Artificial neural network as an incremental non-linear constitutive model for a finite element code

A back propagation artificial neural network (BP ANN) is proposed as a tool for numerical modelling of the constitutive behaviour of a physically non-linear body. Training process of the ANN using experimental data is discussed in details and illustrated with an example. In particular, some difficulties in the constitutive description proposed in consistent, incremental form are discovered and two solutions are proposed to overcome them. Two numerical examples are presented. The first one deals with modelling of elasto-plastic hysteresis, the second shows the application of ANN to approximation of biaxial non-linear behaviour.