机械科学与技术2025，Vol.44Issue(2)：225-235,11.DOI:10.13433/j.cnki.1003-8728.20230177
维修知识图谱与深度学习网络在发动机故障智能推理中的应用
Study on Maintenance Knowledge Graph and Deep Learning Network in Intelligent Reasoning of Engine Fault
摘要
Abstract
Aiming at the unified modeling of the lack of knowledge level of multi-source heterogeneous engine information,an application framework of knowledge graph of automobile engine fault maintenance auxiliary decision-making including source data layer,map construction layer,inference decision layer and fault retrieval layeris proposed.Firstly,the multi-source heterogeneous information in the engine fault maintenance process is condensed into a structured knowledge network,and the knowledge graph is classified according to the engine fault business scenarios and requirements.Then,the BERT-BiLSTM-At deep neural network is used for fault information extraction,and the improved RETE algorithm for fault repair knowledge reasoning,and then the construction of engine fault maintenance knowledge graph is completed.Finally,WPF is used to construct and realize the fault information analysis and retrieval and intelligent auxiliary decision-making system based on knowledge graph.The application and challenges of future knowledge graph in intelligent maintenance of engine faults are summarized and prospected.关键词
发动机故障/知识图谱/神经网络/信息抽取/知识推理/分析检索Key words
engine fault/knowledge graph/neural network/information extraction/knowledge reasoning/analysis and retrieval分类
计算机与自动化引用本文复制引用
蒲昊苒,阴艳超,徐成现..维修知识图谱与深度学习网络在发动机故障智能推理中的应用[J].机械科学与技术,2025,44(2):225-235,11.基金项目
国家自然科学基金项目(52065033)、云南省重大科技专项计划(202002AD080001)及云南省重大科技项目(202202AG050002) （52065033）
