A digital transformation: advancing power standards with knowledge graphs and intelligent applications
Xiaoxin Gao, Qiang Yang, Jiajia Han, Jiale Liu · 2024
This paper addresses the digitization of power standards in the power industry through the development of knowledge graphs and the application of intelligent technologies. We explore the transformation of standard management, leveraging technologies like cloud computing, big data, and AI, and discuss the strategies of international standardization bodies like ISO and IEC towards machine-readable standards. The core of our research is the construction and application of a power standard knowledge graph, aimed at enhancing the management of standardized knowledge. This involves stages from data collection, incorporating OCR technology, to data management with an emphasis on data quality and security. A significant contribution of this work is the development of the PLM-based Knowledge Graph Completion Model, particularly focusing on the KG-SBERT model, which integrates semantic and structured features from pre-trained language models and knowledge graphs. Our experiments demonstrate the superiority of KG-SBERT over traditional KGE models in knowledge graph completion tasks, highlighting the effectiveness of prompt information in improving model performance. surpassing traditional knowledge graph embedding techniques in link prediction accuracy, attaining an MRR of 47.67, and Hits@5 and Hits@10 scores of 25.15, 81.54, and 93.91, respectively. This study underlines the critical role of digital transformation in the power industry, demonstrating how advanced technologies can significantly enhance the management, security, and accessibility of power standards, with broad implications for the industry's future development.