Performance Enhancing Speed Control of a DC Motor Based on an Artificial Neural Network

Abstract

This paper employs artificial neural networks (ANNs) in estimating and controlling the speed of a separately excited DC motor, which is also one of the most essential modern techniques used in control applications and to enhance the effectiveness of separately excited DC motor speed control. For speed control of a separately excited DC motor, the paper compares the performance of the conventional proportional integral (PI) and artificial neural networks (ANNs). A neural network (NNs) based controller is an extremely appreciated technology for accomplishing high-recital speed control. The simulation results are expose to demonstrate the efficacy and advantage of the control system of a separately excited DC motor (SEDCM) using ANNs over the traditional control arrangement. The integration of (ANNs) has shown significant improvements in performance in speed control of DC motors separately excited, chiefly in dynamic response and robustness in contradiction of load variations. This technique leverages the adaptive competences of ANNs to improve traditional control methods. In this manuscript, two controller for speed control of SEDCM are designed and simulated using matlab Simulink program. The first one is the PI controller and the second is the ANN controller, the results of the ANN controller based speed control of the SEDCM give quick response, low overshot, small settling time, zero steady state error when sudden change happened in reference speed or load torque. In addition, as compared with the results of PI controller which give slow response, high overshot, large settling time under the same conditions.