A reduced order smoothing filter for speed estimation of three phase induction motor

In this paper, a single-stage smoothing filter algorithm is used for the speed-sensorless field-oriented control of induction motors and is verified experimentally. Two well-known flux models of induction motor are used for estimation of rotor flux and speed. The uncertainties in measurement and model are accounted for. The experiment is carried out on a closed loop field-oriented control system. An estimate of the state variables in the next instant is made, using the conventional extended Kalman filter (EKF). This estimate is used to smoothen the estimate of the previous instant. This refinement is found to improve the estimates of the previous and next instances, since an additional data point is made use of. Using the measured stator phase voltages and currents, speed is estimated. The results are compared with those with the Extended Kalman Filter. The algorithm is found to make improvement in the transient part of response of the system, for the same values of process and measurement error covariances. The performance of the system for different reference speeds is also analyzed. It is observed that the transient performance is improved and estimation remains good for a range of values of process and measurement error covariances.