Wavelet-Based Filtered-X LMS Algorithm for the Control of Permanent Magnet Synchronous Motors

Permanent Magnet Synchronous Motor (PMSM) can behave chaotically for a certain range of its parameters. To improve its dynamical behavior and enable a robust control of the rotor angular speed, a novel method combining the wavelet transform with the filtered-x LMS algorithm is presented in this paper. Without linearizing the model so as to not advertently misinterpret the underlying dynamics, the method can identify the nonlinear PMSM model with adaptive filters in real-time and guarantee a comprehensive control in both the time and frequency domains. Firstly, the physical PMSM model is analyzed and its chaotic behavior without control is investigated. The wavelet-based filtered-x LMS is then applied to the nonlinear PMSM system subject to desired angular speeds that are constant and varying harmonically in time. Numerical studies show that chaotic behaviors are effectively mitigated and the system output matches the desired angular speed after the initial transient period, thus demonstrating the feasibility of the method for the control of PMSMs.