Stability and dynamic performance improvement of a sensorless IPMSM drive via adaptive estimated-speed-assisted position prediction and current quality evaluation

This paper studies stability issues of sensorless torque control of high power interior permanent magnet synchronous machines (IPMSMs) under large load transients, e.g., complete torque reversals with the highest slew rate allowed. To perform a sensorless vector control, a current model-based sliding-mode observer is utilized to estimate the rotor position. The correlation between current regulation quality and position estimation accuracy is theoretically studied and experimentally verified. Based on the correlation, the accuracy of position estimation can be evaluated online indirectly. Moreover, an adaptive estimated-speed-assisted position prediction (ESAPP) scheme is proposed to improve the stability and performance of the sensorless control systems under large load transients. The ESAPP period is adapted according to the position estimation accuracy. Experimental results on a 150 kW sensorless IPMSM drive system are provided to validate the proposed algorithms.