Presizing Methodology of DC-DC Converters Using Optimization Under Multiphysic Constraints: Application to a Buck Converter

This paper presents an approach to presize dc-dc converters. This approach is carried out in three main steps. The first one helps the designer to select the most adapted architecture answering the specifications and appropriate technologies of passive and active components of the selected architecture minimizing a major constraint (like volume or cost). Models used in this step are developed from manufacturer datasheets. The second step consists on optimizing the selected architecture by considering the appropriate technologies resulting from the first step under multiphysic constraints: thermal, losses (or efficiency), volume, and electromagnetic compatibility. Analytical models to consider the different constraints are developed. The third step carries out a second optimization with integration of a finite element simulation. It allows refining the optimization results of the second step and consists on optimizing the components placement under thermal and EMC constraints together or separately. The developed methodology proposes a progressive approach to answer industrial needs when presizing power converters. It allows reducing significantly the computing time which is suitable in a presizing phase where several parameters must be optimized and multiphysic constraints must be respected.