ML25008A249

From kanterella
Jump to navigation Jump to search

Memorandum - Summary of Regulatory Audit for TRISO-X Integrated Safety Analysis Definitions of Unlikely and Highly Unlikely
ML25008A249
Person / Time
Site: Triso-X
Issue date: 01/23/2025
From: Matt Bartlett
NRC/NMSS/DFM/FFLB
To: Lav S
NRC/NMSS/DFM/FFLB
References
EPID L-2022-PMP-0005
Download: ML25008A249 (1)


Text

MEMORANDUM TO:

Samantha Lav, Chief Fuel Facility Licensing Branch Division of Fuel Management Office of Nuclear Material Safety and Safeguards FROM:

Matthew Bartlett, Sr. Project Manager Fuel Facility Licensing Branch Division of Fuel Management Office of Nuclear Material Safety and Safeguards

SUBJECT:

AUDIT REPORT FOR MEETING HELD ON DECEMBER 3 AND 4, 2024, WITH TRISO-X LLC, TO ALIGN ON THE INTEGRATED SAFETY ANALYSIS DEFINITIONS OF UNLIKELY AND HIGHLY UNLIKELY, (ENTERPRISE PROJECT IDENTIFIER L-2022-PMP-0005)

The U.S. Nuclear Regulatory Commission (NRC) conducted an audit at the TRISO-X, LLC (TRISO-X) site on December 3 and 4, 2024. The purpose of the audit was to discuss appropriate definitions for unlikely and high unlikely. The NRC staff issued requests for additional information (RAI) including RAI-5, Definitions of Highly Unlikely and Unlikely, via letter dated August 28, 2023 (ML23230B191). TRISO-X provided a response via letter dated October 12, 2023 (ML23285A344), which did not resolve the issue. The NRC staff requested TRISO-X to supplement their response via letter dated April 2, 2024 (ML24082A222). The audit was scheduled to allow the NRC and TRISO-X staff to work toward resolution of the issue prior to TRISO-X submittal of a revised license application Chapters 1-13 in December 2024and the updated Integrated Safety Analysis Summary (ISA) Summary in January 2025.

The Director of the Division of Fuel Management opened the meeting by stating the NRC staff is seeking a success path that allows the NRC to make a finding of reasonable assurance of adequate protection for all the accident sequences in the ISA Summary. This means aligning on definitions of unlikely and highly unlikely that are applicable for the semi-quantitative ISA methodology TRISO-X has adopted. The NRC staffs focus is on ensuring safety and adequate defense-in-depth. As long as these goals are achieved, there is flexibility in applying the regulatory framework through exemptions, license conditions, and license commitments. The NRC staff recognize each fuel cycle facility is unique and the NRC staff want the audit discussions to focus on the TRISO-X license application.

CONTACT: Matthew Bartlett, NMSS/DFM 301-415-7154 January 23, 2025 Signed by Bartlett, Matthew on 01/23/25

2 S. Lav TRISO-X stated their goal for the audit was to identify the information the NRC staff needs to make its regulatory finding.

The NRC staff pointed out that the guidance in NUREG-1520, Revision 2, Standard Review Plan for Fuel Cycle Facilities License Applications, provide three general approaches for the ISA methodology including quantitative, semi-quantitative, and qualitative. TRISO-X has opted to use a semi-quantitative methodology, which takes into consideration both qualitative and quantitative information to assess the frequency and risk reduction contribution of safety controls. As a result, the analysis has inherent uncertainty in the risk index numbers. The guidance in NUREG-1520, Revision 2 indicates that acceptable definitions are 10-5 for highly unlikely and 10-4 for unlikely (-5/-4). TRISO-X has proposed definitions of 104 for highly unlikely as and 10-3 for unlikely (-4/-3). The NRC staff pointed out that NUREG-1520, Revision 2, states on page 3-AA-1:

The index method is one acceptable method of demonstrating compliance with the performance requirements. However, taking credit for using this method requires that the applicant follow all of the guidance contained in Appendix A. Otherwise, the applicant should provide additional justification.

The TRISO-X staff provided an overview of the operations involved in the uranium process building. They indicated the detailed design for the facility is under development, but they do not expect this effort to impact the content of the license application. TRISO-X staff confirmed they will submit Chapters 1-13 of the revised license application in December 2024. They highlighted the fact that the update includes a reduced number of accident sequences by reducing the amount of flammable materials required in their process.

The TRISO-X staff indicated their ISA methodology uses a combination of Hazard and Operability study (HAZOP), what if, and the checklist method. TRISO-X is working to finalize their probability hazard assessment in December 2024. Based on the outcome of the audit, they plan to submit a revised ISA Summary by the end of January 2025.

The NRC staff expressed the desire to better understand how TRISO-X is assigning values in the ISA Risk Assessment. The TRISO-X staff worked through several examples including material handling between different phases of operations, furnace operations, and manual operation involving material. The bulk of their accident sequences are criticality-related with a small number of radiological, chemical and fire-related accident scenarios. The NRC staff asked if TRISO-X had additional margin of safety (e.g., additional controls) they could credit in addition to item(s) relied-on for safety (IROFS) to support the justification for using the higher threshold of -4/-3 for their likelihood definitions. The TRISO-X staff said they believed they did have additional margin for all types of accident sequences. The NRC staff explained that since TRISO-X is reducing the margin of safety in their definitions for unlikely and highly unlikely, they need to demonstrate areas that provide for safety margin through other means, which would allow the NRC staff to qualitatively determine that TRISO-X is meeting the strategic goal of

-5/-4. This is needed for the NRC staff to determine reasonable assurance of adequate protection. TRISO-X stated they believe they have incorporated additional margin into their analysis including no sole-IROFS, the concept of see and flee, conservative assumptions for calculations of their risk index numbers, etc.

TRISO-X pointed out that the regulations allow facilities to use a facility specific definition for highly unlikely and unlikely and multiple existing facilities have successfully used -4/-3. The

3 S. Lav NRC staff stated that the thresholds for a semi-quantitate risk assessment include qualitative assumptions, which involve uncertainty on the reliability of the risk index numbers. The NRC staff determined that thresholds of -5/-4 are needed to meet the agencies strategic goals. The NRC has approved -4/-3 thresholds for semi-quantitative ISA methodologies for other facilities on a case-by-case basis, based on additional justification of additional safety margin.

The NRC staff pointed out that TRISO-X has several accident sequences that use only administrative IROFS for accident sequences. This appears contrary to requirements in Title 10 of the Code of Federal Regulations (10 CFR) 70.64(b), which requires a, preference for the selection of engineered controls over administrative controls to increase overall system reliability and the corresponding guidance in NUREG-1520, Revision 2, page 3-21. TRISO-X stated certain accident sequences require the use of administrative IROFS because their operations are based on batch processes and many of their controls only involve administrative actions. Therefore, they often do not have engineered controls to use as IROFS without a significant redesign and cost. The NRC staff observed in several instances TRISO-X appears to be relying on the same individual to perform multiple actions and crediting the administrative IROFS as independent. TRISO-X agreed that this was accurate but pointed out that since the actions are different, they should be able to credit the actions as independent. The NRC staff disagreed and pointed out this issue would be followed up further during the detailed technical review of the revised ISA Summary.

TRISO-X stated they had not previously understood that the -5/-4 threshold was needed to support the NRCs strategic goals. The NRC staff stated that the likelihood targets of highly unlikely and unlikely are based on limiting the integrated risk for a given facility and depends on the number of events that could occur leading to high and intermediate consequences per the ISA. The staff further stated that these principles are documented in public agency records since the early promulgation of the draft rule (SECY-98-185) and the Statements of Consideration for Subpart H (64FR41338). The staff indicated that the likelihood targets are intended to be consistent with the NRC Strategic Plan (NUREG-1614, Vol.1) and the performance goal of having no occurrences of accidental criticality. (SECY-98-185 and 64FR41338).

NUREG-1718, Standard Review Plan for the Review of an Application for a Mixed Oxide (MOX) Fuel Fabrication Facility, published August 2000, Section 5.4.3.2. and Appendix A provide detailed information on the development of the likelihood targets and the nexus to the strategic goals. Additionally, NUREG-1520, Revision 1, first published in March 2002, Section 3.4.3.2 (9) also identifies the nexus to the strategic goals in the development of likelihood targets. TRISO-X proposed that they continue to use the thresholds of -4/-3 for their likelihood definitions and fill in the gap for meeting the strategic goal by providing reliable scoring for the IROFS, ensuring they are robust, and incorporating additional safety margin in the analysis.

The NRC staff agreed that the goal is to ensure appropriate IROFS and safety margin is maintained, so facilities meet the strategic goal of reasonable assurance of adequate protection.

The NRC staff asked TRISO-X staff what their goal was for using the -4/-3 thresholds, since the majority of their accident sequences assign IROFS to reach a risk index threshold of -5 or lower.

TRISO-X stated they want to have similar reporting and inspection requirements compared to other fuel fabrication facilities. Therefore, using the lower threshold (-5/-4 instead of -4/-3) would result in significant additional burden for inspection and reporting without significant increase in safety. This would result in a competitive disadvantage to their facility and result in additional costs and negative public perception due to increased reporting of low safety significant events.

The NRC staff said they understand this perspective.

4 S. Lav The TRISO-X and NRC staff discussed several options to meet the safety goals while maintaining an appropriate level of reporting and oversight. Several different options were discussed including exemptions from certain reporting requirements, license conditions to maintain additional safety margin, license commitments to avoid sole-IROFS, and license commitments to provide a basis for use of administrative IROFS. Ultimately, both parties agreed in principle that the following options provided assurance of appropriate IROFS and defense-in-depth, while avoiding unnecessary reporting and inspection findings. These options included the following:

Commitment: No sole items relied-on for safety (IROFS).

[NOTE: THE FOLLOWING COMMITMENT WAS PROPOSED TO BE ADDED BY TRISO-X VIA EMAIL, FOLLOWING THE AUDIT.]

Commitment: The definitions of highly unlikely and unlikely in Section 9 of the ISA Summary will remain unchanged as -4 (highly unlikely) and -3 (unlikely) for determining reportability when evaluating a process upset against the 10 CFR 70.61 performance requirements.

Commitment: In addition to the statements on hierarchy of controls, provide a justification for relying only on administrative controls as IROFS, which demonstrates the alternatives to achieve the design criteria are not practicable (e.g., consistent with 10 CFR 70.64(b) and NUREG-1520, Revision 2, page 3-21).

Commitments for License Chapter 3 for demonstration of compliance with 10 CFR 70.61:

Criticality accident sequences: Use the double contingency principle (DCP) with the criteria presented in the three bullets shown in NUREG-1520 p. 5-A-7 and 5-A-8 (i.e.,

DCP +3). Use -4 (risk-indexed highly unlikely) for risk indexing1.

Non-criticality accident sequences: Intermediate and high consequence accident sequences will be documented with a likelihood index of at least -5 (risk-indexed highly unlikely) and -4 (risk-indexed unlikely).

Both TRISO-X and the NRC staff agreed in principle with these options for commitments to resolve the definitions of unlikely and highly unlikely. Additional details on implementation would need to be discussed and revised further in the RAI process on the revised license application.

Both parties agreed to hold a follow-up call to continue to refine these options on December 16, 2024.

5 S. Lav Attendees:

NRC TRISO-X Shana Helton, Director, Division of Fuel Management Jennifer Wheeler, Vice President, Regulatory Affairs Dante Johnson, Chief, Nuclear Analysis and Risk Assess Branch Robert Maurer, ISA Lead James Downs, Acting Chief, Fuel Facilities Licensing Branch Burns Cunningham, Licensing Engineer Jonathan Marcano Lozada, Senior Risk and Reliability Analyst Jeremy Munson, Senior Criticality Nuclear Engineer Matthew Bartlett, Senior Licensing Project Manager Docket: 07007027 License: Not Applicable

ML25008A249 OFFICE NMSS NMSS NMSS NMSS NAME MBartlett JGoodridge SLav MBartlett DATE 01/15/2025 01/15/2025 01/23/2025 01/23/2025