Optimization of Marine Pressurized Water Reactor In-Core Fuel Management

Abstract

An advanced optimization method based on the use in tandem of the Gauss-Newton non-linear search method and of a direct steepest descent technique is presented far optimizing the refuelling parameters of a Marine Pressurized Water Reactor. The objective function is formulated as the weighted sum of squared parameters such as the desired cycle length and the deviation of the radial power peaking from average. The decision variables of the optimization problem are the enrichments of the fresh fuel allotted to the various pre-determined refuelling zones in the core. In addition, the problem includes constraints such as maximum enrichments. A Marine PUR was mathematically modelled using the computer codes WIMS-AECL, CITATION and CALYPSO, and the optimization itself was carried out using a specifically written FORTRAN program (MARCORE). The method is typically able to achieve reductions in the objective function of up to 90% in as little as two or three iterations. This corresponds to achieving a cycle length within 1% of that desired, and an overall reduction of the deviation of the radial power density from average of 70%.