Intelligent Optimized Modulation of Triple Active Bridge Converters for Source-Storage-Load Integrated Systems

The triple active bridge (TAB) converter is a typical application in source-storage-load systems due to its high efficiency and high power density. However, due to the existence of coupling and other issues, the analytical method is complex to analyze and not easy to optimize. To solve this problem, this paper adopts the deep deterministic policy gradient (DDPG) algorithm to optimize the TAB converter. A normalized model of the TAB converter is derived through harmonic analysis, from which a parameter-independent optimization problem is extracted. The DDPG algorithm is used to optimize this problem. Finally, an artificial neural network (ANN) applicable to different parameters is obtained for the modulation of the TAB converter, realizing an optimized modulation method that can be deployed on different TAB converter platforms without retraining. Hardware experiments are conducted to verify the correctness of the proposed method.