Inductor Design for Multiphase Bidirectional DC-DC Boost Converter for an EV/HEV Application

Magnetics design is a major factor within the design of any high-power converter due to its large volume, weight, and cost. Typically inductor design is a time-consuming iterative process. Therefore, any technique which speeds this process up will help reducing the time spent for the design. As such this paper discusses typical inductor design, and proposes a population based optimization technique for a high power inductor for an EV/HEV application. The Genetic Algorithm (GA) optimization technique will be used to design inductors implementing powdered iron core (FeSi) with rectangular wire and ferrite core with litz wire at different operating frequencies and for different number of phases for a 40kW nominal, 60kW peak bidirectional DC-DC boost converter. The comparison will show what phase number at what operating frequency provides the lowest inductor volume. This paper will also discuss a lumped parameter thermal network for the core temperature estimation which will be used within the Genetic Algorithm.