ANN Aided Speed Observer and Field Situated Sensor Less Controller for Permanent Magnet Synchronous Engine Drive

Abstract

Sensor less speed control of permanent magnet synchronous motor (PMSM) remains in progress to beat the prevailing drawbacks like speed ripples, torque ripples and poor performance at low speeds with no load or lightly loaded conditions In this paper, Sensor less Field Oriented Control (FOC) system which is based on artificial neural network (ANN) implementation is proposed. The control system uses ANN aided Model Reference Adaptive System Adaptive System (MRAS) for speed estimation. Both the speed estimator and the controller of the proposed system use ANN based implementations. The performance of the proposed system is analyzed for speed control of PMSM, for various speed and load torque values. The simulation results of the proposed ANN based speed estimator and controller model demonstrate that ANN aided MRAS based estimator has the capability to enhance the performance of the system. Specifically, it reduces the overall complexity and initial speed overshoot by 1% as compared to the conventional controllers employed for PMSM speed control. It also achieves 0.002 rad/s improvement on steady state error with acceptable settling and rising time in its dynamic response.