Sensorless Control of Permanent Magnet Synchronous Motor Using Square-root Cubature Kalman Filter

This paper introduces a Square-root Cubature Kalman Filter (SCKF) based speed and position observer for an Interior Permanent Magnet Synchronous Motor (IPMSM). Cubature Kalman Filter (CKF) is a new variety of Kalman filter which uses a third degree spherical-radial cubature rule to numerically compute multivariate moment integrals. Unlike in an Extended Kalman Filter (EKF), mean and covariance are propagated through the non-linear system, which minimizes the errors due to linearization. Square-root Cubature Kalman Filter (SCKF) algorithm is utilized for the sensorless control of an IPMSM of 1.5kW, 3000rpm rating. For the SCKF algorithm, IPMSM is modeled in stationary alpha beta reference. SCKF is preferred as the square root version is supposed to give improved numerical stability as compared to CKF. To get comparatively better transient performance and convergence of the SCKF for a non-zero initial rotor position, system covariance matrix Q is chosen adaptively. Simulation results for a VSI fed IPMSM are presented and the convergence of SCKF is shown for a variation of stator resistance. Experimental results are shown for Matlab-Simulink Real time hardware implementation using National Instruments DAQ card NI PCI-6221.