Sensorless Control for IPM with Adaptive Resistance Estimation Using Flux Norm Correction

This paper proposes a robust sensorless control algorithm for the interior permanent magnet motor (IPM) with online resistance identification. A nonlinear Luenberger observer is built to estimate the PM flux linkage and resistance. Considering that the state pair of rotor position and resistance are not observable with mere voltage and current measurement, we introduce an additional constraining algebraic equation that the Euclidean norm of the permanent magnet (PM) flux linkage is constant during the motor operation. With assumption of the constraining equation, the rotor resistance can be estimated by an adaptive mechanism, which applies a gradient descent algorithm to the error of PM flux linkage Euclidean norm. The stability of the observer is guaranteed via Lyapunov analysis. The observer performance is tested in a close-loop way in an IPM bench.