Optimization of a space vector control using MOAM algorithm and extended Kalman filter

This paper proposes a new application of parameter estimation for induction motors. Induction motor is described by non-linear differential equations and an Extended Kalman Filter (EKF) estimates three parameters needed for the control (rotor resistance, stator resistance and magnetising inductance). A Modified Random Optimisation Method, MOAM, is used to optimize the covariance matrix in Extended Kalman Filter, rotor and stator autoinductances, friction factor and moment of inertia. It is used to achieve the best performance in the Vector control. Simulation studies on a field oriented controller (FOC), under different variations on the parameter model, and experiments with 5.5 kW motor are presented.