Digital deduction methods for ²⁵² Cf production by high-throughput fast reactor irradiation

252Cf is a high-intensity isotope neutron source with significant demand in scientific research, military development, and other fields. Currently, only the United States and Russia have the conditions for large-scale production. By using high-throughput fast reactor design parameters, irradiation target design, and energy spectrum to determine the "step-by-step cycle" irradiation scheme, an RNN recurrent neural network and NSGA-II multi-objective genetic algorithm were used to construct a digital inference method to simulate the full cycle production process of252Cf and accurately predict the yield. Finally, an optimized production plan was proposed that maximizes252Cf production while reducing or limiting the consumption of target materials. This article can provide valuable reference for the design of 252Cf irradiation schemes for high-throughput reactor production in China.