Multi-Objective Optimization Design of Plate-type Fuel via Co-simulation Method.

• By combining the numerical simulation and multi-objective evolutionary algorithm, a co-simulation method is developed to the multi-objective optimization design of plate-type fuel. • The obtainment of Pareto optimal solutions and Pareto front proves the ability of co-simulation in actual engineering design. • Through the qualitative and quantitative analyses of the Pareto fronts, the co-simulation with NSGA-II or CMOEA-MS has better property about the design and thermal–hydraulic performance of the fuel plate. • The co-simulation method could modify the multiphysics model or add objective functions and constraints flexibly to expand the optimization objectives. Plate-type fuel element has been implemented and studied widely because of its good heat transfer capability. For the actual design of fuel assembly, the design objectives usually to be in competition, and the constraints also need to be considered. Therefore, this is a challenging constrained multi-objective optimization problem. A co-simulation method has been developed in this study to optimize the design of fuel plate for the thermal–hydraulic performance. Moreover, due to the versatility of this co-simulation method, extra objectives and constraints could be added as needed. First, a two-dimensional thermal-fluid–solid multiphysics model of plate-type fuel assembly was established and solved by the finite element method. After that, the values of relevant physical variables were verified with the three-dimensional model. Finally, the numerical simulation of multiphysics model was incorporated with the multi-objective evolutionary algorithm (NSGA-II, adaptive NSGA-III and CMOEA-MS) which could handle constraints to optimize the thickness parameters of fuel plate. Pareto optimal solutions were obtained separately after the co-simulation with three algorithms, and the qualitative and quantitative analyses of the Pareto fronts showed that CMOEA-MS and NSGA-II had better performance in this study. This co-simulation could provide comprehensive and reliable design information according to the design objectives and requirements. [ABSTRACT FROM AUTHOR]