APPLICATION OF NONLINEAR NODAL DIFFUSION GENERALIZED PERTURBATION-THEORY TO NUCLEAR-FUEL RELOAD OPTIMIZATION

The determination of the family of optimum core loading patterns for p ressurized wafer reactors (PWRs) involves the assessment of the core a ttributes for thousands of candidate loading patterns. For this reason , the computational capability to efficiently and accurately evaluate a reactor core's eigenvalue and power distribution versus burnup using a nodal diffusion generalized perturbation theory (GPT) model is deve loped. The GPT model is derived from the forward nonlinear iterative n odal expansion method (NEM) to explicitly enable the preservation of t he finite difference matrix structure. This key feature considerably s implifies the mathematical formulation of NEM GPT and results in reduc ed memory storage and CPU time requirements versus the traditional res ponse-matrix approach to NEM. In addition, a treatment within NEM GPT can account for localized nonlinear feedbacks, such as that due to fis sion product buildup and thermal-hydraulic effects. When compared with a standard nonlinear iterative NEM forward flux solve with feedbacks, the NEM GPT model can execute between 8 and 12 times faster. These de velopments are implemented within the PWR in-core nuclear fuel managem ent optimization code FORMOSA-P, combining the robustness of its adapt ive simulated annealing stochastic optimization algorithm with an NEM GPT neutronics model that efficiently and accurately evaluates core at tributes associated with objective functions and constraints of candid ate loading patterns.