Another application case for AI-empowered dynamic reasoning

By Zhang Chenxu

State Grid Henan Information & Communication Company has significantly reduced the time required to identify information system faults by developing an intelligent operations knowledge base and a root-cause analysis engine powered by dynamic reasoning.

The company recently established a Quality Control (QC) team tasked with sorting and categorizing nearly 10,000 sets of hardware and software equipment. The QC team's work follows the Plan-Do-Check-Act (PDCA) cycle, according to the company.

This initiative was launched amid the expansion of new power systems, which has increased the complexity of information system architectures. Engineers now need to sift through extensive alarms, logs, and case records during fault diagnosis, leading to prolonged troubleshooting times for information operations and maintenance teams.

To tackle these challenges, the QC team developed an intelligent operations knowledge base that consolidates over 20,000 entries of risk alerts, cases, and practical experience. Utilizing machine reading comprehension (MRC) and semantic retrieval, the system enables highly accurate Q&A, effectively bridging gaps in fault-handling experience.

Additionally, the team created a fault-analysis tool supported by a knowledge graph and the Guangming large model, enabling dynamic reasoning. This tool has already helped resolve more than 200 incidents and improved the efficiency of handling recurring faults by over 90 percent.

The QC report titled "Reducing Information System Fault-Location Time" received the top award at the 2025 National Electric Power Industry QC Achievements Exchange organized by the China Water Resources and Electric Power Quality Management Association.

zhangchenxu@chinadaily.com.cn