Application of neural network based model predictive controller to power switching converters

Abstract

Neural network based Model Predictive Controller (MPC) for a dc-dc buck converter working in Continuous Conduction Mode (CCM) is presented. The converter operates at a switching frequency of 500 KHz. Although neural networks (NN) have been used in problems involving knotty, non-linearity and uncertainties but here they are applied to a buck converter to control its characteristics. The neural network is trained using ‘trainlm’ method using Neural Network Toolbox. The simulation results show that the neural network model predictive controller depicts better static and dynamic characteristics. The controller is then compared with the classical lead controller. Matlab/Simulink based simulated results validate the design.