Experimental Verifications for a Multiobjective Shielding Design Method on a p-Be Neutron Source With Injection of 7 MeV Protons

NSGA-MC as a multiobjective shielding design method can optimize the shielding parameters aiming at multiple objectives through evolutionary computations by Monte Carlo N-Particle (MCNP) calculations and a genetic algorithm, nondominated sorting genetic algorithm II (NSGA-II). In our previous study, NSGA-MC has been proven to be effective through simulation results by applying to the target shielding design of a compact accelerator-driven neutron source. However, to confirm the effectiveness of NSGA-MC, experimental verifications are critically important. Thus, we planned to experimentally study sets of multilayer shielding structures consisting of PE, B4C, and Pb layers and optimize with NSGA-MC with two objectives of equivalent dose rate and material density. Three cases identified from the Pareto fronts obtained by NSGA-MC and a nonoptimized case based on RANS target station shielding configuration were tested by utilizing RIKEN accelerator-driven compact neutron source (RANS). The dose rates of neutron and gamma ray after neutron beam penetrated the shielding material were measured with commercially available neutron and gamma-ray dosimeters based on these shielding configurations. Discussions were made on experimental results and simulation results. Validation was assured by this experiment within 42%.