Optimization of target materials for intense compact DT neutron generators by BP neural networks

The target end of an intense compact neutron generator has to withstand power ranging from several kilowatts to even dozens of kilowatts, which demands that the target materials should have good thermal properties. Generally, the target materials are composed of a backing, a target film and a protective layer. There are multiple applicable materials for each layer, and the three layers of materials can form thousands of heat transfer combinations. The traditional evaluation of the optimal combination of target materials requires extensive testing. In this paper, the thermal properties of three different materials applied to the target system were evaluated through the back propagation neural network (BPNN), and the optimal combination of materials was obtained. The results were consistent with those from the FLUENT simulation. The SRIM software was utilized to further optimize the material of protective layer and target film. Meanwhile, the neutron yield was evaluated under the optimal combination of target materials by SRIM, and the cooling system of the target with optimal materials was also discussed.