Key points are not available for this paper at this time.
In order to improve the efficiency and accuracy of automobile engine fault maintenance, an accurate retrieval method of automobile engine fault driven by knowledge graph was proposed. Firstly, the definition and framework of knowledge graph are discussed. The entity extraction of engine fault features was carried out by multi-source neural network, and the disambiguation of fault entities was carried out by integrating entity link technologies; Secondly, fault knowledge reasoning is carried out to eliminate the wrong knowledge in the knowledge base and infer new knowledge to form a complete knowledge graph.. On this basis, the retrieval subgraph of engine fault semantics is designed. Combined with the influence of physical distance and proximity, the retrieval result evaluation model is established, and the subgraph matching was carried out based on the similarity calculation of graph structure and semantic information. Finally, four knowledge graphs including entity equipment graph, ontology graph, maintenance rule graph and history graph were constructed by selecting some automobile engine fault cases from 2017 to 2020. Finally, the process architecture of engine fault search and analysis is constructed and the effectiveness of the proposed method was verified by precision rate and recall rata, which provides a new idea for accurate and efficient engine maintenance.
Tang et al. (Fri,) studied this question.