Robust speed estimation for speed sensorless vector control of induction motors

In this paper, some new approaches by using an adaptive binary observer, a reduced order extended Kalman filter and a two degrees of freedom controller to a high performance variable drive system for induction motors without a speed sensor are proposed. The speed sensorless vector controls realized in this paper have robustness to load and parameter variations and excellent dynamic characteristics in comparison with conventional speed sensorless vector control schemes. The binary observer estimates the rotor speed and rotor flux with alleviation of the high frequency chattering, and retains the benefits achieved in the conventional sliding observer. The reduced order extended Kalman filter for sensorless control has advantage of saving computation time and simplifying structure in comparison with the full order extended Kalman filter, and also retains the benefits achieved in the extended Kalman filter, such as good robustness to measurement noise and parameter variation. In the two degrees of freedom controller, a feedback controller to robustly control the variations of the rotor resistance is implemented by H_∞ control theorem, and a feedforward controller to obtain good tracking performance is designed. Newly proposed systems implement the speed sensorless vector control of induction motors by 32-bit floating point TMS 320C31 DSP