Prediction of the eddy-current and temperature distribution in a TFIH device using neural networks in order to improve the convergence of the finite element calculations

This paper presents a method of predicting the eddy-current distribution in transverse flux inductive heating devices (TFIH) with the help of one neural network. A second neural network is used to obtain the temperature distribution in the thin moving conducting sheet caused by the eddy-current losses. Both solutions are initial solutions for the finite element calculations of this nonlinear coupled electromagneto-thermal problem.