Sensorless vector control of permanent magnet synchronous motor using neural network observer

Abstract

The control of a machine requires measurements on a few quantities which are often difficult to measure; usually a speed measurement is required for controlling a permanent magnet synchronous motor. While the use of position and speed sensors can reduce the control performance. The use of observers may be a potential solution for a sensorless control. This paper deals with the use of the neural network approach in speed estimation of Permanent Magnet synchronous Motor (PMSM). To validate the investigated estimation, dynamic model and vector control schemes of the PMSM using space vector modulation technique, to control the DC-AC converter, are implemented in Matlab/Simulink.