Bundle enrichment minimization via multi-objective simulated annealing

This study presents the implementation of a unique multi-objective approach to the simulated annealing optimization algorithm within a light water reactor (LWR) nuclear fuel bundle design application. This multi-objective optimization technique has been tested within the computer code FORMOSA-L, which consists of an optimization driver coupled to a lattice physics evaluator (CPM 2) with the purpose of optimizing the 'pin by pin' placement and/or loading of nuclear fuel, burnable absorbers, or other material regions within a LWR fuel assembly. In this study, two objective functions are minimized simultaneously: the average enrichment and the local power peaking of the fuel bundle. Furthermore, upper and lower constraints on assembly average k∞ are enforced to enable the preservation of a targeted k∞ profile versus burnup by the optimized fuel bundles. (author)