The study on the PMSM sensorless control using the sub-optimal fading extend Kalman filter

EKFs are the very important speed identification methods in PMSM sensorless vector control. With the step changes of the reference speed of the PMSM drive system the residuals in EKFs are not the autocorrelation Gaussian white noise series anymore, which can lead to the EKF lose the ability of tracing state variables and bring out divergence problems in worst case. A sub-optimal fading extend Kalman filter-SFEKF is adopted to identify speed and it can forces the output state variables to tracing the gradual or step changing references. The SFEKF well overcomes varying condition´s impacts caused by the disturbances, and improves the dynamic response and tracking precision. The simulation and experiment results show that the SFEKF has a simple arithmetic, moderate calculation and good robustness.