Implementation Issues of Flux Linkage Estimation on Permanent Magnet Machine Position Sensorless Drive at Low Speed

Abstract

This paper improves the surface permanent magnet (PM) machine position sensorless drive at low speed. Considering the surface PM machine (SPM) drive, EMF voltage or flux linkage should be estimated for the sensorless drive. Different from EMF voltage, the flux linkage based on the voltage integration is theoretically independent to speed which is suited for the low speed position estimation. In this paper, several improvements on the flux-based sensorless drive are proposed to enhance the low speed dynamic performance. First, a modified voltage integration is develop to remove the flux estimation drift caused by voltage or current offset. This integration contains a high-pass filter (HPF) for the DC drift elimination. In addtion, the filter delay is compensated to maintain the flux phase. Second, inverter deadtime harmonics are decoupled with the knowledge of actual machine phase voltages. It is shown that the position estimation error is decreased for the better low speed performance. According to experimental results, SPM machine sensorless drive is enhanced at 4%∼ 6% speed region from many aspects. They include position signal SNR, position error and drive dynamic response. More importantly, the overall current regulation bandwidth can increase to 1kHz at low speed. It is compatible to standard encoder-based field oriented control (FOC) drives. INDEX TERMS Permanent magnet machine, flux observer and position sensorless drive.

A Novel Robust Sensorless Technique for Field-Oriented Control Drive of Permanent Magnet Synchronous Motor

“…Thus, the stator flux linkage estimation could be significantly affected at low speed (i.e., the magnitude attenuation and phase lag) with the LPF flux linkage estimator presented in refs. [18], [26].…”

“…However, the errors of the d-q-axis current and rotor angle estimation in the proposed PP-01 method are still large at speed below 200 rpm. This may be due to the phase lag of the stator flux linkage estimation caused by the LPF estimator [18], [25] and the possible nonlinearity of the inverter and PMSM [26]. This may cause the closed-loop SLC-FOC drive system to be unstable at low speeds.…”

“…To solve the issue, a phase compensator is added to the output angle in the stator flux linkage estimation. If the rotor speed is lower than the cutoff frequency (ω e < ω c ), the compensation angle θ comp is expressed as [25], [26]…”

“…For simplicity, the stator flux linkage estimation technique used here is a typical closed-loop estimator based on low-pass filter (LPF) with the adaptive compensation [20], [26], as re-constructed in Fig. 5.…”

“…Overall, the common difficulty for the flux-linkage-based SLC is that the rotor position estimation relies heavily on the accuracy of the estimated stator flux linkage magnitude and angle, which can be easily influenced by factors such as variation of resistance and potential nonlinearity with low B-EMF at high current. Hence, recent studies mostly focus on enhancing the accuracy of the stator flux linkage estimator [14], [15], [18], [26] rather than developing methods to reduce the dependence on stator flux linkage components (i.e., magnitude and phase angle) in SLC calculations.…”