Research on Key Technologies of AI-Based Source-Load Coordinated Regulation at the Edge of Distribution Networks
DOI:
https://doi.org/10.4108/ew.12162Keywords:
distribution network, source-load coordination, deep reinforcement learning, federated learning, edge computing, multi-timescale optimization, renewable energy integrationAbstract
INTRODUCTION: The integration of a high proportion of renewable energy into distribution networks introduces significant challenges, including increased source-load volatility, uncertainty, and operational difficulties.
OBJECTIVES: This study aims to address these challenges by proposing an artificial intelligence-based framework for source-load coordination at the grid edge.
METHODS: The proposed framework integrates deep reinforcement learning, federated learning, and edge computing technologies. It employs multi-timescale optimization and distributed cooperative control methods to enhance the distribution network's capability to accommodate renewable energy.
RESULTS: Experimental results on the IEEE 33-node system demonstrate that the proposed framework can reduce the wind and solar curtailment rate by 65.5% and decrease operational costs by 12.7% compared to traditional methods, while simultaneously satisfying all distribution network security constraints.
CONCLUSION: This research provides a theoretical foundation and a technical pathway for building a new data-driven intelligent control system for distribution networks.
Downloads
References
[1] Li P, Zhong H M, Ma H W, et al. Multi-time-scale source-load-storage coordinated optimization and regulation of active distribution networks based on deep reinforcement learning. Automation of Electric Power Systems. 2024; 48(5): 1487-1496.
[2] Wu W C. AI-empowered operation and regulation of active distribution networks: collaborative, secure, robust, and llm-enhanced multi-agent reinforcement learning. Electric Power. 2025, 58(8): 1-12.
[3] Wang Z H, Gao H J, Gao Y W, et al. Multi-level dynamic reconfiguration and optimal operation method of urban distribution networks based on deep reinforcement learning. Automation of Electric Power Systems. 2022, 46(14): 1-9.
[4] Li P, Han B, Han Y D, et al. Research on source-grid-load-storage collaborative interactive control of distribution networks for new power systems. Power System Technology. 2025, 49(2): 1-10.
[5] Quan H, Peng X G, Liu H Y, et al. Real-time voltage optimization control method for distribution networks based on deep reinforcement learning. Automation of Electric Power Systems. 2023, 47(19): 1-8.
[6] Li P, Xi W, Li P, et al. Analysis and prospect of key issues in digital transformation of distribution networks based on edge computing. Automation of Electric Power Systems. 2024, 48(1): 1-10.
[7] Xi W, Li P, Li P, et al. Two-stage PMU data compression method for distribution network edge computing devices. Automation of Electric Power Systems. 2023, 47(11): 1-8.
[8] Wang B B, Zhu J, Wang J L, et al. Federated learning-based industrial power load forecasting framework under smart meter data privacy protection. Automation of Electric Power Systems. 2023, 47(15): 1-8.
[9] Yang Y, Chen Y P, Shu Y L, et al. Edge collaboration method for construction of digital twin models of distribution networks. Automation of Electric Power Systems. 2024, 48(3): 1-9.
[10] Jiangxing Intelligence. Practical application cases of edge computing + smart grid solutions. Smart Grid. 2020, 14(5): 1-6.
[11] Sun C X, Yang L, Wang X P, et al. Delay optimization deployment method of edge computing network service function chains combined with deep reinforcement learning. Computer Engineering and Applications. 2024, 60(4): 1363-1372.
[12] Sun C X, Yang L, Wang X P, et al. Fine-tuning method of roberta pre-trained model based on federated partition learning and low-rank adaptation. Computer Engineering and Applications. 2024, 60(1): 1-10.
[13] Tian Y A, Ning L, Wang C, et al. Model lightweight in deep learning and application of specific schemes. Computer Engineering and Applications. 2024, 60(11): 1-8.
[14] Tian Y A, Ning L, Wang C, et al. Research on lightweight algorithm of deep reinforcement learning models. Computer Engineering and Applications. 2024, 60(11): 1-8.
[15] He L Z, Xie Y M, Li L M, et al. Efficient federated learning scheme optimized based on game theory. Computer Engineering and Applications. 2022, 58(1): 1-8.
[16] Duo C H, Qi G L, Mei H W, et al. Task offloading decision and edge resource allocation optimization method in distribution networks. Computer Engineering and Applications. 2023, 59(8): 1-9.
[17] Wang B B, Zhu J, Wang J L, et al. User electricity price response characterization and its application based on federated learning under data silos. Automation of Electric Power Systems. 2024, 48(1): 1-10.
[18] He L Z, Xie Y M, Li L M, et al. Cooperative game optimization scheduling method for prosumer groups based on edge-cloud collaboration. Automation of Electric Power Systems. 2022, 46(12): 1-9.
[19] Qiu G F, He C, Luo Z, et al. Master-slave game economic dispatch of distribution networks considering uncertainties of renewable energy accommodation and demand response. Automation of Electric Power Systems. 2021, 45(2): 1-9.
[20] Caihongjin Industrial Group, et al. Source-grid-load-storage coordination inventory of domestic microgrid enterprises. Automation of Electric Power Systems. 2025, 58(8): 1-12.
[21] Jiangxing Intelligence, et al. Edge-cloud collaboration practice for source-grid-load-storage integration. Automation of Electric Power Systems. 2025, 58(8): 1-14.
[22] Acrel Electric Co., Ltd., et al. Source-grid-load-storage collaborative interaction: breaking the traditional “source follows load” power mode. Automation of Electric Power Systems. 2025, 58(6): 1-9.
[23] China Unicom Group, et al. Edge Computing (Distributed Computing Framework)[J]. Automation of Electric Power Systems, 2025, 58(7): 1-9.
[24] Zhang Wenjun, Hu Hao, Yi Shanjun, et al. Dynamic Optimal Scheduling Model of Distribution Networks Based on Multi-Energy Coordinated Response[J]. Automation of Electric Power Systems, 2022, 46(14): 1-9.
[25] Baidu Cloud, et al. Full-scale Outbreak of Edge Computing: A New Computing Power System Extending from Cloud Centers to "Everything as a Node"[J]. Automation of Electric Power Systems, 2025, 58(11): 1-10.
[26] Zhao Jingjing, Zhu Jiongda, Li Zhenkun, et al. Two-Stage Intra-day Distributed Optimal Scheduling of Distribution Networks Considering Robust Balance of Flexible Supply and Demand[J]. Automation of Electric Power Systems, 2022, 46(22): 1-9.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Jie Chen, Haodong Ren, Jiaxu Zhou, Binyu Liu, Guangru Zhang
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open-access article distributed under the terms of the Creative Commons Attribution CC BY 4.0 license, which permits unlimited use, distribution, and reproduction in any medium so long as the original work is properly cited.
