Design Automation of Power Electronic Converters a Grid Elitist Multiobjective Genetic Algorithm

Power Electronic Design Automation is a methodology using optimization algorithms that selects optimal parameters and components achieving the highest power densities and efficiencies for the lowest prices. The paper presents a MATLAB based tool for design automation of power electronic converter. It implements a grid elitist multiobjective genetic algorithm to achieve optimal design of power converters. It uses libraries of existing components for semiconductors, heat-sinks, capacitors, magnetic materials, magnetic cores, AWG wires, gate drivers, signal isolator to feed the models. The output of the optimization is a Pareto, best tradeoff between several objectives respecting numerous constraints, e.g. EMC standards, thermal limits of components, available volume. The designer can finally pickup a feasible solution among all the optimal ones where each solution is a bill of materials. This reduces dramatically the time to market. Success stories of industrial products automatically designed by the platform are presented at the end of the paper.