Search

The position and speed estimation performance of the sliding mode observer (SMO) for permanent magnet synchronous motor (PMSM) drives under low carrier ratio degrades, due to the inherent chattering and enlarged control delay. To improve estimation performance, this paper proposes a quasi-proportional-resonant (QPR) based adaptive observer to obtain back electromotive force with adaptive harmonics cancellation. In this observer, the QPR controller is employed to eliminate errors between the observed currents and the actual ones in the αβ reference frame, replacing sign function and low pass filter (LPF) of the SMO to cancel the chattering, and harmonics immune performance can also be achieved thanks to frequency selection characteristic of the QPR controller. Besides cancelling the chattering, the proposed observer can also eliminate phase shift produced by the LPF, since the QPR controller is a zero phase-shift amplifier at the resonant frequency. Comparative experiments are carried out in a 3-kW PMSM drive, and the results validate the effectiveness of the proposed adaptive position observer. [ABSTRACT FROM AUTHOR]

In order to reduce the chattering and phase shift, this article proposes an enhanced sliding mode observer (SMO) for the sensorless control of permanent magnet synchronous motor (PMSM). In the conventional SMO, a low-pass filter (LPF) is commonly employed to extract the back electromotive forces (EMFs) from the sliding mode switching function, which brings the phase shift and degrades the precision of position and speed estimations. In this article, a frequency-adaptive complex-coefficient filter (FACCF) is adopted to replace the LPF. Since the FACCF has no phase shift and amplitude attenuation at the central frequency, it can be used to extract the EMFs without distortion while the chattering can also be suppressed. For an easy parameter design, the normalized phase-locked loop is used to obtain accurately the position and speed estimations. The experiments are carried out on a 3-kW PMSM drive, and the results verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]