Finite Element Based Surrogate Modeling and Irradiation Capsule Optimization for Large-Scale Neutron Irradiation Campaigns

摘要

Large-scale material irradiations like the graphite qualification project between ORNL and X-Energy require efficient utilization of all available reactor positions to achieve temperature and fluence experiment goals in a timely manner. Designing and optimizing untnstrumented irradiation capsules for multiple positions using multiple materials and dimensions can be time intensive and unsystematic. This work describes a surrogate model developed to characterize and predict irradiation capsule thermal performance which covers the primary design variables. This model uses a central composite design of experiments to solve FEA models and fit those results to a 2nd order polynomial response surface using linear least square regression. Optimizing the response surfaces for two capsule parameters (average temperature and temperature range) provides confidence that the best capsule designs are being utilized. Future optimization work will expand these techniques to 3-D capsule models and incorporate core dose rate location dependence and discrete reactor position availability.