Performance Analysis of Separately Excited DC Motor with Wavelet Neural Network based Controller
Megha Khatri, Pankaj Dahiya, Amaar Hussain · 2020
The direct current motors are used in various applications such as defence, industries, robotics, which are the major power consumers. The optimal control of input parameters can decrease the power consumption and enhance the performance, reliability and life span of these machines. Although the separately excited DC motors provides constant speed output which is stimulated to the variations and unpredictability in inputs such as presences of nonlinearities such as saturation, friction and noise propagation along the series of unit processes degrade its performance. Thus optimal controlling of the parameters through the motor control unit is required to achieve smooth operation. This article is applied to the Wavelet Neural Network (WNN) based controller which uses the wavelet transforms as the activation function. This helps in continuous training and optimization of the controller parameters with minimum time duration. The different controllers such as proportional-integral, NN based proportional-integral-derivative, WNN based proportional-integral-derivative with filter, WNN based proportional-integral-proportional-derivative (WNN-PIPD) are compared in terms of rise time, maximum overshoot, steady state error etc. The inherent approximation capability and learning abilities of the proposed WNN-PIPD controller improve the degree of tolerance and adaptability concerning other controllers.