On-line estimation of induction motor parameters by extended Kalman filter

In variable speed AC drives it is often important to know the accurate values of motor parameters to realize high performance drive systems. On-line estimation of two parameters of an induction motor, the rotor resistance and magnetizing inductance, which are important in field-oriented control is investigated by using an extended Kalman filter. A continuous form Kalman filter is adopted to reduce the amount of calculation needed. An estimation system based on this Kalman filter algorithm is implemented employing a digital signal processor. Using this estimation system the influence of covariances of noises and initial setting of parameters on the estimation results is first investigated. The variations of rotor resistance with temperature and magnetizing inductance with flux level are then estimated. The influence of load torque variations on the estimated parameters are also shown. These experimental results confirm the usefulness of on-line parameter estimation using a Kalman filter