Use of Statistical Theory of Extremes in Hot Channel Analysis of a Liquid Metal Fast Breeder Reactor

Abstract

In the thermal hydraulic design of nuclear reactor cores, it is of interest to know the probability for 0, 1, 2, . . ., D hot channels and/or cladding and fuel hot spots [i.e., channels (spots) in the core at which temperature limits are exceeded]. A previous paper considered this problem and provided a technique, referred to as the method of correlated temperatures, for obtaining the distribution of the number of hot channels. This method is partly analytical and partly Monte Carlo. In the present paper a special case, that of zero hot channels, is considered and it is shown that by application of the theory of extremes numerical results can still be obtained without the use of Monte Carlo computations proposed earlier. A hot channel factor analysis is carried out using the proposed method on a simplified hypothetical LMFBR-type core and the results are compared with those obtained (a) from the method of correlated temperatures and (b) Amendola’s method. The method based on extreme value theory compares very favorably with the more general method of correlated temperatures.