Position extraction from a discrete sliding-mode observer for sensorless control of IPMSMs

Sliding-mode observers (SMOs) offer a promising solution for sensorless control of interior permanent magnet synchronous machines (IPMSMs) due to their excellent robustness to system structure and parameter uncertainty. However, in practical applications, it is challenging for an SMO to achieve a perfect estimation for the back electromagnetic force (EMF) using a finite or relatively lower sampling frequency, especially for high-speed applications. Phase shift, magnitude variation, and heavy noise in the estimated back EMF will cause unexpected errors in rotor position extraction. Thus, advanced rotor position extraction methods are needed to obtain position information from the estimated back EMF. This paper proposes a novel estimated speed feedback algorithm to work together with the conventional inverse tangent method and angle tracking observer (ATO) to extract the rotor position from the extended back EMF obtained from a discrete SMO. The extracted position has reduced oscillations compared to that obtained from traditional methods. The proposed position extraction methods are validated by simulations in MATLAB Simulink as well as experiments on a practical IPMSM drive system.