Sensorless speed estimation of induction motor based on feed-forward neural network algorithm

Abstract

The squirrel-cage induction motor is widely used in the industry. Its usage is increasing because of possible applications like electric vehicles, which require an efficient motor drive. However, most control techniques obtain speed from the shaft encoder. These are the disadvantages which are unacceptable for most applications. Thus, in present, research in this area is mostly focused on applying speed sensorless and obtains robust induction motor. The ability of Artificial Neural Networks to map arbitrary nonlinear function has been used to investigate estimation of speed in induction motor, where the inputs of the network are the voltage and current of stator. This work has attracted attention in recent years, though very little work has been done so far because of the complexity of the problem. This work makes an important contribution to the area of induction motor by off-line trained speed estimators, using d-q axis dynamic equations. A method that use current and voltage as a input of system proposed whereby artificial neural network is trained off-line to estimate the speed of motor. The result presented in the thesis indicate that the proposed scheme are able to track the speed under load variations. The effectiveness of the proposed method was tested by simulation.