Speed Control of Permanent Magnet DC Motor Using Artificial Neural Network

This paper introduces an approach to regulate the speed of a Permanent Magnet DC (PMDC) motor using an Artificial neural network (ANN), specifically a Feedforward neural network (FNN). Leveraging the inherent adaptability of an ANN which useful for controlling the dynamic speed requirements in practical applications, this study explores the potential substitution of conventional controller, specifically the Proportional-Integral (PI) controller. Currently, the increasing trend of ANN adoption, along with advancements in supporting hardware, reinforces the viability of employing ANN in motor control. Acknowledging limitations on this paper is dismissing some factor in the real system which could occur on the simulation experiments. The establishment of the FNN using Simulink/MATLAB software. The Levenberg-Marquardt with backpropagation is chosen as the learning algorithm due to its computational efficiency. To validate the viability of the FNN, experiments involve testing its performance across varying reference speeds, comparing it with the PI controller. Notably, experiments also test both controllers with an unknown motor. The result aims to illuminate the potential advantages and limitations of employing FNN in regulating PMDC motor control applications.