Coupling PZMI, Neural Network and Genetic Algorithms to solve EMC problems

Solving problems related to electromagnetic radiation of three-dimensional systems is very complicated. This is due to the strong nonlinearity of the mathematical equations related to the radiated field. In this paper, a novel algorithm based on coupling the PZMI and Neural Network with the inverse electromagnetic method based on Genetic Algorithms is proposed to identify radiation sources. The proposed coupling method will be explained and will be applied to a realistic example. It has the advantage to use several times the Genetic Algorithm Method with for each time, a reduced number of parameters to identify. By this way, the convergence of the Genetic Algorithms is assured and the resolution time of the global approach is extremely reduced.