Developing a safe and high performance fuel management optimization for MTRs using stochastic knowledge base searches

• Developing a safe and high performance fuel management optimization.
• Standard safety issues are introduced coherently and influenced effectively.
• High reliable (severally verified and benchmarked) core modeling.
• Involving and interpreting large numbers of core parameters coherently.
• Introducing and analyzing different models of artificial material forming.

Each core configuration of a nuclear Material Testing Reactor (MTR) can be optimized to provide desired irradiating conditions and retain corresponding Operational Limits and safety Conditions (OLCs). In this paper refueling cycle length and thermal neutron fluxes are chosen as the optimizing objectives; also OLCs including total Power Peaking Factor, Shutdown Margin, Reactivity Safety Factor (RSF), and maximum permissible core excess reactivity are chosen as optimizing constraints. This is one of the most complex types of an engineering design problem as a non-continues, combinatorial, multi-modal, multi-objective, and constrained optimization. Stochastic knowledge base searches have been introduced and tried throughout an accurate core modeling. Selection rules are chosen based on a memorial random walk, two different models of glass and crystal transition, a wide duplicated fuzzy Gaussian distribution updating in accordance with the wrong distance, and a Simulated Annealing (SA) and re-annealing process for artificial metals. An expertise and conservative initial solution is fed to the optimization process. States of developing knowledge-based are rechecked and studied according to the best solution. Safety margins are influenced by stepwise penalty functions instead of a direct rejecting method. All parameters and optimization process are visualized and studied coherently; also decision and criterion spaces are implicitly projected and analyzed. Results are very promising. Required iterations and computational costs are decreased. Safety faults are automatically removed; and final results are gained near touch the infeasible frontier formed by safety margins. Refueling cycle length significantly increased. Averaged and maximum values of irradiating neutron fluxes enhanced, while selected OLCs completely passed. On the other hands, formed lattice material and type, lattice defects, appropriate annealing process, and gained material properties are artificially simulated and described.