“Begging the Question” in Licensing Basis Accident Analyses

Abstract

“Begging the question” describes a situation in which the statement under examination is assumed to be true (i.e., the statement is used to support itself). Examples of this can be found in analysis reports that were prepared by analysts who are not mindful (or maybe uninformed) of the analysis criteria they’re required to fulfill. This is generally seen in analyses of anticipated operational occurrences (AOOs). AOOs are defined in Appendix A of 10 CFR §50 [1], and in ANS-N18.2-1973 [2], where they’re also known as American Nuclear Society (ANS) Condition II events. This standard [2] also defines more serious, Condition III and IV events. Analyses of AOOs, or ANS Condition II events are required to show that: (1) reactor coolant system (RCS) pressure will not exceed its safety limit, and (2) no fuel damage will be incurred, and (3) a more serious accident will not develop, unless there is a simultaneous occurrence of another, independent fault. The three requirements are often demonstrated by three different analyses, each of which is designed to yield conservative results with respect to one of the requirements. Accident analyses that are performed to demonstrate compliance with the first two requirements are relatively straightforward. They rely mostly upon the design of safety valves and the timing of reactor trips. “Begging the question” is seen in analyses that are designed to demonstrate compliance with the third requirement. This paper will describe how this logical fallacy has been applied in licensees’ accident analyses, and accepted by the NRC staff.