Low Speed Estimation in Sensorless Direct Torque Controlled Induction Motor Drive Using Extended Kalman Filter

Abstract

Sensorless Direct Torque Control (DTC) is a powerful control scheme for high performance control of induction motor (IM) drives, which provides very quick dynamic response with simple structure and a decoupled control of torque and flux. The performance of the DTC drive depends greatly on the accuracy of the estimated flux components, torque and speed, using monitored stator voltages and currents. Low speed estimation is a great challenge because of the presence of transient offset, drift and domination of ohmic voltage drop.Extended Kalman filter (EKF) is a non linear adaptive filter which performs the process of finding the best estimate from the noisy data based on state space techniques and recursive algorithm. This paper mainly focuses on the accurate estimation of speed ranging from very low speed to rated speed using the equation of motion. A new state space model of the IM is developed for estimation in EKF, with load torque as an input variable and not as an estimated quantity which is the case in most previous studies. The developed algorithm is validated using MATLAB-Simulink platform for speeds ranging from low speed to rated speed at rated torque and at various torque conditions. An exhaustive analysis is carried out to validate the performance of DTC Induction motor drive especially at the low speeds. The results are promising for accurate estimation of speed ranging from low speed to rated speed using EKF.