A Neural-Network-Based Controller for a Three- Phase Dual-Active-Bridge DC-DC Converter

Abstract

Due to its many benefits, the Dual Active Bridge (DAB) DC-DC converter is always easy to find in micro grids applications, energy storage systems applications, vehicles to grid applications, and much more in the current energy architectures. Due to system changes and disturbances on both the input side and on the output side brought by the broad variety of applications, the DAB performs inadequately. The output voltage of the DAB is to be controlled and kept constant during system fluctuation with limited time response using a neural network based adaptive controller. The configuration of the proposed controller is identical to that of a PI controller. The study is done using MATLAB Simulink, where the system is tested under system variations. The proposed controller, a PI controller, and a combination of an AANN in parallel with a PI controller are all subjected to a performance test via time domain analysis. The results of the comparison between the three controllers favored the suggested controller.