ML20083K386
| ML20083K386 | |
| Person / Time | |
|---|---|
| Site: | 99902041 |
| Issue date: | 03/26/2020 |
| From: | John Lehning NRC/NRR/DSS/SFNB |
| To: | Peters G Framatome |
| Rowley J | |
| Shared Package | |
| ML20083J967 | List: |
| References | |
| ANP-10339P Rev. 0, EPID L-2018-TOP-0034 | |
| Download: ML20083K386 (6) | |
Text
REQUEST FOR ADDITIONAL INFORMATION FOR THE REVIEW OF TOPICAL REPORT: ANP-10339P, REVISION 0, ARITA - ARTEMIS/RELAP INTEGRATED TRANSIENT ANALYSIS METHODOLOGY FRAMATOME INC.
REGULATORY BASIS As discussed in Section 3.0 of ANP-10339P, the ARITA methodology is intended to perform analysis used to demonstrate compliance with NRC regulations concerning certain anticipated operational occurrences and postulated accidents. Applicable regulations include General Design Criteria (GDCs) from Appendix A to Title 10 of the Code of Federal Regulations (CFR)
Part 50 that are associated with the integrity of the fuel (GDC 10), the design of the reactor coolant system (GDCs 15 and 31) the design of the emergency core cooling system (GDC 35),
the design of instrumentation, control, and protection systems (GDCs 13, 20, and 25), and reactivity control systems (GDCs 26, 27, and 28). Although the ARITA methodology is not used for performing radiological dose calculations, it is used to determine the number of failed fuel rods for dose assessments to determine compliance with regulatory limits specified in 10 CFR Part 100 or 10 CFR 50.67.
RAI-1
ANP-10339P identifies specified acceptable fuel design limit (SAFDL) figures of merit and non-SAFDL figures of merit, but it does not identify the acceptance criteria for each. Please explicitly state the acceptance criteria for each SAFDL and non-SAFDL figure of merit, with justification. In the case where the limit is plant specific, state how that limit is determined for a given plant.
RAI-2
ANP-10339P indicates that the ARITA evaluation methodology is intended for the analysis of PWR non-LOCA events identified in Chapter 15 of NUREG-800, the Standard Review Plan (SRP). Additionally, ANP-10339P discusses how the ARITA evaluation methodology is a composite of three variant evaluation methodologies (coupled EM, 0D EM, and static EM).
However, not all of the Chapter 15 events are applicable for modeling by the ARITA methodology, and ANP-10339P does not clearly identify which of the variant evaluation methodologies will be used to analyze each of the Chapter 15 events that are applicable.
Provide a table showing every event that will be analyzed using the ARITA evaluation methodology and which of the three variant evaluation methodologies will be used for the analysis of each event.
RAI-3
Section 4.2.4.7.2, Selection Process on Page 4-36 of ANP-10339P states that ['''''''''''''''' '''''
'''''''' '''''''''' ''''''''''''''''' '''''''''''''' ''''''''''''''''''''''] The discussion in this section also makes reference to [''''' '''' '''' ''''''''''''''] and suggests these designs can be ['''''''''''''' '''''''''''''
'' ''''''''' ' ''''''''''''''' ''''' '''' ''''''''''''']. However, the discussion does not explicitly state any criteria for the justification of similar behavior. Provide clarification on the intent of this passage and identify what criteria are used to justify the behavior of [' '''' '''' '' ''''''''''' '''''''' ''''''''''''
''''' ' ''''''''''' ''''''''''].
RAI-4
Section 4.2.4.7.2, Selection Process on Page 4-35 of ANP-10339P and Section 4.2.4.7.3, Calculation of FCM and TCS on Page 4-37 of ANP-10339P respectively provide discussions on the selection of [''''''''''''''''''''' ''''''''''''' ''' '''''] and on how fuel centerline melt (FCM) and transient cladding strain (TCS) are calculated using ['''''''''''''''' ''' '''' '''''''''''''''''''''''''''' ' ''
''''''''''''''''' ''''''''''''''''' '''''''''''' ''''' ''''']. It is not clear that the selection process for ['''''''''''''''''''''
'''''''''' ''''' ''] guarantees the [''''''' ''''''''''''' ''' ''''] will be identified and used in calculation of FCM and TCS. Provide a detailed justification that a ['''' '''' '''' '''''' '''''''''' '''''''' '
''''''] will not be identified in a full-core analysis versus ['''' ''''''''''''''''' '''' '''''''''''''''' '''''''''''''''].
Additionally, indicate whether Framatome has performed a validation of its selection process
[''''''' ' ''''' '''''''''''''''''' ''''' '''''''''''''''''''' ''' '''' ''''''''''''''' '''''''''''] and if not, what analyses have been performed to validate the approach described in Sections 4.2.4.7.2 and 4.2.4.7.3 of ANP-10339P.
RAI-5
Section 4.2.4.7.4, Calculation of TCS at Partial Power Conditions on Page 4-37 of ANP-10339P indicates the [''''''''''''''' ''''''''''''''''''''''''' ''''' ''''''''''' ''''''''' '' ''''''''''''' ''''''''' ''''''''''''
''''''''''''''''''''''' '''' ' '''''''''' ' ''''''''''''''' ' ']. The use of this type of distribution will result in ['' ''''''' ''''''''''''''''' ' ''''''''''''''' ''''' '''''''''' ''''''''' ''''''''''''''''''''''''''''' '''''' ' ''''''''''''' '''''''
''''''''''''''''''''''''''''' '''' ''''''''''' ''' '' '''''''''''''''''' '' '''''' '''''''''' '' '''''''''''''']. Provide justification that use of [ ''''''''''' '''''''''''''''''' '''''''''''''''''''''''' '''''''''''''''''' ''' ''''''''''''''' '''''''''''''' '''''''''' '' '''
''''''''''' ''''''''''' '''''''''''''''' ''''''''''''''''''''''''''' ''''''''] is acceptable.
RAI-6
Section 4.2.4.7.4, Calculation of TCS at Partial Power Conditions on Page 4-38 of ANP-10339P indicates that, during the calculation of TCS at partial power conditions, the final TCS value [' '''''''''''''''''''' ''''' '''''''''''''']. The stated intent of this is conservatism, but the
basis for the magnitude is not discussed. Provide a discussion on the basis for and determination of the [''''''' '''''''''''''' '''''''' ' ''''''''' ''' '''''' ''''''' ' '''''''''''''''''''].
RAI-7
Section 4.5 of ANP-10339P indicates the ARITA evaluation methodology makes use of the Wilks method, a non-parametric statistical approach, to determine a 95 percent probability with 95 percent confidence upper tolerance limit (95/95 UTL). Specifically, Section 4.5.1.1 discusses how the input probability distributions for the ARITA methodology ['''''' ' ''''''''''''''''''' '
'''''''''''''''' ''''''''' '''' ''''' ''']. The accuracy of the UTLs determined by the Wilks method is dependent upon representatively sampling the input probability distributions; independent NRC staff calculations indicate that ['''''''''''''''''' '''' '''''''''''''''' ''''''''''' ' '''' ''''''' '''''''''''''''''
'''''''''''''''''''''' ' '''''] may tend to underestimate the 95/95 UTL. Provide justification that the
['''''''''''''''' '''''' '''''''''''''''''''' ''''''''''''''''''''' ''''] will not result in an adverse tendency to underestimate the 95/95 UTLs.
RAI-8
The statistical approach presented in ANP-10339P uses multiple figures of merit (FOMs) to demonstrate that fuel failure will not occur. The approach proposes demonstrating failure does not occur by showing that each FOM will be individually satisfied 95 percent of the time with a 95 percent confidence. However, demonstrating that each individual FOM is satisfied on a 95/95 basis would mean that the consideration of fuel failure from all possible sources would be less than 95/95. In other words, 95/95 assurance that the event is successfully mitigated would not exist by setting individual 95/95 criterion for each FOM when multiple FOMs are relevant to that event. Justify the adequacy of the reduced probability and confidence level associated with the proposed approach of using multiple FOM UTL statements or propose an alternative means of assuring a sufficiently high probability and confidence level (e.g., 95/95) of successfully mitigating an event (e.g., a simultaneous statement considering all possible sources of failure).
RAI-9
ANP-10339P indicates the ARITA evaluation methodology is comprised of [''''''' ''''''''''
''''''''''''''''' '''''''''''''''''''''''''''''''''] and that a non-parametric statistical approach is used to make a 95/95 UTL statement. But ANP-10339P does not clearly identify all of the uncertainties that will be sampled in the statistical approach, nor which uncertainties will be sampled [' '''''' ' ''
''''''' ''''''''''''' '''''''''''''''''''' ''''''''''''''''''''''''''''']. ANP-10339P further does not identify in all cases the relevant uncertainty distributions and prescribed sampling ranges. Therefore, NRC staff cannot assess the completeness of the statistical approach or its adequacy. Provide a tabulated list of all the uncertainties being sampled in the non-parametric statistical approach and identify which will be used in each SRP Chapter 15 event analysis ['' '''''''' ''' '''''''' ''''''']. This table should identify the type of probability distribution for each uncertainty and the mean and standard deviation (or other relevant parameters as appropriate) that define the selected probability distribution. If the probability distributions are subject to change due to a dependent factor (e.g., fuel type or ['''' ''''''''''']), or if plant-specific values may be used, provide a firm method for how the uncertainty parameters will be determined (e.g., from manufacturing
tolerances, from plant specific information, etc.). In either case, justification should be provided for the chosen probability distribution, defining parameters, and prescribed sampling ranges.
The table should also identify whether each sampled uncertainty is generically applicable or plant-specific.
RAI-10
The statistical statement the ARITA methodology intends to demonstrate is not clearly articulated in ANP-10339P. However, in the post-submittal meeting held on May 8-9, 2019, Framatome clarified that the intended statistical statement is as follows:
For [a given plant], if this event occurred from within the licensed operating space, the limiting MDNBR (FCM, TCS, peak pressure, etc.) margin to the established limit is W (X, Y, Z, N) with 95% probability at 95% confidence.
Confirm the italicized quotation above represents the statistical statement the ARITA methodology intends to demonstrate or provide appropriate modifications. Justify the intended statistical statement for the ARITA methodology is sufficient to demonstrate that a reactor is operating safely at all points within its allowed operating domain.
RAI-11
[''''''''''''''''''''' '''''''''''''''''] deterministic and statistical evaluation models have generally selected bounding initial conditions and event definition parameters to ensure that a plant will satisfy acceptance criteria for analyzed events initiating at all postulated conditions in the permissible operating domain. [ '''''''''''''''' ''' '''''''''' ''''''''''''''''''''''''' ''''''''''''''''''' ' ''''''''''''''''''''' ''''
'''''''''' '''''''''''''''''''''' ''''' '''''''''' ''''' ''''''''''''' ''''''''''''''''''' '' ' ''''''''''''''''''''''''''''']
Thus, Framatomes proposal to sample ['''''''''''''''' '''''''''''''''''''''''' ''''''''' '''''''''''''''''' ''''''''''''''''''''
'''''''''''''' '''''''''''''''''''''' '' ''''''''''''''''''''''''''' ''''''''''''''' '''''''''''''' '''''''' '''''''' ''''' ''''''' ''''''''''''''''''']
For example, ['''''''''''''''''''''''''''''' ''''''' ' ''''''''''''''''''' ''''''''''''''''''' ' '''''''' '''''''''''''' ' '''''''
''' ''''''''''''''''' ' '''''''' '' '''' '''''' '''']
Considering the discussion above, please justify sampling ['''''''''''''''''''' ''''''''''''''''''''''''''''''''''''''''''''
''''''''' '' '''''''''''''' '''''''''''''''''' '''''''''''']
RAI-12
Section 4.5.1.5 of ANP-10339P indicates that if a non-parametric statistical analysis is performed where the outcome shows a failure to meet an acceptance criterion, then a reanalysis is performed with a [''''''''''''' '''''' ''''''''''''''''''' '''''''' '''''''''''''''' ''''''''''''''''' '''''''''''''''''
'' ''''''''''''''''' '''''''''''''''' '''''''''''''] and a different random sampling seed. NRC staffs concern with this approach is that it leaves open the question of whether a reanalysis that satisfies the acceptance criterion does so based upon the substance of the [''''''''''''' '' '''''' '''''''''''''''''''
''''''' ' '''''''''''''''''''''' ''''''''''''''''''''] or merely the selection of a new random seed. Passing the acceptance criterion based primarily upon the selection of a new random seed would be inappropriate because it would effectively degrade the statistical confidence level. In past reviews, NRC staff found that reanalysis under such circumstances should reuse the original random sampling seed. Provide justification that selecting a new random seed when a reanalysis is performed will not degrade the statistical confidence level. Conversely, provide justification that when a reanalysis is performed wherein the acceptance criterion is met, it is the result of a [''''''''''''''' '' ''''''' '''''''''''''''''] and not the result of a more favorable seed.
RAI-13
ANP-10339P does not identify the conditions for which a reanalysis of an operating plants safety analysis of record would be performed (e.g., due to plant operation deviating from the parameter distributions used in determining the 95/95 UTL in the safety analysis of record).
Given that it is possible for plant configuration and operation to deviate sufficiently from the underlying bases and assumptions of the safety analysis of record (e.g., including assumed uncertainty distributions) such that the analysis is no longer applicable, it appears appropriate to
establish a method to account for this. What is the threshold at which the statistical analysis of record is no longer applicable and must be reassessed, and how is it monitored to determine whether a plant has crossed this threshold?