Study on many-objective optimization method for reactor 3D shielding structure based on Genetic Algorithm

BackgroundRadiation shielding design is an important part of reactor design, and the development of new nuclear power technology for various kinds of reactors has put forward new demands on radiation shielding optimization design methods.PurposeThis study aims to overcome the shortcomings of the traditional multi-objective optimization methods for shielding structures in dealing with the optimization problem of 3D shielding structures, such as slow optimization speed, difficulty in convergence, and poor globalization.MethodsBased on the non-dominated sorting genetic algorithm Ⅲ (NSGA-Ⅲ), the many-objective optimization method for 3D shielding structure design for nuclear reactor was proposed. The Monte Carlo N-Particle Transport Code (MCNP) was employed to analyze comparative performance of the NSGA-III optimization method on the basis of the 3D shielding structure model of nuclear reactors, and shield weight, volume and radiation dose rate in specific regions were taken as the optimization targets.Results & ConclusionsThe numerical simulation results show that the NSGA-III based optimization method for 3D shielding structure design can search for the Pareto-optimal front more efficiently and stably, providing a new idea for the optimization of radiation shielding design.