{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,8]],"date-time":"2026-07-08T05:14:37Z","timestamp":1783487677401,"version":"3.55.0"},"reference-count":39,"publisher":"MDPI AG","issue":"7","license":[{"start":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T00:00:00Z","timestamp":1782691200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Science and Technology Program of China Southern Power Grid Co., Ltd.","award":["036000KC23090003 (GDKJXM20231024)"],"award-info":[{"award-number":["036000KC23090003 (GDKJXM20231024)"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Computers"],"abstract":"<jats:p>Voltage and reactive power control in active distribution networks is a safety-critical and highly dynamic problem, where traditional optimization methods often struggle to balance efficiency and robustness under complex operating conditions. Recently, large language models (LLMs) have shown promise in sequential decision-making tasks, but their direct application to power system control remains limited by the lack of physical grounding and safety guarantees. In this paper, we propose SafeVolt, a closed-loop LLM-based framework that integrates multi-candidate action generation, simulator-in-the-loop evaluation, and a fine-tuned expert judge for safety-aware decision making. In addition, a high-level rule distillation mechanism that converts successful control experiences into reusable operational axioms is introduced to enable iterative self-improvement. Experiments on a standard distribution network scenario demonstrate that the proposed method outperforms representative baselines, achieving substantial improvements in average reward, voltage violation rate, reactive power loss, and system stability. In particular, voltage violations and extreme events are substantially reduced, indicating enhanced operational safety. These results suggest that combining LLM reasoning with physical simulation and structured feedback provides a promising direction for reliable and adaptive power system control.<\/jats:p>","DOI":"10.3390\/computers15070422","type":"journal-article","created":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T15:27:37Z","timestamp":1782746857000},"page":"422","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["SafeVolt: Closed-Loop Large Language Model Framework for Safety-Aware Voltage Control in Active Distribution Networks"],"prefix":"10.3390","volume":"15","author":[{"given":"Zhijun","family":"Shen","sequence":"first","affiliation":[{"name":"Power Dispatch Control Center of Guangdong Power Grid Co., Ltd., Guangzhou 510335, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qian","family":"Guo","sequence":"additional","affiliation":[{"name":"Power Dispatch Control Center of Guangdong Power Grid Co., Ltd., Guangzhou 510335, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2602-2219","authenticated-orcid":false,"given":"Kaiyuan","family":"Pang","sequence":"additional","affiliation":[{"name":"Power Dispatch Control Center of Guangdong Power Grid Co., Ltd., Guangzhou 510335, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xinlei","family":"Cai","sequence":"additional","affiliation":[{"name":"Power Dispatch Control Center of Guangdong Power Grid Co., Ltd., Guangzhou 510335, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhenfan","family":"Yu","sequence":"additional","affiliation":[{"name":"Power Dispatch Control Center of Guangdong Power Grid Co., Ltd., Guangzhou 510335, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kunhao","family":"Feng","sequence":"additional","affiliation":[{"name":"School of Electric Power Engineering, South China University of Technology, Guangzhou 510641, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tao","family":"Yu","sequence":"additional","affiliation":[{"name":"School of Electric Power Engineering, South China University of Technology, Guangzhou 510641, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,6,29]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"165","DOI":"10.1016\/j.arcontrol.2023.03.008","article-title":"Voltage Regulation in Distribution Grids: A Survey","volume":"55","author":"Srivastava","year":"2023","journal-title":"Annu. Rev. Control"},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Yu, Z., Zhou, B., Yang, D., Wu, W., Lv, C., and Cui, Y. (2024). Power System Operation Mode Calculation Based on Improved Deep Reinforcement Learning. Mathematics, 12.","DOI":"10.3390\/math12010134"},{"key":"ref_3","first-page":"8842463","article-title":"How to Construct a Power Knowledge Graph with Dispatching Data?","volume":"2020","author":"Fan","year":"2020","journal-title":"Sci. Program."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"2152","DOI":"10.1109\/TSG.2013.2246199","article-title":"Model Predictive Control of Voltages in Active Distribution Networks","volume":"4","author":"Valverde","year":"2013","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"312","DOI":"10.1109\/TSG.2020.3018633","article-title":"Distributed Coordinated Reactive Power Control for Voltage Regulation in Distribution Networks","volume":"12","author":"Tang","year":"2021","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"1544","DOI":"10.1016\/j.joule.2024.05.009","article-title":"Exploring the Capabilities and Limitations of Large Language Models in the Electric Energy Sector","volume":"8","author":"Majumder","year":"2024","journal-title":"Joule"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"8925","DOI":"10.1038\/s41598-025-91940-x","article-title":"A Large Language Model for Advanced Power Dispatch","volume":"15","author":"Cheng","year":"2025","journal-title":"Sci. Rep."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"100688","DOI":"10.1016\/j.egyai.2026.100688","article-title":"Power System Operational Reliability Evaluation with Retrieval-Augmented Generation Enhanced Large Language Model","volume":"24","author":"Cheng","year":"2026","journal-title":"Energy AI"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"100591","DOI":"10.1016\/j.egyai.2025.100591","article-title":"Intelli-Dispatch-SQL: An LLM-Based Agent for Reliable Text-to-SQL in Power Dispatching","volume":"22","author":"Ni","year":"2025","journal-title":"Energy AI"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Li, J., Chen, J., Ren, R., Cheng, X., Zhao, X., Nie, J.-Y., and Wen, J.-R. (2024). The Dawn After the Dark: An Empirical Study on Factuality Hallucination in Large Language Models. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Association for Computational Linguistics.","DOI":"10.18653\/v1\/2024.acl-long.586"},{"key":"ref_11","unstructured":"Zhou, W., Jiang, Y.E., Wilcox, E., Cotterell, R., and Sachan, M. (2023). Controlled Text Generation with Natural Language Instructions. Proceedings of the 40th International Conference on Machine Learning, PMLR. Available online: https:\/\/proceedings.mlr.press\/v202\/zhou23g.html."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Ouyang, S., and Li, L. (2023). AutoPlan: Automatic Planning of Interactive Decision-Making Tasks with Large Language Models. Findings of the Association for Computational Linguistics: EMNLP 2023, Association for Computational Linguistics.","DOI":"10.18653\/v1\/2023.findings-emnlp.205"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"106498","DOI":"10.1016\/j.epsr.2020.106498","article-title":"Coordination of OLTC and Smart Inverters for Optimal Voltage Regulation of Unbalanced Distribution Networks","volume":"187","author":"Li","year":"2020","journal-title":"Electr. Power Syst. Res."},{"key":"ref_14","unstructured":"Liao, W., Bak-Jensen, B., Pillai, J.R., Wang, Y., and Wang, Y. (2021). A Review of Graph Neural Networks and Their Applications in Power Systems. arXiv."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"4086","DOI":"10.1109\/TSG.2023.3239740","article-title":"Spatio-Temporal Graph Convolutional Neural Networks for Physics-Aware Grid Learning Algorithms","volume":"14","author":"Wu","year":"2023","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"109531","DOI":"10.1016\/j.ijepes.2023.109531","article-title":"Multi-agent Graph Reinforcement Learning for Decentralized Volt-VAR Control in Power Distribution Systems","volume":"155","author":"Hu","year":"2024","journal-title":"Int. J. Electr. Power Energy Syst."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"121743","DOI":"10.1016\/j.apenergy.2023.121743","article-title":"Physical-assisted Multi-agent Graph Reinforcement Learning Enabled Fast Voltage Regulation for PV-rich Active Distribution Network","volume":"351","author":"Chen","year":"2023","journal-title":"Appl. Energy"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"444","DOI":"10.1007\/978-981-97-5618-6_37","article-title":"Enhancing Large Language Models with Graph-Based Node Sampling for Fault Attribution in Power Distribution Networks","volume":"Volume 14874","author":"Huang","year":"2024","journal-title":"Advanced Intelligent Computing Technology and Applications"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"111845","DOI":"10.1016\/j.patcog.2025.111845","article-title":"MGRAG: MultiModal Grid-Aware Retrieval Augmentation Generation Framework for Power Grid Work Tickets","volume":"169","author":"Mao","year":"2026","journal-title":"Pattern Recognit."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"361","DOI":"10.1109\/TSG.2020.3010130","article-title":"Deep Reinforcement Learning Based Volt-VAR Optimization in Smart Distribution Systems","volume":"12","author":"Zhang","year":"2021","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"4137","DOI":"10.1109\/TSG.2021.3072251","article-title":"Data-Driven Multi-agent Deep Reinforcement Learning for Distribution System Decentralized Voltage Control with High Penetration of PVs","volume":"12","author":"Cao","year":"2021","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref_22","first-page":"3271","article-title":"Multi-Agent Reinforcement Learning for Active Voltage Control on Power Distribution Networks","volume":"Volume 34","author":"Wang","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_23","unstructured":"Qu, Y., Ma, J., and Wu, F. (2024). Safety Constrained Multi-Agent Reinforcement Learning for Active Voltage Control. Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, International Joint Conferences on Artificial Intelligence (IJCAI). Available online: https:\/\/www.ijcai.org\/proceedings\/2024\/21."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"5965","DOI":"10.1109\/TPWRS.2023.3315543","article-title":"On the Potential of ChatGPT to Generate Distribution Systems for Load Flow Studies Using OpenDSS","volume":"38","author":"Bonadia","year":"2023","journal-title":"IEEE Trans. Power Syst."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"4747","DOI":"10.1109\/TPWRS.2023.3338961","article-title":"Real-Time Optimal Power Flow with Linguistic Stipulations: Integrating GPT-Agent and Deep Reinforcement Learning","volume":"39","author":"Yan","year":"2024","journal-title":"IEEE Trans. Power Syst."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"4297","DOI":"10.1109\/TSG.2025.3583934","article-title":"LLM-Based Adaptive Distribution Voltage Regulation Under Frequent Topology Changes: An In-Context MPC Framework","volume":"16","author":"Jena","year":"2025","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Christou, P., Islam, M.Z., Lin, Y., and Xiong, J. (2025). LLM4DistReconfig: A Fine-Tuned Large Language Model for Power Distribution Network Reconfiguration. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), Association for Computational Linguistics.","DOI":"10.18653\/v1\/2025.naacl-long.208"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"8634","DOI":"10.52202\/075280-0377","article-title":"Reflexion: Language Agents with Verbal Reinforcement Learning","volume":"Volume 36","author":"Shinn","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"19724","DOI":"10.1609\/aaai.v38i17.29946","article-title":"MemoryBank: Enhancing Large Language Models with Long-Term Memory","volume":"Volume 38","author":"Zhong","year":"2024","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"ref_30","unstructured":"Wang, G., Xie, Y., Jiang, Y., Mandlekar, A., Xiao, C., Zhu, Y., Fan, L., and Anandkumar, A. (2024). Voyager: An Open-Ended Embodied Agent with Large Language Models. Trans. Mach. Learn. Res., 1\u201344. Available online: https:\/\/openreview.net\/forum?id=ehfRiF0R3a."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"101211","DOI":"10.1016\/j.patter.2025.101211","article-title":"RePower: An LLM-driven Autonomous Platform for Power System Data-Guided Research","volume":"6","author":"Liu","year":"2025","journal-title":"Patterns"},{"key":"ref_32","unstructured":"Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K., and Cao, Y. (2023). ReAct: Synergizing Reasoning and Acting in Language Models. The Eleventh International Conference on Learning Representations, OpenReview. Available online: https:\/\/openreview.net\/forum?id=WE_vluYUL-X."},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"68539","DOI":"10.52202\/075280-2997","article-title":"Toolformer: Language Models Can Teach Themselves to Use Tools","volume":"Volume 36","author":"Schick","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Saha, B.K., Aarthi, V., and Naidu, O.D. (2025). DrAgent: An Agentic Approach to Fault Analysis in Power Grids Using Large Language Models. 2025 International Conference on Artificial Intelligence in Information and Communication (ICAIIC), IEEE.","DOI":"10.1109\/ICAIIC64266.2025.10920654"},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"51991","DOI":"10.52202\/075280-2264","article-title":"CAMEL: Communicative Agents for \u201cMind\u201d Exploration of Large Language Model Society","volume":"Volume 36","author":"Li","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_36","unstructured":"Wu, Q., Bansal, G., Zhang, J., Wu, Y., Li, B., Zhu, E., Jiang, L., Zhang, X., Zhang, S., and Liu, J. (2024). AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation. First Conference on Language Modeling (COLM), OpenReview. Available online: https:\/\/www.microsoft.com\/en-us\/research\/publication\/autogen-enabling-next-gen-llm-applications-via-multi-agent-conversation-framework\/."},{"key":"ref_37","unstructured":"(2026, June 22). MAPDN Dataset. Multi-Agent Power Distribution Networks Voltage Control Dataset. Available online: https:\/\/huggingface.co\/datasets\/hsvgbkhgbv\/Multi-Agent-Power-Distribution-Networks."},{"key":"ref_38","unstructured":"Qwen Team (2024). Qwen2.5: A Party of Foundation Models, Qwen. Available online: https:\/\/qwenlm.github.io\/blog\/qwen2.5\/."},{"key":"ref_39","unstructured":"Ollama (2026, June 22). Ollama: Run Large Language Models Locally. Available online: https:\/\/ollama.com\/."}],"container-title":["Computers"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2073-431X\/15\/7\/422\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,8]],"date-time":"2026-07-08T04:29:06Z","timestamp":1783484946000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2073-431X\/15\/7\/422"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,29]]},"references-count":39,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2026,7]]}},"alternative-id":["computers15070422"],"URL":"https:\/\/doi.org\/10.3390\/computers15070422","relation":{},"ISSN":["2073-431X"],"issn-type":[{"value":"2073-431X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,29]]}}}