Sensorless vector control of three phase induction motor

Abstract

In this paper, an indirect vector control of induction motor is simulated and the speed is estimated using conventional Model Reference Adaptive System (MRAS). It is modified using neural network PI controller. A conventional mathematical model based MRAS speed estimator can give a relatively precise speed estimation result, but error will occur during low frequency operation. It is also very sensitive to machine parameter variations. Hence instead of PI controller a two-layered neural network PI controller (NNPIC) is used. With the help of projection algorithm, the parameters of the NNPIC are automatically adjusted and the difference between the two models of MRAS is minimized for speed estimation. Neural network-based MRAS estimator gave robust performance during low frequency and parameter variation. Also, this scheme reduced the work of tuning mechanism of PI controller. The estimated speed was taken as a feedback and the speed was controlled by indirect vector control using space vector pulse width modulation (SVPWM). The simulation results showed improvement in the performance of an induction motor drive.