Speed control of neural network based energy efficient DC drive

DC motors have a very fine speed control characteristics and so are widely used in various industrial applications. Their performance can be optimised, if along with fine speed control a motor can be made energy efficient. In this paper an attempt is made to achieve the speed control at the condition of maximum efficiency for a separately excited DC motor using both field control and armature control methods. In recent trends, it is seen that artificial intelligence based controllers have superior performance compared to the power electronic based controllers PID controllers. The controller chosen for the desired task is an Artificial Neural Network, as any non-linear data can be fit in it. The network is trained and a model is generated in MATLAB. The neural network is then connected to a model of a separately excited DC motor and a few simulated results were obtained to verify the task. The whole simulation work is performed in Simulink and the output curves were obtained for various types of load torques and reference speeds.