ML25258A035
| ML25258A035 | |
| Person / Time | |
|---|---|
| Site: | Nuclear Energy Institute |
| Issue date: | 09/12/2025 |
| From: | Amico P, Facemire J, Lee D, Young D Nuclear Energy Institute |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| Download: ML25258A035 (1) | |
Text
Seismic Event Selection for RI EPZ Sizing September 17th, 2025 Paul Amico, Larry Lee, David Young, Jon Facemire
Comment 1. Use of 1g Cutoff Comment 1: Because the fragility determination accounts for a spectral shape and the natural frequencies of an SSC, the arbitrary cut-off of PGA at 1.0g PGA may lead to unconservative screening of the SSCs, meaning screening out of SSCs that should be retained in the analysis.
Response: We propose to include a caution on the use of a 1.0g cutoff that will require a check to confirm that it is applicable to the site. The proposed words are on the next slide.
Comment 1. Use of 1g Cutoff To determine an appropriate upper bound cutoff, the applicant should conduct a review of the seismic design/building code(s) applicable to key offsite infrastructure in the vicinity of the plant site. The infrastructure of interest for this review would typically be local area power transmission lines and substations, roads and bridges, and structures housing law enforcement and fire response functions (e.g., police and fire stations) and emergency operations centers. The goal of the review is to qualitatively identify a PGA at which most of this infrastructure would be significantly degraded or lost.
Consideration of seismic accelerations greater than this level would not meaningfully benefit emergency planning for the site due to the increasing uncertainty that planned prompt protective measures could be effectively implemented given the level of infrastructure damage. This conclusion also acknowledges the fact that planned prompt protective measures may not be the most suitable set of actions given the potential offsite conditions following a severe seismic event. In such situations, State and local emergency managers would need to assess the status of offsite emergency response capabilities and infrastructure as well as other competing response priorities (e.g., search and rescue in collapsed structures), and then determine ad hoc protective actions based on their knowledge and experience.
Comment 2. Failure or Success of SSCs Comment 2a: if this proposed pass-fail criterion were used for todays large light water reactors, the analysis would assume that most of those failures would be successes, and the corresponding sequences would be dropped from the analysis.
Response: The methodology specifically states that it is not applicable to large light water reactors.
Comment 2. Failure or Success of SSCs Comment 2b: NEI could consider in this methodology retaining these human errors and non-seismic failures in the first-stage analysis, study which sequences contain those types of events, and use case-by-case judgment as to whether the sequence should be retained or should be screened out, based on the role of these failures and the importance of the sequence.
Response: The methodology does not consider specific sequences or develop minimal cutsets. It creates a single plant damage state where all SSCs with C10% below the target value fail, thereby bounding the plant damage and maximizing consequences. The consideration of specific sequences and minimal cutsets and associated plant level insights are addressed in the PRA/PRA-based SMA that is required to be included in the application for design certification, COLA or construction permit.
Comment 2. Failure or Success of SSCs Discussion: In the PRA-based seismic margins approach, which developed minimal cutsets for core damage, the retention of random failure modes and human actions had a specific purpose.
Cutset HCLPF might be controlled by a very seismically strong component that was highly unreliable or whose function could fail due to human error. The idea was to protect against missing an important combination.
Random failures and HEPs <= 0.01 were required to be considered.
Example: Cutset A(0.2g) x B(0.22g) x C(0.55g) has controlling HCLPF 0.55g However, what if random failure probability or HEP for component C was 0.07?
Alternative cutset AxBxC would be A(0.2g) x B(0.22g) x C(P=0.07), so it would be the combination of HCLPF 0.22g x 0.07 probability, which could be more important than the seismic-only cutset.
This concept does not make sense for the seismic EPZ methodology, since we are not trying to determine a cutset, sequence, or plant level value and are only providing the basis for a bounding plant damage state (see next slide).
Example using 2 x GMRS = 0.5g Generic version of an actual plant design.
Comment 2. Failure or Success of SSCs Fragility Group Code C10% (g)
Fragility Group Code C10% (g)
Fragility Group Code C10% (g)
SFR_112 0.06 SFR_103 0.18 SFR_124 0.24 SFR_128 0.06 SFR_111 0.18 SFR_125 0.24 SFR_135 0.06 SFR_113 0.18 SFR_134 0.24 SFR_136 0.06 SFR_114 0.18 SFR_145 0.24 SFR_137 0.06 SFR_115 0.18 SFR_102 0.30 SFR_139 0.06 SFR_116 0.18 SFR_106 0.30 SFR_140 0.06 SFR_122 0.18 SFR_107 0.30 SFR_147 0.06 SFR_138 0.18 SFR_109 0.30 SFR_149 0.06 SFR_104 0.24 SFR_110 0.30 SFR_121 0.12 SFR_118 0.24 SFR_123 0.30 SFR_144 0.12 SFR_119 0.24 SFR_127 0.30 SFR_101 0.18 SFR_120 0.24 SFR_143 0.30
- This collection of failures already represents a quite low probability.
It doesnt seem reasonable to simply add additional random/human failure.
It is unclear that there is any value in removing seismic failures and adding these.
Comment 3.
Additional Guidance on Scenario Selection Comment 3a: NEI should consider providing specific guidance on how a plant damage state is determined for a scenario that may lead to radiological consequences given that an SPRA may not be available.
NEI should consider providing an example to clearly demonstrate how a plant damage state is determined from the available information, how the seismic successes and failures are incorporated in the sequences, and guidance on demonstrating that the resulting scenario follows regulatory guidance, meets regulations, and is appropriate for EPZ sizing.
Response: We agree that an example would improve clarity and understanding. We believe that providing that example is adequate to address the NRC comment in its entirety. However, the example, while it should be based on a real situation, must protect proprietary information. The form of our proposed example is shown on the following slides.
Comment 3.
Additional Guidance on Scenario Selection Step 1: Determine the fragility parameters for the SSCs that perform the safety functions that would be credited for a seismic event.
Step 2: Determine the GMRS for the site being considered.
Step 3: Compare the C10% for the SSCs to 2xGMRS and prepare a table of failures and a table of successes (see next slides for the example of a site that has a GMRS of 0.25g and so 2xGMRS of 0.5g)
Comment 3.
Additional Guidance on Scenario Selection
- Table of failures (repeat of table in Slide 7)
Fragility Group Code C10% (g)
Fragility Group Code C10% (g)
Fragility Group Code C10% (g)
SFR_112 0.06 SFR_103 0.18 SFR_124 0.24 SFR_128 0.06 SFR_111 0.18 SFR_125 0.24 SFR_135 0.06 SFR_113 0.18 SFR_134 0.24 SFR_136 0.06 SFR_114 0.18 SFR_145 0.24 SFR_137 0.06 SFR_115 0.18 SFR_102 0.30 SFR_139 0.06 SFR_116 0.18 SFR_106 0.30 SFR_140 0.06 SFR_122 0.18 SFR_107 0.30 SFR_147 0.06 SFR_138 0.18 SFR_109 0.30 SFR_149 0.06 SFR_104 0.24 SFR_110 0.30 SFR_121 0.12 SFR_118 0.24 SFR_123 0.30 SFR_144 0.12 SFR_119 0.24 SFR_127 0.30 SFR_101 0.18 SFR_120 0.24 SFR_143 0.30
Comment 3.
Additional Guidance on Scenario Selection
- Table of Successes Fragility Group Code C10% (g)
Fragility Group Code C10% (g)
SFR_126 0.60 SFR_105 1.79 SFR_146 0.90 SFR_108 1.79 SFR_150 1.20 SFR_132 1.79 SFR_117 1.49 SFR_133 1.79 SFR_129 1.49 SFR_142 1.79 SFR_130 1.49 SFR_131 2.99 SFR_141 1.49 SFR_148 2.99
Comment 3.
Additional Guidance on Scenario Selection Step 4: Based on the tables of failure and successes, identify the functional failures and successes that define the plant damage states:
This combination of failures and successes can be summarized as follows:
Damage to all buildings except the reactor building (SFR_nnn, SFR_nnn, etc)
Degraded reactivity control (SFR_nnn, SFR_nnn, etc)
Loss of all AC and DC power (SFR_nnn, SFR_nnn, etc)
Loss of normal reactor cooling (SFR_nnn, SFR_nnn, etc)
Loss of HVAC (SFR_nnn, SFR_nnn, etc)
Loss of spent fuel cooling (SFR_nnn, SFR_nnn, etc)
Primary piping failures (SFR_nnn, SFR_nnn, etc)
Passive air cooling degraded (SFR_nnn, SFR_nnn, etc)
Reactor building isolation success (SFR_nnn, SFR_nnn, etc)
As indicated, for each functional failure or success, the specific fragility groups that contributed to the conclusion would be identified.
Comment 3.
Additional Guidance on Scenario Selection Step 5: Based on the defined plant damage state (functional failures and success), develop the source term in accordance with the guidance in NEI 24-05.
Step 6: Based on the source term, perform a consequence calculation for the specified site in accordance with the guidance in NEI 24-05.
Step 7: Based on the consequence calculation for the specified site, determine the distance to 1 Rem over 96 hours0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br />.
The example will also include performing the cliff-edge check (see discussion for proposed cliff-edge effect check in response to question 4)
Comment 3.
Additional Guidance on Scenario Selection Comment 3b: NEI should consider describing actions to be taken to update or confirm the seismic scenario for an EPZ once a site-specific SPRA becomes available, for example actions to be taken by an applicant for an operating license with a previous construction permit.
Response: NEI 24-05 provides guidance on the requirement to confirm all the scenarios used for EPZ determination once a site-specific PRA is completed.
Comment 4. Sensitivity Check Comment 4a: The analysis adds additional failures of SSCs whose C10% capacity is within 10% higher than the EPZ earthquake (i.e., 2 x GMRS), to guard against what NEI describes as potential cliff-edge effects. Staff notes that the cliff-edge effects, in the sense that a slight increase in the hazard severity would result in a large increase in consequences Response: This topic came up in the tabletop exercise, and as a result NEI has changed the approach to perform the cliff-edge assessment by increasing the hazard rather than decreasing the component capacity by 10%. This is more consistent with the accepted definition of cliff-edge as stated in the NRC comment.
Comment 4. Sensitivity Check Comment 4b: The use of a mean fragility curve to determine the cliff-edge effects may mask the potential cliff-edge effects because of a large u. The governing failure modes could also be examined, such as a brittle anchorage failure mode or a failure mode associated with a threshold criterion (e.g., buckling).
Response: With the change of approach as discussed on the previous slide, the focus is now rightly on the uncertainty in the hazard rather than the uncertainty in the fragility. As a result, the use of the mean fragility curve and keeping the C10% constant based on that curve is reasonable. In addition, under this approach there is no point in examining governing failure modes since these are all accounted for in the fragility parameters.
Comment 4. Sensitivity Check Comment 4c: Staff note that NEI should consider expanding the white papers discussion on the cliff-edge effects including the need for sensitivity analyses, and that an example for calculating cliff-edge effects could prove useful.
Response: NEI agrees that the cliff-edge effect is a sensitivity study.
As mentioned previously, this is properly the sensitivity of the success or failure of an SSC to a change in the hazard severity and not NEIs originally proposed approach of a sensitivity of the success or failure to a change in the C10% value. NEI will add an example of a cliff-edge effect sensitivity study to the white paper.
Comment 4. Sensitivity Check Comment 4d: A possible approach would be to perform a close examination for potential cliff-edge effects in addition to the use of the proposed criteria. Such an approach could involve examination of the cutsets of various scenarios or sequences for singletons and dominant contributors that may lead to cliff-edge effects.
Response: As discussed in the response to question 2, this methodology does not develop cutsets nor does it look at a range of scenarios or sequences. It is not intended to be a PRA-based seismic margin study to determine a plant-level capacity nor to identify dominant contributors, which is already addressed in the requirements for the severe accident risk chapters of the DCD or SAR. The only purpose is to define a reasonable single bounding plant damage state upon which to base the consequence calculation to support the EPZ distance. The approach proposed by NRC would fundamentally and foundationally change this.
Comment 4. Sensitivity Check Summary The cliff-edge effect check will be performed by a sensitivity analysis based on increasing the hazard level. See next slide for proposal.
The cliff-edge effect sensitivity study will continue to use the C10%
values as currently discussed in the paper as the basis for comparison with this increased hazard.
An example cliff-edge effect sensitivity study based on a genericized version of a known design will be added to the paper.
The methodology will continue to use the bounding single plant damage state approach and will not use sequence-and scenario-based cutsets.
Comment 4. Sensitivity Check Proposal for the cliff-edge sensitivity check hazard level is to use the 84th percentile hazard at the return interval for 2 x GMRS.
Example - Imagine GMRS for plant with hazard to right is 0.3g GMRS 0.3g 2xGMRS (base case) 0.6g Cliff Edge 0.75g