A New Regulation and Control Method for Distribution Networks Based on Human-machine Hybrid Augmented Intelligence

With the widespread application of artificial intelligence in distribution network regulation, efficiency has been significantly improved. However, due to the high penetration of new energy sources integrated into the distribution network, operation modes have become increasingly complex and fragile, posing challenges for existing artificial intelligence in meeting the safety operation requirements of the distribution network. To address this issue, a new distribution network regulation method based on human-machine hybrid augmented intelligence is proposed. Firstly, adversarial generative imitation learning is employed to supplement missing data in the regulation scenario, and the data is annotated using a human-machine hybrid approach. Secondly, regulatory targets are established based on regulatory scenarios, and the model is then trained. On this basis, the experiential advantages of human intelligence and the computational advantages of artificial intelligence are leveraged to optimize regulatory strategies. Finally, the proposed method is validated in a new-type distribution network in a certain area in Sichuan Province of China, and it outperforms both Pareto optimality-based method and bidirectional long short-term memory neural network method in terms of voltage control strategy and cost optimization. This method enhances the practical application level of artificial intelligence in the regulation of new-type distribution networks.