Modeling and Analysis of LLC Resonant Converters under Strong Magnetic Fields: A Multi-Physics Approach with Machine Learning

This paper addresses the interference of strong magnetic fields in controlled nuclear fusion systems on power electronic converters, focusing on the LLC resonant converter. Through experimental testing and multi-physics modeling, the mechanisms by which the external magnetic field affects the parameters of magnetic components and the dynamic characteristics of the converter are revealed. The study demonstrates that the external magnetic field induces nonlinear attenuation of the magnetizing inductance in the LLC transformer, leading to resonant frequency drift, current spikes, and reduced energy transfer efficiency. A magnetic field-circuit coupling model, constructed using finite element simulation and neural networks, validates the dominant role of magnetic component parameter variations in converter failure. This research provides theoretical and methodological support for the design of highly reliable power electronic equipment in strong magnetic field environments.