An artificial neural network system for predicting the deformation of steel plate in triangle induction heating process

This paper presents the development of an artificial neural network (ANN) model to predict the positions and sizes of induction triangle heating based on laminated plate theory combined with Finite Element Method (FEM) solutions. The vertical displacements and transverse shrinkage of nodes on the plate are used as inputs to the network and the position of the heating triangles on the plate, heated surface, and size of the heating triangles are the outputs of the models. The developed feed-forward ANN with 242-100-72 arrangement is capable of estimating all necessary conditions for the induction triangle heating process. The training patterns of the neural network are obtained using an analytical solution derived from the plate laminate theory to predict the plate deformations in the induction heating process. The developed neural network model is tested to demonstrate its feasibility for determining the heating positions on the surface of a flat steel plate in the triangle induction heating process for forming a desired shape.