Estimation of induction motor states and parameters based on Extended Kalman Filter considering parameter constraints

The Extended Kalman Filter (EKF) has been applied to estimate states and parameters of an induction motor. For this application, sometimes, the parameters estimated by the filter may violate their physical ranges. To overcome this drawback, in this paper, motor's parameters constraints are incorporated into the EKF. The proposed technique modifies the EKF computation loop such that if any estimated parameter does not satisfy the physical constraints, the quadratic programming (QP) will be invoked to adjust the estimation. The proposed technique has been implemented in MATLAB environment and tested with the parameter data obtained from a 380 V, 50 Hz, 4 poles, 0.37 kW, squirrel cage induction motor. The numerical experimental results indicate that the proposed algorithm can improve estimation performance over the conventional EKF.