Sensorless Speed Control of Induction Motor Derives Using a Robust and Adaptive Neuro-Fuzzy Based Intelligent Controller Abolfazl Vahedi Farzan Rashidi Electric Machines Research Laboratory, Iran University of Science and Technology, Tehran, Iran Abstract. In this paper a novel sensorless adaptive neurofuzzy speed controller for induction motor derives is formulated. An artificial neural network (ANN) is adopted to estimate the motor speed and thus provide a sensorless speed estimator system. The performance of the proposed adaptive neurofuzzy speed controller is evaluated for a wide range of operating conditions for induction motor. These include startup, step changes in reference speed, unknown load torque and parameters variations. Obtained results show that the proposed ANN provides a very satisfactory speed estimation under the above mentioned operation conditions and also the sensorless adaptive neurofuzzy speed controller can achieve very robust and satisfactory performance and could be used to get the desired performance levels. The response time is also very fast despite the fact that the control strategy is based on bounded rationality. 1. Introduction Variable speed motor drives play an important role in modern industries because they are utilized extensively in factory automation to store energy or to meet stringent load requirements. The use of variable speed motor drives is ever increasing and will maintain its momentum for several decades to come [1]. Among all different kinds of electric motor drives, the induction motor has become the subject of a large body of research in the field of electric motor drives. This is partly because the motor has an intrinsically simple and rugged structure and low manufacturing cost. Moreover, induction motor drives have the wide speed range, high efficiencies, and robustness [2]. All these merits make the motor a good candidate for the industrial applications. Sensorless control of induction motor drives is now receiving wide attention. The main reason is that the speed sensor spoils the ruggedness and simplicity of induction motor. In a hostile environment, speed sensors can not even be mounted. However, due to the high order and nonlinearity of the dynamics of an induction motor, estimation of the angle speed and rotor flux without the measurement of mechanical variables becomes a challenging problem [3]. The advantages of speed sensorless induction motor derives are reduced hardware complexity and lower cost, reduce size of derive machine, eliminate of sensor cable, better noise immunity, increasing reliability and less maintenance requirements [4]. Various speed control algorithms for induction motor derives have been devised in the literature. Among them, PID controllers [5], optimal [6], nonlinear [7,8] and robust [9,10] control strategies, and neural and/or fuzzy [3,11] approaches are to be mentioned. The purpose of this paper is to suggest another control approach, based on adaptive neurofuzzy controller to achieve faster response with reduced overshoot and rise time. Although the proposed method is a mathematical model based approach, it demonstrates some robustness towards model parameter variations due to the fact that the adaptive neurofuzzy based intelligent controllers are usually robust [12]. Proceedings of the 5th WSEAS Int. Conf. on Power Systems and Electromagnetic Compatibility, Corfu, Greece, August 23-25, 2005 (pp304-312)