Two-Stage Active Distribution Network Voltage Control via LLM-RL Collaboration: A Hybrid Knowledge-Data-Driven Approach

2026-02-25

AI Summary

This study addresses the challenges of voltage violations and power quality degradation in active distribution networks with high penetration of distributed photovoltaic (PV) generation. Existing data-driven approaches struggle to effectively integrate multi-source heterogeneous information and often rely heavily on trial-and-error tuning. To overcome these limitations, this work proposes a two-stage voltage control framework that synergistically combines large language models (LLMs) and reinforcement learning (RL) agents. In the day-ahead stage, an LLM generates scheduling strategies for on-load tap changers (OLTCs) and capacitor banks based on regional forecasts; in the intra-day stage, an RL agent optimizes PV inverter reactive power outputs using real-time nodal measurements. By incorporating an LLM self-evolution mechanism and an RL pretraining-finetuning pipeline, the framework achieves deep integration of knowledge-driven and data-driven paradigms, significantly enhancing training efficiency and voltage regulation performance while effectively mitigating voltage violations, as confirmed by ablation studies.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

The growing integration of distributed photovoltaics (PVs) into active distribution networks (ADNs) has exacerbated operational challenges, making it imperative to coordinate diverse equipment to mitigate voltage violations and enhance power quality. Although existing data-driven approaches have demonstrated effectiveness in the voltage control problem, they often require extensive trial-and-error exploration and struggle to incorporate heterogeneous information, such as day-ahead forecasts and semantic-based grid codes. Considering the operational scenarios and requirements in real-world ADNs, in this paper, we propose a hybrid knowledge-data-driven approach that leverages dynamic collaboration between a large language model (LLM) agent and a reinforcement learning (RL) agent to achieve two-stage voltage control. In the day-ahead stage, the LLM agent receives coarse region-level forecasts and generates scheduling strategies for on-load tap changer (OLTC) and shunt capacitors (SCs) to regulate the overall voltage profile. Then in the intra-day stage, based on accurate node-level measurements, the RL agent refines terminal voltages by deriving reactive power generation strategies for PV inverters. On top of the LLM-RL collaboration framework, we further propose a self-evolution mechanism for the LLM agent and a pretrain-finetune pipeline for the RL agent, effectively enhancing and coordinating the policies for both agents. The proposed approach not only aligns more closely with practical operational characteristics but also effectively utilizes the inherent knowledge and reasoning capabilities of the LLM agent, significantly improving training efficiency and voltage control performance. Comprehensive comparisons and ablation studies demonstrate the effectiveness of the proposed method.

Problem

• voltage control
• active distribution network
• distributed photovoltaics
• voltage violations
• power quality

Innovation

• LLM-RL collaboration
• two-stage voltage control
• hybrid knowledge-data-driven
• self-evolution mechanism
• distribution network optimization