A novel method to represent the saturation characteristics of PMSM using Levenberg-Marquardt algorithm

Inclusion of saturation in the permanent magnet synchronous machine modeling significantly affects the accuracy and reliability of the analyses. Therefore, an accurate representation of the saturation characteristics in the machine model is important. In this paper, a non-linear optimization algorithm is proposed to represent the saturation characteristic of electrical machines based on the experimentally measured saturation characteristics data. This method is applied to a laboratory permanent magnet synchronous machine. The results are compared to the saturation characteristic curve using discrete Fourier transform. The error analysis demonstrates the effectiveness of the proposed algorithm.