GridMind: LLMs-Powered Agents for Power System Analysis and Operations

2025-09-02

AI Summary

Conventional power system analysis workflows are cumbersome, hindering efficient decision-making in modern grids. Method: This paper proposes a multi-agent system integrating large language models (LLMs) with deterministic engineering solvers to establish a conversational scientific computing paradigm. It enables natural-language-driven joint reasoning for power flow analysis and N−1 contingency verification. The system couples AC optimal power flow and security analysis solvers via function-calling interfaces and employs a lightweight agent architecture supporting context retention, domain-specific knowledge embedding, and efficient deployment of smaller LLMs. Contribution/Results: Experiments on IEEE benchmark systems demonstrate stable, high-accuracy solutions across LLM scales; notably, compact models reduce inference latency by over 40% while preserving analytical fidelity. This work pioneers native integration of LLM interactive capabilities with deterministic power system computation, substantially enhancing accessibility, robustness, and real-time performance of complex grid analysis.

Technology Category

Multiagent Systems: Agent CommunicationCognitive Modeling & Cognitive Systems: Agent ArchitecturesMachine Learning: Large Multimodal Models (LMMs)

Application Category

Economics, Online Markets and Human Computation: Cost models of using LLMs in production systemsSemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactionsSearch and Retrieval-Augmented AI: Large language models for search

Abstract

The complexity of traditional power system analysis workflows presents significant barriers to efficient decision-making in modern electric grids. This paper presents GridMind, a multi-agent AI system that integrates Large Language Models (LLMs) with deterministic engineering solvers to enable conversational scientific computing for power system analysis. The system employs specialized agents coordinating AC Optimal Power Flow and N-1 contingency analysis through natural language interfaces while maintaining numerical precision via function calls. GridMind addresses workflow integration, knowledge accessibility, context preservation, and expert decision-support augmentation. Experimental evaluation on IEEE test cases demonstrates that the proposed agentic framework consistently delivers correct solutions across all tested language models, with smaller LLMs achieving comparable analytical accuracy with reduced computational latency. This work establishes agentic AI as a viable paradigm for scientific computing, demonstrating how conversational interfaces can enhance accessibility while preserving numerical rigor essential for critical engineering applications.

Problem

• Integrating AI agents with power engineering solvers for grid analysis
• Enabling conversational interfaces for complex power system operations
• Maintaining numerical precision while using natural language interfaces

Innovation

• LLMs integrated with engineering solvers
• Agents coordinate power flow via natural language
• Maintains numerical precision through function calls