Application of Optimized EKF in Direct Torque Control System of Induction Motor

The paper introduces extended Kalman filter (EKF) to estimate state variables and stator resistance of induction motor and employs virtual noise to compensate for model error from system model linearity. As EKF can´t estimate state variables accurately in case of unknown noise statistics, genetic algorithm is proposed to optimize EKF. Study shows that GA has the shortcoming of inefficiency, so the paper proposes immune genetic algorithm (IGA) optimization scheme, which converges to globally optimal solution on the basis of excellent individual group. Experimental system of induction motor based on direct torque control (DTC) is designed, and comparative study is carried out a great deal. The conclusion that system´s performance based on IGA-EKF optimization scheme is better than that based on GA-EKF optimization scheme is drawn