Performance enhancement of AC voltage controller-fed induction motor drive using neural networks

Abstract

The tracking of maximum efficiency point of a lightly loaded induction motor is generally obtained using search technique which is time consuming and the machine is subjected to many step changes of voltage which is undesirable. This paper presents a novel technique for energy efficient operation of voltage controlled induction motor drive using artificial neural networks (ANN). In the proposed method, the time required to converge to the best efficiency point of the drive is minimum and the machine is subjected to only one step change of voltage, irrespective of operating point. The back-propagation neural network is trained off-line to identify the optimal efficiency point of the motor in the entire operating range with a no priori knowledge of the machine parameters.