Knowledge Extraction and Application of Railway Operation Equipment Fault Diagnosis Based on Knowledge Graph

Focusing on the issue where traditional fault diagnosis methods cannot effectively utilize the textual records of railway equipment faults and heavily rely on experts' experience, a knowledge graph-based fault analysis method for railway operation equipment is proposed. By designing preprocessing rules to process the text corpus to be annotated, and combining the attention mechanism (ATT) and bi-directional gated recurrent neural network (BiGRU) to complete the knowledge extraction, the knowledge graph of railway operation equipment faults is constructed. On this basis, a cause recommendation strategy combining query and reasoning is designed, which can give reasonable fault analysis suggestions based on the faults that have occurred. This research utilizes the dataset sourced from a railway administration, examines the effectiveness of the knowledge extraction method through experiments, and analyzes the feasibility of utilizing the constructed knowledge graph to assist in fault analysis through cause recommendation.