Mitigation of Total Harmonic Distortion in the Railway Power System using Bald Eagle Optimized Deep Convolutional Neural Network

Harmonic distortion leads the power losses in the power systems and non-linearity loads generate harmonics, which additionally introduces resistive and eddy current losses in the transformers. These disturbances result in introducing operational problems which cause the interruption in the power distribution system and if mitigation is not derived about the disturbances, it affects the quality or reduces the lifetime of the equipment. Hence, the main objective of the research is to perform the harmonic analysis in the power systems using a Bald Eagle optimized deep convolutional neural network (BEO-deep CNN) classifier. BEO algorithm tackles the shortcomings problems in the Total Harmonic Distortion (THD) and is also used in neural networks for the prediction of harmonics in the power system. Further, the BEO algorithm provides the trade-off exploration and exploitation which improves the convergence speed and in addition, the simulated value is used for the determination of the THD. The error metrics of Mean Squared Error (MSE), R Score (correlation coefficient), and Root Mean Square Error (RMSE) metrics are obtained from the proposed BEO-deep CNN and the efficient rates of error metrics values are 15.16,0.96, and 3.89 which proves the efficacy of the model.