Human–AI Collaborative Decision Model for Low-Voltage Diagnosis and Closed-Loop Mitigation in Distribution Transformer Areas
Low-voltage mitigation in distribution transformer areas is rarely a single-step technical problem. Field diagnosis, plan selection, and post-implementation review are often handled separately, limiting the reuse of operational evidence and professional judgment. We recast mitigation as a six-stage human–AI closed loop linking state sensing, causal diagnosis, collaborative decision-making, implementation, effectiveness evaluation, and knowledge updating. For each candidate measure, the model considers data reliability, AI diagnostic confidence, human judgment reliability, and implementation risk. An expected-loss criterion assigns a machine-led, human–AI collaborative, or human-led mode. Safety and engineering constraints screen the available measures, after which an uncertainty-aware score identifies the preferred plan–mode pair. Evidence collected after implementation updates causal probabilities and the cause–measure knowledge base. Three constructed scenarios illustrate how decision authority moves from machine-led analysis toward professional control as implementation risk rises or diagnostic evidence weakens. These cases establish the internal consistency of the decision logic; they do not demonstrate improved field accuracy. The framework provides a transparent basis for subsequent calibration and validation with operational data.
[1] Shen S, Zhong Q, Xu Z, et al., 2025, Mining User Low-Voltage Violation Patterns Using Hierarchical Affinity-Propagation Clustering. Electric Power Engineering Technology, 44(1): 30–38.
[2] General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China, Standardization Administration of China, 2008, Power Quality—Deviation of Supply Voltage. Standards Press of China.
[3] Guo S, Zhao X, Sun G, et al., 2025, Analysis and Prediction of Over-Limit Characteristics for Regional Distribution-Transformer Voltage Using Bilayer Clustering and Correlation-Feature Screening. Southern Power System Technology, 19(2): 19–27.
[4] Huang Y, Yang S, Cao X, et al., 2023, Topology Identification of Low-Voltage Distribution Networks Using Segmented Current Features and a Random-Forest Algorithm. Power Demand Side Management, 25(2): 63–69.
[5] Wei J, Tian S, Chen W, et al., 2025, Low-Voltage Prediction in Distribution Transformer Areas Using a GA–BP–LSTM Model. Electric Engineering, 46(1): 15–20.
[6] Zhou Y, Zhang B, Huang W, et al., 2023, Low-Voltage Prediction for Distribution Transformer Areas Using an LSTM–BP Combined Model. Journal of Electric Power Science and Technology, 38(5): 177–186.
[7] Chen W, Chen J, Fan Y, et al., 2024, Design and application of smart distribution station areas based on intelligent fusion terminals with cloud-edge coordination. Electric Power, 57(4): 190–199.
[8] Fan Y, Wu H, Li Z, et al., 2024, A flexible orchestration of lightweight AI for edge computing in low-voltage distribution network. Frontiers in Energy Research, 12: 1424663.
[9] Campos V, Klyagina O, Andrade J, et al., 2024, ML-Assistant for Human Operators Using Alarm Data to Solve and Classify Faults in Electrical Grids. Electric Power Systems Research, 236: 110886.
[10] Machlev R, Heistrene L, Perl M, et al., 2022, Explainable artificial intelligence techniques for energy and power systems: review, challenges and opportunities. Energy and AI, 9: 100169.
[11] Naiseh M, Al-Thani D, Jiang N, et al., 2023, How the Different Explanation Classes Impact Trust Calibration: The Case of Clinical Decision Support Systems. International Journal of Human-Computer Studies, 169: 102941.
[12] Yamada S, 2022, Trust Calibration as Rationality for Human–AI Cooperative Decision Making. Cognitive Studies, 29(3): 364–370.
[13] Amershi S, Weld S, Vorvoreanu M, et al., 2019, Guidelines for Human–AI Interaction. Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, 1–13.
[14] Fügener A, Grahl J, Gupta A, et al., 2022, Cognitive Challenges in Human–Artificial Intelligence Collaboration: Investigating the Path Toward Productive Delegation. Information Systems Research, 33(2): 678–696.
[15] Parasuraman R, Sheridan T, Wickens C, 2000, A Model for Types and Levels of Human Interaction with Automation. IEEE Transactions on Systems, Man, and Cybernetics—Part A: Systems and Humans, 30(3): 286–297.
[16] Guo X, Zheng F, Wang S, 2024, Mobile Energy-Storage Scheduling in Distribution Networks Considering Voltage-Violation Risk. Power Demand Side Management, 26(4): 37–42.
[17] Li B, Li Y, Hai Z, et al., 2025, Online Mitigation of Three-Phase Imbalance in Distribution Networks Using Multi-Agent Deep Reinforcement Learning. Proceedings of the CSEE, 45(5): 1729–1741.
[18] Yang Y, Wen F, Zhou X, et al., 2023, Review of Coordinated Voltage Control in Distribution Systems with High Photovoltaic Penetration. Electric Power Automation Equipment, 43(10): 48–58.
[19] Yuan B, Yang P, Wang T, et al., 2025, Multi-Objective Dispatch of Distributed Energy Storage in High-PV Distribution Transformer Areas Considering Multiple Stakeholders. Modern Electric Power, 42(6): 1268–1277.
[20] Zhang X, Wu Z, Sun Q, et al., 2024, Application and Progress of Artificial Intelligence Technology in the Field of Distribution Network Voltage Control: A Review. Renewable and Sustainable Energy Reviews, 192: 114282.
[21] Schemmer M, Kühl N, Benz C, et al., 2023, Appropriate Reliance on AI Advice: Conceptualization and the Effect of Explanations. Proceedings of the 28th ACM International Conference on Intelligent User Interfaces, 410–422.
