A nonlinear estimator for dynamical and robust sensorless control of permanent magnet synchronous machines

This paper proposes an angular position and speed estimation scheme that is based on a direct evaluation of the angle of the voltage induced by the spinning rotor (back-EMF) of a permanent-magnet synchronous motor (PMSM). It is an inverse parallel-model estimation method, meaning the rotor position and speed are directly calculated based on stator current measurements and the stator voltage commands. In contrast to existing schemes, no observer, no integration and no speed or flux estimation is necessary. The estimators are extended with filters to cope with measurement noise, and directly used for field-oriented control. It is shown algebraically and experimentally that parametric robustness is outstanding. The resulting estimated angle is driftless even under uncertainties. The scheme is suitable for encoderless control of a PMSM at high and low speeds. The performance of the scheme is confirmed by experimental results.