Vanadium redox flow batteries are one of the most promising energy storage technologies today due to their low fire hazard, long cycle life, and excellent scalability. However, to unlock the full potential of the technology, further efforts are needed in the development of advanced control strategies.

To enhance the stability and anti-interference capability of vanadium redox flow batteries in microgrids, a group of researchers led by the University of Western Australia has developed a new learning-based data-driven approach. H∞ control approach.

The control strategy uses an integral reinforcement learning algorithm to produce excellent steady-state and dynamic responses through measurement alone. According to the researchers, compared to model-based control methods, it is insensitive to model parameter variation.

Furthermore, compared to most existing artificial intelligent control methods that require a large amount of experimental data for offline neural network training, the proposed control strategy contributes to eliminating the offline training process and, therefore, does not require the expensive and troublesome training data acquisition process. .

Importantly, the proposed control guarantee offers closed-loop control stability, which cannot be achieved by almost all control methods that rely entirely on offline trained neural networks.

“While ensuring excellent online control performance, the proposed method overcomes the drawbacks of existing vanadium redox flow battery control methods, such as model dependence and inability to perform online training,” the researchers said.

In their paper, they provide proof of stability and claim to verify the superiority of the proposed control method through demonstration of simulation results and quantitative analysis. They discussed their findings in “A novel learning-based data-driven H∞ Control Strategy for Vanadium Redox Flow Batteries in DC Microgrids,” which was published recently Journal of Power Sources.

This content is protected by copyright and may not be reused. If you would like to collaborate with us and reuse some of our content, please contact: