Application of Sequential Unconstrained Minimization Technique to LMFBR Core Design Optimization Problem

Abstract

This paper describes an LMFBR core design optimization procedure, where the sequential unconstrained minimization technique (SUMT) using the conjugate direction method for the unconstrained minimizer is adopted as the optimization algorithm. To minimize computing effort in estimating the equilibrium core condition, the procedure utilizes the regression functions of the burn up characteristics in a simplified method of calculation, which is called the “power balance method”. As a numerical example, the procedure was applied to calculations on a 1,000 MWe LMFBR core design, on which an evaluation is also made of the sensitivity shown by the results to changes in critical constraints. It is shown that the optimization algorithm based on SUMT is well applicable to core design optimization, provided careful attention is paid in treating the penalty term of the transformed performance index in order to obviate a discontinuity occurring in the locus of the optimized solution at the constraint boundary. It is also demonstrated that the simplified method of burnup calculation is helpful in reducing computer time. Further, some considerations are presented on such aspects as the behavior of the design variables, computing effort required, and the main characteristics of the optimized core.