A Novel Signal Processing Method of Sliding Mode Observer for Position Sensorless PMSM Drives

In order to improve the performance of permanent magnet synchronous motor (PMSM) driven by magnetic field directional control, an improved sliding mode observer (SMO) signal processing method is proposed. A structure based on second order generalized integrator (SOGI) is designed to extract the positive sequence components of the observed the back electromotive force (EMF). Compared with the traditional low-pass filter (LPF), the proposed structure has better filtering effect and almost no phase lag. Meanwhile, a phase locked loop (PLL) based on tangent function is proposed. Compared with the traditional PLL, the cutoff frequency of the tangent function based PLL is not affected by rotation speed and has a faster convergence rate. Simulation and experiments verify that the improved signal processing method of sliding mode observer can quickly and accurately track the position and speed of motor rotor, and the system has the characteristics of fast response and strong robustness.