Electrical-Sensorless Control of Induction Motor *

The field oriented control (FOC) of the Induction machine (IM) requires the measurement of rotor speed and stator currents. This information is provided from three sensors: one mechanical and two electrical sensors. Generally, we often pay much attention to mechanical sensorless control. Extensive works have already been done on this. Electrical sensors are no less important as mechanical one. As a matter of fact, a failure of only one electrical sensor causes a malfunction of the overall control chain. In this paper we propose a new electrical sensorless control of IM with the same performance as the FOC and lower cost. The proposed control whose the scheme is very simple, is based on the artificial neural networks (ANN). A neural block, with specified inputs, is designed to provide the desired IM control variables at any operation point basing on a minimum amount of information provided from the motor. The effectiveness of the proposed control for IM drive is verified at several operation conditions and highlighted by comparing to the conventional FOC. The results clearly show the merits of the proposed control over conventional one, in terms of cost, simplicity of the control scheme and above all, its interest in the event of the technical defect of one or both stator current sensors.