Comparison Between PI and Neural Network Controller for Dual Active Bridge Converter
Marco Bindi, Carlos Iturrino Garcia, Fabio Corti, Maria Cristina Piccirilli, Antonio Luchetta, Francesco Grasso, S. Manetti
Abstract
In this paper a comparison is presented between two strategies for the control of a Dual Active Bridge (DAB) converter regulated through Single Phase Shift (SPS) modulation. The purpose of the controller is to regulate the phase-shift between the voltage impressed to the inductor from the primary and the secondary converter to obtain the desired output current. To achieve this goal, two control strategies based on different approaches are proposed. The first one performs the regulation using a Proportional-Integral (PI) controller, while the second one, called Model Reference Control (MRC), performs the regulation through Neural Networks (NN). The obtained results show that the Artificial Intelligence (AI) based technique has comparable reference current tracking performance and better dynamic characteristics. Furthermore, this technique does not require the mathematical model of the converter and can be used in a wider range of operating conditions.