A Hybrid Power Decoupling Strategy for Triple Active Bridge

Triple active bridge (TAB) is widely used due to its high efficiency and power density. One key technique of TAB is how to achieve power decoupling among each port. The conventional approach is based on complicated look-up table, which consumes large amount of memories. Although some AI-based method shows some advantages, the decoupling accuracy cannot be guaranteed in whole range. Therefore, in this paper, a hybrid solution is proposed by combining look-up table and artificial neural network (ANN) together. The principle of how to integrate these two methods is well analyzed. The overall decoupling effect is improved while much processor memories is saved. Simulation and experimental results prove the effectiveness of this method.