The hybrid method of modeling the shape and pipe rolling processes

In this work the approach to the simulation of rolling processes, based on the joint application of engineering techniques, the finite element method and neural networks, called "hybrid" is presented. Its implementation in the form of software products for the cases of shape and pipe rolling is considered. A neural network to predict the value of deformation resistance of high-strength steels depending on the chemical composition, temperature, velocity and strain rate is developed. A comparison of experimental and predicted data is investigated.