An artificial neural network for the analysis of profiled corrugated circular horns

An artificial neural network (ANN) analysis of profiled corrugated circular horns is proposed. This technique allows, on the basis of a reduced number of full-wave analysis, for the supervised learning of an ANN able to interpolate accurate results for a large class of horns. The performances in computing time of the ANN are far better than those of the full-wave analysis, and its memory requirements are far lower allowing for the ANN to run on any computer.