A multi-dimensional finite element analysis of magnetic core loss in arbitrary magnetization waveforms with switching converter applications

Estimating losses accurately for magnetic materials in power conversion circuits is a challenging task. Solving the partial differential equations in a complex geometry of inductive components such as filter-integrated transformers is an additional prominent challenge. While there are some commercially available finite element simulation tools that provide loss calculations in time-domain for multi-dimensional geometries of magnetic materials, the principle of analysis is not fully disclosed and less flexible for engineers to use. In this paper, we introduce a comprehensive method for systematically and flexibly applying time-domain dynamic loss equations to geometric finite element methods in multi-dimensions. The developed post-process framework for magnetic core loss calculations is coded in an open-source programming language, Python, and verified by comparison with analytical solutions for a filter integrated isolation transformer under a variety of operating conditions. Parallel processing is used to deal with the large datasets associated with element-by-element numerical calculations in the time domain. A high level of accuracy is achieved and verified.