Core reload optimization for equilibrium cycles using simulated annealing and successive linear programming

Abstract

An algorithm is developed to determine directly all the parameters of the optimal equilibrium cycle. The core reload scheme is described by discrete variables, while the cycle length, as well as uranium enrichment and loading of burnable poison in each feed fuel assembly, are treated as continuous variables. An important feature of the algorithm is that all these parameters are determined by the solution of one big optimization problem. To search for the best reload scheme, simulated annealing is applied. The optimum cycle length as well as uranium enrichment and loading of burnable poison in each feed fuel assembly are determined for each reload pattern examined using successive linear programming. The uranium enrichments and loadings of burnable poison are considered to be distinct in different feed fuel assemblies. The number of batches and their sizes are not fixed, and also determined by the algorithm. As the first step of the numerical investigation of the algorithm, a problem of feed fuel cost minimization for a target equilibrium cycle length and fixed batch sizes is considered. The algorithm developed is demonstrated to provide about 2% less feed fuel cost than the ordinary simulated annealing algorithm.