Speed Regulation of Closed Loop Brush less DC (BLDC) Motor Using Novel Multi-Layer Perceptron Approach

Recently developed advanced applications of Brushless DC (BLDC) motor demands for variable speed operation to equate the assessed performance. In this connection, many speed controlling techniques are introduced for the fruitful operation of BLDC motor at various speed levels. The performance of the existing techniques is up to the mark for a fixed speed reference. During dynamic change in reference speed the existing techniques fail to follow the path. Therefore, there is a need to regulate the speed of the BLDC motor during dynamic conditions. Speed regulation is achieved by minimizing the residual error content. If the content of residual error increases, there will be a deviation between actual and reference speed resulting in poor speed regulation. Hence, this paper presents a novel and simple controlling technique based on Multi-Layer Perceptron (MLP) technique, which is capable of regulating the speed as per the dynamic reference path by minimizing the residual error content. MLP is from the compound of deep learning artificial neural network has two layers, one receives speed error data from the traditional PID controller, and the other layer predicts the output data. Computations are performed through Gradient Descent for arbitrary hidden layers to reduce the speed error. As the computed data approaches the target, optimal data for the speed regulation is attained.