Abstract
Japan Atomic Energy Agency is developing an evaluation system aided by artificial intelligence (AI) named ARKADIA (Advanced Reactor Knowledge- and AI-aided Design Integration Approach through the whole plant lifecycle), aiming to offer solution for challenges to the design and operation of a nuclear plant. A sub-system of it, named ARKADIA-Design, is being developed to support the design optimization study for an advanced nuclear plant including a sodium-cooled fast reactor (SFR). Authors are developing a design optimization process for the structure of the component in SFR. This paper describes the outline of a design optimization process, the brief introduction of evaluation methods for the process, and the result of the demonstration of the optimization process for a feasibility study. The development is being performed in a representative problem considering the thermal transient and seismic motion as a major issue in SFRs. In the optimization process, the failure probability is used as the element of the objective function for optimization, to compare the contribution rate of the loads to the failure that occurred in different mechanism in an equation. Since a number of evaluations for the uncertainty quantification in the estimation of the failure probability is needed, the simplified evaluation method for failure probability in the optimization process is being developed to reduce evaluation cost. And iterative evaluations for updating variables of design parameters to achieve the optimal solution are needed. The automatization function in the optimization process is being developed for efficient execution of the optimization process. Through the demonstration, it was confirmed that the optimization process under development may provide an optimal solution to the representative problem.
