Development of core design optimization process: Feasibility study of multivariable optimization via integrated sequential analyses of neutronics, thermal-hydraulics, and fuel integrity evaluation

• Optimization system for core designs combining integrated sequential analysis and Bayesian optimization is being developed. • Integrated sequential analysis system of neutronics, thermal-hydraulics, and fuel integrity evaluations was developed. • Optimization system was tested with simplified optimization problems of core design. • Optimization system could reveal the optimal solutions of simplified core design problems. The Japan Atomic Energy Agency has been developing an innovative design approach for advanced reactors such as fast reactors, known as Advanced Reactor Knowledge- and AI-aided Design Integration Approach through the whole plant lifecycle (ARKADIA). One task of ARKADIA is to establish a core design optimization process to automatically identify optimal core and fuel design parameters by combining an optimization method and integrated sequential analyses of neutronics, thermal-hydraulics, and fuel integrity evaluations as well as plant dynamics analysis. The optimization process has been developed in stages. In a previous study, the optimal solution consistent with the reference solution was obtained in a simple two-variable optimization problem by focusing only on neutronics. Herein, the optimization process was extended to multivariable optimization, including other analyses. In particular, an integrated sequential analysis system was developed to evaluate thermal-hydraulics and fuel integrity as well as neutronics in the core. The number of core design variables was increased from two to four. The extended optimization process was applied to two problems of three- and four-variable optimization with multiple constraints. In the three-variable problem, the validity of optimization calculation was shown by the optimal solution matched to the reference solution. In the four-variable problem, the solution satisfied all the defined constraints. These results confirmed the feasibility of the core design optimization process combined with integrated analyses up to four variables.