Considerations on a Comparison of Integrated Safety Analysis to Probabilistic Risk Assessment

revising the fuel cycle oversight program directed staff to produce a concise paper that was a critical evaluation and comparison of Integrated Safety Analyses (ISA) and Probabilistic Risk Assessment (PRA). The SRM to SECY-10-0031 stated that the Commission expects the paper " to better inform proposed enhancements to the oversight process. " The following are NRC staff considerations for conducting such a comparison. EXECUTIVE SUMMARY ISA Description ISA is a systematic analysis required by 10 CFR 70 Subpart H for major fuel cycle facilities to identify all accident sequences leading to high or intermediate consequences, and Items Relied On For Safety (IROFS) to prevent or mitigate these accidents. Section 70.61 explicitly defines high and intermediate consequences. For example, a high-consequence event for a worker is a life threatening chemical exposure or a dose greater than 100 rem. An intermediate-consequence event for a worker is a chemical exposure that produces serious health effects, or a dose from 25 to 100 rem. Accidental criticality, radiological, and chemical exposures are to be considered. ISAs are to identify sufficient IROFS to demonstrate that high-consequence events are " highly unlikely " and intermediate are " unlikely " , as defined by the licensee. There is no requirement that accident sequence likelihoods be quantified as frequencies; nor that they be multiplied by consequences and summed to obtain an estimate of risk to individuals. Definition of the terms " unlikely " and " highly unlikely " is left to the licensee to propose, subject to the U.S. Nuclear Regulatory Commission's (NRC) approval. The NRC provided guidance (References 1 and 4) as to some acceptable ways to comply with these requirements. Although the NRC produced these guidance documents, NRC did not develop the basic methods, such as HAZOP or fault trees. They describe, nor perform tests of ISAs, but endorsed techniques already in use for compliance with OSHA regulations. What NRC provided in the Standard Review Plan, Revision 1 of 2010 (Ref. 1) is extensive and detailed guidance on how to execute these methods in practice, based on experience with the actual ISAs over the last 10 years. In particular, likelihood evaluation guidance is provided in Reference 1, Chapter 3 and its appendices for quantitative, qualitative, and risk index methods. Some PRAs may limit the scope of their analysis, for example, to internal events only, because this is sufficient for some particular applications. ISAs are intentionally not limited …