Prediction of the influence of parameters of JWL equation of state on EFP velocity based on BP neural network

DOI:	10.11809/bqzbgcxb2023.06.007
Keywords:	JWL equation of state; neural network; explosively formed projectile (EFP); velocity
Abstract:	To quickly and accurately predict the velocity of an explosively formed projectile (EFP) and calibrate the parameters of the JWL equation of state, this paper uses LS DYNA finite element software to establish a numerical model of Φ 65 mm EFP, and analyzes the influence law of the parameters of the JWL equation of state on EFP velocity. On this basis, the calculation formula of the EFP velocity is fitted based on the parameters of the JWL equation of state, a BP neural network prediction model of the EFP velocity is constructed, and the learning samples are generated to train the network model. The results show that the parameters of the JWL equation of state have a great influence on the EFP velocity. R1 has the greatest influence and the order of the influence of each parameter is R1, A, B, R2, E 0 and ω. Both the calculation formula and the network model can accurately predict the EFP velocity, and the prediction accuracy of the latter is better than that of the former. Most of the errors of the calculation formula can be controlled within 10%, and the prediction errors of the network model are not more than 5%.
Published:	2023-06-28
PDF HTML