Reconstruct fissile material mass of pit based on activation effect of explosive and neural network

Future international nuclear disarmament may involve the dismantlement of nuclear warheads. In nuclear warhead dismantlement verification, the mass information of the fissile material in the pit is an important attribute of nuclear warheads, and can be used to verify that the nuclear warheads demanded by the nuclear disarmament treaty have indeed been dismantled. In this paper, a method of reconstructing the fissile material mass of the pit based on the activation effect of the explosive and the neural network is proposed, and may be applied in the future nuclear warhead dismantlement verification. Firstly, the number and average abundance of C-14 produced by the neutron activation reactions in the explosive inside the nuclear warhead was calculated based on the Monte Carlo numerical simulation. Secondly, it is found that the spatial distribution of the C-14 abundances in the explosive is closely related to the fissile material mass of the pit through the numerical simulation. Then, neural networks were established to reconstruct the fissile material mass of the pit through the training. The testing results show that, the fissile material mass of the pit can be reconstructed accurately based on the activation effect of the explosive and the neural network, and the reconstruction precision is better than 10%.