Research in the International Journal of Information and Communication Technology suggests that so-called knowledge graphs, a form of AI-based data organisation, could improve the reliability and maintenance of power communication systems that help keep the lights on and modern electricity grids running smoothly.
The researchers report that such a system works better than a conventional database in query efficiency, fault diagnosis, and operational decision-making. They explain that this technology could be used to help utility operators anticipate equipment failures earlier and manage increasingly complex power networks more effectively.
Power communication equipment functions as the information backbone of electricity grids, enabling substations, sensors and control centres to exchange data in real-time. However, as grids are becoming more digitalised through smart sensors, distributed energy systems and private 5G networks, operators are generating far larger volumes of interconnected data that somehow has to be managed.
The researchers argue that conventional relational databases struggle with this level of complex data. Relational databases organise information into rigid tables linked by predefined relationships. While suitable for simpler systems, the researchers say they create information silos in large infrastructure networks, where maintenance records, fault reports, environmental conditions, and operational data are fragmented across separate systems.
The proposed AI framework instead uses a knowledge graph, which represents devices, faults, maintenance activities, and communication links as interconnected nodes. By explicitly mapping relationships between all these different pieces of information, the system can identify dependencies and hidden correlations more effectively. In order to integrate this information from different sources, the researchers used natural language processing (NLP), an AI technique that extracts meaning from human language.
NLP enables the system to analyse unstructured materials such as maintenance reports and technical documents alongside structured operational data. The resulting information is stored in the graph database designed specifically for highly connected data. This approach allows the utility operator to have in place predictive infrastructure management. Now, instead of relying mainly on manual inspections and operator experience when faults occur, they can predict failures in advance and carry out preventative maintenance.
Zhang, J., Chen, S., Guo, L., Xie, J., Li¸ B. and Zhong, R. (2026) ‘Research on intelligent management of the full lifecycle of power communication equipment based on knowledge graphs’, Int. J. Information and Communication Technology, Vol. 27, No. 42, pp. 72–92.
No comments:
Post a Comment