Abstract: Current maintenance of smart grid dispatch models faces challenges such as low operational efficiency, high reliance on manual intervention, and difficulties in quality control. A novel intelligent operation and maintenance system is designed and implemented by integrating knowledge graph technology with the Node-RED visual workflow engine, constructing a structured operation knowledge base and dynamic process guidance. It can guide operation and maintenance personnel step by step to accurately and efficiently carry out dispatching model maintenance tasks. The system employs natural language processing techniques to extract entity relationships from operation manuals and dispatch system domain data, forming the inferable operation knowledge graph. Leveraging Node-RED’s visual programming interface, it enables flexible configuration of operation workflows. Real-time verification algorithms covering data integrity, electrical rationality, and other dimensions are developed. Through a practical case study of a municipal-level power grid, it has been demonstrated that the system improves operational efficiency by 62.5%, with a model accuracy rate exceeding 99%, providing a verifiable quality assurance mechanism for the dispatching automation system.

Key words: dispatch model operational maintenance, knowledge graph, Node-RED, multi-dimensional verification, dynamic workflow generation