Design of Interleaved DC-DC SEPIC Converter Based Switching Regulator Using Neural Network Based Machine Learning Algorithm

DC-DC Converter based Switching Regulators are widely utilized in industrial and domestic applications, because of high conversion efficiency with reduced power loss. This research work uses an Interleaved DC-DC SEPIC Converter (ISC) to construct a switching regulator because of following reasons: It is a two-stage SEPIC converter with phase-shifted switches operating with decreased reduced ripples, switching stress, and it also employs small size inductors and capacitors for the same output power because of high operating frequency. Here Neural Network (NN) based Cascade Controller is employed to construct ISC based switching regulator. A NN is a machine learning technique used to train data without the need of human intervention. Three different training techniques are utilized to train the NN: Levenberg-Marquardt (LM), Bayesian Regularization (BR), and Scaled Conjugate Gradient (SCG) Algorithm. The time domain parameters for achieving the regulated output voltage have been lowered with the NN machine learning algorithm when compared to the Conventional Proportional Integral (PI) controller.