A stable machine learning based control for bidirectional interleaved DC/DC converter in battery systems

This paper proposes an innovative machine learning-based control algorithm for the bidirectional interleaved DC/DC converter in battery systems. It offers fast dynamics, gain tuning-free control design, and guaranteed closed-loop stability. Furthermore, compared to the other learning-based control methods, the proposed algorithm shows low computational complexity in both the neural network training and the online digital implementation. The efficacy of the proposed method is substantiated through experimental validation.