Research on Robust Performance of Speed-Sensorless Vector Control for the Induction Motor Using an Interfacing Multiple-Model Extended Kalman Filter

Abstract

The interfacing multiple-model extended Kalman filter (IMM-EKF) is proposed here as a modification of the extended Kalman filter (EKF). In this algorithm, two multiple-model EKF groups are built, one group is the optimum model, and the other is the noise model. Each model group is created by multiple models, and it will get good performance at stable state and robust ability when disturbance occurred. The algorithm gets the estimation value by mixing the outputs of the different model in different weightings, and the calculation of weightings is researched. Whether the IMM-EKF can give better estimation performances and robust ability than the EKF for speed estimation of induction machines is explored in this paper. Via simulations and experiments, estimated error and the change of flux linkage by disturbance based on the IMM-EKF and EKF is compared. The simulation results show that the IMM-EKF has the better estimation performance of antigross error than the EKF.