A Hysteresis Periodic Magnetic Field Solution using Neural Network Model and Fixed Point Technique

摘要

This paper presents a time-periodic finite element procedure for the solution of hysteretic magnetic field problems, under periodic steady-state supply, which is the most usual operating condition for the electrical devices. The approach, based on the neural network hysteresis model, employs the fixed-point technique for handling magnetic nonlinearity. The approach is applied to a two-dimensional model problem, The results are validated by comparison with the exact analytical solution.