Knowledge graph-based alarm management in petrochemical enterprises: A study on fusion and analysis of multi-source heterogeneous information

Abstract

In response to the increasing emphasis on alarm management in the petrochemical industry, there has been an explosive growth in relevant information. However, this information is often scattered across different systems and databases, stored in various forms such as documents, tables, and images, making it challenging to uniformly store, share, and utilize multi-source heterogeneous information. This commonly leads to the problem of “Information Islands.” In order to effectively leverage knowledge in the field of alarm management in the petrochemical industry and overcome the challenge of non-interoperable information, a method for fusing multi-source heterogeneous information in petrochemical enterprise alarm management based on knowledge graph is proposed. This method aims to standardize the management of alarm-related information and achieve information fusion. Initially, the approach utilizes data from petrochemical enterprises and publicly available data in the field of alarm management to establish both local and global ontologies. Subsequently, mapping algorithms are designed to achieve a more accurate construction of the hybrid ontology. Based on this foundation, a knowledge graph for alarm management in the petrochemical industry is established. Additionally, corresponding modules for information storage and retrieval are developed. Through the application demonstration using real alarm management information from a petrochemical enterprise, the results indicate that the proposed method for fusing multi-source heterogeneous information in petrochemical enterprise alarm management can effectively achieve information fusion.