Comparative study of pole placement methods in adaptive flux observers

A robust and accurate observer for IM speed estimation, not only at normal and high speeds but also at very low and zero speeds, is certainly important and desirable for sensorless IM drives. Many pole placement methods have been proposed to design a good adaptive flux observer that is less sensitive to IM speed; however, most of them have poor dynamic performance at very low and zero speeds. Four different pole placement techniques are discussed in this paper. Three of the four techniques are based on the generalization of existing methods. Authors propose the fourth technique, which is an improvement over the existing schemes. The feedback gain is chosen in such a way that the observer poles are unaffected by motor speed variations. Although all four schemes estimate the motor speed and stator resistance in real time, the proposed technique is superior to others in performance and robustness to motor speed and stator resistance variations and noises.