An Innovative Deep Reinforcement Learning Controller for DC/DC DAB Converters Based on Deep Deterministic Policy Gradient

In order to adapt to renewable energy sources and multiple DC loads connected to the power grid, the application of dual active bridge (DAB) converters in the context of DC micro-grids (MGs) has been widely used. Aim to stabilize DC bus output voltage of the dual-active-bridge (DAB) DC-DC converter, an innovative control method is proposed in this paper. Using the deep reinforcement learning (DRL) algorithm with the Actor-Critic architecture sustains voltage stabilization of the DAB converters. More specifically, the deep deterministic policy gradient (DDPG) algorithm is according to the mathematical model of the system's input and output data learning system, the optimal control of the system can be realized according to the given reward. It is guaranteed that the DC converter has strict stability in the face of all kinds of disturbance problems in the system, and the effectiveness of this control strategy is verified by simulation. In order to prove the advantages and practicability of the proposed adaptive method, The effectiveness of this control strategy is verified by simulation outcome.