Speed sensorless control of induction motors using decomposed algorithms

This thesis addresses the problem of estimating the rotor speed and parameters of an induction motor, based on measurements of stator voltages and currents. The speed estimate is further used to design and simulate a speed sensorless AC drive system.; Based on the time-scale property of small induction motors, five linear regression models are derived from the motor electrical equations. The resulting estimator is structured as a five-stage procedure with each stage designed using the recursive least squares estimator for a linear regression model.; The performance of the estimator and the speed sensorless AC drive system is studied by means of computer simulations. In general, the estimator presents good tracking capabilities for both speed and motor parameters. The resulting response for the speed sensorless AC drive compares favorably to that of the control system using tachometer feedback. The effect of measurement noise on estimation and system performance is also investigated.