A new regression modeling method for PMSLM design optimization based on K-Nearest Neighbor Algorithm

Abstract

This paper focuses on the rapid regression modeling research for the design optimization of Permanent magnet synchronous linear motors (PMSLM) which are applied in the linear motion machines. Based on the Finite Element Analysis (FEA), the initial PMSLM model is built and the modeling data of optimization are obtained. Then, the machine learning regression algorithm named K-Nearest Neighbor Algorithm (KNNA) is introduced to mapping the nonlinear relations of structure parameters and performances, and establishes the rapid calculation model for the next optimization. Finally, the superiority and reliability of this method is confirmed by the FEA experiments.