ANN-based direct torque control scheme of an IM drive for a wide range of speed operation

Abstract

Induction motor (IM) drives with direct torque control (DTC) enable fast torque response without the need for complex orientation conversions or inner loop current loop. In the speed estimation responses, however, there is a significant level of torque ripple. The voltage source inverter adds acoustic noise and needs a high sampling frequency since it operates at a high and variable switching frequency. This work describes an ANN-based DTC technique for controlling the speed of an IM drive over a large speed range. To achieve good dynamic performance of induction motor drive, the ANN-based speed controller will replace the speed controller, switching tables, and hysteresis comparators. The neural network was trained using the back-propagation algorithm. The goal of a neural speed controller is to improve the system ability to respond quickly to changes in process variables while also mitigating the impacts of external perturbations. The projected ANN based DTC considerably and simply tracks the reference speed thus improves the efficiency of speed-torque of induction motors with quicker responses for rapid varying of speed reference and torque as that of Electric Vehicles in any uneven roads circumstances. MATLAB/Simulink software is used to evaluate the drive performance for both transient and dynamic operations. The proposed control performance is simulated and compared to a DTC-based traditional PI speed controller. In comparison to PI, the results show that ANN has better and faster effects. The torque ripple gets reduced by 1.5% in ANN (artificial neural network) controller compared to PI controller. The THD (total harmonic distortion) is reduced by 6.38% from PI controller to ANN controller.