Component Loss Analysis Based on Particle Swarm Optimization Algorithm and RFR

The purpose of this paper is to investigate the component loss modeling and optimization methods based on Particle Swarm Optimization (PSO) and Random Forest Regressor (RFR). The core loss model has limited prediction accuracy under complex operating conditions, so this study utilizes a data-driven approach to analyze the core loss of magnetic components by machine learning algorithms such as GBDT, OLS, and RFR, and introduces the PSO algorithm for optimization. A model that can accurately predict the core loss is constructed in the study by combining multiple operating parameters. Through experimental validation, the prediction accuracy of the model under multiple working conditions is significantly improved, and it has good generalization ability. In addition, in order to solve the multi-objective optimization problem in component design, this paper achieves the dual optimization objectives of loss minimization and energy transfer maximization by calculating the weights of core loss and transmitted magnetic energy. The results provide theoretical support for the efficient design of magnetic components, as well as a reference for the modeling and optimization of complex systems.