The purpose of this paper is to improve performance in the estimation of velocity and flux in the sensorless control of induction motors (IMs) over a wide speed range, including low and zero speed.
Temperature and frequency dependent variations of stator (Rs) and rotor (R′r) resistances are very effective on estimation performance in sensorless control over a wide speed range. To this aim, an extended Kalman filter (EKF) is designed, which estimates the stator resistance, Rs, load torque, tL, velocity and flux. To provide robustness against R′r variations, the extended model is also continuously updated with R′r values from a look‐up table, built via EKF estimation.
As demonstrated by the experimental results, the estimated states and parameters undergo a very short transient and attain their steady‐state values accurately, with no need for signal injection due to the inherent noise introduced by EKF.
The value of this study is in the development of an EKF‐based scheme, which solves the Rs−R′r estimation problem in IM sensorless control. The successful experimental results obtained with the combined EKF and look‐up table approach also offer a solution to all EKF‐based estimation schemes which involve a high number of estimated parameters, hence, compromising estimation accuracy.
