Emulation of high-speed plate collision with an artificial neural network

Based on a continuum model of high-speed impact of plates, a set of training data is constructed, according to which an artificial neural network was trained to determine the velocity profile of the rear surface of the target plate from the impact parameters and parameters of the material model. The trained neural network was used as a fast emulator of high speed plate impact. The use of the Bayesian approach to the model calibration made it possible to solve the inverse problem of determining the parameters of the material model from the velocity profile of the rear surface.