ML24215A212
Text
Response to SDAA Audit Question Question Number: A-19.1-23 Receipt Date: 04/24/2023 Question:
Section 6.1 "Overall Results" in the SDA Probabilistic Risk Assessment Main Report states,
(( 2(a),(c) The justification in EPRI 1016741 related to modeling of common cause failures states, Common cause failures as a means of triggering multiple train failures at the same time should be relegated to those either actually observed, or those for which a credible mechanism can be identified. Table 3-1 can be used as an aid for identifying components susceptible to common cause failure. For contributors that are common cause failure related, however, these events are so rare, that there is not usually sufficient historical information upon which to base a frequency. Thus, a more appropriate approach is to use the existing CCF data (alpha factors) as is. If this results in a support system initiating event frequency inconsistent with operating experience, then refine the data as described in Section 2.3.3 until the result is consistent with operating experience.
- 1. In addition to EDAS, provide a list of any other systems that use the guidance provided in EPRI 1016741 Section 3.1.3 for the quantification of single train and dual train initiating events or for the failure probability of mitigating systems.
- 2. EPRI 1016741 was published in 2008. Justify that the approach and conclusions in this report are unchanged due to CCF occurrences and data over the past 15 years.
- 3. Since NuScale does not have operating experience, the EDAS is a first-of-a-kind system, and VLRA-type batteries do not have previous applications in the nuclear industry, describe how (i)
NuScale Nonproprietary NuScale Nonproprietary
(( }} 2(a),(c) was determined to be consistent with operating experience, and (ii) uncertainty was captured in the determination.
- 4. If ((
}} 2(a),(c), describe and justify how EPRI 1016741 is applicable to (( }} 2(a),(c). If EPRI 1016741 is not the basis for (( }} 2(a),(c) and justify its applicability.
Response
- 1. The guidance provided in EPRI 1016741, Support System Initiating Events Identification and Quantification Guideline, Section 3.1.3 is only applied to ((
}}2(a),(c) considered in the loss of DC power initiating event.
- 2. The approach and conclusions are consistent with operating experience data. Only one DC bus failure (a single, independent failure; not a CCF) is identified in INL/EXT-21-65055, Industry-Average Performance for Components and Initiating Events at U.S. Commercial Nuclear Power Plants: 2020 Update. Similarly, as documented in INL/EXT-21-62940, CCF Parameter Estimations, 2020 Update, the data show a steady and significant reduction in CCF event occurrences over the years.
- 3. As described in part (2) of this response, the NRC operating experience includes only one DC bus failure and no DC bus CCFs.
In the NuScale configuration, there are no credible mechanisms that could trigger multiple division failures involving passive equipment at the same time. The augmented DC power system (EDAS) contains four power channels. Power channels A and C, on Division I, are located ((
}}2(a),(c) Power channels B and D, on Division II, are located (( }}2(a),(c) Each room also has an independent heating, ventilation, and air conditioning system.
The EDAS receives power from battery chargers with battery backups, and also include ties between DC buses within the same division. Failures involving active components (e.g., battery chargers) result in different consequences than failures of passive components (e.g., buses). On a loss of supply power, the batteries allow the module protection system to function normally NuScale Nonproprietary NuScale Nonproprietary
and systems to actuate on-demand. Conversely, de-energization of EDAS buses will result in safety system actuation (i.e., fail-safe design). Uncertainty is captured in the initiating event frequency being modeled using a lognormal probability density function with an error factor of 10. An error factor of 10 represents a broadening of the uncertainty in the estimate of the initiating event frequency beyond what would be calculated from the generic data alone. The impact of this uncertainty on the initiating event frequency is reflected in the parametric uncertainty results (ER-102082, Quantification Notebook, Section 6.1.1). A sensitivity study was also performed that doubles the loss of DC power initiating event frequency (ER-102082, Section 6.2.2.1).
- 4. In the PRA modeling of EDAS power as a support system, ((
}}2(a),(c) Consistent with the discussion in Section 4.1 of EPRI 1016741, there are special considerations for electrical bus support system initiators. Because electrical buses are in continuous use, a fault is almost immediately known.
When in normal operation, there are no active components associated with a bus, and therefore ((
}}2(a),(c) In addition, the PRA considers loss of a single alternating current bus (EMVS-EBA-2000X) that supplies power to both divisions of EDAS, and this loss of support system initiating event frequency is much more likely than a CCF of 4 of 4 EDAS buses.
No changes to the SDAA are necessary. NuScale Nonproprietary NuScale Nonproprietary}}