Застосування нейромережевої моделі для прогнозування стану матеріалу при високошвидкісному деформуванні

Abstract

The neural network's model (NN-model) of determining equations of the plastic state of material is in-process presented for the wide range of deformations, speeds of deformations and temperatures. Realization of model is carried out through the neural network-3-th layer of perseptron with one by the hidden layer. It is got optimum architecture, teaching of NN-model is conducted on the base of experimental information for steel 45. For the verification set of data comparing is conducted to the results has been realize with Johnson-Kuke quation.