Neuron network simulation of seismic waves travel time curves
© M.A. Lazarenko, O.A. Gerasimenko
The implementation of artificial neural networks for travel-time model of P- and S-phases of seismic waves arrangement is proposed. The principles of multilayer, fullconnected, feedforward, controlled, and backpropagated neuron network functioning and approach to net topology choice and extrapolation error assessment are considered. The 3D travel-time relationships for various scenario of seismic process using the travel-time inversion based on Ukrainian seismic stations records are considered. The examples of Herglotz-Wiehert inversion for single stations as well as for arbitrary source and station coordinates in circum-Black Sea region are presented.