面向燃驱压缩机组的故障知识本体建模及应用研究  

Research on fault knowledge ontology modeling and application for combustion-driven compressor units

作　　者：王明达[1] 李云飞 吴志生 张榜 WANG Mingda;LI Yunfei;WU Zhisheng;ZHANG Bang(School of Mechanical and Electrical Engineering,China University of Petroleum,Qingdao 266580,Shandong,China)

机构地区：[1]中国石油大学(华东)机电工程学院,山东青岛266580

出　　处：《安全与环境学报》2023年第10期3472-3482,共11页Journal of Safety and Environment

摘　　要：针对当前燃驱压缩机组故障诊断效率低下、故障知识分散、共享和重复使用困难等问题,提出了一种基于本体和故障模式、影响及危害性分析(Failure Mode,Effects and Criticality Analysis,FMECA)的燃驱压缩机组故障诊断方法。首先,该方法以机组FMECA数据为知识源,通过本体建模从知识源中提取深层知识和浅层知识。然后,使用本体开发软件Protégé5.5将抽取得到的类、属性和实例构建故障知识库。在此基础上,通过JESS(Java Expert Shell System)规则进行故障诊断推理,维修人员可以快速找到机组故障原因,并选择合适的处理方法。最后,以某型燃驱压缩机组为对象验证该故障诊断方法的有效性,结果表明,该方法提高了故障诊断的能力和故障知识的利用率,能够为诊断决策提供良好的知识支持。The scattered knowledge and complex knowledge system of fault diagnosis of driven compressor sets make it difficult for current fault diagnosis methods to obtain good results in practical engineering applications,especially in the case of knowledge sharing,reuse,and lack of fault knowledge of newly hired maintenance personnel.Therefore,to solve the above problems,this paper adopts the knowledge representation method of ontology and proposes an ontology-based Failure Mode,Effects and Criticality Analysis(FMECA)fault diagnosis method.Firstly,FMECA and Fault Tree Analysis(FTA)methods are used to obtain fault knowledge,and fault knowledge is divided into deep knowledge and shallow knowledge.Secondly,an improved ontology modeling method based on the seven-step method and the skeleton method is used to extract deep and shallow knowledge from the knowledge sources and defines a formal representation of the fault knowledge ontology,as well as concepts,attributes,relationships,and entities in the ontology.Then,the ontology development software Protégé5.5 and RDF data query technology(SPARQL Protocol and RDF Query Language)are used to construct the knowledge base for fault diagnosis of combustion drive compressor group,and the OWL(Web Ontology Language)language and SWRL(Semantic Web Rule Language)languages are used to describe the conceptual attributes of entities and fault diagnosis rules in the knowledge source respectively.On this basis,the fault diagnosis process is adopted from high to low according to the hierarchy of equipment,and the diagnosis reasoning is carried out by JESS(Java Expert Shell System)rule engine so that the maintenance personnel can quickly find the cause of the fault of the combustion-driven compressor set and choose the appropriate fault treatment method.Finally,the effectiveness of the fault diagnosis method is verified with a certain type of combustiondriven compressor set.The results show that the fault diagnosis method combines deep knowledge and shallow knowledge and stores them in the for

关 键 词：安全工程 燃驱压缩机组 故障诊断 本体 故障模式、影响及危害性分析(FMECA) 

分 类 号：X937[环境科学与工程—安全科学]