Deep Reinforcement Learning Aided Variable-Frequency Triple-Phase-Shift Control for Dual-Active-Bridge Converter

To improve the conversion efficiency of the dual-active-bridge converter, this article demonstrates a variable-frequency triple-phase-shift (TPS) control strategy with the help of the deep reinforcement learning method. More specifically, the twin delayed deep deterministic policy gradient (TD3) algorithm is adopted to train the agent offline with the aim of minimum power losses, under the TPS modulation with varying switching frequency. Moreover, the zero-voltage-switching performance has been considered during the training of the TD3 algorithm. Based on these, the trained TD3 agent acts as a fast surrogate predictor, which can produce appropriate control strategies in real-time for whole continuous operating conditions with soft switching and maximum conversion efficiency. The effectiveness and correctness of the proposed scheme is validated through the experimental results in a laboratory prototype.

Automated summary

This article proposes an improvement approach for the dual-active-bridge converter using a variable-frequency triple-phase-shift control strategy with the help of deep reinforcement learning. The twin delayed deep deterministic policy gradient algorithm is adopted to train the agent offline, aiming to minimize power losses under varying switching frequencies. The training of the algorithm incorporates zero-voltage switching performance. Based on this, the trained agent acts as a fast surrogate predictor, generating appropriate control strategies in real-time for continuous operation with soft switching and maximum conversion efficiency. The effectiveness and correctness of the proposed scheme are validated through experimental results in a laboratory prototype.