A Large Language Model Question Answering System for Power System Equipment Fault Diagnosis based on RAGFlow

This paper aims to construct a knowledge graph-enhanced power system fault diagnosis question-answering system based on Retrieval-Augmented Generation (RAG), addressing challenges in power equipment fault diagnosis such as fragmented data sources, over-reliance on expert experience, low diagnostic accuracy, and insufficient intelligent capabilities. By integrating large language models (LLMs) with domain-specific knowledge bases, the proposed system leverages the open-source framework RAGFlow to retrieve relevant information and generate diagnostic recommendations, thereby enhancing the model’s domain expertise and prediction accuracy to deliver an intelligent solution for power system fault diagnosis. Experimental validation demonstrates that the system achieves significant improvements in diagnostic efficiency and accuracy.