The adjoint problem as physical heuristic for loading pattern optimization

Loading pattern optimization enables longer irradiation cycles while still maintaining the ability to safely operate and shutdown the reactor. The loading pattern optimization problem is multi-objective and is characterized by a huge nonlinear discrete search space. Its complexity grows exponentially with the number of fuel assemblies in the core. However, in-core fuel management is an essential part of routine work in a nuclear reactor, hence optimized usage of the fuel inventory enables the economic and efficient utilization of resources. This study is a proof-of-concept for the hypothesis that adjoint-based neutron importance functions can be used for the optimization process of the core loading pattern. New optimization techniques are developed in order to demonstrate the successful utilization of adjoint-based functions as the optimization driving force. Different importance functions are developed and studied. It is demonstrated that the physical insight obtained from the importance function can be used for the optimization of loading patterns. Eventually, this new technique should be integrated into a stochastic optimization algorithm, e.g., evolutionary algorithms, in order to accelerate and improve existing optimization algorithms.