 of 10 CFR 50.36(c)(2)(ii) (Ref. 4).

Before the issuance of Amendment 338, FitzPatricks TS Bases B 3.6.1.8, MSLC System, Applicable Safety Analysis, stated the following:

The MSLC System mitigates the consequences of a DBA LOCA by ensuring that fission products that may leak from the closed MSIVs are diverted to and filtered by the SGT [standby gas treatment] System. The operation of the MSLC System prevents a release of untreated leakage for this type of event.

The MSLC System satisfies Criterion 3 of 10 CFR 50.36(c)(2)(ii) (Ref. 3).

35

Section 3 Discussion of Criterion 3 A basic concept in the adequate protection of the public health and safety is that if a postulated DBA or transient occurs, SSCs are available to function or to actuate to mitigate the consequences of the DBA or transient. Safety sequence analyses or their equivalent provide a method of presenting the plant response to an accident. These can be used to define the primary success paths.

A safety sequence analysis is a systematic examination of the actions required to mitigate the consequences of events considered in the plants DBA and transient analyses, as presented in Chapters 6 and 15 of the plants FSAR (or equivalent chapters). Such a safety sequence analysis considers all applicable events, whether explicitly or implicitly presented. The primary success path of a safety sequence analysis consists of the combination and sequences of equipment needed to operate (including consideration of the single-failure criterion), so that the plant response to DBAs and transients limits the consequences of these events to within the appropriate acceptance criteria.

It is the intent of criterion 3 to capture into the TS only those SSCs that are part of the primary success path of a safety sequence analysis. This criterion also captures those support and actuation systems that are necessary for items in the primary success path to successfully function.

Section 4 Leakage Pathways that Bypass the Condenser Credited in the NRC Analysis Containment structures, their isolation systems, and components are designed to reduce the radiological consequences and risk to the public from nuclear accidents. When designing these SSCs, the design-basis dose safety analysis plays a critical role in determining the allowable leakage limits from these SSCs. These design-basis analyses consider the potential scenarios and pathways to the environment after an accident. The TS revised by Amendment 338 involve the MSIVs and the MSLCS. The MSIVs and the MSLCS were designed and controlled to ensure that radioactivity after an accident complies the regulations and is consistent with the design requirements, assumptions, and inputs in the FitzPatrick LOCA radiological consequence.

An assumption used in the revised FitzPatrick LARs LOCA analysis and RG 1.183 is that all MSIV leakage goes to the environment via the turbine building without going through the condenser. The NRC staffs analysis does not use this assumption, and explicitly credits the power conversion system. However, the power conversion system has no TS LCO to ensure that the mitigation of radioactivity credited in the NRC staffs analyses remains consistent with the way the plant is maintained and operated. For example, leakage though isolation valves, connected to steam equalization header, and not controlled by TS could bypass the condenser. In addition, the removal of the MSLCS would allow outboard MSIV stem leakage to leak into the main steam tunnel also bypasses the power conversion system credited in the NRC staffs analysis.

36

A discussion is provided below of the leakage pathways to the environment and their relationship to the LOCA analysis and TS. These credible MSIV leakage pathways include leakage from valve stems, other direct leakage, SSCs not tested for leakage, and SSCs that are open, by design, to the environment.

Section 5 Valve Stem and Other Direct Leakage Pathways Because SSCs cannot be assured to be leaktight and are not required to be leaktight, traditional nuclear safety designs consider leakage from the MSIV leakage pathway to be ultimately leaked to the environment. To ensure that the health and safety of the public is maintained consistent with these designs, the amount of leakage is required to be tested periodically. What components are tested, and how they are tested is important to understanding which credible leakage pathways to the environment via the turbine building are assumed.

The MSIVs and the piping between the MSIVs are the only components in the MSIV leakage pathway that must be tested to ensure that regulations used to protect the health and safety of the public and operators are met. These tests are required by a TS surveillance requirement. The test determines the amount of leakage from the MSIVs and the piping between the MSIVs by pressurizing the volume between the MSIVs. The amount of leakage is determined from the loss of pressure over time. This test is good for finding the amount of leakage, but it does not determine where the leakage occurs.

A common misunderstanding in MSIV leakage testing is that it measures only valve seat leakage. However, the MSIV leakage surveillance testing is also used to test other credible leakage pathways. Other than through the valve seats, leakage can occur from the valve stem packing or other direct leakage from the valve body or steamline piping.

The valve stem packing leakage and other direct leakage from components outside of containment were highlighted as staff concerns in the development of leakage control systems used to resolve Generic Safety Issue 16 (GSI-16), BWR Main Steam Isolation Valve Leakage Control Systems (NUREG-0933, Main Report with Supplements 1-34).

Licensees also had these concerns as shown in the FSARs describing the purpose of MSIV leakage control systems. In these FSARs, licensees stated that these MSLCSs were designed, in part, to collect stem packing leakage. Amendment 338 allowed the removal of the requirements for the MSLCSs used to reduce uncontrolled or untreated MSIV leakage to the environment.

Releases through the valve packing or other direct leakage from MSIVs, TSVs, and turbine control valves or piping contained in the steam tunnel would be discharged to the turbine building or to the environment via the steam tunnels. Any MSIV leakage releases from stem packing and other direct leakage from SSCs contained in the turbine building would be let go directly to the turbine building, where they can be released to the environment. Therefore, if there is no MSLCS or components and piping systems to the 37

condenser to decrease the driving force out of the valve packing or other direct leakage pathways, then this leakage should be assumed to leak to the environment as an unprocessed, ground- level release. (see RG 1.183, Appendix A, Regulatory Positions 6.4 and 6.5).

Contrary to the RG 1.183 guidance, the NRC staff credited the power conversion system for mitigating the accident without an MSLCS or docketed information requesting credit for, and statements from FitzPatrick that, these components, piping systems, and procedures exist and create a release pathway capable of performing the safety function of transporting this radiation to the condenser. The licensee also must commit to using these pathways during the LOCA in order for them to be credited in the accident analysis.

Therefore, contrary to the requirements of 10 CFR 50.36, the NRC staff credited the power conversion system without any of the credited fission product barriers being linked to a TS LCO.

38

Summary of Issue 4 The NRCs approval of the FitzPatricks LOCA analysis, based upon an NRC staff analysis that credited portions of the power conversion system not credited in the FitzPatrick analysis, does not ensure that a clear, logical, and consistent design basis exists to support evaluations of future modifications, SEs and NRC inspections for FitzPatrick.

Supporting Details for Issue 4 The details supporting Issue 4 are broken into two sections. Section 1 provides applicable parts of the 10 CFR 50.59 regulation, RG 1.183 and SRP 15.0.1 pertaining to Issue 4. Section 2 provides a discussion of Issue 4.

Section 1 Regulations and Guidance - 10 CFR 50.59, Regulatory Guide 1.183 and SRP 15.0.1 Title 10 of the Section CFR 50.59, Changes, tests and experiments, states:

(2) A licensee shall obtain a license amendment pursuant to Sec. 50.90 prior to implementing a proposed change, test, or experiment if the change, test, or experiment would:

(iii) Result in more than a minimal increase in the consequences of an accident previously evaluated in the final safety analysis report (as updated);

RG 1.183, Regulatory Position 1.1.1, in part states:

Once the initial AST implementation has been approved by the staff and has become part of the facility design basis, the licensee may use 10 CFR 50.59 and its supporting guidance in assessing safety margins related to subsequent facility modifications and changes to procedures.

RG 1.183, Regulatory Position 1.5, in part, states the following:

The NRC staffs finding that the amendment may be approved must be based on the licensees analyses, since it is these analyses that will become part of the design basis of the facility 39

In addition, the SRP 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, states:

Independent calculations should be performed as necessary to conclude, with reasonable assurance, that the applicants analyses are acceptable. The staffs approval of the application is to be based on the licensees docketed information. If differences are discovered between the licensees methods and assumptions and those deemed acceptable to the staff, the reviewer should resolve the differences with the licensee. If necessary, the licensee should update the disputed assumptions and resubmit the affected analyses.

Section 2 Discussion The regulatory issue described in Issue 1 and the technical issues described in Issue 2 directly and negatively impact future plant or design modifications using 10 CFR 50.67 and 10 CFR 50.59, and NRC inspections. As discussed in Issue 1, the NRC staff did not follow the 10 CFR 50.67 regulation when the NRC approved the LAR. The NRC staff also did not follow SRP 15.0.1 and work out differences between the licensees docketed information and assumptions and methods deemed acceptable to the NRC staff (such as RG 1.183). The NRC based its acceptance, in part, upon its own independent analysis that differed significantly from the FitzPatrick analysis. The NRCs analyses used different inputs and assumptions and yielded significantly different qualitative results than the licensees now approved analysis of record. 41 By approving the analysis of record proposed by FitzPatrick without addressing these differences, or the differences in the licensees analyses (now the analysis of record) and the regulatory guidance, future 10 CFR 50.59 (for changes that impact the DBA LOCA) and revisions to the DBA LOCA analysis to show compliance with 10 CFR 50.67 are impacted. The now approved analysis of record contains flawed inputs and assumptions that are now part of the design basis for FitzPatrick, and assumptions that staff used that are not required to be maintained by FitzPatrick.

The approval of the flawed FitzPatrick analysis, put technically indefensible aerosol deposition models (inconsistent with known science regarding aerosol deposition) into the licensing basis for FitzPatrick and other assumptions that are clearly not consistent with RG 1.183. That means that future analyses that use these assumptions and inputs will not agree with scientific knowledge or the NRC staffs guidance. It is not clear how an NRC inspector would have any means to enforce that the licensees SSCs remain in the configurations relied upon in the NRCs analysis. Likewise, future 10 CFR 50.67 and 10 CFR 50.59 evaluations involving changes that impact the LOCA are likely invalid since the starting point of these evaluations is now a flawed LOCA analyses.

41 As stated in the NRCs independent assessment documented in the NRCs SE, doses will be significantly lower than those estimated using deterministic methods that do not credit holdup and retention of the MSIV leakage within the PCS.

40

Summary of Issue 5 Even if it was appropriate for the staff to consider their risk and engineering insights to determine the licensees compliance with 10 CFR 50.67, those risk and engineering insights would not compensate for the uncertainties and errors in the FitzPatrick LOCA analysis and provide reasonable assurance that 10 CFR 50.67 is met.

Supporting Details for Issue 5 Issue 5 is based upon an analysis by an author of this DPO using information provided by FitzPatrick. The analysis determined the impact of crediting the NRCs risk and engineering insights. Specifically, FitzPatricks now approved LOCA analysis was revised to credit transport of MSIV leakage to the condenser and deposition and holdup of MSIV leakage in the condenser prior to it being released to the environment.

First, an NRC staff model of the MSIV leakage pathway was developed. The NRC model replicated the now approved FitzPatrick LOCA model that does not consider credit for mitigation of the MSIV leakage by the condenser. The NRC staff model used inputs and assumptions from the now approved FitzPatrick LOCA analysis, and the NRC computer code RADTRAD to calculate the radiological consequences. The NRC MSIV leakage model results were benchmarked against the approved FitzPatrick LOCA analysis. The NRCs analysis results matched the licensees control room dose from the MSIV leakage pathway within 1%.

Next, the NRCs model was revised to consider mitigation of the MSIV leakage by the condenser consistent with the NRCs risk and engineering insights. In the revised model, MSIV leakage is sent to the condenser prior to being released to the environment, via the main steam line drains. 42 The model was also revised to address the issues discussed in Issue 2, Section 2.1, 2.3 and 2.5 above, so that the model conforms with RG 1.183. The revised model used the condenser volume and condenser aerosol deposition provided by FitzPatrick in their March 30, 2020 response to an NRC staffs request for additional information. The condenser free volume was taken from Table RAI-2b (page 11) and the aerosol removal in the condenser was taken from Table 2d (page 14). 43 For simplicity all elemental iodine that entered the condenser was assumed to be deposited in the 42 Using the drain lines is consistent with the FitzPatrick, March 30, 2020 response (page 6) to an NRC staff request for additional information. In this response FitzPatricks sensitivity model assumed the MSIV leakage to travel to the condenser through the drain lines from the main steam line piping between the MSIVs. This modeling is consistent with other plants in the Exelon fleet (e.g. LaSalle and Limerick).

43 The NRC staff notes that the licensee assumed a 2-micron aerosol mass median diameter and a geometric standard deviation of 2.0 particle size in its calculation of the condenser removal coefficients. These assumptions are not conservative given that the range of particles is expected to be as low as 1 micron and the deposition rate of smaller particles is less than for larger particles.

Therefore, the condenser removal coefficients provided by FitzPatrick are likely non-conservative because they do not consider the lower range of particle sizes expected for this accident. However, for the purpose of the DPO authors analysis, using the non-conservative removal coefficients yields the same conclusion that the regulations are not met.

41

condenser.

The revised models control room dose for the MSIV pathway was combined with the licensees control room doses from the other contributors to control room dose (containment leakage, ESF leakage, etc.) provided in Table 2 of this DPO. The results of this analysis show that even when the condenser is credited, the control room dose for the LOCA would exceed the 10 CFR 50.67 regulation criterion of 5 rem TEDE.

42

Document 2: Memo Establishing DPO Panel UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 February 12, 2021 MEMORANDUM TO: Theresa Clark, Panel Chair Office of Nuclear Material Safety and Safeguards Donnie G. Harrison, Panel Member Office of Nuclear Material Safety and Safeguards Ann Marie Stone, Panel Member Region III THRU: George A. Wilson, Director Office of Enforcement FROM: Ian A. Gifford, Differing Views Program Manager Office of Enforcement

SUBJECT:

AD HOC REVIEW PANEL - DIFFERING PROFESSIONAL OPINION ASSOCIATED WITH THE FITZPATRICK AMENDMENT CONCERNING AN ALTERATE SOURCE TERM FOR CALCULATING LOSS-OF-COOLANT ACCIDENT DOSE CONSEQUENCES (DPO-2021-001)

In accordance with Management Directive (MD) 10.159, The NRC Differing Professional Opinion Program; and in my capacity as the Differing Views Program Manager (DVP PM); and in coordination with George Wilson, Director, Office of Enforcement, Andrea Veil, Acting Director, Office of Nuclear Reactor Regulation and the Differing Professional Opinion (DPO) submitters; you are appointed as members of a DPO Ad Hoc Review Panel (DPO Panel) to review a DPO submitted by two U.S. Nuclear Regulatory Commission (NRC) employees.

The DPO (Enclosure 1) involves the FitzPatrick amendment concerning an alternate source term for calculating loss-of-coolant accident (LOCA) dose consequences. The DPO has been forwarded to Ms. Veil for consideration and issuance of a DPO Decision.

The DPO Panel plays a critical role in the success of the DPO Program. Your responsibilities for conducting the independent review and documenting your conclusions in a report are addressed in the handbook for MD 10.159 in Section II.F and Section II.G, respectively. The DPO Web site also includes helpful information, such as a Differing Views Best Practices Guide, tables with status information and timeliness goals for open DPO cases, and closed DPO case files (which include previous DPO panel reports). We will also send you additional information that should help you implement the DPO process.

CONTACT: Ian Gifford, OE (301) 287-9216

T. Clark, et al. 2 Timeliness is an important DPO Program objective. Thus, the disposition of this DPO should be considered an important and time sensitive activity. Although MD 10.159 identifies a timeliness goal of 75 calendar days for the DPO panel review and report and 21 additional calendar days for the issuance of a DPO Decision, the DPO Program also sets out to ensure that issues receive a thorough and independent review. Therefore, the overall timeliness goal will be based on the significance and complexity of the issues, schedule challenges, and the priority of other agency work. Process milestones and timeliness goals specific to this DPO will be discussed and established at a kick-off meeting.

Communication of expected timelines and status updates are important in the effectiveness and overall satisfaction with the Differing Views Program. If you need an extension beyond the timeliness goal, please send an e-mail to Mr. Wilson, Ms. Veil, the DPO submitters, and DPOPM.Resource@nrc.gov that includes the reason for the extension request and a proposed completion date.

An important aspect of our organizational culture includes maintaining an environment that encourages, supports, and respects differing views. As such, you should exercise discretion and treat this matter appropriately. To preserve privacy, minimize the effect on the work unit, and keep the focus on the issues, you should simply refer to the employees as the DPO submitters. Avoid conversations that could be perceived as hallway talk on the issue and refrain from behaviors that could be perceived as retaliatory or chilling to the DPO submitters or that could potentially create a chilled environment for others. It is appropriate for employees to discuss the details of the DPO with their co-workers as part of the evaluation; however, as with other predecisional processes, employees should not discuss details of the DPO outside the agency. If you have observed inappropriate behaviors, heard allegations of retaliation or harassment, or receive outside inquiries or requests for information, please notify the Office of Enforcement.

On an administrative note, please ensure that all DPO-related activities conducted by staff are charged to Activity Code ZG0007. Managers should report time to their Management/Supervisor Activity Code. Administrative Assistants should report time to their Secretary/Clerical Activity Code.

We appreciate your willingness to serve on the DPO Panel and your dedication to completing a thorough and objective review of this DPO. Successful resolution of the issues is important for the NRC and its stakeholders. If you have any questions or concerns, please feel free to contact me. We look forward to receiving your conclusions and recommendations.

Enclosures:

1. DPO-2021-001 Submittal

2. Process Milestones and Timeliness Goals cc: A. Veil, NRR J. Tappert, NRR R. Taylor, NRR M. King, NRR M. Blumberg, NRR M. Markley, NRR

T. Clark, et al. 3 G. Wilson, OE F. Peduzzi, OE D. Solorio, OE I. Gifford, OE

T. Clark, et al. 4

SUBJECT:

AD HOC REVIEW PANEL - DIFFERING PROFESSIONAL OPINION ASSOCIATED WITH THE FITZPATRICK AMENDMENT CONCERNING AN ALTERATE SOURCE TERM FOR CALCULATING LOSS-OF-COOLANT ACCIDENT DOSE CONSEQUENCES (DPO-2021-001)

DATE: 02/12/2021 Non-public ADAMS Package: ML21042B838 MEMO: ML21042B862 - ML21042B867 - ML21042B876 OE-011 OFFICE OE: DPO/PM OE: D NAME IGifford GWilson DATE 02/12/2021 02/12/2021 OFFICIAL RECORD COPY

Document 3: DPO Panel Report DPO-2021-001: JAMES A. FITZPATRICK NUCLEAR POWER PLANT -

ISSUANCE OF AMENDMENT NO. 338 RE: ALTERNATIVE SOURCE TERM FOR CALCULATING LOSS-OF-COOLANT ACCIDENT DOSE CONSEQUENCES DIFFERING PROFESSIONAL OPINION AD HOC REVIEW PANEL REPORT Authors:

• Theresa Clark

• Donnie Harrison

• Ann Marie Stone Dated:

June 9, 2021

EXECUTIVE

SUMMARY

On January 19, 2021, staff of the U.S. Nuclear Regulatory Commission submitted a differing professional opinion related to the recent issuance a license amendment for the James A.

FitzPatrick Nuclear Power Plant. This amendment adopted an alternative source term for accident analyses and increased the allowable amount of main steam isolation valve leakage.

On February 12, 2021, a review panel was established to review the issues raised by the submitters. The panel members work in other offices than the submitters and were not involved in the review of the FitzPatrick amendment.

The issues raised by the submitters are summarized in Section 2 of this report. Section 3 of this report provides background on the issues discussed in this report. Section 4 of this report presents the results of the panels review of the issues raised by the submitters. The panel substantiated several of the submitters findings, as summarized in Section 5 of this report and expanded upon in each subsection titled Findings. These conclusions resulted in several recommendations, as enumerated below and expanded in each subsection titled Recommendations.

1. The panel recommends that the staff request the FitzPatrick licensee to submit JAF-CALC 19-00005, Revision 1 on the docket. If appropriate, the staff could issue a correction to the July 2020 safety evaluation to add a reference to the docketed version, for greater public clarity.

2. The panel recommends that appropriate processes be followed to document the staffs conclusion on whether backfitting is warranted on this topic, as discussed in Section 4.2.2 of this report.

3. The panel recommends providing clarification in LIC-101 with respect to whether confirmatory analysis or evaluations performed by the staff in support of an amendment are considered working files meeting the criteria to be considered official agency records.

4. The panel also recommends reviewing and revising LIC-101 as needed to ensure that expectations for requesting updated documentation (i.e., revised versions of enclosures or analyses) are clear for future applications.

5. The panel recommends revising LIC-206 to reflect the proper use of risk insights to scope a review, but not to alleviate the need to pursue known flaws in a licensee analysis.

6. The panel recommends that trainingperhaps in conjunction with the LIC-206 revisionbe provided to all staff and management involved in licensing actions (i.e.,

technical reviewers, project managers, and supervisors) to emphasize the philosophical points mentioned in Section 4.3.5 of this report.

7. The panel recommends that the staff revise guidance for crediting the condenser leakage treatment pathway as part of alternative source term licensing actions to reflect any findings and recommendations of this report that are accepted by the decisionmaker, as discussed further in Section 4.4.4 of this report.

The panel appreciates the openness of all the staff and management approached during its review. These individuals accommodated meetings, answered all questions, provided i

supporting documents that illuminated the review, and universally explained their good regulatory intent in conducting the activities they discussed. Most importantly, the panel appreciates the determination of the submitters in bringing these issues to light through the agencys established processes for raising different views. They provided thorough information that gave the panel a good basis for beginning its review and were consistently open and accessible to further the panels work, even in a situation that could have felt uncomfortable.

Although the panel finds flaws in the staffs conclusions on the FitzPatrick review, it is with an understanding that the agency is transitioning to greater risk-informed decision-making. Given the panel members extensive risk-assessment experience, the panel has great sympathy for the desire to apply the agencys limited resources to issues of the highest safety significance.

The staff is currently underserved, however, by unclear messaging and guidance on exactly how and when to apply these risk insights. The agency can expect some growing pains as it navigates this path to becoming a modern risk-informed regulator, and the panels conclusion is that public health and safety remains assured in this instance despite the issues raised.

ii

TABLE OF CONTENTS

1. Introduction ....................................................................................................................... 1

2. Summary of Issues ........................................................................................................... 1 2.1. Issue 1: The licensee dose analysis does not demonstrate compliance with 10 CFR 50.67 1 2.2. Issue 2: Staff issued an amendment using an analysis not consistent with the licensees analysis .................................................................................................................. 2 2.3. Issue 3: The LOCA analysis credits structures systems and components not in technical specifications, inconsistent with 10 CFR 50.36........................................................ 3 2.4. Issue 4: Errors in the design basis could adversely impact future plant changes and regulatory actions .................................................................................................................... 3 2.5. Issue 5: The NRC staff analysis does not show compliance with 10 CFR 50.67........ 3

3. Background ....................................................................................................................... 3 3.1. Original accident source term framework ................................................................... 3 3.2. Alternative source term framework ............................................................................. 5 3.3. MSIV leakage control systems .................................................................................... 7 3.3.1. Initial requirements .............................................................................................. 7 3.3.2. Generic Issue ....................................................................................................... 7 3.3.3. Review of regulatory requirements ...................................................................... 8 3.3.4. Consideration of the condenser as an alternative MSIV leakage pathway .......... 9 3.4. Treatment of fission product concentration and deposition in alternative source term applications ........................................................................................................................... 12 3.4.1. Perry pilot ........................................................................................................... 12 3.4.2. Regulatory Issue Summary 2006-04 ................................................................. 12 3.4.3. Sandia analysis .................................................................................................. 13 3.4.4. Draft Revision to RG 1.183 ................................................................................ 13 3.4.5. Differing view on RG 1.183 ................................................................................ 14 3.5. 2019-2020 Exelon Alternative Source Term Applications ......................................... 15 3.6. Conduct of panels review ......................................................................................... 19

4. Results of Panel Review ................................................................................................. 20 i

4.1. Issue 1: Errors in licensee dose analysis .................................................................. 20 4.1.1. Volume in MSIV line is not homogenous and plate out is assumed .................. 20 4.1.2. Credible release paths not considered (stem/packing leaks) ............................ 22 4.1.3. Assumed dilution in turbine building, not ground release .................................. 23 4.1.4. Credited nonsafety-related components resulting in omission of release pathways impacting the atmospheric dispersion factors. ................................................ 25 4.1.5. Assumed worst case is not worst case .............................................................. 27 4.1.6. Removal coefficients and impact of aerosol particle size distribution ................ 29 4.1.7. Elemental iodine removal constant issues ......................................................... 35 4.1.8. Findings ............................................................................................................. 37 4.1.9. Recommendations ............................................................................................. 38 4.2. Issue 5: Cumulative effect of errors .......................................................................... 39 4.2.1. Findings ............................................................................................................. 39 4.2.2. Recommendations ............................................................................................. 40 4.3. Issue 2: Use of staff evaluation ................................................................................. 40 4.3.1. 10 CFR 50.67 Regulatory Requirement ............................................................ 40 4.3.2. Identification of technical/analytical issue with analysis of record...................... 41 4.3.3. Approach to resolving the issue with analysis of record .................................... 42 4.3.4. Findings ............................................................................................................. 45 4.3.5. Recommendations ............................................................................................. 47 4.4. Issue 3: Technical specifications .............................................................................. 47 4.4.1. Technical specification requirement .................................................................. 48 4.4.2. Structures, systems, and components relied upon for approval of license amendment ..................................................................................................................... 48 4.4.3. Findings ............................................................................................................. 49 4.4.4. Recommendations ............................................................................................. 50 4.5. Issue 4: Clarity of licensing and design basis ........................................................... 50 4.5.1. Clarity of analysis of record ............................................................................... 50 ii

4.5.2. Availability of analysis of record ......................................................................... 51 4.5.3. Accuracy of analysis of record ........................................................................... 51 4.5.4. Findings ............................................................................................................. 51 4.5.5. Recommendations ............................................................................................. 52

5. Conclusion and Recommendations ................................................................................ 52 iii

1. INTRODUCTION On January 19, 2021, staff of the U.S. Nuclear Regulatory Commission (NRC) submitted a differing professional opinion (DPO) related to the issuance of Amendment No. 338 1 for the James A. FitzPatrick Nuclear Power Plant (FitzPatrick). This amendment adopted an alternative source term for FitzPatricks accident analyses and increased the allowable amount of main steam isolation valve (MSIV) leakage.

On February 12, 2021, a DPO review panel was established to review the issues raised by the submitters. The panel members work in other offices than the submitters and were not involved in the review of the FitzPatrick amendment.

The issues raised by the submitters are summarized in Section 2 of this report. Section 3 of this report provides background on accident source terms, MSIV leakage control, treatment of fission product concentration in accident analyses, and the FitzPatrick review (along with other similar reviews). Section 4 of this report presents the results of the panels review of the issues raised by the submitters. The panel grouped these issues into three key areas: (1) the significance of errors that the submitters identified in the licensees analysis, (2) the appropriateness of using an NRC staff evaluation related to holdup of radionuclides in the power conversion system, and (3) the clarity of the licensing and design basis.

2.

SUMMARY

OF ISSUES The panel developed following summary of issues identified in the DPO. It served as the starting point for the scope to be addressed by the panel. On March 17, 2021, the submitters agreed that this summary accurately addressed their concerns.

2.1. Issue 1: The licensee dose analysis does not demonstrate compliance with 10 CFR 50.67 FitzPatricks requested NRC approval for adopting an alternative source term in accordance with 10 CFR 50.67. FitzPatrick stated that its loss-of-coolant-accident (LOCA) analysis for showing compliance with 10 CFR 50.67 followed the guidance in Regulatory Guide (RG) 1.183 2, NUREG-0800, Standard Review Plan (SRP) Section 15.0.1 3, and RG 1.194 4. However, the submitters contend the LOCA analysis does not follow this guidance and contains errors.

The submitters assert that, if the licensee had corrected these errors and followed the guidance, the actual control room dose would exceed the 5 rem total effective dose equivalent (TEDE) requirement in 10 CFR 50.67(b)(2)(iii). Specifically, the submitters consider the following aspects of the licensees analysis were in error and/or did not conform to guidance in the associated RGs and SRP.

• By not including the air space in the main steam line up to the closed MSIVs, the applicants analysis did not assume the radioactive release was instantaneously and homogenously mixed throughout all the free air volume in the drywell containment.

1 Agencywide Documents Access and Management System (ADAMS) Accession No. ML20140A070.

2 RG. 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Plants, dated July 2000. ADAMS Accession No. ML003716792.

3 SRP Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, dated July 2000. ADAMS Accession No. ML003734190.

4 RG 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants, issued June 2003. ADAMS Accession No. ML031530505.

1

Despite the licensees statements indicating that the containment elemental and particulate (natural) deposition/plateout is not credited, this deposition is credited in the space upstream of the MSIVs that is part of the drywell containment. (RG 1.183, Appendix A, Regulatory Position 3.1)

• The analysis did not consider credible release pathways that would increase the amount of radioactivity released to the environment. Concerns over these pathways were raised by the Advisory Committee on Reactor Safeguards during the initial licensing of certain facilities and were addressed by the MSIV leakage control system design. (RG 1.183, Appendix A, Regulatory Position 6.4 and 6.5)

• By assuming a release point at a specific location in the turbine building, rather than on the surface of the turbine building, dilution was credited in the turbine building. This dilution reduced the atmospheric dispersion factor, thereby improperly assessing the control room dose. (RG 1.183, Appendix A, Regulatory Position 6.4)

• By crediting nonsafety-related structures systems and components such as vents and doors that are not controlled by technical specifications, the licensee omitted release pathways that would be closer to the control room and would impact control room doses.

(RG 1.183, Regulatory Position 5.1.2 and RG 1.194, Regulatory Position 2.0)

• The analysis did not consider the limiting LOCA break location by selecting a recirculation line break, rather than a break in the reactor coolant system just prior to the MSIVs. In addition, when the licensee modeled the break just prior to the MSIVs (as a sensitivity study in response to the staffs request for additional information (RAI)), the licensee did not use the assumptions that were used in the LOCA analysis supporting the license amendment request (the licensees proposed design and licensing basis).

(RG 1.183, Appendix A)

• By using removal coefficients for aerosol settling that are nonconservative (based both on established science and a licensee reevaluation as part of a sensitivity study), and by incorrectly modeling the impact of changing aerosol particle size distribution as the radioactive material moves down the steam line, the LOCA model overestimates the deposition in the steam lines and, therefore, underestimates the dose in the control room. (RG 1.183, Regulatory Position 5.1.2 and RG 1.194, Regulatory Position 2.0)

• By using the elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis, the analysis modeled elemental removal greater than the SRP Section 6.5.2 5 limit on removal. This deviation is discussed in the NRC safety evaluation, but the safety evaluation does not assess the aggregate impacts of the deficiencies in the analysis described above. (RG 1.183, Appendix A, Regulatory Position 3.3 and SRP Section 6.5.2) 2.2. Issue 2: Staff issued an amendment using an analysis not consistent with the licensees analysis Alternative source term requirements in 10 CFR 50.67(b)(2) state that the NRC may issue an amendment only if the applicants analysis demonstrates with reasonable assurance compliance with the dose criteria in 10 CFR 50.67. The NRC staff review did not address the 5

SRP Section 6.5.2, Containment Spray as a Fission Product Cleanup System, Revision 4, dated March 2007. ADAMS Accession No. ML070190178.

2

concerns identified in Issue 1. Instead, the submitters assert that the NRCs approval was based on the licensees flawed analysis combined with risk and engineering insights that, in part, credit the capability of the power conversion system and main condenser to serve as a holdup volume for MSIV leakage. Because the licensees analysis of record does not credit transport of the MSIV leakage to the condenser or this holdup volume, the submitters contend that the NRC staff rationalized a basis for approval that was not based on the analysis in the licensees submittal.

2.3. Issue 3: The LOCA analysis credits structures systems and components not in technical specifications, inconsistent with 10 CFR 50.36 As noted in Issue 2, the safety evaluation for the FitzPatrick license amendment included staff-generated risk and engineering insights. Some of the systems and specific equipment alignments credited in the staffs and the licensees evaluation are not controlled by technical specifications. The submitters contend that this credit conflicts with technical specification requirements in 10 CFR 50.36(c)(2)(ii), Criterion 3 that requires limiting conditions for operation be established for structures, systems, and components, as well as alternative source term guidance in RG 1.183, that would assure these structures, systems, and components are available and consistent with the integrity assumed in the analysis.

2.4. Issue 4: Errors in the design basis could adversely impact future plant changes and regulatory actions As noted in Issue 1, the submitters identified several errors in the licensees analysis. As noted in Issue 2, the safety evaluation for the FitzPatrick license amendment included staff-generated risk and engineering insights and analysis. The submitters contend that the NRC staff did not follow SRP 15.0.1 and work out these differences to reconcile errors in the licensees analysis and or the factual accuracy of docketed information. Therefore, the submitters contend that the design and licensing basis of the FitzPatrick was modified and approved based on erroneous and incorrect information in a manner that will adversely impact future changes at the plant using 10 CFR 50.59, Changes, tests, and experiments, challenge inspectors in performing inspections and regulatory findings, and challenge licensing reviews for future license actions that may use it as a precedent.

2.5. Issue 5: The NRC staff analysis does not show compliance with 10 CFR 50.67 As noted in Issue 1, the submitters identified several errors in the licensees analysis. The submitters assert that the errors and uncertainties in the licensees analysis are so large that, even it was appropriate to use the staffs risk and engineering insights, the control room doses would exceed the regulatory values necessary for making a finding of reasonable assurance that 10 CFR 50.67 is met.

3. BACKGROUND 3.1. Original accident source term framework As part of the licensing review for a nuclear power plant, applicants must submit a safety analysis report that contains assessments of the radiological consequences of potential accidents and an evaluation of the proposed facility site. 6 The NRC (and its predecessor, the 6

Much of this background information is derived from the alternative source term final rule notice (64 FR 72001; December 23, 1999).

3

U.S. Atomic Energy Commission) used these assessments to determine that both the design and the site met applicable requirements in 10 CFR Parts 50 and 100, respectively.

Source term requirements were promulgated in 1962 in 10 CFR 100.11, Determination of exclusion area, low population zone, and population center distance. 7 In establishing these areas, applicants needed to assume the source term from a major accident that would result in substantial release of appreciable quantities of fission products from the core to the containment atmosphere. 8 The accident source term includes the composition and magnitude of the radioactive material, the chemical and physical properties of the material, and the timing of the release from the reactor core.

The design-basis accident was not intended to be an actual event sequence, but rather a surrogate to use in evaluating the performance of engineered safety features. The design-basis accident was intentionally conservative to address uncertainties in accident progression, fission product transport, and atmospheric dispersion. The resulting source term was used as a design parameter for accident mitigation features, equipment qualification, control room radiation doses, and post-accident vital area access doses. Dose assessments considered a whole-body dose (primarily from noble gases in the source term) and a thyroid dose (from inhaling radioiodines). The thyroid dose was generally limiting and was used to design engineered safety features such as the containment spray system, charcoal filters in containment, reactor building exhaust, and control room ventilation systems. The source term is also relevant to changes made by licensees under 10 CFR 50.59. 9 In the 1962 issuance of 10 CFR 100.11, Technical Information Document (TID) 14844 10 was referenced as a source of guidance. This source term represents a major accident involving significant core damage, typically in conjunction with a large LOCA. In addition, accident dose calculation methods for boiling-water reactors were outlined in RG 1.3. 11 These calculations were developed to be consistent with the TID-14844 source term and the dose guidelines in 10 CFR 100.11.

Key features of this original source term framework are:

• Three categories of radionuclides 7

27 FR 3509; April 12, 1962. The control room dose guidelines were later included in the 1971 issuance of the general design criteria 10 CFR Part 50, Appendix A, Criterion 19.

8 They also needed to consider the expected containment leak rate and the site meteorological conditions.

9 Criterion (c)(2)(iii) in 10 CFR 50.59 states that changes need a license amendment if they cause a more than minimal increase in the consequences of an accident previously evaluated in the final safety analysis report.

10 TID-14844, Calculation of Distance Factors for Power and Test Reactors, dated March 23, 1962.

ADAMS Accession No. ML021720780.

11 RG 1.3, Assumptions Used for Evaluating the Potential Radiological Consequences of a Loss of Coolant Accident for Boiling Water Reactors, first issued in 1970. Parallel methods for pressurized-water reactors were found in RG 1.4, Assumptions Used for Evaluating the Potential Radiological Consequences of a Loss of Coolant Accident for Pressurized Water Reactors. Today, both guides have been withdrawn, as the guidance has been updated and incorporated into RG 1.183, "Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors," and RG 1.195, "Methods and Assumptions for Evaluating Radiological Consequences of Design Basis Accidents at Light-Water Nuclear Power Reactors." All versions of these guides are available at https://www.nrc.gov/reading-rm/doc-collections/reg-guides/power-reactors/rg/index.html.

4

• Radioactive iodine in predominately elemental form, with small fractions of particulate and organic forms

• Immediate release of the source term to the containment at the start of the postulated accident

• Doses calculated at the exclusion area boundary for the first 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> and at the low population zone for the assumed 30-day duration of the accident 3.2. Alternative source term framework Major research efforts initiated after the accident Three Mile Island, Unit 2 (TMI-2) advanced the NRCs understanding of the timing, magnitude, and chemical form of fission product releases from severe reactor accidents. The culmination of these efforts was NUREG-1465 12, which defined revised, more realistic accident source terms for future light-water reactors. The NUREG-1465 source term differed from the original source term framework in several significant ways:

• Eight categories of radionuclides based on similarity of chemical behavior

• Radioactive iodine predominately in the form of aerosolized cesium iodide (CsI)

• Release over roughly 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> rather than instantaneously at time zero 13 Operating reactor licensees were not required to use this revised source term in their accident analyses, as the TID-14844 source term continued to be adequate to protect public health and safety. However, some licensees might wish to use this more realistic source term (later referred to as the alternative source term) in analyses to support operational flexibility and cost-beneficial licensing actions. For example, plant systems originally designed to actuate very rapidly to mitigate an instantaneous source term would not be required to perform under such stringent requirements. Systems designed to remove elemental iodine would be less important if the iodine were aerosolized. The NRC saw that some of these applications could improve overall safety and reduce overall occupational radiation exposure.

After a process of developing industry guidance 14, conducting NRC assessments 15, and reviewing pilot license amendments, the resulting requirements for licensees requesting to adopt the alternative source term were promulgated in 10 CFR 50.67, Accident source 12 NUREG-1465, Accident Source Terms for Light-Water Nuclear Power Plants, dated February 1995.

ADAMS Accession No. ML041040063.

13 Specifically, there were five release phases over several hours, with major core damage beginning at 30 minutes. In considering design-basis-accident analysis for evolutionary and passive light-water reactor designs, the NRC determined that only the first three phases (coolant, gap, and in-vessel) needed to be analyzed. The ex-vessel and late in-vessel releases could only result from core damage accidents with vessel failure and core-concrete interactions. (SECY-94-302, Source Term-Related Technical and Licensing Issues Pertaining to Evolutionary and Passive Light-Water-Reactor Designs, dated December 19, 1994. ADAMS Accession No. ML003708141.)

14 Electric Power Research Institute Technical Report TR-105909, Generic Framework for Application of Revised Accident Source Term to Operating Plants. Interim version dated November 1995 available at ADAMS Accession No. ML20094K909.

15 SECY-98-154, Results of the Revised (NUREG-1465) Source Term Rebaselining for Operating Reactors, dated June 30, 1998. Available at https://www.nrc.gov/reading-rm/doc-collections/commission/secys/1998/secy1998-154/1998-154scy.pdf.

5

term. 16 Licensees revising their source terms must evaluate the consequences of applicable design-basis accidents previously analyzed in the safety analysis report. The applicants analysis must demonstrate with reasonable assurance that:

• An individual located at any point on the boundary of the exclusion area for any 2-hour period following the onset of the postulated fission product release, would not receive a radiation dose in excess of 0.25 Sv (25 rem) TEDE.

• An individual located at any point on the outer boundary of the low population zone, who is exposed to the radioactive cloud resulting from the postulated fission product release (during the entire period of its passage), would not receive a radiation dose in excess of 0.25 Sv (25 rem) TEDE.

• Adequate radiation protection is provided to permit access to and occupancy of the control room under accident conditions without personnel receiving radiation exposures in excess of 0.05 Sv (5 rem) TEDE for the duration of the accident.

The NRC issued guidance for alternative source terms: RG 1.183 17 and SRP Section 15.0.1.

(The development of these documents was discussed in a 2000 Commission paper. 18) In addition to providing technical guidance on the use of the alternative source term, the guidance clarified that the NRC would approve such requests only if the facility, as modified, would continue to provide sufficient safety margins with adequate defense in depth to address unanticipated events and to compensate for uncertainties in accident progression and analysis assumptions and parameter inputs.

The guidance outlines pathways for full implementation and selective interpretation. Under a full implementation, all characteristics of the alternative source term are used, and at least the design-basis-accident LOCA needs to be analyzed. Other design-basis radiological analyses may be completed at the time, but all future analyses must use the alternative source term.

Under a selective implementation, one or more characteristics of the alternative source term may be implemented, or only a subset of design-basis radiological analyses may be updated.

For example, licensees might want to use the timing features of the alternative source term to increase the required closure time for containment isolation valves.

Section 5 of RG 1.183 introduces several important analysis assumptions and methodologies, including the following.

• Design-basis analyses provide a conservative set of assumptions to test the performance of the facility design, and the alternative source term provides an appropriate and prudent safety margin against unpredicted events and compensates for uncertainties. Licensees should be cautious in using data based on a specific accident sequence, because it may not be conservative for other sequences.

• Mitigating features can be credited if they are safety-related, required to be operable by technical specifications, powered by emergency power sources, and either automatically 16 64 FR 72001; December 23, 1999.

17 See note 11.

18 SECY-00-0156, Final Regulatory Guide 1.183 (Formerly DG-1081), Alternative Radiological Source Terms for Evaluating Design-Basis Accidents at Nuclear Power Plants, and Standard Review Plan Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, dated July 19, 2000. ADAMS Accession No. ML003728017.

6

actuated or with manual actions included in emergency procedures. The most limiting single active component failure should be assumed. Loss of offsite power occurrence and timing should be assumed to maximize radiological consequences.

3.3. MSIV leakage control systems 3.3.1. Initial requirements General Design Criterion (GDC) 54, Piping Systems Penetrating Containment, of 10 CFR Part 50, Appendix A, General Design Criteria, requires, in part, that piping systems penetrating primary containment be provided with leak detection, isolation, and containment capabilities having redundancy, reliability, and performance capabilities that reflect the importance to safety of isolating these piping systems.

As part of initial licensing, most boiling-water reactors (BWRs) were required to install systems to control leakage past the MSIVs in the event of a postulated design-basis LOCA. These systems were designed to collect and filter fission products to assure that doses would not exceed the 10 CFR Part 100 limits, as described in RG 1.96. 19 For example, the NRCs Advisory Committee on Reactor Safeguards recommended as part of its review of the FitzPatrick construction permit application 20 that the applicant study the effects of leakage through the MSIVs and measures to deal with such leakage.

3.3.2. Generic Issue These requirements were reconsidered under the NRCs Generic Safety Issue (GSI) program. 21 GSI C-8 incorporated several issues related to the MSIV leakage control system 22, as summarized below.

• NRC dose calculations in 1975 indicated that higher offsite doses could result by using the MSIV leakage control system, which would result in relatively little holdup time or cold-trapping of iodine and volatiles, compared to not using the system and maintaining the integrity of the steam lines and condenser. While the steam lines and condenser were not designed for the safe shutdown earthquake (SSE), the probability of failure was considered to be small.

• Operating experience indicated a relatively high failure rate of the MSIV leakage control system, with a variety of failure modes. 23 There was concern that some measured MSIV leakages were so high that a steam line break could have the consequences of a small 19 RG 1.96, Design of Main Steam Isolation Valve Leakage Control Systems for Boiling Water Reactor Nuclear Power Plants, dated June 1977. ADAMS Accession No. ML003740263.

20 Report on James A. FitzPatrickNuclear Power Plant, dated January 27, 1970. ADAMS Accession No. ML20248E410.

21 Item C-8, Main Steam Line Leakage Control Systems, as described at https://www.nrc.gov/sr0933/Section%202.%20Task%20Action%20Plan%20Items/c08r1.html.

22 Sometimes referred to as the MSIV leakage collection system.

23 Main Steam Isolation Valve (MSIV) Survey, dated July 1, 1982. ADAMS Accession No. ML19255D007 (available to the NRC staff but not publicly available in ADAMS at the time of this report).

7

LOCA with a simultaneous loss of containment. The NRC considered whether backfitting of these systems was necessary. 24 The NRC evaluated the costs and benefits of various potential changes and determined that they were not cost-effective or otherwise warranted. The regulatory analysis for the resolution of this issue was published in NUREG-1372 25 and related studies were documented in NUREG-1169 26 and NUREG/CR-5397. 27 NUREG-1169 summarized the use of a realistic fission product transport model to assess the offsite dose consequences of alternate means of treating MSIV leakage using nonsafety-related systems that could be available after a LOCA.

One of these approaches used the large volume of the main condenser to hold up the release of fission products leaking from the MSIV and down the main steam line, either through the turbine bypass valves or through the main steam line condensate drains. The report noted that the latter pathway would be completely passive at some plants because the drain valves fail open on a loss of power.

As part of the resolution of GSI C-8, the NRCs Office of Nuclear Regulatory Research recommended that any further generic work be continued as part of ongoing programs to eliminate unnecessary requirements and to update source terms. 28 3.3.3. Review of regulatory requirements In the 1980s, the NRC reviewed regulatory requirements to determine if deleting or modifying them could improve the effectiveness or efficiency of the NRC regulatory program for nuclear power plants without adversely affecting safety. 29 Pacific Northwest Laboratory (PNL) prepared a report to support the NRC staffs evaluation. 30 Industry representatives and NRC staff interviewed by PNL raised issues with the design of the MSIV leakage control systemspecifically, that radionuclides would be routed to the process gas control system and from there to the plant vent, rather than to the condenser. There was a suggestion that eliminating this system could reduce the risks of plant operation slightly. PNL concluded that there could be a marginal effect on risk if requirements were streamlined, and that greater than $1 million could be saved.

24 Request for Prioritization of BWR Main Steam Line Isolation Valve Leakage as a Generic Issue, dated July 30, 1982. ADAMS Accession No. ML19253F858 (available to the NRC staff but not publicly available in ADAMS at the time of this report).

25 NUREG-1372, Regulatory Analysis for the Resolution of Generic Issue C-8: Main Steam Isolation Valve Leakage and LCS Failure, U.S. Nuclear Regulatory Commission, June 1990. ADAMS Accession No. ML20044A428.

26 NUREG-1169, Resolution of Generic Issue C-8: An Evaluation of Boiling Water Reactor Main Steam Isolation Valve Leakage and the Effectiveness of Leakage Treatment Methods, dated August 1986.

ADAMS Accession No. ML20210N036.

27 NUREG/CR-5397, Value-Impact Analysis of Regulatory Options for Resolution of Generic Issue C-8:

MSIV Leakage and LCS Failure, dated May 1990. ADAMS Accession No. ML20044E672.

28 Resolution of Generic Safety Issue C-8, dated March 15, 1990. ADAMS Accession No. ML20086R032.

29 This program was introduced in an October 3, 1984, Federal Register (FR) notice (49 FR 39066).

30 NUREG/CR-4330, Identification of Regulatory Requirements that May Have Marginal Importance to Risk, dated April 1986. https://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr4330/index.html 8

3.3.4. Consideration of the condenser as an alternative MSIV leakage pathway In 1982, in response to MSIV leakage test failures, the Boiling Water Reactor Owners Group (BWROG) formed an MSIV Leakage Committee. This committee developed recommendations to minimize leakage and calculational methods to estimate dose rates due to MSIV leakage. 31 A follow-on MSIV Leakage Closure Committee was established in 1986. This group updated dose calculations and prepared typical licensing submittals that would justify eliminating technical specification requirements for the MSIV leakage control system and to modifying the MSIV leakage limit in technical specifications. The committee outlined these proposals in a July 1988 meeting with the NRC. 32 Making these changes, according to the committee, would reduce dose to maintenance personnel, reduce outage time, reduce repair costs, and extend the service life of the MSIVs.

In addition, the committee concluded that current plant designs provided highly reliable alternate treatment methods (compared to the MSIV leakage control system) through an isolated condenser with either the main steam bypass valves or the main steam drain valves open. It was best to have automatic alignment of this flow path to the condenser upon containment isolation, but the committee noted that some plants may administratively open either the bypass or drain line valves. Figure 1 provides a schematic of the MSIV leakage control system and condenser pathways.

31 These calculations were evaluated in NUREG-1169 (see note 26).

32 Summary of Meeting with BWR Owners Group MSIV Leakage Closure Committee, dated August 17, 1988. ADAMS Accession No. ML20207G480.

9

Figure 1. Schematic of the offsite dose pathways using the leakage control system and main condenser pathways, respectively. From BWROG presentation (see note 32).

The NRC staff in attendance at the meeting noted the significant fraction of overall core damage frequency that resulted from seismic events. Therefore, they saw a need to show that the non-10

following [an SSE]. The amount of reduction allowed will be evaluated on an individual case basis. References A-9 and A-10 provide guidance on acceptable models.

3.4. Treatment of fission product concentration and deposition in alternative source term applications 3.4.1. Perry pilot In 1996, the Perry Nuclear Power Plant 37 requested that the NRC approve three changes to its licensing basis: 38

1. removal of the MSIV leakage collection system (with reference to the prior NRC-led evaluation of this system)

2. increase of the allowable MSIV leakage ratea burden reduction that could also reduce occupational exposure from MSIV maintenance

3. change to emergency operating procedures to add containment pH control chemicals in the event of a LOCAa safety enhancement that would prevent iodine dissolved in the suppression pool from revolatilizing The NRC documented its assessment of the radiological consequences of this request in a report designated as AEB 98-03. 39 This report includes an evaluation of thermal hydraulics following a LOCA, based on both the licensees analysis and NRC staff confirmatory calculations using MELCOR analyses for the Grand Gulf Nuclear Station. 40 It also includes an NRC staff evaluation of fission product deposition using RADTRAD. The NRC staff considered spray scrubbing in the wetwell and natural deposition in the drywell, including natural deposition in the main steam line given the licensees proposal to remove the MSIV leakage control system.

With the support of these confirmatory calculations, the NRC approved the Perry amendment in March 1999. 41 Multiple additional applications followed the Perry model, including the Hope Creek application referenced in Section 3.3.4.

3.4.2. Regulatory Issue Summary 2006-04 The NRC later issued Regulatory Issue Summary (RIS) 2006-04 to identify issues that had arisen during the review of the first alternative source term applications. 42 The RIS clarified that licensees could reference AEB 98-03, but that the report was based on specific plant 37 Perry is a boiling-water reactor of General Electric Type 6 design (BWR-6) with a Mark III containment.

https://www.nrc.gov/info-finder/reactors/perr1.html 38 License Amendment Request: Revision of Main Steam Line Leakage Requirements and Elimination of the Main Steam Isolation Valve Leakage Control System, dated August 27, 1996. ADAMS Accession No. ML20117E782.

39 AEB 98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application using the Revised (NUREG-1465) Source Term, dated December 9, 1998. ADAMS Accession No. ML011230531.

40 Grand Gulf is of the same general reactor and containment design as Perry. https://www.nrc.gov/info-finder/reactors/gg1.html 41 Amendment No. 103, issued March 26, 1999. ADAMS Accession No. ML021840462.

42 RIS 2006-04, Experience with Implementation of Alternative Source Terms, dated March 7, 2006.

Available at https://www.nrc.gov/reading-rm/doc-collections/gen-comm/reg-issues/2006/ri200604.pdf.

12

parameters that determined a plant-specific removal rate constant. Licensees referencing the AEB 98-03 assumptions were advised to justify that they were applicable to their design.

The RIS also stated that the alternative source term is defined as a fission product release from the core into containment, without specifying the amount and composition of fission products in the reactor vessel or attached piping. Regulatory Position 6.0 in Appendix A to RG 1.183 allows licensees to treat fission product concentration in the drywell as representative of that near the MSIV. The RIS noted that some alternative source term amendment requests reduced drywell activity by assuming mixing with the free air volume of the wetwell, which could be acceptable if the licensee conducted analyses of the mixing mechanism.

The RIS went on to discuss particle size distributions, noting that aerosol deposition in piping may differ substantially from that in containment. Piping models were observed to include the volume between the reactor vessel and inboard MSIV, the volume between MSIVs, and the volume outside the outboard MSIV. Analyses should address different particle size distributions and settling velocities in each volume, as well as the effects of steam flow. The RIS also advised not to model deposition of gaseous iodine in piping unless appropriate justification is provided.

3.4.3. Sandia analysis Sandia National Laboratories conducted an analysis in 2008 43 to long-standing technical question as to the applicability of alternative source term. The report pointed out that, while the alternative source term assumes a specific source term release to containment, a realistic scenario would continue radiation release to the vessel, containment, and steam lines throughout fuel damage, vessel reflooding, and accident recovery. The analysts used the MELCOR code to model airborne radioactivity after core damage for recirculation line break and steam line break scenarios. The RADTRAD code was then used to develop dose estimates using site-specific dispersion values.

Sandia concluded that airborne concentrations in the vessel steam dome could significantly exceed the drywell airborne concentrations during the first 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of the accident. Since the vessel is the source of effluent that would leak through the MSIVs, Sandia asserted that use of the drywell concentration would be nonconservative and conceptually inaccurate when evaluating the dose from MSIV leakage. In addition, Sandia stated that the use of drywell sprays reduces containment pressure and thereby MSIV leakage flow rates. However, Sandia did not recommend crediting sprays for any reduction of the airborne concentration of radionuclides associated with MSIV leakage, given the distinction it drew between the containment and the vessel volumes. Sandia provided additional commentary on other parameters modeled in these analyses.

3.4.4. Draft Revision to RG 1.183 In 2009, the NRC issued a draft revision to RG 1.183 44, which added significant detail to the section on MSIV leakage modeling in BWRs, as well as revisions to the sections on release fractions and meteorology assumptions. One of the changes in this draft revision was to state, 43 SAND2008-6601, Analysis of Main Steam Isolation Valve Leakage in Design Basis Accidents Using MELCOR 1.8.6 and RADTRAD, issued October 2008. ADAMS Accession No. ML083180196.

44 DG-1199, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors. ADAMS Accession No. ML090960464. Published for comment on October 14, 1999 (74 FR 52822).

13

as suggested by Sandia, that the source of MSIV leakage is assumed to be the reactor vessel steam domenot the drywell as previously stated in RG 1.183, Revision 0. Section A-5 of the draft revision referenced the Sandia report as an acceptable model for estimating the radioactivity available for release via MSIV leakage.

The draft revision also added to the position on holdup in main steam piping and the condenser, with reference to the provision of emergency power sources, safety-grade piping segments, and either seismic Category I classification or with demonstrated seismic ruggedness.

Significant public comment was received on this draft guide and it was not issued as a final revision. A 2018 periodic review by the NRC staff concluded that a revision to RG 1.183 was still warranted. 45 3.4.5. Differing view on RG 1.183 In August 2020, an NRC staff member submitted a DPO regarding RG 1.183. DPO-2020-002 addressed three issues that were reviewed by an independent panel:

1. The regulation requires an analysis of a large release to an intact containment and does not specify an accident scenario, but RG 1.183 describes analysis assumptions for a LOCA which leads to: (1) an inappropriate need to impose non-physical assumptions pertaining to the operation of safety-related [structures, systems, and components] to mechanistically account for the [design-basis accident] source term; (2) inconsistency with LOCA analysis for 10 CFR 50.46; and (3) over-conservative assumptions intended to maximize fission product release (e.g., RG 1.183 assumptions on Mark Ill drywell to containment mixing).

2. Ongoing licensing issues resulting from the staff proffering physical explanations for the deterministic source term. For example, licensees have been instructed, via the [RAI]

process, to delay the mixing of the drywell and suppression pool free air volumes in any type of BWR (not just those with Mark Ill containments, as noted in RG 1.183, Appendix A, Assumption 3. 7) until 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> after accident initiation with respect to the containment leakage pathway.

3. Inconsistent treatment of crediting safety-related [structures, systems, and components]

within the deterministic radiological consequence analyses when determining compliance with the various [design-basis accident] radiological performance criteria.

For Issue 1, the DPO-2020-002 panel noted that imposition of other requirements pertaining to the distribution and mixing of the source term (e.g., RG 1.183, Appendix A, position 3. 7 pertaining to mixing within the containment) can lead to contradictory treatment of the operation of safety-related structures, systems, and components, driving the analysis beyond the design basis. The DPO-2020-002 panel concluded that:

imposition of a core damage source term provides sufficient defense-in-depth to assess, as noted in 10 CFR 50.34 for example, the safety features "engineered into the facility and those barriers that must be breached as a result of an accident before a release of radioactive material to the environment can occur." To defeat these same safety features for the purposes of maximizing offsite radiological dose appears contradictory to regulatory objective of assessing the capability of the plant design to 45 Results of Periodic Review of Regulatory Guide 1.183, dated June 25, 2018. ADAMS Accession No. ML18159A069.

14

mitigate an accident. Further, the imposition of assumptions that defeat safety features that can otherwise be credited can result in an arbitrary approach that lacks technical defensibility, particularly as our knowledge of severe-accident behavior continues to improve.

For Issue 2, the DPO-2020-002 panel noted that safety system actuation should be credited if the system meets proper requirements. The assumptions in the draft revision to RG 1.183 and the Sandia analysis were predicated on safety-related systems not operating during the 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> that the source term would need to develop; if they operated as designed, they would reflood the core and mix the released activity into the containment atmosphere. The DPO-2020-002 panel, however, found that staff did not instruct licensees to unwillingly adopt these modeling practices; while the Perry precedent may have been overly conservative, it was used correctly.

For Issue 3, the DPO-2020-002 panel supported the submitters proposal to overlay the alternative source term onto the designed operation of the facility during the design-basis accident. The panel also supported clarifying in guidance that provisions related to crediting safety-related structures, systems, and components should appropriately consider their assurance of safety functions and the design bases of the facility.

The DPO-2020-002 panel made several recommendations regarding crediting safety-related systems, structures, and components; applying precedent and guidance without unnecessary conservatism; and assessing the need for integrated analyses in certain scenarios. In February 2021, the Director of the Office of Nuclear Regulatory Research issued a decision, agreeing with the panel's conclusions and recommending that RG 1.183 and the SRP be updated accordingly.

Details on DPO-2020-002 (including background information developed by that panel that is relevant to this review) can be found in the associated case file. 46 3.5. 2019-2020 Exelon Alternative Source Term Applications In 2019, Exelon submitted license amendment requests for four of its nuclear power plants, 47 all related to the allowable amount of MSIV leakage. Three of these plants already had the alternative source term as part of their licensing basis through prior amendments; only FitzPatrick requested to adopt the alternative source term as part of this 2019 application.

During the NRC staffs review of the FitzPatrick license amendment request, several requests for additional information were issued by the NRC and answered by the licensee. 48

1. August 27, 2019: submittal of meteorological data for calculation of atmospheric dispersion factors (/Qs) and ARCON96 49 computer files 46 ADAMS Accession No. ML21067A645.

47 FitzPatrick, dated August 8, 2019 (ADAMS Accession No. ML19220A043); Quad Cities, dated March 5, 2019 (ADAMS Accession No. ML19064B369); Nine Mile Point, dated May 31, 2019 (ADAMS Accession No. ML19151A537); Dresden, dated October 21, 2019 (ADAMS Accession No. ML19294A304). Each application included supplements with various other dates, as noted in the safety evaluations (see note 53).

48 Licensee response letters dated August 27, 2019, January 16, 2020, and March 30, 2020 (ADAMS Accession Nos. ML19261A168, ML20017A052, and ML20090E279, respectively) 49 ARCON96 is a general code for assessing atmospheric relative concentrations in building wakes under a wide range of situations. See RG 1.194 (note 4) for more details.

15

2. January 16, 2020: responses to questions on environmental qualification and personnel dose estimates

3. March 30, 2020: responses to questions on the accident and dose analysis, including the aerosol removal issue mentioned above The March 2020 response resulted in a revision to the accident analysis of record, designated as JAF-CALC-19-00005, Revision 1. The results of this analysis are shown in Figure 2.

Figure 2. Revised offsite and onsite doses using the revised RADTRAD spray removal coefficient, as documented in JAF-CALC-19-0005, Revision 1. Page 3 of RAI response dated March 30, 2020 (ADAMS Accession No. ML20090E279). Note: ESF stands for engineered safety features and CR stands for control room.

Each NRC review included an evaluation of aerosol removal factors. To address this area of review, the licensee conducted sensitivity studies 50 to evaluate the impact of sprays on the aerosol settling velocity and to identify other inputs with well-defined uncertainty or conservatism that could be used to offset the uncertainty associated with the current aerosol deposition model.

The licensee stated that the conservatisms addressed by the sensitivity study were:

50 Quad Cities supplement dated March 31, 2020 (ADAMS Accession No. ML20091H576); Nine Mile Point supplement dated May 14, 2020 (ADAMS Accession No. ML20135G951); Dresden supplement dated May 6, 2020 (ADAMS Accession No. ML20127H890).

16

• Inclusion of all four main steamlines for holdup and deposition (as separately modeled nodes)

• More realistic control room operator breathing rate, using the RG 1.183 value for the first 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> but a light-intensity work breathing rate thereafter

• Aerosol impaction on the first closed MSIV, as had been previously approved for Nine Mile Point

• Condenser holdup and deposition, assuming that MSIV leakage travels to the condenser via the drain lines that attach to the main steam piping between the MSIVs In discussing the condenser credit, the licensee stated that components and piping systems typically used in this release path are sufficiently rugged to ensure they are capable of performing some level of radioactivity removal during and following [an SSE]. Thus, it is reasonable to assume that the condenser pathway could be made available for mitigating the consequences of MSIV leakage.

The results of the FitzPatrick sensitivity study are shown in Figure 3. The sensitivity cases that included credit for condenser deposition had results below the regulatory limit for TEDE (5 rem control room dose). 51 51 In Sensitivity 8, the licensee showed that including 50 cubic feet of the ruptured inboard pipe volume in the holdup model would reduce the result below the regulatory limit (to 4.99 rem). This is the only non-condenser case below 5 rem, and it was not discussed in the NRC safety evaluation. The analogous calculations for Quad Cities and Dresden also showed sensitivity results above the regulatory limit, but less significantly. The results for Nine Mile Point were below the regulatory limit for all sensitivity cases.

17

Figure 3. Sensitivity study results. Page 15 of RAI response dated March 30, 2020 (ADAMS Accession No. ML20090E279).

The NRC issued the FitzPatrick amendment on alternative source term and MSIV leakage in July 2020 52 and the other three plants amendments in June and October 2020. 53 In the FitzPatrick safety evaluation, the NRC staff made several key findings regarding these sensitivity cases, as discussed below.

• The NRC staff disagreed with the use of a reduced breathing rate. While a light-intensity breathing rate might be appropriate for normal working conditions, it should not be considered under 10 CFR 50.67 for determining radiation exposures from access to and occupancy of the [control room] under accident conditions, when [control room]

personnel may be expected to be at a higher level of stress and engaged in increased activities.

• The NRC staff disagreed with crediting MSIV impaction. The licensee referenced an approved analysis for Nine Mile Point as precedent. The NRC staffs approval in that case noted that the NRC staff does not generally endorse taking credit for impaction 52 Amendment No. 338, dated July 21, 2020. ADAMS Accession No. ML20140A070.

53 Quad Cities Amendment No. 281 and 277, dated June 26, 2020 (ADAMS Accession No. ML20150A328); Nine Mile Point Amendment No. 182, dated October 20, 2020 (ADAMS Accession No. ML20241A190); Dresden Amendment Nos. 272 and 265, dated October 23, 2020 (ADAMS Accession No. ML20265A240).

18

when modeling removal of particulates in main steam lines following a LOCA. The NRC found it only acceptable in that case.

• The NRC staff found it reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage. In a section on risk and engineering insights, the staff found it reasonable to conclude that the

[structures, systems, and components] in the [power conversion system] would be available following an SSE and that the likelihood of them being unavailable to serve as a volume for holdup and retention is very low. On this basis, the staff determined that there is high confidence that the [main steam lines] and the [power conversion system]

will be available for fission product dilution, holdup, and retention, especially at the seismic accelerations at a plants design-basis SSE. The NRC staff determined that these risk and engineering insights support its reasonable assurance finding based on its deterministic review. 54 3.6. Conduct of panels review As noted in Section 1 of this report, an independent DPO review panel was established in February 2021 to review the issues raised by the submitters. The panels review included the following activities.

• To understand the issues raised in the DPO, the panel met with the DPO decisionmaker (Andrea Veil, director of the Office of Nuclear Reactor Regulation) and her support staff, the DPO program leads, and the submitters.

• The panel developed the summary of issues (included as Section 2 of this report) and agreed on it with the submitters. This step was completed on March 17, 2021.

• The panel conducted interviews with staff and managers involved with the FitzPatrick review, the other Exelon amendment reviews, and the generic consideration of risk and engineering insights regarding holdup of fission products in the power conversion system. The panel interviewed 12 individuals (including the submitters) in March and April 2021. The panel also reinterviewed one of the submitters and one of the FitzPatrick reviewers to address specific technical follow-up questions.

• The panel obtained and reviewed numerous reference documents. Important references are included in this report as footnotes. In addition, the panel received interim drafts and calculation details from individuals involved with the FitzPatrick review. These documents informed the panels understanding of the decision-making process but were not the basis for the panels conclusions, and they are not referenced or included in this report.

• The panel coordinated extensively to develop the content and findings presented in this report.

The panel expects to discuss its findings and recommendations with the DPO decisionmaker after issuance of this report, following normal DPO processes.

54 Similar risk and engineering insights associated with the power conversion system were referenced in each of the other Exelon safety evaluations.

19

4. RESULTS OF PANEL REVIEW The following subsections provide the results of the panels review of the five issues identified by the submitters. Because the issues are interrelated, they are presented in a different order than included above in Section 2. The first pair of issues (Issues 1 and 5) have to do with errors identified in the licensees dose analysis and the significance of those errors with respect to the conclusions drawn by the NRC staff. The second pair of issues (Issues 2 and 3) have to do with whether using a staff evaluation in support of the license amendment approval was appropriate.

Finally, Issue 4 relates to the other issues and addresses whether the errors and staff evaluation resulted in an unclear licensing and design basis. The subsection titles below identify which of the submitters issues is being addressed in that subsection.

4.1. Issue 1: Errors in licensee dose analysis 4.1.1. Volume in MSIV line is not homogenous and plate out is assumed The submitters were concerned that by not including the air space in the main steam line up to the closed MSIVs, the applicant did not assume that the radioactive release was instantaneously and homogenously mixed throughout all the free air volume in the drywell containment. Despite the licensees statements indicating that the containment elemental and particulate (natural) deposition/plate out is not credited, this deposition is credited in the space upstream of the MSIVs that is part of the drywell containment.

The submitters contend the licensee was not in conformance with RG 1.183, Appendix A, Regulatory Position 3.1 since the licensee assumed a flow rate from the reactor pressure vessel to these volumes. If Regulatory Position 3.1 were followed, the radioactivity released from the fuel would instantaneously and homogeneously be mixed through the drywell, steam dome, and main steam line piping up to the MSIV containment isolation valves.

The panel did not substantiate this concern. Figure 4 from the licensees calculation shows how the licensee divided the main steam line piping for use in determining deposition. As stated in the license amendment request and safety evaluation, the licensee modeled two of the four main steam lines and provided a justification for this approach, which the staff accepted. The first main steam line (denoted as B) assumes the inboard MSIV fails to close. Volume 1 represents the volume of piping from the vessel to the closed outboard MSIV. Volume 2 represents the volume of piping from the closed outboard MSIV to the closed turbine stop valve (TSV). As stated in Regulatory Position 3.1, the radioactivity released from the fuel is assumed to mix instantaneously and homogeneously throughout the free air volume in the drywell. Since Volume 1 is part of the vessel/drywell atmosphere, at time = 0, the radioactivity in Volume 1 is the same as that assumed in the drywell. Seat leakage occurs at the closed outboard MSIV at a predetermined flow rate, based on the maximum allowable technical specification leakage rate for an individual MSIV, and as a result, the atmosphere from the vessel/drywell is drawn into Volume 1 at a designated flow rate. Similarly, main steam line (denoted as C) assumes the outboard MSIV fails to close. Volume 3 represents the volume of piping from the vessel to the closed inboard MSIV. Volume 4 represents the volume of piping from the closed inboard MSIV to the closed TSV. As stated in Regulatory Position 3.1, the radioactivity released from the fuel is assumed to mix instantaneously and homogeneously throughout the free air volume in the drywell. Since Volume 1 is part of the vessel/drywell atmosphere, at time = 0, the radioactivity in Volume 1 is the same as that assumed in the drywell. Seat leakage occurs at the closed outboard MSIV at a predetermined flow rate, based on the maximum allowable technical 20

specification leakage rate for an individual MSIV, and as a result, the atmosphere from the vessel/drywell is drawn into Volumes 1 and 3 at a designated flow rate.

Regulatory Position 6.3 allows crediting deposition and plate out on steam system piping upstream of the outboard MSIVs as evaluated on an individual case basis. The licensee stated conformance with Regulatory Position 6.3, citing their formulation for determining activity removal was based on that developed in AEB-98-03. The NRC staff did not challenge the licensee as to the assumptions regarding crediting deposition and plate out on steam system piping upstream of the outboard MSIVs.

The panel reviewed AEB-98-03 55, as well as the original Perry license amendment request 56 and NRC approval to better understand crediting of deposition. Perry has three isolation valves:

inboard MSIV, outboard MSIV, and main steam stop valve. The piping from the vessel to the main steam stop valve is seismically qualified. Perry relied only on deposition in piping upstream of the main steam stop valve and did not credit any piping that is not seismically qualified.

Because the calculations were not available for the panel to review, the panels conclusions are based on Appendix A of AEB-98-03 and a summary provided in the section titled In the Evaluation of Fission Procedure Deposition of AEB-98-03. The panel noted that natural deposition in the main steam line was modeled for the Perry plant, assuming the distance from the vessel to valves and a flow rate within the volumes. The panel did not attempt to validate the appropriateness of the flowrates assumed for FitzPatrick nor compare the values between sites.

Figure 4. Original Figure 3 from calculation JAF-CALC-19-00005, Revision 0.

55 See note 39.

56 Submittal dated August 27, 1996. ADAMS Accession No. ML20117E782.

21

4.1.2. Credible release paths not considered (stem/packing leaks)

The submitters were concerned the analysis did not consider credible release pathways that would increase the amount of radioactivity released to the environment. The submitters contend concerns over these pathways were raised by the Advisory Committee on Reactor Safeguards during the initial licensing of certain facilities and were addressed by the MSIV leakage control system design. Therefore, the licensee did not meet RG 1.183, Appendix A, Regulatory Positions 6.4 and 6.5.

As stated in Section 4.1.1, the assumed radioactive release pathway is from the vessel, through seat leakage from either of the closed inboard or outboard MSIV through the TSV. The licensee assumes piping downstream of the TSV is lost since the piping and turbine building are not capable of performing their safety function during and following an SSE. It is not clear whether the licensee assumed the TSV as open or closed (whereby the release would be through a leaking TSV). The licensee assumed this was the only or most limiting release pathway. In addition, as described in Section 4.1.1 and in accordance with Regulatory Position 6.5, the licensee assumed natural deposition in the piping between the MSIVs and TSV.

The panel agrees with the submitters that the licensee did not consider the most limiting or other available, viable release pathways in their analysis, nor did the staff question the licensees assumptions. Had the licensee assumed leakage up through the stem of the outboard MSIV instead of past the valve seat, deposition in the piping between the outboard MSIV and TSV could not be credited. This would have an impact on control room dose - not only due to the radioactivity but also the distance to the control room intake would be less.

As described in FitzPatrick Updated Final Safety Analysis Report (UFSAR) Section 9.19 57 (prior to the removal of the MSIV leakage control system) and shown in UFSAR Figure 4.6-3 58 (similar to Figure 5 below), stem leakage from the outboard isolation valve was processed through the MSIV leakage control system when the main steam pressure was less than 16 psig and manually initiated from the control room. The panel reviewed the procedures used for leak testing 59 the MSIVs and concluded the leakage collection system would not be in service during the test. With this system isolated, leakage up the stem could occur; therefore, the leakage determined by the surveillance test included leakage across the valve seat as well as up the stem. The panel noted the leak rate test cannot distinguish where the actual leak originates, making it possible that all leakage is through the stem pathway. The licensee did not assess this leakage path nor provide justification that leakage through this pathway would not have a detrimental impact on dose rates.

The panel noted the licensee has since isolated and capped off the piping from the MSIV leakage control system to the standby gas treatment system. 60 The panel also noted the licensee did not evaluate nor were they queried on other leakage pathways.

57 Section appears on page 174 of 190 of the PDF version of UFSAR Chapter 9. ADAMS Accession No. ML20038A354 [not publicly available].

58 Figure appears on page 71 of 126 of the PDF version of UFSAR Chapter 4. ADAMS Accession No. ML20038A349 [not publicly available].

59 ST-39B-X7B, Type C Leak Test Main Stem Line B MSIVs (IST) [not publicly available]

60 FP-97B Main Steam Isolation Valves, Leakage Collection Piping, Sheet 2, Note 6 states lines were capped per Engineering Change, EC 625093.

22

Figure 5. Typical BWR MSIV structure, from Figure 3.1 of NUREG/CR-6246, Effects of Aging and Service Wear on Main Steam Isolation Valves and Valve Operators, published March 1996. Available at https://www.osti.gov/servlets/purl/204274 (page 30 of 114 of PDF).

4.1.3. Assumed dilution in turbine building, not ground release The submitters were concerned that by assuming a release point at a specific location in the turbine building, rather than on the surface of the turbine building, dilution was credited in the turbine building. This dilution would have reduced the atmospheric dispersion factor, thereby reducing the calculated control room dose.

Section 2.3.4 of NUREG/CR-6331, Revision 1 61 states that the distance from the source to the receptor refers to the horizontal distance between the release point and the air intake.

61 NUREG/CR-6331, Revision 1, Atmospheric Relative Concentrations in Building Wakes, published May 1997. ADAMS Accession No. ML17213A187.

23

Section 3.4 of RG 1.194 62 affirms the source-to-receptor distance is the shortest horizontal distance between the release point and intake. It further states that for an area source such as building surfaces, the shortest horizontal distance from a wall to the control room intake is used whereas releases within building complexes, the shortest horizontal distance between the release point and the intake could be through intervening buildings. In these cases, it is acceptable to take the length of the shortest path (e.g., taut string length) around or over the intervening building as the source-to-receptor distance.

Section 3.2.4 of RG 1.194 states that some release sources may be better characterized as area sources, specifically postulated releases from the surface of a reactor or a secondary containment building. As such, these assessments typically used the shortest distance between the building surface and the control room intake and have treated the building as a point source.

It further states that diffuse source modeling should be used only for those situations in which the activity being released is homogeneously distributed throughout the building and when the assumed release rate from the building surface would be reasonably constant over the surface of the building.

Calculation JAF-CALC-19-00004, Revision 0 63 provided the bases for the atmospheric dispersion factors. Assumption 1 states the release point for MSIV leakage occurs at the TSVs, since the piping between the outboard MSIV and TSV is categorized as quality assurance Category I. The licensee assumes the piping downstream of the TSVs (including the condenser) and the turbine building itself are lost in this design bases scenario. In this scenario, the licensee also assumes the TSVs leak at a greater rate than the MSIVs. The licensee considered three other pathways as described in Section 4.1.4 below; however, the licensee concluded that using the TSVs as the release location to the environment was both conservative and appropriate.

In Section 3.2 of the safety evaluation, the staff noted that the licensee proposed new design-basis-accident atmospheric dispersion values (/Qs) for the new ground level MSIV leakage release pathway to the control room using the ARCON96 atmospheric dispersion model (NUREG/CR-6331). The staff acknowledged the licensee utilized a straight-line trajectory between the release points (TSVs) and receptor (control room) and that the licensee conservatively did not credit wake effects from obstructions within the turbine building. Previous design calculations 64 assumed the minimum distance to an equivalent circular area projection of the turbine building (diffuse source). However, the licensee believed a more realistic release location is at the TSVs, with a straight path to the control room intake.

In determining the distances, the licensee noted that the TSVs and intake to the control room ventilation system are roughly at the same y-coordinate such that only the x-coordinate distances are needed. 65 As illustrated below in Figure 6, the licensee determined the distance from the TSVs to the control room intake by adding the distances from the TSVs to Column 12 (turbine building edge).

62 See note 4.

63 JAF-CALC-19-00004, Control Room Atmospheric Dispersion for Turbine Building Release, Revision 0, was included as Attachment 9 to FitzPatrick license amendment request (ADAMS Accession No. ML19220A043), as referenced in note 47 (page 1056 of 1082 of PDF).

64 The licensee referenced JAF-CALC-RAD-04409 and JAF-CALC-RAD-00007. These calculations were not provided as part of the license amendment request.

65 For the distance to the Technical Support Center, the determined the distance of the hypotenuse formed by the x- and y-coordinates.

24

building roof; (3) a turbine building exhaust duct that is 60 from the control room intake. The vent from this duct is monitored for radiation and would be closed to prevent a radiological release to the environment. In the event of a loss of offsite power in combination with the LOCA, there would be considerable holdup in the turbine building. Therefore, the licensee concluded that using the TSVs as the release location to the environment was both conservative and appropriate.

It is important to determine an acceptable or realistic /Q because the safety-related related Control Room Emergency Ventilation Air Supply (CREVAS) System is not assumed to be operating until 30 minutes into the event when operators manually align the system. Specifically, during this design base accident, the control room intake will consist of unprocessed outside air at either the primary or secondary ventilation intakes until operators manually place CREVAS in the isolated mode of operation to prevent infiltration of contaminated air into the control room. At this point, outside air would be taken in at either the primary or secondary ventilation intakes and passed through one of the charcoal adsorber filter subsystems for removal of airborne radioactive particles.

The submitters were concerned that by crediting nonsafety-related structures systems and components such as vents and doors that are not controlled by technical specifications, the licensee omitted release pathways that would be closer to the control room and would impact control room doses. Crediting nonsafety-related structures, systems, and components is not consistent with RG 1.183, Regulatory Position 5.1.2 and RG 1.194, Regulatory Position 2.0.

RG 1.194 provides guidance for determining the atmospheric relative concentration values (/Q) when determining the design-basis control room radiological habitability. Regulatory Position 2.0 of this document states:

For each of the source-to-receptor combinations, the 95th percentile /Q should be determined. Values for parameters used as input to the /Q assessment should be selected consistent with achieving this confidence level. Selection of conservative, bounding source-to receptor combinations and less detailed site parameters for the /Q evaluation may be sufficient to establish compliance with regulatory guidelines.

Regulatory Position 3.2 states in part that it may be possible to identify bounding combinations in order to reduce the needed calculational effort. In determining the bounding combinations it will be necessary to consider the distance, direction, release mode, and height of the various release points to the environment in relation to the various control room intakes.

The panel noted the licensee dismissed the three pathways and did not establish why a release only at the TSVs was conservative or bounding such that a detailed evaluation of these pathways would not be necessary to establish compliance with the regulatory guidelines. This oversight or lack of justification could be viewed as taking credit for nonsafety-related structures, systems, and components as interpreted by the submitters.

The panel confirmed that the smoke ejector vents, security access door, and turbine building exhaust ducts are not safety-related and are not governed by technical specifications. For example, as stated in UFSAR Section 9.9.3.4 67, exhaust air from the turbine building operating floor and lower floors is discharged to the atmosphere by exhaust fans located at the 292-foot elevation. All exhaust air is discharged to the atmosphere through the building vent and is 67 Section appears on page 92 of 190 of the PDF version of UFSAR Chapter 9. ADAMS Accession No. ML20038A354 [not publicly available].

26

monitored by radiation monitors that, when set points are exceeded, automatically shut down the system supply and exhaust fans. Neither the verification of the setpoints nor the leak-tightness of this path is monitored nor measured. The panel did not see a limitation in place in the event a radiation monitor was incapable of functioning (shutting off a fan).

The panel agrees with the submitters that the licensee did not provide a bounding source-to-receptor combination; therefore, the atmospheric dispersion factors are not assured. The turbine building exhaust release location is closer to the control room intake than the distance assumed to be traversed in the environment when assuming the release is from the TSVs. This difference in distance will impact the calculated dispersion factor. The panel agrees with the submitters that by not following the RG 1.183 methods, the licensees analysis underestimated the atmospheric dispersion factor values; therefore, the LOCA doses estimated by the licensee would be larger.

4.1.5. Assumed worst case is not worst case The submitters were concerned that the accident analysis of record 68 does not maximize the postulated radiological consequences as provided for in RG 1.183. This is based on:

First, the licensees calculation provided in the response to the RAI is not consistent with the now approved analysis of record submitted as its new design basis. Four steam lines are considered rather than the two in the now approved analysis of record. Therefore, it is invalid for determining the impact of the licensees lack of modeling the break in the piping up to the first MSIV. When checking the impact of a questionable assumption, the impact of that assumption cannot be determined by making multiple changes that are inconsistentwith the proposed licensing basis. These other changes can obfuscate the impact of the issue being studied.

Secondly, while the NRC staffs [safety evaluation] concluded that the licensees analysis demonstrated that the impact of including a main steam line break does not significantly impact the dose consequences, this conclusion is not valid when the many other modeling issues in the LOCA analysis described in Section 2 [of this DPO] are considered. These issues and the factor discussed in the paragraph above mask the significance of not considering a main steam line break in the analysis.

In Section 7.5.2 of calculation JAF-CALC-19-00005, Revision 0, the licensee defines a MSIV failed line as the shortest steam line in which the inboard MSIV failed to close, resulting in a well-mixed volume boundary from the reactor pressure vessel nozzle to the outboard MSIV. The licensee further states that the line is not assumed to be breached as a breach (main steam line break upstream of the inboard MSIV during a LOCA) would be considered a different design-basis accident; therefore, not considered in their analysis. The staff questioned this position and noted that assuming a ruptured main steam line would maximize the dose contribution from MSIV leakage and would be consistent with the guidance from RG 1.183. As a result, the staff issued ARCB-RAI-3, requesting the licensee to provide additional information to justify not assuming a main steam line rupture in their analysis.

In its letter dated March 30, 2020, 69 the licensee responded by first stating that its analysis (JAF-CALC-19-00005, Revision 0) submitted with the license amendment request only modeled 68 JAF-CALC-19-00005, Revision 1 69 ADAMS Accession No. ML20090E279.

27

two main steam lines (B and C) and not four, as this assumption would result in less deposition, thus is more conservative. The licensee performed an evaluation to address the question posed in the RAI. The licensee stated this sensitivity evaluation also incorporated the changes associated with two other RAIs related to correcting fall height and spray flow assumptions. 70 The licensees results showed that the control room dose consequences was slightly lower when assuming a break in main steam line B, leakage past the MSIVs in the remaining three lines, and adjustments to the spray height and flow rates. The results are shown below in Table 1.

Table 1. Summary of dose estimates from licensee calculations.

Post-LOCA TEDE Dose (rem) low exclusion area population control Post-LOCA Release Pathway boundary zone room Original license amendment request: Two main steam lines, failed inboard MSIV in one 0.22 0.27 3.41 line JAF-CALC-19-00005, Revision 0 71 Base Case: Two main steam lines, failed inboard MSIV in one line with correction for fall 0.22 0.27 3.44 height and spray flow JAF-CALC-19-00005, Revision 1 72 Rupture in main steam line B, leakage in all four main steam lines 0.29 0.27 3.34 JAF-CALC-19-00005, Revision 1 (as manipulated) 73 In the safety evaluation, the NRC staff concluded the analysis of the dose consequences resulting from an assumed main steam line break provided by the licensee demonstrates that the impact of including a main steam line break does not significantly impact the dose consequences from MSIV leakage for FitzPatrick.

The panel agrees with the submitters that the analysis of record was based on leakage through two main steam lines while the evaluation for a ruptured main steam line was performed assuming leakage through four main steam lines. The table above shows the impact of correcting for fall height and spray flow, each assuming leakage through two main steam line.

The result was an increase in the control room dose from 3.41 rem to 3.44 rem. However, when assuming leakage through four main steam lines, the control room dose decreased when compared to the original calculation, from 3.41 rem to 3.34 rem.

The panel agrees, in general, that the impact of an assumption cannot be determined by making multiple changes to assumptions or methodologies in a proposed licensing basis. The panel agrees these changes can obscure the impact of the issue being studied. In the case 70 See Section 4.1.7 of this report for the analysis of fall height and containment spray flow. It should be noted that the licensee did not provide JAF-CALC-19-00005, Revision 1 nor the sensitivity evaluation associated with assuming a breached main steam line for staff review.

71 Attachment 1 to license amendment request, Section 3.11.15, as referenced in note 47 (page 29 of 1082 of PDF).

72 From the March 3, 2020 letter.

73 From the March 3, 2020 letter.

28

above, the impact of assuming a breach in main steam line B is not clear, had the licensee assumed leakage through two steam lines.

With respect to the validity of the dose consequence conclusions due to other modeling issues identified by the submitters, the panel addresses this issue in Section 4.2 of this report.

4.1.6. Removal coefficients and impact of aerosol particle size distribution The submitters were concerned that by using removal coefficients for aerosol settling that are nonconservative, and by incorrectly modeling the impact of changing aerosol particle size distribution as the radioactive material moves down the steam line, the LOCA model overestimates the deposition in the steam lines and, therefore, underestimates the dose in the control room. Specifically, the submitters asserted the licensee (1) treated aerosol removal by sprays and aerosol removal in the main steam lines as independent processes; (2) used a 20-group probabilistic settling velocity distribution for modeling aerosol gravitational settling in the main steam lines based on AEB 98-03; and (3) did not appropriately model the impact of changing aerosol particle size distributions as the radioactivity moves down the steam line. As a result, the new model proposed by FitzPatrick and as accepted by the NRC, overestimated the deposition in the steam lines and, therefore, does not accurately demonstrate compliance with 10 CFR 50.67.

Background

As described in Section 3.4.2, the NRC issued RIS 2006-04 to reduce the need for RAIs and help improve the planning for and implementation of an alternative source term. Section 2 of the RIS provides a discussion on the treatment of fission product deposition in the main steam line piping. It states that it is acceptable to reference AEB 98-03; however, the removal rate constant is site-specific, based on plant parameters. It further states that licensees should provide appropriate justification that the AEB 98-03 assumptions are applicable to their particular design. The RIS also acknowledges differences in aerosol particulate sizes, noting that a majority of large (i.e., heavier) particles will deposit in the inboard volume and that the distribution of the aerosol that leaks to the subsequent volume is smaller (i.e., lighter) particles. Lastly, it advises licensees that the choice of an effective settling velocity in any volume should account for the distribution of particle sizes in that volume.

As stated in Section 3.1.1.4 of the safety evaluation, the issue of how the change in the aerosol size due to drywell sprays would impact assumptions made in the subsequent main steam line aerosol deposition was discussed in the FitzPatrick preapplication meeting held on June 20, 2019. 74 As stated in the meeting summary, [s]ince the precedent cited for the proposed [main steam line] aerosol deposition did not include drywell sprays, the licensee should consider including a detailed discussion of how the use of sprays is accounted for in the subsequent steam line aerosol deposition.

In the license amendment request 75, the licensee stated that a 20-group probabilistic settling velocity distribution for MSIV leakage was implemented in the alternative source term LOCA dose analysis rather than using the AEB 98-03 single, median value, model. The licensee further stated that the 20-group probabilistic distribution methodology has been previously approved at Clinton Power Station, 76 Limerick Generating Station, 77 and LaSalle County Station. 78 74 Summarized in a meeting summary dated July 15, 2009. ADAMS Accession No. ML19183A128.

75 Attachment 1 to license amendment request, Section 3.11.11, as referenced in note 47 (page 25 of 1082 of PDF).

76 Amendment No. 167, issued September 19, 2005. ADAMS Accession No. ML052570461.

77 Amendment Nos. 185 and 146, issued August 23, 2006. ADAMS Accession No. ML062210214.

78 Amendment Nos. 197 and 184, issued September 6, 2010. ADAMS Accession No. ML101750625.

29

In Section 3.1.1.4 of the safety evaluation, the staff elaborates on the licensees referencing of the above amendments. Their concerns and conclusions are summarized below.

Clinton Originally, the Clinton licensee modeled three settling volumes with the same settling velocity in each. As a result, the staff was concerned the removal through aerosol settling was overestimated. Clinton changed the model to assume only two settling volumes in an unbroken main steam line and one settling volume for the broken main steam line. Clinton only assumed leakage through two of the four main steam lines. Additionally, Clinton calculated a weighted average for the aerosol settling velocity from the AEB-98-03 distribution, converted that average settling velocity to an effective aerosol filtration efficiency for each main steam line, and applied the applicable effective filtration efficiency to the leakage rate out of each main steam line. The settling area was assumed to be the projected horizontal area in the horizontal sections of the qualified main steam piping.

The staff acknowledged difficulty in determining how much deposition (i.e., settling velocity value) is appropriate and noted Clinton used a model based on the methodology of AEB-98-03, but applied additional conservatism to address the staffs comments. Based on the staffs sensitivity analysis, which found the total LOCA dose from all pathways acceptable, the staff concluded the main steam line aerosol settling model to be reasonable and appropriate.

Limerick The Limerick licensees modeling differs from Perrys in that Limerick credited five volumes including a pathway to the condenser whereas Perry assumed three volumes and did not credit the availability of the condenser. Similar to Clinton, the staff was concerned the removal through aerosol settling was overestimated, since the model assumed the same settling velocity on all five volumes. Limerick changed its model to combine penetration piping and downstream piping into a single outboard node. Additionally, Limerick used a 20-group probability distribution of settling velocities with efficiencies determined for each group and a net weighted average efficiency.

The staff acknowledged the licensee used a model based on the methodology of AEB-98-03 and included some additional conservatism to attempt to address the NRC staffs questions on the applicability of the AEB-98-03 methodology. The staff performed a sensitivity analysis.

Similar to the Clinton analysis, the staff found that the total LOCA dose from all pathways would continue to be acceptable and concluded the main steam line aerosol settling model to be reasonable and appropriate.

LaSalle LaSalles modeling of aerosol settling in the MSIV leakage pathway was different from that for Perry in that piping downstream from the outboard MSIV to and including the condenser was credited. Additionally, similar to Limerick, LaSalle used a 20-group probability distribution of settling velocities with efficiencies determined for each group and a net weighted average efficiency. The licensee used a model based on the methodology in AEB-98-03 and included some additional conservatism to address the staffs concerns about the applicability of the AEB-98-03 methodology to LaSalle. The 10th percentile aerosol settling velocity is a smaller value (and estimates less aerosol settling) than 90 percent of the calculated settling velocities in AEB-98-03. Based upon AEB-98-03, use of the 10th percentile settling velocity is more 30

conservative than the use of the median settling velocity noted as reasonable in AEB-98-03.

Given the aforementioned conservatism and the presence of a seismically qualified condenser, the staff found the LaSalle main steam line aerosol settling model to be reasonable and appropriate.

Application to FitzPatrick As stated in Section 3.1.1.4 of the FitzPatrick safety evaluation, the NRC staff acknowledged that aerosol settling was expected to occur in the main steam line piping. The NRC staff further acknowledged that the 20-group method was used in previous license amendment requests and were found to be acceptable by the staff. However, these previous evaluations did not credit aerosol removal from drywell sprays and the licensee did not provide a discussion of the impact of drywell sprays on the subsequent main steam line deposition. As such, the aerosol removal by sprays and aerosol removal in the main steam lines were handled as independent removal mechanisms. The staff acknowledged the interdependence of aerosol removal and the use of drywell spraysspecifically that larger aerosol particles in the containment atmosphere will be preferentially removed, making subsequent removal by deposition in downstream piping more challenging. The staff issued ARCB-RAI-2 requesting that FitzPatrick provide additional information describing how the gravitational settling credited in the main steam lines considered the changing aerosol characteristics (i.e., aerosol size and density distributions) due to the preferential removal of larger aerosols because of the credit assigned to containment sprays.

In its letter dated March 30, 2020, the licensee responded by submitting the results of a sensitivity analysis (introduced above in Section 3.5), which was intended to evaluate the impact of sprays on the aerosol settlingvelocity and the impacts on the analysis of including a ruptured steam line. The licensee identified a base case and then adjusted the base sensitivity case to adjust control room operator breathing rates, include aerosol impaction on the first closed MSIV, and credit holdup and deposition in the condenser. The licensee stated that its sensitivity analysis concluded that other identified considerations are sufficient to offset the uncertainty introduced by the drywell spray effects on the aerosol deposition model.

In Section 3.1.1.4.2 of the safety evaluation, the staff presented its analysis of the sensitivity analysis (as noted above in Section 3.5). A summary is presented below.

Base case: concern with assumed median diameter In the licensees base case, a simplified model was developed using first principles as identified in NUREG/CR-5966. 79 The licensee accounted for the different rates aerosols are removed by spray depending on their particle size by adjusting the spray removal rate by collection efficiency variation as provided in Figure 19 of NUREG/CR-5966. The licensee quantified the suspended aerosol mass for 20 distinct particle size groups based on a probability distribution.

Each particle size group was evaluated independently, not assuming agglomeration resulting from interaction of particles. In addition, the licensee assumed a 2-micron aerosol mass median diameter and geometric standard deviation of 2.0 particle size in the 20-group method to recalculate the aerosol removal rates in its sensitivity analysis.

As noted in Section 3.1.1.4.2 of the safety evaluation, the staff did not review and evaluate this assumption because: (1) no basis was provided for the assumption, (2) this assumption was not 79 NUREG/CR-5966, A Simplified Model of Aerosol Removal by Containment Sprays, published June 1993. ADAMS Accession No. ML063480542.

31

used in the licensees proposed analysis of record, and (3) it was not used by the NRC staff to determine reasonable assurance for complying with 10 CFR 50.67 for this particular license amendment request.

Control room operator breathing rate For its sensitivity analysis, the licensee used the RG 1.183 recommended for the first 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />, followed by a reduced value 80 of 3.28x10-4 m3/sec from 2 to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> and 3.06x10-4 m3/sec from 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> to 30 days. The staff disagreed with the licensees selection of a reduced breathing rate in the sensitivity analysis since, under accident conditions, control room personnel may be expected to be at a higher level of stress and engaged in increased activities. However, the staff concluded this factor has only a small effect, as the licensees sensitivity analysis shows a small effect of the breathing rate on the dose consequence (5 percent reduction).

Aerosol MSIV impaction In its letter dated March 30, 2020, the licensee referenced the Nine Mile Point 1 alternative source term LOCA licensing basis described in Nine Mile Point 1 alternative source term calculation H21C092, 81 which credits the phenomenon of impaction where some of the aerosol particles will be deposited on the first closed MSIV sealing surface. The safety evaluation associated with the Nine Mile Point, Unit 1 alternative source term 82 stated that the NRC staff does not generally endorse taking credit for impaction when modeling removal of particulates in main steam lines after a LOCA. The staff did not consider credit for MSIV impaction to be appropriate for use in the FitzPatrick MSIV leakage sensitivity.

Condenser holdup and deposition In its letter dated March 30, 2020, the licensee stated that aerosol holdup and deposition provided by the condenser is not modeled in JAF-CALC-19-00005 for FitzPatrick and that depending on the event scenario, multiple pathways could exist to route activity to the condenser, including the drain lines and the turbine itself.

Since the licensee did not credit the condenser in their proposed new licensing basis (and in fact assumed the turbine building and piping downstream of the TSVs was lost), the panel, in considering this error, did not credit any impact the condenser may have on dose consequences.

Panel review Based on the above, the panel concludes the licensee did not address the impact of sprays on the aerosol settling velocity since the staff appeared to have rejected the adjustments to the base case. The staff acknowledged that the base case was not intended to replace the accident analysis of record (revised analysis as a result of ARCB-RAI-1B); however, the staff then concluded that [t]he analysis of record indicates that dose consequences comply with all applicable dose acceptance criteria. The panel agrees with the submitters that the analysis of record does not address the interdependence of aerosol removal and the use of drywell sprays, 80 EPA/600/R-09/052F, Exposure Factors Handbook: 2011 Edition. Available at https://ofmpub.epa.gov/eims/eimscomm.getfile?p download id=522996.

81 ADAMS Accession No. ML070110240.

82 Amendment No. 194, dated December 19, 2007. ADAMS Accession No. ML073230603 32

specifically, larger aerosol particles in the containment atmosphere will be preferentially removed, making subsequent removal by deposition in downstream piping more challenging.

Impact As stated in the DPO, an example of the previous concerns with addressing the impact of sprays was discussed in the safety evaluation for the Nine Mile Point, Unit 2 alternative source term. 83 The submitters wrote:

In this SE, the NRC staff stated that to address historically documented NRC concerns about the use of AEB 98-03 and concerns about crediting drywell sprays, the NMP2 containment bypass model (1) limited the assumed credit for the aerosol settling velocity in the steam line (6.6E-5 meters/second for NMP2 versus 1.17E-3 meters per second for Perry (AEB 98-03, Revision 0)a reduction by a factor of 18), (2) limited the deposition credited to only one pipe segment between the closed isolation valves, and (3) did not credit deposition in a steam line with the assumed stuck-open valve. This is a clear example of how the NRC acknowledged that crediting drywell sprays will impact the aerosol deposition in the steam lines and that additional conservatisms are needed in the

[Nine Mile Point, Unit 2 license amendment request] to show compliance with 10 CFR 50.67.

By comparison, in the FitzPatrick model of record:

1. the assumed settling velocities ranged from 1.17E-04 meters/second to 3.48E-03 meter/second 84, imposing no limit as in Nine Mile Point Unit 2 analysis;

2. credit for deposition was used in all four pipes, essentially from the vessel to the TSVs; and

3. credit was assumed for the steam line with the assumed stuck-open valve.

In the DPO and reproduced below as Table 2, the submitters provided a table comparing the aerosol removal coefficients as presented in the license amendment request, a reevaluation performed by NRC staff, Revision 0 of AEB 98-03, response to the RAI, and SAND2008-6601. 85 (The table row including non-public information was excluded from this report.)

83 Amendment No. 125, issued May 29, 2008. ADAMS Accession No. ML081230439.

84 Table 4 of Calculation JAF-CALC-19-00005, Revision 0. The licensee did not submit Revision 1; however, the changes made in Revision 1 to account for spray flow and fall heights should not have changed the settling velocities.

85 See page 30 of the DPO for the footnotes referenced in Table 2.

33

Table 2. Comparison of aerosol removal coefficient models. (See page 30 of the DPO for notes 28, 29, 30, 32, and 33 shown in this table.)

Removal Coefficients or Lambdas (hr-1)

Removal Model Main Steam Line B Main Steam Line C Inboard Outboard Inboard Outboard Condenser FitzPatrick LAR28 (2020) 22.03 22.00 22.23 21.96 N/A29 NRCs Reevaluation30(2020) 11.27 11.11 11.40 9.93 N/A AEB 98-03, Rev.0 (1998) 9.04 9.04 9.04 9.04 N/A FitzPatrick RAI32 (2020) 0-0.33 hrs 0.402-0.598 0.176 0.402-0.598 0.176 0.0160 0.33-2 hrs 0.187-0.443 0.161 0.187-0.443 0.161 0.0102 2-24 hrs 0.158-0.342 0.109 0.158-0.342 0.109 0.0807 24-720 hrs 0.144-0.255 0.060 0.144-0.255 0.060 0.00629 SAND2008-660133 (2008) 0-2 hrs 0.034-2.9 1.3 0.0-2.9 1.3 0.02 2-12 hrs 0.0-1.8 1.0 0.0-1.8 1.0 0.18 12-720 hrs 0.0-1.0 0.7 0.0-1.0 0.7 0.12 As shown above, the now approved calculation of record (JAF-CALC-19-00005, Revision 1) estimated the removal coefficients in both credited steam line volumes to be approximately 22 hr-1 for 30 days.

This is over twice the value in the AEB 98-03, Revision 0, model that estimated these values to be about 9.0 hr-1 for the inboard and outboard steam line volumes. The difference correlates to the licensees model crediting approximately four times more radiation deposition than even the erroneous AEB 98-03, Revision 0.

More importantly, the FitzPatrick reanalysis analysis showed that the removal coefficients when considering drywell sprays were significantly less than in the now approved analysis of record LOCA that FitzPatrick used to justify the amendment. As shown above, the values projected by an NRC reevaluation 86 showed that the NRC reevaluation calculated the removal coefficients to be between approximately 9.9 hr-1 and 11.2 hr-1. The base sensitivity case in the licensees reanalysis resulted in an estimated control room dose of 7.35 rem. An estimated control room dose was not provided in the NRC reevaluation. However, considering the influence of the coefficients in the calculation of dose, the panel believes modifying the coefficient in the calculation of record to account for drywell sprays will likely result in a control room dose rate above 5 rem.

Regulatory Position 6.3 of RG 1.183 presents a position on crediting a reduction in airborne radioactivity in the containment by containment spray system. Footnote 1 states: The removal rate constants selected for use in design-basis calculations should be those that will maximize the dose consequences. The panel concludes the FitzPatrick main steam line removal coefficients as assumed in the calculation of record do not represent values that maximize the dose consequences, because the calculation of record does not incorporate the impact drywell spray on aerosol deposition.

86 The submitters requested support to estimate the proposed removal coefficients provided in FitzPatricks now approved analysis of record. The results were provided in two e-mails dated August 31, 2020, and November 10, 2020, from the NRC reviewer who wrote the safety evaluation section on FitzPatricks removal coefficients.

34

4.1.7. Elemental iodine removal constant issues The submitters stated that by using the elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis, the analysis modeled elemental removal greater than the SRP Section 6.5.2 87 limit on removal. This deviation is discussed in the NRC safety evaluation, but the safety evaluation does not assess the aggregate impacts of the deficiencies in the analysis described other sections of this DPO. The submitters concluded the analysis does not conform to RG 1.183, Appendix A, Regulatory Position 3.3 and SRP Section 6.5.2.

As stated in RG 1.183, Appendix A, Regulatory Position 3.3, a licensee may credit a reduction in airborne radioactivity in the containment that results from operation of containment spray systems, as long as such systems were designed and maintained in accordance with SRP Section 6.5.2. It further states that acceptable models for the removal of iodine and aerosols are described in SRP Section 6.5.2 and NUREG/CR-5966.

SRP Section 6.5.2 states that an applicant should identify differences between the design features, analytical techniques, and procedural measures proposed for its facility and the SRP acceptance criteria and evaluate how the proposed alternatives to the SRP acceptance criteria provide acceptable methods of compliance with the NRC regulations. For elemental iodine removal while spraying fresh solution, the SRP provides a formula for the first-order removal coefficient, s, assuming a well-mixed droplet model. The coefficient is a function of volume flow rate of the spray, containment building net free volume, and mass-mean diameter of the spray drops. The time it takes for a droplet to fall is estimated as the ratio of the average fall height to the terminal velocity of the mass-mean drop. A limit on the elemental iodine removal coefficient, s by spray is 20 hr-1.

In a similar manner, the first-order removal coefficient for particulates, p, can be estimated by:

where h is the spray drop fall height, V is the containment building net free volume, F is the spray flow, and E/D is the ratio of a dimensionless collection efficiency E to the average spray drop diameter D. Since the removal of particulate material chiefly depends on the relative sizes of the particles and the spray drops, it is convenient to combine parameters that cannot be known. It is conservative to assume E/D to be 10 per meter initially, then changing abruptly to 1 spray drop per meter after the aerosol mass has been depleted (decontamination factor is 50).

As stated in Section 7.2 of calculation JAF-CALC-19-00005, Revision 0, the licensee calculated the particulate aerosol removal coefficients to be 42.45 hr-1; however, the licensee chose to use 30.0 hr-1 in their analysis. The lower value is more conservative. The licensee also believed it was conservative to assume the same value for the elemental iodine removal coefficient.

The panel noted the staff was concerned that the calculation for determining particulate aerosol removal coefficient did not take into account obstructions in the drywell that would impact the spray drop fall height, h, and the containment spray flow rate, F. As such, the staff issued ARCB-RAI-1B which requested the licensee to provide justification for using 31 ft as h and 5,600 gpm as F, as these terms did not account for obstructions in the drywell.

In a letter dated March 30, 2020, the licensee responded that the 31 feet represented the difference between the lower spray header elevation (287 6) and the bottom of drywell 87 See note 5.

35

elevation (256 6). The licensee discussed methodologies used by Nine Mile Point Nuclear Station, Unit 1 88 and Oyster Creek Nuclear Generating Station 89, which made specific reductions based on obstructions in the drywell or blocked nozzles. The licensee stated that Nine Mile Point, Unit 1 used an average spray header elevation and the full design flow rate along with a 33.3-percent reduction in the fall height to account for obstructions in the drywell and a 33.3-percent reduction in the flow rate to account for potentially blocked nozzles. The 33.3-percent reduction in fall height to account for obstructions was based on 3D modeling of the drywell performed for Oyster Creek, and the 33.3-percent reduction in flow rate is based on Modular Accident Analysis Program analysis performed for Oyster Creek that showed that the design flow rate was lower than the actual flow rate that would be present. Nine Mile Point, Unit 1 used the same Oyster Creek assumption because it is expected that the obstructions would be similar for BWR Mark I containments. To account for obstructions, the FitzPatrick licensee revised calculation JAF-CALC-19-00005 (creating Revision 1) and included two multipliers of 0.67, resulting in a new value for the particulate aerosol spray removal coefficient equal to 26.36 hr-1. The licensee noted that this reduction in p (and thereby s), changed the control room dose from 4.64 to 4.67 rem with the MSIV contribution changing from 3.41 to 3.44 rem.

In Section 3.1.1.1.4 of the FitzPatrick safety evaluation, the staff provided an explanation of the licensee's response by stating the licensee referred to the methodology used by Nine Mile Point, Unit 1 and Oyster Creek, which made specific reductions in the spray removal coefficient calculation based on obstructions in the dry well or block nozzles that may impede flow. The staff then referenced the Nine Mile Point, Unit 1 safety evaluation and stated that to address the drywall construction congestion the licensee had multiplied the spray rate by 0.67 for additional conservativism and the fall height by 0.67 to account for drywall congestion.

The staff acknowledges this revision to the methodology caused a reduction in the assumed particulate spray removal coefficient from 30 hr-1 to 26.36 hr-1. The staff made a conclusion that the adjustments made to account for the presence of obstructions in the drywell were reasonable and therefore acceptable. The staff acknowledged that SRP Section 6.5.2 limits the elemental iodine removal constant to a value of 20 hr-1. It further stated that the staff performed confirmatory calculations using the RADTRAD program using 20 hr-1 for the elemental iodine removal constant and concluded the impact was not significant. In addition, the staff found this deviation from the SRP acceptable for this license amendment request and that, the staffs finding in this safety evaluation was only applicable to this license amendment request.

The panel requested to review the staffs confirmatory calculations which used 20 hr-1 for the elemental iodine removal. However, the staff did not retain this calculation and responded as stated in the safety evaluation, that the impact was not significant.

The panel reviewed the license amendment requests referenced by the licensee in their response to ARCB-RAI-1B and noted that while Nine Mile Point, Unit 1 and Oyster Creek did reduce the fall height and spray flow rate each by 33 percent, the FitzPatrick licensee incorrectly inferred FitzPatrick was consistent with the two plants. Specifically, the FitzPatrick licensee stated that Nine Mile Point, Unit 1 and Oyster Creek used an average fall height in their calculations. While Nine Mile Point, Unit 1 and Oyster Creek used the term average fall height in their calculation, the determination for these values is significantly different from the 88 ADAMS Accession No. ML070110231.

89 ADAMS Accession No. ML050940234.

36

methodology used by FitzPatrick, and this difference could have a significant impact when calculating the coefficients, as discussed below.

Nine Mile Point, Unit 1 90 The average fall height was obtained by determining the flow-weighted average by the summation of the flow through the header multiplied by the height of the header divided by the summation of the header flows. This plant has primary and secondary spray which use multiple spray headers but with different flows being delivered to each header. It should be noted this licensee limited the elemental iodine removal rate to 20 hr-1 in accordance with SRP Section 6.5.2.

The straightforward (non-flow-weighted) average fall height would have been more than 1.5 times greater than the calculated flow-weighted averages for the primary and secondary headers.

The licensee then multiplied the calculated flow-weighted averages by 0.67 to obtain the fall height used in the coefficient calculation.

Oyster Creek The average fall height was a function of the fraction of drops removed in each interval times the fall height for that fraction. Using this method for each interval gives an average fall height of 39.98 and 19.8 feet for the upper and lower headers, respectively. The licensee weighted these values to account for the differences in nozzles from each other. With 32 of the 88 nozzles at 39.9 ft, 14.5 ft was assumed for that elevation. Likewise, 56 of the 88 nozzles are at 19.8 ft for weighted height of 12.6. The Oyster Creek licensee added the weighted values to obtain an average fall height of 27.1 ft. Using the FitzPatrick methodology of ring header to floor, the average fall height would have been 41.21 ft.

Impact If similar methodologies as Nine Mile Point, Unit 1 and Oyster Creek were used for FitzPatrick, the average fall height for FitzPatrick would decrease. Changing this one parameter would cause the coefficient to decrease and would increase the calculated control room dose.

The panel concluded the licensee did not justify using the elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis. The panel agrees with the submitters that the analysis of record does not conform to RG 1.183, Appendix A, Regulatory Position 3.3 and SRP 6.5.2.

4.1.8. Findings The panel agrees with the submitters that the licensee did not consider the most limiting or other available, viable release pathways in their analysis, nor did the staff question the licensees assumptions. Had the licensee assumed leakage up the stem of the outboard MSIV, deposition in the piping between the outboard MSIV and TSV could not be credited. This would have an impact on control room dose - not only due to the radioactivity but also the distance to the control room intake would be less.

90 The calculation itself is non-public as it contains proprietary information.

37

The panel is neutral on whether it was appropriate to assume a point receptor (TSV to intake) or area receptor (turbine building wall to intake). However, the panel did note that the licensee changed its modeling of the release from that assumed in previous design calculations without justification. The staff did not question this change nor request a comparison between the atmospheric dispersion factor obtained through straight-line trajectory and assuming the minimum distance to an equivalent circular area projection of the turbine building.

The panel agrees with the submitters that the licensee did not provide a bounding source-to-receptor combination; therefore, the atmospheric dispersion factors are not assured. The turbine building exhaust release location is closer to the control room intake than the distance assumed to be traversed in the environment when assuming the release is from the TSVs. This difference in distance will impact the calculated dispersion factor. The panel agrees with the submitters that by not following the RG 1.183 methods, the licensees analysis underestimated the atmospheric dispersion factor values; therefore, the LOCA doses estimated by the licensee would be larger.

The panel agrees with the submitters that the analysis of record was based on leakage through two main steam lines while the evaluation for a ruptured main steam line was performed assuming leakage through four main steam lines. The table above shows the impact of correcting for fall height and spray flow, each assuming leakage through two main steam line.

The result was an increase in the control room dose from 3.41 rem to 3.44 rem. However, when assuming leakage through four main steam lines, the control room dose decreased when compared to the original calculation, from 3.41 rem to 3.34 rem.

The panel agrees, in general, that the impact of an assumption cannot be determined by making multiple changes to assumptions or methodologies in a proposed licensing basis. The panel agrees these changes can obscure the impact of the issue being studied. In the case above, the impact of assuming a breach in main steam line B is not clear, had the licensee assumed leakage through two steam lines.

The panel agrees with the submitters that the analysis of record does not address the interdependence of aerosol removal and the use of drywell sprays, specifically, larger aerosol particles in the containment atmosphere will be preferentially removed, making subsequent removal by deposition in downstream piping more challenging.

The panel concludes the FitzPatrick main steam line removal coefficients as assumed in the calculation of record do not represent values that maximize the dose consequences because the calculation of record does not incorporate the impact of drywell spray on aerosol deposition.

The panel concludes the licensee did not justify using the elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis. The panel agrees with the submitters that the analysis of record does not conform to RG 1.183, Appendix A, Regulatory Position 3.3 and SRP 6.5.2.

4.1.9. Recommendations LIC-101, 91 Appendix B, Section 4.0 states that [g]iven that the [safety evaluation] serves as the record of the staff's disposition of an application for amendment, the information relied upon in the [safety evaluation] and supplied by the licensee must be docketed and under oath or 91 LIC-101, License Amendment Review Procedures, Revision 6, dated July 31, 2020. ADAMS Accession No. ML19248C539.

38

affirmation. It further states [h]owever, if the information is important in the staff's decision-making process and is not otherwise in the public domain or reasonably inferred by the staff, it must be formally provided by the licensee. Section 4.5.5 of this report includes the panels recommendations on the clarity of the licensing basis.

In addition, LIC-101, Appendix B, Section 10.2 states that working files may meet criteria to be considered [official agency records]The written guidance associated with the license amendment process, such as this office instruction, clearly states that the basis and reasons for granting a license amendment must be contained in the SE issued with the license amendment.

The panel recommends providing clarification in LIC-101 with respect to whether confirmatory analysis or evaluations performed by the staff in support of an amendment are considered working files meeting the criteria to be considered official agency records.

4.2. Issue 5: Cumulative effect of errors The submitters were concerned that the errors and uncertainties they identified in the licensees analysis as described in Section 4.1 of this report are so large that, even it was appropriate to use the staffs risk and engineering insights, the control room doses would exceed the regulatory values necessary for making a finding of reasonable assurance that 10 CFR 50.67 is met.

The submitters performed an independent analysis to determine the impact of crediting the staffs risk and engineering insights. Specifically, the submitters revised the FitzPatricks now-approved LOCA analysis 92 to credit transport of MSIV leakage to the condenser and deposition and holdup of MSIV leakage in the condenser prior to it being released to the environment. As a first step, the submitters created a model that replicated the analysis of record, that is, incorporated licensee inputs and assumptions, including the radiological release at the TSVs.

The submitters analysis results matched the licensees control room dose from the MSIV leakage pathway within 1 percent.

The submitters revised their model to assume that the MSIV leakage travels to the condenser through the drain lines from the main steam line piping between the MSIVs as described in the licensees March 30, 2020, RAI response. The model was also revised to address the submitters concerns with (1) containment mixing and deposition [Section 4.1.1]; (2) leakage pathway to environment [Section 4.1.2]; and (3) turbine building dilution [Section 4.1.3]. As indicated in Table 3 presented on page 32 of the DPO, the calculated control room dose increased from 5 rem to 466 rem. Therefore, the submitters contended that even when the condenser is credited, the control room dose for the LOCA would exceed the 10 CFR 50.67 criterion of 5 rem TEDE.

4.2.1. Findings The panel did not attempt to validate the analysis performed by the submitters; however, based on the issues identified in Section 4.1 of this report, the panel believes the licensee did not provide sufficient information to demonstrate that their deterministic analysis complies with 10 CFR 50.67. Considering the very small margin available in the licensees analysis of record 92 Calculation JAF-CALC-19-00005, Revision 1 39

(4.67 rem), the panel believes the estimated control room dose is likely to exceed 5 rem during this assumed design-basis accident.

The panel did not find that there was an undue risk to public health and safety such that the adequate protection provisions of 10 CFR 50.109 need be invoked to require revisions to plant operation or procedures (e.g., use of the condenser). The panel also did not find that there were errors or omissions that would warrant use of the compliance-focused provisions of 10 CFR 50.109. The NRC staff clearly identified potential technical issues related to the calculations provided by the licensee, the licensee addressed these issues using sensitivity studies, and the NRC staff concluded in full awareness of these issues that the licensees request was acceptable. While the panel might have come to different conclusions, that is not the intent of compliance backfitting.

Within the scope of its review, the panel did not identify a substantial safety enhancement afforded by revising these calculations. These scenarios are bounding with inherent conservatisms. The safety impact of changing a calculated dose from such a scenarioeven by an order of magnitudewas not viewed by the panel to be substantial. However, the decision on backfitting is not within the scope of the panels review. For completeness, the staff could consider developing a backfitting evaluation 93 to ensure that the resolution of this potential backfit is fully documented.

4.2.2. Recommendations The panel recommends that appropriate processes be followed to document the staffs conclusion on whether backfitting is warranted on this topic.

4.3. Issue 2: Use of staff evaluation Alternative source term requirements in 10 CFR 50.67(b)(2) state that the NRC may issue an amendment only if the applicants analysis demonstrates with reasonable assurance compliance with the dose criteria in 10 CFR 50.67. The submitters assert that the NRC staff review did not fully address the concerns identified in Issue 1. Instead, the submitters assert that the NRCs approval was based on the licensees flawed analysis combined with risk and engineering insights that, in part, credit the capability of the power conversion system and main condenser to serve as a holdup volume for MSIV leakage. Because the licensees analysis of record does not credit transport of the MSIV leakage to the condenser or this holdup volume, the submitters contend that the staff rationalized a basis for approval that was not based on the analysis in the licensees submittal.

4.3.1. 10 CFR 50.67 Regulatory Requirement The regulatory requirements in 10 CFR 50.67 are straightforward and clear:

(b) Requirements.

(1) A licensee who seeks to revise its current accident source term in design basis radiological consequence analyses shall apply for a license amendment under § 50.90.

93 Backfitting needs to be considered consistent with Management Directive 8.4, Management of Backfitting, Forward Fitting, Issue Finality, and Information Requests, dated September 20, 2019.

ADAMS Accession No. ML18093B087.

40

The application shall contain an evaluation of the consequences of applicable design basis accidents1 previously analyzed in the safety analysis report.

(2) The NRC may issue the amendment only if the applicant's analysis demonstrates with reasonable assurance that:

(i) An individual located at any point on the boundary of the exclusion area for any 2-hour period following the onset of the postulated fission product release, would not receive a radiation dose in excess of 0.25 Sv (25 rem) total effective dose equivalent (TEDE).

(ii) An individual located at any point on the outer boundary of the low population zone, who is exposed to the radioactive cloud resulting from the postulated fission product release (during the entire period of its passage), would not receive a radiation dose in excess of 0.25 Sv (25 rem) total effective dose equivalent (TEDE).

(iii) Adequate radiation protection is provided to permit access to and occupancy of the control room under accident conditions without personnel receiving radiation exposures in excess of 0.05 Sv (5 rem) total effective dose equivalent (TEDE) for the duration of the accident.

Footnote 1 to 10 CFR 50.67 provides additional clarity regarding the use of design-basis accidents previously analyzed in the licensees safety analysis for the consequence analysis and that this hypothesized accident is intended to bound the consequences of any credible accident.

The fission product release assumed for these calculations should be based upon a major accident, hypothesized for purposes of design analyses or postulated from considerations of possible accidental events, that would result in potential hazards not exceeded by those from any accident considered credible. Such accidents have generally been assumed to result in substantial meltdown of the core with subsequent release of appreciable quantities of fission products.

4.3.2. Identification of technical/analytical issue with analysis of record Section 3.1.1.4, Assumptions on Main Steam Isolation Valve Leakage, of the FitzPatrick safety evaluation discusses a number of issues associated with the licensees analysis and provides some historical perspective regarding how these issues were addressed in previous alternative source term applications by other licensees. Toward the end of this section, the staff notes:

The issue of how the change in the aerosol size due to drywell sprays would impact assumptions made in the subsequent [main steam line] aerosol deposition was discussed in the FitzPatrick preapplication meeting held on June 20, 2019 (ADAMS Accession No. ML19183A128). As the staff stated in the meeting summary, Since the precedent cited for the proposed [main steam line] aerosol deposition did not include drywell sprays, the licensee should consider including a detailed discussion of how the use of sprays is accounted for in the subsequent steam line aerosol deposition.

The NRC staff acknowledges that aerosol settling is expected to occur in the main steamline piping. The NRC further acknowledges that the 20-group method was used in previous [license amendment requests] that were found to be acceptable by the staff.

However, since these previous evaluations did not credit aerosol removal from drywell sprays and the licensee did not provide a discussion of the impact of drywell sprays on the subsequent [main steam line] deposition, the NRC staff concluded that additional information was needed regarding the licensees MSIV leakage assessment.

41

In Section 3.1.1.4.1, Aerosol Removal by Sprays and in Main Steam Lines, of the FitzPatrick safety evaluation, the NRC staff noted that the licensee considered aerosol removal by containment sprays and aerosol removal in the main steam lines as independent removal mechanisms. The NRC staff also noted that, regardless of the specific removal mechanisms involved, larger aerosol particles in the containment atmosphere will be preferentially removed, making subsequent removal by deposition in downstream main steam line piping more challenging. To address this concern with the analysis of record, the NRC staff issued ARCB-RAI-2 requesting the licensee to provide additional information describing how the gravitational settling credited in the main steam lines considers the changing aerosol characteristics (i.e., aerosol size and density distributions) due to the preferential removal of larger aerosols by containment sprays.

4.3.3. Approach to resolving the issue with analysis of record In addressing this issue with their analysis of record, the licensee response to ARCB-RAI-2 includes a discussion of what the licensee viewed as conservatisms in its analysis and also presented a series of analyses addressing the staff-identified issue and these conservatisms.

The results of the licensees base analysis, which directly addresses the staff-identified issue regarding preferential aerosol removal, exceeds the regulatory dose limit to the control room.

The licensees subsequent analyses address their proposed conservatisms in the analysis.

These analyses lower the control room doses below the regulatory limit. However, as described in Section 3.1.1.4.1 of the FitzPatrick safety evaluation, the staff rejected the licensees perspective that these issues were appropriate to consider as conservatisms. The third conservatism involved the potential availability of the condenser pathway. The licensee indicated that this pathway could be made available without any discussion of how this action is achieved under the design-basis dose accident conditions, which assume the loss of piping downstream of the TSVs.

While this section of the FitzPatrick safety evaluation notes that the credit for the condenser pathway would lower the control room dose to an acceptable range, the safety evaluation does not have any discussion by the staff regarding what the licensee means by could be made available. Instead, the conclusion of this section discusses the disconnect in the design-basis paradigms between the core damage mitigation aspects that are implemented to ensure core damage is averted and the dose assessment aspects, which assumes core damage occurs.

This discussion regarding the different paradigms and approaches of the deterministic approach does not consider the regulations and RG 1.183 language that recognizes the bounding approach to the deterministic dose analysis.

As shown above, the footnote to 10 CFR 50.67 clearly states that the consequence/dose analysis is intended to be bounding of any credible accidents and is to be based on the design-basis accidents previously analyzed in the safety analysis report. Consistent with this bounding concept of the regulations, Section 3.11 of the FitzPatrick safety evaluation states:

The radiological consequences design-basis LOCA analysis is a deterministic evaluation based on the assumption of a major rupture of the primary reactor coolant system (RCS) piping. As applied in nuclear technology the deterministic approach generally deals with evaluating the safety of a nuclear power plant in terms of the consequences of a predetermined failure of the [emergency core cooling system] to provide adequate core cooling, which results in a significant amount of reactor core fuel damage as specified in RG 1.183. This general scenario does not represent any specific accident sequence but is representative of a class of severe damage incidents that was evaluated in the development of RG 1.183 source term characteristics.

42

This approach is somewhat analogous to the probabilistic risk assessment approach to grouping initiating events, in which initiating events that have similar impacts and challenges to a plant (and operational responses) are analyzed as a group. The analyses (e.g., thermal hydraulic, system success criteria, operator response timing) are performed based on a representative initiating event that is typically the most challenging. However, the frequency of the group of initiating events is the sum of all the events within the group, not just the representative event. This approach provides sufficient risk insights that are recognized as somewhat conservative; the analyst, however, need not model numerous initiating events and scenarios. In the deterministic design-basis approach, the LOCA accident scenario is the representative event that bounds the impacts and challenges of numerous other potential scenarios. Since this is a deterministic analysis (i.e., the scenario is predetermined, not derived) a specific frequency is not used. If a frequency were to be estimated, it would need to include all scenarios for which the LOCA accident scenario represents, not just the LOCA scenario itself. Any risk insights regarding the design-basis LOCA analysis would need to be careful not to solely consider the frequency of the design-basis LOCA. Instead, they would need to reflect the entire scope of events being represented by this bounding scenario.

Multiple NRC staff interviewed by the panel questioned the bounding approach of the LOCA design-basis analysis. This perspective is echoed a number of times in the FitzPatrick safety evaluation, including twice in Section 3.1.1.4.4, NRC Staff Evaluation of Licensees Sensitivity Analysis, where it states the probability of an event resulting in substantial fuel melt such as a LOCA followed by a significant unrelated seismic event is very low. This perspective does not recognize the intentional bounding aspect of the analysis and, similar to grouped events in probabilistic risk assessment, the frequency should not be characterized by the representative/bounding event, but rather all events that the LOCA scenario bounds. This frequency is likely orders of magnitude greater than considered by the staff reviewers in forming their risk insights and engineering judgments.

The NRC staffs discounting of the design-basis LOCA analysis is reflected in the NRC staff risk insights and engineering judgment that credits the condenser pathway to achieve acceptable control room doses. The staff, however, did not require that the condenser pathway be incorporated into the analysis of record. The staff also did not confirm relevant operational issues, such as:

• how the pathway is established under these accident conditions (e.g., whether the valves align to the condenser and isolate divergent pathways on emergency, safety-related power or whether they fail open or closed on a loss of offsite power)

• that the actions to align the pathway are incorporated into procedures

• that the pathway is incorporated into testing and maintenance programs to demonstrate and maintain this use of the condenser, which is different than its normal operational design purpose As a result, the staff conclusion does not address the implications of the licensee analysis, which needs deposition in the condenser to achieve an acceptable control room dose, or the actions needed to assure this pathway would be effectively implemented.

The panel notes that the 1999 BWROG topical report (discussed in Section 3.3.4 of this report) is already an NRC-approved approach for crediting the condenser pathway in an alternative source term license application. This NRC-approved approach includes conditions that need to 43

uncertainty in the fact that the FitzPatrick analysis of record does not address the aerosol removal aspects that led to the ARCB-RAI-2. It is also unclear in the safety evaluation what specific uncertainties are being addressed since all aspects of the licensee response to the RAI are addressed at some level; without credit for the condenser pathway, the licensees control room doses are unacceptable. This allusion to uncertainties appears to have been leveraged to resolve the issue identified by ARCB-RAI-2.

Further, in a section on risk and engineering insights, the staff found it reasonable to conclude that the [structures, systems, and components] in the [power conversion system] would be available following an SSE and that the likelihood of them being unavailable to serve as a volume for holdup and retention is very low. On this basis, the staff determined that there is high confidence that the [main steam lines] and the [power conversion system] will be available for fission product dilution, holdup, and retention, especially at the seismic accelerations at a plants design-basis SSE. The NRC staff determined that these risk and engineering insights support its reasonable assurance finding based on its deterministic review. 95 This position of the staff is almost entirely based on the inherent seismic robustness of the condenser and its associated piping. While there is operating experience that could be leveraged to support the inherent seismic robustness of the condenser and its piping, the staff position is not supported by past alternative source term application seismic walkdowns that identified the need for plant modifications to address a variety of seismic-related issues, such as block wall impacts on instrument lines that could cause a divergent release path away from the condenser, as well of other types of interactions and anchorage issues. This staff perspective also does not address what actions are needed to make the condenser pathway available, including the need to isolate potential divergent pathways, and what programs, procedures, and plant modifications (e.g.,

establishing emergency power sources to the pathway valves and boundary valves that must isolate) must be established to ensure this pathway will be maintained as a highly reliable mitigation. Only after these actions are addressed and taken by the licensee should the staff conclude there is high confidence that the pathway will be available and functional to warrant credit for the FitzPatrick alternative source term application.

In summary, the panel concludes that the approval of the FitzPatrick alternative source term license application is not based on the licensees submitted analysis of record, but rather is based on the staffs unconfirmed view of the availability of the condenser to mitigate the licensee-provided analysis that shows unacceptable control room doses.

4.3.4. Findings The panel agrees with the submitters that 10 CFR 50.67 clearly establishes that the staff can approve an alternative source term license application only if the licensees analysis demonstrates with reasonable assurance that the public and control room doses limits are met, which are also clearly established within 10 CFR 50.67. In addition, the DPO notes that even if the regulation was not this precise (i.e., other licensing actions where the regulatory language is not this clear), the expectation of the staff is that they review and engage the licensee based on the licensees submitted information and, while the staff can use confirmatory analyses and insights (including risk perspectives) to aid in their review, these staff analyses and insights cannot be substituted for the licensees analysis, especially when licensee information indicates unacceptable results. Consistent with its prime responsibility for safety, the licensee must provide an application that demonstrates the regulations are met. The staff reviewers are to 95 Similar risk and engineering insights associated with the power conversion system were referenced in each of the other Exelon safety evaluations.

45

verify the acceptability of the licensee submittal, not base the acceptance of a licensing action based primarily on their own perspectives and views.

The panel agrees with the submitters that the staff approval of the FitzPatrick alternative source term application is based on the staff perspective that the power conversion system (i.e.,

condenser pathway) would be available as a holdup volume to achieve acceptable control room doses (i.e., results do not meet the regulatory limits without this credit) when the staff-identified issue identified above associated with aerosol removal and drywell spray operations is properly addressed and that this credit to achieve acceptable control room doses is not incorporated into the licensees analysis of record nor are any programmatic elements included that would assure the availability and reliability of this pathway when needed. As discussed in Section 4.2 of this report, the panel concluded that the aggregate errors and issues identified in this report, if fully addressed by the licensee in a licensing basis analysis, could result in dose values that exceed the criteria in 10 CFR 50.67. In addition, if the RAI response base case had been used as the analysis of record, the dose criteria in 10 CFR 50.67 would not have been met.

The panel finds it acceptable for staff reviewers to draw generic conclusions about the inherent seismic robustness of the condenser and the associated piping and valves by considering operating experience and design aspects. The licensee need not necessarily perform detailed seismic design analyses to confirm the inherent seismic robustness of the condenser and its pathway given this experience. This staff perspective is adequately presented in the safety evaluation.

However, the panel disagrees with the NRC staff reviewers that the inherent seismic robustness of the pathway provides high confidence that a pathway to the condense will be available in and of itself. Past seismic walkdowns related to this type of application for some licensees have identified the need for physical plant modifications to ensure the pathway is not failed by seismic-related failure mechanisms, such as system and piping interactions and anchorage issues. Without confirming via a walkdown that the pathway will not be affected by these types of seismic-related failure mechanisms, the panel does not agree that the staff can have high confidence that the pathway if fully intact and functional.

The panel disagrees with the NRC staff reviewers perspective, presented numerous times in the safety evaluation, that there is high confidence that a pathway to the condenser would be available. While the NRC staff provided some detailed discussion and bases for the seismic robustness of the pathway, there is no justification for having that same high confidence that the pathway will actually be open or properly aligned (with or without an SSE). In fact, the licensees RAI response states the pathway could be made available without any discussion of what actions are needed to make it available. Since the staff did not pursue any clarification in this area, but rather declares the pathway would be available, the safety evaluation contains no information and there is no licensee commitment that would support the premise that the pathway will actually be available when needed. Without any specific considerations to support the availability of the pathway, the panel has no confidence in this operational aspect of the application. The panel notes that the NRC staff did not address the alignment and operational aspects given the occurrence of the initiating event and did not provide any discussion of the risk triplet. Consideration of this risk triplet would include attributes such as:

• What could go wrong: no safety-related power to open pathway or failed closed valves in this configuration

• How likely is it: expected, even at or below the SSE, there might be a loss of site power 46

• What are the consequences: unavailability of the pathway without the operating crew recognizing actions are needed to align the pathway, given that this is not the normal use of the system In addition, this condenser pathway would not be a preplanned approach to mitigate a well-understood core damage event. It would likely not be considered to have any confidence of being established under the accident conditions considered (including in a risk assessment) without licensee actions that are proceduralized, tested, and verified.

4.3.5. Recommendations The panels recommendations in this area stem from two fundamental regulatory concepts that appear to have been departed from during the FitzPatrick amendment review.

• The licensee has the primary responsibility for safety and must provide an application that demonstrates the regulations are met (especially when a regulation is as clear as 10 CFR 50.67 regarding licensee analysis). Therefore, staff reviewers must review the licensees submitted information and draw conclusions based on this information. While the staff can use confirmatory analyses and insights (including risk perspectives) to aid in their review, these staff analyses and insights cannot be substituted for the licensees analysis, especially when licensee information indicates unacceptable results. Staff reviewers should not base the acceptance of a licensing action based primarily on their own perspectives and views.

• The first element of risk-informed decision-making is to meet the regulations, unless the application is an exemption request. Therefore, risk insights and engineering judgments, while providing valuable insights on the scope and focus of the staffs review, cannot be used as a substitute for meeting the clear wording of the regulations.

The panel recommends that LIC-206 be clarified to reflect the proper use of risk insights to scope a review, but not to alleviate the need to pursue known flaws in a licensee analysis.

The panel recommends that trainingperhaps in conjunction with the LIC-206 revisionbe provided to all staff and management involved in licensing actions (i.e., technical reviewers, project managers, and supervisors) to emphasize the philosophical points mentioned above.

4.4. Issue 3: Technical specifications As discussed in Section 4.3 of this report, the safety evaluation for the FitzPatrick license amendment included staff-generated risk and engineering insights. Some of the systems and specific equipment alignments credited in the staffs and the licensees evaluation are not controlled by technical specifications. The submitters assert that this credit conflicts with technical specification requirements in 10 CFR 50.36(c)(2)(ii), Criterion 3, as well as alternative source term guidance in RG 1.183, that requires limiting conditions for operation be established for structures, systems, and components that would assure these are available and maintained consistent with the assumed performance of the analysis.

47

4.4.1. Technical specification requirement The regulation in 10 CFR 50.36 establishes the regulatory requirements for licensee technical specifications, specifically that:

The technical specifications will be derived from the analyses and evaluation included in the safety analysis report, and amendments thereto...

Further, the criteria in 10 CFR 50.36(c)(2)(ii) establish when a technical specification limiting condition of operation must be provided, including:

A structure, system, or component that is part of the primary success path and which functions or actuates to mitigate a design basis accident or transient that either assumes the failure of or presents a challenge to the integrity of a fission product barrier.

Further, 10 CFR 50.36(c)(3) establishes surveillance requirements to assure the quality of these structures, systems, and components are maintained and that the limiting conditions of operation will be met.

In addition to the regulatory requirements above, RG 1.183, Section 5.1.2, Credit for Engineered Safeguards Features, specifically states:

Credit may be taken for accident mitigation features that are classified as safety-related, are required to be operable by technical specifications, are powered by emergency power sources, and are either automatically actuated or, in limited cases, have actuation requirements explicitly addressed in emergency operating procedures. The single active component failure that results in the most limiting radiological consequences should be assumed. Assumptions regarding the occurrence and timing of a loss of offsite power should be selected with the objective of maximizing the postulated radiological consequences.

As discussed in Section 4.3 of this report, the panel agrees with the submitters that the approval of the FitzPatrick alternative source term license amendment is based on the availability of the condenser leakage treatment pathway. However, the staff reviewers credited this pathway based on their own views that the condenser pathway is seismically robust. The staff did not engage the licensee to ensure this credit is incorporated into the analysis of record or to confirm what automatic or manual actions would be necessary to make the pathway available under the design-basis dose analysis accident conditions. As a result, the licensee did not propose, and the staff did not pursue, any technical specifications, emergency operating procedures, or any other type of commitment, condition, or action to assure the availability and reliability of this pathway.

4.4.2. Structures, systems, and components relied upon for approval of license amendment As described in Section 4.3 of this report, there already exists an NRC-accepted BWROG topical report approach for crediting the condenser leakage treatment pathway for alternative source term license applications. Based on the staff review at the time, the safety evaluation report for the topical report required plant-specific seismic analyses because the staff did not have familiarity and confidence in the late-1990s-era methods being employed by the industry to generically demonstrate the seismic robustness of the condenser and the associated piping.

Based on staff interviews and the preapplication meeting summary, it is clear that the licensee did not want to implement this topical report approach primarily due to the analytical costs 48

associated with validating the seismic capability of the non-safety-related condenser and associated piping.

As part of the licensees response to ARCB-RAI-2, the licensee states that crediting the condenser pathway would lower the control room doses to less than the regulatory limit and that this pathway could be made available. The staff reviewed seismic operating experience to reach a conclusion that the condenser and its piping were seismically robustness. In numerous places throughout the FitzPatrick safety evaluation, the staff refers to having high confidence that the condenser would be available and often refers to its inherent seismic robustness. It appears that the staff connected the inherent seismic robustness of the condenser and its piping with it being available, even though the licensee did not take this position or provide any information on how the pathway could be made available.

In assuming that the condenser leakage treatment pathway would be available, the NRC staff made a number of unidentified assumptions about the availability of the pathway beyond just the inherent seismic robustness of the piping. While the condenser and its piping may have inherent seismic robustness, a review by the panel of past alternative source term applications that credited the condenser pathway identified a number of seismic-related and non-seismic-related modifications that had to be implemented to ensure the pathway (including the isolation of boundary/interface piping) would be available when needed. These have included block wall issues, anchorage issues, emergency power or manual capabilities to operate valves to open the pathway, and similar capabilities to close other valves to ensure the leakage goes to the condenser and is not diverted through a non-credited pathway. None of these past recognized issues were considered or addressed by the staff in determining that the condenser leakage treatment pathway would be available.

The NRC staff clearly recognized that credit for the condenser leakage treatment pathway was necessary to achieve acceptable control room dose results to resolve the technical/analytical issue identified with the licensees analysis of record, as discussed in Section 4.3 of this report.

A typical review process would have included engaging the licensee on how they would address the fact that their own analysis indicates that control room doses would not meet the regulatory limits without crediting the condenser. This engagement may have resulted in the pathway being incorporated into the licensees analysis of record, if the licensee could not provide an alternative approach to addressing the issue. If the condenser leakage treatment pathway had been incorporated into the licensees analysis of record, then it would have necessitated that the condenser leakage treatment pathway be controlled appropriately to assure its functionality when needed. The staff did not pursue this issue to closure by the licensee, meaning that this functionality is not assured.

The BWROG topical report does not address the specific programs the licensee needs to implement to assure the pathway is reliable and available when needed. A search of past applications revealed that some licensees created new technical specification action statements and surveillance frequencies, and some incorporated the valves into their inservice test programs, as well as appropriate emergency procedures.

4.4.3. Findings The panel agrees with the submitters that the programmatic aspects of ensuring the availability, reliability, and performance of the condenser leakage treatment pathway, including boundary valve isolation, should have been identified and managed via established licensee programs consistent with previous license amendments that utilized the NRC-approved BWROG topical report NEDC-31858P. This approach would be consistent with 10 CFR 50.36 49

and the specific guidance in RG 1.183. The panel does not take a position on whether the implemented programs must specifically include technical specifications, or if the staff might determine that other licensee-controlled programs achieve the desired level of assurance that the pathway components will be reliable and available when needed. That decision would be part of an NRC review. However, in the case of FitzPatrick, the panel notes that the components associated with this staff-credited pathway to achieve acceptable control room doses were not required to be included in any licensee program (e.g., technical specifications, surveillance testing, inservice testing, emergency operating procedures, maintenance programs) and as such their reliability and availability in the context of this application cannot be assured.

4.4.4. Recommendations The panel recommends that the staff revise guidance for crediting the condenser leakage treatment pathway as part of alternative source term licensing actions to reflect any findings and recommendations of this DPO report that are accepted by the DPO decisionmaker. In particular, the panel recommends including specific references to the operational provisions of the NRC-approved BWROG topical report and RG 1.183 (e.g., valves that must change position are supplied by emergency power sources). The panel does not view it necessary given operating experience that licensees perform analyses related to the inherent seismic robustness of the condenser and the associated piping. The revised guidance should provide for walkdowns or other appropriate methods, as well as programmatic controls, to initially validate and then subsequently assure the availability of credited components.

The panels recommendations on regulatory philosophy included in Section 4.3.5 above apply to this section as well.

4.5. Issue 4: Clarity of licensing and design basis The submitters contended that the NRC staff did not reconcile errors in the licensees analysis and or the factual accuracy of docketed information. Therefore, the submitters contended that the design and licensing basis of the FitzPatrick was modified and approved based on erroneous and incorrect information in a manner that will adversely impact future changes at the plant using 10 CFR 50.59, challenge inspectors in performing inspections and regulatory findings, and challenge licensing reviews for future license actions that may use it as a precedent.

4.5.1. Clarity of analysis of record The original alternative source term analysis associated with the FitzPatrick application was clear: JAF-CALC-19-00005, Revision 0. The licensee stated in its submittal that this was the calculation upon which this [license amendment request] is based.

The March 2020 RAI response 96 refers to a revised analysis (JAF-CALC-19-00005, Revision 1).

The licensee, however, did not provide any explicit statements that the analysis of record associated with the amendment request was being updated.

Page 33 of the FitzPatrick safety evaluation states: the accident analysis of record is the revised analysis discussed in the letter dated March 30, 2020, in response to the RAI 96 ADAMS Accession No. ML20090E279.

50

concerning obstructions (ARCB-RAI-1B). The analysis of record indicates that dose consequences comply with all applicable dose acceptance criteria.

4.5.2. Availability of analysis of record JAF-CALC-19-00005, Revision 0 was submitted as Attachment 6 to the August 2019 submittal. 97 The licensee did not submit Revision 1 as part of the March 2020 RAI response.

4.5.3. Accuracy of analysis of record The panels review of potential errors within the analysis of record is summarized above in Section 4.1. Consistent with Section 4.2 of this report, the panel is not able to determine the aggregate impact of errors on the licensees analytical conclusions. These issues were not pursued during the license amendment review, and the licensees approach was determined to be an acceptable method for the alternative source term. Calculational changes could theoretically be made that would compensate for these errors and demonstrate that accident doses would be within regulatory limits. However, the licensees analysis conducted in response to an RAI included results above regulatory limits, and correcting some of the errors would also increase the results. The licensee might address these errorsto the extent allowed under 10 CFR 50.59when applying this method to other accident analyses. The NRCs ability to require licensee action at this stage, however, is limited by the regulatory requirements of 10 CFR 50.109, Backfitting.

4.5.4. Findings The panel agrees in part with the submitters. While the basis for the NRCs approval is technically clearthe NRC staff specifically references the analysis of record as JAF-CALC-19-00005, Revision 1the documentation lacks regulatory rigor.

This revised calculation was not submitted on the docket, as Revision 0 was, so the details of the calculation are not readily available to inspectors, license reviewers, or members of the public. Presumably the changes in Revision 1 are associated with the change in spray removal coefficients (26.36 hr-1 for a decontamination factor less than or equal to 50 and 2.636 hr-1 thereafter for the remaining duration of spray operation), but this could not be verified.

While licensees may not be required to submit full calculations to support their amendment reviews, it is reasonable for the NRC staff to request a description of all changes to an analysis of record that were made during the review of an amendment request. The NRC staff should then summarize this information in its safety evaluation. This paper trail will support future NRC inspectors in understanding exactly what was approved at the time, and on what basis.

Because FitzPatrick was approved for full alternative source term implementation, this calculation becomes the foundation for all future analysis. Any errors in the LOCA analysis of record, if not corrected by the licensee, would be carried forward into other accident analyses without any NRC review. As discussed in Section 4.2.2 above, the NRC would need to invoke its processes under 10 CFR 50.109 to raise any issues with analyses conducted under this framework.

97 Page 230 of 1082 in ADAMS Accession No. ML19220A043.

51

4.5.5. Recommendations The panel recommends that the staff request the FitzPatrick licensee to submit JAF-CALC-19-00005, Revision 1 on the docket. If appropriate, the NRC could issue a correction to the July 2020 safety evaluation to add a reference to the docketed version, for greater public clarity.

The panel recommends that NRC license review guidance be reviewed and updated as needed to ensure that expectations for requesting updated documentation (i.e., revised versions of enclosures or analyses) are clear for future applications.

5. CONCLUSION AND RECOMMENDATIONS The panel substantiated several of the DPO submitters findings, as summarized below.

For Issue 1 on errors in the licensees dose analysis, the panel agrees with the submitters on most of the issues they identified. The panel was neutral on one point regarding the appropriateness of point or area receptors. The panel does not draw a conclusion on whether the staff would normally be expected to identify these calculation issues during its review. The licensee is responsible for quality assurance of its own calculations, and the NRC needs to review submittals to the extent that it can make a finding that there is reasonable assurance that regulations are met, including following up on any errors identified in the course of the review. Depending on the scope and depth of the review, which can be informed by risk insights, there could be errors in licensee calculations that are never identified by the NRC. The risk insights, however, should inform the staff of the consequences of potential errors in areas of an analysis that are not reviewed in depthan effective risk achievement worth of each element of the staffs review.

For Issue 2 on the use of staff evaluation, the panel agrees with the submitters that staff analyses and insights were used to draw conclusions without the licensee providing corroborating information. The staff appears to have extended its confidence in the overall safety of plant operation and the well-known conservatisms in design bases analyses (both true in the philosophical sense) to a specific finding on a regulation that directly invokes the licensees analysis. The panel does not take a position on whether the licensee could have demonstrated compliance with the 10 CFR 50.67 dose criteriathe panels concern is that the licensee was not thoroughly asked to do so. The panel also notes that the NRC-approved BWROG approach to crediting condenser holdup and deposition includes outdated and overly conservative assumptions about the fragility of secondary systems. A modernized approach could reduce the burden of licensees conformance to this approach and make alternative source term calculations more realistic while remaining within a structured regulatory regime.

For Issue 3 on proper use of technical specifications, the panel agrees with the submitters that neither technical specifications nor any other programmatic assurances were put into place to assure that the condenser pathway would be available following a core damage event. This result is unsurprising since the licensee did not request to credit this pathway in its analysis of record and only mentioned it as part of a sensitivity study. Under the modernized approach suggested above, assuring the availability of the condenser pathway would be paramountand it was not done in the FitzPatrick review.

For Issue 4 on the clarity of the licensing and design basis, the panel concluded that the licensees intended analysis of record is clearly referenced, but agreed with the 52

submitters that this analysis is inaccessible and contains errors thatpresumably still unresolvedwould be included in all future alternative source term calculations under the original NRC approval. The panel discussed whether the licensee should be required to update its analysis because of these issues. The panel came to a preliminary conclusion that the backfitting requirements of 10 CFR 50.109 would likely limit the staffs action to simply requesting the analysis of record to be docketed. The conclusion on this matter rests with the office, not with the panel.

For Issue 5 on the cumulative effect of the errors identified in Issue 1, the panel believes the licensee did not provide sufficient information to demonstrate that their deterministic analysis meets 10 CFR 50.67 and that the small analytical margin means the estimated control room dose may exceed 5 rem during this assumed design-based accident. The panel did not find that there was an undue risk to public health and safety such that the adequate protection provisions of 10 CFR 50.109 need be invoked to require revisions to plant operation or procedures (e.g., use of the condenser).

These conclusions resulted in several recommendations, as enumerated below:

1. The panel recommends that the staff request the FitzPatrick licensee to submit JAF-CALC 19-00005, Revision 1 on the docket. If appropriate, the NRC could issue a correction to the July 2020 safety evaluation to add a reference to the docketed version, for greater public clarity.

2. The panel recommends that appropriate processes be followed to document the staffs conclusion on whether backfitting is warranted on this topic, as discussed in Section 4.2.2 above.

3. The panel recommends providing clarification in LIC-101 with respect to whether confirmatory analysis or evaluations performed by the staff in support of an amendment are considered working files meeting the criteria to be considered official agency records.

4. The panel also recommends reviewing and revising LIC-101 as needed to ensure that expectations for requesting updated documentation (i.e., revised versions of enclosures or analyses) are clear for future applications.

5. The panel recommends revising LIC-206 to reflect the proper use of risk insights to scope a review, but not to alleviate the need to pursue known flaws in a licensee analysis.

6. The panel recommends that trainingperhaps in conjunction with the LIC-206 revisionbe provided to all staff and management involved in licensing actions (i.e.,

technical reviewers, project managers, and supervisors) to emphasize the philosophical points mentioned in Section 4.3.5 above.

7. The panel recommends that the staff revise guidance for crediting the condenser leakage treatment pathway as part of alternative source term licensing actions to reflect any findings and recommendations of this DPO report that are accepted by the DPO decisionmaker, as discussed further in Section 4.4.4 above.

The panel appreciates the openness of all the staff and management approached during its review. These individuals accommodated meetings, answered all questions, provided 53

supporting documents that illuminated the review, and universally explained their good regulatory intent in conducting the activities they discussed. Most importantly, the panel appreciates the determination of the submitters in bringing these issues to light through the NRCs established processes for raising different views. They provided thorough information that gave the panel a good basis for beginning its review and were consistently open and accessible to further the panels work, even in a situation that could have felt uncomfortable.

Although the panel finds flaws in the staffs conclusions on the FitzPatrick review, it is with an understanding that the NRC is currently transitioning to greater risk-informed decision-making.

Given the panel members extensive risk-assessment experience, the panel has great sympathy for the desire to apply the NRCs limited resources to issues of the highest safety significance.

The staff is currently underserved, however, by unclear messaging and guidance on exactly how and when to apply these risk insights. The NRC can expect some growing pains as it navigates this path to becoming a modern risk-informed regulator, and the panels conclusion is that public health and safety remains assured in this instance despite the issues raised.

54

Document 4: DPO Decision UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 November 17, 2021 MEMORANDUM TO: Mark Blumberg, Senior Reactor Engineer Radiation Protection and Consequence Branch Division of Risk Assessment Office of Nuclear Reactor Regulation Michael Markley, Branch Chief Plant Licensing Branch II-1 Division of Operating Reactor Licensing Office of Nuclear Reactor Regulation Signed by Veil, Andrea FROM: Andrea D. Veil, Director on 11/17/21 Office of Nuclear Reactor Regulation

SUBJECT:

DIFFERING PROFESSIONAL OPINION CONCERNING THE JAMES A. FITZPATRICK NUCLEAR POWER PLANT ISSUANCE OF AMENDMENT NO. 338 RE: ALTERNATIVE SOURCE TERM FOR CALCULATING LOSS-OF-COOLANT ACCIDENT DOSE CONSEQUENCES (DPO-2021-001)

The purpose of the memorandum is to respond to your differing professional opinion (DPO) submitted on January 19, 2021, in accordance with Management Directive 10.159, The Nuclear Regulatory Commission Differing Professional Opinions Program (Agencywide Documents Access and Management System (ADAMS) Accession No. ML15132A664).

DPO-2021-001, (ADAMS Accession No. ML21042B867), documents your concerns with the staffs issuance of Amendment No. 338 for James A. FitzPatrick Nuclear Power Plant related to their alternative source term for calculating loss of coolant accident dose consequences (ADAMS Accession No. ML20140A070).

I commend you for your commitment and dedication to the Nuclear Regulatory Commissions mission. Your willingness to raise concerns with your colleagues and managers and ensure that your concerns are heard and understood is admirable and vital to ensuring a healthy safety culture within the Agency. I also want to thank you for the substantial amount of time you dedicated to having additional discussions with me and my staff to ensure that I fully understood CONTACT: Caroline Tilton, NRR (301) 415-0990

your perspectives on these very complex issues. My response to your DPO, including associated follow-up actions, is described in the Enclosure.

Enclosures:

1. Directors decision for differing professional opinion

2.

Attachment:

Detailed evaluation of issue 1, concerns (a) through (b)

DIRECTORS DECISION FOR DIFFERING PROFESSIONAL OPINION JAMES A. FITZPATRICK NUCLEAR POWER PLANT ISSUANCE OF AMENDMENT NO. 338 RE: ALTERNATIVE SOURCE TERM FOR CALCULATING LOSS-OF-COOLANT ACCIDENT DOSE CONSEQUENCES (DPO-2021-001)

Background

In differing professional opinion (DPO) 2021-001, you expressed concerns with Amendment No.

338 1 for James A. FitzPatrick Nuclear Power Plant (FitzPatrick). This amendment adopted an alternative source term (AST) for accident analyses, increased the allowable amount of main steam isolation valve (MSIV) leakage and removed the MSIV leakage collection system (MLCS) from the Technical Specifications (TS). You stated that the licensees analysis was in error, did not follow applicable agency guidance, and did not demonstrate compliance with NRC regulations; that the staffs safety evaluation (SE) was not consistent with the licensees analysis and did not demonstrate compliance with the Nuclear Regulatory Commission (NRC) regulations; and, that the resultant errors in the licensees design basis could adversely impact future plant changes and regulatory actions.

The DPO ad hoc review panel (the panel) issued their report to me on June 9, 2021, after reviewing the applicable documents, conducting internal interviews with relevant individuals, and completing their deliberations. On June 22, 2021, I discussed the panel report with the DPO panel Chair and the members of the panel. On June 11, 2021, we met to discuss your insights and comments based on the panel report findings.

To inform my decision regarding your DPO, I reviewed your DPO submittal, the panels report, and considered our discussion on June 11, 2021. To better understand your concerns, I assigned the Deputy Office Director for Engineering and a Technical Assistant from my office to assist in the evaluation and documentation of my decision. This Office of Nuclear Reactor Regulation (NRR)-led team gathered information through discussions with you, the DPO panel, and other knowledgeable staff and reviewed documents pertinent to your DPO submittal. In addition, we tasked an independent reviewer with calculating the doses in the control room using an NRC-approved code. The information collected provided independent insights and perspectives that were vital to reaching my conclusion.

Summary of Issues You and the panel agreed on a summary of issues identified in the DPO that accurately addressed your concerns. This summary of issues can be found in Section 2 of the DPO panel report.

1 FitzPatrick Amendment No. 338 dated July 21, 2020 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML20140A070).

Enclosure

Summary of My Decision Our independent assessment found that the FitzPatricks approved analysis of record is acceptable and meets the regulatory requirements. Our assessment also identified that the licensees assumptions for one parameter (elemental removal coefficient) were outside of the boundaries established by the standard review plan (SRP). We also acknowledge that the particulate removal coefficient values used in the calculation, both for the drywell and the main steam lines (MSLs), may not have been the most conservative. The NRC staffs SE for FitzPatrick also noted these non-conservatisms and concluded that the assumptions used in the licensees analysis were acceptable and did not significantly impact the dose results.

It is important to note that your concerns are centered around control room doses not meeting the regulatory limit of 5 rem imposed by 10 CFR 50.67 after a loss-of-coolant-accident (LOCA).

The results for offsite doses have significant margin and therefore, are not an area of concern.

As such, our review focused on control room doses. It is also important to note that the licensees approved calculation of record also contains significant conservatisms. A few of the many conservatisms include: (1) using the volume of only two of the four MSLs and only half of the reactor building volume to account for deposition of activity; (2) not crediting containment elemental iodine and particulate deposition/plateout and activity reduction by the steam separators or the iodine partitioning in the reactor vessel; and (3) not considering suppression pool fission product scrubbing and retention.

We performed our own independent modeling of the licensees analysis using the NRC-developed RADionuclide Transport, Removal And Dose Estimation, (RADTRAD) code. We modeled the licensees analysis of record adjusting various parameters to incorporate information we learned after the FitzPatrick amendment was issued. Specifically, we used a value within the SRP range for the elemental removal coefficient, a reasonable value for the particulate removal coefficient in the drywell and more accurate values for the particulate removal coefficients in the MSLs. Even though the dose calculated in the control room in this independent analysis exceeded the regulatory limit (5.11 rem), considering the substantial conservatisms noted above, we have reasonable assurance the dose will remain below 5 rem.

In fact, we performed a RADTRAD run using the same inputs mentioned above but crediting suppression pool fission product scrubbing and retention and the control room dose value resulted in 3.03 rem. Although our independent analysis using the analysis of record and incorporating information learned after FitzPatricks review resulted in exceeding the regulatory limit, the staffs conclusion regarding the FitzPatricks review was based on the best information available at the time. Therefore, a backfit is not justified given that this new information does not indicate that the licensees or staffs methods were clearly in error and any further effort from the licensee would not provide an increase in safety as compared to the status quo based on significant conservatisms in the licensees analysis. In addition, the licensees analysis of record also demonstrates control room doses will remain below 5 rem. Although there were non-conservatisms in the licensees analysis, the NRC staff concluded in the SE that the significant conservatisms contained in the analysis and mentioned above provide reasonable assurance that the control room dose would remain below 5 rem.

Our independent assessment also found that the NRC staff based their regulatory finding of adequate protection on the licensees deterministic analysis rather than a separate analysis completed by the NRC staff. The NRC staff performed an assessment of the seismic ruggedness of the power conversion system (PCS) and the main condenser to achieve high confidence that these systems will remain available after a safe shutdown earthquake (SSE) for fission product dilution, holdup, and retention. The NRC staff used engineering judgement and

risk insights to support the deterministic conclusion, and to balance any uncertainties or approved non-conservatisms. The NRC staff did not use these risk insights as the basis for their regulatory finding, nor was an assumption of holdup in the condenser credited by the NRC staff. Consistent with the statements of consideration for 10 CFR 50.67, the NRC staff leveraged these risk insights, in a manner that complements the deterministic approach and supports the traditional defense-in-depth philosophy. This approach also follows the NRCs policy statement on the use of probability methods. These risk insights are in accordance with the Commissions direction in SRM-SECY-19-0036, Application to the Single Failure Criterion to Nuscale power LLCs Inadvertent Actuation Block Valves which stated: [i]n any licensing review or other regulatory decision, the staff should apply risk-informed principles when strict, prescriptive application of deterministic criteria such as the single failure criterion is unnecessary to provide for reasonable assurance of adequate protection of public health and safety.

I agree with the panel in their recommendations to provide additional guidance to clarify the appropriate use and documentation needed to refer to confirmatory and independent analyses and risk insights in our SEs. In addition, I found that there was a lack of clarity and transparency with regard to the basis for the staffs conclusions in some areas of the SE. As such, I recommend that the internal guidance updates include the importance of clearly and transparently articulating the basis of the staffs conclusion. This is consistent with the Openness NRC Principle of Good Regulation. It will also help to ensure that the public has clarity on the basis for our decisions.

It is important to note that neither the panel nor I identified a public health and safety issue associated with the concerns raised in the DPO.

My Assessment of the DPO and the Panels Findings The panel performed a thorough evaluation of the DPO, and related technical areas. Their report is a good source of background and reference information, was very well written, and provided thoughtful conclusions. It also documented the significant and valuable effort of the DPO panel. The next several paragraphs, outline the basis for my decisions, and delineate where my conclusions differ from the panels findings.

Issue 1: You stated that the licensees dose analysis does not demonstrate compliance with 10 CFR 50.67. You also stated that FitzPatricks LOCA analysis, for showing compliance with 10 CFR 50.67, does not follow the guidance in Regulatory Guide (RG) 1.183, 2 NUREG-0800, SRP Section 15.0.1, 3 and RG 1.194. 4 You contended that the LOCA analysis contains errors. You also asserted that, if the licensee had corrected these errors and followed the guidance, the actual control room dose would exceed the 5 rem total effective dose equivalent (TEDE) requirement in 10 CFR 50.67(b)(2)(iii).

You provided seven distinct examples that support your concerns under Issue 1. We have labeled them (a) through (g). To summarize our findings for Issue 1 and concerns (a) through (g), our independent assessment determined that the licensees dose analysis is in compliance 2

RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Plants, dated July 2000. ADAMS Accession No. ML003716792.

3 SRP Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, dated July 2000. ADAMS Accession No. ML003734190.

4 RG 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants, issued June 2003. ADAMS Accession No. ML031530505.

with 10 CFR 50.67, follows the regulatory guidance, and it is not in error. We did identify three parameters used as inputs to the licensees calculation that may not have been the most conservative (i.e., elemental removal coefficient and the particulate removal coefficient used for the drywell and the MSLs). The NRC staff addressed and approved these parameters in the FitzPatricks SE. We agree that more conservative values would increase calculated doses.

We performed an independent analysis using more conservative values for the elemental removal coefficient, the particulate removal coefficient in the drywell and particulate removal coefficients in the MSLs and found that the dose in the control room results is fractions above the 10 CFR 50.67 limit (5.11 rem). It is important to note that this independent analysis used a new method to calculate deposition in the MSLs that was not available at the time of the review.

We agree with you that considering these more conservative values, in FitzPatricks analysis of record, results in a dose that exceeds the regulatory limit. However, when considering the substantial conservatisms noted above, we have reasonable assurance the dose will remain well below 5 rem if the postulated scenario were to occur. In addition, the methodology used by the licensee was an acceptable method that was consistent with the regulatory guidance at the time of the review. The conservatisms in the assumptions are evident in our independent RADTRAD run which, using more conservative values for the parameters found non-conservative, but crediting suppression pool fission product scrubbing and retention, the control room dose value resulted in 3.03 rem. Although credit for suppression pool fission product scrubbing and retention should be approved on a case-by-case basis and with appropriate justification, it is a physical phenomenon that would be expected to happen during the postulated scenario. For the reasons stated above, I do not believe a backfit is justified because any further effort from the licensee would not provide an increase in safety as compared to the status quo based on significant conservatisms in the licensees analysis. We also did not agree with your interpretation of the regulatory guidance. Specifically, it would have requested that the licensee make assumptions that would be inconsistent with the postulated scenario, which has been determined to be the most limiting. Please refer to Attachment 1 for details on how we dispositioned concerns (a) through (g).

Issue 2: You stated that the NRC staff issued an amendment using an analysis not consistent with the licensees analysis. You explained that AST requirements in 10 CFR 50.67(b)(2) state that the NRC may issue an amendment only if the applicants analysis demonstrates, with reasonable assurance, compliance with the dose criteria in 10 CFR 50.67. You asserted that the NRCs approval was based on the licensees flawed analysis combined with the staffs risk and engineering insights that, in part, credit the capability of the PCS and main condenser to serve as a holdup volume for MSIV leakage. Because the licensees analysis of record does not credit transport of the MSIV leakage to the condenser or this holdup volume, you contended that the NRC staff rationalized a basis for approval that was not based on the analysis in the licensees submittal.

In the report, the panel concluded that the approval of the FitzPatricks alternate source term (AST) license amendment request (LAR) is not based on the licensees submitted analysis of record, but rather is based on the staffs unconfirmed view of the availability of the condenser to mitigate the licensee-provided analysis that shows unacceptable control room doses. The panel also agreed with you that 10 CFR 50.67 clearly establishes that the staff can approve an AST license application only if the licensees analysis demonstrates with reasonable assurance that the public and control room doses limits are met, which are also clearly established within 10 CFR 50.67. The panel concluded that the aggregate errors and issues identified in their report, if fully addressed by the licensee in a licensing basis analysis, could result in control room dose values that exceed the criteria in 10 CFR 50.67. The panel added that if the request for additional information (RAI) response base case had been used as the analysis of record, the

dose criteria in 10 CFR 50.67 would not have been met.

I disagree with the panel and this specific concern. The licensees analysis is not in error and it demonstrates that the regulatory limits are met based on regulatory guidance available at the time the amendment was reviewed and approved. While there were some non-conservative parameters in the licenses analysis, the NRC staff determined that these parameters did not significantly impact the results. A detailed assessment of the RAI base case included in the licensees sensitivity analysis, is contained in the Appendix to this memorandum under Issue1(f).

In the SE, the NRC staff did not credit the PCS or the main condenser to make the regulatory finding of adequate protection. Section 3.5 of the SE titled NRC Staff Risk and Engineering Insights, explains that, by Commission direction in the SRM to SECY-19-0036, 5 the staff should apply risk-informed principles in any licensing review or other regulatory decision when strict, prescriptive application of deterministic criteria is unnecessary to provide for reasonable assurance of adequate protection of public health and safety. In Section 3.5 of the SE, the NRC staff concluded that these risk and engineering insights support its reasonable assurance finding based on its deterministic review. The NRC staff performed an independent assessment of the seismic robustness of the condenser which was referred to in the SE for FitzPatrick but not made publicly available. While this was a different approach from previous applications, in which a licensee credits the condenser through referencing a Boiling Water Reactor Owners Group (BWROG) topical report, 6 this assessment gave the NRC staff high confidence the PCS and the main condenser would remain available for fission product dilution, holdup, and retention when exposed to seismic accelerations corresponding to FitzPatricks design basis SSE. In making this assessment, the staff leveraged recent relevant operating experience such as that obtained from the Great East Japan Earthquake. The NRC staffs assessment balanced the impact that less-conservative assumptions had on the doses, with noted conservatisms (e.g., modeling 2 vs 4 MSLs and the high confidence that the PCS and main condenser would be available). These risk insights resulted in additional safety margin but were not necessary to make the regulatory finding. In addition, as mentioned in the statements of consideration for 10 CFR 50.67 (74 FR 71990), defense in depth was addressed by using a deterministic dose calculation which fundamentally accounts for uncertainties in equipment and human performance. Rather than relying on crediting the PCS and main condenser to make a regulatory finding that the control room doses would be below 5 rem, the NRC staff based their regulatory finding on the licensees application and the responses to the RAIs, which indicated that control room doses would remain under 5 rem. This regulatory finding was informed by the risk insights noted above (i.e., the risk insights were not the basis for the finding). While some scenarios modeled by the licensee in their RAI response indicated doses greater than 5 rem, the RAI response which included the sensitivity analysis is not the official licensing basis, and in the NRC staffs judgment, the additional conservatisms in the licensees analysis provided assurance that the regulatory limits were met. Consequently, regardless of whether a regulatory requirement calls for a deterministic evaluation, risk insights can be applied to inform regulatory decisions as supported by the aforementioned Commission direction.

My assessment that the NRC staff did not credit the PCS or the main condenser to get 5

SRM to SECY-19-0036, Application of the Single Failure Criterion to NuScale Power LLCs Inadvertent Actuation Block Valves, dated July 2, 2019. ADAMS Accession No. ML19183A408.

6 BWROG Licensing Topical Report NEDC-31858P-A, Increasing MSIV Leakage Rate Limits and Elimination of Leakage Control Systems, dated August 1999. ADAMS Accession Nos. ML993350208 (transmittal) and ML993440253 (report - non-public).

reasonable assurance of adequate protection is supported by the fact that the licensees analysis of record, JAF-CALC-19-0005, Revision 1, Post-LOCA EAB, LPZ, and CR Dose -

AST Analysis, does not credit these nonsafety-related SSCs for dilution, holdup and retention.

In addition, it is also supported by the following statements in the NRC staffs SE:

• Since the application is not a fully risk-informed submittal (with probabilistic risk information), the staff does not apply risk as the basis for acceptance of a request; however, the following risk and engineering insights inform the technical review by supporting the deterministic safety conclusions and enhance the technical reviewers confidence in their technical evaluations.

• In addition, as mentioned in the statements of consideration for 10 CFR 50.67, defense in depth is addressed using a DBA in the deterministic dose calculation. Therefore, consistent with the statements of consideration for 10 CFR 50.67, the principles of risk-informed decisionmaking, and the Commission direction to the staff in the SRM to SECY-19-0036, the NRC staff has determined these risk and engineering insights support its reasonable assurance finding based on its deterministic review.

I did find the communication that the NRC staff used to inform how these risk insights were utilized lacked clarity and transparency. The NRC staffs SE for FitzPatrick should have included details of the independent assessment used to conclude the PCS and the main condenser would be available after an SSE. The level of detail provided in an SE should be sufficient for a member of the public to understand how we arrived at our conclusions. By not including this detail in the NRC staffs SE, we created the perception that, without crediting the PCS and the main condenser, the doses based on the licensees sensitivity analysis would be above 5 rem, and that the staff used its knowledge of the inherent seismic robustness of these nonsafety related SSCs as the basis for its decision.

Issue 3: You stated that the LOCA analysis credits SSCs not in TS, inconsistent with 10 CFR 50.36. As noted in Issue 2, you asserted the NRC staffs SE for FitzPatrick included staff-generated risk and engineering insights. You added that some of the systems and specific equipment alignments credited in the staffs and the licensees evaluation are not controlled by TS. You contended that this credit conflicts with TS requirements in 10 CFR 50.36(c)(2)(ii),

Criterion 3, that requires limiting conditions for operation (LCO) be established for SSCs. You stated this credit also conflicts with AST guidance in RG 1.183, that would assure these SSCs are available and maintained consistent with the integrity assumed in the analysis.

The panel agreed with you that the programmatic aspects of ensuring the availability, reliability, and performance of the condenser leakage treatment pathway, including boundary valve isolation, should have been identified and managed via established licensee programs consistent with previous license amendments that utilized the NRC-approved BWROG topical report. The panel stated that this approach would be consistent with 10 CFR 50.36 and the specific guidance in RG 1.183. The panel also stated that, in the case of FitzPatrick, components associated with this staff-credited pathway to achieve acceptable control room doses were not required to be included in any licensee program (e.g., TS, surveillance testing, inservice testing, emergency operating procedures, maintenance programs) and as such their reliability and availability in the context of this application cannot be assured.

I disagree with the panel and this specific concern. Under Issue 2, I explain that the NRC staff did not credit the PCS or the main condenser to make the regulatory finding of adequate protection. Rather, the NRC staffs independent assessment found that it is reasonable to

conclude that the likelihood of the PCS and the main condenser not being available to serve as a volume for holdup and retention is very low. These insights were used to support, not in place of, the deterministic safety conclusion. The criteria in 10 CFR 50.36(c)(2)(ii) establishes that a TS LCO must be provided for [a] structure, system, or component that is part of the primary success path and which functions or actuates to mitigate a design basis accident or transient that either assumes the failure of or presents a challenge to the integrity of a fission product barrier. In addition, RG 1.183, Regulatory Position 5.1.2, allows credit for accident mitigation features that are classified as safety-related and required to be operable by TS (among other requirements). Since credit is not taken for these nonsafety related SSCs, nor are they part of the primary success path in the approved design basis dose consequence analysis, they do not need to be included in the licensees TS.

Issue 4: You stated that errors in the design basis could adversely impact future plant changes and regulatory actions. As noted in Issue 1, you stated that you identified several errors in the licensees analysis. As noted in Issue 2, you stated the NRC staffs SE for FitzPatrick included staff-generated risk and engineering insights and analysis. You contended that the NRC staff did not follow SRP 15.0.1 and did not work out these differences to reconcile errors in the licensees analysis or the factual accuracy of docketed information. For this reason, you also contended that the design and licensing basis of FitzPatrick was modified and approved based on erroneous and incorrect information in a manner that will adversely impact future changes at the plant using 10 CFR 50.59, Changes, tests, and experiments, challenge inspectors in performing inspections and regulatory findings, and challenge reviews for future licensing actions that may use it as a precedent.

The panel agreed with you, in part. The panel stated that while the basis for the NRCs approval is technically clear (the NRC staff specifically references the analysis of record as JAF-CALC-19-00005, Revision 1), the documentation lacks regulatory rigor. The panel also explained that since Revision 1 was not submitted on the docket, as Revision 0 was, the details of the calculation are not readily available to inspectors, license reviewers, or members of the public. In addition, the panel stated that since the approved calculation becomes the foundation for all future analysis, any errors, if not corrected by the licensee, would be carried forward into other accident analyses without any NRC review. The panel explained that consequently, for the NRC staff to raise any issues with these analyses, a backfit would need to be invoked.

I disagree with the panel and this specific concern. As an attachment to their LAR, FitzPatrick included JAF-CALC-19-00005, Revision 0. In their March 30, 2020 response to the NRC staffs RAI, the licensee refers to JAF-CALC-19-00005, Revision 1, stating it revised offsite and onsite doses using new removal coefficients and spray flow rate that account for obstructions. The licensee clearly stated the revised values in their response. As supported by the internal guidance contained in LIC-101 7, it is not necessary for the licensee to submit Revision 1 on the docket if they sufficiently described the changes made in their docketed response. In addition, as stated by the panel in their report, the NRC staff clearly references JAF-CALC-19-00005, Revision 1, in the SE as the calculation of record being approved.

Issue 5: You stated that the NRC staffs analysis does not show compliance with 10 CFR 50.67. As noted in Issue 1, you expressed that you identified several errors in the licensees analysis. You asserted that the errors and uncertainties in the licensees analysis are so large that, even if it was appropriate to use the staffs risk and engineering insights, the control room 7

LIC-101, License Amendment Review Procedures, Revision 6, issued August 3, 2020. ADAMS Accession No. ML19248C539.

doses would exceed the regulatory values necessary for making a finding of reasonable assurance that 10 CFR 50.67 is met.

You performed an independent analysis of the doses in the control room accounting for the concerns that you raised under Issue 1 and that included other assumptions. The calculated control room dose in your independent analysis increased from 5 rem to 466 rem.

The panel did not attempt to validate your analysis. The panel stated that the licensee did not provide sufficient information to demonstrate that their deterministic analysis complies with 10 CFR 50.67. In addition, the panel stated that considering the very small margin available (0.33 rem), the estimated control room dose is likely to exceed 5 rem during this assumed design-basis accident. The panel concluded that an adequate protection backfit was not warranted because there was no undue risk to public health and safety and that, within the scope of the panels review, they did not identify a substantial safety enhancement afforded by revising the calculations based on the inherent conservatisms in the analyses.

I disagree with this specific concern. Our independent assessment finds that the inputs and parameters discussed in this memorandum and included in the approved analysis of record are not in error. These inputs and parameters were reviewed and approved by the NRC staff based on the regulatory guidance available at the time. In addition, our independent assessment showed that correcting for the non-conservative parameters identified in Issue 1 and removing a few of the many conservatisms built into the analysis of record, control room dose values continue to meet the regulatory limit.

I agree with the panel that requesting a revision to these calculations provides no safety enhancement. I also agree with the panel that these design basis scenarios are bounding and contain inherent conservatisms. In addition, I agree with the panel that no backfit is warranted.

Additional Issues: During our discussions, you expressed concerns with draft Interim Staff Guidance (ISG) DRA-ISG-2021-XX 8 and the recent draft revision to RG 1.183 by stating that these documents used the FitzPatrick approach and amendment as a model. These documents are undergoing the necessary review process including a review by the Advisory Committee on Reactor Safeguards (ACRS). As supported by the conclusions in this document, I do not take objections with the guidance contained in either of these documents.

My Assessment of the Panels Recommendations I appreciate the panels thoughtful assessment of the concerns raised in the DPO as well as your perspectives. The panel has the following recommendations for consideration by NRC management. My responses to the panels recommendations are provided below.

Response to Recommendation 1 (Panel Recommendation 1) The panel recommends that the staff request the FitzPatrick licensee to submit JAF-CALC 19-00005, Revision 1 on the docket. If appropriate, the staff could issue a correction to the July 2020 safety evaluation to add a reference to the docketed version, for greater public clarity.

8 ISG DRA-ISG-2021-XX, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms. ADAMS Accession No. ML21078A051.

I disagree with Recommendation 1. The information provided by the licensee on the docket clearly stated the parameters that were revised in Revision 1 of JAF-CALC-19-00005 and was sufficient for the NRC staff to make their regulatory finding.

Response to Recommendation 2 (Panel Recommendation 2) The panel recommends that appropriate processes be followed to document the staffs conclusion on whether backfitting is warranted on this topic, as discussed in Section 4.2.2 of [the panel] report.

I disagree with Recommendation 2. The licensees approved analysis of record meets the regulation. While our independent analysis indicated a dose of 5.11 rem in the control room accounting for the non-conservatisms in the licensees analysis and applying current day methods, accounting for even a small number of the many conservatism in the analysis of record provides reasonable assurance that the regulatory requirements were met. In addition, the Agencys current policies on backfitting preclude the NRC staff from issuing a backfit when the understanding for what constituted proper implementation of the regulations, standards, and practices is not widely known or understood by professionals at the time. In particular, this understanding is not restricted to the regulatory positions of the NRC but includes any industry or professional standards and practices in existence at the time the original determination was made. At the time the FitzPatrick LAR was being reviewed, and in the present time for that matter, the NRC did not have a specific reduction criteria or any approved guidance licensees could use to address the impact of drywell sprays on deposition in the MSLs. In addition, the use of the 20-group method to calculate particulate removal coefficient values was an acceptable method. Further, based on inherent conservatisms in the analyses, the panel concluded in the DPO panel report that a backfit is not warranted.

Response to Recommendation 3 (Panel Recommendation 3) The panel recommends providing clarification in LIC-101 with respect to whether confirmatory analysis or evaluations performed by the staff in support of an amendment are considered working files meeting the criteria to be considered official agency records.

I agree with Recommendation 3. I task the NRR Division of Operating Reactor Licensing (DORL) with addressing this recommendation to consider the use of engineering and risk insights as documented in Section 3.5 of the FitzPatrick SE and to determine the appropriate level of documentation for subsequent amendments.

Response to Recommendation 4 (Panel Recommendation 4) The panel also recommends reviewing and revising LIC-101 as needed to ensure that expectations for requesting updated documentation (i.e., revised versions of enclosures or analyses) are clear for future applications.

I disagree with Recommendation 4. LIC-101 provides clear guidance in this area by stating that information important in the staff's decision-making process and not otherwise in the public domain or reasonably inferred by the staff, must be formally [on the docket] provided by the licensee. In addition, I have determined that all appropriate information was submitted on the docket during the Fitzpatrick review.

Response to Recommendation 5 (Panel Recommendation 5) The panel recommends revising LIC-206 9 to reflect the proper use of risk insights to scope a review, but not to alleviate the need to pursue known flaws in a licensee analysis.

I agree with Recommendation 5, in part. I support updating LIC-206 to reflect the proper use of risk insights, but risk insights can and should be used beyond scoping a review. Risk insights can also be used to support our regulatory conclusions, as in the case of FitzPatricks amendment. I task NRR DORL with updating LIC-206 to provide clarifying guidance on the proper use of risk insights including an explanation of the proper terminology and the appropriate level of documentation to be used in subsequent amendments.

Response to Recommendation 6 (Panel Recommendation 6) The panel recommends that trainingperhaps in conjunction with the LIC-206 revisionbe provided to all staff and management involved in licensing actions (i.e., technical reviewers, project managers, and supervisors) to emphasize the philosophical points mentioned in Section 4.3.5 of [the panel] report.

In Section 4.3.5 of the panel report, the panel states that the recommendations in this area stem from two fundamental regulatory concepts that appear to have been departed from during the FitzPatrick amendment review:

• The licensee has the primary responsibility for safety and must provide an application that demonstrates the regulations are met (especially when a regulation is as clear as 10 CFR 50.67 regarding licensee analysis). Therefore, staff reviewers must review the licensees submitted information and draw conclusions based on this information. While the staff can use confirmatory analyses and insights (including risk perspectives) to aid in their review, these staff analyses and insights cannot be substituted for the licensees analysis, especially when licensee information indicates unacceptable results. Staff reviewers should not base the acceptance of a licensing action based primarily on their own perspectives and views.

• The first element of risk-informed decision-making is to meet the regulations, unless the application is an exemption request. Therefore, risk insights and engineering judgments, while providing valuable insights on the scope and focus of the staffs review, cannot be used as a substitute for meeting the clear wording of the regulations.

I agree with the regulatory philosophies above and with Recommendation 6, in part. The NRC staff did not use Section 3.5 of the SE as the basis for their regulatory finding. The NRC staff also clearly documented that their approval was based on the deterministic calculation, Revision 1 of JAF- CALC 19-00005. In lieu of training, I task the NRR Division of Risk Assessment (DRA) with hosting an Executive Team (ET) Chat to provide a forum for NRR management and staff to discuss these regulatory philosophies. This ET Chat will be recorded for future reference and can serve as a just-in-time reference for all levels of the organization.

9 LIC-206, Integrated Risk-Informed Decision-Making for Licensing Reviews, Revision 1, issued June 26, 2020. ADAMS Accession No. ML19263A645.

In addition, I agree with the panel that the NRC staff could have been more transparent by sharing the details that were used to support the regulatory finding in certain parts of the SE.

However, this observation will be addressed by Recommendation 3.

Response to Recommendation 7 (Panel Recommendation 7) The panel recommends that the staff revise guidance for crediting the condenser leakage treatment pathway as part of alternative source term licensing actions to reflect any findings and recommendations of this report that are accepted by the decisionmaker, as discussed further in Section 4.4.4 of this [the panel] report.

In Section 4.4.4 of the panel report, the panel recommends including specific references to the operational provisions of the NRC-approved BWROG topical report and RG 1.183 (e.g., valves that must change position are supplied by emergency power sources). The panel does not view it necessary, given operating experience, that licensees perform analyses related to the inherent seismic robustness of the condenser and the associated piping. The panel also stated that the revised guidance should provide for walkdowns or other appropriate methods, as well as programmatic controls, to initially validate and then subsequently assure the availability of credited components.

I agree with Recommendation 7. The BWROG topical report provides a framework that the NRC finds acceptable when assessing the seismic robustness of the main condenser and associated components. I task NRR DRA with evaluating the need to incorporate some or all portions of the BWROG topical report in the next revision to RG 1.183.

Concluding Remarks I found that your DPO positions were of notable technical merit and well documented in your submittal. A summary of the DPO will be included in the Weekly Information Report (when the case is closed) to advise employees of the outcome. Thank you for raising your DPO and for your active participation in this process. I commend you for your commitment and dedication to the NRC mission. An open and thorough exploration of how we carry out our regulatory processes is essential to the effectiveness of our programs. Your willingness to raise concerns is admirable and vital to ensuring a healthy safety culture within the Agency.

In closing, I want to reiterate two points that I think underscore what I believe to be a critical aspect surrounding this DPO. First, we may not agree on how the staff dispositioned flaws in the licensees analysis using risk insights. This fact underscores the transition that is happening in the Agency as we make greater use of risk-informed decision-making. The efforts to apply the Agencys limited resources to issues of the highest safety significance will continue to evolve. During this transition, it is expected that some disagreements on how to apply these risk insights will result. It is important that we learn from these disagreements and improve our own processes to add clarification on exactly how and when to apply these risk insights.

Second, applying risk-informed principles in licensing reviews or other regulatory decisions when strict, prescriptive application of deterministic criteria are unnecessary to provide for reasonable assurance of adequate protection of public health and safety is consistent with the Commissions direction in the SRM to SECY-19-0036 and is an acceptable practice. Therefore, the NRC staff is encouraged to utilize risk insights to support or inform a reasonable assurance finding.

ATTACHMENT: DETAILED EVALUATION OF ISSUE 1, CONCERNS (a) THROUGH (g)

For issue 1, you considered the following aspects of the licensees analysis to be in error and/or did not conform to guidance in the associated regulatory guides (RGs) and the standard review plan (SRP):

(a) : You stated that FitzPatricks analysis did not assume the radioactive release was instantaneously and homogeneously mixed throughout all the free air volume in the drywell containment since the analysis did not include the air space in the main steam lines (MSLs) up to the closed main steam isolation valves (MSIVs).

The differing professional opinion (DPO) panel did not substantiate this concern. Specifically, they stated that at time=0, the radioactivity in the volume of piping from the vessel to the modeled closed MSIVs is the same as that assumed in the drywell. In our independent assessment, we discovered that the radioactivity calculated upstream of the closed MSIVs was smaller than in the drywell because it accounts for natural deposition (allowed by RG 1.183 Appendix A, Regulatory Position 3.2) and removal by containment sprays in the drywell (allowed by RG 1.183 Appendix A, Regulatory Position 3.3), in addition to other reducing mechanisms such as decay. During further discussions with you to clarify this concern, you stated that since the MSIVs are part of the drywell, in order for the drywell to represent an instantaneous and homogeneously mixed volume, the activity upstream the MSIVs should be the same as in the drywell. You added that this is required by RG 1.183 Appendix A, Regulatory Position 3.1 which states the radioactivity released from the fuel should be assumed to mix instantaneously and homogeneously throughout the free air volume of the primary containment in pressurized water reactors or the drywell in BWRs as it is released. However, RG 1.183 Appendix A, Section 3.0, Assumptions on Transport in Primary Containment, provides criteria for licensees to consider when evaluating containment leakage. The criteria to consider when evaluating MSIV leakage is provided in Section 6.0, Assumptions on Main Steam Isolation Valve Leakage in BWRs. RG 1.183 Appendix A, Section 6, allows credit, on an individual case basis, for a reduction of the amount of released radioactivity on steam system piping upstream of the outboard MSIVs by deposition and plateout.

I agree that the MSIVs are considered part of the containment boundary. However, for the purposes of the alternative source term (AST) analysis, Section 3.0 of RG 1.183 does not apply to the modeling of MSIV leakage and therefore, a reduced activity in the volume of piping upstream of the modeled closed MSIVs to account for natural deposition and removal by containment sprays is allowed.

I agree with the panel that Regulatory Position 6.3 of RG 1.183, allows crediting deposition and plateout on steam system piping upstream of the outboard MSIVs on an individual case basis and as evaluated by FitzPatrick and that the licensees analysis assumed that the release was instantaneous and homogenously mixed in accordance with Nuclear Regulatory Commission (NRC) guidance.

(b) : You stated that FitzPatricks analysis did not consider credible release pathways that would increase the amount of radioactivity released to the environment. You also stated that concerns over these pathways were raised by the Advisory Committee on Reactor Safeguards during the initial licensing of certain facilities and were addressed by the MSIV leakage control system (MLCS) design.

Attachment

As the panel report explains, the need for an MLCS has evolved since original licensing of plants. NRCs Generic Safety Issue (GSI) program GSI C-8 10 raised several issues related to the MLCS including potentially higher offsite doses resulting from its use and its high failure rate in a variety of failure modes. There is also precedence for licensees applying an AST to increase their allowed MSIV leakage and to remove the need to credit the MLCS and eliminate the associated Technical Specifications (TS) requirements. As stated in their license amendment request (LAR), FitzPatrick proposed an AST analysis that does not credit the MLCS while still meeting the requirement of 10 CFR 50.67. This is an acceptable approach.

You were concerned that there could be leakage up through the stem of the MSIV, resulting in a release that is closer to the control room. You were also concerned that leakage through the stem would not credit deposition in the piping and would result in higher radioactivity in the control room. The panel agreed with your concern that the licensee did not consider the most limiting or other available, viable release pathways in their analysis, nor did the staff question the licensees assumptions. The panel also agreed with you that had the licensee assumed leakage up through the stem of the outboard MSIV instead of past the valve seat, deposition in the piping between the outboard MSIV and the turbine stop valves (TSV) could not be credited and therefore would lead to a more limiting scenario.

I disagree with the panel and this specific concern. The MLCS collected leakage from the stem packing and downstream of all outboard MSIVs. Any release due to leakage past the stem of the outboard MSIVs at the time the LAR was being reviewed, would have to assume failure of safety-related, Seismic Class I components. In addition, three months after the amendment was approved, FitzPatrick isolated and capped off the piping from the MLCS to the standby gas treatment system. If a direct release from the stem of the outboard MSIVs to the environment were to occur, it would release in the reactor building, which has been evaluated for the safe shutdown earthquake (SSE) and it is assumed to remain structurally intact. Any leakage in the reactor building gets processed by the standby gas treatment (assumed to be in operation 20 minutes after a loss of coolant accident (LOCA)). This results in doses to the control room that are less than the evaluated scenario of MSIV leakage through the seat of the valves and all the way to the TSVs. In fact, our independent assessment, using the licensees analysis of record, revealed doses in the control room under this scenario would be around 1.27 rem, much less than the 4.67 rem calculated by the licensee. Therefore, our independent evaluation confirms the licensees assumption that the release that will occur at the TSV is more limiting and appropriate.

(c) : You stated that FitzPatricks analysis assumed a release point at a specific location in the turbine building (TB), rather than on the surface of the TB. You also stated that this assumption allowed dilution to be credited in the TB as it reduced the atmospheric dispersion factor, thereby improperly assessing the control room dose.

In their LAR and included in the calculation of record, the licensee stated that TSVs are assumed to be the release point for MSIV leakage and that the MSLs have been seismically analyzed up to the TSVs. The licensee also stated that this effectively assumes a seismic event causes the LOCA, the TB is lost, and the release occurs directly from the TSVs to the environment. The licensee explained that this is a more realistic release location than the minimum distance to an equivalent circular area projection of the TB (closest point on the surface of the building) and is consistent with other calculations.

10 Item C-8, Main Steam Line Leakage Control Systems

RG 1.194 Section 3.4 Determination of Source-Receptor Distances and Directions, states:

The source-to-receptor distance is the shortest horizontal distance between the release point and the intake.

The panel was neutral on what would be the appropriate assumption in this case: a release point at a specific location (i.e., TSV) or on the surface of the TB. The panel also stated that the surface of the building is 32.81 closer to the control room intake than the TSV. The shorter distance would have an impact in the atmospheric dispersion values and increase the calculated dose.

In their report, the panel stated that FitzPatrick had assumed the minimum distance to an equivalent circular area projection of the TB in previous design calculations. In their LAR, the licensee stated that a more realistic release location is at the TSVs, with a straight path to the control room intake. In the safety evaluation (SE), the NRC staff acknowledged the licensee utilized a straight-line trajectory between the release points (TSVs) and receptor (control room intake) and that the licensee conservatively did not credit wake effects from obstructions within the TB.

During our discussion, the panel stated that their assessment found that releases have been modeled in the past as both area and point sources.

I disagree that the licensee should have chosen the release point to be the closest point on the surface of the TB. The licensee chose the TSVs as the release point and used the calculated distance between the TSV and the control room intake to determine the atmospheric dispersion factor. Since the licensee assumed the TB was lost due to the seismic event, this is a reasonable assumption and does not deviate from RG 1.194. While assuming a release from the surface of the TB would result in the release being closer to the control room, I find that it is not reasonable to request that a licensee evaluate a parameter that is not consistent with the scenario evaluated and that is determined to be the most limiting. Assuming a loss of the TB is conservative in that it does not credit deposition and wake effects in the building. As such, assuming a release from the surface of the TB, while more conservative, would result in using conflicting parameters in the postulated scenario.

(d) : You stated that FitzPatricks analysis credited nonsafety-related structures, systems, and components (SSCs) such as vents and doors that are not controlled by their TS and that by omitting these release pathways that would be closer to the control room, doses could be impacted.

In the calculation of record, the licensee stated that additional release pathways out of the TB in the event of a LOCA were evaluated. These release pathways included: the smoke ejector vents on top of the TB; the access door from the TB to the administrative building roof; and, a TB exhaust duct that is 60 from the control room intake.

The panel agreed with you that the pathways noted in the DPO are not governed by TS, the set points are not monitored, and there are no limitations in place to ensure that these pathways are not available for a release to occur. Therefore, the panel agreed the licensee did not provide a bounding source-to-receptor combination, underestimating the atmospheric dispersion factor.

The panel stated that the TB exhaust release location is closer to the control room intake and therefore LOCA doses could be larger.

I disagree with the panel and this specific concern. The three pathways mentioned by the licensee are all part of the TB, which the analysis assumes is lost. If the pathways were to maintain their integrity post-LOCA, the licensee could make the argument that deposition in the TB could be credited and this would balance the dose results. Even if deposition in the TB is not credited, the following judgement could be made on the acceptability of these pathways to transport activity to the control room intake: the smoke ejector vents are normally closed and would not open in a post-LOCA scenario; the access door is normally closed, locked, and monitored by security; and, the TB exhaust duct, which is closer to the control room intake than the TSVs, has a gravity damper which only opens when the ventilation fan is in operation.

Since the most limiting scenario assumes a loss of offsite power (LOOP) and the TB ventilation fan is not powered by the emergency buses, it is reasonable to assume the damper is closed during this event. In addition, while these three pathways are not safety-related and are not in TS, they are not being credited for accident mitigation but rather assumed to retain their normal operation position. During follow up discussions with you on this concern, you indicated that the licensee should assume the most conservative scenario, and this includes assuming scenarios beyond what the licensee and the NRC have determine to be conservative and appropriate. In this specific case, you believe the licensee should assume a LOOP does not occur and therefore the gravity damper is open and available as a release pathway. Similar to concern 1(c), such assumptions would result in using parameters that conflict with the postulated scenario that was determined to be conservative and appropriate.

(e) : You stated that FitzPatricks analysis did not consider the limiting LOCA break location by selecting a recirculation line break, rather than a break in the reactor coolant system (RCS) just prior to the MSIVs. In addition, when the licensee modeled the break just prior to the MSIVs (as a sensitivity study in response 11 to the staffs request for additional information (RAI)), the licensee did not use the assumptions that were used in the LOCA analysis supporting the LAR.

The panel agreed with your concern that the analysis of record was based on total leakage through two MSLs while the evaluation for a ruptured MSL completed in response to the staffs RAI was performed assuming the total leakage was through four MSLs, making the comparison between the sensitivity study and the submittal of record challenging. The panel also agreed with you that the most limiting scenario cannot be determined by making multiple changes to the assumptions since this can obscure the impact. The panel indicated that the impact of assuming a breach in the MSL versus the recirculation line was not clear.

I agree with you and the panel that when evaluating the most limiting scenario, the integrity of the assumptions is crucial for a direct comparison. In responding to NRC staffs request to demonstrate the limiting scenario, the licensee provided a sensitivity study that changed the input parameters and in some of the cases did not meet the regulatory requirements. I believe the NRC staff should have followed up with the licensee and requested a more direct comparison. I also believe the NRC staff should have provided a detailed evaluation of the acceptability of the licensees docketed response and sensitivity study results in the FitzPatricks SE. In our independent assessment, we used the assumptions of the analysis of record to model a break in the RCS just prior to the MSIVs, and the dose results remain under regulatory limits.

(f) : You stated that FitzPatricks analysis used removal coefficients for aerosol settling that are non-conservative (based both on established science and a licensee reevaluation as part of a 11 FitzPatricks response to NRC staffs RAI dated March 30, 2020. ADAMS Accession No. ML20090E279.

sensitivity study). You added that by incorrectly modeling the impact of changing aerosol particle size distribution as the radioactive material moves down the steam line, the LOCA model overestimates the deposition in the steam lines and, therefore, underestimates the dose in the control room.

Particulate Removal Coefficients in the Drywell In their LAR, the licensee stated that the particulate removal coefficient in the drywell is 42.45hr-1 but 30.0 hr-1 is conservatively used in the analysis. The licensee also stated that the removal efficiency will be reduced to 3.0hr-1 when the decontamination factor (DF) is less than

50. In addition, the licensee stated that based on the equations in SRP 6.5.2, 12 the elemental iodine removal coefficient is conservatively assumed to be the same as the particulate aerosol removal coefficient.

In its March 30, 2020, letter, the licensee responded to a request from the NRC staff to provide a justification for their proposed fall height value which apparently did not consider obstructions present in the drywell. In their response, the licensee explained that a reduction in spray removal coefficient to account for possible obstructions to the spray coverage was not included.

The licensee also explained that by using a similar methodology as Nine Mile Point 1 and Oyster Creek to account for these obstructions, (an averaged spray header elevation and a reduction in spray flow by 1/3 each), the new particulate aerosol spray removal coefficients are 26.36hr-1 for DF less than or equal to 50 and 2.636hr-1 thereafter and for the remaining duration of spray operation.

The NRC staff evaluated the particulate aerosol spray removal coefficients in the SE for FitzPatrick by stating: The NRC staff reviewed the changes made to the initial LAR regarding the adjustments made to account for the presence of obstructions within the drywell and concludes that the modifications are reasonable and, therefore, acceptable.

The DPO panel found that FitzPatricks methodology, in fact, deviated from the methodology used by Nine Mile Point 1 and Oyster Creek. In addition, the DPO panel concluded that if FitzPatrick would have employed similar methodologies as Nine Mile Point, Unit 1 and Oyster Creek, the average fall height would decrease which would cause an increase in the calculated control room dose.

I agree with the submitters and the panel that the particulate removal coefficient values FitzPatrick used for the drywell in their analysis may not have been the most conservative.

However, the NRC staff reviewed and approved these values and concluded that the regulatory requirements were met, despite the licensees use of non-conservative removal coefficients.

Our independent analysis, using particulate removal coefficient values in the drywell as low as 20 hr-1 and correcting for the other non-conservatisms identified in the memorandum, resulted in control room doses that exceed the regulatory limit (5.11 rem). However, considering the substantial conservatisms also noted in the memorandum, we have reasonable assurance the dose will remain below 5 rem. In fact, we performed a RADTRAD run using the same inputs mentioned above but crediting suppression pool fission product scrubbing and retention, and the control room dose value resulted in 3.03 rem. Therefore, the staffs conclusions that the regulatory requirements were met, considering the licensees analysis, using methods that were appropriate at the time of the review, and accounting for the additional risk insights related to 12 SRP Section 6.5.2, Containment Spray as a Fission Product Cleanup System, Revision 4, dated March 2007. ADAMS Accession No. ML070190178.

other conservatisms in the analysis, were appropriate. Backfitting the licensee and requesting use of lower particulate removal coefficient values is not justified because any further effort from the licensee would not provide an increase in safety as compared to the status quo based on significant conservatisms in the licensees analysis.

Particulate Removal Coefficients in the MSLs The NRC staff acknowledged in FitzPatricks SE, that their modeling of aerosol settling is based on the methodology described in NRCs Office of Nuclear Regulatory Research (RES) AEB 03 13, with deviations. The NRC staffs SE also mentions that while AEB-98-03 uses a median distribution of aerosol settling velocities, FitzPatrick employed the 20-group probabilistic distribution to assess the MSL aerosol settling velocity, which directly influences the aerosol deposition. In their LAR, the licensee stated that this deviation from AEB-98-03 addresses a long-established NRC staff concern, documented in RIS 2006-004 14, with the selection of a single value for MSL aerosol settling velocity to evaluate the deposition of aerosol particles which, in reality, contain a wide range of sizes and weights. Specifically, the degree of MSL deposition is expected to decrease as the leakage progresses through the lines, since the larger and heavier aerosols would have already settled out of the MSLs in upstream sections of piping.

The NRC staff approved FitzPatricks use of the 20-group method based on best estimates available at the time and relying on previously approved LARs with similar approaches (i.e.,

Clinton, Limerick and LaSalle.) After issuing FitzPatricks amendment, the NRC staff learned that the 20-group method is not conservative for modeling the changing aerosol settling velocity distribution as it moves through the MSL volumes. The NRC staff is updating RG 1.183 15 (DG-1389) to include three options that licensees could use to more accurately calculate aerosol settling velocity physics parameters. This proposed revision also incorporates recommendations from an NRR/RES study 16. It is important to note that regulatory guides are not regulatory requirements and provide an acceptable way to meet NRC regulations.

Licensees could still propose alternative methodologies with proper justifications.

In our discussions, you also expressed concerns with the licensees assumptions that aerosol removal by sprays and aerosol removal in the MSLs are independent removal mechanisms. In the SE, the NRC staff acknowledged the interdependence of these two mechanisms.

Specifically, that larger aerosol particles in the containment atmosphere will be preferentially removed, making subsequent removal by deposition in downstream piping more challenging.

In their response to NRC staffs RAI on how the gravitational settling credited in the MSL considered the changing aerosol characteristics (i.e., size and density distributions) due to the preferential removal of larger aerosols because of the credit assigned to containment sprays, the licensee provided the results of a sensitivity analysis. This sensitivity analysis identified a base case and then adjusted this base case to consider the impact of sprays and other variables the licensee stated were conservatisms in the calculation. In their response, the 13 AEB 98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application using the Revised (NUREG-1465) Source Term, dated December 9, 1998. ADAMS Accession No. ML011230531.

14 RIS 2006-04, Experience with Implementation of Alternative Source Terms, dated March 7, 2006.

Available at https://www.nrc.gov/reading-rm/doc-collections/gen-comm/reg-issues/2006/ri200604.pdf.

15 DG-1389, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors. ADAMS Accession No. ML21292A319.

16 Informal Assistance Request 2020-11-30, Re-evaluation of the AEB 98-03 Monte Carlo Settling Velocity Description Parameters, dated March 19, 2021. ADAMS Accession No. ML21078A155.

licensee stated that the base case is essentially the JAF-CALC-19-00005, Revision 1 model including the nodalization adjustments and the revised aerosol removal factors. In their response, the licensee also stated that the base case, which resulted in a control room dose of 7.35 rem, indicates the conservative modeling of the drywell spray impact on the aerosol removal in the MSLs without adjusting for any other inherent conservatisms in the RADTRAD inputs (e.g. using only half of the reactor building volume, not crediting containment elemental iodine and particulate deposition/plateout and not crediting activity reduction by the steam separators or the iodine partitioning in the reactor vessel). It is important to note that the purpose of a sensitivity analysis is to demonstrate the impact of varying parameters on the results. As such, licensees may choose to use a range of parameters in these analyses. The fact that a case in the sensitivity analysis resulted in doses that were above the regulatory limits does not undermine the licensees approved analysis of record which showed that the doses were within regulatory limits.

The licensee did not modify their analysis of record as a result of this sensitivity analysis. In the FitzPatrick SE, the NRC staff documented their disapproval of portions of the sensitivity analysis (e.g., reduced control room operator breathing rate and impaction in the MSIV). Furthermore, the NRC staff concluded in the SE that [t]he analysis of record indicates that dose consequences comply with all applicable dose acceptance criteria. The NRC staff did not explicitly address the sensitivity analysis base case result which exceeded the regulatory limit when the impact of sprays in the particulate removal coefficient in the MSLs was accounted for.

The panel report explained that Footnote 1 to Regulatory Position 6.3 of RG 1.183 presents a position on crediting a reduction in airborne radioactivity in the containment by the containment spray system by stating: The removal rate constants selected for use in design-basis calculations should be those that will maximize the dose consequences. The panel concluded that the particulate removal coefficients assumed in the FitzPatrick calculation of record do not represent values that maximize the dose consequences. The panel also noted that the licensees sensitivity analysis included assumptions that the staff disagreed with and for which the basis were not documented. The panel also agreed with you that the analysis of record does not address the interdependence of aerosol removal and the use of drywell sprays. In addition, the panel concluded that the FitzPatricks MSL removal coefficients, as assumed in the calculation of record, do not represent values that maximize the dose consequences because the calculation of record does not incorporate the impact of drywell spray on aerosol deposition.

It is worth noting that at the time the FitzPatrick LAR was being reviewed, and in the present time for that matter, the NRC did not have a specific reduction criteria or any approved guidance licensees could use to address the impact of drywell sprays on deposition in the MSLs. DG-1389 indicates, conservatively, that sprays should not be credited when using any of the three methodologies described in the proposed regulatory guidance to compute removal coefficients in the MSL. It is also important to consider that, at this time, the current Revision 0 of RG 1.183 is not being rescinded. RG 1.183, Revision 0 continues to provide an acceptable option for licensees to use in their AST applications even after Revision 1 is issued. Even though RG 1.183, Revision 0 is silent on the appropriate use of drywell sprays when determining deposition in the MSLs, the substantial conservatisms contained in the guidance compensate for this deficiency.

I agree with you that the 20-group method overestimates the deposition in the MSL. I also agree with you and the panel that the analysis of record does not address the interdependence of aerosol removal and the use of drywell sprays. Our independent analysis, using the multi-group method (an acceptable option in the proposed DG-1389) and correcting for the other non-

conservatisms identified in the memorandum, resulted in control room doses that exceed the regulatory limit (5.11 rem). The dose result of 5.11 rem credits drywell sprays because, currently, an appropriate model for calculating MSL deposition when sprays are running has not been included in regulatory guidance. Drywell spray is a safety-related system and, therefore, it is reasonable to assume that sprays are initiated and remain operable during the postulated event. After a LOCA, the drywell sprays will assist in substantially lowering pressure in the drywell which will significantly reduce actual control room dose due to the decrease in leak rate for all pathways. This dose-reducing mechanism was not accounted for in the licensees analysis or in our independent analysis. This is an additional conservatism built into the licensees analysis. Drywell spray will also reduce the activity that is transported through the leakage pathways by removing heavier particles as described in the memorandum. Therefore, considering these substantial conservatisms and others noted in the memorandum, we have reasonable assurance the dose will remain below 5 rem. In addition, the staffs conclusions that the regulatory requirements were met, considering the licensees analysis, using methods that were appropriate at the time of the review, and accounting for the additional risk insights related to other conservatisms in the analysis, were appropriate. Consequently, backfitting the licensee and requesting a more accurate model for calculating MSL deposition when sprays are running is not justified because any further effort from the licensee would not provide an increase in safety as compared to the status quo.

(g) : You stated that FitzPatricks analysis used elemental iodine removal coefficients greater than 20 hr-1 in their LOCA analysis which is greater than the SRP Section 6.5.2 limit. You further stated that this deviation is discussed in the NRC staffs SE, but the SE does not assess the aggregate impacts of the deficiencies in the analysis described above.

SRP Section 6.5.2 states: [The elemental removal coefficient by spray] must be limited to 20 per hour to prevent extrapolation beyond the existing data for boric acid solutions with a pH of

5. In their LAR, the licensee proposed to credit control of the pH level above 7.0 in the suppression pool following a LOCA by means of injecting sodium pentaborate into the reactor core with the standby liquid control system. The licensee also proposed conforming changes to the TS to ensure pH level is maintained above 7.0. Regardless of pH control, licensees have historically used values that are lower than the value in the SRP.

The NRC staff evaluated the licensees deviation from the SRP value of 20 hr-1 by stating in the SE for FitzPatrick that the NRC staff performed confirmatory calculations using the RADTRAD program to assess the impact of including this limitation on the licensees dose analysis and concluded that the impact was not significant for this case. The NRC staff concluded that this deviation from the SRP is acceptable, it is case-specific and is only applicable to the FitzPatrick LAR.

The panel agreed with the submitters that the licensee did not justify the use of elemental iodine removal constant greater than 20 hr-1 and that the licensees analysis does not conform to the SRP.

Our independent assessment found that the NRC staffs confirmatory results referred to in the NRC staffs SE indicated that, had the licensee limited the elemental iodine removal constant to 20 hr-1, the control room dose would have increased by approximately 3 percent (or about 0.14 rem). It also appears that this issue was discovered late in the review process, during a peer review of the drafted SE and after the incorporation of the RAI responses. Since the impact on the calculated result was small, the NRC staff determined that an additional round of RAIs to resolve this issue would not be necessary to make a reasonable assurance determination. The

review team added language in the SE to discourage use of this assumption in subsequent LARs.

I agree with you and the panel that the licensee did not justify using the elemental iodine removal constant greater than 20 hr-1 in their analysis and that this is a deviation from SRP 6.5.2 and past applications. However, the NRC staff acknowledged this deviation and approved it by stating that the impact was not significant. In our independent analysis, we determined that using an elemental iodine removal constant of 20 hr-1, results in a control room dose increase of roughly 2% or 0.1 rem. Therefore, we also agree with the SEs finding that the licensees deviation from SRP 6.5.2 related to the elemental iodine removal constant, does not significantly impact the analysis. Consequently, we conclude that backfitting the licensee and requesting the use of a lower value is not justified.

Document 5: DPO Appeal Submittal

Differing Professional Opinion Appeal for DPO-2021-001 Continuation of NRC Form 690 Basis for the Appeal In Differing Professional Opinion (DPO) 2021-001, two U.S. Nuclear Regulatory Commission (NRC) senior experts in the areas of radiological engineering and licensing raised several concerns. 1 An independent and impartial DPO review panel, made up of experienced and senior-level employees knowledgeable in the relevant subject areas, performed a thorough, complete, and objective review of the DPO. 2 The DPO Panel substantiated most of the concerns raised in the DPO, as documented in the findings and recommendations in the DPO Panel report dated June 9, 2021. 3 However, each of the Directors decisions on DPO-2021-001 4 indicates that the Director disagreed with almost all of the findings and many of the recommendations made by the DPO review panel.

After an extensive consideration, we, the DPO submitters, have concluded that the Directors decision has many significant flaws in its technical rationale and statements used to disagree with the DPO and DPO Panel report. This DPO Appeal provides some of our concerns for consideration by the Executive Director for Operations (EDO).

Because of the significance of these flaws, we request that the EDO pursue an additional review of and response to the DPO issues. In the response to this DPO Appeal, we request that the EDO reconsider the information contained in the DPO and DPO Panel reports, including the DPO Panels recommendations that were rejected in the NRR Directors decision.

This DPO Appeal has identified the following recommendations that are within the scope of the DPO and the subsequent DPO Panel report:

• Adopt all the DPO Panels recommendations in its evaluation of the DPO. In a preliminary assessment of the DPO Panel recommendations, we provided comments on

1. One of the DPO submitters was honored in 2003 as the NRCs top engineer and one of the top 10 engineers in the Federal Government by the National Society of Professional Engineers. For more than 20 years, he has served as the NRC lead for the development of the accident analysis regulatory guidance and computer models related to this DPO. The other DPO submitter is a Branch Chief in the Division of Reactor Licensing for the last 13 of his 34 years at the NRC.

2 The DPO Panel was made up of three senior-level NRC employeesa member of the Senior Executive Service, one of the agencys few Senior Level Advisors for Risk, and a regional Branch Chief for engineering.

3 The DPO review panel report on DPO-2021-001 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML21160A232) is included in an enclosure to a memorandum dated June 9, 2021 (ADAMS Accession No. ML21160A234). We, the submitters of the DPO, agree with most all the conclusions and findings in the DPO Panel report and find it is consistent with the requirements in Management Directive and Handbook 10.159, NRC Differing Professional Opinion Program, dated August 11, 2015 (ADAMS Accession No. ML15132A664), to be independent, thorough, complete, and objective. In the few instances where the panel did not agree with the concerns raised in the DPO, we believe this is the result of misunderstandings that could have been remedied if the process allowed for more time to resolve these complex issues.

4 The Directors decision on DPO-2021-001 is provided in Enclosure 1 to a memorandum from the Director of the Office of Nuclear Reactor Regulation to the DPO submitters, dated August 27, 2021 (identified as Major Version 8 at ADAMS Accession No. ML21236A254). After it was identified that the decision did not address all the DPO issues, the decision was revised and reissued approximately 12 weeks later on November 17, 2021, under the same accession number. It is distinguished from the original issuance as Major Version 9, added to ADAMS on November 18, 2021.

1

specific changes to documents impacted by these recommendations to the Director in an e-mail on June 17, 2021. 5 These changes should be considered for implementation.

• Use the insights from the DPO Panel report and concerns raised in the DPO to correct the documents impacted by the issues raised in the DPO (i.e., safety evaluations noted in the DPO, Draft Regulatory Guide 1389 (DG-1389) (identified as Major Version 12 at ADAMS Accession No. ML21204A065) (a draft revision of Regulatory Guide 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, issued July 2000)).

• Do not allow the Directors decisions assessment to prejudge whether backfitting should be applied to the issues identified in the DPO. Task the NRC staff to use the backfit process to assess whether backfitting is warranted.

• Direct the NRC staff to not issue Interim Staff Guidance (ISG) DRA-ISG-2021-XX, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms, draft issued May 2021 (ADAMS Accession No. ML21078A051), as discussed below. This ISG is based upon the methods used in the flawed safety evaluation for the James A. FitzPatrick Nuclear Power Plant (FitzPatrick) that does not meet NRC regulations.

• Provide an EDO Appeal Decision that corrects errors in the Directors decision so that flawed information is not propagated into the licensing and engineering review processes.

In accordance with Management Directive and Handbook 10.159, the DPO Appeal should describe the basis clearly and succinctly and should focus on perceived flaws in the DPO Directors decision. This appeal summarizes the following significant perceived flaws:

Flaw 1 The Directors decision and directed assessment supporting the decision did not consider a limiting credible accident scenario that bypasses the pathway to the environment assumed in the license amendment and Directors decision. This scenario was previously mitigated by the design of the main steam leakage control system that is no longer credited in the FitzPatrick design.

Flaw 2 The Directors decision and directed assessment are based upon a model that credits deposition and holdup in the main steamline upstream of the inboard closed main steam isolation valve (MSIV). This modeling is based upon an incorrect statement in the Directors decision that Section 3.0 (i.e., Regulatory Guide 1.183, Appendix A, Regulatory Position 3) does not apply to the modeling of MSIV leakage.

Flaw 3 The technical rationale in the Directors decision used to disagree with the DPO submitters concern that the analysis of record does not consider the limiting atmospheric dispersion factors is inconsistent with the regulatory guidance used by the licensee to demonstrate compliance with Title 10 of the Code of Federal Regulations (10 CFR) 50.67, Accident source term.

5 The e-mail dated June 17,202 is not in ADAMS but is transmitted with this appeal.

2

Flaw 4 The technical rationale in the Directors report, used to disagree with many of the DPO issues and the findings and recommendations in the independent DPO Panel report, relies heavily on the results of modeling, assessments, and analysis of the control room doses that are characterized as being independent. These models, assessments, and analyses are not independent, nor are they impartial, because they to the FitzPatrick safety evaluation that is the subject of the DPO.

Flaw 5 Significant nonconservatism exists in the modeling of the nonlimiting release pathway assumed in the Directors decision. Despite the Directors decision acknowledging the interdependence between the aerosol removal in the steamline and the crediting of drywell sprays, the Directors decision model uses a steamline model that does not model the effect of the sprays on steamline deposition. The model also assumes a grossly nonconservative MSIV leakage source term aerosol size distribution that is inconsistent with state-of-the-art science. These flaws would significantly increase the estimates of the control room dose for this nonlimiting pathway.

Flaw 6 Arguments provided in the Directors decision Summary of My Decision prejudge whether backfitting screening and/or evaluation (i.e., a backfit assessment) is warranted and avoids consideration of backfit procedures and guidance. These arguments are misleading, inaccurate, and should not be used to bypass backfitting processes and procedures for the known errors in the licensees analyses and design basis.

Flaw 7 The Directors decision asserts that the condenser was not credited and the staff used engineering judgment and risk and engineering insights to support the deterministic conclusion and balance any uncertainties and nonconservatisms rather than as a basis for the decision. These statements conflict with information contained in the NRC staffs safety evaluation, meetings held with one of the DPO submitters, versions of the ISG that existed at the time the DPO was submitted, and extensive information discussed in Enclosure 1 to the DPO that provided supporting details for DPO Issue 1.

Flaw 8 The Directors decision alleges that conservatisms in the licensees model provide reasonable assurance that the regulations are met. These statements are based upon calculations that do not reflect the licensees design and licensing bases. Reasonable assurance of compliance with 10 CFR 50.67 is based upon the licensees licensing and design bases, not an analysis containing assumptions that the NRC deems appropriate. The Directors decision includes a contrived rationale that conflicts with the regulatory requirements in 10 CFR 50.67; staff review procedures in Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Revision 0, issued July 2000, of NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants: LWR Edition (SRP); and other statements made in the Directors decision.

Flaw 9 The Directors decision contains numerous additional flaws, inaccurate statements, and incorrect technical rationale, as described in the Enclosure to this DPO Appeal.

The enclosure to this appeal contains more details supporting the basis for these concerns.

3

Impacts of These Flaws on the Directors Decision The Directors decision is heavily based upon a directed assessment performed on behalf of the Director that conflicts with the results of the independent expert DPO Panel that reviewed the DPO. If Flaw 1 was addressed correctly in the Directors decision directed assessment, the results would reveal that the design-basis loss-of-coolant accident dose for the control room design greatly exceeds the dose limits in 10 CFR 50.67. If Flaws 1, 2, and 3 were corrected, the results of the Directors decision directed assessment would show that the dose limits in 10 CFR 50.67 are exceeded. The doses would be approximately a factor of 100 times greater than those calculated in the licensees analysis of record.

Additional Issues from the Directors Decision and DPO Panel Report Issue 1 In the section entitled Additional Issues, 6 the Directors decision stated that the Director did not have any objections with regard to the guidance in DG-1389 (a draft revision of Regulatory Guide 1.183) and DRA-ISG-2021-XX. The draft ISG and DG-1389 should be informed by the results of DPO-2021-001 just as DG-1389 was informed by DPO-2020-002. 7 However, those involved in approving DG-1389 and DRA-ISG-2021-XX, including the Office of the General Counsel and the Advisory Committee on Reactor Safeguards, may never see the DPO input or the results of the DPO Appeal during their reviews.

Making the ISG and DG-1389 reviewers aware of the DPO and the results of the independent Panels review is critical and consistent with the foundations of the DPO process. 8 The DPO contains issues that were largely confirmed by the independent DPO Panel. This confirmation showed that the methods used as the basis for developing the ISG and revising Regulatory Guide 1.183 (i.e., the FitzPatrick safety evaluation, which is the subject of DPO-2021-001),

indicate that the regulations were not met. The information concerning this DPO should be provided to those involved in decisions on these documents to inform their technical understanding and decision making. The ISG was based primarily upon the methods used to approve the FitzPatrick license amendment, which has fundamental technical and regulatory errors and is the subject document of concern in our DPO and this DPO Appeal. The FitzPatrick amendment was also used to inform DG-1389. Therefore, we request that the EDO delay the final approval of DRA-ISG-2021-XX and DG-1389 until the results of this DPO Appeal are 6

The Director decision states, in part, the following: During our discussions, you expressed concerns with draft Interim Staff Guidance (ISG) DRA-ISG-2021-XX and the recent draft revision to RG [Regulatory Guide] 1.183 by stating that these documents used the FitzPatrick approach and amendment as a model.

These documents are undergoing the necessary review process including a review by the Advisory Committee on Reactor Safeguards (ACRS). As supported by the conclusions in this document, I do not take objections with the guidance contained in either of these documents.

7 The case file for DPO-2020-002 (ADAMS Accession No. ML21067A645) provides the DPO decision that was used to inform and make several changes to Regulatory Guide 1.183 (see DG-1389).

8 Page 6 of Management Directive 10.159 states, in part, that the EDO publishes periodic announcements that affirm that the DPO Program strengthens the NRC and that the DPO Program is a potential source of valuable ideas. On page 3, it also states, in part, that the objective of the DPO process is to ensure that all employees have the opportunity to have their views heard and considered. Page 10 of the Directors decision states, in part, that an open and thorough exploration of how we carry out our regulatory processes is essential to the effectiveness of our programs and that raising concerns is vital to ensuring a healthy safety culture in the agency. Based upon these foundations, it is critical to provide the information in this DPO to other decisionmakers to inform their decisions and ensure that this potential source of valuable ideas is heard and considered by all involved in approving documents that are based upon the subject of the DPO.

4

completed and made available. Approval and processing of these documents should be withheld pending completion of the EDO Appeal and any associated tasking(s).

It is worthy to note that the method approved in the ISG circumvents the intent of 10 CFR 50.67 by allowing the NRC staff to approve license amendments, like the FitzPatrick amendment, that are flawed and do not provide reasonable assurance of meeting the regulations based on licensee information submitted on the docket. NRC approval of the Fitzpatrick AST amendment was based upon the staffs beliefs and conjecture rather than the merits of the licensees docketed analyses that contains errors. Formalizing this process will normalize a process that circumvents the regulations in 10 CFR 50.67.

Issue 2 The numerous flaws in the Directors decision could be referenced by licensees in future license amendment requests under 10 CFR 50.67. Also, because the Fitzpatrick AST amendment provides a precedent, licensees may use this as a basis for making improper changes under 10 CFR 50.59. In either case, licensees would likely make future changes that have an unrecognized adverse impact on dose and safety. This DPO Appeal requests the EDO to direct these flaws be corrected, if necessary, through backfit and/or Orders, to reconcile the errors before they become propagated throughout the nuclear industry analysis of record for applicable plants.

Issue 3 The Directors decision directed assessment replicated errors in the FitzPatrick licensees analysis of record. These errors represent safety-significant issues. It is difficult to argue otherwise as these errors allowed a safety-significant system (the main steam leakage control system) to be removed from the plant design and also allowed the relaxation of leakage requirements for the safety-related MSIVs (which provides a barrier for mitigating the release of fission products). If the EDO determines that the Directors decision contains the flaws described in this appeal, the NRC staff should be directed to consider backfit procedures or Orders for the license to fix these safety-significant issues.

5

Enclosure to the DPO-2021-001 Appeal Flaw 1 The Directors decision 1 and directed assessment supporting the decision did not consider a limiting credible accident scenario that bypasses the pathway to the environment assumed in the license amendment and Directors decision. This scenario was previously mitigated by the design of the main steam leakage control system (MSLCS) that is no longer credited in the design of the James A. FitzPatrick Nuclear Power Plant (FitzPatrick).

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision U.S. Nuclear Regulatory Commission (NRC) Management Directive and Handbook 10.159, NRC Differing Professional Opinion Program, dated August 11, 2015 (ADAMS Accession No. ML15132A664), states that the scope of a DPO Appeal must be limited to the scope of the DPO and the Summary of Issues. Therefore, for each flaw in this DPO Appeal, the appeal shows the relationship of the flaw to the DPO. The relationship of the flaw to the DPO review panel report 2 and the Directors decision is also provided.

Flaw 1 relates to a DPO issue documented in Section 2.2 (page 16) of the DPO. The issue relates to how the analysis of record loss-of-coolant accident (LOCA) does not consider credible release pathways to the environment. The DPO Panel reviewed this issue 3 and agreed with this concern, as documented in Section 4.1.2 (page 22) of the DPO Panel report. However, page 2 (paragraph 3) of the attachment to the Directors decision (item b) discusses the technical rationale used by the Director and the directed assessment to disagree with the DPO Panel and the submitters about this concern.

Summary of the Issue The limiting accident scenario must be considered in order to demonstrate that with the modifications requested (and now approved), the FitzPatrick facility would comply with the regulations (Title 10 of the Code of Federal Regulations (10 CFR) 50.67, Accident source term). 4 However, the Directors decision, the NRC staffs safety evaluation 5, and the licensees analyses did not consider a credible limiting accident scenario necessary to show compliance with 10 CFR 50.67.

1. The Directors decision on Differing Professional Opinion (DPO)-2021-001 is provided in Enclosure 1 to a memorandum from the Director of the Office of Nuclear Reactor Regulation to the DPO submitters, dated August 27, 2021 (identified as Major Version 8 in the Agencywide Documents Access and Management System (ADAMS) at Accession No. ML21236A254). After it was identified that the decision did not address all the DPO issues, the decision was revised and reissued on November 17, 2021, under the same accession number. It is distinguished from the original issuance as Major Version 9 added to ADAMS on November 18, 2021.

2 The DPO review panel report on DPO-2021-001 (ADAMS Accession No. ML21160A232) is included in an enclosure to a memorandum dated June 9, 2021 (ADAMS Accession No. ML21160A234).

3 Please note that DPO Issues 1 and 2 correlate to DPO Panel report Issues 2 and 1, respectively.

4 As stated in 10 CFR 50.67, The fission product release assumed for these calculations should be based upon a major accident, hypothesized for purposes of design analyses or postulated from considerations of possible accidental events, that would result in potential hazards not exceeded by those from any accident considered credible. [emphasis added]

5 The NRC staffs safety evaluation for FitzPatrick is contained in Enclosure 2 to the license amendment dated July 21, 2020 (ADAMS Accession No. ML20140A070).

1

Because MSIVs (stem packing, valve bonnets, and seats) and downstream piping and components (including the power conversion system (PCS) 8) cannot be made leaktight and are allowed some leakage during operations (and thus during accidents), the original design approved by the NRC included an MSLCS. The purpose of the MSLCS was to treat allowed MSIV leakage by creating a negative pressure to collect and filter this leakage before releasing it to the environment. 9 In accordance with Regulatory Guide (RG) 1.96, Design of Main Steam Isolation Valve Leakage Control Systems for Boiling Water Reactor Nuclear Power Plants, Revision 1, issued June 1976, a design objective of these systems was to reduce and control stem packing leakage or other direct leakage to the steam tunnel, because this leakage would escape to the turbine building and the environment through the steam tunnel. 10 In accordance with 10 CFR 50.36, Technical specifications, the technical specifications are derived from the analyses and evaluations included in the UFSAR. To justify the change to the FitzPatrick technical specifications, the licensee provided a revised UFSAR radiological assessment for a design-basis LOCA. 11 The NRC staff stated that its approval of these changes was based primarily on 10 CFR 50.67.

The original plant accident design credited the MSIVs and the MSLCS to protect public health and safety. In particular, Section 5 the original FitzPatrick safety evaluation report (ADAMS Accession No. ML19182A200) states the following:

Leakage through the closed main steam line isolation valves following a postulated LOCA presently relies on the low leakage characteristic of the valve

[at the time, 11.5 standard cubic feet per hour]. The acceptability of the present leakage limits and the need for an auxiliary sealing system are under study by 8 Section 1.6.1.4, Power Conversion System, of the FitzPatrick updated final safety analysis report (UFSAR) states the following: The unit utilizes a Power Conversion System which includes a turbine-generator, a main condenser, condensate pumps, a steam jet air ejector, Turbine Sealing System, Turbine Bypass System, condensate demineralizers, Reboiler System, condensate booster pumps, reactor feed pumps, feedwater heaters, drain coolers and Condensate Storage System to produce electrical power from the steam coming from the reactor, condense the steam into water, and return the heated feedwater to the reactor. The Circulating Water System removes the heat rejected to the main condenser. Section 1.6.1.4 appears on page 44 of 67 of the PDF version of UFSAR Chapter 1 (ADAMS Accession No. ML20038A346)

[not publicly available].

9 UFSAR Section 9.19.2, Safety Design Bases, states, in part, the following: The MSLCS is designed to collect and process leakage across the seats of the MSIVs and to collect and process stem packing leakage from the outboard containment MSIVs following a design basis LOCA [emphasis added]. The effluent of the MSLCS is processed by the Standby Gas Treatment System (SGTS) and is exhausted through the Stack.

The negative pressure in the SGTS is sufficient to provide the required flow through the MSLCS to collect all postulated leakage. The MSLCS is designed to meet the requirements of a Seismic Class I system, as defined in Section 12.2. Section 9.19 appears on page 174 of 190 of the PDF version of UFSAR Chapter 9 (ADAMS Accession No. ML20038A354) [not publicly available].

10 RG 1.96, Revision 1, states the following: It should be noted that any leakage from the stem packing of the outboard isolation valve would contribute to the 2-hour dose, since in most designs such leakage would escape to the turbine building and the environment via the steam tunnel [emphasis added]. Reduction and control of steam packing leakage or other direct leakage to the steam tunnel from the outboard isolation valve should be a design objective of the leakage control system or of other systems provided for this purpose. This quotation appears on page 2 of 5 of the PDF version of RG 1.96, Revision 1 (ADAMS Accession No. ML003740263).

11 UFSAR Section 9.19.1, Safety Objective, states, in part, the following: The safety objective of the Main Steam Leakage Collection System (MSLCS) is to collect and process leakage past the main steam isolation valves (MSIVs) following a Loss-Of-Coolant Accident (LOCA) so that resultant exposures are maintained below the values specified by 10 CFR 100. Section 9.19 appears on page 174 of 190 of the PDF version of UFSAR Chapter 9 (ADAMS Accession No. ML20038A354) [not publicly available].

3

the staff. There is nothing in the existing design which would preclude incorporation of an additional sealing feature if such is determined necessary.

The applicant will continue to study developments in this area.

Section 5.2.3 of Supplement 1 to the FitzPatrick safety evaluation report (ADAMS Accession No. ML15205A100) states the following:

On the basis of Staff calculations of the effects of main steam line isolation valve leakage and a continuing concern by the ACRS [Advisory Committee on Reactor Safeguards 12], PASNY [Power Authority of the State of New York] was requested to submit plans for installation of a supplementary sealing system [the MSLCS].

PASNY has committed, in a letter dated January 11, 1973, to demonstrate and install an acceptable sealing system at the time of the plants first refueling outage. 13 Discussion The NRCs safety analysis for the alternative source term amendment (AST) approved the removal of the MSLCS from the technical specifications. This change allowed the MSLCS to be capped off. 14 Therefore, the MSLCS is no longer available to perform its safety function to monitor and route leakage from the MSIV stem or bonnet to the SGTS for processing. 15 However, the Directors decision directed assessment continues to credit the SGTS despite the MSLCS no longer functioning because the assessment incorrectly states the stem (packing) leakage would be released to the reactor building rather than to the steam tunnel.

The outboard MSIVs are located in the steam tunnel. Now that the MSLCS is capped off, leakage from the outboard MSIV stem or bonnet, previously routed to the SGTS by the MSLCS, would leak into the steam tunnel. Consistent with RG 1.96, without the MSLCS stem (or bonnet), packing of the outboard isolation valve would escape to the turbine building and the environment through the steam tunnel. 16 It would not leak into the reactor building and be 12 In its letter on the construction permit review of the FitzPatrick plant (January 27, 1970), the ACRS noted that additional features to control MSIV leakage should be considered.

13 It is important to note that when the NRC staff performed these calculations, the regulation used to evaluate the FitzPatrick MSIVs and the MSLCS design was 10 CFR Part 100, Reactor Site Criteria. Unlike 10 CFR 50.67, 10 CFR Part 100 does not specify that the approval of the design must be based only on the applicants analyses.

14 According to page 2 of the attachment to the Directors decision, 3 months after the NRC approved the amendment, the FitzPatrick licensee isolated and capped off the piping from the MSLCS to the SGTS. This plant change is a result of the NRCs flawed safety evaluation for FitzPatricks amendment that did not fully address the impact on the accident analyses of removing the MSLCS.

15 UFSAR Section 4.11.4 describes the MSLCS functions. It states that the MSLCS was installed to properly monitor and route (by remote manual operations) the noncondensibles (i.e., gases that will not condense into a liquid within the temperatures expected) of the packing gland leak off of outboard MSIVs to the SGTS and the main condenser. Section 4.11.4 also references UFSAR Section 9.19. Section 4.11.4 appears on page 121 of 126 of the PDF version of UFSAR Chapter 4 (ADAMS Accession No. ML20038A349) [not publicly available]. Section 9.19 appears on page 174 of 190 of the PDF version of UFSAR Chapter 9 (ADAMS Accession No. ML20038A354) [not publicly available].

16 In accordance with RG 1.96, a design objective of these systems was to reduce and control stem packing leakage or other direct leakage to the steam tunnel, because this leakage would escape to the turbine building and the environment through the steam tunnel. (The location of the outboard MSIVs in the steam tunnel and the fact that the steam tunnel is open to the turbine building at FitzPatrick were confirmed in a conversation on September 9, 2021 with a former resident inspector of the FitzPatrick facility.) Consistent with RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear 4

processed by the SGTS, as assumed in the Directors decision directed assessment. If the Directors decision technical rationale were used and the MSIV leakage goes to the reactor building and then to the SGTS, the MSLCS would be unnecessary and would not have needed to be installed in the original plant design and credited in the accident analysis.

Rather than considering a credible release pathway from the MSIV stem or bonnet leakage to the turbine building through the steam tunnel, the Directors decision directed analysis modeling assumed all the MSIV leakage was seat leakage. Assuming all MSIV leakage is from seat leakage allows for crediting the removal of radioactivity in the steamline piping to the turbine stop valve, which, if appropriate, could significantly mitigate the release. However, this pathway would not be available to credit mitigation of the MSIV leakage if the leakage was entirely through the MSIV packing seals (stem and bonnet leakage). 17 Therefore, the Directors decision directed assessment, NRC staff safety evaluation, and the licensees calculations are flawed because they did not consider this credible accident scenario in which all the MSIV leakage is through the MSIV stem or bonnet. When this scenario is considered, calculations using the method given in RG 1.183 indicate that the accident dose criterion in 10 CFR 50.67 would be exceeded by a large margin.

Even if all the leakage were assumed to be through the MSIV seats, crediting the deposition in these volumes downstream of the MSIVs is not consistent with the guidance in RG 1.183 the licensee said it used to show compliance with 10 CFR 50.67, and the justification for crediting this pathway is incomplete. These issues are discussed below.

Assessment of Compliance with Applicable Regulatory Positions in Regulatory Guide 1.183 The licensee used RG 1.183 to show compliance with 10 CFR 50.67. On page 61 of of the license amendment request dated August 8, 2019 (ADAMS Accession No. ML19220A043), the licensee showed how its analyses conform to RG 1.183. The licensee stated that the analyses comply with Regulatory Positions A-6.4 and A-6.5, which are directly related to Flaw 1.

RG 1.183, Regulatory Position A-6.4, states the following:

In the absence of collection and treatment of releases by ESFs such as the MSIV leakage control system, or as described in paragraph 6.5 below, the MSIV leakage should be assumed to be released to the environment as an Power Reactors, Revision 0, issued July 2000, Regulatory Position 5.1.2, since the steam tunnel is not safety-related and required to be operable by technical specifications, it would not be credited. This is further reinforced in RG 1.183, Appendix A, Regulatory Position 6.4, which states the following: In the absence of collection and treatment of releases by ESFs [engineered safety features] such as the MSIV leakage control system, or as described in paragraph 6.5 below, the MSIV leakage should be assumed to be released to the environment as an unprocessed, ground-level release. Holdup and dilution in the turbine building should not be assumed.

17 Page 22 of the DPO Panel report (ADAMS Accession No. ML21160A232) states the following: The panel noted the leak rate test cannot distinguish where the actual leak originates, making it possible that all leakage is through the stem pathway. The licensee did not assess this leakage pathway nor provide justification that leakage through this pathway would not have detrimental impact on dose rates. On page 49 of Attachment 2 to the license amendment request, dated August 8, 2019 (ADAMS Accession No. ML19220A043), the licensee states that their analysis conforms to RG 1.183, Regulatory Position 5.1.3. This regulatory position states, in part, that: The numeric values that are chosen as inputs to the analyses required by 10 CFR 50.67 should be selected with the objective of determining a conservative postulated dose. Contrary to the licensees stated conformance to Regulatory Position 5.1.3, the licensees inputs to their analysis did not reflect the leakage path through the MSIV valve packing (stem and bonnet leakage) which would provide a conservative postulated dose.

5

unprocessed, ground-level release. Holdup and dilution in the turbine building should not be assumed.

RG 1.183, Regulatory Position A-6.5, states the following:

A reduction in MSIV releases that is due to holdup and deposition in main steam piping downstream of the MSIVs and in the main condenser, including the treatment of air ejector effluent by offgas systems, may be credited if the components and piping systems used in the release path are capable of performing their safety function during and following a Safe Shutdown Earthquake (SSE). The amount of reduction allowed will be evaluated on an individual case basis. References A-9 and A-10 provide guidance on acceptable models. , page 61, of the license amendment request also states the following:

Since [the] MSLC system is no longer credited, no ESFs are assumed to be available to collect or treat MSIV leakage. Releases are assumed to be from the Seismic Class I Turbine Stop Valves (TSVs) without credit for holdup or dilution in the condenser or turbine building [emphasis added]. The release is treated as a ground level release for dose assessment.

It also includes the following statements:

Main steam piping downstream of the MSIVs to the TSVs is credited [emphasis added]. No credit is taken for holdup and deposition in piping downstream of the TSVs, or in the condenser.

The licensees analyses do, in part, seem to conform to Regulatory Position A-6.4 in that, with no ESF credited, the releases enter the environment as ground-level releases. However, they do not conform with Regulatory Position A-6.5 for crediting a reduction of the MSIV releases in the main steam piping downstream of the MSIVs, as explained below.

RG 1.183, Regulatory Position A-6.5, allows for a reduction in MSIV leakages due to holdup and deposition in the main steamline piping downstream of the MSIVs and in the condenser if the components and piping systems are capable of performing their safety functions. Regulatory Position A-6.5 also states that acceptable models providing guidance for using this regulatory position are contained in a safety evaluation for a Boiling Water Reactor Owners Group report noted as Reference A-1018 in Appendix A to RG 1.183.

As discussed above, the FitzPatrick licensee cited conformance with Regulatory Position A-6.5 as the reason for crediting holdup and deposition in the steamline piping up to the turbine stop valve. However, it did not cite Reference A-10 as the guidance it used, and the alternative method used did not provide a complete justification for crediting deposition in this downstream piping. To understand why this justification was not complete, one must understand the basis for crediting piping downstream of the MSIVs in the context of this regulatory position.

Reference A-10 provides guidance on how to credit deposition and holdup in an alternative leakage pathway and in the condenser. The alternative leakage pathway provides a pathway to 18 Safety Evaluation of GE Topical Report, NEDC-31858P (Proprietary GE report), Revision 2, BWROG Report for Increasing MSIV Leakage Limits and Elimination of Leakage Control Systems, September 1993, dated March 3, 1999 (ADAMS Accession No. 9903110303) 6

the condenser, which is at lower pressure than the MSIV leakage, ensuring the MSIV leakage pathway is down the pathway credited in the accident analysis. This is the same method used when crediting the MSLCS, which in this design previously created a negative pressure to ensure the MSIV leakage was routed through the pathway credited in the previous FitzPatrick accident analysis of record.

Reference A-10 provides numerous staff conditions for crediting an alternative leakage pathway, including the existence of a maintenance and testing program for the active components (such as valves) in the alternative drain path, demonstration of how the alternative drain pathway addresses the single failure of active components to verify its availability to convey MSIV leakage to the condenser, the existence of a secondary path to the condenser, emergency operating procedures to identify necessary operator actions to mitigate MSIV leakage consequences using the alternative drain pathway, and assurances that the plants inservice testing program includes the valves required to open the drain path to the condenser.

The FitzPatrick licensee did not use Reference A-10 and did not credit a pathway to the condenser. 19,20 The licensees perceived compliance with Regulatory Position A-6.5 only seems to focus on the fact that the piping up to the turbine stop valve is seismic Class 1. 21 While the structure of the piping could be credited, there is no justification to assume the MSIV leakage will continue to flow down the piping to the turbine stop valves throughout the duration of the accident. The analysis provides no assurance that are even close or comparable to those called for by Reference A-10 or for a safety-related system such as the MSLCS to ensure the flow goes down the pipe downstream of the MSIVs. In fact, if the turbine stop valves function as designed and close as assumed in these analyses, 22 and there is no leakage from the turbine stop valves, 23 the pressure downstream of the MSIVs would increase to the point that the 19 Slide 32 of the NRC staffs presentation on DRA-ISG-2021-XX that was given to the ACRS on July 23, 2021 (ADAMS Accession No. ML21223A034), provides information on four license amendment requests that included the FitzPatrick submittal. The slide states that, for these analyses, A pathway to the condenser was not credited in the analyses of record (AOR).

20 In several places, the Directors decision states that neither the NRC staff nor the licensee credited the PCS or the condenser to make a regulatory finding of adequate protection. Page 5 of the Directors decision states the following: In the SE [safety evaluation], the NRC staff did not credit the PCS or the main condenser to make the regulatory finding of adequate protection. Page 5 states the following: My assessment that the NRC staff did not credit the PCS or the main condenser to get reasonable assurance of adequate protection is supported by the fact that the licensees analysis of record, JAF-CALC-19-0005, Revision 1, Post-LOCA EAB, LPZ, and CR DoseAST Analysis, does not credit these non-safety-related SSCs [structures, systems, and components] for dilution, holdup and retention. Page 6 states the following:

Under Issue 2, I explain that the NRC staff did not credit the PCS or the main condenser to make the regulatory finding of adequate protection.

21 In addition to the comments in Attachment 2 cited above about compliance with RG 1.183, Attachment 1, page 23, of the license amendment request dated August 8, 2019 (ADAMS Accession No. ML19220A043),

states the following: All four MSL [main steamline] headers are seismically qualified from the RPV [reactor pressure vessel] nozzle to the seismic boundary break at the TSV; therefore, they are qualified to withstand the SSE [safe-shutdown earthquake], and they comply with the RG 1.183, Appendix A, Section 6.5 requirement to be credited for aerosol deposition. Therefore, the MSIV leakage pathway boundary is extended up to the TSV.

22 Figure 3 on page 75 of Calculation No. JAF-CALC-19-00005, contained in Attachment 6 to the license amendment request dated August 8, 2019 (ADAMS Accession No. ML19220A043), shows that the analysis performed by the FitzPatrick licensee credited the volume upstream of closed turbine stop valves (volumes V2 and V4 in the drawing), and the turbine stop valves are closed.

23 Such an assumption is consistent with RG 1.183, Regulatory Position 5.1.1, which states the following:

These design basis analyses were structured to provide a conservative set of assumptions to test the performance of one or more aspects of the facility design. Many physical processes and phenomena are represented by conservative, bounding assumptions rather than being modeled directly. It is also consistent 7

blocked pathway and pressure downstream would inhibit flow down the pipe downstream of the outboard MSIV. Any inboard MSIV seat leakage would then go out the allowable outboard MSIV valve stem or bonnet pathways, bypassing the volumes downstream of the outboard MSIV. 24 Therefore, the flawed FitzPatrick analysis and assessment performed for the Directors decision directed assessment do not provide assurance or confirmation that the leakage is directed down the pathway credited in these safety analyses.

Because credit for deposition downstream of the MSIVs is not justified, Regulatory Position A-6.4 would be used to model the credible release pathway from the MSIV stems and bonnet. This credible pathway would yield design doses well in excess of 10 CFR 50.67, and the NRC should not have approved the license amendment. In addition, the directed assessment for the Directors decision relies heavily upon the results of an analysis that is flawed because it did not consider the credible release pathway through the MSIV stem and bonnet.

It is also important to note that the staffs alleged risk and engineering insights 25 would have no bearing on the accident scenario where all the MSIV leakage goes through the MSIV packing seal because this scenario would bypass the PCS and condenser credited in the risk and engineering insights.

Flaw 2 The Directors decision and directed assessment are based upon a model that credits deposition and holdup in the main steamline upstream of the inboard closed MSIV. This modeling is based upon an incorrect statement in the Directors decision that Section 3.0 (i.e., RG 1.183, Appendix A, Regulatory Position 3) does not apply to the modeling of MSIV leakage.

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision This flaw relates to a DPO issue documented in Section 2.1 (page 14) of the DPO. The issue concerns how the analysis of record LOCA does not conform with the regulatory guidance used to model the source of the MSIV leakage. These inconsistencies underestimate the LOCA analysis releases and, thus, underestimate the design-basis accident doses in the analysis of record. The DPO Panel reviewed this issue and did not substantiate the concern, as documented in Section 4.1.1 (page 20) of the DPO Panel report. 26 However, page 1 of the with Regulatory Position 5.1.3, which states the following: The numeric values that are chosen as inputs to the analyses required by 10 CFR 50.67 should be selected with the objective of determining a conservative postulated dose. In some instances, a particular parameter may be conservative in one portion of an analysis but be nonconservative in another portion of the same analysis. As a conservative alternative, the limiting value applicable to each portion of the analysis may be used in the evaluation of that portion.

Assuming no leakage in the turbine stop valve is selected with the objective of determining a conservative postulated dose.

24 Note that when performing these analyses, the various failure modes of the TSV must be considered to find the limiting release pathway. The FitzPatrick analysis assumed that the TSVs would leak to the environment.

A more limiting scenario exists when the TSVs are assumed not to leak. This scenario would inhibit any MSIV seat leakage from flowing through the credited piping downstream of the MSIVs and would consider leakage through more limiting upstream pathways such as through the valve stem packing or bonnet.

25 It is important to note that the NRC staff did not perform a risk analysis and did not use any risk insights beyond using operational experience to argue the seismic robustness of the condenser. The only engineering insights appear to be those that deviated and removed conservatism from the RG 1.183 methods used by the licensee to show compliance with the regulations.

26 Based on a discussion with a DPO Panel member, it is our understanding that the panel did not substantiate the concerns because it did not have enough time to complete its review of all the issues.

8

attachment to the Directors decision discusses the Directors agreement that the MSIVs are part of the containment boundary but disagreed with the DPO about this concern.

Background

Page 1, paragraph 4, of the attachment to the Directors decision states that a reduced activity (due to deposition) in the volume upstream of the modeled closed MSIVs is allowed to account for natural deposition. The Directors decision states the following:

During further discussions with you [DPO submitters] to clarify this concern, you stated that since the MSIVs are part of the drywell, in order for the drywell to represent an instantaneous and homogeneously mixed volume, the activity upstream the MSIVs should be the same as in the drywell. You added that this is required by RG 1.183 Appendix A, Regulatory Position 3.1 which states the radioactivity released from the fuel should be assumed to mix instantaneously and homogeneously throughout the free air volume of the primary containment in pressurized water reactors or the drywell in BWRs as it is released. However, RG 1.183 Appendix A, Section 3.0 [Regulatory Position 3 of Appendix A],

Assumptions on Transport in Primary Containment, provides criteria for licensees to consider when evaluating containment leakage. The criteria to consider when evaluating MSIV leakage is provided in Section 6.0, Assumptions on Main Steam Isolation Valve Leakage in BWRs. RG 1.183 Appendix A, Section 6, allows credit, on an individual case basis, for a reduction of the amount of released radioactivity on steam system piping upstream of the outboard MSIVs by deposition and plateout [emphasis added].

Page 1 also states the following:

I agree that the MSIVs are considered part of the containment boundary.

[emphasis added]. However, for the purposes of the alternative source term (AST) analysis, Section 3.0 [Regulatory Position 3] of RG 1.183 does not apply to the modeling of MSIV leakage [emphasis added] and therefore, a reduced activity in the volume of piping upstream of the modeled closed MSIVs to account for natural deposition and removal by containment sprays is allowed.

Discussion The reasoning in the Directors decision that Section 3.0 (i.e., Regulatory Position 3 of Appendix A to RG 1.183) does not apply to the modeling of MSIV leakage is flawed. RG 1.183, Appendix A, Regulatory Position 6.1, is contained within the section entitled Assumptions on Main Steam Isolation Valve Leakage in BWRs. Regulatory Position A-6.1 refers directly to Regulatory Position 3 in Appendix A, stating the following:

For the purpose of this analysis, the activity available for release via MSIV leakage should be assumed to be that activity determined to be in the drywell for evaluating containment leakage (see Regulatory Position 3) [emphasis added].

No credit should be assumed for activity reduction by the steam separators or by iodine partitioning in the reactor vessel.

Appendix A, Regulatory Position 3, in the section entitled Assumptions on Transport in Primary Containment, states the following:

9

Acceptable assumptions related to the transport, reduction, and release of radioactive material in and from the primary containment in PWRs or the drywell in BWRs are as follows:

The information that follows includes Appendix A, Regulatory Position 3.1. As such, the technical rationale in the Directors decision is fundamentally flawed. Regulatory Position 3.1 is not separate and distinct from the method in RG 1.183 for modeling MSIV leakage; rather, it is an integral part of this method. However, neither the licensee nor the Directors decision directed assessment used the guidance in Regulatory Position A-3.1. This error results in an underestimation of the radiological doses in the licensees analysis of record and the Directors decision directed assessment.

Flaw 3 The technical rationale in the Directors decision used to disagree with the DPO submitters concern that the analysis of record does not consider the limiting atmospheric dispersion factors is inconsistent with the regulatory guidance used by the licensee to demonstrate compliance with 10 CFR 50.67.

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision This flaw relates to a DPO issue documented in Section 2.4 (page 19) of the DPO. The issue relates to how the FitzPatrick analysis of record LOCA does not conform with the regulatory guidance used to model the limiting release pathways to the environment. The DPO Panel documented its review of this issue in Section 4.1.4 (page 25) of the DPO Panel report. On page 27 of the DPO Panel report, the Panel agreed with the DPO submitters that the licensee did not provide a bounding source-to-receptor combination; therefore, the atmospheric dispersion factors are not assured. However, on page 4 of the attachment to the Directors decision, the Director disagreed with the Panel on this concern.

Discussion Flaw 3 relates to how the analysis of record LOCA does not conform with the regulatory guidance used to model the limiting release pathways to the environment. These inconsistencies underestimate the atmospheric dispersion factors used in the LOCA analysis and, consequently, underestimate the design-basis accident doses used to approve the amendment. The DPO Panel reviewed this issue and agreed with the DPO concern on this issue, as documented in Section 4.1.4 (page 25) of the DPO Panel report. However, page 4 of the attachment to the Directors decision discusses the reasons the Director disagreed with the DPO Panel and the submitters about this concern:

I disagree with the panel and this specific concern. The three pathways mentioned by the licensee are all part of the TB [turbine building], which the analysis assumes is lost. If the pathways were to maintain their integrity post-LOCA, the licensee could make the argument that deposition in the TB could be credited and this would balance the dose results [emphasis added].

Even if deposition in the TB is not credited, the following judgement could be made on the acceptability of these pathways to transport activity to the control room intake: the smoke ejector vents are normally closed and would not open in a post-LOCA scenario; the access door is normally closed, locked, and monitored by security; and, the TB exhaust duct, which is closer to the control room intake than the TSVs, has a gravity damper which only opens when the ventilation fan is in operation. Since the most limiting scenario assumes a loss of 10

offsite power (LOOP) and the TB ventilation fan is not powered by the emergency buses, it is reasonable to assume the damper is closed during this event [emphasis added]. In addition, while these three pathways are not safety-related and are not in TS [technical specifications], they are not being credited for accident mitigation but rather assumed to retain their normal operation position.

During follow up discussions with you on this concern [emphasis added], you indicated that the licensee should assume the most conservative scenario, and this includes assuming scenarios beyond what the licensee and the NRC have determine to be conservative and appropriate. In this specific case, you believe the licensee should assume a LOOP does not occur and therefore the gravity damper is open and available as a release pathway. Similar to concern 1(c),

such assumptions would result in using parameters that conflict with the postulated scenario that was determined to be conservative and appropriate.

The above justifications for disagreeing with the concern and the DPO Panel report are flawed for the following reasons:

• The statements that deposition in the TB could be credited and this would balance the dose results and that the most limiting scenario assumes a loss of offsite power are unsupported subjective analytical judgment.

• The statement that the licensee could make the argument that the deposition in the TB could be credited is inconsistent with the assumption made by the licensee in the analysis that the turbine building is lost. 27 In accordance with 10 CFR 50.67, the approval of the amendment must be based upon the licensees analysis, and this assumption is not part of the licensees analysis.

• The above statements about judgments that can be made and what is reasonable to assume about crediting the integrity of non-safety barriers (such as the smoke ejector vents, access doors, and turbine building exhaust) and, therefore, not considering them as release pathways directly conflict with RG 1.183, Regulatory Position 5.1.2 and Regulatory Position A-6.4, 28 used by the licensee to demonstrate compliance with 10 CFR 50.67. To credit these non-safety-related systems and barriers and not consider releases through them, although such releases could be consistent with their normal operation, would provide credit for non-safety-related systems that is equivalent to that for a safety-related system required by technical specifications to be operable. The above statement that these three pathways are not safety-related and are not in TS, they are not being credited for accident mitigation but rather assumed to retain their normal operation position is, therefore, contradictory. In performing deterministic safety 27 Page 8 of 26 of Calculation No. JAF-CALC-19-00005, contained in Attachment 6 to the license amendment request dated August 8, 2019 (ADAMS Accession No. ML19220A043), states that the Turbine Building is lost, and the release is directly from the TSVs to the Environment.

28 Calculation No. JAF-CALC-19-00005, contained in Attachment 6 to the license amendment request dated August 8, 2019 (ADAMS Accession No. ML19220A043), provides a table entitled Conformance with Regulatory Guide 1.183. In that table, on pages 86 and 96 of the calculation, the licensee stated that the analysis conforms with RG 1.183, Regulatory Position 5.1.2 and Regulatory Position A-6.4. Regulatory Position 5.1.2 states, in part, the following: Credit may be taken for accident mitigation features that are classified as safety-related, are required to be operable by technical specifications, are powered by emergency power sources, and are either automatically actuated or, in limited cases, have actuation requirements explicitly addressed in emergency operating procedures. Regulatory Position A-6.4 states, in part, the following: Holdup and dilution in the turbine building should not be assumed. Holdup is necessary for deposition to occur.

11

analyses, no credit is taken for non-safety equipment unless it is conservative to do so.

Crediting the normal operating position rather than positioning the system in the potential open condition is to credit the operation of the non-safety--related equipment as a mitigation barrier.

• The above statements about follow-up discussions do not consider the full content of information in the DPO and the DPO Panel report. Statements made about the consideration of the LOOP being beyond what the licensee and the NRC have determined to be conservative and appropriate and that DPO concern 1(c) would conflict with the postulated scenario that was determined to be conservative and appropriate are based upon opinions that are inconsistent with RG 1.183, Regulatory Position 5.1.2, used by the licensee to demonstrate compliance with the regulations.

Regulatory Position 5.1.2 states that the occurrence and timing of a LOOP should be selected with the objective of maximizing the postulated radiological consequences and only credits accident mitigation features that are classified as safety- related and meet other conditions.

However, the perception that the analysis methods in RG 1.183 are conservative is correct. RG 1.183, Regulatory Position 5.1, provides a perspective on the RG 1.183 methods pertinent to this discussion:

These design basis analyses [in RG 1.183] were structured to provide a conservative set of assumptions to test the performance of one or more aspects of the facility design. Many physical processes and phenomena are represented by conservative, bounding assumptions rather than being modeled directly. The staff has selected assumptions and models that provide an appropriate and prudent safety margin against unpredicted events in the course of an accident and compensate for large uncertainties in facility parameters, accident progression, radioactive material transport, and atmospheric dispersion. Licensees should exercise caution in proposing deviations based upon data from a specific accident sequence since the DBAs [design-basis accidents] were never intended to represent any specific accident sequencethe proposed deviation may not be conservative for other accident sequences.

In addition, it is common to consider the impact of both LOOP and non-LOOP conditions and to not credit non-safety-related systems when performing design-basis calculations.

Such assumptions are not only used in the dose analyses but also in other accident analyses. These assumptions are used to find the limiting accident conditions.

Flaw 4 The technical rationale in the Directors report, used to disagree with many of the DPO issues and the findings and recommendations in the independent DPO Panel report, relies heavily on the results of modeling, assessments, and analysis of the control room doses that are characterized as being independent. These models, assessments, and analyses are not independent, nor are they impartial, because they were to the FitzPatrick safety evaluation that is the subject of the DPO.

12

Flaw 5 Significant nonconservatism exists in the modeling of the nonlimiting release pathway assumed in the Directors decision. Despite the Directors decision acknowledging the interdependence between the aerosol removal in the steamline and the crediting of drywell sprays, the Directors decision model uses a steamline model that does not model the effect of the sprays on steamline deposition. The model also assumes a grossly nonconservative MSIV leakage source term aerosol size distribution that is inconsistent with state-of-the-art science. These flaws would significantly increase the estimates of the control room dose for this nonlimiting pathway.

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision This flaw relates to a DPO issue documented in Section 2.6 (page 22) of the DPO. The issue relates to how the analysis of record LOCA grossly overestimated the amount of deposition in the steamlines. These errors underestimate the LOCA analysis releases and, consequently, underestimate the design-basis accident doses used to approve the amendment. The DPO Panel reviewed this issue in Section 4.1.6 (page 29) of the Panel report. On page 34 of the Panel report, the DPO Panel concluded that the FitzPatrick main steamline removal coefficients as assumed in the calculation of record do not represent values that maximize the dose consequences, because the calculation of record does not incorporate the impact of drywell spray on aerosol deposition. On page 7 of the attachment to the Directors decision, the Director agreed with this concern that the method used overestimated the deposition in the main steamline, but then used flawed methods to assess the control room accident dose for a nonlimiting pathway to inform the Directors decision.

Background

Pages 7 and 8 of the attachment to the Directors decision state, in part, the following:

I agree with you that the 20-group method overestimates the deposition in the MSL. I also agree with you and the panel that the analysis of record does not address the interdependence of aerosol removal and the use of drywell sprays.

Our independent analysis, using the multi-group method (an acceptable option in the proposed DG-1389)resulted in control room doses that exceed the regulatory limit (5.11 rem) [emphasis added]. The dose result of 5.11 rem credits drywell sprays [emphasis added] because, currently, an appropriate model for calculating MSL deposition when sprays are running has not been included in regulatory guidance.

The NRC staffs RAI ARCB-RAI-2 further describes this interdependency in the analysis of record, stating, in part, the following: 32 The issue of how the shift in the aerosol size because of drywell sprays would impact assumptions made in the subsequent main steam line aerosol deposition was discussed in the pre-application meeting held on June 20, 2019, (ADAMS Accession No. ML19183A128). As stated in the meeting summary, Since the precedent cited for the proposed main steam line aerosol deposition did not include drywell sprays, the licensee should consider including a detailed 32 The FitzPatrick RAIs are located at ADAMS Accession No. ML20035D576.

14

discussion of how the use of sprays is accounted for in the subsequent steam line aerosol deposition.

From an examination of the submitted information it appears that the licensee considers the aerosol removal by sprays and aerosol removal in the main steam lines as independent removal mechanisms. The NRC staff notes that regardless of the specific removal mechanisms involved, larger aerosol particles in the containment atmosphere will be the preferentially removed therefore making subsequent removal by deposition in downstream piping more challenging

[emphasis added].

Appendix A to DG-1389 (a draft revision of RG 1.183 currently under development) contains a multigroup model. The model does not consider the impact of sprays and, therefore, DG-1389, Regulatory Position A-5, states that the models are not valid if credit has been taken for aerosol removal from drywell sprays.

In addition, DG-1389, Regulatory Position A-5.6.2 cites a range of aerosol sizes for the reactor coolant system from the State-of-the-Art Report on Nuclear Aerosols (SOAR), dated December 17, 2009. 33 The SOAR contains our best available knowledge about the sizes of the aerosols for a light-water-reactor accident. The SOAR recommends a log-normal distribution for aerosols in the reactor coolant system (aerodynamic mass median diameter (AMMD) of 1.0 m with of 2.0). Larger values are were also measured for the containment (AMMD of 3.0 m and of 2.0). Since the MSIVs are connected to the reactor coolant system and not directly to the containment located outside the reactor coolant system 34, and 10 CFR 50.67 requires an accident dose that is not to be exceeded by any accident considered credible, 35 the smaller aerosols for reactor coolant system are most appropriate for modeling this accident. The smaller sized aerosols in the reactor coolant system will yield higher doses since they are harder to remove and slower to be removed from the leakage to the environment.

An e-mail dated December 8, 2021 36, about specific details of the Directors analysis stated that the aerosol deposition in the steamline is based upon an MSIV leakage source term that is assumed to have an AMMD of 3.0 m with of 2.

Discussion As described with respect to Flaw 1, the Directors decision directed assessment is based upon a flawed model that only analyzes the MSIV seat leakage pathway, which is not the limiting scenario. In addition, the Directors decision directed analysis has other flaws that are nonconservative. In the cited quotations in the Background section above, the Directors 33 NEA/CSNI/R(2009)5, State-of-the-Art Report on Nuclear Aerosols, Nuclear Energy Agency, Committee on the Safety of Nuclear Installations, December 17, 2009.

34 Attachment 1 to the August 8, 2019 license amendment request (Section 3.11.9 on page 20) describes how the MSIVs are functionally part of the primary containment boundary. 10 CFR 50.2, Definitions states: For nuclear power reactors of the direct cycle boiling water type, the reactor coolant system extends to and includes the outermost containment isolation valve in the main steam and feedwater piping. Therefore, for the purposes of these design basis calculations, the reactor coolant system is part of the containment, 35 As stated in 10 CFR 50.67, The fission product release assumed for these calculations should be based upon a major accident, hypothesized for purposes of design analyses or postulated from considerations of possible accidental events, that would result in potential hazards not exceeded by those from any accident considered credible [emphasis added].

36 The e-mail dated December 8, 2021 is not in ADAMS but is transmitted with this appeal [not publicly available].

15

decision acknowledges the interdependence between the aerosol removal in the steamline and the crediting of drywell sprays. However, the directed assessment supporting the Directors decision used a steamline aerosol deposition model from DG-1389 that does not model the effect of sprays. If sprays are modeled, the downstream deposition would be more challenging 37 (less removal) and, therefore, doses would be larger.

In addition, the e-mail cited above in the Background section states that the source of the aerosols for the MSIV leakage in Directors decision model is based upon the assumption that the aerosols have an AMMD of 3.0 m and of 2.0. Since the MSIVs are connected to the reactor coolant system that is considered to be part of the containment and because 10 CFR 50.67 requires an accident dose that is not to be exceeded by any accident considered credible, the assumed size (AMMD of 3.0 m and of 2.0) is very nonconservative. The SOAR recommends a log-normal distribution for aerosols in the reactor coolant system (AMMD of 1.0 m with of 2.0). Since these aerosols connected to the MSIV are much smaller than those assumed in the Directors decision directed assessment analysis model (AMMD of 1.0 m vs.

the 3.0 m assumed in the Directors decision model) and the deposition rate is inversely proportional to the square of the aerosol sizes, this nonconservatism in the Directors decision is very large.

In addition, page 8 of the attachment to the Directors decision provides additional information and reasons for crediting the sprays, along with a model that does not consider the effects of the sprays. That information and reasoning is, however, flawed for the following reasons:

• The attachment notes that the draft guidance does not have an appropriate (aerosol) model for calculating main steamline deposition when sprays are running. However, the absence of an appropriate model in DG-1389 (a proposed draft revision to RG 1.183) does not justify the use of an invalid model. RG 1.183 has never contained an explicit model for aerosol deposition in the main steamlines, but licensees have proposed, and the NRC has reviewed, models for this deposition.

• The attachment notes that drywell sprays are a safety system and, therefore, it is reasonable to credit them. That may be true, but it does not justify use of an invalid and nonconservative model when drywell sprays are credited.

• The attachment notes that the sprays will lower the pressure in the drywell and decrease the leak rate for all pathways. It is true that the sprays will reduce the pressure, however, it is improper to use this reduction to lower the design-basis dose or determine whether a backfit is appropriate because the licensees docketed analysis of record for the allowable containment releases is not consistent with the reduction credited in the Directors decision. The discussion of Flaw 8 gives more detail on our reasoning for this disagreement.

37 Taken from the FitzPatrick RAIs (ADAMS Accession No. ML20035D576). The NRC staffs RAI ARCB-RAI-2 states: The NRC staff notes that regardless of the specific removal mechanisms involved, larger aerosol particles in the containment atmosphere will be the preferentially removed therefore making subsequent removal by deposition in downstream piping more challenging.

16

Flaw 6 Arguments provided in the Directors decision Summary of My Decision prejudge whether backfitting screening and/or evaluation (i.e., a backfit assessment) is warranted and avoids consideration of backfit procedures and guidance. These arguments are misleading, inaccurate, and should not be used to bypass the backfitting process for the known errors in the licensees analyses and design basis.

Background

Page 2 of the Directors decision states that a backfit is not justified and explains, in part, the following:

Although our independent analysis using the analysis of record and incorporating information learned after FitzPatricks review resulted in exceeding the regulatory limit, the staffs conclusion regarding the FitzPatricks review was based on the best information available at the time. Therefore, a backfit is not justified given that this new information does not indicate that the licensees or staffs methods were clearly in error [emphasis added] and any further effort from the licensee would not provide an increase in safety as compared to the status quo based on significant conservatisms in the licensees analysis. In addition, the licensees analysis of record also demonstrates control room doses will remain below 5 rem. Although there were non-conservatisms in the licensees analysis, the NRC staff concluded in the SE that the significant conservatisms contained in the analysis and mentioned above provide reasonable assurance that the control room dose would remain below 5 rem [emphasis added].

Discussion The above justification for not performing a backfit is discussed below, followed by the flaws that we see in each reason provided in the Directors decision:

1. the FitzPatrick review was based upon the best information available at the timegiven that this new information does not indicate that the licensees or staffs methods were clearly in error Perceived Flaws:

a. Use of Erroneous Methods The licensee and staff methods were erroneous. The DPO asserts and the DPO Panel report confirmed that several methods used by the licensee and the NRC staff methods were chosen in error. The Directors decision agreed with the identification of three of these errors involving the spray and steamline removal coefficients, referring to the selection of these methods as being a deficiency and nonconservative. 38 Additionally, 38 Page 4, paragraph 1, of the Directors decision states, in part, We did identify three parameters used as inputs to the licensees calculation that may not have been the most conservative (i.e., elemental removal coefficient and the particulate removal coefficient used for the drywell and the MSLs). Paragraph 1 also refers to these errors as parameters found [to be] non-conservative [emphasis added]. Page 7 of the attachment to the Directors decision refers to the failure to address the impact of sprays as a deficiency.

The fourth paragraph states, in part, Even though RG 1.183, Revision 0 is silent on the appropriate use of drywell sprays when determining deposition in the MSLs, the substantial conservatisms contained in the 17

there are more errors than these three errors. This DPO Appeal provides a sampling of some of the flaws in the Directors disagreement regarding these additional errors. So, to assert that methods used were not clearly in error is flawed.

b. Mischaracterizing Existing Information as new information and that the best available information was used for the review It is misleading and incorrect to call the information that correctly models the spray and main steamline removal new information. Section 6.5.2, Containment Spray as a Fission Product Cleanup System, of NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants: LWR Edition (SRP),

describes methods to model the spray removal, based upon experimental work done in the 1970s. The NRC issued the most recent version of SRP Section 6.5.2 in March 2007 (ADAMS Accession No. ML070190178) which was before issuance of the subject licensing action. 39 Even the licensees better, but still flawed, estimates of the steamline deposition contained in its response to the NRCs request for information demonstrates that the technology and science existed to model the steamline deposition issues. 40 It is misleading and incorrect to call the use of grossly nonconservative steamline removal coefficients as using the best available information at the time. Extensive information and technology existed at the time of review to model these coefficients. This issue is discussed extensively in DPO Section 2.6 starting on page 22 of Enclosure 1 to the DPO. The NRC staffs analysis incorrectly perceived facts and failed to recognize the flawed coefficients.

In addition, one of the DPO submitters has worked on these issues for over 24 years at the NRC. In his extensive experience, they are not new issues. This experience included helping to develop a standalone computer code with Dr. Dana Powers and MELCOR models to estimate the aerosol removal in the steamline when sprays are credited. This work was completed more than a decade ago.

At the time the NRC staff wrote the FitzPatrick safety evaluation, the staff incorrectly perceived facts and performed or failed to recognize the flawed analyses. These omission or mistakes of fact of the science and technology that existed at the time of review were not due to new or modified interpretations of what constitutes compliance.

Therefore, consistent with Handbook Section III.B.6 of Management Directive 8.4, Management of Backfitting, Forward Fitting, Issue Finality, and Information Requests, dated September 20, 2019 (ADAMS Accession No. ML18093B087), compliance backfitting could be used to address these issues.

guidance compensate for this deficiency [with respect to not modeling a reduction in the steam line removal to account for crediting drywell sprays]. Page 2, paragraph 1, of the Directors decision, states in part, Our assessment also identified that the licensees assumptions for one parameter (elemental removal coefficient) were outside of the boundaries established by the standard review plan (SRP).

39 Page 2, paragraph 1, of the Directors decision, states, in part, Our assessment also identified that the licensees assumptions for one parameter (elemental removal coefficient) were outside of the boundaries established by the standard review plan (SRP). This refers to SRP Section 6.5.2 (ADAMS Accession No. ML070190178).

40 Page 5 of 17, paragraph 3, of Attachment 1 to the memorandum dated March 30, 2020, providing FitzPatricks response to the NRC staffs request for additional information (RAI) ARCB-RAI-2 (ADAMS Accession No. ML20090E279), provides an assessment of the impact of drywell sprays on the settling velocity. The settling velocity was used to recalculate the aerosol removal rate (deposition) in the steamline.

18

2. any further effort from the licensee would not provide an increase in safety as compared to the status quo based on significant conservatisms in the licensees analysis.

Perceived Flaws:

a. Unsupported Statements This statement and reasoning are not clear and lack the justification necessary for an independent reviewer to come to the same conclusion. In accordance with Management Directive 8.4, Handbook Section III.B.7, such a determination should be made in consultation with other appropriate offices, including the Office of Enforcement or the Office of the General Counsel, and there only needs to be an attributable, not a significant, increase in safety for a compliance backfit. It is not clear whether this consultation was done or the basis for the statement, but an attributable increase in safety does exist, as described below.

• Increase in Safety There are increased safety and regulatory compliance benefits when the flawed, nonconservative, and deficient design-basis analysis is corrected and used to demonstrate compliance with the regulations. Without correcting these issues, future design changes that are based upon the design-basis LOCA analysis will incorporate these errors to make safety determinations. These errors are so large that they will definitely and negatively impact these safety determinations. In addition, these errors allowed a safety system to be taken out of technical specification and the leakage from the safety-related MSIV to be increased based upon a flawed analysis.

b. Assertion of Significant Conservatisms The discussion of Flaw 8 addresses the concerns with the assertions of significant conservatism in the licensees analyses.

3. the licensees analysis of record also demonstrates control room doses will remain below 5 rem.

Perceived Flaw: The licensees analysis of record contains multiple deficiencies and nonconservative assumptions, as asserted in the DPO and confirmed in the independent DPO Panel report. This appeal contains further information on these flaws.

When some of these errors are corrected, our analysis shows that the analysis of record is non conservative by 2 orders of magnitude (i.e., the control room dose is 466 rem TEDE when the regulations require control room doses to 5 rem TEDE or less). Even the nonconservative and flawed directed analysis informing the Directors decision and the licensees sensitivity analysis (provided in response to the staffs RAIs) shows that the regulatory limit for the control room is exceeded when using the analysis of record assumptions and attempting to correct for only 19

these two errors. 41 Given this information, the statement that the analysis of record demonstrates control room doses below 5 rem is not correct.

4. Although there were non-conservatisms in the licensees analysis, the NRC staff concluded in the SE that the significant conservatisms contained in the analysis and mentioned above provide reasonable assurance that the control room dose would remain below 5 rem.

Perceived Flaws: This DPO Appeal discusses the flaws in this argument in more detail with respect to Flaw 8. It is inappropriate to use the alleged conservatisms stated to offset errors in the licensees analyses.

5. Therefore, a backfit is not justified Perceived Flaw - Incomplete Backfit Assessment:

a. The Commission directed the staff in MD 8.4 to determine the following when considering a backfit:

• Are either of the adequate protection exceptions to performing a backfit analysis applicable?

• Is the compliance exception to performing a backfit analysis applicable?

• Does the proposed action constitute a cost-justified substantial increase in overall protection?

The Directors decision does appear to provide an indirect and cursory focus on determining whether the compliance exception to performing a backfit analysis applies.

However, the Directors decision does not provide a defensible and complete discussion of whether correcting the licensing basis is needed for adequate protection or constitutes a cost-justified substantial increase in overall protection. The Directors decision neither describes how the backfit screening steps were applied nor contains a documented evaluation (for use of the adequate protection or compliance exceptions) or a backfit analysis (for a cost-justified substantial increase in overall protection backfit) that supports the conclusion that a backfit is not justified.

b. Although NUREG-1409, Revision 1 (ML21006A431) is with the Commission for a vote, it reflects the staffs current understanding of how to interpret and implement the Commissions policy in MD 8.4 and the Solicitors memorandum about using the compliance exception. Appendix B of that NUREG has a checklist for determining if the compliance exception to performing a backfit analysis applies. The Directors decision does not appear to have correctly applied the NUREGs checklist. For example, the 41 On page 2 of Enclosure 1 to the Directors decision, paragraph 3 states, in part, We modeled the licensees analysis of record adjusting various parameters to incorporate information we learned after the FitzPatrick amendment was issued. Specifically, we used a value within the SRP range for the elemental removal coefficient, a reasonable value for the particulate removal coefficient in the drywell and more accurate values for the particulate removal coefficients in the MSLs. Even though the dose calculated in the control room in this independent analysis exceeded the regulatory limit (5.11 rem) [emphasis added] . As discussed on page 7 of Attachment 1 to the Directors decision, paragraph 1 states that the licensees base case that considered the impact of modeling the drywell spray impact on aerosol removal in the main steamlines resulted in a control room dose of 7.35 rem.

20

following steps from the compliance exception checklist are not addressed in the Directors decision: 42

• Step 3 of the compliance exception checklist has the staff determine whether the NRC consistently interpreted and applied the identified requirement. Staff review procedures are contained in SRP Section 15.0.1, issued July 2000 and an acceptable method for complying with 10 CFR 50.67 is contained in RG 1.183, issued July 2000. RG 1.183 refers to other SRPs such as SRP Section 6.5.2.

These documents existed at the time of the staffs review. For each of the errors identified in the DPO and in the Directors decision no assessment of these errors with respect to this step are present in the Directors response.

• Step 5 of the compliance exception checklist has the staff determine whether it has identified at least one error or omissioneither the NRCs own error, or the omission or error of the licensee, applicant, or a third party (e.g., a vendor or another government agency), through any of the following: incorrect perception or understanding of the facts; failure to recognize flawed analyses; or failure to draw direct inferences from those facts or analyses. Numerous errors exist in the licensees analyses as described in the DPO. The Directors decision agreed that three of these errors existed. Additional errors are discussed in the DPO and this DPO Appeal. For each of the errors, no assessment with respect to this backfitting step are present in the Directors response. The errors associated with not modeling the limiting release pathway and using erroneous removal coefficients in the nonlimiting pathway are examples of possible errors of omission and should be considered for backfit evaluation.

• Step 7 of the compliance exception checklist has the staff determine whether the existence of the error(s) or omission(s) were determined by standards and practices that were prevailing among professionals or experts in the relevant area at the time of the NRC determination that the NRC requirement or commitment was satisfied and a regulatory approval was issued.

• Step 9 of the compliance exception checklist has the staff determine whether the NRC would likely not have issued its approval had it known of the error(s) or omission(s). Once the errors are addressed, the doses would greatly exceed the 10 CFR 50.67 limit for the control room operators and would, consequently, not support adequate protection in that the regulation is not met. Addressing only Flaw 1 would have precluded approval of the amendment.

Based upon the above, the Directors decisions directed assessment should not be used to prejudge whether backfitting should be applied to the issues identified in the DPO. The backfit process should be used to determine whether backfitting is warranted. The backfit review 42 Our intent in this discussion is not provide a full backfit analysis, but to point out the deficiencies with respect to not pursuing a backfit determination based upon the information provided in the Directors decision report. Additional information and examples can be provided upon request.

Therefore, some steps listed dont include any examples and only cite a summary of the step.

21

should be performed using the Commissions policy in MD 8.4 and with input from the agencys Backfitting and Forward Fitting Community of Practice.

Flaw 7 The Directors decision asserts that the condenser was not credited and the staff used engineering judgment and risk and engineering insights to support the deterministic conclusion and balance any uncertainties and nonconservatisms rather than as a basis for the decision. These statements conflict with information contained in the NRC staffs safety evaluation, meetings held with one of the DPO submitters, versions of the interim staff guidance (ISG) that existed at the time the DPO was submitted, and extensive information discussed in Enclosure 1 to the DPO that provided supporting details for DPO Issue 1.

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision Although this flaw relates to all documented Issues in the DPO, it is highlighted in DPO Issue 1.

The DPO Panel reviewed all these issues in their report. On pages 2 and 5 of Enclosure 1 to the Directors report, the Director provided statements that conflict with various agency documents and information provided in meetings with a DPO submitter

Background

Page 2 of the Directors decision states, in part, the following:

Our independent assessment also found that the NRC staff based their regulatory finding of adequate protection on the licensees deterministic analysis rather than a separate analysis completed by the NRC staff. The NRC staff performed an assessment of the seismic ruggedness of the power conversion system (PCS) and the main condenser to achieve high confidence that these systems will remain available after a safe shutdown earthquake (SSE) for fission product dilution, holdup, and retention. The NRC staff used engineering judgement and risk insights to support the deterministic conclusion, and to balance any uncertainties or approved non-conservatisms. The NRC staff did not use these risk insights as the basis for their regulatory finding, nor was an assumption of holdup in the condenser credited by the NRC staff [emphasis added].

In disagreeing with DPO Panel Report Issue 2 (Issue 1 of the DPO), the Director states the following on page 5 of the Directors decision:

I disagree with the panel and this specific concern. The licensees analysis is not in error and it demonstrates that the regulatory limits are met based on regulatory guidance available at the time the amendment was reviewed and approved.

While there were some non-conservative parameters [emphasis added] in the licenses analysis, the NRC staff determined that these parameters did not significantly impact the results.

In addition, in the SE, the NRC staff did not credit the PCS or the main condenser to make the regulatory finding of adequate protection [emphasis added].

22

However, page 33, paragraph 3, of the NRC staffs safety evaluation (ADAMS Accession No. ML20140A070) states, in part, the following:

However, the staff also considers it reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage. The NRC staffs consideration of risk and engineering insights is discussed in Section 3.5 of this SE.

Several drafts of DRA-ISG-2021-XX, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms, up until the time the DPO was submitted, contained versions of the following language:

The staff, through this ISG, should acknowledge the presence of the PCS [power conversion system] and its ability to provide a large hold-up and retention volume for MSIV leakage to resolve differences between the licensees methods and assumptions and those deemed acceptable to the staff [emphasis added]. In doing so, the staff should recognize that there is a high probability that doses will be lower than those estimated using deterministic methods which include accepted assumptions but do not credit hold-up and retention of the MSIV leakage within the power conversion system. Acknowledgement of the presence of the PCS can be used by the staff as part of the information for its reasonable assurance finding [emphasis added]. This ISG does not change the acceptable methods used by the licensee to demonstrate conformance with 10 CFR 50.67 and is consistent with the Commission direction in the SRM to SECY-19-0036.

Discussion The quotations contained in the Background section above assert that the NRC staff used engineering judgment and risk and engineering insights to support the deterministic conclusion and balance any uncertainties and nonconservatisms. Additionally, they assert that the NRC staff did not use the risk insights as the basis for its regulatory finding, nor did the staff credit an assumption of holdup in the condenser.

However, these assertions conflict with information contained in the NRC staffs safety evaluation, meetings held with one of the DPO submitters, agency documents including versions of the ISG that existed at the time the DPO was submitted, and extensive information discussed in Enclosure 1 to the DPO (see the supporting details for Issue 1, starting on page 1).

Conflicting Information in the Safety Evaluation Page 33, paragraph 3, of the staffs safety evaluation states that it is reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays. At the time of the analysis, the failure to consider the impact of the drywell sprays was an error in the licensees analysis because it involved differences between the licensees methods and assumptions and those deemed acceptable to the staff. Offsetting these errors and non-conservatisms is very different than and conflicts with the statements made on page 2 of the Directors decision that the engineering judgement and risk insights support the deterministic conclusion and balance uncertainties and approved non-conservatisms.

23

The results of the analyses used to demonstrate compliance with 10 CFR 50.67 are quantitative dose values. The errors in the FitzPatrick analysis impact these dose values. To offset these errors changes the results (basis) used to make our regulatory finding and is a way of crediting the power conversion system. So, to offset these errors conflicts with statements made in the Directors decision that the NRC staff did not use these risk insights as the basis of their regulatory finding.

Conflicting Information Provided in Meetings During meetings on the draft ISG with one of the DPO submitters, the submitter was informed that the ISG would be based upon the review method used for the FitzPatrick review. He was told that the ISG would provide a mechanism to resolve licensee issues related to compliance with methods and assumptions in its license amendments (including compliance with regulatory positions in RG 1.183). It was thought that this would benefit the NRC and the licensee by decreasing the review time and that crediting the condenser, since it had such a large volume, would compensate for issues or errors in the licensing analysis. When one DPO submitter told other meeting participants that this was not necessarily true, he was informed that the condenser would surely make up for errors of at least a factor of 2.

During an Office of Nuclear Regulatory Regulation All Hands Meeting on December 7, 2020, a slide summarizing the staffs review of four MSIV LARs was provided. One of the reviews summarized is the NRCs review for FitzPatricks Amendment 338. The summary slide stated that: Conclusions of reasonable assurance were based on a combination of deterministic methodology and risk insights.

Conflicting Information Provided in Agency Documents The technical rationale that doses will be lower than those estimated using the traditional deterministic methods (in RG 1.183) when the PCS is credited and that this information could be used by the staff to make a finding of reasonable assurance is documented in an internal letter dated December 21, 2020,

. 43 That same technical rationale was documented in multiple versions of the ISG by the same authors over a period of over 2 months. At the time the DPO was submitted, a version of the ISG was undergoing final revision before being submitted for internal concurrence. 44 The version cited in the Background section above stated that the staff could use the presence of the PCS as part of the information for its reasonable assurance finding as well as to resolve differences between the licensees methods and assumptions and those deemed acceptable to the staff. In other words, the presence of the PCS could be used to offset errors in the licensees analysis. This information conflicts with the statements in the Directors decision that the NRC staff did not credit the PCS or the main condenser to make a finding of adequate protection.

43 An internal memorandum dated, December 21, 2020 (ADAMS Accession No. ML20314A107) provides details on the development of the ISG. The memorandum states: Specifically, the high probability that doses will be lower than those estimated strictly using traditional deterministic methods, which include accepted assumptions that do not credit hold-up and retention of the MSIV leakage within the PCS, can be used by the staff as part of the information for its reasonable assurance finding.

44 An e-mail on February 24, 2021 provided the current version of the ISG cited. An excerpt from this ISG version is provided in the Background section for Flaw 8. The e-mail provides stated that the version of the ISG in the e-mail was being cleaned and was to be put into concurrence. The e-mail is not in ADAMS and is provided separately with our appeal [not publicly available].

24

The current draft version of the ISG also states, in part, the following:

The NRC staff concludes with reasonable assurance, based [emphasis added] in part on the risk and engineering insights to compensate for uncertainties in the evaluation of the dose consequences from the MSIV release pathway, that the licensees dose estimates will comply with the acceptance criteria.

The ISG reflects the methods used by the NRC staff to approve the FitzPatrick amendment. The above statements conflict with the statements in the Directors Directors decision that the NRC staff did not use these risk insights as the basis of their regulatory finding.

Note that after the DPO was submitted and during the DPO Panels interviews, the language regarding the use of the power conversion system to resolve differences between the licensees methods and assumptions and those deemed acceptable to the staff was removed from the ISG. However, even the revised general guidance in the current version of the ISG would still allow for the staff to offset errors in the licensees analyses using the ISG.

Conflicting Information Provided in Enclosure 1 to the DPO The descriptions provided for Issue 1 of our DPO provide extensive information regarding the use of alleged risk and engineering insights. In Section 4 for Issue 1 on page 8 of Enclosure 1 to the DPO we discuss the insights used in the NRC staffs safety evaluation that are not in the FitzPatrick analysis and provide information on the flaws in the logic that assert that these insights were not used by the staff. To highlight some of those flaws, it makes no sense that the licensees fundamentally flawed analysis could be used to demonstrate compliance with the regulation without some other considerations such as these insights. It is also not logical that the staff charged the licensee for extremely large amounts of staff hours to develop insights that were not necessary to approve the amendment.

Flaw 8 The Directors decision alleges that conservatisms in the licensees model provide reasonable assurance that the regulations are met. These statements are based upon calculations that do not reflect the licensees design and licensing bases. Reasonable assurance of compliance with 10 CFR 50.67 is based upon the licensees licensing and design bases, not an analysis containing assumptions that the NRC deems appropriate. The Directors decision includes a contrived rationale that conflicts with the regulatory requirements in 10 CFR 50.67; staff review procedures in SRP Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Revision 0, issued July 2000; and other statements made in the Directors decision.

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision This flaw relates to Issues 1, 2, 4 and 5 documented in the DPO with respect to how the analysis of record LOCA does not comply with the regulations. The DPO Panel reviewed this issue regarding compliance with 10 CFR 50.67 in various parts of the DPO Panel report. On page 9 of the Directors decision, the Director agreed with a related statement made on page 47 of the Panel report. However, the technical rationale in the Directors decision used to show compliance with 10 CFR 50.67 is based upon assumptions that differ from those used in the licensees docketed analyses.

25

=

Background===

RG 1.183 provides a method acceptable to the NRC staff for demonstrating compliance with 10 CFR 50.67. RG 1.183 was used by the licensee in the FitzPatrick amendment request and by the NRC staff to perform its review.

RG 1.183 provides a very systematic and purposefully conservative method that is well vetted by others (the NRC staff, industry, the ACRS) and gives a specific quantitative result. The NRC staff selected these assumptions and models to provide an appropriate and prudent safety margin against unpredicted events in the course of an accident and to compensate for large uncertainties in facility parameters, accident progression, radioactive material transport and atmospheric dispersion. RG 1.183 states that licensees should exercise caution in proposing deviations because these analyses were never intended to represent any specific accident sequence. Yet, the Directors decision made the same mistakes as the NRC staffs analysis supporting the review of the AST amendment. The Director used NRC presumed conservatisms that were not in the docketed design and licensing basis to assert that there is reasonable assurance of compliance with the 10 CFR 50.67.

The absolute values of these analyses are relied upon to approve the amendment and for future evaluations under 10 CFR 50.59, Changes, tests and experiments. The assumptions, inputs, and methods used to create these results become part of the design bases and are used in future design changes and 10 CFR 50.59 evaluations. The integrity of the assumptions, inputs, methods, and results are paramount to the regulatory process.

Discussion The Directors decision states that RG 1.183 and the licensees analyses contain substantial and significant conservatisms. 45 The perceived conservatisms are related to the assumptions the licensee used in its calculations for demonstrating compliance with 10 CFR 50.67.

Throughout, the Directors decision uses these alleged conservatisms to assert that reasonable assurance is met and to justify the reasoning for not backfitting the license to correct the deficiencies in the licensees analyses. This conclusion is based upon the assertion that changing the licensees analysis assumptions to use those the Director finds acceptable provides reasonable assurance of compliance with 10 CFR 50.67. This technical rationale is flawed.

Reasonable assurance of compliance with 10 CFR 50.67 is based upon the licensees licensing and design bases, not an analysis containing assumptions that the NRC deems appropriate.

The Directors decision includes a contrived rationale that conflicts with the regulatory requirements in 10 CFR 50.67, the staff review procedure in SRP Section 15.0.1, and other statements made in the Directors decision.

The requirements in 10 CFR 50.67 call for licensees to revise their source terms to evaluate the consequences of applicable design-basis accidents previously analyzed in their safety analysis reports. The licensees must demonstrate with reasonable assurance that certain doses in the control room and off site are not exceeded. SRP Section 15.0.1 contains a staff review 45 For example, page 6 of the Directors decision states that substantial [emphasis added] conservatisms contained in the guidance compensate for this deficiency. Page 5 states, Although there were non-conservatisms in the licensees analysis, the NRC staff concluded in the SE that the significant [emphasis added] conservatisms contained in the analysis and mentioned above provide reasonable assurance that the control room dose would remain below 5 rem.

26

procedure and reiterates the importance of using the licensees docketed information.

Specifically,Section III.6.c states the following:

The staffs approval of the application is to be based on the licensees docketed information. If differences are discovered between the licensees methods and assumptions and those deemed acceptable to the staff, the reviewer should resolve the differences with the licensee. If necessary, the licensee should update the disputed assumptions and resubmit the affected analyses.

However, when the Directors decision directed assessments modified the assumptions used in the licensees analysis to consider these alleged conservatism, the Director used assumptions that are different than those in the licensing basis. This conflicts directly with 10 CFR 50.67 and SRP Section 15.0.1.

In addition, the technical rationale in the Directors decision with respect to these alleged conservatisms conflicts with the Directors agreement with the DPO Panel report on a related issue about fundamental regulatory concepts. On page 9 of the Directors decision, the Director agreed with a statement made on page 47 of the Panel report. 46 The Director agreed that the licensee has the primary responsibility for safety, that staff analyses and insights cannot be substituted for the licensees analyses, and that the staff reviewers should not base the acceptance of the licensing action primarily on their own perspectives and views. However, the technical rationale in the Directors decision used to show compliance with 10 CFR 50.67 is based upon assumptions that differ from those used in the licensees docketed analyses. The assumptions made by the licensee that are deemed conservative were changed to lower the control room doses to less than those required in the regulation.

Not complying with these fundamental regulatory concepts can have a long lasting and detrimental effect on safety. It creates differences between the design and licensing basis used by the licensee to comply with the regulations and what the NRC staff relies on to verify compliance with the regulations. The impact of these differences is similar to Issue 4 of the DPO. These differences can have a negative impact on safety future changes at the plant using 10 CFR 50.59, creates challenges to inspectors in performing inspections and developing regulatory findings, and problematic for licensing actions that rely on actions approved in error.

It is important to point out that, even if it were appropriate to use these revised assumptions, the cited conservatisms should not be characterized as conservatisms for many reasons. For example, based upon our experience, it could be demonstrated that the alleged conservatisms would not be appropriate to use, have little impact on the results and/or conflict with other assumptions or the regulatory guidance used. While to go through this list of alleged conservatisms and provide proof would take too much time and effort to be included in the timeframe of this appeal, we will provide a few high-level examples.

46 Page 10 of the Directors decision, states, in part, I agree with the regulatory philosophies above. These policies include the following: The licensee has the primary responsibility for safety and must provide an application that demonstrates the regulations are met (especially when a regulation is as clear as 10 CFR 50.67 regarding licensee analysis). Therefore, staff reviewers must review the licensees submitted information and draw conclusions based on this information. While the staff can use confirmatory analyses and insights (including risk perspectives) to aid in their review, these staff analyses and insights cannot be substituted for the licensees analysis, especially when licensee information indicates unacceptable results.

Staff reviewers should not base the acceptance of a licensing action based primarily on their own perspectives and views.

27

In Directors report it states: A few of the many conservatism include... (3) not considering suppression pool scrubbing and retention. It also states the following: In fact, we performed a RADTRAD run using the same inputs mentioned above but crediting suppression pool fission product scrubbing and retention and the control room dose value resulted in 3.03 rem. and Although credit for suppression pool fission product scrubbing and retention should be approved on a case-by-case basis and with appropriate justification, it is a physical phenomenon that would be expected to happen during the postulated scenario.

We reviewed the analysis crediting suppression pool scrubbing and determined the following issues with the use of this alleged conservatism:

First and foremost, this assumption is not in the design basis for the licensees analyses and therefore, should not be used to show compliance with 10 CFR 50.67.

RG 1.183, Regulatory Position 3.5, which the licensee used to demonstrate compliance with 10 CFR 50.67 states the following:

Reduction in airborne radioactivity in the containment by suppression pool scrubbing in BWRs should generally not be credited [emphasis added]. However, the staff may consider such reduction on an individual case basis. The evaluation should consider the relative timing of the blowdown and the fission product release from the fuel, the force driving the release through the pool, and the potential for any bypass of the suppression pool (Ref. 7). Analyses should consider iodine re-evolution if the suppression pool liquid pH is not maintained greater than 7.

To our knowledge no licensee has ever been approved for suppression pool scrubbing under 10 CFR 50.67. The relative timing of the blowdown compared to the release of the source term make it hard to justify and that is why the regulatory position does not provide a regulatory position with automatic credit for some suppression pool scrubbing. No adequate justification is provided in the Directors decision for asserting this is a conservatism. 47 The analysis supporting the Directors decision incorrectly treats suppression pool scrubbing, sprays removal and deposition in the steamline as independent processes when they are not. In addition, continued removal of aerosols due to multiple removal processes is unlikely. At some point the aerosol will become very difficult to remove once a self-sustaining aerosol size is reached. Therefore, it is not expected to provide much additional mitigation when sprays are credited, and likely would not overcome the factor of 100 error in doses discussed above.

Arguments made by the authors of the ISG maintain that the State-of-the-Art Reactor Consequence Analyses (SOARCA) project analyses prove that suppression pool scrubbing can be credited for 10 CFR 50.67 analyses. The SOARCA analyses are very different than those used for 10 CFR 50.67 compliance (design-basis accident analyses). The SOARCA analyses are best available and more realistic analyses, whereas the design-basis accident analyses are intentionally conservative and bounding analyses. These differences must be considered when applying SOARCA insights to design-basis accident analyses.

In conclusion, based upon the above information, the use of alleged conservatism in the Directors decision to offset deficiencies in the licensees analyses and to allege reasonable 47 The statements in the Directors decision that suppression pool scrubbing is a physical phenomenon that is expected to happen in the postulated scenario is not enough to justify its use.

28

assurance of meeting the requirements in 10 CFR 50.67 is flawed. Issue 4 of the DPO discusses the implications of the erroneous use of these conservativisms on future compliance and regulatory actions.

Flaw 9 The Directors decision contains numerous additional flaws, inaccurate statements, and incorrect technical rationale in its statements and logic, as described in the enclosure to the DPO-2021-001 appeal.

Relationship of the Flaw to the DPO, DPO Panel Report, and Directors Decision The descriptions below of the flaws in the Directors decision discuss the relationship of these flaws to the DPO.

Discussion Flaw 9.1 This flaw relates to DPO Issue 4. On page 7 of the Directors decision, the Director disagreed with the DPO Panel and DPO submitters on Issue 4 of the DPO. In DPO Issue 4, the submitters stated that errors in the now approved design basis could adversely impact future plant licensing actions and inspections. The reasoning in the Directors decision used to disagree with Issue 4 narrowly focuses only on the issue of containment spray removal coefficients. The Directors response with respect to Issue 4 did not discuss or consider other erroneous models and assumptions now in the FitzPatrick design basis. These include a major issue identified in Section 2.6 of the DPO about the approval of a technically indefensible and grossly nonconservative steamline aerosol deposition model (inconsistent with known science on aerosol deposition) and those assumptions identified in the DPO that are not consistent with RG 1.183. These issues are discussed above with respect to the other flaws described in this appeal and in the DPO. Therefore, the reasoning in the Directors decision is deficient and flawed because it did not consider the adverse impact of all these errors, now in the FitzPatrick design and licensing basis, on future licensing actions and the potential use of these methods by other licensees under 10 CFR 50.59 and/or amendments.

Flaw 9.2 This flaw relates to DPO Issue 2, Section 2.6 (see page 26 of the DPO). Page 6 of the attachment to the Directors decision states the following:

The NRC staff approved FitzPatricks use of the 20-group method based on best estimates available at the time and relying on previously approved LARs [license amendment requests] with similar approaches (i.e., Clinton, Limerick and LaSalle.) After issuing FitzPatricks amendment, the NRC staff learned that the 20-group method is not conservative for modeling the changing aerosol settling velocity distribution as it moves through the MSL volumes.

The idea that the 20-group method used in the analysis of record represents the best estimates at the time and that these estimates relied upon in previously approved LARs with similar approaches is misleading and flawed. In RAI ARCB-RAI-2, the NRC staff acknowledges that the 20-group method was used at Clinton Power Station, Limerick Generating Station, and LaSalle County Station, but these precedents are not similar to FitzPatrick because those 29

facilities did not credit drywell sprays. 48 The licensee also acknowledges issues with the steamline deposition credited in the analysis of record. 49 Despite acknowledging these issues, neither the licensee nor the NRC corrected the steamline removal coefficients in the analysis of record used to approve the license amendment.

It is worth noting that the science involved in modeling this aerosol deposition is not new. The NRC and the nuclear industry have known for more than a decade that the sprays impact the steamline deposition factors. 50 The basic science to model the impact of sprays on steamline deposition has existed for several decades. Lastly, the licensees attempt at modeling the steamline deposition in its sensitivity analysis (in response to NRC staff RAIs) shows that the methodologies and science existed at the time to model the impact of sprays on steamline deposition. Therefore, the statements that what the NRC approved represents best estimate models at the time are misleading and incorrect.

Flaw 9.3 This flaw relates to DPO Issue 2, Section 2.6 (see page 26 of the DPO). Page 7 of the attachment to the Directors decision states the following:

It is worth noting that at the time the FitzPatrick LAR was being reviewed, and in the present time for that matter, the NRC did not have a specific reduction criteria or any approved guidance licensees could use to address the impact of drywell sprays [emphasis added] on deposition in the MSLs.

The idea that specific reduction criteria did not exist or that approved guidance is needed to approve the FitzPatrick amendment is incorrect. Reduction criteria for the impact of sprays have been used and approved in license amendments. 51 As discussed above with respect to Flaw 9.2, the theory to model the impact of sprays does not represent new science. The licensee even created a method and described it in its response to RAI ARCB-RAI-2. Methods and theory for modeling the impact of sprays have existed for a very long time.

48 The regulatory basis for RAI ARCB-RAI-1C states, the NRC staff notes that the analyses cited as precedents did not credit drywell sprays.

49 The licensees response to RAI ARCB-RAI-2 (ADAMS Accession No. ML20090E279) states, in part, A sensitivity analysis was performed to evaluate the impact of sprays on the aerosol settling velocity

[emphasis added] and to identify other inputs with well-defined uncertainty or conservatism that could be used to offset the uncertainty associated with the current aerosol deposition model. This sensitivity analysis concludes that conservatisms are sufficient to offset the uncertainty introduced by the drywell spray effects on the aerosol deposition model [emphasis added].

50 Page 28 of the FitzPatrick safety evaluation contained in Enclosure 2 to the license amendment dated July 21, 2020 (ADAMS Accession No. ML20140A070), discusses the impact of the sprays on the aerosols and references a 1993 NUREG. Computational models have existed for several decades in codes such as the NRCs MELCOR code and the MAEROS code.

51 Page 17 of Enclosure 2 to the NRC staffs safety evaluation for Nine Mile Point Nuclear Station, Unit 2, dated May 29, 2008 (ADAMS Accession No. ML081230439), states, However, for additional conservatism, and to address concerns historically documented by the NRC staff, the licensee used the 3rd percentile [sic; one-half the third percentile] settling velocity of 0.000066 m/sec. The NRC staff agrees that this 3rd percentile settling velocity value is sufficiently conservative to reflect the effectiveness of drywell spray activity removal in containment upstream of this pathway. Though, as discussed earlier, the LOCA activity leak rates are reduced by a factor of 2 after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, based on decreasing containment pressure, the licensee conservatively does not credit this reduction to increase removal efficiency by natural deposition in the bypass lines.

30

Lastly, the idea that approved guidance is needed before a licensee can submit and the NRC can approve a license amendment is misleading, incorrect, and conflicts with the following statements on page 6 of the attachment to Directors decision:

It is important to note that regulatory guides are not regulatory requirements and provide an acceptable way to meet NRC regulations. Licensees could still propose alternative methodologies with proper justifications.

31

Document 6: Statement of Views on DPO Appeal Submittal

UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 February 3, 2022 MEMORANDUM TO: Daniel H. Dorman Executive Director for Operations Signed by Veil, Andrea FROM: Andrea D. Veil, Director on 02/03/22 Office of Nuclear Reactor Regulation

SUBJECT:

STATEMENT OF VIEWS REGARDING APPEAL OF DIFFERING PROFESSIONAL OPINION CONCERNING DPO-2021-001 The purpose of this memorandum is to provide my statement of views on the appeal of differing professional opinion (DPO)-2021-001.

On January 19, 2021, a Senior Reactor Engineer from the Office of Nuclear Reactor Regulation (NRR) submitted a DPO concerning the U.S. Nuclear Regulatory Commission (NRC) staffs issuance of Amendment No. 338 for James A. FitzPatrick Nuclear Power Plant related to their alternative source term for calculating loss of coolant accident dose consequences (Agencywide Documents Access and Management System (ADAMS) Accession No. ML20140A070). On February 12, 2021, a DPO Ad Hoc Review Panel was established and tasked to meet with the submitters, review the DPO submittal, and issue a DPO panel report, including conclusions and recommendations, regarding the disposition of the issues presented by the submitters in the DPO. On June 9, 2021, after reviewing the applicable documents, performing internal interviews of relevant individuals, and completing their deliberations, the panel issued their report to me, the NRR Office Director (ADAMS Accession No. ML21160A229). On November 17, 2021, I issued the Directors Decision memorandum to the DPO submitters documenting my assessment and decision regarding the DPO (ADAMS Accession No. ML21236A254). On January 7, 2022, the DPO submitters sent an appeal to you, the Executive Director for Operations, and expressed views on the Directors Decision memorandum.

CONTACT: Caroline Tilton, NRR 301-415-0990

In the basis for the appeal, the submitters stated that I disagreed with most of the findings and recommendations made by the differing professional opinion (DPO) panel. I would like to clarify this statement. In the DPO, the submitters document a total of 11 issues: seven under Issue 1 and four in Issues 2 through 5. In my decision, I agreed with the panel on five out of the 11 issues. On one of the concerns that I didnt agree with, the panel was neutral. In my decision for this concern, I offered clarifying information based on interactions with technical experts on the subject. During my review of the concerns, we had multiple conversations with FitzPatricks resident inspector to understand the plant configuration and how different components, referred to in the DPO, connected to the credible leakage paths. Subsequently, we shared this information and my preliminary findings with the panel. As a result of this discussion and when presented with drawings, pictures and supporting data, the panel indicated that their findings were based on the information they had gathered acknowledging their lack of access to an independent dose analysis expert and detailed plant-specific information. It is important to note that my decision was aligned with the panel on the overarching and most important issue: that backfitting the licensee to require a revision of related calculations provides no safety enhancement and that these design basis scenarios are bounding and contain inherent conservatisms. In summary, I agreed with several of the DPO panels conclusions, provided clarifying information during the review, and agreed with the panel on its conclusion that there is no safety issue.

As part of the recommendations included in the appeal, the submitters referenced their preliminary assessment of the panels recommendations provided via e-mail on June 17, 2021 and transmitted the assessment with the appeal. Noting that this information is not part of the original DPO, the submitters make several recommendations which essentially echo the panels Recommendations 5 and 6. My assessment of these recommendations are included on page 10 of my decision.

In the appeal, the submitters identified nine points of disagreement, referred to as flaws, with my decision. My views on these nine items are as follows:

1. In Item 1, the submitters stated that the assessment supporting my decision did not consider a limiting credible accident scenario that bypasses the pathway to the environment and was previously mitigated by the design of the main steam leakage control system. This item is equivalent to Issue 1(b) in the DPO. The submitters did not provide any new information in the appeal in support of this issue. You can find my assessment of this issue on page 1 of the attachment to my decision under Issue 1, Concern (b).

2. In Item 2, the submitters stated that the assessment supporting my decision is based upon a model that credits deposition and holdup in the main steamline upstream of the inboard closed main steam isolation valve (MSIV), and that this modeling is based on an incorrect statement in my decision that a section of the Regulatory Guide applicable to Alternate Source Term modeling does not apply to modeling of MSIV leakage. This item is equivalent to Issue 1(a) in the DPO. The submitters did not provide any new information in the appeal in support of this issue. You can find my assessment of this issue on page 1 of the attachment to my decision under Issue 1, Concern (a).

3. In Item 3, the submitters stated that the technical rationale used in my decision to assess the atmospheric dispersion factors is inconsistent with the regulatory guidance used by the licensee to demonstrate compliance with the regulatory requirements. This item is equivalent to Issue 1(c) in the DPO. The submitters did not provide any new information

in the appeal in support of this issue. You can find my assessment of this issue on page 2 of the attachment to my decision under Issue 1, Concern (c).

4. In Item 4, the submitters stated that my decision was heavily based on the analysis performed by an impartial independent reviewer. The member of my staff who performed the independent analysis was involved in the review of the subject FitzPatrick safety evaluation (SE) but was not involved in the dose analysis review and was, therefore, independent.

5. In Item 5, the submitters stated that the model used to support my decision does not consider the impact of sprays on steamline deposition and assumes a nonconservative aerosol size distribution. In support of this item (page 15 of the enclosure to the appeal),

the submitters stated that DG-13891 (the recent draft revision of Regulatory Guide (RG) 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors) refers to the State-of-the-Art Report on Nuclear Aerosols 2 (SOAR) as recommending a range of sizes for the reactor coolant system and in-containment aerosols. Specifically, the SOAR recommends a log-normal distribution for aerosols with an aerodynamic mass median diameter (AMMD) of 1.0 m and a geometric standard deviation of 2.0 in the reactor coolant system and an AAMD of 3.0 m with a geometric standard deviation of 2.0 for in-containment aerosols. DG-1389, however, recommends a AMMD value of 2.0 m based on engineering judgement. DG-1389 has not been approved and, because it is regulatory guidance, it is not a regulatory requirement. Licensees can determine an appropriate value for their plant-specific scenario when requesting a license amendment if supported by an appropriate justification. In addition, using the SOAR recommendation is one of the acceptable methods described in DG-1389 that licensees can consider in developing an aerosol size distribution model acceptable to the NRC. Licensees are also able to propose other methodologies in their submittals. Our independent assessment used in-containment values. This approach is supported by the regulations pertaining to radiological consequence analyses and established regulatory practices. You can find additional details of my assessment, as it relates to this issue, on page 6 of the attachment to my decision under Issue 1, Concern (f), Particulate Removal Coefficients in the MSLs.

6. In Item 6, the submitters stated that my decision bypassed backfitting processes and procedures. My assessment is based on the policy described in Management Directive (MD) 8.4,3 Management of Facility-Specific Backfitting and Information Collection in consultation with NRR staff that are experts in the subject. A backfit using the adequate protection exception is not justified because the facility is meeting existing requirements that provide adequate protection of public health and safety. A backfit using the compliance exception is not justified because, considering information we had at the time we issued the FitzPatrick amendment, our assessment concluded that the NRC staff did not make an error or omission, and our SE clearly documented our understanding and approval of the licensees submittal, including the identified non-1 DG-1389, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors (ADAMS Accession No. ML21292A319).

2 NEA/CSNI/R(2009)5, State-of-the-Art Report on Nuclear Aerosols, Nuclear Energy Agency, Committee on the Safety of Nuclear Installations, dated December 17, 2009.

3 MD 8.4, Management of Backfitting, Forward Fitting, Issue Finality, and Information Requests, dated September 20, 2019 (ADAMS Accession No. ML18093B087).

conservatisms. In addition, the licensee employed a methodology endorsed by the NRC with previously approved deviations that addressed a long-established NRC staff concern documented in RIS 2006-004,4 Experience with Implementation of Alternative Source Terms. The submitters also stated that, at the time the amendment was issued, an acceptable methodology existed to model aerosol deposition which the licensee should have employed. The existing methodology had not been endorsed by the NRC when the license amendment was issued, nor had the NRC or any licensees used that methodology in any licensing actions or requests. For this reason, the licensee was not required to adopt this methodology. Therefore, while the points of view surrounding the appropriate methodology have changed, the methodology used by the licensee was not in error at the time of the licensing action. Most importantly, our independent assessment confirmed that, considering new information learned after FitzPatricks amendment was issued and taking into considerations the non-conservatisms in the licensees analysis, the regulatory requirements continue to be met. A cost justified substantial increase in overall protection backfit is not justified because any further effort from the licensee would not provide a substantial increase in overall protection as compared to the status quo based on significant conservatisms in the licensees analysis.

7. In Item 7, the submitters stated that the assessment documented in my decision under Issue 2, conflicts with statements in the subject SE and the version of the draft Interim Staff Guidance (ISG) DRA-ISG-2021-XX5, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms that existed at the time the DPO was submitted. It is important to note that the interim staff guidance referenced in support of this item was, and continues to be, in draft form. Under Issue 2 (page 4) of my decision, I explained that the NRC staff used engineering judgement and risk insights to support the deterministic conclusion, and to balance any uncertainties or approved non-conservatisms. The NRC staff did not use these risk insights as the basis for their regulatory finding. In the amendment, the NRC staff acknowledged the non-conservatisms identified in the licensees analysis and determined these were not significant enough to call into question the conclusion of reasonable assurance.

Consistent with the statements of consideration for 10 CFR 50.67, the NRC staff leveraged these risk insights, in a manner that complements the deterministic approach and supports the traditional defense-in-depth philosophy. This approach also follows the NRCs policy statement on the use of probability methods. These risk insights are in accordance with the Commissions direction in SRM-SECY-19-0036,6 which states: [i]n any licensing review or other regulatory decision, the staff should apply risk-informed principles when strict, prescriptive application of deterministic criteria such as the single failure criterion is unnecessary to provide for reasonable assurance of adequate protection of public health and safety.

8. In Item 8, the submitters stated that my decision was based on calculations that do not reflect the licensees licensing basis. My decision was based on a review of the staffs 4 RIS 2006-04, Experience with Implementation of Alternative Source Terms, dated March 7, 2006.

Available at https://www.nrc.gov/reading-rm/doc-collections/gen-comm/reg-issues/2006/ri200604.pdf.

5 ISG DRA-ISG-2021-XX, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms (ADAMS Accession No. ML21078A051).

6 SRM to SECY-19-0036, Application of the Single Failure Criterion to NuScale Power LLCs Inadvertent Actuation Block Valves, dated July 2, 2019 (ADAMS Accession No. ML19183A408).

analysis and their conclusion of reasonable assurance of adequate protection, which were based on the licensees docketed submission. While I did utilize risk insights in my decision that were not included in the licensees analysis, these insights were not necessary to make a reasonable assurance finding. Rather, they were used to decide if a backfit was justified. These risk insights were based on well-known and widely understood nuclear power systems. Therefore, it was not necessary that the presence of these systems be documented in the licensees licensing basis. The purpose of my review was not to reapprove compliance with 10 CFR 50.67 and recreate the staffs prior reasonable assurance finding. As explained in my decision, the submitters believe that the licensee should assume the most conservative scenario, including scenarios beyond what the licensee and the NRC have determined to be conservative and appropriate, even if it conflicts with the design basis scenario.

9. In Item 9, the submitters stated that my decision contains numerous flaws, inaccurate statements, and incorrect technical rationale. In the enclosure to the appeal, the submitters also stated that the combined impact of these nine items result in doses 100 times greater than those calculated in the licensees analysis of record. Refer to page 7 of my decision, under Issue 5, for additional details of my assessment of this issue.

In the appeal, the submitters also identified three additional issues from my decision and the panel report and associated requests. My views on these issues are as follows:

1. Under Issue 1 of the appeal, the submitters requested that the EDO delay the final approval of DRA-ISG-2021-XX and DG-1389 until the completion of the DPO review and any associated tasking. I continue to have no objections with the guidance contained in these documents. It is important for the review schedule of DG-1389 to remain on track to support future high burnup fuel licensing actions. We have established a target date of January 2023 for issuing this revision to RG 1.183 to support licensees assessments, license amendment requests submittals and approvals and development of 10 CFR 50.59 packages (as appropriate). We have established a target date of September 2022 for finalizing the ISG.

2. Under Issue 2, the submitters stated that the numerous flaws in my decision could be referenced by licensees in future license amendment requests under 10 CFR 50.67, in addition to changes under 10 CFR 50.59. The submitters also requested that the EDO direct these issues to be corrected to reconcile the errors before they become propagated. The concerns raised for this issue are equivalent to Issue 4 in the DPO.

The submitters did not provide any new information in the appeal to support this issue.

You can find my assessment of this issue on page 7 of my decision, under Issue 4.

3. Under Issue 3, the submitters stated that the assessment associated with my decision replicated errors in the licensees analysis of record. The submitters also stated that these errors represent safety-significant issues. The submitters requested that, if the EDO determines that my decision contains the flaws described in the appeal, the NRC staff should be directed to consider backfit procedures or Orders for the license to fix these safety significant issues. This issue raises similar concerns as those in Issue 1(b) and Issue 5 of the DPO. You can find my assessment of these issues on page 1 of the attachment to my decision and page 7 of my decision, respectively.

Accordingly, I conclude that the submitters have not raised any issues that impact the conclusions of my assessment. For this reason, my decision regarding DPO 2021-001 remains

unchanged.

Document 7: DPO Appeal Decision UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 August 25, 2023 MEMORANDUM TO: Mark Blumberg, Senior Reactor Engineer Radiation Protection and Consequence Branch Division of Risk Assessment Office of Nuclear Reactor Regulation Michael Markley, Branch Chief Plant Licensing Branch II-1 Division of Operating Reactor Licensing Office of Nuclear Reactor Regulation Signed by Dorman, Dan FROM: Daniel H. Dorman on 08/25/23 Executive Director for Operations

SUBJECT:

DIFFERING PROFESSIONAL OPINION APPEAL DECISION INVOLVING DPO-2021-001 The purpose of this memorandum is to inform you of my review and conclusions regarding the Differing Professional Opinion (DPO) appeal (DPO Appeal) you submitted on January 7, 2022 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML22039A062 case file). The DPO Appeal raised concerns about the U.S. Nuclear Regulatory Commissions (NRCs) issuance of James A. FitzPatrick nuclear power plant operating license amendment No. 338, regarding use of the Alternative Source Term (AST) for calculating loss-of-coolant accident (LOCA) dose consequences, and the November 17, 2019, NRR Office Directors decision (Directors Decision) on the initial DPO.

A summary of concerns raised in your DPO Appeal and my conclusions for each are listed below. In my review of your concerns, I took into consideration the DPO Appeal Panels analysis and recommendations as documented in its independent and thorough review report (ML23160A209). For the detailed DPO Appeal Panel evaluation of each concern, please refer to the enclosed report.

Issue 1: Determine whether the FitzPatrick AST license amendment is in compliance with 10 CFR 50.67, Accident Source Term, and is based on the applicants analysis as required by 10 CFR 50.67(b)(2), and whether the information discussed in Issue 2 below would result in control room doses higher than those allowed under 10 CFR 50.67.

CONTACT: Hector Rodriguez-Luccioni, OEDO 301-415-6004

M. Blumberg. et. al.

Conclusion 1: The staffs evaluation of the AST amendment did not demonstrate that the application complied with the criteria specified in 10 CFR 50.67(b)(2). Specifically, the applicants analysis did not demonstrate with reasonable assurance that calculated control room dose would not exceed 5 rem TEDE for the duration of the accident.dorm Issue 2: Determine if the LOCA assumptions and methods used by FitzPatrick1 and the NRC staff to evaluate the FitzPatrick AST license amendment request (LAR) for a revised LOCA analysis are accurate and consistent with the stated regulations and standard review plan guidance (or if deviations from guidance by the licensee were documented and included an adequate technical basis). This determination will take into account the technical issues that were described in the DPO Appeal, which are summarized below:

a. The licensees LAR, the NRC staffs safety evaluation, and the Directors Decision did not consider a limiting credible accident scenario that bypasses the assumed pathway to the environment (i.e., leakage from the main steam isolation valves (MSIVs) packing, body, or mechanical joints directly to the environment). (RG 1.183, Alternate Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Appendix A, Regulatory Position 6.4 and 6.5)

b. By crediting the air space in the main steam line up to the closed MSIVs, the licensees analysis did not assume the radioactive release was instantaneously and homogenously mixed throughout all the free air volume in the drywell containment.

Despite the licensees statements that the containment elemental and particulate (natural) deposition/plateout is not credited, this deposition is credited in the space upstream of the MSIVs that is part of the drywell containment. (RG 1.183, Appendix A, Regulatory Position 3.1)

c. By assuming a release point at a specific location inside the turbine building, rather than on the surface of the turbine building, improper dilution was credited in the turbine building. This dilution reduced the atmospheric dispersion factors, thereby underestimating the control room dose. (RG 1.183, Appendix A, Regulatory Position 6.4)

d. By crediting nonsafety-related structures, systems, and components such as vents and doors that are not controlled by technical specifications, the licensee omitted release pathways that would be closer to the control room and would result in higher control room doses. As such, the accident analysis no longer aligns with, or is bounded by the plant design basis. (RG 1.183, Regulatory Position 5.1.2 and RG 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants, Regulatory Position 2.0)

e. The analysis did not consider the limiting LOCA break location by selecting a recirculation line break, rather than a break in the reactor coolant system just prior to 1 In its AST license amendment request (ML19220A043), FitzPatrick indicated that its application conformed with the relevant guidance of RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Rev. 0, issued July 2000 (ML003716792), and was consistent with other guidance contained in NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants LWR [Light-Water Reactor] Edition, Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Rev. 0, issued July 2000 (ML003734190) and RG 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants, issued June 2003 (ML031530505).

M. Blumberg. et. al. the MSIVs. In addition, when the licensee modeled the break just before the MSIVs (provided as a sensitivity study in response to the staffs request for additional information), the licensee did not use the assumptions that were used in its LOCA analysis supporting its LAR. (RG 1.183, Appendix A)

f. By using removal coefficients for aerosol settling that are nonconservative and by incorrectly modeling the impact of changing aerosol particle size distribution as the radioactive material moves down the steam line, the LOCA model overestimates the deposition in the steam lines and, therefore, underestimates the dose in the control room. (RG 1.183, Appendix A, Regulatory Positions 6.3 and 6.5)

g. By using an elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis, the analysis resulted in elemental removal greater than the SRP Section 6.5.2, Containment Spray as a Fission Product Cleanup System, Revision 4, dated March 2007 (ML070190178) limit on removal. Although this deviation was discussed in the NRC staffs safety evaluation, the NRC staff did not assess the aggregate impacts of the deficiencies in the analysis described above. (RG 1.183, Appendix A, Regulatory Position 3.3, and SRP Section 6.5.2)

Conclusion 2:

a. The LAR review process did not adequately address potential leakage from the MSIV packing. Control of packing leakage from the outboard MSIV was a safety function described in the FitzPatrick licensing basis, but an assessment of the impact of the removal of the Main Steam Leakage Collection System (MSLCS) on this function was not provided by the licensee nor documented in the staffs safety evaluation. In addition, outboard MSIV packing leakage into the steam tunnel area represents a potentially unfiltered release path to the environment.

b. Crediting deposition upstream of the inboard MSIV is consistent with the guidance contained in RG 1.183 and is reasonable.

c. Assuming a release point within the turbine building is consistent with RG 1.183 and is reasonable.

d. The licensees approach for determining the release pathway from the turbine building to the control room and the associated treatment of nonsafety-related structures, systems, and components are reasonable.

e. The AST amendment did not consider the most limiting LOCA break location.

f. The removal coefficients for aerosol settling used in the AST LAR are nonconservative and the use of an aerosol settling velocity based on AEB-98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application using the Revised (NUREG-1465) Source Term (ML011230531), overestimates main steam line deposition.

g. The use of an elemental iodine removal constant greater than 20 hr-1 is not consistent with guidance documents and was not adequately justified in the staffs safety evaluation.

M. Blumberg. et. al. Issue 3: Evaluate potentially conflicting statements between the Directors Decision and the NRC staffs safety evaluation on crediting portions of the power conversion system condenser and determine if any mitigation function assigned to the condenser and associated equipment is appropriately documented in the licensing basis.

Conclusion 3: There is a lack of appropriate regulatory documentation and justification for crediting the power conversion system for aerosol deposition. Crediting aerosol deposition in the condenser, a capability not requested by the licensee, was necessary to address known non-conservatism in the licensees analysis supporting the AST LAR. Absent this credit, the staffs evaluation is not sufficient to demonstrate compliance with the control room dose limit.

Issue 4: Determine if the approval basis of the FitzPatrick AST amendment was clearly documented in the licensing basis (e.g., calculations of record, updated final safety analysis report updates, and safety evaluation report, and whether they appropriately reflect the assumptions and mitigating equipment credited in the NRC staffs analysis).

Conclusion 4: The rationale for approving the FitzPatrick AST LAR is not clearly documented in the licensing basis.

In sum, after careful consideration of your appeal, I concluded that even though there is no immediate safety concern, there are issues associated with the FitzPatrick AST license amendment that have significant relevance to the regulatory process that must be addressed.

As such, I am directing NRR to:

1. Take appropriate regulatory action to ensure FitzPatricks compliance with 10 CFR 50.67 and resolve the licensing basis clarity issues for the AST license amendment, including the impact of outboard MSIV packing leakage, the basis for the limiting break location, and the aerosol deposition credit for the main condenser.

2. Develop an implementation plan for the other recommendations included in the DPO Appeal Panel Analysis Report.

I want to thank you for bringing your concerns to my attention, and I appreciate you taking the time and effort to document and share them using our established DPO process. Your willingness to raise concerns through the DPO process is consistent with our organizational values of Openness and Commitment. Our agency relies on dedicated professionals, such as yourself, who are willing to raise concerns that could impact the NRC mission.

A more in-depth analysis of each of the concerns you raised is provided in the enclosure.

Enclosure:

DPO Appeal Panel Analysis

ML23177A274 OFFICE RES/SL NSIR/SL OEDO/ETA NAME KCoyne CJones HRodriguez DATE 06/ 28 /23 06/28 /23 06/27/23 OFFICE OGC(NLO) OGC (NLO) RIII/RA NAME MCarpentier HBenowitz JGiessner DATE 06/ 28 /23 06/ 28 /23 06/30 /23 OFFICE OEDO/DEDO OEDO/EDO NAME SMorris DDorman DATE 07/21 /23 08/25/23 DPO APPEAL PANEL ANALYSIS To review and address the issues raised in the Differing Professional Opinion (DPO) Appeal, the Executive Director for Operations (EDO) assigned the Regional Administrator (RA) for Region III as Chair of this DPO Appeal Panel. He was assisted by an Executive Technical Assistant from the EDOs office, two senior level technical advisors from the Office of Nuclear Regulatory Research (RES) and the Office of Nuclear Security and Incident Response (NSIR) who are also subject matter experts in the topics central to the issues raised in the DPO, and two attorneys from the Office of the General Counsel (OGC). This RA-led DPO Appeal Panel gathered information through discussions with the submitters, the Director of the Office of Nuclear Reactor Regulation (NRR), the initial DPO Panel, the licensee for the James A.

FitzPatrick Nuclear Power Plant (FitzPatrick), and other technical staff with expertise and knowledge in these areas. In addition, the DPO Appeal Panel reviewed key documents pertinent to this Appeal. The DPO Appeal Panel also gathered additional information through its own independent reviews of the FitzPatrick amendment request and other agency documents, as well as performing additional calculations to confirm the technical basis for the subject amendment approval. Collectively, the information gathered and presented here provides independent assessments, additional technical information, and the basis for the DPO Appeal Panel decision.

I. BACKGROUND

1. FitzPatrick Alternate Source Term Amendment On August 8, 2019, Exelon Generation Company, LLC (the licensee) submitted a license amendment request (LAR) for approval of an alternate source term for FitzPatrick in accordance with 10 CFR 50.67, Accident source term (ML19220A043). This application was supplemented on August 27, 2017 (ML19261A168) and responses to staff Requests for Additional Information (RAI) were received on January 16, 2020 (ML20017A052) and March 30, 2020 (ML20090E279). This LAR requested the following:

• Revise the radiological assessment calculational methodology for the Design Basis Accident (DBA) Loss-of-Coolant Accident (LOCA) to use NUREG-1465

• Revise the technical specifications (TS) total main steam line leakage rate limit from 46 standard cubic feet per hour (scfh) to 200 scfh for all four steam lines and adds a leakage limit of 100 scfh for a single steam line (when tested at 25 pounds per square inch gauge (psig))

• Delete Main Steam Leakage Control (MSLC) system from TS

• Revise TS ventilation filter testing program and secondary containment vacuum limit On July 21, 2020, the staff approved the alternate source term (AST) license amendment and issued the safety evaluation (ML20140A070).

2. Differing Professional Opinion (DPO) and Appeal On January 19, 2021, two staff filed a DPO (DPO-2021-001) on the James A. FitzPatrick Nuclear Power Plant - Issuance of Amendment No. 338, Re: Alternative Source Term For 1

ENCLOSURE 1

Calculating Loss-Of-Coolant Accident Dose Consequences (ML21042B867; non-public). The specific concern was that the FitzPatrick AST amendment did not comply with the NRC regulations contained in 10 CFR 50.67. On February 12, 2021, a DPO Ad Hoc Review Panel (DPO Panel) was formed and tasked by the NRC Differing Views Program to review DPO-2021-001. The DPO Panel subsequently issued its findings report to the NRR Office Director on June 9, 2021 (ML21160A229; non-public). With respect to the concerns presented in DPO-2021-001, the DPO Panel concluded that:

1. With regard to alleged errors in the licensees dose analysis, the DPO Panel agreed with the submitters on most of the issues that were identified. The DPO Panel did not draw a conclusion on whether the staff would normally be expected to identify these calculation issues during its review.

2. With regard to the staffs safety evaluation, the DPO Panel agreed with the submitters that staff analyses and insights were used to draw conclusions without the licensee providing corroborating information. The DPO Panel did not take a position on whether the licensee could have demonstrated compliance with the 10 CFR 50.67 dose criteria -

the DPO Panels concern was that the licensee was not asked to do so.

3. Regarding the proper use of technical specifications, the DPO Panel agreed with the submitters that neither technical specifications nor any other programmatic assurances were put into place to assure that the condenser pathway would be available following a core damage event.

4. Regarding the clarity of the licensing and design basis, the DPO Panel noted that the licensees intended analysis of record was clearly referenced, but they agreed with the submitters that this analysis was inaccessible and contained errors that - presumably still unresolved - would be included in all future alternative source term calculations under the original NRC approval.

5. With regard to the cumulative effect of the errors identified by the DPO submitters, the DPO Panel stated that while they did not believe that a health and safety issue existed with respect to issuance of the FitzPatrick amendment, the DPO Panel believed that the licensee did not provide sufficient information to demonstrate that its deterministic analysis meets 10 CFR 50.67 and that the small analytical margin means the estimated control room dose may exceed the 5 rem limit in 10 CFR 50.67 during this assumed design-based accident.

Following receipt of the DPO Panel report, the NRR Office Director discussed the findings with the DPO submitters on June 11, 2021. However, to better understand the DPO submitters concerns, the NRR Office Director assigned the Deputy Office Director for Engineering and a Technical Assistant from NRR to assist in an evaluation and documentation of the final NRR Office Directors Decision. The NRR Office Director issued a decision on November 17, 2021, regarding the DPOs concerns as informed by the DPO Panel report and the NRR evaluation (ML21236A254; non-public). The Directors Decision stated that NRRs assessment found that the FitzPatrick approved analysis of record was acceptable and met the regulatory requirements. The Director agreed with the DPO Panel recommendations to provide additional guidance to clarify the appropriate use and documentation related to the use of confirmatory and staff independent analyses and risk insights in safety evaluations. Also, the Director found that there was a lack of clarity and transparency regarding the basis for the staffs conclusions in some areas of the safety evaluation. For this reason, the Director also recommended that the internal guidance updates include the importance of clearly and transparently articulating the basis of the staffs conclusion.

2

Following issuance of the Directors Decision on the DPO, the submitters appealed the Directors Decision to the OEDO on January 7, 2022 (ML22039A062 case file; non-public). The submitters identified several broad categories of concerns similar to the issues initially raised in the DPO, including:

• NRC issued an amendment based on staffs analysis rather than the licensees analysis;

• Assumptions and methods used in the analysis were not appropriately justified (e.g.,

release points, radionuclide removal, limiting break location, credit for non-safety systems);

• Conflicting statements regarding how the power conversion system (PCS) was credited for radionuclide holdup;

• Lack of clarity of the licensing basis documentation; and

• A concern that three other LARs reviewed concurrently with the FitzPatrick LAR may also be impacted by these issues: Quad Cities increase in MSIV leakage path leak rate (ML20150A328), Nine Mile Point increase in MSIV leakage rate (ML20241A190), and Dresden increase in MSIV leakage path leak rate (ML20265A240).

Following receipt of the submitters appeal, the NRR Office Director reviewed the appeal package and issued its February 3, 2022, statement of views regarding the appeal of DPO-2021-001 that reaffirmed the initial Directors Decision (ML22031A053; non-public).

In July 2022, a DPO Appeal Panel was formed and included agency senior subject matter experts knowledgeable in radiological consequence analysis, probabilistic risk assessment, reactor licensing, and legal aspects of the issues presented in the DPO. The DPO Appeal Panel was led by the NRCs Region III Regional Administrator. The DPO Appeal Panel review team met with the submitters on October 4, 2022, and orally established a statement of concerns with the submitters. Following this meeting, the DPO Appeal Panel created a Summary of Concerns (SOC) for issues raised in the DPO Appeal and asked, in writing, for a review of the SOC by the submitters. The submitters provided comments to the SOC proposed by the DPO Appeal Panel and following a review of all the comments, the Appeal Panel and submitters agreed on a final SOC on January 13, 2023. The final SOC, which summarizes the DPO Appeal into four main issues that are each addressed in the following sections, is provided in Appendix 1.

3. Applicable Regulations 10 CFR 50.67 includes the following specific requirements that a licensee would request when proposing to use a revised accident source term:

(b) Requirements.

(1) A licensee who seeks to revise its current accident source term in design basis radiological consequence analyses shall apply for a license amendment under § 50.90. The application shall contain an evaluation of the consequences of applicable design basis accidents1 previously analyzed in the safety analysis report.

(2) The NRC may issue the amendment only if the applicant's analysis demonstrates with reasonable assurance that:

3

(i) An individual located at any point on the boundary of the exclusion area for any 2-hour period following the onset of the postulated fission product release, would not receive a radiation dose in excess of 0.25 Sv (25 rem) total effective dose equivalent (TEDE).

(ii) An individual located at any point on the outer boundary of the low population zone, who is exposed to the radioactive cloud resulting from the postulated fission product release (during the entire period of its passage), would not receive a radiation dose in excess of 0.25 Sv (25 rem) total effective dose equivalent (TEDE).

(iii) Adequate radiation protection is provided to permit access to and occupancy of the control room under accident conditions without personnel receiving radiation exposures in excess of 0.05 Sv (5 rem) total effective dose equivalent (TEDE) for the duration of the accident.

10 CFR 50.67(b)(1), Footnote 1 states:

The fission product release assumed for these calculations should be based upon a major accident, hypothesized for purposes of design analyses or postulated from considerations of possible accidental events, that would result in potential hazards not exceeded by those from any accident considered credible.

Such accidents have generally been assumed to result in substantial meltdown of the core with subsequent release of appreciable quantities of fission products.

This footnote is similar to other footnotes related to the fission product release for design basis accidents included in 10 CFR 50.34, Contents of applications; technical information, various sections in 10 CFR Part 52, Licenses, Certifications, and Approvals for Nuclear Power Plants, and 10 CFR 100.11, Determination of exclusion area, low population zone, and population center distance.

The DPO Appeal Panel notes that although 10 CFR 50.67 refers to a substantial meltdown of the core for each of the three criteria (exclusion area boundary (EAB), low population zone (LPZ), and control room), the regulations used for initial licensing have differences in how the accident is described, specifically:

• For the EAB and LPZ, 10 CFR 50.34(a)(1)(ii)(D) and 100.11 have similar footnotes referring to a core melt source term.

• For the control room, 10 CFR Part 50, Domestic Licensing of Production and Utilization Facilities, Appendix A, General Design Criteria for Nuclear Power Plants, Criterion 19, Control Room, has no similar footnote, but instead refers to accident conditions.

The difference in assumed source terms for EAB, LPZ, and control room calculated doses was examined during the NuScale design certification review (SECY 2019-0079). While the NRC staff noted this difference, NuScale calculated control room doses were reviewed using a bounding core melt source term, in part to accommodate review of certain Three Mile Island (TMI) action items. These items (10 CFR 50.34(f)(2)(vii) and (xxviii)) include provisions for adequate shielding and accessibility and reference (via footnote) the use of a core melt source term. These additional TMI-related requirements currently only apply to certain new license applicants under 10 CFR Part 52.

4

However, the accident source term final rule (64 FR 71990, Dec. 23, 1999) preamble has additional information regarding the assumed source term and DBA analysis. The preamble notes the following:

• The source term only includes the first three of five release phases (i.e., coolant, gap, and in-vessel). Ex-vessel and late in-vessel are considered inappropriate for design basis analysis purposes.

• The DBA conditions assumed in these analyses, although credible, generally do not represent actual accident sequences but are specified as conservative surrogates to create bounding conditions for assessing the acceptability of engineered safety features.

• These accident analyses are intentionally conservative in order to address uncertainties in accident progression, fission product transport, and atmospheric dispersion.

• Although probabilistic risk assessments (PRAs) can provide useful insights into system performance and suggest changes in how the desired defense in depth is achieved, defense in depth continues to be an effective way to account for uncertainties in equipment and human performance.

Further, the preamble for the accident source term final rule explicitly notes that the DBA fission product source term is applied to the control room operator radiation dose evaluations. For example, the preamble notes that this criterion is provided only to assess the acceptability of design provisions for protecting control room operators under postulated DBA conditions. The DBA conditions assumed in these analyses, although credible, generally do not represent actual accident sequences but are specified as conservative surrogates to create bounding conditions for assessing the acceptability of engineered safety features.

4. Immediate Safety Concern Evaluation With regard to the immediacy of a safety impact associated with the submitters issues, the DPO Appeal Panel notes that the calculated offsite dose consequences are a small fraction of the regulatory requirement of 25 rem TEDE. Specifically, the EAB and LPZ doses are 0.83 rem TEDE and 0.60 rem TEDE, respectively, and are not being disputed by the submitters. Although several of the submitters concerns could increase the offsite doses (e.g., aerosol deposition modeling, limiting DBA selection), the current dose results provide substantial margin to the regulatory limit. The DPO Appeal Panel considered the potential impact to the calculated control room dose and notes that the additional concerns could further increase the calculated control room dose of 4.67 rem TEDE above the regulatory limit of 5 rem TEDE. Specifically, issues associated with release pathways from the turbine building and consideration of MSIV packing leakage become more significant for the calculated control room dose due to the close proximity of the release point to the control room ventilation intake. However, the DPO Appeal Panel notes that the design basis accident described in Regulatory Guide (RG) 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Rev. 0, issued July 2000 (ML003716792), is a stylized scenario intended to assess the capability of engineered safety features, rather than an expectation of actual doses following a design basis event. In addition, additional defense-in-depth features such as the use of portable FLEX1 equipment, severe accident management guidelines, and emergency plans would 1 Following the accident at the Fukushima Dai-ichi Nuclear Power Plant, the NRC issued Order EA-12-049 to require nuclear power plant licensees to use an approach for adding diverse and flexible mitigation 5

provide further mitigation of operator and public doses. On this basis, the DPO Appeal Panel concluded that, while the issues described in the DPO Appeal are serious and have significant relevance to the regulatory process, they do not represent immediate safety issues requiring expedited agency action such as an order or use of the NRCs Integrated Risk-Informed Decisionmaking Process for Emergent Issues (LIC-504; ML19253D401) evaluation process.

Experience with Control Room Doses During Actual Accidents The immediate safety question aside, the DPO Appeal Panel notes that actual radiation doses following a severe accident can represent a substantial safety challenge to the operators.

Experience from the Fukushima Dai-ichi accident provides insights for actual control room doses during an event causing significant core damage. As reported in a Sandia National Laboratory Report, SAND2017-9329R,2 the operator, Tokyo Electric Power Company (TEPCO),

was contacted for information on doses measured in the control room during the accident at Fukushima Dai-ichi. Information from TEPCO was provided on the "shine" dose in the joint control rooms for Units 1 & 2 and for Units 3 & 4:

• In the Units 1 & 2 joint control room, the dose was between 1 mrem/hr and 350 mrem/hr.

• In the Units 3 & 4 joint control room, the dose was between 350 mrem/hr and 1500 mrem/hr.

However, the DPO Appeal Panel notes that the extreme accident conditions at Fukushima were far in excess of those that are considered in an AST design basis event. In particular, the AST only considers the gap and in-vessel release phases equivalent to arrested core damage accident sequence.

5. Summary of Concerns and Review Activities A full SOC for the DPO Appeal is included in Appendix 1. The DPO Appeal Panel conducted the following activities to assess these concerns:

• Assessed the immediate safety impact for the issues described in the DPO Appeal.

• Met with the DPO submitters on October 4, 2022.

• Met with the NRR Office Director on November 14, 2022.

• Interviewed the dose consequence analyst who provided additional analysis supporting the Directors Decision (December 21, 2022), the lead reviewer for PRA (February 7, 2023), and the lead dose consequence reviewer (February 10, 2023).

• Submitted two rounds of information requests to NRR to further clarify key points in the Directors Decision and the AST LAR decision-making process (included in Appendices 5 and 6).

• Conducted a detailed review of the DPO materials, including associated guidance documents and technical reports.

strategiesor FLEXthat will increase defense in depth for nuclear power plants beyond design basis events (ML12054A735).

2 https://www.osti.gov/biblio/1762012-treatment-reactor-systems-within-draft-regulatory-guide-dg 6

II. Evaluation of Specific Issues Identified in the Summary of Concerns The DPO Appeal Panels review and assessment of these issues described in the SOC is provided below.

1. SOC Issue 1 - Compliance With 10 CFR 50.67 Issue 1: Determine if the FitzPatrick AST license amendment is in compliance with 10 CFR 50.67, if it is based on the applicants analysis as required by 10 CFR 50.67(b)(2), and whether the information discussed in SOC Issue 2 (Section II.2) would result in control room doses higher than those allowed under 10 CFR 50.67.

Response to Issue 1:

The DPO Appeal Panel notes that the applicant submitted in its AST LAR an analysis with offsite and control room doses calculated to be below the 10 CFR 50.67(b) acceptance criteria.

The NRC granted the FitzPatrick AST amendment on July 21, 2020 (ML20140A070),

presumptively on the basis that the licensees analysis demonstrates with reasonable assurance per 10 CFR 50.67(b)(2) that these acceptance criteria were met. However, the DPO Appeal Panel carefully considered the issues raised by the submitters and, as described under the evaluation of SOC Issues 2, 3, and 4, was unable to independently reach a conclusion, based on the best available information and current knowledge, that the licensees analysis, when appropriately adjusted to account for non-conservatisms and potential errors, demonstrates with reasonable assurance that the calculated control room dose would remain below 5 rem TEDE for the duration of the accident.

Specific issues considered by the DPO Appeal Panel include the following:

• Failure to consider the potential dose impact associated with MSIV packing leakage (discussed in SOC Issue 2.a). This leakage path was previously mitigated by the MSIV leakage control system and packing leakage from the outboard MSIV was, and continues to be, a recognized release path to the environment under the FitzPatrick licensing basis. Consideration of this leakage path would increase the calculated control room dose.

• Identification of the limiting LOCA break location as the recirculation piping vice a main steam line break (discussed in SOC Issue 2.e). Given that the source term associated with the AST LOCA is deterministic, the main impact of considering a recirculation line break is to enhance fission product removal by main steam line deposition. Therefore, assuming a main steam line break would reduce fission product deposition and increase control room doses.

• Non-conservative modeling of aerosol deposition in the main steam lines, which if adjusted to reflect a more modern understanding of aerosol physics, would increase the control room dose (discussed in SOC Issue 2.f). While this is an extremely complex and nuanced topic, the DPO Appeal Panel notes that providing credit for both drywell spray and main steam line deposition early in the accident is not supported by the physical configuration of the reactor coolant system, and the licensee relied on assumptions that 7

have been known to be non-conservative since at least 2006 (e.g., AEB-98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application using the Revised (NUREG-1465) Source Term, (ML011230531) and an August 23, 2006, memo from the RES Director of the Division of Risk Assessment & Special Projects to the NRR Associate Director for Risk Assessment & New Projects (ML062010315; nonpublic), regarding revised aerosol settling velocities). To offset uncertainties associated with aerosol modeling, the staff credited the existence of a pathway to the power conversion system (e.g., condenser) in considering the dose consequences of MSIV leakage (ML20140A070). However, there is a lack of a clearly defined licensing basis and regulatory controls for fission product removal assumptions associated with the power conversion system (discussed in SOC Issue 3). Further, credit for the availability of the power conversion system for fission product deposition was not requested by the licensee in its AST LAR.

Although the Directors Decision indicated that certain conservatisms not modeled in the LAR analysis would offset these issues, the DPO Appeal Panel could not substantiate that these represented either actual conservatisms (e.g., reactor building dilution volume, in-vessel fission product holdup, suppression pool credit) or that the conservatisms were appropriately captured in the facility licensing basis (deposition in all four main steam lines or the power conversion system). Further, the DPO Appeal Panel believes that conservatisms that are needed to support a regulatory decision, are inherently part of the decision basis and as such need to be appropriately captured in the licensing basis (LB) (e.g., if the condenser is desired to provide dose margin, then control in Technical Specifications or the Technical Requirements Manual, as well as procedures and program controls would need to be captured in the LB). These issues are discussed further under SOC Issues 2.g and 3.

The DPO Appeal Panel further notes that the analysis supporting licensing decisions under 10 CFR 50.67 are intended to be conservative in order to provide a sufficient level of defense in depth and consideration for the high degree of uncertainties associated with severe accident behavior. As described by the Commission in the preamble to the alternate source term final rule (64 FR 71990), the intention of the analysis is to use conservative DBA surrogates to create bounding conditions for assessing the acceptability of engineered safety features. The Commission further states that these analyses are intentionally conservative in order to address uncertainties in accident progression, fission product transport, and atmospheric dispersion. While it recognizes that PRAs provide useful insights, the Commission states that defense in depth continues to be an effective way to account for uncertainties in equipment and human performance. Although the staffs safety evaluation for the FitzPatrick AST amendment includes references to risk-informed decision-making and the direction in SRM-SECY-19-0036 (ML19183A408), the use of reasonably conservative analytical assumptions is consistent with the Commissions stated purpose for the AST DBA analysis.

The DPO Appeal Panel acknowledges that the NRC issued the AST amendment for FitzPatrick in July 2020. However, in light of the issues noted above, the DPO Appeal Panel concludes that issuance of the AST amendment does not comply with the criteria specified in 10 CFR 50.67(b)(2) because the applicants analysis does not demonstrate with reasonable assurance that calculated control room dose does not exceed 5 rem TEDE for the duration of the accident.

Therefore, the DPO Appeal Panel recommends potential agency actions to resolve this issue in Section IV.1 of this Enclosure.

8

2. SOC Issue 2 - Use of Accurate and Consistent Evaluation Methods Issue 2: Determine if the LOCA assumptions and methods used by FitzPatrick3 and the NRC staff to evaluate the FitzPatrick AST license amendment for a revised LOCA analysis are accurate and consistent with the stated regulations and standard review plan guidance (or if deviations from guidance by the licensee were documented and included an adequate technical basis).

Response to Issue 2:

In assessing if the LOCA assumptions and methods used by FitzPatrick and the NRC staff to evaluate the AST license amendment were accurate and consistent with the stated regulations, associated guidance, and standard review plan guidance, the DPO Appeal Panel considered the seven specific technical issues raised by the submitters as summarized in the SOC. The evaluation of each of these issues (identified as Issues 2.a through 2.g) is summarized in the following sections:

a. SOC Issue 2.a - Leakage from the MSIV Packing, Valve Body, and Mechanical Joints Issue 2.a: The licensee amendment review process and the Directors Decision did not consider leakage from the MSIVs packing, body or mechanical joints directly to the environment.

Response to Issue 2.a: The DPO Appeal Panel agrees with the submitters that the license amendment review process and Directors Decision did not adequately address potential leakage from the MSIV packing. Further, the DPO Appeal Panel determined that consideration of packing leakage from the outboard MSIV was included in the FitzPatrick licensing basis prior to the AST amendment and was previously controlled by a leakage control system that would direct MSIV packing leakage to a filtered release path. The DPO Appeal Panel also agrees that leakage in the vicinity of the outboard MSIV into the steam tunnel area would potentially result in an unfiltered release to the environment.

In the analysis supporting the FitzPatrick AST LAR, the licensee did not credit use of the Main Steam Leakage Control System (MSLCS). As a result, the licensee requested that the MSLCS be removed from the FitzPatrick licensing basis, including from Technical Specifications. Section 2.1.3 of the staffs safety evaluation states that the MSLCS collects leakage from the stem packing of all four outboard MSIVs and directs the leakage from the MSIVs to the standby gas treatment system (SGTS) for processing. This description is consistent with the Fitzpatrick updated final safety analysis report (UFSAR) Section 9.19, Main Steam Leakage Collection System, prior to approval of the AST license amendment, which noted that the MSLCS is designed to collect and process leakage across the seats of 3 In its AST LAR (ML19220A043), FitzPatrick indicated that its application conformed with the relevant guidance of RG 1.183 and was consistent with other guidance contained in NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants LWR [Light-Water Reactor] Edition, Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Rev. 0, issued July 2000 (ML003734190) and RG 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants, issued June 2003 (ML031530505).

9

the MSIVs and to collect and process stem packing leakage from the outboard containment MSIVs following a design basis LOCA. The effluent of the MSLCS is processed by the Standby Gas Treatment System (SGTS) and is exhausted through the Stack. The TS Bases for the MSLCS noted that both the stem packing and the downstream portion of each subsystem contribute to reducing uncontrolled or untreated MSIV leakage. Further, UFSAR Figure 4.6-3, Main Steam Isolation Valve 29-AOV-80A-D,86A-D, (which has been retained in the UFSAR following implementation of the AST amendment) notes that the interface between the MSIV body, stem, and packing is a potential leakage path but that these potential leak paths do not exist when the MSLCS is lined up.

Therefore, the UFSAR acknowledged that: (1) the MSIV packing was a potential leakage path and (2) a safety function of the MSLCS was to collect leakage from stem packing of the MSIVs. However, the staffs safety evaluation did not address how the removal of the MSLCS, and subsequent loss of the ability to control possible outboard MSIV packing leakage, was assessed when approving the AST license amendment.

In assessing this issue during the initial DPO review, the DPO Panel agreed with the submitters that the licensee and NRR staff did not address the MSIV packing leakage path and noted the following in Section 4.1.2 of the DPO Panel report:

10

[L]eakage up the [MSIV] stem could occur; therefore, the leakage determined by the surveillance test included leakage across the valve seat as well as up the stem. The panel noted the leak rate test cannot distinguish where the actual leak originates, making it possible that all leakage is through the stem pathway. The licensee did not assess this leakage path nor provide justification that leakage through this pathway would not have a detrimental impact on dose rates.

The NRR Office Director disagreed with the DPO Panel on this assessment and stated the following:

If a direct release from the stem of the outboard MSIVs to the environment were to occur, it would release in the reactor building, which has been evaluated for the safe shutdown earthquake (SSE) and it is assumed to remain structurally intact. Any leakage in the reactor building gets processed by the standby gas treatment (assumed to be in operation 20 minutes after a loss of coolant accident (LOCA)). This results in doses to the control room that are less than the evaluated scenario of MSIV leakage through the seat of the valves.

In their DPO Appeal, the submitters noted that the Directors Decision regarding MSIV packing leakage may have been based on miscommunication with a DPO Panel member and that the conclusion that MSIV packing leakage would be directed to the reactor building and SBGT system was flawed. Although the submitters included this new information in their Appeal (e.g., footnote 7 in the Enclosure to their Appeal), the NRR Office Director in their February 3, 2022, statement of views regarding the appeal of the DPO4 reaffirmed the initial Directors Decision.

The outboard MSIVs are located in the steam tunnel area that runs from the outside of the primary containment building to the turbine building. A review of the FitzPatrick UFSAR indicates that the steam tunnel area appears to have free air flow with the turbine building (which is outside the reactor building boundary and not connected to the SBGT system).

Therefore, the DPO Appeal Panel requested clarification from NRR on the flow path outboard MSIV packing leakage would take to reach the SBGT system. NRR representatives reaffirmed the belief that if release from the stem of the outboard MSIVs were to occur, it would release in the reactor building and be processed by the SBGT system (assumed to be in operation 20 minutes after a LOCA).

The submitters assertion that MSIV packing leakage would flow to the turbine building is consistent with RG 1.96, Design of Main Steam Isolation Valve Leakage Control Systems for Boiling Water Reactor Nuclear Power Plants, Revision 1 (ML003740263), which states, It should be noted that any leakage from the stem packing of the outboard isolation valve would contribute to the 2-hour dose, since in most designs such leakage would escape to the turbine building and the environment via the steam tunnel.

The DPO Appeal Panel reviewed licensing basis information contained in the FitzPatrick UFSAR and could not locate any information supporting NRRs assertion that MSIV packing leakage would be directed to the reactor building and SBGT system following removal of the 4 Memo from Director, Office of Nuclear Reactor Regulation, to the Executive Director for Operations, Statement of Views Regarding Appeal of Differing Profession Opinion Concerning DPO-2021-001, February 3, 2022 (ML22031A053, non-public).

11

MSLCS. The following UFSAR information review by the DPO Appeal Panel is pertinent to this issue:

• UFSAR Section 12.3.2, Turbine Building, states that the main steam lines to the turbine generator from the reactor are housed in a reinforced concrete tunnel that enters the Turbine Building after passing under the adjacent Administration Building.

The reinforced concrete tunnel walls and roof are designed for structural loadings and radiation shielding.

• UFSAR Figure 12.3-7, Reactor Building Cross Section 1-1, shows the outboard main steam MSIVs located within the steam tunnel, an area that appears to be separate and distinct from the reactor building. This is further supported by UFSAR Figures 12.3-2 and 12.3-3, which appear to show no gross flow area between the steam tunnel and reactor building. Of note is that Figure 12.3-3 shows several removeable slabs that form a portion of the boundary between the steam tunnel and the motor generator set room.

• UFSAR Section 9.14.4, Post Accident Sampling System, notes that the reactor water recirculation pump motor generator set room (a room located just above the main steam pipe tunnel and shown on UFSAR Figure 12.3-7), is located outside the secondary containment structure.

• UFSAR Section 9.9.3.4, Turbine Building Ventilation System, notes that [t]he main steam tunnel is exhausted by two full capacity fans located on the roof of the Administration Building. Supply air to the main steam tunnel is induced from the Turbine Building. This indicates that there is an air flow path from the turbine building to the steam tunnel area.

• UFSAR Section 5.2.4.6, Containment Isolation, notes that for a steam leak outside the primary containment between the thermal sleeve and the outer main steam line isolation valve, steam blows directly into the pipe tunnel and into the turbine building, further support that the air volume within the tunnel area has a flow path into the turbine building.

Collectively, this information supports the submitters perspective that outboard MSIV packing leakage, that is no longer controlled by the MSLCS, would follow a pathway into the steam tunnel area and into the environment via the turbine building or through the unfiltered turbine building ventilation exhaust (if operating). In order to confirm the DPO Appeal Panels understanding of steam tunnel ventilation, two Panel members met with engineering representatives from FitzPatrick, who confirmed that steam leakage into the steam tunnel area would be released to the environment through an unfiltered path. Further, FitzPatrick engineering staff confirmed that there is no path for steam leakage in the steam tunnel to be routed to the SBGT system following removal of the MSIV leakage control system.

FitzPatrick staff also noted that there is temperature instrumentation in the steam tunnel to detect a main steam line break (and actuate automatic closure of the MSIVs) and that the MSIVs have the capability to be fully backseated during normal operation. The DPO Appeal Panel does recognize that there is BWR plant-to-plant variability with regard to steam tunnel design and operation, with some plants having a different ventilation system configuration (including the use of blowout panels to the turbine building), which could change the impact of outboard MSIV packing leakage. However, the DPO Appeal Panel concluded that for 12

FitzPatrick, the available evidence indicates that outboard MSIV packing leakage would be vented to the environment through an unfiltered release path.

The DPO Appeal Panel was unable to assess the amount of potential leakage from the MSIV packing, but offers the following observations:

• Despite recognition in the FitzPatrick UFSAR of the MSIV outboard valve packing being a potential release point, no assumptions concerning MSIV packing leakage appear to have been included in either the AST LAR or the staffs safety evaluation.

• Consideration of packing leakage is routinely considered for engineered safety features and included in potential dose consequences. For example, the FitzPatrick AST safety evaluation (Section 3.1.1.3) states, in part:

ECCS leakage develops when ESF [engineered safety features]

systems circulate suppression pool water outside containment, and leaks develop through packing glands, pump shaft seals, and flanged connections. ESF leakage releases into the secondary containment during RB [reactor building] drawdown are assumed to leak directly to the environment as a ground level release with no filtration. After the assumed 20-minute drawdown period, ESF releases are assumed to be mixed in the secondary containment with a 50 percent mixing efficiency filtered by the SGTS and released by the SGTS vent.

• The DPO Appeal Panel believes that licensees would normally be expected to promptly identify MSIV packing leakage that occurs during normal operation and address such circumstances under their operations and maintenance programs.

However, certain factors may limit the ability to identify latent packing leaks during normal operations. For example, the rapid closure operation of the MSIVs during the assumed accident may impact the leak tightness of the stem-packing interface and result in the initiation of leakage following automatic valve closure. Further, backseating the MSIVs during normal operation would reduce or prevent leakage through the packing when the MSIV is fully open, but result in packing leakage once the valve moves off the backseat during a closure operation. The DPO Appeal Panel also notes that the outboard MSIV at FitzPatrick is oriented such that reactor vessel pressure is applied to both the packing and the main disc when the valve is closed.

• Draft DRA-ISG-2021-01, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms (ML21278A372)5, recognizes the potential for MSIV packing leakage, but states that the leakage through the packing represents a small fraction of the leakage, because such leakage must follow a tortuous path through the packing. Further, the flow area through the packing is small, resulting in a small leak rate because the leak rate is dependent on the flow area. The DPO Appeal Panel notes that this conclusion is consistent with the TS Bases B 3.6.1.8, Main Steam Leakage Collection (MSLC) System, which noted that 5 Draft Interim Staff Guidance (ISG) DRA-ISG-2021-XX, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms, was released for public comment in May 2021 (ML21078A051). Draft DRA-ISG-2021-01 (ML21278A372) is an updated version of the ISG that was publicly released in October 2021 to support a November 2021 Advisory Committee on Reactor Safeguards meeting and includes the staff resolution of public comments on DRA-ISG-2021-XX.

13

while the MSLCS reduced mitigated packing leakage, the primary function was to mitigate MSIV seat leakage (e.g., see TS 3.6.1.8 Bases page markups contained in Attachment 4 of the FitzPatrick AST LAR submittal; ML19220A043).

The DPO Appeal Panel reviewed recent Licensee Event Reports (LERs) to assess if MSIV packing leakage constituted a credible leakage path that could challenge main steam line leakage limits. This review identified several examples of excessive MSIV leakage due to packing failures since 2000:

LER Number Plant Comments LER 50-333/2000-015-02 FitzPatrick Identified loss of packing preload and packing degradation on outboard MSIV for main steam lines B and C.

LER 50-271/2004-001-00 Vermont Inboard main steam line B MSIV Yankee leakage exceeded TS allowed leakage with a leakage rate of 219.86 scfh due to packing leakage caused by binding of the stem and packing follower from scoring and galling.

LER 50-277/2006-001-00 Peach Bottom Valve 86A outboard main steam line A Unit 2 MSIV had as-found packing leakage of sufficient magnitude that prevented pressurizing the test volume to the target value during surveillance testing.

LER 0-296/2012-002-00 Browns Ferry Valve 3B Outboard MSIV leak rate Unit 3 exceeded TS limit with as-found leak rate of 781 scfh due to inadequate packing.

LER 50-259/2014-004-00 Browns Ferry 1B outboard MSIV exceed TS limit with Unit 1 a leakage rate of 114.7 scfh due to inadequate packing.

LER 50-293/2016-010-01 Pilgrim On December 15, 2016, operators identified packing leakage on the 2C IR 05000293/2017003 MSIV and a body to bonnet leak on the 2D MSIV during a steam tunnel IR 05000293/2016004 walkdown. As-found leakage test not conducted, but cause of leakage for the IR 05000293/2017007 2C MISV was worn packing from scoring on the valve stem. The leakage for the 2D MSIV leakage in December 2016 was due to the gasket used in the valve body to bonnet interface being out of specification resulting in excessive valve body to bonnet gap.

LER 50-440/2023-002 Perry On March 4, 2023, an evaluation of data from Local Leak Rate Testing (LLRT) determined that the as found maximum pathway leakage for all four MSIVs to be in excess of the Technical Specification (TS) Surveillance 14

LER Number Plant Comments Requirement (SR) 3.6.1.3.10 leakage limit of less than or equal to 250 scfh. In addition, when the B outboard MSIV was tested, the leakage was indeterminate due to a large packing leak. The plausible cause of the outboard MSIV B leakage from the packing gland was attributed to stem misalignment during the work on the valve in 2021 (1R18). This misalignment caused gouges on the stem and ultimately the packing gland leakage. The resulting radiological dose at the Exclusion Area Boundary (EAB) would be 44.29 rem.

LER 50-263/2023-001-00 Monticello A packing leak on valve AO-2-86C, 13 Outboard Main Steam Isolation Valve (MSIV), exceeded the TS leakage limit on March 4, 2023. The condition was discovered on March 21, 2023, when it was concluded that a reasonable expectation of operability was not met, and the MSIV was declared inoperable in accordance with TS. During a maintenance outage on March 26, 2023, the Local Leak Rate Test as-found leakage was 987.1 scfh which exceeded the TS SR limit for individual MSIV leakage of 100 scfh. The cause of the packing leakage is attributed to scoring on the valve stem.

In addition to LERs, the DPO Appeal Panel also reviewed recent Inspection Reports and 10 CFR 50.72 notifications because it was not clear that excessive MSIV packing leakage was consistently reported under the requirements of 10 CFR 50.73. A review of the Inspection Reports issued since 2016 identified the following additional operating experience:

Inspection Report Plant Comments 05000237/2022003 AND Dresden Unit Failure to select appropriate packing for 05000249/2022003 3 use in the Unit 3 1B (inboard) MSIV that resulted in leakage from the valve 05000237/2022002; during operation and eventual shutdown 05000249/2022002 AND of Unit 3 to perform emergent 07200037/2022001 maintenance on April 10, 2022. The licensee had installed a packing set that was later determined to be unsuitable 15

Inspection Report Plant Comments for the application and had torque values less than the vendors recommended minimum.

05000263/2018002 Monticello June 21, 2018Power was reduced to approximately 80 percent for a packing adjustment and partial valve stroking on C outboard main steam isolation valve.

05000277/2015004 AND Peach Bottom On December 13, 2015, Unit 2 05000278/2015004 Unit 2 commenced a shutdown from 100 percent rated thermal power and entered into a forced outage to repair packing on the 2 C inboard MSIV.

05000220/2019003 AND Nine Mile Prior to the most recent refueling 05000410/2019003 Point Unit 1 outage, Nine Mile Point Unit 1 had been experiencing a rising trend in unidentified leakage. It was determined during the outage that the cause was steam leakage from the main steam isolation valve IV-01-01 packing.

The DPO Appeal Panel found that this operating experience demonstrates that packing leakage, including packing leakage capable of exceeding main steam line leakage requirements, occurs and is a potentially credible leakage path to the environment. This appears to contradict statements in both draft DRA-ISG-2021-XX and DRA-ISG-2021-01 that leakage through the packing represents a small fraction of the leakage because packing leakage follows a tortuous path and through a small flow area. While packing leakage may normally be a small fraction of total leakage, recent operating experience demonstrates that packing leakage can represent a significant contribution to main steam line leakage. Furthermore, it is not clear to the DPO Appeal Panel that licensees universally consider outboard MSIV packing leakage during operation a condition that could challenge TS limits for main steam line leakage.

In summary, the DPO Appeal Panel agrees with the submitters that control of packing leakage from the outboard MSIV was a safety function described in the FitzPatrick licensing basis, but an assessment of the impact of the removal of the MSLCS on this function was not provided by the licensee nor documented in the staffs safety evaluation. The DPO Appeal Panel could not confirm NRRs assertion that leakage from the MSIV packing would be directed to the SBGT system and, based on a review of the FitzPatrick UFSAR, agrees with the submitters that any outboard MSIV packing leakage should be assumed to flow to the turbine building and represent an unfiltered release path to the environment. The DPO Appeal Panel believes that this issue was an error in the staffs review of the FitzPatrick AST LAR In addition, operating experience indicates that MSIV packing leakage, including leakage in excess of TS limits, is a credible failure mechanism for the MSIVs, but may not be generally recognized as potentially impacting the containment safety function. This issue may be exacerbated by the lack of discussion of how MSIV packing leakage was considered within the specific licensing basis documents such as the UFSAR and staff safety evaluations. The 16

DPO Appeal Panel believes this is an area where issuance of generic communications may be useful to ensure a consistent understanding of the potential for MSIV packing leakage to represent potential containment bypass pathway.

b. SOC Issue 2.b - Crediting Air Space Between Reactor Vessel and Closed MSIVs Issue 2.b: By crediting the air space in the main steam line up to the closed MSIVs, the applicants analysis did not assume the radioactive release was instantaneously and homogenously mixed throughout all the free air volume in the drywell containment. Despite the licensees statements that the containment elemental and particulate (natural) deposition/plateout is not credited, this deposition is credited in the space upstream of the MSIVs that is part of the drywell containment. (RG 1.183, Appendix A, Section 3.1)

Response to Issue 2.b: The DPO Appeal Panel could not substantiate the submitters concerns with regard to this issue. The DPO Appeal Panels conclusion is similar to the conclusion reached by the initial DPO Panel, which also did not substantiate the concern.

As noted by the submitters, RG 1.183, Appendix A, Section 3.1, states, in part:

The radioactivity released from the fuel should be assumed to mix instantaneously and homogeneously throughout the free air volume of the primary containment in PWRs or the drywell in BWRs as it is released.

However, RG 1.183, Appendix A, Section 6.3, states:

Reduction of the amount of released radioactivity by deposition and plateout on steam system piping upstream of the outboard MSIVs may be credited, but the amount of reduction in concentration allowed will be evaluated on an individual case basis. Generally, the model should be based on the assumption of well-mixed volumes, but other models such as slug flow may be used if justified.

Although no additional guidance is provided for how to assess an individual case basis, NRC Regulatory Issue Summary (RIS) 2006-04, Experience with Implementation of Alternative Source Terms, (ML053460347) provides some additional context for this issue:

Modeling of [Main Steam Line] MSL piping may include volumes between the reactor pressure vessel and the inboard MSIV (inboard volume), between the inboard and outboard valves (in-between volume), and outside of the outboard valve (outboard volume). Since a majority of large (i.e., heavier) particles deposit in the inboard volume, the distribution of the aerosol that leaks to the subsequent volume is smaller (i.e., lighter) particles.

The clarification provided in RIS 2006-04 acknowledges that deposition may be credited within the inboard volume (between the reactor vessel and the inboard MSIV). Therefore, the DPO Appeal Panel concluded that crediting deposition upstream of the inboard MSIV is consistent with the guidance contained in RG 1.183. In its LAR, the licensee states that deposition between the reactor pressure vessel nozzle and the inboard MSIV is credited using the volumetric model described by the initial DPO Panel (i.e., each steam line is represented by a two volume well-mixed model between the reactor pressure vessel nozzle 17

and the turbine stop valve with the breakpoint between volumes either at the inboard or outboard MSIV).

The DPO Appeal Panel notes that SAND2008-6601, Analysis of Main Steam Isolation Valve Leakage in Design Basis Accidents Using MELCOR 1.8.6 and RADTRAD, Section 6.3, was referenced in DG-1199, Draft Revision 1 to RG 1.183, as an acceptable means for modeling main steam line deposition. Section 6.3 of SAND2008-6601 recommends that no credit for aerosol deposition be taken for the portion of the main steam lines upstream of the inboard MSIV. The basis for this recommendation is two-fold: (1) concerns that high temperatures in the in-board portion of the piping may cause vaporization of fission products already deposited on the piping, and (2) natural convection driven bi-directional flow in the main steam lines could enhance bulk transport of fission products from the steam dome to the steam lines. The DPO Appeal Panel concluded that while these factors increase the uncertainty in the amount of fission products deposited in the inboard portion of the main steam lines, it does not preclude some level of deposition. The DPO Appeal Panel notes that the licensee provided an assessment of the potential for revaporization of fission products in the main steam lines in section 7.5.3 of calculation JAF-CALC-19-00005 submitted with its AST LAR. This assessment concluded that the maximum temperature in the main steam lines would remain substantially below the vaporization threshold for cesium iodide (CsI).

The DPO Appeal Panel also considered the timing aspects of fission product releases from the main steam lines. RG 1.183, Regulatory Position 3.3, provides guidance for the timing of fission product releases during the LOCA design basis accident. For BWRs, this guidance states that the gap release phase begins at two minutes with a 0.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> duration, followed by the early in-vessel release phase commencing at 0.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> and lasting 1.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> in duration. The guidance further states that the release should be modeled as increasing in a linear fashion over the duration of the phase, but in lieu of treating the release as a linear ramp, the release can be assumed to be instantaneous at the start of the release phase.

The FitzPatrick AST LAR modeled LOCA releases in a linear fashion using RADTRAD,6 rather than assuming the entire fission product release occurred at the beginning of the release period phase. Further, use of the well-mixed volume assumption, as endorsed in Appendix A, Section 6.3 of RG 1.183, is based on a steady state concentration model. This well-mixed assumption ignores transient effects associated with the buildup of fission product concentrations within the modeled volume.

c. SOC Issue 2.c - Turbine Building Release Point Issue 2.c: By assuming a release point at a specific location inside the turbine building, rather than on the surface of the turbine building, improper dilution was credited in the turbine building. This dilution reduced the atmospheric dispersion factors, thereby underestimating the control room dose. (RG 1.183, Appendix A, Section 6.4)

Response to Issue 2.c: The DPO Appeal Panel did not substantiate the submitters concern that assuming a release point in the turbine building was improper or resulted in an underestimate of the calculated control room dose. While the calculation of control room 6 The RADionuclide, Transport, Removal, and Dose Estimation (RADTRAD) code is a licensing analysis tool used to show compliance with nuclear plant siting criteria for the site boundary radiation doses at the EAB and the LPZ and to assess the occupational radiation doses in the control room and/or Emergency Offsite Facility for various LOCA and non-LOCA DBAs.

18

dose has a high sensitivity to assumptions related to the location of the specific release point, the crediting of a specific release point inside the turbine building does not contradict the guidance in RG 1.183. The DPO Appeal Panel notes that assumptions regarding the specific release point either on the surface or inside the turbine building have a negligible impact on calculation of offsite dose results and for the emergency planning zone or LPZ.

The initial DPO Panel was neutral on this issue and the Directors Decision concluded that it was appropriate to model the release point at the turbine stop valves.

In their initial DPO and Appeal, the submitters asserted that by crediting the length of piping between the outboard MSIVs and the turbine stop valves (TSVs), the distance between the point of release and the intake to the control room ventilation system was increased by approximately 33 feet (from 63 feet to 96 feet). This increased distance resulted in lower atmospheric dispersion factors, which reduced the calculated control room dose due to MSIV leakage by a factor of 2. The submitters stated that by extending the release point an additional 33 feet from the edge of the turbine building to the turbine stop valves, the total calculated control room dose would increase from approximately 4.7 rem TEDE to 8 rem TEDE. In support of the initial DPO review, the NRR staff performed additional calculations to examine the sensitivity of control room dose to release point location. Specifically, the following cases for the impact of distance between the release point and the control room ventilation intake were compared:

Table 1: Calculated Control Room Dose Sensitivity to Release Point Sensitivity Licensee (rem TEDE) NRR Analysis Supporting Case Directors Decision DPO (rem TEDE) Submitter MSIV Total MSIV Total (rem TEDE)

Leakage Leakage Baseline 3.44 4.67 3.50 4.73 -

(~96)

Reduced - - 6.7 7.92 8 Distance

(~63)

These results are based on a two-steam line deposition model, consistent with the documented licensing basis for the AST amendment as discussed further under SOC Issue 2.e. Changing the distance to the control room ventilation intake results in a change to the

/Q atmospheric dispersion factors. In evaluating these results, the DPO Appeal Panel notes that: (1) the NRR assessment is in agreement with the submitters estimate of control room dose impact for a reduced distance; and (2) locating the release point approximately 33 feet closer to the control room ventilation intake does increase the dose contribution from MSIV leakage by approximately a factor of 2. The DPO Appeal Panel believes that the similarity in sensitivity results from the submitters and the NRR analyst provides confidence in the accuracy and reliability of the NRR-developed model and its use.

The DPO Appeal Panel reviewed the pertinent guidance documents related to this issue, including RG 1.183, Appendix A, Sections 6.4 and 6.5. Section 6.4 states that the MSIV leakage should be assumed to be released to the environment as an unprocessed, ground-level release and that holdup and dilution in the turbine building should not be assumed.

Section 6.5 notes that a reduction in MSIV releases that is due to holdup and deposition in 19

main steam piping downstream of the MSIVs and in the main condenser, including the treatment of air ejector effluent by offgas systems, may be credited if the components and piping systems used in the release path are capable of performing their safety function during and following a safe shutdown earthquake (SSE). As documented in both the license amendment request and the staffs safety evaluation, the piping from the reactor pressure vessel (RPV) nozzle to the seismic boundary break at the TSVs is Seismic Class I and Quality Assurance Category I (safety-related). Further, the DPO Appeal Panel agrees that crediting this piping with providing additional distance for the release point, which impacts the calculated /Q atmospheric dispersion factors, is not equivalent to assuming either holdup or dilution in the turbine building. Therefore, the DPO Appeal Panel agrees with the Directors Decision that assuming a release point within the turbine building is well supported, reasonable, and not inconsistent with RG 1.183 guidance.

d. SOC Issue 2.d - Crediting Nonsafety-Related Structures, Systems, and Components Issue 2.d: By crediting nonsafety-related structures, systems, and components such as vents and doors that are not controlled by technical specifications, the licensee omitted release pathways that would be closer to the control room and would result in higher control room doses. As such, the accident analysis no longer aligns with, or is bounded by the plant design basis. (RG 1.183, Regulatory Position 5.1.2 and RG 1.194, Regulatory Position 2.0).

Response to Issue 2.d: The DPO Appeal Panel did not substantiate the submitters concerns regarding the lack of regulatory controls over equipment such as vents and doors in the turbine building. The DPO Appeal Panel determined that, based on a review of UFSAR Figures 12.3-15, Administration BLDG Plan EL 300 FT - 0 IN and Roof, and 12.3-17, Turbine Building Plan - EL 272 FT - 0 IN, that the licensee considered the shortest straight-line distance between the turbine stop valve release point and the control room ventilation intake. This conclusion differs from the determination made by the initial DPO Panel, which concluded that the licensee did not provide a bounding source-to-receptor distance and that the licensees analysis underestimated the atmospheric dispersion factors and resulting calculated control room dose. However, the DPO Appeal Panel conclusion is consistent with the Directors Decision in that the licensees approach is reasonable.

In Calculation No. JAF-CALC-19-00004, Section 3, Assumptions, the licensee noted the following with regard to consideration of alternate release paths in the turbine building:

Additional release pathways out of the Turbine Building in the event of a LOCA were evaluated in ECR 5000012124 (Ref. 16). The smoke ejector vents on top of the Turbine Building are normally closed and would not open in a post-LOCA scenario. The access door from the Turbine Building 323 elevation to the Administrative Building roof is normally kept closed, locked, and monitored by security. There is also a Turbine Building exhaust duct that is 60 from the Control Room Intake. This was also determined not to be an appropriate release location because the vent is monitored for radiation, and would be closed to prevent a radiological release to the environment. In the event of a loss of offsite power (LOOP) in combination with the LOCA, there would be considerable holdup in the Turbine Building. Therefore, using the 20

TSVs as the release location to the environment is both conservative and appropriate.

The DPO Appeal Panel notes that while the alternate release paths described by the licensee would be pertinent to the evaluation if the licensee were to assume the release point was located on the surface of the turbine building (as discussed under SOC Issue 2.c),

the need to make this assumption was not substantiated. Further, given that the licensee used the shortest straight-line distance between the turbine stop valves and the control room intake when evaluating the /Q atmospheric dispersion factors, considering the other release paths would serve to further increase the distance and/or result in the need to make assumptions not supported by the engineering design of the turbine building. For example, transport of fission products through the turbine building ventilation exhaust would result in substantial mixing and dilution. While assuming such mixing would be contrary to RG 1.183, Appendix A, Regulation Position 6.4, which notes that [h]oldup and dilution in the turbine building should not be assumed, the only credible means for fission products to be released via turbine building exhaust is when the ventilation system is operating and such mixing is technical justifiable.

Although the DPO Appeal Panel does agree with the Directors Decision conclusion for this issue, the DPO Appeal Panel is not in agreement with several assumptions stated in the Directors Decision. For example, the Directors Decision noted that it is acceptable to assume that nonsafety-related equipment functionality is lost or that the normal configuration of nonsafety-related equipment is acceptable for DBA analysis. The DPO Appeal Panel does not agree with this position. Instead, consistent with the definition of safety-related equipment in 10 CFR 50.2, nonsafety-related equipment should normally be assumed to provide no accident mitigation function. In this case, the DPO Appeal Panel believes that use of the straight-line distance between the turbine stop valves and the control room ventilation intake for calculating /Q atmospheric dispersion factors is bounding and no mitigation function is provided by the turbine building, smoke ejector vents, doors, or ventilation operation. This view is not consistent with the decision basis stated in the Directors Decision, which postulated that nonsafety-related equipment (smoke ejector vents and doors) would be in its normal position or unavailable due to a loss of offsite power (turbine building ventilation). With regard to offsite power, given the assumed deterministic nature of the fission product release, there is no need to assume that a loss of offsite power is the most limiting condition. In fact, RG 1.183, Regulatory Position 5.1.2 states that

[a]ssumptions regarding the occurrence and timing of a loss of offsite power should be selected with the objective of maximizing the postulated radiological consequences.

In addition, the Directors Decision does not appear to acknowledge that an impact of the AST LAR for FitzPatrick was a requirement to re-establish turbine building ventilation for the purposes of vital area access. Specifically, Section 3.11.16, Vital Area Accessibility, of to the AST LAR, notes the following with regard to Turbine Building ventilation:

However, removal of the MSLC system introduces higher post-LOCA doses in the Turbine Building due to the increased MSIV leakage now having a travel path to the Turbine Building. In order to minimize post-LOCA doses in the Turbine Building and to ensure all releases in the Turbine Building are monitored, the Turbine Building ventilation system must be restarted (if it trips due to a loss of offsite power (LOOP) coincident with a LOCA). This action is already part of the plant response to high gaseous radiation signals in the Reactor Building, Refuel Floor, Turbine Building, or Radwaste Building per 21

site emergency procedures. The post-LOCA Turbine Building dose analysis assumes that the Turbine Building ventilation system is restarted within 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> of initiation of the LOCA.

The LAR further notes that it may be necessary for operators to remove relays in order to override the high radiation trip of the turbine building ventilation system. For this case, FitzPatrick would be expected to establish turbine building ventilation within 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> of a design basis accident. Even if a loss of offsite power were to be assumed, FitzPatrick stated that the 4-hour station blackout coping time is consistent with offsite power recovery within 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. Therefore, statements contained in the Directors Decision relating to turbine building ventilation system operation do not appear to be consistent with the current FitzPatrick licensing basis.

e. SOC Issue 2.e - Limiting LOCA Break Location Issue 2.e: The analysis did not consider the limiting LOCA break location by selecting a recirculation line break, rather than a break in the reactor coolant system just prior to the MSIVs. In addition, when the licensee modeled the break just before the MSIVs (provided as a sensitivity study in response to the staffs request for additional information), the licensee did not use the assumptions that were used in its LOCA analysis supporting its license amendment request. (RG 1.183, Appendix A)

Response to Issue 2.e: The DPO Appeal Panel agrees with the submitters that the AST amendment did not consider the most limiting LOCA break location. In its license amendment request, the licensee assumed a recirculation line break as the most limiting break location. Specifically, in Section 3.11.1 of the LAR, the licensee states:

The LOCA, as with all DBAs, is a conservative surrogate accident that is intended to challenge selective aspects of the facility design. With regard to radiological consequences, a large-break LOCA is assumed as the design basis case for evaluating the performance of release mitigation systems and the containment response. Therefore, a recirculation line rupture is considered as the initiating event rather than a main steam line rupture.

RG 1.183, Appendix A, states that [l]eaks up to a double-ended rupture of the largest pipe of the reactor coolant system are included. The LOCA, as with all design basis accidents (DBAs), is a conservative surrogate accident that is intended to challenge selective aspects of the facility design. However, the recirculation line does not represent the largest pipe or the most limiting break location in the reactor coolant system. The portions of the main steam lines within the drywell are larger than the recirculation lines and are more limiting as a break location. Specifically, a main steam line break within the drywell can result in a concurrent failure of the inboard MSIV in the associated steam line and reduce the available piping area for radionuclide deposition in the main steam line. In the safety evaluation for the AST amendment, the staff questioned the licensees basis for selecting a recirculation line break vice a main steam line break during the review and issued a request for additional information ARCB-RAI-3. The licensee responded to the RAI by providing the results of a sensitivity study comparing the results of the recirculation line break to an assumed main steam line break (ML20090E279). In these sensitivity cases, the licensee did not credit aerosol deposition or plateout in the line segment from the RPV nozzle to the inboard MSIV for this main steam line associated with the break location. The staff concluded in the safety evaluation that, based on the additional analysis of the dose consequences resulting from 22

an assumed main steam line break performed by the licensee, the recirculation line break as submitted in the LAR resulted in a slightly higher calculated dose for the [control room]

than the analysis which modeled an MSL break in one line. In assessing this conclusion, the DPO Appeal Panel concluded that the comparison between the recirculation break and a main steam line break was based on a different set of assumptions. Further, when the impact of the break location was compared using assumptions consistent with the licensing basis defined in the AST amendment, a main steam line break is more limiting for calculated control room doses and may exceed the 5 rem TEDE criteria stated in 10 CFR 50.67(b)(2).

This DPO Appeal Panel conclusion is consistent with findings of the initial DPO Panel, which noted the following in Section 4.1.5 of the DPO Panel report:

The panel agrees with the submitters that the analysis of record was based on leakage through two main steam lines while the evaluation for a ruptured main steam line was performed assuming leakage through four main steam lines.

The panel agrees, in general, that the impact of an assumption cannot be determined by making multiple changes to assumptions or methodologies in a proposed licensing basis. The panel agrees these changes can obscure the impact of the issue being studied.

In the Directors Decision on the DPO, the Director agreed that when evaluating the most limiting scenario, the integrity of the assumptions is crucial for a direct comparison. The Director further noted that the NRC staff should have followed up with the licensee and requested a more direct comparison and provided a detailed evaluation of the acceptability of the licensees docketed response and sensitivity study results in the safety evaluation. The Director further stated that their offices independent assessment, using assumptions of the analysis of record to model a break in the reactor coolant system (RCS) just prior to the MSIVs, resulted in dose results under regulatory limits.

The DPO Appeal Panel reviewed the Directors Decision including its dose assessment and associated sensitivity cases and interviewed the NRR dose analyst who performed those assessments. The Appeal Panel concluded that the dose analyst that conducted the independent assessment was knowledgeable, objective, and unbiased in their conduct of these calculations, noting that the analyst had not been made privy to the initial DPO or its contents. In discussions with this dose analyst, the DPO Appeal Panel determined that no sensitivity cases assuming deposition in only two steam lines had been performed for a main steam line break. At the request of the DPO Appeal Panel, the dose analyst ran a main steam line break case using only two main steam lines for deposition. The results of this analysis, along with the analysis that credits deposition in all four steam lines are provided in Table 2. The licensees analysis provided in response to the staffs RAI on this issue is also shown.

23

Table 2: Main Steam Line Break Sensitivity Cases Control Room Calculated Total Number of Dose due to Control Room Case main steam MSIV Leakage Dose lines credited (rem TEDE) (rem TEDE)

Licensee:

Recirculation Line Break 2 3.44(1) 4.67 (baseline)

Licensee: Break in Main Steam Line (MSL) A - 4 3.34 4.57(2)

ARCB-RAI-3 Response Licensee: Break in Main Steam Line (MSL) B - 4 3.37 4.60(2)

ARCB-RAI-3 Response NRR Analysis:

Recirculation Line Break 2 3.50 4.73(2)

(baseline)

NRR Analysis: Break in 4 3.88 5.11(2)

MSL A (4 MSL model)

NRR Analysis: Break in 4 3.87 5.10(2)

MSL B (4 MSL model)

NRR Analysis: Break in 2 4.76 5.99(2)

MSL B (2 MSL model)

(1) The licensee revised the dose values in their calculation of record based on their response to ARCB-RAI-1b (ML20090E279).

(2) The total calculated control room dose includes an additional 1.23 rem TEDE to account for containment leakage, ESF leakage, containment shine, and the external cloud release pathways.

The licensee provided the results for two steam line break sensitivity cases: the first was a break in main steam line A and the second was a break in main steam line B. Both of the licensees sensitivity cases assumed that all four main steam lines were available for aerosol deposition. However, the calculation of record for the AST LAR performed by the licensee assumed leakage only in the two shortest steam lines (B and C) and no leakage (and therefore no aerosol deposition) assumed in the longer A and D main steam lines.

Therefore, for the two-steam line deposition case (which is consistent with the AST licensing basis), only the results for a main steam line break in line B are provided. As shown in Table 2 above, the total calculated control room dose for all main steam line break sensitivity cases performed by NRR exceed 5 rem TEDE; this appears to be contrary to statements in the Directors Decision stating that the NRR assessment used the assumptions of the analysis of record to model a break in the RCS just prior to the MSIVs, and the dose results remain under regulatory limits.

Although the NRR dose analyst main steam line break results for the four-steam line deposition case are slightly higher than the licensee sensitivity results, they identify that reducing deposition credit from four steam lines to two steam lines increases the calculated control room dose due to MSIV leakage by approximately 0.9 rem TEDE. Further, the 0.9 rem TEDE difference between deposition credit for two versus four main steam lines was consistent for several different sensitivity studies prepared by the dose analyst. The DPO Appeal Panel determined that the safety evaluation conclusion that the recirculation line break resulted in a slightly higher calculated dose for the control room than a main steam line break was dependent on assuming radionuclide deposition in all four main steam 24

lines, which offset the increase associated with assuming a main steam line break rather than a recirculation line break. As such, the control room dose impact of assuming a break in the main steam line versus recirculation line would potentially result in a calculated increase of 0.9 rem TEDE if other conditions are left unchanged. Given the limited margin to the 5 rem TEDE control room dose requirement, the calculated 0.9 rem TEDE increase would cause control room dose to exceed 5 rem TEDE for a limiting main steam line break.

The DPO Appeal Panel reviewed the licensing basis for FitzPatrick, as described in the AST amendment, to verify that the conditions assumed for the DBA analysis considered deposition in only two main steam lines. The AST amendment revised the Surveillance Requirement (SR) related to main steam line leakage (SR 3.6.1.3.10) to read: [v]erify combined main steam line leakage rate is 200 scfh, and 100 scfh for any one steam line, when tested at 25 psig. 10 CFR 50.36, Technical specifications, defines technical specification limiting conditions for operation (LCO) as the lowest functional capability or performance levels of equipment required for safe operation of the facility. Further, 10 CFR 50.36 defines surveillance requirements as requirements relating to test, calibration, or inspection to assure that the necessary quality of systems and components is maintained, that facility operation will be within safety limits, and that the limiting conditions for operation will be met. The main impact of apportioning leakage to additional main steam lines in the AST calculation is to provide more radionuclide deposition area. Assuming the MSIV leakage is distributed over all four main steam lines would maximize deposition area and not define the lowest functional capability of the system. Therefore, the lowest functional capability for main steam line deposition would be the availability of the minimum number of steam lines. In the case of FitzPatrick, that would be two steam lines each with 100 scfh of leakage (at 25 psig) and the remaining two lines with no leakage. Note that at the design pressure for containment (45 psig), the MSIV leakage scales up to 135 scfh for any steam line and a total of 270 scfh for all main steam lines.

In its LAR, the licensee assumed only two main steam lines were available for deposition, consistent with the lowest TS functional capability. Specifically, LAR Section 3.11.10, Determination of MSIV Leak Rates, states the following:

A total of 270 scfh MSIV leakage is assumed to occur as discussed below. Of the total leakage, 135 scfh is assumed to be through the steam line with the "failed" MSIV. The "MSIV failed" line means that the inboard MSIV in one of the shortest MSLs fails to close and remains open during the accident, which extends the well mixed volume boundary from the RPV nozzle to the outboard MSIV. This MSIV failure complies with a single active component failure requirement that results in the most limiting radiological consequences (RG 1.183, Section 5.1.2). The failure is non-specific and the consequences of the failure are that a single steam line is assumed to have a disproportionately high flow to artificially increase the total allowed MSIV leakage.

The shortest of the remaining three intact steam lines is assumed to leak at 135 scfh in accordance with Regulatory Guide 1.183, Appendix A, Section 6.2, which states: "all of the MSIVs should be assumed to leak at the maximum leak rate above which the technical specifications would require declaring the MSIVs inoperable." For the shortest intact main steam line, one well-mixed volume is between the RPV nozzle and the inboard MSIV with a second volume between the inboard MSIV and the TSV. As in the "failed" 25

line, the deposition removal of aerosol in the horizontal pipe, and the deposition removal of elemental iodine in both the horizontal and vertical pipes, are credited in the steam line between the RPV nozzle and the TSV.

The remaining two intact steam lines are assigned a leakage of 0 scfh in the analysis. The remaining two intact steam lines are therefore not explicitly modeled in the analysis but are represented by the leakage through the second shortest main steam line. This is conservative because the shorter MSL length results in less removal.

The staff reviewed this assumption and noted it in Section 3.1.1.4, Assumptions on Main Steam Isolation Valve Leakage of the staffs safety evaluation report:

The licensee took credit for aerosol deposition in both the inboard and outboard section of the MSL between the outboard MSIV and the TSV. The licensee evaluated the total of 270 scfh MSIV leakage by assuming that the leakage would be equally divided between the two shortest MSLs. The licensee assigned an MSIV leakage of 135 scfh to the MSL with the failed open inboard MSIV by assuming one well-mixed volume between the RPV nozzle and outboard MSIV. The licensee assigned a second well-mixed volume in this line between the outboard MSIV and TSV.

The licensee credited aerosol deposition in the second shortest MSL between the RPV nozzle and inboard MSIV. The licensee modeled this section as one well-mixed volume between the RPV nozzle and inboard MSIV. The licensee modeled a second well-mixed volume between the inboard MSIV and TSV.

The licensee limited credit for aerosol deposition to the horizontal sections of the MSLs, while the airborne elemental iodine in this release path is assumed to be adsorbed on the entire MSL surface area. The NRC staff finds that the licensee used modeling assumptions consistent with previous NRC-accepted practice for the evaluation of MSL aerosol deposition, and this aspect of its MSL aerosol deposition model is, therefore, acceptable.

Therefore, the licensing basis associated with the AST amendment includes the assumption that the total main steam leak rate is distributed across two shortest of the four main steam lines. This assumption is consistent with the TS SR, the LAR submittal, and RG 1.183.

Given this licensing basis, the staff conclusion, as documented in Section 3.1.1.4 of the safety evaluation, that the analysis of the dose consequences resulting from an assumed MSL break provided by the licensee demonstrates that the impact of including an MSL break does not significantly impact the dose consequences from MSIV leakage for FitzPatrick is based on an assumption (i.e., removal of radionuclides by deposition in four main steam lines) not included in the AST licensing basis.

The DPO Appeal Panel notes that the licensees previous design basis LOCA assessment described in UFSAR Sections 14.6.1.3 and 14.8.2 identify a recirculation line break as the limiting LOCA location. For example, UFSAR Section 14.6.1.3 states that [t]he most severe Reactor Coolant System effects and the greatest release of radioactive material to the primary containment results from a complete circumferential break of one of the recirculation loop lines. However, as discussed in the staffs AST safety evaluation, the licensee identified in its AST LAR that postulating a main steam line break in one steam line inside the drywell would maximize the dose contribution from the MSIV leakage. Further, the assumed source term for the AST is deterministically specified and does not depend on the LOCA break location. The use of the less limiting recirculation line break is not consistent 26

with RG 1.183, Appendix A, in that the recirculation line break is not a conservative surrogate accident. Further, the basis the staff used to approve a less limiting break location was inconsistent with the licensees AST LAR request and the resultant NRC-approved licensing basis, which used different assumptions for main steam line deposition. The staffs safety evaluation did not describe how the offsetting factors of a more limiting main steam line break location, when combined with optimistic assumptions for main steam line aerosol deposition, were assessed in making a licensing decision. The DPO Appeal Panel considers this to be an error in the staffs review and approval basis of the FitzPatrick AST LAR.

f. SOC Issue 2.f - Aerosol Deposition in the Main Steam Lines Issue 2.f: By using removal coefficients for aerosol settling that are nonconservative and by incorrectly modeling the impact of changing aerosol particle size distribution as the radioactive material moves down the steam line, the LOCA model overestimates the deposition in the steam lines and, therefore, underestimates the dose in the control room.

(RG 1.183, Appendix A, Sections 6.3 and 6.5)

Response to Issue 2.f: The DPO Appeal Panel agrees that the credit for deposition of fission products in the main steam lines was addressed in a non-conservative manner in the FitzPatrick AST license amendment. This issue is further exacerbated by the combined credit for drywell spray removal and main steam line deposition, which does not appear to appropriately address the preferential removal of larger aerosol particles in the drywell. This conclusion is consistent with the initial DPO Panel, which concluded that the panel agrees with the submitters that the analysis of record does not address the interdependence of aerosol removal and the use of drywell sprays specifically, larger aerosol particles in the containment atmosphere will be preferentially removed, making subsequent removal by deposition in downstream piping more challenging. The DPO Appeal Panel has provided a historical summary of aerosol deposition modeling in Appendix 2.

With regard to this issue, the Directors Decision stated:

[O]ur independent assessment determined that the licensees dose analysis is in compliance with 10 CFR 50.67, follows the regulatory guidance, and it is not in error. We did identify three parameters used as inputs to the licensees calculation that may not have been the most conservative (i.e., elemental removal coefficient and the particulate removal coefficient used for the drywell and the MSLs). The NRC staff addressed and approved these parameters in the FitzPatricks SE. We agree that more conservative values would increase calculated doses. We performed an independent analysis using more conservative values for the elemental removal coefficient, the particulate removal coefficient in the drywell and particulate removal coefficients in the MSLs and found that the dose in the control room results is fractions above the 10 CFR 50.67 limit (5.11 rem). It is important to note that this independent analysis used a new method to calculate deposition in the MSLs that was not available at the time of the review. We agree with you that considering these more conservative values, in FitzPatricks analysis of record, results in a dose that exceeds the regulatory limit. However, when considering the substantial conservatisms noted above, we have reasonable assurance the dose will remain well below 5 rem if the postulated scenario were to occur. In addition, the methodology used by the licensee was an 27

acceptable method that was consistent with the regulatory guidance at the time of the review.

The DPO Appeal Panel notes that the Directors Decisions independent analysis of the calculated control room dose of 5.11 rem was later found to be an underestimate that resulted from a calculational error (discussed further under SOC Issue 2.g). Although the Directors Decision acknowledged that the FitzPatrick AST LAR overestimates the deposition in the MSL and did not address the interdependence of aerosol removal and the use of drywell sprays, the Director stated that there was reasonable assurance the dose will remain below 5 rem. This calculation referenced by the Directors Decision included several assumptions:

• Use of the multigroup method endorsed by DG-1389 (draft Revision 1 to RG 1.183)

(ML21204A065) to model main steam line deposition using an initial aerosol particle size distribution consistent with the 2009 Nuclear Energy Agency (NEA) state of the art report (NEA/CSNI/R(2009)57)

• Limiting drywell spray iodine removal coefficients to 20 hr-1 versus the FitzPatrick AST licensing basis of 26.36 hr-1

• Deposition in four main steam lines versus deposition in two main steam lines The DPO Appeal Panel noted that the calculated control room dose referenced in the Directors Decision was based on a model crediting aerosol deposition in four main steam lines, not two main steam lines as described in the FitzPatrick licensing basis. Therefore, the DPO Appeal Panel requested the independent RADTRAD analyst who previously prepared the calculations supporting the Directors Decision to provide an additional sensitivity study crediting deposition in only the two shortest main steam lines to provide a comparison case that was consistent with the FitzPatrick AST licensing basis. In performing this additional sensitivity study, the RADTRAD analyst immediately identified a non-conservative error in the previous multi-group calculation results.8 This error increased the 5.11 rem TEDE value referenced in the Directors Decision to 7.49 rem TEDE (a value approximately 50% greater than the 5 rem TEDE control room design criterion). Further, the sensitivity analysis for the two main steam line deposition case (which is consistent with the licensing basis) resulted in a calculated control room dose value of 10.4 rem TEDE, a value more than twice the 10 CFR 50.67(b)(2) control room dose requirement under design basis accident conditions.

These results are summarized In Table 3 below:

7 https://www.oecd-nea.org/upload/docs/application/pdf/2021-03/csni-r2009-5.pdf 8 The DPO Appeal Panel notes that this error was immediately discovered by the analyst when running the additional sensitivity case for deposition in two steam lines rather than four steam lines. The analyst noted that they had prepared the two steam line case previously for NRR, but NRR staff did not request that this additional sensitivity run be performed. The DPO Appeal Panel believes that had NRR previously requested a full range of sensitivity studies, which could have been readily performed by the analyst, this error would have been discovered earlier.

28

Table 3: Sensitivity Analysis of Various MSL Deposition Cases Number of Main Steam Line Iodine Removal Calculated Steam Lines Deposition Model Coefficient Total Credited for Control Deposition Room Dose (rem TEDE) 4 Directors Decision: 20 hr-1 5.11 Re-evaluated AEB-98-03 (found to be with multi-group in error) 4 Corrected: 20 hr-1 7.49 Re-evaluated AEB-98-03 with multi-group 2 Corrected: 20 hr-1 10.42 Re-evaluated AEB-98-03 with multi-group The basis for the Directors Decision regarding aerosol deposition appears to be that the NRR sensitivity study showed that calculated control room doses were only slightly above 5 rem TEDE and that substantial conservatisms not included in the calculation provide assurance that the control room dose is below 5 rem TEDE. The DPO Appeal Panel assessed the conservatisms referred to in the Directors Decision and determined, with the exception of not crediting the containment pressure reduction function of the safety-related sprays, the reference conservatisms either diverged from the specific licensing basis for the FitzPatrick AST amendment (e.g., crediting deposition in four main steam lines), lacked a documented technical basis (e.g., crediting dilution in the entire reactor building volume, suppression pool scrubbing), or were inconsistent with the early in-vessel release associated with the AST evaluation (deposition on reactor vessel internals). These issues are described in more detail under the assessment of SOC Issue 2.g. Given that the corrected sensitivity result provided a control room dose of 10.4 rem TEDE, the DPO Appeal Panel believes that the calculated control room dose would be above the 5 rem TEDE requirement unless additional mitigation factors were credited.

Current Guidance for Determining Aerosol Deposition Credit for fission product deposition in the main steam lines is included in RG 1.183, but no specific guidance is provided for acceptable methods. Specifically, RG 1.183, Appendix A, Section 6.3, states that:

Reduction of the amount of released radioactivity by deposition and plateout on steam system piping upstream of the outboard MSIVs may be credited, but the amount of reduction in concentration allowed will be evaluated on an individual case basis. Generally, the model should be based on the assumption of well-mixed volumes, but other models such as slug flow may be used if justified.

RG 1.183 does not reference acceptance methods for calculating deposition. RIS 06-04, Experience with Implementation of Alternative Source Terms, provides additional guidance. Specifically, RIS-06-04 states that use of AEB-98-03 was acceptable. AEB-98-03, issued in December of 1998, documented the staff's evaluation of the thermal-hydraulics, 29

fission product deposition, and resulting radiological consequences using the revised source term for the Perry nuclear power plant given the proposed plant changes. However, the RIS further noted that any licensee who chooses to reference the AEB-98-03 assumptions should provide appropriate justification that the assumptions are applicable to their particular design.

Although staff has issued draft RG 1.183 updates to provide more specific guidance for aerosol setting (e.g., DG-1199 in 2009 and DG-1389 in 2022), the only approved guidance is the limited guidance in RG 1.183 and RIS-06-04. However, the staff and industry have continued to increase the knowledge base regarding aerosol deposition since 1999 when 10 CFR 50.67 was promulgated. In particular, the staff has known since 2005 that the method described in AEB-98-03 contains errors and non-conservatively predicts aerosol deposition. The evolution of the staffs understanding of these areas is summarized in the following timeline (additional detail provided in Appendix 2):

• December 1998 - AEB-98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application using the Revised (NUREG-1465) Source Term, issued

• July 2000 - RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactor, Revision 0, issued

• March 2006 - RIS-06-04, Experience with Implementation of Alternative Source Terms, issued

• August 2006 - Concerns regarding the analysis in AEB-98-03 were communicated to the NRR Associate Director for Risk Assessment & New Projects in an August 23, 2006, memo from the RES Director of the Division of Risk Assessment & Special Projects (ML062010315, nonpublic). These issues included recognition, as early as April 2005, that the AEB-98-03 approach included potentially nonconservative aerosol density and diameter distributions (ML062010341, nonpublic). As a result, the memo noted that the original AEB-98-03 report overestimated aerosol settling velocity.

• January 2007 - NRR management noted that there was insufficient information available to perform a backfit analysis or revise regulatory positions based on the issues identified with AEB-98-03 (ML070180233, nonpublic). The memo stated that NRR intended to prepare a User Need memorandum requesting an update to AEB-98-03 and that this information would be incorporated into an update to RG 1.183.

NRR also noted that the information provided in the 2006 RES memorandum does not present an immediate safety concern. There is sufficient conservatism in the accident dose analyses and assumptions to accommodate minor increases in dose based on the errors in AEB-98-03.

• June 2007 - NRR transmitted to RES the User Need request for the work referenced in the January 2007 memo on June 24, 2007 (ML071620396, nonpublic). This request reiterated that NRR recognized in 2005 that some of the assumptions in AEB-98-03 needed to be re-evaluated and that supplemental information was needed to support a backfit analysis of previous uses of AEB-98-03. This request 30

was identified as UNR NRR-2007-005 and requested MELCOR9 analysis of postulated recirculation line and main steam line breaks to determine the accident source term and fission product removal in containment, the main steam lines, and the condenser.

• July 2009 - RES completed the work specified in UNR NRR-2007-005 with the issuance of Sandia technical report SAND2008-6601, which included the conduct of an independent review by Dycoda, a consulting firm with extensive experience in severe accident analysis (ML083180181, nonpublic; SAND2008-6601, publicly available). SAND2008-6601 provided an updated approach to model aerosol deposition, noted that fission product concentrations in the main steam lines can be higher than the drywell during the early phase of a LOCA, and stated that drywell spray and deposition in the main steam lines should not be simultaneously credited early prior to reactor vessel reflood.

• October 2009 - DG-1199, Draft Revision 1 to RG 1.183 issued for public comment (74 FR 52822 ). The draft guidance endorsed portions of SAND2008-6601 (i.e.,

Section 6.3) as an acceptable means for modeling main steam line deposition. The NRC received 150 comments on DG-1199 (summarized in draft, non-public, ML112340473). As noted during a public meeting held in November 2020, work on finalizing DG-1199 was deferred as staff reviewed several AST related licensing submittals, including the AST submittal for FitzPatrick (ML20296A425). The NRC staff elected not to finalize DG-1199 and instead issued DG-1389 in 2022 as the new proposed revision 1 to RG 1.183.

• December 2009 - NEA/CSNI/R(2009)5, State-of-the-Art Report on Nuclear Aerosols, issued. NEA/CSNI/R(2009)5 provided information on the status of existing experimental data and analytical capabilities required for predicting aerosol source terms from LWR accidents. Of particular note is that the fission product aerosol particle size distributions referenced in this publication are smaller than those calculated by AEB-98-03, indicating that AEB-98-03 may overpredict aerosol deposition.

• April 2022 - DG-1389, Draft Revision 1 to RG 1.183 issued for public comment (87 FR 23891). The DG-1389 guidance for addressing fission product deposition in the main steam lines provides for three different acceptable options, but such credit is contingent upon not crediting fission product removal by drywell sprays.

Specifically, Appendix A, Section A-5, of DG-1389 states:

Three methods are presented below to compute aerosol deposition within main steamlines. Each method computes similar removal coefficients suitable for radiological consequences calculations, however, these methods are not valid if credit has been taken for aerosol removal from drywell sprays. The three MSIV leakage models are the:

1. Direct adoption of the SAND 2008-6601 (Ref. A-11) recommendations without scaling R*-factors; 9 MELCOR is a fully integrated, engineering-level computer code developed by Sandia National Laboratories for the NRC to model the progression of severe accidents in nuclear power plants.

31

2. Re-evaluated AEB-98-03 with multi-group; and

3. Numerical Integration.

When DG-1389 was released for public comment, the staff also requested comment on a draft staff technical assessment: Technical Assessment of Hold-up and Retention of Main Steam Isolation Valve Leakage within the Main Steam Lines and Main Condenser (ML20085J042). In addition, DG-1389 included reference to a memorandum providing further guidance on performing the re-evaluated AEB-98-03 with multi-group and numerical integration (i.e., Technical Basis for Draft RG 1.183 Revision 1 (2021) Re-evaluated AEB-98-03 Settling Velocity Method, the Multi-Group Method, and the Numerical Integration Method, July 2021 (ML21141A006)).

As noted in DG-1389, Appendix A, Section A-5.6.2, both the re-evaluated AEB-98-03 with multi-group and the numerical integration method should assume a log-normal aerosol diameter distribution with an aerodynamic mass median diameter (AMMD) of 2.0 micrometers and geometric standard deviation of 2.0. Equation A-5 of DG-1389 is used to determine the median aerosol diameter in the log-normal distribution based on these parameters. The aerosol size distribution parameters specified in DG-1389 are supported by data included in NEA/CSNI/R(2009)5.

DPO Appeal Panel Assessment of Aerosol Deposition Guidance Based on a review of this history, the DPO Appeal Panel determined the following:

• Non-conservatisms and errors in AEB-98-03 were identified as early as 2005. An initial assessment of these issues was formally communicated to NRR in 2006.

• By 2007, NRR had recognized the need to evaluate the impact of the AEB-98-03 issues on prior approvals of AST LARs using the backfit process. At that time, the staff basis for not treating these issues as a more significant safety issue was that the staff believed that there were sufficient conservatisms in the accident dose analyses, assumptions, and modeling to accommodate minor increases in dose based on the errors in AEB-98-03 (ML070180233). The DPO Appeal Panel notes that additional information and trends toward reduced conservatisms in DBA evaluation may have eroded the basis for this conclusion. Additionally, despite the indication that NRR had intended to perform a backfit analysis for the AEB-98-03 issues, the DPO Appeal Panel could locate no backfit analysis in ADAMS.

• By 2009, the NRC had developed an updated understanding of aerosol behavior based on the issuance of SAND2008-6601 and NEA/CSNI/R(2009)5.

• Although the staff provided updated draft guidance to address aerosol deposition with the issuance of DG-1199 in 2009, this guidance has never been finalized.

• Although the draft guidance contained in DG-1389 was not issued until 2022, the staff was (or should have been) aware of errors in the AEB-98-03 approach related to estimates of aerosol particle size distributions as early as 2005 and concerns related to crediting both spray and aerosol deposition in the main steam lines by 2009. In fact, industry shared the concerns with crediting both dry well sprays and 32

main steam line deposition in a 2010 paper10 when recommending [w]hen applying the AST for MSIV leakage dose assessment, credit may be taken either for deposition in the steam lines between the reactor vessel and the inboard MSIV or for drywell sprays but not for both.

• The DPO Appeal Panel could locate no generic communication or other guidance document formally issued by the NRC informing the industry of the known non-conservatisms and errors associated with AEB-98-03 or issues associated with dual credit of sprays and main steam line deposition.

FitzPatrick AST License Amendment and Licensing Precedent Context In the Fitzpatrick AST LAR, the licensee used a 20-group probabilistic distribution of aerosol settling velocity for MSIV leakage based on AEB-98-03. The licensee applied the parameter distributions for aerosol density, particular aerosol diameter, and shape factors as described in AEB-98-03 (e.g., Section 3.11.11 of Attachment 1 to the LAR) to determine aerosol particle size and resulting settling velocity. The licensee noted that a similar approach had been previously approved for Clinton (ML052570461), Limerick (ML062210214), and LaSalle (ML101750625). The DPO Appeal Panel noted that the 20-group approach used for the FitzPatrick AST LAR did not address the issues described in the 2006 RES memo that highlighted the fact that AEB-98-03 had a mistake regarding the shape factor distribution in addition to other non-conservatisms that overpredicted the original settling velocities (i.e.,

ML062010315, non-public).11 A comparison between the LAR, the original 1998 AEB-98-03 settling velocity distribution, the 2006 update to the AEB-98-03 settling velocity distribution, and the numeric integration approach recommended in DG-1389 are provided in the below figure:

10 J.E. Metcalf, P.B. Perez, BWR Steam Line Radionuclide Concentration Distribution following a DBA LOCA, presented at the 31st Nuclear Air Cleaning Conference, International Society of Nuclear Air Treatment Technologies, Charlotte, NC, 19-21 July 2010.

http://isnatt.org/Conferences/31/09.%20Metcalf%20NACC-Paper-Final.pdf 11 The DPO Appeal Panel notes that the memos written in the 2006 and 2007 timeframe regarding non-conservatisms in the AEB-98-03, were, and continue to be, non-publicly available. Therefore, given the absence of any additional guidance or generic communication regarding this issue, the DPO Appeal Panel does not believe it would be reasonable for a licensee to address these issues directly in their AST submittal. However, these issues should have been well known to the staff and should have been addressed as part of the review and the resolution of RAIs, as applicable.

33

The main differences between these various settling velocity distributions are assumptions regarding aerosol diameter, density, and shape factor. Smaller aerosol diameters, lower density, and higher shape factor all tend to decrease settling time. Higher settling velocities increase fission product removal by deposition and result in higher fission product removal efficiencies and higher removal coefficients. A key factor influencing the overall removal efficiency and removal coefficient is that as aerosol particles with higher settling velocities are removed, the aerosol distribution shifts toward smaller particle sizes and slower settling velocities. If this factor is not accommodated in the model, fission product removal by deposition can be overestimated. The settling velocities assumed in the FitzPatrick AST LAR are, in most cases, more than two orders of magnitude greater than those determined using a recommended method referenced in DG-1389. Although the DG-1389 approach had not been issued when the FitzPatrick AST LAR was under review, the non-conservatisms associated with original AEB-98-03 report had been identified as early as 2005. Based on this, the DPO Appeal Panel concluded that the FitzPatrick AST LAR overestimated aerosol deposition in the main steam lines. This is a similar conclusion noted in the Directors Decision, though the sensitivity calculation of control room dose was underestimated in the Directors Decision (described further under SOC Issue 2.g).

An additional issue for the Fitzpatrick review was the credit taken by the licensee for fission product removal in the drywell due to containment sprays. The staff recognized the potential for drywell sprays to preferentially remove larger aerosol particles, another factor which would shift the aerosol size distribution toward smaller diameter particles and slower settling velocities. In ARCB-RAI-2, the staff specifically requested additional information describing how the gravitational settling credited in the main steam lines considers the changing aerosol characteristics (i.e., aerosol size and density distributions) due to the preferential removal of larger aerosols because of the credit assigned to containment sprays. In response, the licensee submitted several sensitivity cases, including use of different size distribution and adjustment to drywell spray removal rate based on aerosol particle size. In addressing the RAI, the licensee assumed deposition in all four main steam lines, an assumption that differs from its requested AST licensing basis (see SOC Issue 2.e for more discussion on this issue), as well as additional cases for a reduced operator breathing rate, 34

aerosol impaction on the MSIV, and condenser hold up and deposition. The results of these cases were summarized in the following table provided in the response to ARCB-RAI-2:

Note that the results presented for JAF-CALC-19-00005 include a reduction in the drywell spray removal coefficient described in the licensees response to ARCB-RAI-1 (ML20090E279) and model deposition in the shortest two main steam lines; the remaining cases include deposition in all four main steam lines. In the staff safety evaluation, the staff evaluated the licensees response to ARCB-RAI-2. With regard to use of the 20-group method and the revised aerosol size distribution, the staff stated that they did not review and evaluate this assumption because the licensee provided no basis for the assumption, it was not used in the analysis of record, and it was not used to determine reasonable assurance for complying with 10 CFR 50.67. However, the staff concluded in the safety evaluation that the licensees base case sensitivity indicates that the conservative modeling of the drywell spray on the aerosol removal in the MSLs without making other adjustments results in increased doses. The DPO Appeal Panel reviewed the aerosol distribution described in the response to ARCB-RAI-2 and compared it to the size distribution referenced in the AST LAR and the more recent draft guidance contained in DG-1389. The DPO Appeal Panel recognizes that while the draft guidance in DG-1389 was not formalized at the time the FitzPatrick LAR was under review, the guidance represents the staffs current best practices for aerosol modeling and the technical issues supporting the guidance had been known for several years prior to submittal of the FitzPatrick amendment. The main impact of using different distributions is the impact on settling velocities, with higher velocities resulting in increased fission product deposition in the main steam lines. The 35

results of this comparison are shown below and indicate that even when considering the impact of drywell spray on shifting the aerosol distribution toward smaller aerosol diameters, the settling velocities presented in the ARCB-RAI-2 response are still significantly greater than the velocities calculated using the DG-1389 numeric integration method:

Table 4: Settling Velocity Comparisons Settling Velocity (m/s)

Percentile AST LAR ARCB-RAI-2 ARCB-RAI-2 DG-1389 (Table 4, JAF- (No Spray) (With Spray - (Numeric CALC-19-00001) After Release) Integration) 10 3.70E-04 1.59E-05 6.19E-06 1.07E-06 50 1.28E-03 9.55E-05 2.73E-05 6.35E-06 90 3.04E-03 3.66E-04 6.25E-05 3.75E-05 The staff evaluated the other sensitivity cases provided by the licensee (i.e., control room operator breathing rate, aerosol impaction on the MSIV, and condenser holdup and deposition) and concluded the following in the staffs safety evaluation for the AST LAR:

• Breathing Rate: The staff disagreed with the licensees selection of the breathing rate, noting while the use of a breathing rate for light intensity work might be justified during time periods of normal working conditions, it should not be considered under 10 CFR 50.67 for determining radiation exposures from access to and occupancy of the [control room] under accident conditions.

• Aerosol Impaction on the MSIV: The staff noted that such credit is not generally endorsed and did not consider such credit appropriate for use in the FitzPatrick MSIV leakage sensitivity analysis.

• Condenser Holdup and Deposition: The staff noted that RG 1.183 permits such credit under certain conditions and that the staff considers it reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage.

The staff also evaluated the licensing precedents cited by FitzPatrick for use of the 20-group aerosol settling probabilistic distribution. The staff reached the following conclusions regarding the use of AEB-98-03 for the Clinton, Limerick, and LaSalle precedents cited in the FitzPatrick AST LAR:

• Clinton (AST amendment issued in 2005, ML052570475): The Clinton AST analysis assumed a main steam line break (which reduced the available settling volume in the broken main steam line), calculated a weighted average for settling velocity from the AEB-98-03 distribution, assumed MSIV leakage (and deposition) only for the two shortest of the four main steam lines, and no credit was taken for aerosol settling after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. The staff performed a sensitivity study using the AEB-98-03 10th percentile settling velocity, and confirmed that with this more conservative value, the total dose from all pathways would remain acceptable.

36

• Limerick (AST amendment issued in 2006, ML062210207): The Limerick AST analysis assumed a main steam line break (which reduced the available settling volume in the broken main steam line), assumed MSIV leakage (and deposition) only for the two shortest of the four main steam lines, credited the condenser for deposition, and no credit was taken for aerosol settling after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> (ML053330355). Additionally, a 20-group probability distribution on settling velocities was used to determine removal efficiencies for each group and a net weighted average efficiency used for main steam line deposition removal. Similar to the approach used for Clinton, the staff performed a sensitivity study using the AEB 03 10th percentile settling velocity, and confirmed that with this more conservative value, the total dose from all pathways would remain acceptable.

• LaSalle (AST amendment issued in 2010, ML101750625): The LaSalle AST analysis assumed a main steam line break (which reduced the available settling volume in the broken main steam line), assumed MSIV leakage (and deposition) only for the two shortest of the four main steam lines, credited the condenser for deposition, and no credit was taken for aerosol settling after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Further, no reduction in MSIV leakage after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> was assumed in the analysis for conservatism. Additionally, a 20-group probability distribution on settling velocities was used to determine removal efficiencies for each group and a net weighted average efficiency used for main steam line deposition removal. In the safety evaluation, the staff specifically evaluated the alternate leakage treatment pathway to the condenser in the safety evaluation.

Clinton, Limerick, and LaSalle all assumed that a portion of the fission products released into containment would plate out in the drywell due to natural deposition processes. This was modeled using the 10th-percentile values in the model described in NUREG/CR-6189, A Simplified Model of Aerosol Removal by Natural Processes in Reactor Containments (ML100130305, July 1996). However, no credit for drywell spray as a fission product removal mechanism was credited in these AST amendments.

In the FitzPatrick safety evaluation, the staff noted that the licensee credited fission product removal by the drywell sprays but recognized that [b]ecause the sprays shift the size of the aerosols to smaller sizes, the aerosols settling in the steam lines would decrease due to these smaller diameter aerosols. The staff then described an earlier licensing precedent for the Nine Mile Point Nuclear Station AST LAR (ML071580314) that also considered fission product removal by the drywell spray in addition to deposition in the main steam lines. In the case of Nine Mile Point, a penalty of the aerosol settling velocity was applied by the licensee to account for the sprays preferentially removing larger aerosol particles in primary containment drywell. Specifically, the licensee used a settling velocity of 6.6 x 10-5 m/sec to address the staffs issues regarding the use of AEB-98-03 (ML20149K682) in conjunction with the fission product removal credit provided by drywell spays. The licensee noted that the settling velocity value of 6.6 x 10-5 m/sec represents the 3rd percentile of the AEB-98-03 distribution.12 In the staffs safety evaluation (ML081230439), the staff determined that the AEB-98-03 3rd percentile settling velocity is sufficiently conservative to reflect the 12 A later staff safety evaluation associated with approving an increase in main steam line leakage limits notes that the settling velocity of 6.6 x 10-5 m/s corresponds to one half of the third percentile value (see ML20241A190). However, the staffs safety evaluation initially approving the Nine Mile Point AST amendment noted this as the 3rd percentile value.

37

effectiveness of drywell spray activity removal in containment upstream of this pathway. It should be noted that the 3rd percentile settling velocity based on the initial AEB-98-03 is approximately equivalent to approximately the 90th percentile in the 2006 update to the AEB-98-03 distribution (which has a mean of 2.01 x 10-5 m/s). In addition, Nine Mile Point limited the spray fission product removal coefficient to 19.8 hr-1 (compared to the 26.35 hr-1 credited for FitzPatrick). Nine Mile Point credited natural deposition in the main steam lines only in pipe segments between closed containment isolation valves. All four steam lines were assumed to leak consistent with technical specification limits (i.e., total of 96 scfh with each steam line limited to 24 scfh). No deposition credit was applied to the main steam line with one MSIV assumed to be stuck open and no deposition credit was applied downstream of the outboard MSIVs. Further, the licensee did not credit any increase in deposition effectiveness due to lower main steam leak rates after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> as a result of an assumed decrease in containment pressure.

The DPO Appeal Panel noted that the licensing precedents referenced in the FitzPatrick safety evaluation included several conservatisms not present in the FitzPatrick licensing basis, including: (1) limiting the time that deposition was credited; (2) limiting the area available for main steam line deposition; (3) using a conservative settling velocity to reflect the impact of drywell sprays; and (4) use of conservative staff sensitivity studies.

To gain further insights into how the AST LAR review for FitzPatrick had been conducted, the DPO Appeal Panel interviewed the lead reviewers for PRA and dose consequences.

Both the PRA and dose consequence reviewers noted that the initial DPO Panel did not interview them. The dose consequence reviewer highlighted the high degree of uncertainty associated with crediting drywell spray in conjunction with main steam line deposition. This reviewer noted that assumptions regarding containment pressure represent a large conservatism because safety related pressure control systems are capable of reducing containment pressure below the design pressure and that calculated offsite doses for FitzPatrick were well below the regulatory limits. However, the dose consequence reviewer noted that the calculation of control room dose can be a limiting factor. The reviewer acknowledged that in asking the licensee to address the aerosol size question presented in ARCB-RAI-2, the staff did not have a complete understanding about the phenomena associated with aerosol dynamics when both spray and deposition are credited. The reviewer stated that the use of risk insights associated with deposition in the main condenser (a capability that had been credited in other LARs) compensated for the uncertainty with sprays and deposition. Further, the reviewer indicated that the risk insights for the main condenser were essential for approving the FitzPatrick AST LAR. The lead PRA analyst noted that the risk insights examined included not only deposition in the main condenser, but other factors such as use of a conservative containment pressure (which would maximize leakage), no credit for suppression pool scrubbing, and other deposition mechanisms (e.g., deposition in the reactor vessel). However, the lead PRA analyst also acknowledged the uncertainties associated with crediting drywell spray and steam line deposition. Finally, the PRA analyst noted that their Branch Chief decided to only include risk insights related to condenser holdup in the FitzPatrick safety evaluation. Both the lead PRA analyst and the lead dose consequence reviewers noted that logic similar to what was used for FitzPatrick was used in approving three other contemporaneous license amendments associated with an increase in allowable main steam line leakage. Finally, the lead consequence analysis reviewer noted that, consistent with the uncertainties associated with combined credit for drywell sprays and main steam line deposition, DG-1389, Appendix A, Section A-5, states that the recommended methods for computing aerosol deposition are not valid if credit has been taken for drywell spray. In addition, DG-1389, 38

Appendix A, Section A-5.4, states that proposed main steam line leak rates in excess of 400 scfh would only be considered on a case-by-case basis with sufficient justification.

The DPO Appeal Panel reviewed the other three LARs and notes that sensitivity studies similar to what was done for FitzPatrick were performed for each. Of note, the sensitivity studies credited deposition in all four main steam lines, even if the licensing basis assumed deposition in fewer than four lines. The licensee for each amendment stated that an initial aerosol distribution of 2 AMMD and a geometric standard deviation of 2 were used for each case; however, the staff did not perform a detailed review of this modeling. For all but Nine Mile Point (which has a much lower contribution to calculated control room dose than the other two plants), the base sensitivity study indicates that the 5 rem TEDE criteria could be exceeded when using an aerosol model more consistent with updated methodologies that were available at the time of the staff reviews (e.g., use of an initial aerosol distribution more consistent with NEA/CSNI/R(2009)5).

Table 5: Sensitivity Results for Contemporaneous Licensing Reviews Calculated Calculated Control Room Control Room Nuclear Safety Associated Dose Dose (base Plant Evaluation RAI (licensing sensitivity basis) case)

(rem TEDE) (rem TEDE)

Nine Mile ML20241A190 ML20135G951 2.27 2.31 Point Quad Cities ML20150A328 ML20091H576 3.66 6.47 Dresden ML20265A240 ML20127H890 4.86 5.80 Each of these sensitivity studies credits both drywell spray and main steam line deposition, credit which would not be provided under the proposed guidance in DG-1389. Therefore, if only spray or main steam line deposition were to be credited, it would be expected that the calculated control room dose would further increase.

Conclusions for Issue 2.f The DPO Appeal Panel agrees with the submitters that the removal coefficients for aerosol settling used in the FitzPatrick AST LAR are nonconservative and do not reflect the state of knowledge that has been developed since issuance of AEB-98-03. Furthermore, these non-conservatisms do not appear to be offset by other factors that have been applied in the past, such as limiting available deposition area, using less than the full calculated credit for main steam line deposition, or use of more limiting settling velocities than provided by AEB-98-03.

With regard to providing credit for both spray and main steam line deposition, the DPO Appeal Panel notes that despite the differences in main steam line deposition modeling recommended by SAND2008-6601 and Metcalf and Perez, both sources agree that crediting both drywell sprays and main steam line deposition as mechanisms for fission product removal during the initial phase of the DBA is not appropriate. An additional complicating factor is that the FitzPatrick AST LAR did not assume a main steam line break as the design basis LOCA (described further under SOC Issue 2.e). The lack of an assumed steam line break makes the assumption that fission products in the drywell would have a direct pathway to the main steam lines early in the accident unlikely. In addition, the DPO 39

Appeal Panel concluded that the use of an aerosol settling velocity based on AEB-98-03 overestimates main steam line deposition.

The DPO Appeal Panel also recognizes that, despite the NRC being aware of non-conservatisms in AEB-98-03 since at least 2005, and other issues such as inappropriate crediting spray simultaneously with main steam line deposition, the NRC has not issued timely updates, revised guidance, or generic communications to licensees to address these issues. For the FitzPatrick case, while the staff had appeared to ask appropriate questions during the review; they did not follow through in addressing the potentially significant non-conservatisms used by the licensee prior to issuing the license amendment. Lastly, the DPO Appeal Panel notes that similar issues appear to be associated with the recent reviews conducted for Dresden, Nine Mile Point, and Quad Cities.

g. SOC Issue 2.g - Elemental Iodine Removal Constant Issue 2.g: By using an elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis, the analysis resulted in elemental removal greater than the SRP Section 6.5.2, Containment Spray as a Fission Product Cleanup System, Revision 4, dated March 2007 (ML070190178) limit on removal. Although this deviation was discussed in the NRC safety evaluation, it did not assess the aggregate impacts of the deficiencies in the analysis described above. (RG 1.183, Appendix A, Section 3.3, and SRP Section 6.5.2).

Response to Issue 2.g: The DPO Appeal Panel agrees with the submitters that the use of an elemental iodine removal constant greater than 20 hr-1 is not consistent with guidance documents and was not adequately justified in the staffs safety evaluation. The initial DPO Panel also agreed with the submitters that the licensee did not justify using an elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis and that the the analysis of record does not conform to RG 1.183 and SRP 6.5.2. With regard to this issue, the Directors Decision memorandum states the following:

I agree with the submitters and the panel that the particulate removal coefficient values FitzPatrick used for the drywell in their analysis may not have been the most conservative. However, the NRC staff reviewed and approved these values and concluded that the regulatory requirements were met, despite the licensees use of non-conservative removal coefficients. Our independent analysis, using particulate removal coefficient values in the drywell as low as 20 hr-1 and correcting for the other non-conservatisms identified in the memorandum, resulted in calculated control room doses that exceed the regulatory limit (5.11 rem). However, considering the substantial conservatisms also noted in the memorandum, we have reasonable assurance the dose will remain below 5 rem.

RG 1.183, Appendix A, Section 3.2, states the following:

Reduction in airborne radioactivity in the containment by natural deposition within the containment may be credited. Acceptable models for removal of iodine and aerosols are described in Chapter 6.5.2, Containment Spray as a Fission Product Cleanup System, of the Standard Review Plan (SRP),

NUREG-0800.

40

NUREG-0800, Chapter 6.5.2, Revision 4, provides a formula for calculating the elemental iodine spray removal coefficient based on several parameters including the fall time of spray drops, the spray flow rate, the mass-mean diameter of spray drops, the gas-phase mass transfer coefficient, and the free volume of the containment. In addition, SRP 6.5.2 states that s must be limited to 20 hr-1 to prevent extrapolation beyond the existing data for boric acid solutions with a pH of 5. In its license amendment request (i.e., Section 7.2 of JAF-CALC-19-00005), the licensee stated that the elemental iodine removal coefficient is conservatively assumed to be the same as the particulate aerosol removal coefficient. The licensee used an initial particulate iodine removal coefficient of 26.36 hr-1 as described in its response to ARCB-RAI-1B. By extension of the assumption stated in the LAR, the elemental iodine removal coefficient would have been assumed to be equivalent to this value. In its safety evaluation, the staff acknowledged that the elemental iodine removal coefficient was greater than 20 hr-1 (i.e., 26.36 hr-1), contrary to SRP 6.5.2. However, rather than justifying the departure from the SRP guidance on its technical merits, the staff only stated that the impact [of using an elemental iodine removal coefficient great than 20 hr-1] was not significant for this case. The staff conclusion was based on independent RADTRAD calculations using a particulate and elemental initial iodine removal coefficient of 20 hr-1 (vice 26.36 hr-1) and modeling main steam line deposition using an approach that accounts for known non-conservatisms of the AEB-98-03 method (i.e., the re-evaluated AEB-98-03 with multi-group method described in DG-1389). The DPO Appeal Panel notes that, although the multi-group method provides a more accurate estimate of aerosol deposition, its use is not required under the approved FitzPatrick licensing basis. The DPO Appeal Panel reviewed the results of NRRs RADTRAD calculations and noted that the 5.11 rem results referenced in the Directors Decision was based on assuming that all four main steam lines were available for radionuclide deposition rather than just two steam lines as described in the licensing basis for the AST amendment.

The DPO Appeal Panel requested RADTRAD results for a two steam line deposition model to compare with the results referenced in the Directors Decision. The RADTRAD analyst who previously prepared the calculations referenced by the NRR Office Director provided these additional results, but in performing this additional sensitivity analysis for the DPO Appeal Panel, immediately identified a non-conservative error in the previous multi-group calculation results.13 This error increased the 5.11 rem value referenced in the Directors Decision to 7.49 rem (a value approximately 50% greater than the 5 rem control room design criterion). The analyst provided the following results for deposition (calculated using the re-evaluated AEB-98-03 with multi-group method) in two and four main steam lines with particulate and elemental iodine removal coefficients of 26.36 hr-1 and 20 hr-1:

13 The DPO Appeal Panel notes that this error was immediately discovered by the analyst when running the additional sensitivity case for deposition in two steam lines rather than four steam lines. The analyst noted that they had prepared the two steam line case previously for their analysis for the Directors Decision, but NRR staff did not request that this additional sensitivity run be performed. The DPO Appeal Panel believes that had NRR staff that were developing the Directors Decision previously requested a full range of sensitivity studies, which the analyst could have readily performed, this error would have been discovered earlier.

41

Table 6: Calculated Control Room Dose Sensitivity to Iodine Removal Coefficient Number of Main Steam Deposition Iodine Removal NRR Steam Lines Model Coefficient Calculated Credited for Total Deposition Control Room Dose (rem TEDE) 4 Re-evaluated AEB-98-03 20 hr-1 7.49 with multi-group 2 Re-evaluated AEB-98-03 20 hr-1 10.42 with multi-group 4 Re-evaluated AEB-98-03 26.36 hr-1 6.45 with multi-group 2 Re-evaluated AEB-98-03 26.36 hr-1 8.99 with multi-group 4 Modified AEB-98-031 20 hr-1 3.77 2 Modified AEB-98-031 20 hr-1 Not calculated 4 Modified AEB-98-031 26.36 hr-1 3.71 2 Modified AEB-98-031 26.36 hr-1 4.73 1

Consistent with FitzPatrick AST Licensing Basis The DPO Appeal Panel notes that, given these values, the impact of changing the iodine removal coefficient is more pronounced for the updated multi-group method, which provides improved modeling of aerosol settling and is based on a more representative size distribution. The impact of changing the removal coefficient is much less significant for the modified AEB-98-03 approach, which is currently included in the FitzPatrick AST licensing basis. Further, these sensitivities are based on adjusting both the particulate and elemental iodine removal coefficients, rather than adjusting only the elemental iodine removal coefficient. As noted in the initial DPO Panel report, an additional potential non-conservatism was how the spray fall height for the particulate iodine removal coefficient was calculated. The licensee averaged the distance between the upper and lower spray headers and the floor of the drywell to obtain an average fall height of 13.07 meters, which was then reduced by a factor of one-third to correct for potential obstructions. The initial DPO Panel noted that the average fall height should be based on a flow weighted average of the fall height, which for other similar plants further reduced the average fall height by a factor of approximately one-third. However, the DPO Appeal Panel did not have sufficient information on the flow rates for the upper and lower drywell spray headers to assess the impact of this difference and therefore was unable to determine if applying a removal coefficient of 20 hr-1 for particulate iodine was conservative or non-conservative.

The conclusion documented in the Directors Decision that assumed that a more realistic analysis using a multigroup method for calculating deposition in the main steam lines resulted in a control room value only slightly above the 5 rem TEDE criterion (i.e., 5.11 rem TEDE) does not appear to be substantiated. Further, while the corrected value for this case, which assumed deposition in all four main steam lines, is approximately 7.5 rem TEDE, the calculated value for the two steam line deposition case (consistent with the AST licensing basis) is 10.4 rem TEDE. Use of the 26.36 hr-1 iodine removal coefficient reduces the 42

calculated control room dose to approximately 9 rem TEDE for the two steam line case, but this is still significantly greater than the 5 rem TEDE acceptance criteria.

The Directors Decision identified significant conservatisms in the licensees approved calculation of record, including the use of only two of four steam lines for deposition, crediting only half the reactor building volume for mixing, not crediting iodine and particulate deposition in the reactor vessel, and not considering suppression pool fission product scrubbing and retention. The DPO Appeal Panel reviewed each of these identified conservatisms and determined the following:

• Use of only two of four steam lines for deposition: As discussed under the disposition of SOC Issue 2.e., crediting only two steam lines for radionuclide deposition is an assumption that is consistent with both the approved technical specification for MSIV leakage and the AST licensing basis. Crediting deposition in all four main steam lines would be non-conservative and inconsistent with the currently approved TS which would allow continued operation if the total main steam line leak rate were distributed among only two steam lines. Further, the licensee clearly stated that their calculation of record supporting the AST LAR was based on deposition only in the two shortest main steam lines.

• Crediting only half of the reactor building volume for mixing: RG 1.183, Appendix A, Section 4.4 states that:

Credit for dilution in the secondary containment may be allowed when adequate means to cause mixing can be demonstrated. Otherwise, the leakage from the primary containment should be assumed to be transported directly to exhaust systems without mixing. Credit for mixing, if found to be appropriate, should generally be limited to 50%. This evaluation should consider the magnitude of the containment leakage in relation to contiguous building volume or exhaust rate, the location of exhaust plenums relative to projected release locations, the recirculation ventilation systems, and internal walls and floors that impede stream flow between the release and the exhaust.

The DPO Appeal Panel noted that the staff safety evaluation credited 50% of the reactor building for mixing, which effectively dilutes the primary containment and ESF leakage sources into the secondary containment. However, in their license amendment request, the licensee only stated that the small amount of containment leakage into the secondary containment would diffuse through the secondary containment prior to being exhausted by the SGTS to the environment and that significant mixing would occur. In Section 3.11.6, Reactor Building, of the LAR, the licensee noted that 1,200 cfm is drawn from each of five elevations below the 369 foot level (i.e., elevations of 262 feet; 280 feet, 6 inches; 307 feet; 334 feet; and 351 feet, 6 inches) and 5000 cfm is drawn from above elevation of 369 feet. This sums to a SGTS flowrate of approximately 11,000 cfm, which is consistent with two trains of SGTS operation. However, as noted in Table 3.11-1 of the LAR, the assumed SGTS flow is 6,000 cfm +10% (i.e., one train of SGTS). At a SGTS flow rate of 6,000 cfm, it would take more than 6.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> to exhaust a volume of air equivalent to a reactor building free volume of 2,370,000 cubic feet. Additionally, if the SGTS draws air from locations in close proximity to ESF or containment leakage, it would be expected that there would be reduced opportunities for mixing and dilution. The DPO Appeal Panel 43

further notes that the FitzPatrick Technical Specifications does not include any other ventilation systems that provide capabilities for recirculating or mixing the secondary containment free air volume. The licensee did not provide any additional information regarding the location of exhaust plenums, internal walls or floors that could impede stream flow, or the presence of recirculation ventilation systems. Further, the DPO Appeal Panel could locate no information in the staffs safety evaluation that assessed and justified the licensees basis for assuming 50% mixing. In reviewing several plant layout drawings contained in Chapter 12.3 of the UFSAR, the DPO Appeal Panel noted numerous floors, walls, and rooms that could potentially limit the amount of secondary containment volume available for mixing. Therefore, in the absence of justification to substantiate crediting even half the reactor building free volume for mixing, the DPO Appeal Panel could not conclude that this was a conservative assumption.

• Not crediting iodine and particulate deposition in the reactor vessel: RG 1.183, Appendix A, Section 6.1, states that [n]o credit should be assumed for activity reduction by the steam separators or by iodine partitioning in the reactor vessel.

This guidance is consistent with the Commissions intent for the AST DBA to include the first three severe accident radionuclide release phases (coolant, gap, and early in-vessel) as described in the preamble to the 10 CFR 50.67 final rule (64 FRN 71990). As noted in NUREG-1465, Section 3.3, during the early in-vessel release phase, the fuel as well as other structural materials in the core reach sufficiently high temperatures that the reactor core geometry is no longer maintained and fuel and other materials melt and relocate to the bottom of the reactor pressure vessel.

Given the characteristics of the early in-vessel release phase (including high temperatures within the reactor vessel which may limit total deposition within vessel),

the mechanisms by which crediting fission product deposition on internal vessel structures during the assumed DBA was not clear to the DPO Appeal Panel.

Therefore, the assumption that there is no deposition within the reactor vessel, as included in the current licensing basis for the FitzPatrick AST amendment, appears to be consistent with the Commissions intent of addressing radioactive release through the early in-vessel release phase during a design basis accident. In addition, during the in-vessel release phase, the reactor vessel would experience high temperatures that would limit possible deposition and reflood of the core following in-vessel release would lead to high steam flows and mixing leading to transport of deposited fission products from internal vessel structures, if it were to occur. As such, the DPO Appeal Panel does not believe this is a conservatism, but instead a consequence of the design basis accident assumptions prescribed for the AST analysis.

• Not considering suppression pool fission product scrubbing and retention: RG 1.183, Appendix A, Section 3.5, states that [r]eduction in airborne radioactivity in the containment by suppression pool scrubbing in BWRs should generally not be credited. The section also states that such credit may be provided on an individual case basis and that the evaluation should consider the relative timing of the blowdown and the fission product release from the fuel, the force driving the release through the pool, and the potential for any bypass of the suppression pool. In Attachment 1, Section 3.11.5, of its license amendment application, FitzPatrick stated that [i]odine removal by Suppression Pool scrubbing is not credited because the bulk core activity is released to containment well after the initial mass and energy release. However, consistent with RG 1.183, Appendix A, Sections 5.1 and 5.3 44

regarding modeling of ESF system leakage, all fission products released from the fuel to the containment were assumed to instantaneously and homogeneously mix in the suppression pool water at the time of release from the core. An assessment of the capability to maintain suppression ph greater than 7.0 to prevent revolution of iodine from the suppression pool water was provided in the LAR and found acceptable in the safety evaluation.

Based on the above, the DPO Appeal Panel concluded that these issues did not represent significant conservatisms, but instead were consistent with the functional capability defined by technical specifications, the modeling of an early in-vessel release source term, or not otherwise justified in the LAR. As such, the DPO Appeal Panel was unable to conclude that these factors would offset the calculated control room dose in excess of 5 rem TEDE.

3. SOC Issue 3 - Conflicting Statements in the Directors Decision Issue 3: Evaluate potentially conflicting statements between the Directors Decision and the NRC staffs safety evaluation on crediting portions of the power conversion system condenser and determine if any mitigation function assigned to the condenser and associated equipment is appropriately documented in the licensing basis.

Response to Issue 3: The DPO Appeal Panel concluded that crediting the power conversion system for aerosol deposition was necessary to address known non-conservatisms in the analysis supporting the FitzPatrick AST LAR. In the safety evaluation, the DPO Appeal Panel determined that the NRC staff implicitly credited aerosol deposition in the condenser; for example, the safety evaluation notes that it is reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage. Therefore, the DPO Appeal Panel does not agree with the conclusions stated in the Directors Decision stating the power conversion system (e.g., condenser) was not credited by the staff in order to make a regulatory decision on the FitzPatrick AST LAR. Further, because the licensee did not request credit for fission product holdup and deposition in the power conversion system nor did it include this credit in its calculation of record, no regulatory controls for the condenser aerosol deposition function were identified.

The DPO Appeal Panels conclusion on this issue is consistent with the determination of the initial DPO Panel. For example, that panel noted the following:

The panel agrees with the submitters that the staff approval of the FitzPatrick alternative source term application is based on the staff perspective that the power conversion system (i.e., condenser pathway) would be available as a holdup volume to achieve acceptable control room doses (i.e., results do not meet the regulatory limits without this credit) when the staff-identified issue identified above associated with aerosol removal and drywell spray operations is properly addressed and that this credit to achieve acceptable control room doses is not incorporated into the licensees analysis of record nor are any programmatic elements included that would assure the availability and reliability of this pathway when needed.

The Directors Decision did not agree with the perspectives of the initial DPO Panel. Specifically, the Directors Decision noted the following:

45

In the SE, the NRC staff did not credit the PCS or the main condenser to make the regulatory finding of adequate protection. Section 3.5 of the SE titled NRC Staff Risk and Engineering Insights, explains that, by Commission direction in the SRM to SECY-19-0036, the staff should apply risk-informed principles in any licensing review or other regulatory decision when strict, prescriptive application of deterministic criteria is unnecessary to provide for reasonable assurance of adequate protection of public health and safety. In Section 3.5 of the SE, the NRC staff concluded that these risk and engineering insights support its reasonable assurance finding based on its deterministic review. The NRC staff performed an independent assessment of the seismic robustness of the condenser which was referred to in the SE for FitzPatrick but not made publicly available. While this was a different approach from previous applications, in which a licensee credits the condenser through referencing a Boiling Water Reactor Owners Group (BWROG) topical report this assessment gave the NRC staff high confidence the PCS and the main condenser would remain available for fission product dilution, holdup, and retention when exposed to seismic accelerations corresponding to FitzPatricks design basis SSE.

The DPO Appeal Panel does not believe that the reference to SRM-SECY-19-0036 is relevant in the context of the decision regarding the FitzPatrick AST LAR. This is not a question of whether or not the power conversion system can be credited for aerosol holdup and deposition; it is instead a question of the staffs approval basis of the license amendment and whether or not the assumptions and factors used in making that decision are appropriately controlled within the licensing basis. With regard to aerosol modeling, associated uncertainties, and the role of the power conversion system, the staffs safety evaluation stated the following:

The NRC staff notes that the guidance provided in RG 1.183 for design-basis LOCA radiological analysis states that structures, systems, and components (SSCs) may be credited with creating a pathway to the condenser if they are able to withstand an SSE. However, the staff also considers it reasonable to include the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage. The NRC staffs consideration of risk and engineering insights is discussed in Section 3.5 of this SE.

Section 3.5 of the staffs safety evaluation further notes:

The staffs independent assessment found it is reasonable to conclude that the SSCs in the PCS would be available following an SSE and that the likelihood of them being unavailable to serve as a volume for holdup and retention is very low.

The assessment provides an insight when addressing uncertainties in the calculation of the dose consequences of MSIV leakage. Specifically, the staff recognizes that there is a high probability that doses will be significantly lower than those estimated using deterministic methods that do not credit holdup and retention of the MSIV leakage within the PCS.

The DPO Appeal Panel determined that the uncertainties in crediting aerosol removal from drywell sprays referenced in the staffs safety evaluation recognizes that the methods and approach used in the FitzPatrick analysis of record overestimates fission product removal. As such, the calculation of record underestimates the control room dose compared to approaches that more accurately model fission product removal mechanisms. Therefore, based on 46

information included in the safety evaluation and gained through interviews with reviewers for the AST LAR, the credit for condenser holdup and deposition was needed to offset non-conservatism in the licensees calculation of record. However, the DPO Appeal Panel also acknowledges that the licensee did not request approval for fission product holdup and deposition in the power conversion system and the staffs safety evaluation did clearly describe the credit provided by the condenser or place any additional conditions or requirements on use of the condenser pathway for handling MSIV leakage. The DPO Appeal Panel notes that other licensees that have requested and been approved to credit holdup and deposition of fission products in the condenser have included configuration control of the release paths in licensee controlled documents such as Technical Requirements Manuals (e.g., LaSalle, ML22111A256) or Technical Specifications and have been subject to enforcement action when these paths have not been maintained in accordance with regulatory requirements (e.g., LaSalle, ML22222A091; and Duane Arnold, ML17123A087).

Lastly, the DPO Appeal Panel notes that consideration of uncertainties is inherent in all regulatory decisions. Within a deterministic framework, the conservative selection of inputs, use of bounding methods, and other assumptions intended to provide additional safety margins help to ensure that uncertainties are appropriately considered in the decision-making process. Within a risk-informed framework, some uncertainties can be more directly assessed and mitigated, but other uncertainties must be addressed more deterministically through the use of appropriate conservatisms. Regardless, the manner in which uncertainties are addressed does not obviate the need to include such considerations within the licensing basis. The DPO Appeal Panel believes that if uncertainties are great enough to significantly erode the level of confidence that a regulatory requirement is met, then the additional considerations that offset these uncertainties provide a basis for the decision and, as such, should be explicitly captured in the licensing basis. In the case of FitzPatrick, the DPO Appeal Panel believes that the availability of the power conversion system for fission product deposition is a key assumption that was necessary to demonstrate that the FitzPatrick AST LAR could meet the radiological acceptance criteria for the control room. Therefore, credit for fission product deposition in the condenser should have been included in the licensing basis with appropriate configuration controls. The DPO Appeal Panel considers this to be an error in the staffs review and approval basis of the FitzPatrick AST LAR.

4. SOC Issue 4 - Approval Basis for the FitzPatrick AST Amendment Issue 4: Determine if the approval basis of the FitzPatrick AST amendment was clearly documented in the licensing basis (e.g., calculations of record, updated final safety analysis report updates, SER, and whether they appropriately reflect the assumptions and mitigating equipment credited in the NRC staffs analysis).

Response to Issue 4: The DPO Appeal Panel carefully considered the issues raised by the submitters and concluded that the approval basis for the FitzPatrick AST LAR was not clearly documented in the licensing basis. Specific issues considered by the DPO Appeal Panel include the following:

• A licensing basis function identified for the MSIV Leakage Control System was collecting packing leakage for the outboard MSIV. Although the staff approved removal of the MSIV Leakage Control System when issuing the FitzPatrick AST amendment, no evaluation of the impact of removal of this function was included in the safety evaluation.

47

Further, the outboard MSIV packing continues to be identified in UFSAR as a leakage path to the environment. The DPO Appeal Panel found that MSIV packing leakage is a credible failure mechanism based on operating experience. This is discussed further under the assessment of SOC Issue 1.

• The FitzPatrick AST LAR assumed that the limiting break location was the recirculation line rather than a main steam line. The impact of this assumption is that a larger portion of the main steam line remained available to be credited for fission product deposition and removal. Although the staff requested additional information on the licensees basis for the limiting break location, the basis for the staffs approval of the recirculation line break was not clearly stated in the safety evaluation. Specifically, the safety evaluation noted that the licensees sensitivity analysis considered four main steam lines, rather than two, for fission product deposition (a configuration less limiting than the requested licensing basis), but the staff accepted the recirculation break location by noting that the impact of including an MSL break does not significantly impact the dose consequences from MSIV leakage. The DPO Appeal Panel notes that additional sensitivity studies suggest that the impact of assuming a main steam line break would increase calculated control room dose by at least 0.9 rem TEDE.14 This is further discussed under the assessment of SOC Issue 2.e.

• The DPO Appeal Panel acknowledges that the AEB-98-03 approach for assessing aerosol deposition has been approved for many licensees over the last 25 years.

However, significant concerns associated with non-conservatisms inherent in the AEB-98-03 approach have been known since at least 2005. Further, in 2007, NRR requested, supplemental information [from RES on revised aerosol parameters] to perform a backfit analysis and to revise [its] regulatory positions in Regulatory Guide 1.183, (ML071620396, non-public). Despite the NRC gaining significant understanding of aerosol behavior through the issuance of SAND2008-6601 and NEA/CSNI/R(2009)5, including insights into the interaction of drywell sprays and main steam line deposition in the 2009 time frame, the staff approved use of AEB-98-03 for the FitzPatrick AST LAR without explaining how these known issues were addressed. This is further discussed under the assessment of SOC Issue 2.f.

• The DPO Appeal Panel found that the basis for approval of an elemental iodine removal coefficient greater than the 20 hr-1 guideline provided in SRP Section 6.5.2 was not clearly described in the safety evaluation. This is further discussed under the assessment of SOC Issue 2.g.

• The DPO Appeal Panel concluded that credit for fission product aerosol deposition in the power conversion system was necessary to offset known non-conservatisms and uncertainties in how aerosol modeling and deposition was addressed in the FitzPatrick AST LAR. Although the safety evaluation described that the staff determined that there is high confidence that the MSLs and the PCS will be available for fission product dilution, holdup, and retention, the safety function being credited for the power conversion system was not clearly captured in the licensing basis. Further, the AST 14 The DPO Appeal Panel notes that the sensitivity studies described in the assessment of SOC Issue 2.e were based on the AEB-98-03 aerosol model used by FitzPatrick for their AST LAR. If a more updated aerosol modeling approach were used, such as those recommended in draft DG-1389 (e.g., re-evaluated AEB-98-03 with multigroup or numerical integration) and that were available at the time of the staff review of the FitzPatrick AST LAR, the dose impact of break location may increase further.

48

amendment lacked any configuration controls for the leakage path to the condenser, unlike other AST LARs that credited this path. This is further discussed under the assessment of SOC Issue 3.

In the view of the DPO Appeal Panel, these issues taken collectively result in gaps between the approval basis used for the FitzPatrick AST LAR and documented licensing basis information contained in the licensees application, the safety evaluation, and the calculation of record.

These gaps represent an omission of design basis15 information for the facility. As noted by the Commission in its policy statement, Availability and Adequacy of Design Bases Information at Nuclear Power Plants (57 FR 35455; August 10, 1992):

The Commission has concluded that maintaining current and acceptable design documentation is important to ensure that (1) the plant physical and functional characteristics are maintained and are consistent with the design basis as required by NRC regulation, (2) systems, structures, and components can perform their intended functions, and (3) the plant is operated in a manner consistent with the design basis.

The Commission believes a licensee should be able to show that it has sufficient documentation, including calculations or pre-operational, startup or surveillance test data to conclude the current facility configuration is consistent with Its design bases. The Commission further believes the design basis must be understood and documented to support operability determination and 10 CFR 50.59 evaluations that may need to be made quickly in responding to plant events.

This perspective is reflected in SRP Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Revision 0 (ML003734190), which states the following:

The staffs approval of the application is to be based on the licensees docketed information. If differences are discovered between the licensees methods and assumptions and those deemed acceptable to the staff, the reviewer should resolve the differences with the licensee. If necessary, the licensee should update the disputed assumptions and resubmit the affected analyses.

It appeared to the DPO Appeal Panel that, in order to offset known non-conservatisms, it was necessary to credit fission product deposition in the main condenser in order to have reasonable assurance that the calculated control room dose was less than 5 rem TEDE. However, this credit was not requested by the licensee nor captured within the licensing basis. The DPO Appeal Panel also notes that the staff requested additional information on the limiting break location and containment spray impacts on aerosol settling, thereby indicating a belief that these issues were salient to the staffs review. Unfortunately, the staffs safety evaluation did not explain how the assessment of the licensees response to these issues, particularly information 15 10 CFR 50.2, Definitions, defines design bases as that information which identifies the specific functions to be performed by a structure, system, or component of a facility, and the specific values or ranges of values chosen for controlling parameters as reference bounds for design. These values may be (1) restraints derived from generally accepted state of the art practices for achieving functional goals, or (2) requirements derived from analysis (based on calculation and/or experiments) of the effects of a postulated accident for which a structure, system, or component must meet its functional goals.

related to factors that would increase the control room dose, was adjudicated during the AST LAR review.

The DPO Appeal Panel also considered the findings of the initial DPO Panel regarding the clarity of licensing and design basis issues. The initial DPO Panel recommended that the licensees updated calculation of record should be submitted to the NRC and that guidance be updated to ensure that expectations for requesting updated information during licensing reviews is clear. While the DPO Appeal Panel supports the objective of achieving clarity in staff guidance, the DPO Appeal Panel does not agree that there is a need for the licensee to submit an updated calculation of record to the NRC. Such submittals represent a snapshot in time and could be taken out of context if used to evaluate a current licensing issue. Under its NRC-approved Quality Assurance program, the licensee maintains documentation control measures and the licensees records provide the most up to date information.

5. Additional DPO Appeal Panel Observations

a. Dose Sensitivity Studies Provided by the Submitters The DPO Appeal Panel notes that the submitters provided dose estimates for several sensitivity cases for control room doses in their initial DPO submission (e.g., Table 3 of the original DPO submission (ML21042B867; non-public). Among the cases examined is one that considered the drywell boundary extending to the MSIV boundary, no credit for deposition between the reactor vessel and inboard MSIV, a more limiting release path (e.g.,

through the outboard MSIV packing with more limited deposition area in the main steam piping), and adjustments to the atmospheric dispersion factors for the calculated control room dose (e.g., associated with reduced distance form release point to control room ventilation intake). The submitters stated they calculated a control room dose value of 466 rem TEDE when addressing these issues, but the details of the calculation were not provided in the submittal. Independent calculations performed by an NRR dose consequence analyst similarly resulted in high control room doses when extremely conservative assumptions are made. The DPO Appeal Panel considered this information and notes the following:

• The DBA is a stylized evaluation intended to determine reference dose values to demonstrate the capability of certain engineered safety features and support siting determinations. These calculations are not intended to be a realistic assessment of expected doses to the control room operators or the public.

• Substantial margin exists for calculated offsite reference doses, therefore issues addressed under this DPO would not be expected to result in calculated reference doses above 25 rem TEDE for the public. In addition, other layers of defense in depth exist (notably emergency planning programs) to further mitigate potential doses to the public.

• The preamble for the final rule for use of alternative source terms for operating reactors (64 FR 71990), notes that the control room design criteria do not imply that control room doses above 5 rem TEDE would be an acceptable exposure during emergency conditions, or that other radiation protection standards of 10 CFR Part 20, Standards for Protection Against Radiation, including individual organ dose limits, might not apply.

Instead, the control room design criteria are provided only to assess the acceptability of 50

design provisions for protecting control room operators under postulated DBA conditions. Further, the final rule preamble notes that DBA conditions assumed in these analyses, although credible, generally do not represent actual accident sequences but are specified as conservative surrogates to create bounding conditions for assessing the acceptability of engineered safety features.

• More realistic analysis conducted for the State of the Art Reactor Consequences Analysis (e.g., NUREG-1935, State-of-the-Art Reactor Consequence Analyses (SOARCA) Report) and the Level 3 PRA project (SRM-SECY-11-0089, Options for Proceeding with Future Level 3 Probabilistic Risk Assessment (PRA) Activities) indicate that risk of severe accident doses to the public is well below the Commissions safety goals (51 FR 30028 (August 21, 1986)).

Therefore, the DPO Appeal Panel concluded that the dose sensitivity calculated by the submitters (i.e., 466 rem TEDE to control room operators) does not represent the expected dose to the control room operators during a design basis accident. Further, the potential control room dose increase associated with the issues raised by the submitters, if substantiated and modeled in a manner consistent with the recent state of practice, is unlikely to result in a control room dose of the magnitude projected by the submitters.

b. Application of Risk Insights to the Design Basis Accident The DPO Appeal Panel acknowledges that application of risk insights is a good practice whenever possible. However, application of risk information and insights in a coherent and consistent manner, which relies on representing the as-operated and as-designed plant as realistically as possible, can be challenging for situations that rely on inherently deterministic or stylized regulatory assumptions. For example, the DBA analysis imposes a fission product source term in the containment without a mechanistic explanation for the cause of core damage. As noted in RG 1.183, the design basis accidents were not intended to be actual event sequences, but rather, were intended to be surrogates to enable deterministic evaluation of the response of a facilitys engineered safety features. In assessing the capability of the condenser for fission product deposition, Section 3.1.1.4.4 of the safety evaluation states:

The NRC staff notes that, by design, an SSE would not result in any core damage. The design-basis radiological assessment of an MHA (referred to in guidance as a LOCA) deterministically imposes a fuel melt source term into the containment to test the ability of the plant to meet predetermined dose acceptance criteria. Since the SSE would not cause fuel damage, the exclusion of non-SSE qualified SSCs in the dose analysis implies that two independent extraordinary events could occur during the analysis period: an event resulting in substantial fuel melt followed by a significant unrelated seismic event. The probability of this sequence of events is very low.

While the DPO Appeal Panel agrees that the LOCA provides a reasonably bounding surrogate event for analysis, the DBA analysis is not intended to only address LOCA conditions. Other accident scenarios can lead to fission product release coincident with a loss of the condenser without requiring two independent extraordinary events. For example, based on a review of PRA results for the FitzPatrick plant available on the SPAR 51

Dashboard,16 the DPO Appeal Panel notes that a design basis seismic event (i.e., a 0.1 -

0.3 g seismic event) contributes approximately 6% to the combined internal event, high wind, and seismic PRA core damage frequency (CDF). Therefore, the occurrence of a design basis seismic event and core damage should not be considered as two independent extraordinary events. Higher intensity seismic events, such as a Bin 2 seismic event (0.3 -

0.5 g), a Bin 3 seismic event (0.5 - 1.0 g), and a Bin 4 seismic event (1.0 - 1.5 g) contribute 10%, 17%, and 3%, respectively, to the CDF total. Further, the loss of condenser heat sink initiating event contributes approximately 12% to the CDF total. Collectively, these initiators account for over 40% of the core damage frequency and can represent a challenge to the availability of the condenser. The DPO Appeal Panel believes that a more coherent risk-informed approach would address plant-specific core damage accident sequences and consider their impact on the ability to provide fission product deposition in the condenser.

c. DPO Review Process The Appeal Panel also notes that there were several challenges associated with the review and assessment of the submitters concerns. The timeline for the FitzPatrick AST amendment and associated DPO activities highlight some aspects of this challenge:

• August 8, 2019 - FitzPatrick AST LAR submitted (ML19220A043)

• July 21, 2020 - AST Amendment and Safety Evaluation issued (ML20140A070)

• January 19, 2021 - DPO-2021-001 Received by OE

• June 9, 2021 - DPO-2021-001 Panel Report Issued

• November 17, 2021 - DPO-2021-001 Directors Decision Issued

• January 7, 2022 - DPO-2021-001 Appeal Received by OE

• February 3, 2022 - NRR Office Director reviewed the appeal package and reaffirmed the initial Directors Decision

• July 2022 - DPO Appeal Review Panel formed

• November 17, 2022 - First round of DPO Appeal Panel Questions to NRR issued; NRR responses received December 15, 2022

• December 22, 2022 - Second round of DPO Appeal Panel Questions to NRR issued; NRR responses received Jan 20, 2023

• January 13, 2023 - Summary of Concerns finalized

• June 2023 - Provide the DPO Appeal Panel Report to OEDO This timeline represents almost two and a half years between the submittal of the initial DPO and resolution of the DPO Appeal process. Further, the DPO Appeal Panel discovered several technical issues of concern associated with the NRR Office Directors secondary assessment of the DPO Panel report, including a failure to recognize that outboard MSIV packing leakage had been initially included in FitzPatricks licensing basis and that there were discoverable errors in its consequence analysis that was used to support the Directors Decision. In some cases, the DPO Appeal Panel believes that the assessment supporting the Directors Decision considered a narrow range of assessment options by directing the independent consequence analysis reviewer to run only a limited set of sensitivity cases, rather than holistically assessing the facts developed by the initial DPO Panel. This is 16 The Standardized Plant Analysis Risk (SPAR) Dashboard provides a summary of key risk insights from the NRCs SPAR models. While the SPAR Dashboard may not have been fully implemented at the time of the FitzPatrick review, risk insights were available from the SPAR model and other sources such as the FitzPatrick Plant Information Risk eBook.

52

highlighted by several of the DPO Appeal Panels conclusions on technical issues that align with the initial DPO Panel findings.

The DPO Appeal Panel believes that the DPO process is an important program and that a significant strength for the NRC is establishment of the NRCs credibility as an objective and technically adept regulator. However, significant delays between the issuance of a licensing action and resolution of an associated DPO limit the ability of the agency to make timely and effective changes when needed and erodes the confidence of staff in its management to understand complex technical issues presented when differing technical views are presented to effectively adjudicate them. Therefore, the issues associated with this DPO Appeal Panel should be examined to ensure that the generic topical decisions associated with the DPO process are based on complete and accurate information and are rendered in a timely manner.

III. Conclusions

1. Technical Issues The DPO Appeal Panel reached the following conclusions as a result of this review:

• MSIV Packing Leakage - The DPO Appeal Panel substantiated the submitters concerns regarding the removal of the MSIV leakage control system function to control packing leakage from the outboard MSIV. The DPO Appeal Panel believes that this issue was an error in the staffs review of the FitzPatrick AST LAR. Further, packing leakage contributes to the main steam line total leak rate and vents directly into the steam tunnel area and to the environment. Despite control of MSIV packing leakage being identified as a licensing basis function prior to the AST LAR, the impact of the removal of the leakage control system on packing leakage was not described in the staffs safety evaluation. Furthermore, information contained in draft DRA-ISG-2021-01 indicating that packing leakage is a small fraction of MSIV leakage is not substantiated by a review of recent operating experience. Additionally, based on a review of LERs and Inspection Reports, it was not clear to the DPO Appeal Panel that licensees recognize MSIV packing leakage as a potential challenge to technical specification requirements.

• Release Assumptions and Deposition Between the Reactor Vessel and Inboard MSIV

- The DPO Appeal Panel did not substantiate the submitters concerns related to crediting the air space between the reactor vessel and the inboard MSIV. The DPO Appeal Panel believes that this issue was addressed appropriately during the review of the FitzPatrick AST LAR. While the DPO Appeal Panel notes that SAND2008-6601 includes recommendations to not credit fission product removal between the reactor vessel and inboard MSIV due to thermal hydraulic uncertainties, the DPO Appeal Panel notes that such credit is recognized in RG 1.183 and RIS 06-04. The approach taken by the licensee to address this issue, particularly in light of the stylized nature of the DBA analysis, appeared to be reasonable.

• Location of Release Point in the Turbine Building and Treatment of Nonsafety-Related SSCs - The DPO Appeal Panel did not substantiate the submitters concerns related to placement of the release point in the turbine building at the turbine stop valve location, a location which relies solely on safety-related and seismically qualified portions of the main steam system. The DPO Appeal Panel believes that this issue was addressed appropriately 53

during the review of the FitzPatrick AST LAR. The licensee based the dispersion parameters used to calculate control room dose on the shortest straight-line distance between the release point and the control room ventilation intake and did not credit holdup or deposition in the turbine building. Further, the DPO Appeal Panel determined that this was a bounding approach, and while the licensee did provide an assessment of alternate release points from the turbine building to the control room ventilation intake, these paths would extend the distance between the turbine stop valves and receptor; therefore, the configuration of nonsafety-related smoke ejectors, doors, and ventilation exhaust is not consequential to the analysis. The DPO Appeal Panel believes that nonsafety-related SSCs in the turbine building were addressed appropriately during the review of the FitzPatrick AST LAR.

• Limiting Break Location - As noted in the preamble to the alternate source term final rule (64 FR 71990, 71997), The DBA conditions assumed in these analyses, although credible, generally do not represent actual accident sequences but are specified as conservative surrogates to create bounding conditions for assessing the acceptability of engineered safety features. Contrary to this intention, the DPO Appeal Panel concluded that the selection of the recirculation line break does not represent a bounding condition. When assessing this issue, the staff failed to identify that assuming a main steam line break was a more limiting condition due, in part, to not ensuring that a sensitivity study reflected the licensing basis conditions for the amendment. The DPO Appeal Panel believes that this issue was an error in the staffs review of the FitzPatrick AST LAR. The main impact of a main steam line break is a reduction in volume available for deposition in the broken line; this impact would be expected to increase dose projections.

• Aerosol Modeling and Related Guidance - The DPO Appeal Panel noted the NRC has failed to adequately address known nonconservative issues associated with the AEB-98-03 approach since at least 2005. While AEB-98-03 had been used for earlier AST amendments, it was generally applied with conservatisms (e.g., limited deposition area, limited time frame for deposition, selection of conservative settling velocity, independent staff confirmatory calculations) to provide margin for known uncertainties. For the FitzPatrick review, AEB-98-03 was applied with limited application of these previous conservatisms and in conjunction with use of drywell spray, a configuration that has been known for well over a decade to overestimate deposition. However, given the conflicting regulatory history regarding the use of AEB-98-03, the DPO Appeal Panel could not definitely determine what guidance for aerosol deposition would have been appropriate for an AST LAR review.

The DPO Appeal Panel determined that the modeling of aerosol deposition for the FitzPatrick AST overpredicts fission product removal when compared with models based on data that has been available since at least 2009, including information contained in NEA/CSNI/R(2009)5 and SAND2008-6601. Given the limited margin to the control room dose criteria, it is likely that if the known non-conservatisms of AEB-98-03 were addressed, calculated control room dose could exceed the 5 rem TEDE criteria unless additional mitigation credit (e.g., condenser deposition) were applied. The DPO Appeal Panel believes that continued use of the original AEB-98-03 guidance, which is known to be incorrect, is inappropriate and will result in inaccurate dose calculations. However, the DPO Appeal Panel concluded, due to the conflicting regulatory history and lack of clear guidance available to licensees, the use of the original AEB-98-03 approach, by itself, does not 54

represent a definitive error17 in the AST LAR review. However, it is also noteworthy that the NRC has not issued updated guidance or generic communications to alert licensees of these known issues. The DPO Appeal Panel believes it is important that the NRC clarify, through the use of an appropriate regulatory mechanism, the applicability of older guidance that is known to include mistakes, be outdated, and/or be nonconservative.

• Credit for Fission Product Deposition in the Power Conversion System - The DPO Appeal Panel determined that credit for fission product deposition in the power conversion system (e.g., main condenser) was necessary to offset known non-conservatisms in the licensees AST analysis. Specific issues include credit for both drywell sprays and main steam line deposition and use of the AEB-98-03 approach for determining aerosol settling velocity, which has been known to be nonconservative since at least 2005. However, credit for fission product deposition in the main condenser was not requested by the licensee and the staff did not impose any additional controls on the configuration of the path to the condenser similar to what has been done for other licensing actions. The DPO Appeal Panel concluded that the lack of appropriate regulatory documentation and justification for assumed fission product deposition in the condenser was an error in the staffs review of the FitzPatrick AST LAR. The DPO Appeal Panel notes that a stated objective of the DBA analysis is to enable deterministic evaluation of the response of the plant engineered safety features (64 FR 71990, 71991). Although the NRC has provided credit for fission product deposition in the nonsafety-related condenser, providing such credit may indicate that the plant engineered safety features are insufficient, on their own, to mitigate the design basis accident. Therefore, the DPO Appeal Panel believes that additional emphasis should be placed on the accident mitigation provided by the plants engineered safety features rather than on nonsafety-related SSCs. An example of such an engineered safety feature not routinely credited is the containment pressure suppression function provided by safety-related spray systems.

• NRC Guidance - The DPO Appeal Panel believes the draft guidance provided in DG-1389 is a significant enhancement to the guidance previously available in Revision 0 to RG 1.183 and RIS 06-04. However, it is also important that the NRC clarify, through the use of an appropriate regulatory mechanism, the applicability of older guidance that is known to be outdated and/or nonconservative, such as the aerosol particle sizing distribution derived from AEB-98-03. Further, any relevant guidance contained in DRA-ISG-2021-01, once updated to account for the issues identified by the DPO Appeal Panel, should be incorporated into the upcoming Revision to RG 1.183.

• Licensing Basis Clarity - As discussed in the resolution of Issue 4, Commission policy highlights the importance of clarity and accessibility of licensing basis information (57 FR 35455). The DPO Appeal Panel identified a number of issues that were not clearly addressed within the licensing basis for the FitzPatrick AST amendment. These include how previous safety functions related to the collection of MSIV packing leakage were accomplished, the basis for the selection of the bounding surrogate for the design basis 17 The DPO Appeal Panel can envision circumstances where the AEB-98-03 approach can be applied, in conjunction with appropriate conservatisms, to provide sufficient assurance that regulatory criteria can be met. Therefore, use of AEB-98-03, particularly in light of the conflicting regulatory history associated with its use, does not necessarily constitute an error. However, given the large uncertainties associated with aerosol deposition modeling, the conservatisms and context in which AEB-98-03 is applied in future assessments should be appropriately documented in the licensing basis.

55

accident, credit for fission product deposition in the main condenser, and use of methods known to be nonconservative for aerosol deposition.

• Conservatisms - As noted in the preamble to the alternate source term rule, These accident analyses are intentionally conservative in order to address uncertainties in accident progression, fission product transport, and atmospheric dispersion (64 FR 71990, 71991).

In the Directors Decision, several conservatisms are highlighted, including using the volume of only two of the four MSLs and only half of the reactor building volume to account for deposition of activity; not crediting containment elemental iodine and particulate deposition/plateout and activity reduction by the steam separators or the iodine partitioning in the reactor vessel; and not considering suppression pool fission product scrubbing and retention. In the discussion related to SOC Issue 2.g., the DPO Appeal Panel evaluated each of these factors and determined that they were: (1) not consistent with the approved licensing basis (crediting deposition in four main steam lines); (2) not supported by information provided in the AST LAR (credit for reactor building dilution and suppression pool scrubbing); or (3) a consequence of the in-vessel release modeled by the DBA (no credit for deposition on reactor vessel internals or iodine portioning). The DPO Appeal Panel notes that care must be used when determining what assumptions constitute actual conservatisms vice basic assumptions underpinning a deterministic and somewhat stylized evaluation such as the DBA analysis. However, the DPO Appeal Panel notes that other potential conservatisms (such as use of the pressure reduction function of the containment spray system) could be potentially credited and could reduce calculated doses.

• Use Of Confirmatory Calculations - The DPO Appeal Panel believes that use of confirmatory calculations, such as those performed by NRR staff in evaluating incoming LARs, are essential and provide valuable insights into complex licensing questions.

Furthermore, based on discussions with the NRR analyst who performed the confirmatory calculations for the NRR DPO review and the DPO Appeal Panels review of these calculations, the DPO Appeal Panel found the analyst to be highly knowledgeable, skilled in performing design basis dose calculations, objective, and professional. In the course of the DPO Appeal Panel review, requests for additional sensitivity studies prompted the analyst to identify an error in earlier calculations used to support the Directors Decision on the DPO.

The identification of an error is not unexpected when complex calculations are being performed without additional quality assurance activities such as a second checker or independent reviewer. However, as a result of this issue, the Directors Decision used a sensitivity result of 5.11 rem TEDE, an incorrect value, to demonstrate that various issues identified in the DPO were not significant. The updated dose result, with the identified errors corrected, results in a calculated control room dose of 7.5 rem TEDE. The calculated control room dose increases to 10.4 rem TEDE if deposition is credited in only two main steam lines, consistent with the FitzPatrick licensing basis. In light of this revised result, it does not appear that the basis supporting the Directors Decision remains valid given this higher sensitivity case result. Further, the DPO Appeal Panel believes that the good practice of performing confirmatory calculations could be enhanced for DPO reviews by having a peer review check by independent reviewers.18 18 The DPO Appeal Panel acknowledges that NRR Office Instruction ADM-401, NRR Technical Work Product Quality and Consistency, (ML20288A606) addresses the review of technical work products including confirmatory analyses that are important to a regulatory decision. However, it is not clear to the DPO Appeal Panel that, because of the reliance on line-management to implement the technical review process, this quality and consistency process is used for the evaluation of DPOs.

56

The DPO Appeal Panel greatly appreciates the support provided by NRR staff in supporting interviews and other information gathering, as well as the cooperation of Constellation engineering support staff.

2. Compliance with the Alternate Source Term Rule (10 CFR 50.67)

The DPO Appeal Panel considered the technical issues described under Section III.1 with respect to the requirements of 10 CFR 50.67. 10 CFR 50.67, requires, in part, that the applicants analysis of the consequences of applicable design basis accidents demonstrate with reasonable assurance that adequate radiation protection is provided to permit access and occupancy of the control under accident conditions without personnel receiving radiation exposure in excess of 5 rem TEDE. The DPO Appeal Panel determined that the licensees AST design basis accident analysis for FitzPatrick does not demonstrate with reasonable assurance that adequate radiation protection is provided to permit access to and occupancy of the control room under accident conditions without personnel receiving radiation exposures in excess of 5 rem TEDE. The basis for the DPO Appeal Panels determination includes the following:

• Failure to consider the potential dose impact from the outboard MSIV packing leakage following removal of the MSIV leakage control system. Leakage from the outboard MSIV packing would increase the calculated control room dose. The DPO Appeal Panel believes this is an error in the review and approval of the FitzPatrick AST LAR.

• Failure to assume a main steam line break as the limiting LOCA location. Assuming the LOCA break location was the main steam line, vice the recirculation line, would reduce main steam line fission product deposition and increase the calculated control room dose. The DPO Appeal Panel believes this is an error in the review and approval of the FitzPatrick AST LAR.

• Offsetting non-conservatisms associated with fission product deposition modeling, including credit for both drywell spray and main steam line deposition, with credit for the ability of the condenser to remove fission products prior to release to the environment. However, the licensee did not request or credit condenser fission product deposition in their AST LAR.

The DPO Appeal Panel believes implicit crediting of the condenser by the NRC staff without this mitigation capability included in the licensees calculation of record and in the absence of appropriate regulatory controls (e.g., procedural controls for alignment of a flowpath to the condenser, testing and environmental qualification of components needed to align flowpath) is an error in the staffs review and approval of the FitzPatrick AST LAR.

The licensees current calculation of record states that calculated control room doses under design basis accident conditions is 4.67 rem TEDE. However, given the limited margin to the 5 rem TEDE regulatory limit, the DPO Appeal Panel believes that, when the above errors are considered, the licensees AST calculation of record does not demonstrate with reasonable assurance that that control room dose limit of 5 rem TEDE during an accident is met. Therefore, the FitzPatrick AST license amendment does not meet the requirements of 10 CFR 50.67.

Further, the licensee has removed the MSIV leakage control system, so returning to the prior licensing basis (i.e., pre-AST licensing basis) would require design changes and changes to technical specifications. The DPO Appeal Panel recommends potential actions to address this compliance issue in Section IV.1.

57

IV. Recommendations The DPO Appeal Panel identified the following recommendations:

1. Address FitzPatrick Compliance Issues with 10 CFR 50.67 Take appropriate regulatory action to address FitzPatricks compliance with 10 CFR 50.67 and licensing basis clarity issues for the AST license amendment, including impact of outboard MSIV packing leakage, the basis for the limiting break location, credit for the main condenser, and aerosol modeling issues.

The DPO Appeal Panel recognizes that there are several options that, if implemented, may provide reasonable assurance that the control room dose limit could be met. These options include, but are not limited to:

• Reducing the allowable main steam line leakage limits in Technical Specifications to an appropriate lower value;

• Imposing appropriate regulatory controls on the condenser alternate release pathway to ensure that it would be available for fission product deposition (e.g., Technical Specification or technical requirements manual controls, appropriate environmental qualification of supporting SSCs, etc.);

• Requiring that the MSIV leakage control system be reinstalled with appropriate technical specification controls; or

• Crediting other mitigation capabilities and imposing appropriate regulatory controls to provide reasonable assurance that these other mitigation capabilities can perform their intended functions. For example, crediting actual containment pressure response during a postulated design basis accident rather than assuming peak containment pressure during the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of the accident can reduce leakage rates and reduce resulting doses. The DPO Appeal Panel requested a sensitivity case to illustrate that a reduction in containment pressure can reduce control room dose (Appendix 3).

Further, the DPO Appeal Panel recognizes that the DBA analysis is a complex calculation and imposing an appropriate regulatory correction would ultimately require that the licensee update their AST consequence analysis to demonstrate that the criteria in 10 CFR 50.67 were met.

Additionally, in updating their calculation of record, the licensee may utilize one or more of the above options in order to demonstrate that dose requirements are met. Therefore, the DPO Appeal Panel identified four potential options for addressing the 10 CFR 50.67 non-compliance issue:

a. Issue a 10 CFR 50.54(f) letter followed by a Confirmatory Action Letter (DPO Appeal Panel recommended option)

b. Issue an order to correct the 10 CFR 50.67 Non-Compliance

c. Issue an exemption to FitzPatrick to approve the existing AST licensing basis

d. Take no action Each of these options is described in the following sections. Regardless of the option ultimately selected by the OEDO, the DPO Appeal Panel believes that a near term action should be to hold a meeting with the licensee, consistent with the requirements of MD 3.5, Attendance at 58

NRC Staff-Sponsored Meetings, to inform the licensee of the OEDOs conclusions and actions resulting from the DPO Appeal.

a. Issue 10 CFR 50.54(f) Letter and Confirmatory Action Letter Although the DPO Appeal Panel has sufficient information to support a conclusion that the current FitzPatrick AST amendment does not comply with 10 CFR 50.67, requesting additional information from the licensee would accomplish several objectives: (1) provide formal notice to the licensee of the issues identified by the DPO Appeal Panel; (2) allow the licensee to respond to the identified errors; and (3) foster transparency in addressing the non-compliance. The DPO Appeal Panel acknowledges that the reasons for the non-compliance with 10 CFR 50.67 are due in large part to staff action; therefore, an issuance of a notice of violation would not be an appropriate response for this unique set of circumstances. In this case, the use of a 10 CFR 50.54(f) letter provides an opportunity for the licensee to provide their perspectives on an appropriate path forward to resolve the compliance issue.

10 CFR 50.54(f) states:

The licensee shall at any time before expiration of the license, upon request of the Commission, submit, as specified in § 50.4, written statements, signed under oath or affirmation, to enable the Commission to determine whether or not the license should be modified, suspended, or revoked. Except for information sought to verify licensee compliance with the current licensing basis for that facility, the NRC must prepare the reason or reasons for each information request prior to issuance to ensure that the burden to be imposed on respondents is justified in view of the potential safety significance of the issue to be addressed in the requested information. Each such justification provided for an evaluation performed by the NRC staff must be approved by the Executive Director for Operations or his or her designee prior to issuance of the request.

The 10 CFR 50.54(f) request for information would request the licensee to provide information related to their plans to restore compliance with 10 CFR 50.67. Because this information would be sought to verify licensee compliance with 10 CFR 50.67, the justification described in 10 CFR 50.54(f) would not be required. The information obtained from the licensee may then be used as the basis for a Confirmatory Action Letter (CAL). As described in the NRCs Enforcement Manual, Section 4.3, a CAL is an administrative action and is a letter issued to a licensee to emphasize and confirm a licensee's agreement to take certain actions in response to specific issues. The Enforcement Manual states several limitations on the use of CALs, including:

• The NRC expects licensees and non-licensees to adhere to any obligations and commitments addressed in a CAL.

• CALs should only be issued when there is a sound technical and/or regulatory basis for the desired actions discussed in the CAL.

• CALs must meet the threshold defined in the Enforcement Policy, i.e., "to remove significant concerns about health and safety, safeguards, or the environment.

Further, the level of significance of the issues addressed in a CAL should be such that if a licensee did not agree to meet the commitments in the CAL, the staff would likely proceed to issue an order.

5

The Enforcement Manual also notes that because a CAL confirms an agreement by the licensee, a CAL may require some negotiation and the licensee must agree to take the action. As discussed in the analysis for option b, related to issuance of an order (Section IV.1.b), the DPO Appeal Panel believes that, because this issue appears to be significant enough to meet the guidelines for a compliance backfit, it is suitable for the CAL process.

The DPO Appeal Panel also believes that this option provides several benefits, including providing an opportunity for the licensee to voluntarily correct the errors in the DBA analysis, provide for regulatory oversight of the corrective actions, and avoid the need to process this issue as a potential backfit.

The DPO Appeal Panel recommends this option.

b. Issue an Order to Correct the 10 CFR 50.67 Non-Compliance Under this option, the NRC would issue an order to require the licensee to address the non-compliance with 10 CFR 50.67. The DPO Appeal Panel believes that this action would constitute backfitting under 10 CFR 50.109. Therefore, the DPO Appeal Panel considered this option within the context of the agencys backfitting guidelines in MD 8.4, Management of Backfitting, Forward Fitting, Issue Finality, and Information Requests (DT-19-15; September 20, 2019).

The DPO Appeal Panel notes that this option has several challenges, including identifying specific actions that would need to be taken by the licensee. As noted earlier in Section IV.1, there are a variety of means by which the compliance issues for 10 CFR 50.67 could be addressed. Ultimately, the licensee would need to prepare and resubmit a DBA analysis that demonstrated with reasonable assurance that the requirements of 10 CFR 50.67 are met. If the reanalysis relies on the need to credit functions not currently within the licensing basis (e.g., fission product deposition in the main condenser), the licensees resubmittal should include appropriate means to ensure these credited functions have appropriate controls (e.g., procedures, functional requirements, testing). Following review of the updated analysis, the staff could update and reissue the safety evaluation. Therefore, the DPO Appeal Panel determined that the issuance of an order to require the licensee to reanalyze the DBA for the alternate source term and resubmit the AST application per 10 CFR 50.67(b), based on a change in the staffs position from 2020 that the licensee meets 10 CFR 50.67, to be a staff action that would constitute backfitting.

In evaluating this potential backfit, the DPO Appeal Panel determined that this action would not be necessary to ensure that the licensees facility provides adequate protection to the health and safety of the public and is in accord with the common defense and security. The basis for the DPO Appeal Panels determination includes the following considerations:

• For the DBA accident analyzed, there still remains significant margin to the offsite criteria (i.e., calculated offsite doses are at least a factor of 20 below the 10 CFR 50.67 dose criteria). The technical issues identified in the analysis would not be expected to exceed this level of margin for the offsite dose criteria.

• The main DBA dose impact is demonstrating that the control room dose criteria of 5 rem TEDE is met. However, the DPO Appeal Panel believes that, during an actual accident, control room doses would be mitigated by multiple defense-in-depth actions by the licensee as described earlier (e.g., see discussion in Section I.4).

60

Therefore, the DPO Appeal Panel believes that NRC actions to address the issues substantiated by its review would not meet the adequate protection threshold for backfitting.

However, NRC action to issue an order to restore compliance with 10 CFR 50.67 meets the compliance exception to the requirement to perform a backfit analysis, as set forth in 10 CFR 50.109(a)(4)(i). The compliance exception is associated with omissions or mistakes of fact in the licensing process. The DPO Appeal Panel believes the issues related to the potential for unanalyzed MSIV packing leakage (SOC Issue 2.a); selection of the limiting LOCA accident conditions (SOC Issue 2.e); and the crediting deposition in the power conversion (SOC Issue 3) to offset uncertainties in the treatment of aerosols (SOC Issues 2.f and 2.g) result in the failure of the applicants AST analysis to demonstrate, with reasonable assurance, that calculated control room dose does not exceed 5 rem TEDE for the duration of the accident. The DPO Appeal Panel determined that these errors result in a non-compliance with the requirements of 10 CFR 50.67. Accordingly, the DPO Appeal Panel concludes that the appropriate justification for the backfit is the compliance exception to the requirement to perform a backfit analysis. The MD 8.4 Handbook,Section III.B.6 provides the following guidance:

Use of the compliance exception is limited to the following situations that define omission or mistake of fact:

(a) The NRC staff, whether by its own error or by licensee or third-party error or omission at or before the time of its determination that a known and established standard of the Commission was satisfied: (1) incorrectly perceived facts, (2) performed or failed to recognize flawed analyses, or (3) failed to properly draw inferences from those facts or analyses, as judged by the standards and practices that were prevailing among professionals or experts in the relevant area at the time of the determination in question, and (b) Those facts, analyses, or inferences have now been properly perceived, performed, or drawn.

(c) The error or omission may have been committed by any involved party and must be traced to (i) The original LB or to a change to the LB, (ii) The regulations reflected in the LB at the time of implementation that were applicable to the licensee whose LB is at issue, and (iii) Any standards and practices in existence at the time the original determination was made.

The understanding for what constituted proper implementation of the regulations, standards, and practices must have been widely known or understood by professionals at the time. This is not restricted to the regulatory positions of the NRC but includes any industry or professional standards and practices in existence at the time the original determination was made.

61

An evaluation of the compliance backfit is documented in Appendix 4. The evaluation concluded that the proposed order to ensure compliance with 10 CFR 50.67 constitutes backfitting. The proposed order would require the licensee to correct and resubmit the AST license amendment to demonstrate compliance with 10 CFR 50.67. Further, the proposed order meets the criteria to invoke the compliance exception to performing a backfit analysis.

As such, the NRC would not need to prepare a backfit analysis to support the proposed order.

The DPO Appeal Panel does not recommend this option as it does not provide an opportunity for the licensee to voluntarily address the noncompliance in a manner most efficient and cost beneficial for them or allow for a fulsome consideration of licensee perspectives prior to issuance of an order.

c. Issue an Exemption Under this option, the staff would process an exemption under 10 CFR 50.12, Specific exemptions. The exemption would acknowledge the errors in the licensees analysis of record and associated staff safety evaluation, and approve, via exemption from 10 CFR 50.67, the licensing basis approved in the FitzPatrick AST amendment. In order to process such an exemption, the staff would need to meet one of the special circumstance provisions of 10 CFR 50.12. In considering the findings of the DPO Appeal Panel, a potential special circumstance could be the undue hardship that would result from addressing the non-compliance in recognition of the staffs role in these circumstances, combined with the determination that the current situation does not represent an immediate safety issue. This option would be used if the OEDO determines that the errors are minor in nature and do not set any precedent for future reviews.

The DPO Appeal Panel does not recommend this option as it does not correct the underlying noncompliance issues, could set an inappropriate precedent for future reviews, and may diminish public confidence in the agency.

d. Take No Action Under this option, the staff would take no action to address the non-compliance with 10 CFR 50.67 under Recommendation IV.1. The other recommendations in Section IV of this report would still be endorsed by the DPO Appeal Panel. This option could only be used if the conclusion is reached that the licensee is reasonably in compliance with the rule currently.

The DPO Appeal Panel does not recommend this option because it would not clarify ambiguities in the FitzPatrick licensing basis (e.g., credit for fission product deposition in the condenser) and the approval basis for the FitzPatrick AST LAR could be perceived to establish a new licensing precedent that includes the errors noted by the Appeal Panel.

Recommended action for addressing compliance issues The DPO Appeal Panel recommends the issuance of a 10 CFR 50.54(f) letter followed by a CAL as the preferred means to address the issue of non-compliance.

62

2. Clarify Purpose and Objectives of the DBA Analysis Evaluate the purpose and objectives of the DBA analysis conducted for AST under 10 CFR 50.67 to more clearly articulate the underlying assumptions, equipment capability, and basis for crediting safety-related engineered safety features and non-safety-related equipment. This will better focus future regulatory guidance and licensing decisions on the underlying purpose of this evaluation. This recommendation should be coordinated with implementation of the recommendations from DPO-2020-002, Differing Professional Opinion Regarding RG 1.183 Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors (ML21060A972, nonpublic; ML21067A645, public). The DPO Appeal Panel believes this could be a longer-term effort compared to the other recommendations.

3. Enhancements to DPO Review Process Enhance the process used to evaluate differing professional opinions to ensure that Director Decisions are based on factual, accurate, and timely information. This includes guidance for an appropriate level of independent peer review or quality assurance on confirmatory calculations.

4. Updated Guidance for DBA Reviews In light of the issues identified in this report, in the near term, revise and consolidate the staffs updated guidance (DG-1389 and DRA-ISG-2021-01). Ensure that guidance clearly states the regulatory positions associated with acceptable methods for the design basis accident analysis, including guidance for alternate methods that are reviewed on a case-by-case basis or require plant-specific analysis. The DPO Appeal Panel believes any update to RG 1.183 should be consolidated into a single revision to the regulatory guide and not include companion interim staff guidance. Specific issues that should be addressed in the guidance update include:

• Enhanced focus on the overall intent of regulations related to the DBA analysis (e.g., focus on assessing the acceptability of engineered safety features rather than overreliance on non-safety-related features (e.g., deposition in power conversion systems).

• Provide more specific guidance for evaluating alternative options in Regulatory Guidance and SRPs.

• Revise language in draft DRA-ISG-2021-01 relative to MSIV packing leakage to better reflect actual operating experience (i.e., events indicate that packing leaks may not be small contributors to overall main steam line leak rate).

• The current staff position in use of methods previously approved by the staff but now known to contains errors and is non-conservative (e.g., AEB-98-03).

• Use of an appropriate communication vehicle (e.g., Regulatory Issue Summary), to inform industry stakeholders of the appropriate use of AEB-98-03 and applicability of RIS 2006-04 to future reviews.

• Future consideration of withdrawal of RG 1.183, Revision 0, following issuance of updated guidance (i.e., RG 1.183, Revision 1). Note that withdrawal of RG 1.183, Rev. 0, may have backfitting and forward fitting implications.

5. Enhancements to the License Amendment Review Process Enhance the guidance for licensing amendment reviews to:

63

• Provide improved guidance on how items that are required to address areas of high uncertainty should be incorporated and documented clearly into the licensing basis.

• Provide guidance on addressing licensing precedents when the underlying technical basis for the previous precedent is no longer valid or is reasonably believed to be non-conservative.

• Improve processes to enhance the clarity and logic used in safety evaluations. Ensure that the basis of the staffs approval is clearly described and establishes an unambiguous licensing basis for the licensee.

• When differing technical views arise during the assessment of a requested license amendment, recommend elevating the concerns to management in order to effectively address and resolve technical issues, which may necessitate an appropriate level of independent peer review.

6. Address Potential Generic Implications for BWR Outboard MSIV Packing Leakage Evaluate if the failure to consider outboard MSIV packing leakage for the FitzPatrick AST LAR represents a potential generic issue and evaluate issue within appropriate process. Additionally, consider the need for issuance of generic communication on MSIV packing leakage and the need to evaluate operational leakage for its impact to the compliance with main steam line leakage Technical Specification requirements.

7. Address Extent of Condition for Other AST License Amendments The DPO Appeal Panel notes that methods similar to those used for the FitzPatrick AST license amendment have been applied to other nuclear power plants. Most recently, license amendments were approved for Nine Mile Point (ML20241A190), Quad Cities (ML20150A328),

and Dresden (ML20265A240). The DPO Appeal Panel recommends that the approval basis for these license amendments and other similar AST amendments be reviewed to determine if errors similar to those in the FitzPatrick AST amendment are present and appropriate corrective action be taken.

64

I. APPENDIX 1: DPO-2021-001 APPEAL

SUMMARY

OF CONCERNS Request:

1. Adopt all the DPO Panels recommendations in its evaluation of the DPO.

2. Correct the following documents:

a. Plant-specific amendments and Safety Evaluation Reports (SERs) impacted by a revised control room dose analysis as described in the DPO1

b. Draft revision to RG 1.183, Alternative Radiological Source Terms for Evaluating Design Bases Accidents at Nuclear Power Reactors, (DG-1389, ML21204A065).

3. Task the NRC staff to use the backfit process to assess whether backfitting is warranted for the issues raised in the DPO (e.g., restoration of main steam line isolation valve (MSIV) leakage control system, reduction in MSIV allowable leakage, correction of licensees calculation of record).

4. Direct the NRC staff to not issue Interim Staff Guidance (ISG) DRA-ISG-2021-XX, Supplemental Guidance for Radiological Consequence Analyses Using Alternative Source Terms, draft issued May 2021.

5. Provide an EDO Appeal Decision that corrects errors in the Directors Decision so that flawed information is not propagated into the licensing and engineering review process.

Issues the DPO Appeal Review Team will address:

1. Determine if the FitzPatrick Alternate Source Term (AST) license amendment is in compliance with 10 CFR 50.67, is based on the applicants analysis as required by 10 CFR 50.67(b)(2), and whether the information discussed in Issue 2 below would result in control room doses higher than those allowed under 10 CFR 50.67.

2. Determine if the LOCA assumptions and methods used by FitzPatrick2 and the NRC staff to evaluate the FitzPatrick AST license amendment for a revised LOCA analysis are accurate and consistent with the stated regulations and standard review plan guidance (or if deviations from guidance by the licensee were documented and included an 1 In addition to the Safety Evaluation Report for the FitzPatrick AST license amendment (ML20140A070),

the submitters noted that three other safety evaluations associated with increases to allowable MSIV leakage may be impacted: Quad Cities (ML20150A328), Nine Mile Point (ML20241A190), and Dresden (ML20265A240).

2 In their AST license amendment request (ML19220A043), FitzPatrick indicated that their application conformed with the relevant guidance of RG 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Rev. 0, issued July 2000 (ML003716792), and was consistent with other guidance contained in NUREG 0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants LWR [Light-Water Reactor] Edition, Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Rev. 0, issued July 2000 (ML003734190) and RG 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nuclear Power Plants, issued June 2003 (ML031530505).

I-1 APPENDIX 1

adequate technical basis). This determination will take into account the technical issues that were described by the submitters in their DPO which are summarized below:

a. The licensee amendment, NRC staff and Directors Decision did not consider a limiting credible accident scenario that bypasses the assumed pathway to the environment (i.e., leakage from the MSIVs packing, body or mechanical joints directly to the environment). (RG 1.183, Appendix A, Section 6.4 and 6.5)

b. By crediting the air space in the main steam line up to the closed MSIVs, the applicants analysis did not assume the radioactive release was instantaneously and homogenously mixed throughout all the free air volume in the drywell containment. Despite the licensees statements that the containment elemental and particulate (natural) deposition/plateout is not credited, this deposition is credited in the space upstream of the MSIVs that is part of the drywell containment. (RG 1.183, Appendix A, Section 3.1)

c. By assuming a release point at a specific location inside the turbine building, rather than on the surface of the turbine building, improper dilution was credited in the turbine building. This dilution reduced the atmospheric dispersion factors, thereby underestimating the control room dose. (RG 1.183, Appendix A, Section 6.4)

d. By crediting nonsafety-related structures systems and components such as vents and doors that are not controlled by technical specifications, the licensee omitted release pathways that would be closer to the control room and would result in higher control room doses. As such, the accident analysis no longer aligns with, or is bounded by the plant design basis. (RG 1.183, Regulatory Position 5.1.2 and RG 1.194, Regulatory Position 2.0).

e. The analysis did not consider the limiting LOCA break location by selecting a recirculation line break, rather than a break in the reactor coolant system just prior to the MSIVs. In addition, when the licensee modeled the break just before the MSIVs (provided as a sensitivity study in response to the staffs request for additional information), the licensee did not use the assumptions that were used in its LOCA analysis supporting its license amendment request. (RG 1.183, Appendix A)

f. By using removal coefficients for aerosol settling that are nonconservative and by incorrectly modeling the impact of changing aerosol particle size distribution as the radioactive material moves down the steam line, the LOCA model overestimates the deposition in the steam lines and, therefore, underestimates the dose in the control room. (RG 1.183, Appendix A, Sections 6.3 and 6.5)

g. By using an elemental iodine removal constant greater than 20 hr-1 in the licensees LOCA analysis, the analysis resulted in elemental removal greater than the SRP Section 6.5.2, Containment Spray as a Fission Product Cleanup System, Revision 4, dated March 2007 (ML070190178) limit on removal. Although this deviation was discussed in the NRC safety evaluation, it did not assess the aggregate impacts of the deficiencies in the analysis described above. (RG 1.183, Appendix A, Section 3.3, and SRP Section 6.5.2)

I-2 APPENDIX 1

3. Evaluate potentially conflicting statements between the Directors Decision and the NRC staffs safety evaluation on crediting portions of the power conversion system condenser and determine if any mitigation function assigned to the condenser and associated equipment is appropriately documented in the licensing basis.

• In their Appeal, the DPO submitters stated that the Directors decision asserts that the condenser was not credited and the staff used engineering judgment and risk and engineering insights to support the deterministic conclusion and balance any uncertainties and non-conservatisms rather than as a basis for the decision. The submitters further noted that these statements conflict with information contained in the NRC staffs safety evaluation, meetings held with one of the DPO submitters, versions of the ISG that existed at the time the DPO was submitted, and additional information as discussed in Enclosure 1 of the DPO.

4. Determine if the approval basis of the FitzPatrick AST amendment was clearly documented in the licensing basis (e.g., calculations of record, updated final safety analysis report updates, SER, and whether they appropriately reflect the assumptions and mitigating equipment credited in the NRC staffs analysis).

Note: The DPO Appeal Panel does not intend to address how potential remedies to any identified issues would be addressed through potential revisions to interim guidance documents under development, changes to the draft revision to RG 1.183, or under the agencys backfit regulations in 10 CFR 50.109. However, under the DPO Appeal process, if issues raised by the submitters are substantiated, it is anticipated that direction from the OEDO to NRR would include making conforming changes to these documents, in addition to any other necessary remedies.

With regard to backfitting, if technical or regulatory issues are identified during the review that require corrective action, there is a range of potential actions that could be taken, some of which may not require formal backfitting (e.g., request for a license amendment or exemption, addressing issues under the licensees corrective system, and/or revisions to the SER).

Therefore, the question of backfitting will be deferred until the OEDO and NRR have an opportunity to review the outcome of the DPO Appeal Panel review and decide on an appropriate course of action. Although these issues will not be resolved during the DPO Appeal Panel review, they will be addressed, as appropriate, as part of the overall DPO process and subsequent follow-up actions.

I-3 APPENDIX 1

II. APPENDIX 2: Background on Main Steam Line Fission Product Deposition Guidance The DPO Appeal Panel reviewed the historical context for the treatment of aerosol deposition in BWR main steam lines in order to gain an appreciation for the state of practice and available guidance in this area. This appendix summarizes the results of this review and highlights many of the aerosol related issues raised in the FitzPatrick AST LAR DPO.

1. RG 1.183 and AEB-98-03 Modeling of Aerosol Deposition Regulatory Guide (RG) 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, Revision 0, was issued in July 2000. Credit for fission product deposition in the main steam lines is included in RG 1.183, but no specific guidance is provided for acceptable methods. Specifically, RG 1.183, Appendix A, Section 6.3, states that:

Reduction of the amount of released radioactivity by deposition and plateout on steam system piping upstream of the outboard MSIVs may be credited, but the amount of reduction in concentration allowed will be evaluated on an individual case basis. Generally, the model should be based on the assumption of well-mixed volumes, but other models such as slug flow may be used if justified.

Similarly, RG 1.183, Appendix A, Section 6.5, states:

A reduction in MSIV releases that is due to holdup and deposition in main steam piping downstream of the MSIVs and in the main condenser, including the treatment of air ejector effluent by offgas systems, may be credited if the components and piping systems used in the release path are capable of performing their safety function during and following a safe shutdown earthquake (SSE). The amount of reduction allowed will be evaluated on an individual case basis.

Although RG 1.183 does not reference acceptance methods for calculating main steam line deposition, RIS-06-04, Experience with Implementation of Alternative Source Terms, issued in March 2006, provided additional guidance. Specifically, RIS-06-04 stated that use of Accident Evaluation Report (AEB) 98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application Using the Revised (NUREG-1465) Source Term, (ML011230531) was acceptable. AEB-98-03, issued in December of 1998, documented the staff's evaluation of the thermal-hydraulics, fission product deposition, and resulting radiological consequences associated with use of a revised source term for the Perry nuclear plant. However, the RIS further noted that any licensee who choose to reference the AEB-98-03 assumptions should provide appropriate justification that the assumptions are applicable to their particular design.

A key parameter derived from the AEB-98-03 report is aerosol settling velocity, which is used to determine aerosol deposition in the main steam lines. AEB-98-03 used a Monte-Carlo technique to develop a probabilistic distribution for settling velocity based on assumed variability in input parameters such as aerosol density, aerosol diameter, and aerosol shape factor. However, shortly after RIS-06-04 was issued, concerns regarding mistakes in the analysis in AEB-98-03 II-1 APPENDIX 2

were communicated to the NRR Associate Director for Risk Assessment & New Projects in an August 23, 2006, memo from the RES Director of the Division of Risk Assessment & Special Projects (ML062010315, nonpublic). These issues included recognition, as early as April 2005, that the AEB-98-03 approach included potentially nonconservative aerosol density and diameter distributions (ML062010341, nonpublic). As a result, the memo noted that the original AEB 03 report overestimated aerosol settling velocity. Additional analysis (ML062010276, nonpublic) further refined the aerosol density and shape factor distributions. RES provided the following updates to NRR with regard to the AEB-98-03 settling velocity:

Settling Velocity (m/s)

Cumulative AEB-98-03 Updated AEB- Updated Probability (December 1998) 98-03 (July Analysis of Distribution 2005)1 Aerosol Percentile Deposition (August 2006)2 10 2.1E-04 2.95E-05 5.32E-06 (most likely) - 3.60E-05 5.34E-06 40 8.1E-04 8.81E-05 1.52E-05 50 (median) 1.17E-03 1.17E-04 2.01E-05 60 1.48E-03 1.55E-04 2.64E-05 1

Includes updates to the aerosol density and diameter distributions 2 Includes additional updates to the aerosol diameter and shape factor distributions based on input from Sandia National Lab (ML062010296, nonpublic)

These results indicate that the original 1998 AEB-98-03 report overpredicted settling velocity by over an order of magnitude compared to the 2006 updated analysis. Overprediction of settling velocity would lead to non-conservative estimates of radionuclide removal by deposition. It was recognized that the original AEB-98-03 report had been used to eliminate MSIV leakage control systems in BWRs, but no assessment of the impact was included in the August 2006 memo.

In a memo (ML070180233, nonpublic) to RES dated January 17, 2007, NRR noted that there was insufficient information available to perform a backfit analysis or revise regulatory positions based on the issues identified with AEB-98-03. The memo further stated that NRR intended to prepare a User Need memorandum requesting an update to AEB-98-03 and that this information would be incorporated into an update to RG 1.183. NRR also noted that the information provided in the 2006 RES memorandum does not present an immediate safety concern. There is sufficient conservatism in the accident dose analyses and assumptions to accommodate minor increases in dose based on the errors in AEB-98-03.

NRR transmitted the User Need request for additional in this area work to RES on June 24, 2007 (ML071620396, nonpublic). This request reiterated that NRR recognized in 2005 that some of the assumptions in AEB-98-03 needed to be reevaluated and that supplemental information was needed to support a backfit analysis of previous uses of AEB-98-03. This request was identified as UNR NRR-2007-005 and requested MELCOR analysis of postulated recirculation line and main steam line breaks to determine the accident source term and fission product removal in containment, the main steam lines, and the condenser. RES completed the work specified in UNR NRR-2007-005 with the issuance of Sandia technical report SAND2008-6601, which included the conduct of an independent review by Dycoda, a consulting firm with extensive experience in severe accident analysis (ML083180181, nonpublic).

II-2 APPENDIX 2

2. SAND2008-6601 - MELCOR evaluation of Design Basis Accident In response to UNR NRR-2007-005, the NRC contracted with Sandia National Laboratories (SNL) to perform a reassessment of the methods in AEB-98-03 (e.g., see ML091520056). The results of this reassessment are contained in SNL report SAND2008-6601, Analysis of Main Steam Isolation Valve Leakage in Design Basis Accidents Using MELCOR 1.8.6 and RADTRAD (ML083180196 and ML113400138 (errata)). Conclusions documented in the SAND2008-6601 report include the following:

• MELCOR best-estimate analyses of two design basis accidents (recirculation line break and steam line break LOCAs) show that during the first two hours of the accidents, aerosol concentrations in the reactor vessel significantly exceed those in the drywell.

• Current regulatory guidance permitting use of the fission product concentration in the drywell atmosphere for the first two hours is non-conservative and conceptually inaccurate because fission products entering the steam lines originate in the reactor vessel steam dome rather than the drywell.

• Recognition that drywell sprays reduce fission product release by reducing the drywell pressure, but also noted that sprays did not appreciably reduce fission product concentration in the steam dome during the first hour for the accidents evaluated.

Following reactor vessel reflood at two hours, the report noted that sprays could be credited for reducing fission product concentration flowing back into the reactor vessel and main steams from the drywell. This issue is discussed further in Section 3 of this Enclosure.

• Recommendation that no credit be taken for removal of fission products in the in-board segments of main steam lines due to uncertainties associated with potential for revaporization of deposited fission products and the influence of thermally driven flow that could enhance the transport of fission products to the main steam lines.

SAND2008-6601 calculated updated steam line removal coefficients for a recirculation line break accident based on updated distributions for several parameters associated with aerosol settling velocity (e.g., including the aerosol shape factor, the Cunningham slip factor) and other aerosol physics not previously considered in AEB-98-03 that influence particle agglomeration, diffusion, and thermophoretic deposition. These uncertainty distributions allowed the calculation of 5 h, 50th, and 95th percentile removal coefficients for main steam line aerosol deposition. The 5th percentile values were considered by the authors to be conservative and the 50th percentile values were intended to represent mean tendencies and could be used to avoid conservatism.

II-3 APPENDIX 2

Note: Table numbers are from SAND2008-6601.

3. Impact of Drywell Sprays on Main Steam Line Deposition The SAND2008-6601 report challenged the viewpoint that the initial source of fission products to the main steam line release pathway (for unbroken main steam lines) would be from the drywell rather than the reactor vessel. The report noted that during the in-vessel release phase, the source of fission products to the main steam lines would be the reactor vessel rather than the drywell. Following accident recovery and release termination (assumed to be two hours after the start of the accident), the assumed reflooding of the vessel would sweep fission products into the drywell. In discussing this issue, Section 1 of the report stated:

This misconception or oversimplification in viewing fission product transport from overheated fuel has led to subsequent important conceptual errors in analysis such as proposed use of drywell sprays to reduce airborne radioactivity (as illustrated in Figure 1-4) or equilibrating drywell and wetwell airspace volumes to achieve the same effect, when neither of these processes can directly affect the airborne concentration within the reactor vessel where a continuous source of fission products issues from the overheated fuel.

II-4 APPENDIX 2

Consistent with this view, SAND2008-6601 recommended that the drywell sprays not be credited for fission product removal associated with main steam line leak path during the first two hours of the accident. After two hours, the drywell sprays could be credited for fission product removal associated with the MSIV leakage path because reflood of the vessel, which is assumed to terminate the release, would sweep fission products from the reactor vessel and promote mixing within the drywell. SAND2008-6601 also provided recommended scaling factors to adjust the steam dome fission product concentration relative to the drywell for the first two hours of the accident. The report did note that credit for sprays for pressure reduction within containment would seem reasonable provided it was supported by adequate engineering analysis.

In 2010, the recommendations of SAND2008-6601 were challenged by industry in a paper written by J.E. Metcalf and P.B. Perez21. Metcalf and Perez noted that the scaling factors proposed by SAND2008-6601 would have a significant impact on operating BWRs. They then assessed the potential for trapping clean steam between the reactor vessel and the inboard MSIVs and considered the potential conservatisms associated with assuming that the drywell concentration as the source of MSIV leakage. Based on this work, Metcalf and Perez concluded that assuming the source for the main steam line leakage was the drywell is acceptable.

However, the paper also stated:

When evaluating the conservatism of choosing either the steam lines or the drywell to be the source of steam line leakage, the relative benefits of these effects needs to be taken into account, because both could be credited in the DBA-LOCA analysis. One thing seems certain, however: it would be non-conservative to give full credit for both effects. That is, either the activity makes its way to the leak point of the MSIV via the drywell (e.g., by a MSLB as the initiating event) or the activity makes its way to the MSIV leak point via the main steam line. Its unlikely (because of the thermal-hydraulic conditions within the reactor vessel during in-vessel core degradation) that a large fraction of the MSIV leakage would reside for a while in the drywell (to be depleted by drywell sprays) 21 J.E. METCALF, P.B. PEREZ, BWR Steam Line Radionuclide Concentration Distribution following a DBA LOCA, presented at the 31st Nuclear Air Cleaning Conference, International Society of Nuclear Air Treatment Technologies, Charlotte, NC, 19-21 July 2010.

http://isnatt.org/Conferences/31/09.%20Metcalf%20NACC-Paper-Final.pdf II-5 APPENDIX 2

and then reach the MSIV leak point via the steam lines (experiencing deposition in that pathway).

This aligns with the recommendation provided by Metcalf and Perez that [w]hen applying the AST for MSIV leakage dose assessment, credit may be taken either for deposition in the steam lines between the reactor vessel and the inboard MSIV or for drywell sprays but not for both.

The DPO Appeal Panel notes that this recommendation is consistent with DG-1389 proposed guidance that credit for main steam line fission product deposition is not valid if credit has been taken for aerosol removal from drywell sprays.

In response to the Metcalf and Perez 2010 paper, N.C. Andrews and R. O. Gauntt summarized the Metcalf and Perez position in SAND2017-932R, Treatment of Reactor Systems within Draft Regulatory Guide 1.183 DG-1199,22 as follows:

Metcalf argues that aerosol deposition and the absence of mixing in the steam line are sufficient, such that using the drywell aerosol concentration as a surrogate for the concentration at the MSIV is both bounding and conservative.

Accordingly, he states that the concentration of aerosols in the drywell is higher than that just ahead of the inboard MSIV within the steam lines. In the paper, it is acknowledged that this treatment is not in fact the most physical treatment but claimed as a conservative surrogate for the actual situation. (Metcalf and Perez, 2010)

Metcalf then examines the total deposition that can occur in the main steam line before the MSIV in order to provide further justification. Using a two volume calculation, Metcalf calculates a decontamination of 80% in the first control volume and a decontamination of 60% in the second control volume. The leads to an overall removal efficiency of 92% for the system. This, Metcalf claims, is higher than the reduction in leakage as a result of drywell sprays, which is claimed to be a factor of 5 to 6. Per this result he argues that it would be conservative to credit either deposition in the main steam lines or drywell sprays, but not both.

SAND2017-9329R noted that a key parameter for the assessment of main steam deposition is the assumed aerosol size distribution and resulting settling velocity. Metcalf and Perez indicated that they applied a median settling velocity for each deposition volume, but determined that the median value provided in AEB-98-03 of 1.17E-03 m/s was too high for control volumes in series with the drywell as the source. Therefore, Metcalf and Perez used median settling velocities of 6.8E-04 m/s for the first volume and 2.6E-04 m/s for the second volume in series.

The DPO Appeal Panel notes that even these reduced median settling velocities were a factor of approximately 13 to 34 times greater than the revised median provided in the 2006 RES memorandum to NRR (ML062010315) and would potentially over-predict fission product removal by deposition. SAND2017-9329R instead recommended an aerosol distribution with an AMMD range of 1.0 to 2.0 m with a geometric standard deviation of 2.0 based on NEA/CSNI/R(2009)5. SAND2017-9329R than compared the maximum decontamination factor based on the NEA/CSNI/R(2009)5 recommended aerosol size to AEB-98-03 and noted that decontamination factors calculated using AEB-98-03 can be more than an order of magnitude greater than those calculated using an AMMD range of 1 to 2 m. These results called into question the conservatism associated with using the drywell fission product concentration to estimate fission product deposition in the main steam lines as suggested by Metcalf and Perez.

22 https://www.osti.gov/biblio/1762012 II-6 APPENDIX 2

In considering the differing perspectives offered by Sandia and Metcalf and Perez, the DPO Appeal Panel notes that both agree on two key modeling issues: (1) credit for both drywell spray and main steam line deposition is not technically justified, and (2) AEB-98-03 overpredicts aerosol settling velocities and resultant fission product removal.

4. DG-1199 - Initial Draft Update for RG 1.183 In October 2009, the NRC issued DG-1199, a proposed revision 1 to RG 1.183, for public comment (74 FR 52822 ). With regard to main steam line aerosol deposition, DG-1199 included the following acceptable recommendations:

• DG-1199, Appendix A, Section A-5.1, stated that the source of MSIV leakage is assumed to be the concentration in the reactor vessel steam dome. At the end of the in-vessel release phase (i.e., arrest of core damage due to reflood), the reactor vessel steam dome should be assumed to be equal to the containment (or drywell activity). The modeling and scaling adjustment factors contained in Section 5.2 of SAND2008-6601 were referenced in DG-1199 as an acceptable approach for estimating the increase in reactor vessel steam dome concentration compared to the drywell for the initial phases of a design basis accident.

• SAND2008-6601, Section 6.3, was referenced as an acceptable means to model aerosol deposition. In particular, DG-1199, Appendix A, Section A-5.8, referred to SAND2008-6601 Tables 6-1 and 6-2. Table 6-2 (50th percentile values) was recommended for use from the start of the accident until termination of the early in-vessel release phase, and Table 6-1 (5th percentile values) following the in-vessel release phase.

Prior to the issuance of DG-1199 for public comment, an NRR staff member submitted a non-concurrence on the DG-1199 (ML091520056). Specific concerns raised in the nonconcurrence included:

• The proposed change used a beyond design bases analysis by incorporating a MELCOR in-vessel source term that maximized radio-aerosol concentration.

Experience would show that for some or all of the first hour after a plant transient, flow may actually be into the vessel from residual steam from the turbine stop valves back to the steam dome rather than being biased out to maximize dose to the control room.

• The Sandia report clamped (normalized) the in-vessel concentration after the first hour of the accident instead of taking the analyzed best estimate data that showed the concentration in-vessel substantially declining after the first hour of the accident. This could substantially decrease the dose from the MSL pathway for the duration of the accident and may prove that the AST/TID containment source term would, in actuality, be more conservative over the total duration of a DBA LOCA for BWRs.

• In addition to these non-realistic biases used to develop the current guidance, the analysis did not model or assume that the BWR vessel separators and dryers would not reduce dose consequence by deposition. These assumptions were based on a PWR study that is not necessarily applicable to BWR designs.

II-7 APPENDIX 2

The response to the nonconcurrence evaluated these concerns, but no changes were made to DG-1199 as a result of the non-concurrence review. However, the Director of NRRs Division of Risk Analysis instructed the cognizant Branch Chief to keep a record of the concerns and to reevaluate whether any changes are needed to DG-1199 the public comment period was complete. The NRC received 150 comments on DG-1199 (summarized in draft, non-public, ML112340473). As noted during a public meeting held in November 2020, work on finalizing DG-1199 was deferred as staff reviewed several AST related licensing submittals, including the AST submittal for FitzPatrick (ML20296A425). The NRC staff elected not to finalize DG-1199 and instead issued DG-1389 as the new proposed revision 1 to RG 1.183.

5. DG-1389 - Second Draft Update to RG 1.183 DG-1389 incorporated resolution of DG-1199 public comments, recommendations from DPO-2020-002 (ML21067A645), and lessons learned from recent licensing experience. DG-1389 was issued for public comment in April 2022 (87 FR 23891). The DG-1389 guidance for addressing fission product deposition in the main steam lines provides for three different acceptable options, but such credit is contingent upon not crediting fission product removal by drywell sprays. Specifically, Appendix A, Section A-5, of DG-1389 states:

Three methods are presented below to compute aerosol deposition within main steamlines. Each method computes similar removal coefficients suitable for radiological consequences calculations, however, these methods are not valid if credit has been taken for aerosol removal from drywell sprays.

The three MSIV leakage models are the:

1. Direct adoption of the SAND 2008-6601 (Ref. A-11) recommendations without scaling R*-factors;

2. Re-evaluated AEB 98-03 with multi-group; and

3. Numerical Integration.

When DG-1389 was released for public comment, the staff also requested comment on a draft staff technical assessment titled, Technical Assessment of Hold-up and Retention of Main Steam Isolation Valve Leakage within the Main Steam Lines and Main Condenser (ML20085J042). In addition, DG-1389 includes reference to a memorandum providing further guidance on performing the re-evaluated AEB-98-03 with multi-group and numerical integration (i.e., Technical Bases for Draft RG 1.183 Revision 1 (2021) Re-evaluated AEB-98-03 Settling Velocity Method, the Multi-Group Method, and the Numerical Integration Method. July 2021; ML21141A006). As noted in DG-1389, Appendix A, Section A-5.6.2, both the re-evaluated AEB-98-03 with multi-group and the numerical integration method should assume a log-normal aerosol diameter distribution with an aerodynamic mass median diameter (AMMD) of 2.0 micrometers and geometric standard deviation of 2.0. Equation A-5 of DG-1389 is used to determine the median aerosol diameter in the log-normal distribution based on these parameters. The aerosol size distribution parameters specified in DG-1389 are supported by data included in NEA/CSNI State-of-the-Art Report on Nuclear Aerosols, II-8 APPENDIX 2

NEA/CSNI/R(2009)523. The analysis of experimental results described in NEA/CSNI/R(2009)5 state that it would seem realistic for aerosols in the hot leg to comprise a near-lognormal population of particles with AMMD around 1 m or less and standard deviation around 2 (Section 9.4) and for the containment source term aerosol mass distribution might be described by a log-normal function characterized by a geometric standard deviation around 2.0. Aerosol Mass Median Diameters (AMMD) of about 3.5 - 4.0 m was measured (Section 6.2). As a result, a recent NRR reevaluation of the AEB-98-03 referenced the use an AMMD of either 1.0 m (for aerosols originating from the RCS) or 3.0 m (for those from containment) with a geometric standard deviation of 2.0 in either case (ML21141A006). The use of an assumed aerosol distribution is not needed for licensees using the SAND2008-6601 approach because recommended removal coefficients are provided in the report. Regardless of the main steam line fission product deposition approach, Section A-5.1 states that the source of MSIV leakage should be assumed to be the containment drywell activity concentration and that the containment boundary should be assumed to extend to the MSIVs.

The DPO Appeal Panel notes that the draft guidance provided in DG-1389 is a significant enhancement compared to the individual case basis guidance currently contained in RG 1.183. However, the fission product main steam line deposition methods described in DG-1389 are largely based on research findings developed over a decade ago, notably the NEA/CSNI/R(2009)5 State-of-the-Art report on nuclear aerosols. The multi-group method (which uses a Monte Carlo approach to determining deposition parameters) and the numeric integration approach, essentially solve the basic equations used in AEB-98-03 with an updated aerosol diameter distribution based on the 2009 NEA State of the Art report NEA/CSNI/R(2009)5.

23 https://www.oecd-nea.org/upload/docs/application/pdf/2021-03/csni-r2009-5.pdf II-9 APPENDIX 2

III. APPENDIX 3: Reduced Containment Pressure Sensitivity Study The DPO Appeal Panel believes that certain design and accident response factors that are not currently credited in DBA analysis could reduce both offsite and control room calculated dose results. The most significant factor considered by the DPO Appeal Panel was the assumption that containment pressure remains at the peak value of 45 psig for the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of the DBA (ML20140A070). This assumption is consistent with RG 1.183, Appendix A, Section 3.7, which states The primary containment (i.e., drywell for Mark I and II containment designs) should be assumed to leak at the peak pressure technical specification leak rate for the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.

While the DPO Appeal Panel believes that this assumption provides additional conservatism consistent with the intent of the DBA analysis, for the purposes of a backfit evaluation, a more realistic expectation for the plant response could be considered. The DBA accident response described in the UFSAR (i.e., UFSAR Figure 14.6-22, Containment Pressure Response DBA LOCA Power Uprate Conditions), shows that the peak containment pressure is reached within the first 10 seconds of the DBA. Containment pressure then drops to approximately 30 psig for the next several minutes, and then decreases to and remains below approximately 15 psig for the remainder of the DBA. The DPO Appeal Panel also considered that the onset of the in-vessel release phase begins at 30 minutes, consistent with RG 1.183, Table 4. Therefore, containment pressure would be expected to be less than approximately 15 psig when significant fission products release from the core begins during the in-vessel release phase.

In order to assess the potential dose reduction associated with use of a more realistic containment pressure response, the DPO Appeal Panel requested the analyst who previously performed sensitivity studies in support of the Directors Decision to perform an additional sensitivity case. The analysis assumed use of the DG-1389 multigroup method, a removal coefficient of 20 hr-1, and credit for deposition in only two main steam lines, but reduced MSIV and containment leakage by a factor of 0.6 for the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to reflect an assumed containment pressure of 15 psig versus 45 psig (i.e., the square root of 15/45, or approximately 0.6). The results of this sensitivity, as well as the initial case, are provided below:

Sensitivity of Calculated Control Room Dose to Containment Peak Pressure Assumed Peak Number of Main Steam Iodine Removal NRR Containment Steam Lines Deposition Coefficient Calculated Pressure Credited for Model Total Control (0-24 hours) Deposition Room Dose (rem TEDE)

Re-evaluated 45 psig 2 AEB-98-03 with 20 hr-1 10.42 multi-group Re-evaluated 15 psig 2 AEB-98-03 with 20 hr-1 5.63 multi-group II -1 APPENDIX 3

Although this sensitivity case assumed a recirculation line break vice a more limited steam line break, the results indicate that use of a more realistic containment pressure accident response during a DBA LOCA can reduce calculated control room dose by approximately 45%. While this sensitivity was done only for scoping evaluation purposes, it does indicate that application of more realistic assumptions to the DBA analysis can significantly reduce dose impacts.

II -2 APPENDIX 3

IV. APPENDIX 4: Compliance Backfit Documented Evaluation COMPLIANCE BACKFIT DOCUMENTED EVALUATION ORDER TO RE-EVALUATE AND SUBMIT LICENSE AMENDMENT REQUEST EXELON GENERATION COMPANY, LLC JAMES A. FITZPATRICK NUCLEAR POWER PLANT DOCKET NO. 50-333 CONTENTS I. INTRODUCTION II. BACKGROUND III. PROPOSED BACKFITTING IV. ADEQUATE PROTECTION EXCEPTION APPLICABILITY V. COMPLIANCE EXCEPTION APPLICABILITY A. Known and Established Standard B. Errors C. Summary VI. COST CONSIDERATIONS VII. CONCLUSION IV-1 APPENDIX 4

I. INTRODUCTION The U.S. Nuclear Regulatory Commission (NRC) proposes to issue an order to Exelon Generation Company, LLC (licensee), the licensee for the James A. Fitzpatrick Nuclear Power Plant (FitzPatrick), to require the licensee to resubmit an alternate source term (AST) license amendment request (LAR). The NRC issued the licensee an AST license amendment in 2020 that was based, in part, on errors in the supporting analyses and safety evaluation that were identified by the NRC in 2023. These errors result in the NRCs lack of reasonable assurance of the licensees compliance with 10 CFR 50.67, Accident source term. The order would require the licensee to correct and resubmit the AST license amendment to demonstrate compliance with 10 CFR 50.67. The order would constitute the imposition of a staff position interpreting the Commission's regulations that is different from a previously applicable staff position and would result in the licensee changing systems, structures, components (SSCs) or the design of FitzPatrick or procedures necessary to operate FitzPatrick. Therefore, the order would meet the definition of backfitting in 10 CFR 50.109, Backfitting. This documented evaluated examines the proposed backfitting action and the basis for justifying the backfitting action with the compliance exception to the requirement to perform a backfit analysis.

II. BACKGROUND On August 8, 2019, the licensee submitted a license amendment request for approval of an alternate source term for FitzPatrick in accordance with 10 CFR 50.67 (ML19220A043). On July 21, 2020, the NRC staff approved the AST license amendment and issued its safety evaluation to support issuance of the amendment (ML20140A070). On January 19, 2021, two members of the NRC staff filed a differing professional opinion (DPO) on the issuance of the AST license amendment to the licensee (ML21042B867; non-public). The main concern in the DPO was that the AST license amendment did not comply with the NRC regulations contained in 10 CFR 50.67. The Office of Nuclear Reactor Regulation (NRR) Office Director issued a decision on November 17, 2021, stating that the licensee-approved analysis of record was acceptable and met the regulatory requirements (ML21236A254; non-public). The DPO submitters appealed the NRR Office Directors Decision to the NRCs Executive Director for Operations (EDO) on January 7, 2022 (ML22039A062 case file; non-public). Following receipt of the submitters appeal, the NRR Office Director reviewed the appeal package and issued their February 3, 2022, statement of views regarding the appeal of DPO-2021-001 that reaffirmed the initial Directors Decision (ML22031A053; non-public). In July 2022, a DPO Appeal Panel was formed to assess the DPO concerns. In the course of conducting its review of the LAR, the NRC staffs safety evaluation, and other supporting information, the DPO Appeal Panel identified several errors that collectively result in a lack of reasonable assurance of the licensees compliance with 10 CFR 50.67. Specifically, the DPO Appeal Panel concludes that the errors result in a failure of the licensees analysis of applicable design basis accidents to demonstrate with reasonable assurance that adequate radiation protection for the control room is provided to permit access and occupancy without personnel receiving radiation exposure in excess of 5 rem TEDE, contrary to 10 CFR 50.67(b)(2).

One option for the NRC to address the noncompliance with 10 CFR 50.67(b)(2) is to issue an order to require the licensee to correct the errors and resubmit the AST LAR for NRC approval.

IV-2 APPENDIX 4

III. PROPOSED BACKFITTING As relevant here, backfitting is defined in 10 CFR 50.109 as the imposition of a regulatory staff position interpreting the Commission's regulations that is different from a previously applicable staff position and results in the modification of or addition to the SSCs or design of a nuclear power plant, or the procedures or organization required to operate a nuclear power plant. The proposed order must meet this definition to constitute backfitting.

The order would be issued to a holder of an operating license for a nuclear power plant under 10 CFR Part 50. Thus, the order would affect an entity within the scope of 10 CFR 50.109.

The order would reflect a change in the NRC staffs position on whether the licensee complies with 10 CFR 50.67. When it issued the AST license amendment in 2020, the NRC determined that the NRCs issuance and the licensees implementation of that license amendment provided reasonable assurance that the licensee would comply with the 10 CFR 50.67. Even though the licensee has complied with the license amendment, the NRC staff can no longer conclude that it has reasonable assurance that the licensee is complying with 10 CFR 50.67. Thus, the NRC has changed its position on whether the licensee complies with 10 CFR 50.67.

Issuance of an order would impose on the licensee certain requirements that reflect the staffs changed position. The NRC no longer has reasonable assurance of the licensees compliance with 10 CFR 50.67, so the order would require a new LAR and supporting analyses that show that a revised AST license amendment would demonstrate reasonable assurance of compliance with 10 CFR 50.67. Thus, the NRC would be imposing its changed staff position on the licensee.

Correcting the errors and resubmitting the AST LAR would result in the issuance of an AST license amendment, a denial of the LAR, or the withdrawal of the LAR. In each case, the licensee would be required to modify SSCs or the design of FitzPatrick or the procedures or organization required to operate FitzPatrick.

Therefore, the imposition of a changed NRC staff position through an order directing the nuclear power plant licensee to correct and resubmit the AST license amendment would result in the modification of SSCs or design of a facility or procedures or organization required to operate a facility. The proposed order would meet the definition of backfitting under 10 CFR 50.109.

IV. ADEQUATE PROTECTION EXCEPTION APPLICABILITY The DPO Appeal Panel concludes that the issues it identified with the 2020 AST license amendment do not present a condition of undue risk to public health and safety or the common defense and security. For the design basis accident analyzed for the FitzPatrick AST LAR, significant margin to the offsite criteria still remains (i.e., calculated offsite doses are at least a factor of 20 below the 10 CFR 50.67 dose criteria of 25 rem TEDE). The technical issues identified in the analysis should not result in exceeding this level of margin for the offsite dose criteria. The main design basis accident dose impact is that the licensees analysis supporting the AST LAR fails to demonstrate with reasonable assurance that the 10 CFR 50.67 control room dose criteria of 5 rem TEDE is met. Although the DPO Appeal Panel finds that the issues substantiated by its review could increase calculated control room dose for a postulated design basis accident by approximately a factor of two, this issue does not represent an adequate IV-3 APPENDIX 4

protection issue for the operators. The DPO Appeal Panel believes that, during an actual accident, control room doses could be mitigated by multiple defense-in-depth actions such as the use of portable FLEX equipment, severe accident management guidelines, and emergency plans.

Therefore, the proposed order is not necessary to ensure that FitzPatrick provides adequate protection to the public health and safety and is in accord with the common defense and security, or to define or redefine the level of protection to the public health and safety or the common defense and security deemed to be adequate. As a result, the NRC cannot invoke the adequate protection exception to the requirement to perform a backfit analysis to justify the proposed backfitting action.

V. COMPLIANCE EXCEPTION APPLICABILITY Under Management Directive 8.4, Management of Backfitting, Forward Fitting, Issue Finality, and Information Requests, the compliance exception is used to justify backfitting actions to address situations in which the licensee has failed to meet known and established standards of the Commission because of omission or mistake of fact. The following criteria from Management Directive 8.4 define when the compliance exception can be invoked:

(a) The NRC staff, whether by its own error or by licensee or third-party error or omission at or before the time of its determination that a known and established standard of the Commission was satisfied: (1) incorrectly perceived facts, (2) performed or failed to recognize flawed analyses, or (3) failed to properly draw inferences from those facts or analyses, as judged by the standards and practices that were prevailing among professionals or experts in the relevant area at the time of the determination in question, and (b) Those facts, analyses, or inferences have now been properly perceived, performed, or drawn.

(c) The error or omission may have been committed by any involved party and must be traced to (i) The original LB or to a change to the LB, (ii) The regulations reflected in the LB at the time of implementation that were applicable to the licensee whose LB is at issue, and (iii) Any standards and practices in existence at the time the original determination was made.

A. Known and Established Standard The licensee needs to comply with 10 CFR 50.67. Specifically, under 10 CFR 50.67(b)(2), the licensee needs to provide an LAR and supporting analysis that demonstrates with reasonable assurance that [a]dequate radiation protection is provided to permit access to and occupancy of the control room under accident conditions without personnel receiving radiation exposures in IV-4 APPENDIX 4

excess of 0.05 Sv (5 rem) total effective dose equivalent (TEDE) for the duration of the accident.

This requirement existed and was known and established at the time that the NRC issued the FitzPatrick AST license amendment. The NRC issued the final rule establishing 10 CFR 50.67 on December 23, 1999 (64 FR 71990). On July 21, 2020, the NRC staff approved the FitzPatrick AST license amendment, thereby approving the licensees method of compliance with 10 CFR 50.67.

The NRC has substantial experience applying the requirements of 10 CFR 50.67 in a consistent manner. Following issuance of the 1999 final rule, the NRC issued more than 70 AST-related license amendments before issuing the FitzPatrick AST license amendment in 2020. Guidance supporting 10 CFR 50.67 reviews is contained in Regulatory Guide (RG) 1.183, Revision 0, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors (issued July 2000); NUREG-0800, Standard Review Plan, Section 15.0.1, Radiological Consequence Analyses Using Alternative Source Terms, Revision 0 (issued July 2000); and RIS 2006-04, Experience with Implementation of Alternative Source Terms (issued March 7, 2006). Regarding aerosol settling, RG 1.183 states that reduction in fission product by deposition in the main steam line can be credited, but the reduction in fission product concentration will be evaluated on an individual case basis. RIS 2006-04 further clarified the RG 1.183 guidance by noting that reference to AEB 98-03, Assessment of Radiological Consequences for the Perry Pilot Plant Application using the Revised (NUREG-1465) Source Term, was acceptable, but licensees should provide appropriate justification that the assumptions used are applicable to their plant. The licensee utilized the AEB 98-03 method, as did several other applicants referenced in the NRC staffs AST safety evaluation.

However, before the NRC staffs issuance of the FitzPatrick AST license amendment, AEB 98-03 was generally applied in a conservative manner by including additional limitations such as: limiting the time frame that deposition was credited, not crediting removal by drywell spray and deposition simultaneously, crediting the main condenser for fission product deposition, or using more conservative aerosol settling velocities. In the FitzPatrick AST license amendment, these additional conservatisms were not applied, and non-conservatisms associated with the AEB 98-03 methodknown at that timebecame more significant in the calculation of dose consequences. Further, as evidenced by the NRC staffs requests for additional information (RAIs) associated with aerosol settling and selection of the limiting break location, the staff was aware of how these issues had been considered during past reviews (see ML20090E279). Although the issue associated with outboard main steam line isolation valve (MSIV) packing leakage may be somewhat unique to FitzPatrick, the licensing basis was clearly described in Section 9.19 of the licensees UFSAR and Technical Specifications.

B. Errors The DPO Appeal Panel identified three sets of NRC staff errors through the incorrect understanding of facts, the failure to recognize flawed analysis, and the failure to draw direct inferences from facts or analysis:

• Incorrect understanding of facts: The NRC staff failed to recognize that the interface between the packing and stem in the outboard MSIV was a potential fission product IV-5 APPENDIX 4

leakage path to the environment as described in the FitzPatrick UFSAR.

Furthermore, the FitzPatrick TS Bases and UFSAR description of the MSIV leakage control system, whose removal was approved by the FitzPatrick AST license amendment, included a function to direct outboard MSIV packing leakage to a filtered release path. However, these facts were not understood by the reviewers and consequently, following removal of the MSIV leakage control systems, the potential dose impact from the outboard MSIV packing leakage was not addressed in the staffs safety evaluation. Fission product release from this location would increase calculated control room dose.

• Failure to recognize flawed analysis: The NRC staff failed to recognize a main steam line break as the limiting loss of coolant accident location based on a misinterpretation of sensitivity results provided by the licensee. Assuming the main steam line as the break location would reduce main steam line fission product deposition and increase the calculated control room dose.

• Failure to draw direct inferences from facts or analysis: The NRC staff offset non-conservatisms associated with fission product deposition modeling, including credit for both drywell spray and main steam line deposition, with credit for the ability of the condenser to remove fission products prior to their release to the environment.

However, the licensee did not request or credit condenser fission product deposition in their AST LAR. Further, when the staff credited the fission product removal capabilities of the main condenser, they failed to infer that such credit should be accompanied by appropriate regulatory controls. These regulatory controls include procedures to align a flowpath to the main condenser and testing and equipment qualification of components needed to align the flowpath. The DPO Appeal Panel believes implicitly crediting the mitigation capability of the condenser by the NRC staff without appropriate regulatory controls is an error.

The FitzPatrick AST license amendment included the staffs safety evaluation, which contained these errors. Thus, these errors occurred at the time the NRC staff issued the amendment.

The errors would have been deemed errors as judged by the standards and practices that were prevailing among professionals and experts when the FitzPatrick AST license amendment was issued:

• Outboard MSIV packing leakage: The FitzPatrick UFSAR recognizes the outboard MSIV packing as a fission product release point during a design basis accident.

Prior to issuance of the license amendment, leakage from this location was controlled by the MSIV leakage control system. This system directed leakage to a filtered release path, and FitzPatricks Technical Specifications included requirements for system operability. The AST license amendment allowed removal of this system but did not address how leakage would be handled. The need to address MSIV packing leakage has been included in regulatory guidance since IV-6 APPENDIX 4

1976 (e.g., RG 1.96, Design of Main Steam Isolation Valve Leakage Control Systems for Boiling Water Reactor Nuclear Power Plants, Revision 1, Section B).

• Limiting break location: RG 1.183 includes guidance in Appendix A to consider leaks up to a double-ended rupture of the largest reactor coolant system pipe.

Contrary to this guidance, the licensee selected a recirculation line break, representing a smaller break size and less limiting location. The NRC staff recognized that the licensee did not select the more limiting main steam line pipe when issuing an RAI on this topic (ML20090E279); however, in evaluating the licensees response, the staff failed to recognize that the licensee did not provide a sensitivity case that reflected actual design basis conditions for the AST LAR.

• Aerosol removal uncertainty offset by credit for the main condenser: The NRC staff has recognized since at least 2005 that the methods in AEB 98-03 underpredicted aerosol settling in main steam piping. Further, research documented in the 2008 report SAND2008-6601, Analysis of Main Steam Isolation Valve Leakage in Design Basis Accidents Using MELCOR 1.8.6 and RADTRAD, highlighted potential non-conservatisms associated with the use of AEB 98-03 and simultaneous credit for drywell spray and main steam line deposition as fission product removal mechanisms. The NRC staff recognized these issues when issuing an RAI to the licensee (ML20090E279), but despite sensitivity studies provided by the licensee demonstrating that calculated control room dose could exceed the limits of 10 CFR 50.67, no further action was taken by the staff to resolve these issues.

These errors could have been readily corrected using practices and knowledge available when the NRC issued the FitzPatrick AST license amendment. The correct analysis methods have been known to experts for over a decade and, with the exception of addressing MSIV packing leakage, were demonstrated by the licensee in its response to staff RAIs (e.g., ML20090E279).

MSIV packing leakage was previously controlled by the licensee using the leakage control system. The licensee could either reinstall a system to control MSIV packing leakage or consider such leakage within its design basis analysiseither approach is well within the standards and practices that were prevailing among relevant professionals and experts at the time of the issuance of the FitzPatrick AST license amendment.

If the errors were recognized at the time the FitzPatrick AST license amendment was issued, then the NRC staff would likely have taken action to address the errors before issuing the license amendment.

C. Summary The discussion above shows that each element of the compliance exception is met, so it is appropriate to invoke the compliance exception to performing a backfit analysis to justify issuing the order to the license.

VI. COST CONSIDERATIONS IV-7 APPENDIX 4

Management Directive 8.4 directs the NRC staff to include at least some consideration of costs in its documented evaluation for a compliance backfit. The instant backfitting action would be the issuance of an order for the licensee to reanalyze the design basis accident analysis for the AST in order to address identified errors and resubmit the license amendment for NRC staff review. Reperforming the design basis accident analysis would need to address MSIV packing leakage not previously considered in the original AST LAR. Although the costs of this reanalysis cannot be quantified by the NRC staff, reanalysis to consider a more limiting break location, more accurate aerosol deposition removal modeling, and credit for the condenser can be readily accomplished by the licensee based on the previous sensitivity studies submitted in response to a staff RAI (see ML20090E279). If additional controls for the main condenser are needed, the licensee has developed such controls for other sites (e.g., LaSalle), including environmental qualification of necessary equipment, procedures, and regulatory controls (e.g., Technical Specification or other technical requirement controls). Addressing MSIV leakage is subject to greater cost uncertainty, and the licensee previously had capabilities to control such leakage at FitzPatrick but removed the system following approval of the AST LAR.

VII. CONCLUSION The NRC has determined that the proposed order constitutes backfitting and is needed to ensure compliance with 10 CFR 50.67. The order would require the licensee to correct and resubmit the AST license amendment to demonstrate compliance with 10 CFR 50.67. This action is necessary because the NRC committed errors in its review of the licensees previous AST LAR, and the NRC would not have issued that AST license amendment had it not committed these errors. The proposed order meets the criteria to invoke the compliance exception to performing a backfit analysis. As such, the NRC has not prepared a backfit analysis to support the proposed order.

IV-8 APPENDIX 4

V. APPENDIX 5: DPO-2021-01 Appeal Panel Questions and NRR Answers (Round 1)

1. Treatment of Conservatisms The Directors decision notes that particulate removal coefficient values used in the calculation may not have been the most conservative. In RG 1.183, specific guidance is provided for selecting removal coefficients. Given that particulate removal coefficients represent actual physical processes, rather than parameters that can be selected at the discretion of the analyst to establish design safety margins, how did you justify that the removal coefficients identified in the licensees calculation of record are consistent with the current state of knowledge as documented in NRC research studies, consensus standards or methods?

Generally, if a licensee departs from the guidance in RG 1.183, how does the staff ensure and document that the use of such parameters meet the requirements of 10 CFR Part 50, Appendix B (e.g., design control) and/or peer-reviewed scientific literature? The DPO Appeal panel notes that in some cases both the submitters and the initial DPO panel noted that the licensees selected removal coefficient values were significantly greater than typical values applied to these analyses.

NRR Answer:

Answer to the question in the first paragraph: In the Directors Decision Memo, NRR acknowledged the particulate removal coefficients used in the FitzPatrick analysis for the drywell and the main steam line (MSL) volumes were not conservative and were greater than typical values.

RG 1.183 offers two options for calculating particulate removal coefficients in the drywell. It refers to the guidance contained in (Section C.iv) SRP Section 6.5.2 (ML070190178) and it also references NUREG/CR5966 (ML063480542).

To this date, NRC does not have a specific reduction criteria or any approved guidance licensees can use to address the impact of drywell sprays on deposition in the main steam lines (MSLs) (i.e., calculate particulate removal coefficients in the MSLs). Historically, licensees have mirrored methodologies approved at other plants to inform their own submittals. This practice is captured in RIS 2006-04, Experience with Implementation of Alternative Source Terms, dated March 7, 2006, where NRC stated AEB 98-03 (ML011230531) provides an acceptable method for calculating particulate removal coefficients in the MSLs. RIS 2006-04 does remind licensees that AEB 98-03 is specific to a particular plant (i.e., Perry) and that licensees should use plant- specific assumptions and appropriate justification that are applicable to their design.

Since there is no accurate and precise way of calculating particulate removal coefficients in V-1 APPENDIX 5

the MSLs, industry and the NRC have relied heavily on best estimates. As documented in RIS 2006-04, the AEB 98-03 method uses a single value for MSL aerosol settling velocity to evaluate the deposition of aerosol particles which, in reality, contain a wide range of sizes and weights. Using a single value overestimates deposition since the rate of deposition in the MSLs is expected to decrease as the leakage progresses through the lines, as the larger and heavier aerosols would have already settled out of the MSLs in upstream sections of piping. To address this deficiency, licensees (including FitzPatrick, Clinton, Limerick, and LaSalle) have employed a 20-group probabilistic distribution of particle sizes in the MSL volumes (conservatively deviating from AEB 98-03 based on newly discovered information, at the time of the submittals).

Therefore, the NRC staff approved FitzPatricks analysis based on current state of knowledge and consistent with existing methodologies found acceptable in generic communications.

Answer to the question in the second paragraph: At times, licensees will propose different source terms and transport models from those in the regulatory guidance. Some departures from the guidance will have a negligible impact on the result while others are known to have a large impact. The staff, in general, knows how a parameter impacts the result. If a licensee presents a method or model that departs from the regulatory guidance, the staff will review it, as the guidance presents an acceptable way, not the only way, to meet the regulation. During this review, the staff will go through the regulatory process, such as by hosting public meetings and submitting requests for additional information. If necessary, the staff will also perform literature review, confirmatory analyses, audits, and site visits. If the proposed change is technically challenging or the NRR staff does not have the time to review it, NRR will formally reach out to the technical experts in RES through the user-need process. If the issue is outside the expertise of RES staff, the Agency then typically goes to the national laboratories or contractors. This process is handled by RES through contracts. The results of these efforts are typically captured in reports which are then reviewed by RES and NRR staff. After reviewing the report, the staff will reengage with the licensee to discuss the acceptability of the proposed method in light of the results of the research. As it pertains to this DPO, FitzPatrick followed RG 1.183. They deviated in their calculation for elemental removal coefficient and after performing confirmatory calculations, the NRC staff determined this deviation had a negligible impact on dose results.

2. Methods to Assess Functionality of Non-Safety SSCs Credited in Safety-Related DBA Calculations and Addressing Uncertainty in the Licensing Basis Record The Directors decision noted that the NRC did not credit the PCS or the main condenser to make the regulatory finding of adequate protection. However, upon review, it does appear that the release pathways to the environment described in the staffs SER did assume functionality of certain aspects of non-safety equipment in the power conversion system (PCS), main condenser, and surrounding structure. Are we reading this right?

Since the staff did not impose any additional licensing requirements on these non-safety systems, how was potential failure of these systems assessed/treated in the staffs evaluation? Should these non-safety SSCs be considered part of the licensing basis since V-2 APPENDIX 5

they supported the staffs consideration of uncertainty in the risk-informed decision-making (RIDM) process used for the LAR review?

The Directors decision notes that the staff only considered the capability of the PCS/condenser following a safe shutdown earthquake. Given that the regulatory finding was informed by the risk insights, did the staff consider a broader range of risk insights to evaluate these non-safety systems beyond the SSE? For example, did the staff evaluate severe accident insights from Fitzpatrick or similar facilities when deciding what challenges to evaluate (in addition to SSE capability) and the availability of these SSCs during a range of severe accident scenarios?

For the Fitzpatrick AST LAR review, what was the staffs basis for why a different approach was selected other than that described in the approved BWROG Topical Report (NEDC-31858P-A)? What were the significant differences between the staffs approach for Fitzpatrick and this Topical Report?

The Directors decision notes that the staff performed an assessment of the seismic ruggedness of the power conversion system (PCS) and main condenser to achieve high confidence that these systems will remain available after a safe shutdown earthquake (SSE) for fission product dilution, holdup, and retention. The decision also notes that that this was used, in part, to balance any uncertainties or approved non-conservatisms. This implies that the assessment for PCS and the condenser was needed to support the staffs regulatory finding, and as such should be documented appropriately in the licensing basis.

The DPO Appeal panel notes that integration of risk information into decision-making practices is a key feature of NRRs Risk-Informed Decision-Making Plan (RIDM, ML18116A023). In addition, consideration of uncertainties is an important aspect of RIDM as described in RG 1.174, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, and included in the Commissions PRA policy statement (60 FR 42622). For deterministic reviews (which the Directors decision notes the NRC staff based their regulatory finding of adequate protection and the licensees deterministic analysis) the staff normally applies conservative selection of parameters and use of safety margins. Could additional background be provided as to why consideration of non-safety SSCs in a risk-informed manner should not be considered part of the staffs regulatory decision basis, and therefore part of the licensing basis, for this issue?

As a related issue, the Directors decision notes that the PCS/main condenser analysis provided high confidence. Why was a high confidence level of functionality of these systems needed if they are non-safety related and not credited for the DBA analysis (for example, non-safety SSCs are typically assumed to failure in the worst- case failure mode for safety-related assessments)?

NRR Answer:

FitzPatricks analysis assumes the turbine building is lost to a seismic event and that the turbine stop valves (TSVs) are the release point of radiation to the environment. This assumption is consistent with RG 1.183 and is supported by sound engineering judgement V-3 APPENDIX 5

given that all components upstream the TSVs are seismically qualified, safety-related and, therefore, not assumed to fail. The NRC staffs approval of FitzPatricks analysis was based on this information and did not account for all or any portion of the PCS/main condenser being available for deposition.

The pathways described in Section d) of the Attachment to the Directors Decision Memo are all part of the turbine building, which the analysis assumes is lost. Assuming the turbine building is lost yields more conservative results than assuming radiation is released through these pathways. In addition, if we were to consider these pathways as available for radiation release, then we could assume the turbine buildings volume is available for deposition of activity which will yield more realistic results (considerably lower than current results) but would be contrary to existing regulatory positions.

The assessment of the seismic ruggedness of the power conversion system (PCS) and main condenser was not needed to support the staffs regulatory finding. FitzPatricks analysis did not request credit for PCS/main condenser activity deposition. For this reason, this credit is not part of their licensing basis and adoption of the approach documented in the BWROG Topical Report (NEDC-31858P-A) was not needed.

The high confidence afforded by the seismic ruggedness of the PCS/main condenser balanced the uncertainty in the calculation. Every estimated value carries with it uncertainty. In these analyses, the models used to calculate particulate removal coefficients in the MSL carry some of the most uncertainty. Deposition in the MSL is a physical phenomenon which is guaranteed to happen. To date, there is no approved test to provide a precise and accurate model to calculate it. The model used by the licensee was developed for Perry, was approved by the NRC, and deemed acceptable in RIS 2006-04. However, knowing this model wasnt an exact representation of deposition in the MSL, the NRC staff described the seismic ruggedness to address any uncertainties in the model. Given that the purpose of the PCS/main condenser reference was to address any uncertainties, the NRC staff did not find it necessary to consider seismic events beyond the safe shutdown earthquake.

3. Impact of Removal of MSIV Leakage Collection Systems It is our understanding that the Fitzpatrick LAR allowed the licensee to remove the MSIV leakage collection system (LCS), but no longer assumed that leakage via the MSIV valve packing and MSIV mechanical joints was a credible leak path (which was the leakage path that the LCS was intended to mitigate). What was the basis for the assumption to no longer consider MSIV packing or mechanical joint leakage in the DBA consequence analysis? How is this basis captured/documented in the licensing basis going forward?

How was operating experience factored into the determination that leakage at the MSIVs was not credible? If instead this leakage path is considered to be bounded by another leakage path, how are the factors that support this conclusion maintained going forward (i.e., the leakage path at the MSIVs to the environment may be currently bounded, but what will ensure that future changes do not undermine this conclusion)?

V-4 APPENDIX 5

Note - The Directors decision does note certain factors that were considered in evaluating the MSIV leakage paths to the environment, including operation of the standby gas treatment system, capping of the leakage collection system LCS piping, etc. While this leak path may be bounded by the TSV leak path based on the configuration at the time of the license amendment, what will ensure that this will remain the case over time (and that this decision basis is clearly contained in the licensing basis)?

NRR Answer:

An alternative source term (AST) analysis provides a licensee with a more realistic and, therefore, more accurate result of dose, when compared to previous analyses used (e.g.,

TID- 14844 (ML021750625), RG 1.3 (ML003739601) and RG 1.4 (ML003739614)). It is favorable for a licensee to get approval to use an AST because, if the analysis meets the regulatory requirements in 10 CFR Part 50.67 (i.e., doses within the limits) without crediting the main steam isolation valve (MSIV) leakage collection system (MLCS), then NRC allows removal of these systems from TS and from service. Refer to Section b) of the Director Decision Memo Attachment for additional information on how MLCSs can result in potentially higher offsite doses resulting from its use and its high failure rate in a variety of failure modes. Since the licensees analysis met the regulatory requirement, the licensee was able to declassify the MLCS to non-safety related. The licensee was also able to remove the MLCS from their technical specifications. In addition, the licensee performed a safety-related modification to isolate and cap the MLCS. If a direct release from the stem of the outboard MSIVs to the environment were to occur, it would release in the reactor building, which has been evaluated for the safe shutdown earthquake (SSE) and it is assumed to remain structurally intact. Any leakage in the reactor building gets processed by the standby gas treatment (SBGT) system (assumed to be in operation 20 minutes after a loss of coolant accident (LOCA)). This results in doses to the control room that are less than the evaluated scenario of MSIV leakage through the seat of the valves and all the way to the TSVs.

Leakage from the stem of the MSIVs could occur but as mentioned in the paragraph above, it would be processed by the SBGT system, which is a safety-related system and required to be operable per the technical specifications (TS 3.6.4.3).

The licensee would need NRC approval to change the configuration described above.

4. Quality of the Calculation of Record The Directors decision documented the basis and reasoning for concluding that the issues raised by the submitters do not constitute safety issues and there is reasonable assurance that the facility can be operated in a manner consistent with regulatory requirements. The DBA consequence calculation is a safety-related calculation that must meet the requirements of 10 CFR 50, Appendix B (as implemented through the licensees approved Quality Assurance Plan). In light of the issues noted in the initial DPO panel report and your review, does the staff believe that the licensees current calculation of record accurately reflects the design and operation of the plant, the current state of practice in V-5 APPENDIX 5

modeling radionuclide behavior, and is consistent with research results and operating experience applicable to the facility?

In the Directors decision, the Director agreed with certain concerns related to the identification of the limiting break location [i.e., recirculation line vs main steam line (MSL)].

The Director noted that the staff should have provided a detailed evaluation but that the staffs independent evaluation confirmed that doses remained below regulatory limits. How is the assessment of the MSL break currently documented in the licensing basis for the facility, including the licensees and staffs calculation of record?

NRR Answer:

Answer to the question in first paragraph: Yes, the NRC staff has reasonable assurance that the licensees revised calculation of record (Rev 1) reflects the design and operation of the plant, the current state of practice in modeling radionuclide behavior, and is consistent with research results and operating experience applicable to the facility. While newer methods have been developed that may be more accurate, they would also reduce conservatisms in the analysis and yield more realistic results.

Answer to the question in second paragraph: The licensees analysis postulates a traditional large break LOCA (double-ended guillotine break of the recirculation line). In a section of FitzPatricks submittal, the licensee stated: Although postulating a main steam line break in one steam line inside the drywell would maximize the dose contribution from the MSIV leakage, the steam line break is not a credible event during a LOCA, because the Seismic Class 1 main steam piping up to the TSVs is designed to withstand the SSE (Ref.

9.62). In reviewing this statement, the NRC staff requested the following information from the licensee via ARCB-RAI-3: Please provide additional information to justify that assuming a recirculation line rupture instead of a main steam line rupture is consistent with the guidance from RG 1.183 that assumptions should be selected with the objective of maximizing the postulated radiological consequences. The licensees response included an assessment of the dose consequence modeling all four MSLs (vs only two MSLs volumes available for activity deposition) with one broken MSL and a failed open MSIV.

The results indicated that the dose consequences in the original analysis were bounding.

The scenario provided in the licensees response assumes multiple failures (i.e., the recirculation break, an MSIV stuck open, an MSL break) and are overly conservative. The licensees current analysis of record provides the bounding scenario assuming the most limiting failure.

5. Clarity of Staff Position and Review Criteria for Design Basis Accident Dose Criteria The Directors decision noted that there was a lack of clarity and transparency with regard to the basis for the staffs conclusions and internal updates that would include the importance of clearly and transparently articulating the basis of the staffs conclusion were recommended. In light of this, and that the methods used for Fitzpatrick may have been applied to other license amendments, could you clarify the staff position with regard to the current, approved, revision to RG 1.183 and associated SRP sections? Do these still V-6 APPENDIX 5

represent the staff positions for DBA consequence analysis? If not, or if other methods are now acceptable to the staff, how were these new positions communicated to internal and external stakeholders (e.g., issuance of Generic Communications, ISGs, notification that new review criteria would be used when the notice of no significant hazards is issued for the amendment request)? If the staff position was evolved since issuance of RG 1.183, Revision 0, was there any opportunity for stakeholder comment (both by members of the public and industry) in a manner equivalent to the process used for the approved version of RG 1.183?

In the Directors decision, the term regulatory finding of adequate protection is used several times. For regulations that were promulgated to meet an adequate protection standard, what criteria does the staff use to determine that adequate protection is met?

What guidance documents this process? For regulations that are not specifically provided for adequate protection, but were instead imposed on a different basis (such as a cost-beneficial safety improvements per 50.109), what guidance is available for staff to consistently apply regulatory acceptance criteria consistent with the intent of the associated regulation (e.g., providing a level of safety above the adequate protection standard)?

NRR Answer:

Answer to the questions in first paragraph: Yes, the guidance contained in the current version of RG 1.183 continues to represent the staffs position. In the draft revision of RG 1.183 (DG-1389), the NRC staff will provide additional options for performing these calculations, however, to this date, the currently approved version will not be rescinded.

DG-1389 was issued for public comments. Regulatory guides provide one acceptable way to meet regulatory requirements.

The NRC staff remains open to accept methodologies outside those documented in regulatory guides, as long as they provide reasonable assurance of adequate protection.

Such different methodologies would be communicated to stakeholder through the public licensing process.

Answer to the questions in second paragraph: Generally, the NRC will issue generic guidance (e.g., standard review plan, regulatory guidance, NuRegs, etc.) documenting an acceptable way (not the only way) to meet regulatory requirements. Unless specified otherwise, the NRC staff generally uses reasonable assurance as the standard to determine if adequate protection has been met. Reasonable assurance is not complete and total assurance. A degree of engineering judgement is used, at times, to give the staff high confidence that adequate protection has been accomplished.

For regulations that are not specifically provided for adequate protection but were instead imposed on a different basis (such as a cost-beneficial safety improvements per 50.109),

management directives (e.g., for backfits) and policy statements (e.g., use of PRA) also provide guidance acceptable to the NRC.

V-7 APPENDIX 5

6. Request for Meeting with NRR Independent Reviewer The DPO Appeal panel members Cynthia Jones and Kevin Coyne would like to meet individually with the independent reviewer that NRR used for DPO 2021-001 that performed the RADTRAD control room dose calculations as discussed in the NRR Directors Decision Memo. Based on time availability, we would ask that this be scheduled in the mid- December timeframe. This time would be used to go over in detail the basis for selection of the input parameters for the RADTRAD runs, entail a general discussion of the technical results, and possibly run the code with other various parameters recommended by the DPO Appeal panel members to observe the variation in results.

NRR Answer:

The meeting with the NRR independent reviewer is scheduled for Monday, December 19 at 2:00 p.m.

V-8 APPENDIX 5

VI. APPENDIX 6: DPO-2021-01 Appeal Panel Questions and NRR Answers (Round 2)

1. With regard to calculation of particulate removal coefficients in the main steam lines, the response to Question 1 (Treatment of Conservatisms) noted:

licensees (including FitzPatrick, Clinton, Limerick, and LaSalle) have employed a 20-group probabilistic distribution of particle sizes in the MSL volumes (conservatively deviating from AEB 98-03 based on newly discovered information, at the time of the submittals). Therefore, the NRC staff approved FitzPatricks analysis based on current state of knowledge and consistent with existing methodologies found acceptable in generic communications.

The staffs safety evaluation (ML20140A070) provided detailed discussions of the previous licensing precedents for Clinton (2005), Limerick (2006), Nine Mile Point (2007) and LaSalle (2010) and noted several factors which were applicable to the staffs acceptance of the approach to evaluating particulate removal in the main steam lines, including:

- No credit for aerosol settling after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> (Clinton, Limerick, LaSalle)

- No credit for hold up or plate out in the main condenser (Clinton)

- Failure of an outboard MSIV on each line to close

- Presence of a seismically qualified condenser (Limerick, LaSalle)

- No credit for drywell sprays (Clinton, Limerick, LaSalle)

- Application of a penalty on aerosol settling velocity to account for removal of larger particles by the sprays resulting in a conservative estimate (Nine Mile Point)

In FitzPatricks response to ARCB-RAI-2 (ML20090E279), the licensee modeled aerosol deposition more conservatively, resulting in a sensitivity case (Case, S0, Base Sensitivity Case) that exceeded the control room 10 CFR 50.67 acceptance criteria. As stated in the staffs SE, NRC staff compared the licensees aerosol removal constants used in the base case with the values used in other MSIV leakage analyses submitted by other licensees for NRC staff review and agrees that the values used in the licensees base case are conservative.

a) In light of the additional conservatisms associated with previous licensing precedents and the significant sensitivity of the radiological consequence results to selection of aerosol deposition parameters (as demonstrated by sensitivity case S0), what was the basis for the staff accepting the aerosol deposition parameters that were initially questioned in ARCB-RAI-2?

NRR Answer:

The licensee submitted their license amendment for review and approval by the NRC staff. The sensitivity analysis was submitted as additional information in response to VI-1 APPENDIX 6

the RAI. The sensitivity analysis base-case was not submitted by the licensee, nor viewed as by the NRC staff, as a supplement to the original submittal to be used as their licensing basis. It was presented to the NRC staff to address concerns related to how the gravitational settling credited in the main steam lines (MSLs) considers the changing aerosol characteristics (uncertainty) due to the preferential removal of larger aerosols because of the credit assigned to containment sprays. In addition, the sensitivity analysis also showed the significant reduction in results if a pathway to the condenser was credited in the analysis.

The sensitivity analysis was performed and submitted in response to ARCB-RAI-2e following a close-door meeting with the licensee held January 29, 2020. (see ADAMS ML20044F188). The sensitivity analysis model differed from the LAR model with two important changes. These changes represent their sensitivity study base case. The primary modeling change which impacts the results the most was utilizing different settling velocity assumptions than those assumed in AEB-98-03 and endorsed as Agency positions in RIS 2006-04. The sensitivity study base-case used NUREG CR 5966 which is believed to be overly conservative and has not been adopted to date.

The sensitivity study base-case also replaced the AEB-98-03 assumption for aerosol size with the recommendation from the Nuclear Energy Agency report State-of-the-Art Report on Nuclear Aerosols, NEA/CSNI/R(2009), (SOAR). The SOAR recommends a range of aerosol AMMDs based on experiments which range from 1 AMMD to 3 AMMD. As discussed in the SOAR, a value of 1 AMMD would best represent aerosols in the reactor coolant system and a value of 3 AMMD represents aerosols in the containment. The NRC staff understands that the dose analysis required by regulation, and reflected in regulatory guidance, is to use a source term in the containment. However, the licensee selected a value of 2 AMMD. This is overly conservative as a containment source term, however the NRC staff felt it reasonable for a sensitivity analysis intended to provide insights to understand the uncertainty within the Fitzpatrick LAR licensing basis model utilizing AEB 98-03 assumptions. The SOAR recommendations are not endorsed by the NRC for use in design basis accident analyses. The NRC staff working on the Fitzpatrick LAR were aware of the sensitivity analysis approach and the licensees use of the SOAR report assumptions as opposed to the AEB 98-03 assumptions. The NRC staff also understood that they could not impose on the licensee the use of the SOAR as their licensing basis analysis of record. The use of the SOAR recommendations in the sensitivity study were discussed with the licensee in the closed-door meeting and were only intended to help the NRC staff understand the uncertainty in the model.

b) The SE seems to indicate that the staffs acceptance of the less conservative aerosol removal constants was based, at least in part, on the capabilities associated with the power conversion system and condenser (e.g., Section 3.1.1.4.4 of the SE states

[t]he licensees sensitivity results demonstrate that the condenser is very effective in substantially reducing the dose consequences from MSIV leakage). However, as noted in the response to Question 2 the staffs approval of the FitzPatrick analysis did not account for all or any portion of the PCS/main condenser being available for deposition. Please clarify if the approval of the aerosol removal constants initially VI-2 APPENDIX 6

questioned in ARCB-RAI-2 were approved on a purely deterministic basis or if the risk and engineering insights described in the SE were needed to support approval of non-conservative parameters.

NRR Answer:

As stated in Section 3.5 of the NRC staffs SE, Since the application is not a fully risk-informed submittal (with probabilistic risk information), the staff does not apply risk as the basis for acceptance of a request; however, the following risk and engineering insights inform the technical review by supporting the deterministic safety conclusions and enhance the technical reviewers confidence in their technical evaluations. Therefore, we confirm that these parameters were approved on a purely deterministic basis.

2. Although the response to Question 1 (Treatment of Conservatisms) acknowledges that the particulate removal coefficients were non-conservative, the staff also approved potentially non-conservative values for the elemental iodine removal coefficients. With regard to elemental iodine removal, Standard Review Plan (SRP) Chapter 6.5.2, Rev 4 (2007, ML070190178),Section III.4.C.ii, states:

s must be limited to 20 per hour to prevent extrapolation beyond the existing data for boric acid solutions with a pH of 5.

The SE noted that the staff concluded that the impact of the deviation was not significant based on the staffs confirmatory calculation. However, in approving the use of a value greater than 20 per hour in the SE, how did the staff consider the limitations stated in SRP Section 6.5.2 (e.g., extrapolation beyond existing data) and what that the technical basis for deviating from the SRP?

NRR Answer:

The technical basis for deviating from the SRP with regards to the elemental iodine removal coefficient was that, after performing confirmatory calculations, the staff verified that using more conservative values would have a negligible impact in the final results (i.e., control room dose). During the interviews performed on behalf of the NRR Office Director with FitzPatricks lead reviewer, he explained that elemental iodine was believed to be 95% of the dose contribution (the old source term from TID-14844 (RGs 1.3 and 1.4) assumed that 95% of the iodine released would be in elemental form). However, the revised alternative source term from NUREG-1465 (RG 1.183) assumes that 95% of the iodine released is in particulate form. Therefore, differences in the removal coefficients used for elemental iodine would be reduced in significance.

3. In the response to Question 1 (Treatment of Conservatisms), it was stated that:

In the Directors Decision Memo, NRR acknowledged the particulate removal coefficients used in the FitzPatrick analysis for the drywell and the main steam line (MSL) volumes were not conservative and were greater than typical values.

FitzPatricks quality assurance program description is included in Exelon Nuclear NO-AA-10, Quality Assurance Topical Report (QATR) (e.g., see ML22031A026). Section 3 of VI-3 APPENDIX 6

the QATR provides process and control measures for identification of design inputs, calculations, and analysis. These measures cover quality related activities such as identification of design inputs, control of the design process (including safety analysis accident scenarios), design analyses, and design verification. These QATR design control measures are needed to meet the requirements of 10 CFR 50, Appendix B, Criterion III, Design Control, related to the establishment of design control measures for verifying or checking the adequacy of design items including accident analysis. The Appeal Panel notes that its independent analysis described in the Directors Decision indicates that use of more reasonable and accurate values for these removal coefficients increases the control room radiological results. Further, RG 1.183, Section 5.1.3, states that numeric values that are chosen as inputs to the analyses required by 10 CFR 50.67 should be selected with the objective of determining a conservative postulated dose.

Considering that the calculation of record (JAF-CALC-19-00005) supporting the Alternate Source Term (AST) license amendment is appropriately identified as safety-related calculation and covered by the QATR and 10 CFR 50, Appendix B, why is the use of acknowledged non-conservative calculational inputs consistent with the QATR and 10 CFR 50, Appendix B?

NRR Answer:

Design inputs approved as part of the AST adoption become part of the sites licensing basis (CLB). Any changes to these design inputs need to be assessed to ensure regulatory requirements (e.g., 10 CFR 50, Appendix B, Criterion III, 10 CFR 50.59) and self- imposed standards (i.e., QATR) continue to be met.

4. In the response to Question 2 (Methods to Assess Functionality of Non-Safety SSCs), it was stated that:

FitzPatricks analysis assumes the turbine building is lost to a seismic event and that the turbine stop valves (TSVs) are the release point of radiation to the environment. This assumption is consistent with RG 1.183 and is supported by sound engineering judgement given that all components upstream the TSVs are seismically qualified, safety-related and, therefore, not assumed to fail. The NRC staffs approval of FitzPatricks analysis was based on this information and did not account for all or any portion of the PCS/main condenser being available for deposition.

The Appeal Panel agrees the licensees LAR submittal assumed release from the turbine stop valves without credit for hold up or dilution in the turbine building (e.g., see Attachment 2 of the LAR). However, the staffs Safety Evaluation noted that the staff include[d] the probability of the existence of a pathway to the condenser to offset uncertainties in crediting aerosol removal from drywell sprays in calculating the dose consequences of MSIV leakage. As noted in the response to Question 1, the treatment of particulate removal is an acknowledged non-conservatism in the analysis rather than an uncertainty. Further, the Directors decision noted that more reasonable and accurate values for these parameters were determined and used in the staffs own independent modeling conducted after issuance of the FitzPatrick AST amendment. As noted in Section 3.5 of the staffs SER, the staffs findings are primarily based on traditional VI-4 APPENDIX 6

deterministic review approaches but risk and engineering insights associated with holdup and deposition in the condenser inform the technical review by supporting the deterministic safety conclusions. The staffs evaluation of the crediting the PCS and condenser concluded by stating that these risk and engineering insights support its reasonable assurance finding based on its deterministic review.

A plain language reading of the staffs conclusions as documented in the SE support a view that:

(1) the licensees deterministic evaluation used certain design inputs that were non-conservative; (2) use of these non-conservative inputs did not support LAR approval based on the deterministic analysis alone; (3) consideration of risk and engineering insights associated with the PCS and condenser were needed to offset these non-conservatisms and associated uncertainties; and (4) the totality of the information presented in the SE provided the staffs regulatory basis for demonstrating compliance with 10 CFR 50.67 and approving the AST LAR.

This view is reinforced in the Section 3.6 of the SE where the staff states there is reasonable assurance supported by risk and engineering insights, that the licensees estimates for the EAB, LPZ, and CR doses comply with the cited acceptance criteria. It is unclear to the Appeal Panel how the staff could reach could approve the amendment given the non-conservative inputs without reliance on the risk and engineering insights documented in Section 3.5 of the SE. Further, since the staff had not quantified the degree of non-conservative in removal parameters or the potential offsetting mitigation credit, it is not clear how the staff ensured that non-conservatism were appropriately offset or balanced by other mitigation credit or conservatisms. Considering the statements contained in the SE, could the staff have approved the AST LAR for FitzPatrick without consideration of the risk and engineering insights associated with credit for the PCS and condenser? If so, how were known non-conservatisms documented and addressed by the deterministic analysis in a manner consistent with the guidance in RG 1.183? If not, how is the staffs reliance on these risk and engineering insights documented in the licensing basis?

Note: LIC-100, Control of Licensing Bases for Operating Reactors, Section 6.5, states that NRC safety evaluations provide the regulatory bases for NRC decisions in licensing actions. LIC-100 further notes that [i]t is important that the licensees provide the licensing bases information so that there is no confusion following the licensing action and to avoid a perception of staff-imposed backfits.

NRR Answer:

The NRC staff would have approved the AST LAR for FitzPatrick without consideration of the risk and engineering insights associated with credit for the PCS and condenser because the deterministic analysis provided reasonable assurance the regulatory VI-5 APPENDIX 6

requirements would be met. In fact, if the NRC staff were to receive a similar LAR today, it would very likely be approved without any credit for the PCS and condenser.

The non-conservatisms known at the time of the review were reviewed and addressed in their respective sections of the NRC staffs SE (i.e., Section 3.1.1.1.4).

5. In the response to Question 3 (Impact of Removal of MSIV Leakage Collection Systems),

it was stated that:

If a direct release from the stem of the outboard MSIVs to the environment were to occur, it would release in the reactor building, which has been evaluated for the safe shutdown earthquake (SSE) and it is assumed to remain structurally intact.

Any leakage in the reactor building gets processed by the standby gas treatment (SBGT) system (assumed to be in operation 20 minutes after a loss of coolant accident (LOCA)).

Given the location of the outboard MSIVs in the pipe tunnel area (e.g., see UFSAR Figure 12.3-7), can you please explain the flow path that would direct potential leakage from the outboard MSIV packing or mechanical joints to the SBGT system rather than to the turbine building? If leakage from the MSIV packing or mechanical joints begins, how would this flow path be established and maintained?

NRR Answer:

The response to this question was developed with the assistance of FitzPatricks resident inspector. The resident inspector walked down the MSLs, entered the pipe tunnel area and determined that given the location, the path of least resistance for a potential release would be the reactor building.

6. In the response to Question 4 (Quality of the Calculation of Record), it was stated that:

The licensees analysis postulates a traditional large break LOCA (double-ended guillotine break of the recirculation line). In a section of FitzPatricks submittal, the licensee stated: Although postulating a main steam line break in one steam line inside the drywell would maximize the dose contribution from the MSIV leakage, the steam line break is not a credible event during a LOCA, because the Seismic Class 1 main steam piping up to the TSVs is designed to withstand the SSE (Ref.

9.62). In reviewing this statement, the NRC staff requested the following information from the licensee via ARCB-RAI-3: Please provide additional information to justify that assuming a recirculation line rupture instead of a main steam line rupture is consistent with the guidance from RG 1.183 that assumptions should be selected with the objective of maximizing the postulated radiological consequences. The licensees response included an assessment of the dose consequence modeling all four MSLs (vs only two MSLs volumes available for activity deposition) with one broken MSL and a failed open MSIV. The results indicated that the dose consequences in the original analysis were bounding. The scenario provided in the licensees response assumes multiple failures (i.e., the recirculation break, an MSIV stuck open, an MSL break) and are overly VI-6 APPENDIX 6

conservative. The licensees current analysis of record provides the bounding scenario assuming the most limiting failure.

Please clarify the following:

a) While the response to ARCB-RAI-3 (ML20090E279) supports the conclusion that the recirculation line break is more limiting than a main steam line (MSL) break for the control room dose, the MSL break scenario appears to be more limiting for the exclusion area boundary (EAB) and the low population zone (LPZ) analysis. For example, the MSL break radiological consequence contribution from MSIV leakage for the EAB analysis is more than 30% greater than the recirculation line break case.

As specified in 10 CFR 50.67, the AST analysis must meet three radiological criteria:

dose at the exclusion area boundary; dose at the low population zone boundary; and dose in the control room. However, in the analysis presented in the LAR that was subsequently approved, the staff did not appear to include the bounding case for offsite (EAB and LPZ) dose consequences in the licensing basis. Why was the analysis that was non-bounding for two of these three radiological criteria selected as the sole licensing basis case?

NRR Answer:

The analysis performed in response to ARCB-RAI-3 was not submitted as a supplement to the LAR or as a substitute to the original analysis. The purpose of the analysis was to demonstrate that the original analysis is conservative for the most limiting case relative to the acceptance criteria which is the control room dose. While there was an increase in the off-site estimates in this analysis, the NRC staff considered these increases to be minor relative to the large margin to the off-site acceptance criterion. When reviewing dose consequence analyses, the NRC staff tends to focus more attention on analyses that estimate doses approaching the regulatory acceptance criteria. In the case of FitzPatrick, the off-site results for the analysis of record as well as all the sensitivity cases are more than a factor of 25 below the acceptance criteria.

b) The licensees response to ARCB-RAI-3 does not include a discussion of the multiple failures discussed in the above response (i.e., the recirculation break, an MSIV stuck open, an MSL break). Specifically, the licensee does not discuss assuming a concurrent recirculation line break. Further, the assumption of a MSL break coincident with an MSIV failure is analogous to similar assumptions made in the LAR (i.e., LOCA due to recirculation line break and a single stuck open MSIV). As discussed in UFSAR Section 5.2.4.6, Containment Isolation, a MSL line break rendering the inner isolation valve inoperable is considered within the containment isolation design basis. Can you clarify what conditions were assumed for the staffs MSL break analysis and why the staff believes that the MSL break coincident with a failed open MSIV is overly conservative (particularly in comparison to the recirculation line break used in the calculation of record)?

VI-7 APPENDIX 6

NRR Answer:

ARCB-RAI-3 reads as follows: Please provide additional information to justify that assuming a recirculation line rupture instead of a main steam line rupture is consistent with the guidance from RG 1.183 that assumptions should be selected with the objective of maximizing the postulated radiological consequences. There is nothing in the request that suggests analyzing multiple failures. Rather the question asks for information regarding an analysis with the initiating event being a break in a MSL rather than a recirculation line break. As with all design basis dose consequence analyses, in addition to the initiating event, the worst single failure is assumed which in this case is a failure of one MSIV to close. The NRC staff does not consider that the assumption of a MSL break coincident with a failed open MSIV is overly conservative.

c) The MSL break analysis described in the response to ARCB-RAI-3 assumed that 135 scfh of MSIV leakage was equally distributed among the three non-faulted steam lines (45 scfh per steam line) rather than assuming that this leakage was concentrated on a single non-faulted steam line (similar to the assumption made by the licensee for the recirculation line break). As discussed in Section 3.11.10 in Attachment 1 of the licensees LAR submittal (ML19220A043), the licensee assigned a leakage rate of 135 scfh to both of the shortest of the steam lines, consistent with the guidance in RG 1.183, Appendix A, Section 6.2 (the remaining longer steam lines were assumed to have no leakage). The licensee noted that this was conservative because it resulted in less radionuclide removal in the main steam lines. The Appeal Panel notes it is difficult to directly compare the recirculation and MSL break cases since they applied different MSIV leakage assumptions, with the MSL break analysis using a potentially non-conservative assumption for MSIV leakage relative to the recirculation line break analysis.

Considering these differences in addressing MSIV leakage, would applying the same assumption used for the recirculation line break case to the MSL break case (i.e., assigning the total allowable MSIV leakage to the two shortest steam lines) result in higher radiological consequences than described in the response to ARCB-RAI-3?

NRR Answer:

RG 1.183 Appendix A, Assumption 6.2 states the following: All the MSIVs should be assumed to leak at the maximum leak rate above which the technical specifications would require declaring the MSIVs inoperable. The leakage should be assumed to continue for the duration of the accident. Postulated leakage may be reduced after the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, if supported by site-specific analyses, to a value not less than 50% of the maximum leak rate. In their LAR, the licensee chose to only credit MSL deposition in the two shortest MSLs. There is nothing in RG 1.183 that restricts the number of MSLs that can be credited for deposition. In response to the RAI, the licensee determined that assuming a ruptured MSL as the initiating event and crediting deposition in the remaining intact MSLs yielded results VI-8 APPENDIX 6

demonstrating that the original analysis is conservative for the most limiting case relative to the acceptance criteria which is the control room dose. It is conceivable that results could be higher if credit was restricted to the two shortest MSLs however there is nothing in RG 1.183 that restricts the number of MLS that can be credited for deposition.

d) The licensees response to ARCB-RAI-3 only provides the MSIV leakage contribution to the post-LOCA radiological consequences. Please confirm that the remaining contributors (e.g., containment leakage and ESF leakage as noted in the response to ARCB-RAI-1B) are the same as in the recirculation line break case. If these values change as a result of assuming a MSL break, please provide the updated values.,

NRR Answer:

ARCB-RAI-3 was asked in relation to the dose consequence from MSIV leakage. The NRC staff did not expect that the licensee would verify that other release pathways would not be impacted by the requested analysis. The NRC staff uses professional judgement to allocate resources when performing reviews. Since there is no reason to expect that changes in assumptions for the MSIV leakage pathway would have an impact on the dose consequence from other pathways, such confirmation was not requested of the licensee or performed by the NRC staff. The NRC staff notes that when performing a complex dose consequence analysis with releases from different pathways, each pathway is analyzed independently. The dose consequence from each pathway is then added to produce the composite dose for comparison to the acceptance criteria.

7. In Question 5 (Clarity of Staff Position and Review Criteria for Design Basis Accident Dose Criteria), the Appeal Panel requested the title or procedure of the guidance documents that the staff used to determine a regulatory finding of adequate protection, specifically the question was:

In the Directors decision, the term regulatory finding of adequate protection is used several times. For regulations that were promulgated to meet an adequate protection standard, what criteria does the staff use to determine that adequate protection is met? What guidance documents this process? For regulations that are not specifically provided for adequate protection, but were instead imposed on a different basis (such as a cost-beneficial safety improvements per 50.109), what guidance is available for staff to consistently apply regulatory acceptance criteria consistent with the intent of the associated regulation (e.g., providing a level of safety above the adequate protection standard)?

The staff response stated, in part, that:

Regulatory guides provide one acceptable way to meet regulatory requirements. The NRC staff remains open to accept methodologies outside those documented in VI-9 APPENDIX 6

regulatory guides, as long as they provide reasonable assurance of adequate protection.

Reasonable assurance is not complete and total assurance. A degree of engineering judgement is used, at times, to give the staff high confidence that adequate protection has been accomplished.

How does NRR ensure that the staffs technical evaluations of incoming LARs for use of ASTs provide consistent repeatable results amongst different reviewers? In other words, given that the staff is using RG 1.183 for evaluation of incoming LARs, which has various parameters for each evaluation, how does NRR management ensure consistency in the staffs review if one uses more conservative or realistic results for these evaluations?

NRR Answer:

Each license amendment request is plant-specific and uses plant-specific proposed parameters. The NRC staff, with supervision from NRC management, reviews the information provided in the LAR, regardless of regulatory guidance adoption, and decides on its acceptability. Management review, knowledge management shared between reviewers, and precedent are used to ensure consistency where appropriate. Regulatory guidance allows for a more consistent approach in reviewing these evaluations, but the NRC staff welcomes deviations from the guidance with proper justifications.

VI-10 APPENDIX 6