Single Phase PFC Topology Selection Based on Neuron Network Algorithms

The current paper presents a specialized algorithm for an expert system for converter topology selection. The algorithm is aimed at providing suggestions for an optimal topology for a single phase power factor correction converter, in terms of preferred levels of efficiency, cost and footprint size. The algorithm is developed using neuron networks and procedurally generated solution training data base build using data sets for existing semiconductor and magnetic devices. Results regarding verification and accuracy assessment of the algorithm are presented.