An adaptive DE algorithm and its preliminary application in intelligent design of space reactor

Abstract

BackgroundThe space reactor has broad application prospects, and its structure optimization design has an extremely high economic and technical application value. Due to the high coupling degree of variables and complex design in the optimal design of the space nuclear reactor structure, the manual solution process is inefficient and time-consuming, and the optimal solution is hard to achieve.PurposeThis study aims to propose an optimal algorithm for structure design of space reactor.MethodsAn adaptive differential evolution (DE) algorithm based on the classic DE algorithm was proposed. The strategy of balancing global and local search capabilities by dynamically adjusting the control parameters was adopted to enhance the algorithm optimization ability. Finally, with structural parameters as variables and the core quality as objective function the Kilopower reactor model with a thermo power of 4.3 kW was optimized, and the neutron transport was calculated by combing the neutron transport design and safety evaluation software SuperMC simultaneously.ResultsThe core quality of the optimized space reactor was reduced by 38.2% when the design requirements were met.ConclusionsThe correctness and effectiveness of the adaptive DE algorithm proposed in this paper are verified in the design of space reactor core.

Keywords

• differential evolution algorithm
• adaptive
• core optimization
• kilopower