Position and speed estimation for high speed permanent magnet synchronous motors using wideband synchronous fundamental-frequency extraction filters

The non-ideal factors such as the inherent chattering of a sliding mode observer (SMO), the delay of a control algorithm, and the dead-time effect give rise to the harmonic error of position estimation. To improve the performance of the position observation of high-speed permanent magnet synchronous motors (HSPMSMs), a wideband synchronous fundamental-frequency extraction filter (WSFEF) is proposed. On this basis, a novel signal processing method consisting of a WSFEF-PLL is applied to extract the fundamental frequency signal of the estimated back electromotive force (EMF). The application of a phase-locked loop (PLL) ensures that the resonance frequency of the WSFEF is adaptive, which is essential for the variable-speed operation in sensorless HSPMSM drive systems. Using the WSFEF-PLL in a SMO-based position estimator, the rotor position estimation error caused by the harmonics contained in the back EMF can be effectively eliminated, which contributes to improving the accuracy and dynamic performance of the rotor position estimation. Simulations and experiments verify the feasibility and effectiveness of this method.