摘要
对旋转机械进行故障诊断时,在输入征兆条件缺失的情况下,采用传统的基于规则的故障诊断推理方法,其诊断结果会出现较大的偏差。针对这一问题,提出了一种基于知识图谱的旋转机械故障诊断推理方法。首先,阐述了基于本体的旋转机械故障诊断知识表示方法,通过构建旋转机械故障诊断知识本体表示模型进行了知识表达,在此基础上构建了旋转机械故障诊断的知识图谱;然后,结合基于关系路径的知识图谱推理方法,提出了基于知识图谱的旋转机械故障诊断方法,利用旋转机械设备结构间关系进行了故障原因推理;最后,以核电厂主泵为例,构建了主泵故障诊断的知识图谱,对基于知识图谱的旋转机械故障诊断方法进行了验证。研究结果表明:在输入征兆缺失的条件下,采用该方法得到的故障诊断准确率达到92.1%,远高于传统的基于规则的故障诊断推理方法的准确率,有效地解决了以往征兆缺失时故障诊断准确率低的问题;同时,知识图谱的应用也可为其他机械设备智能诊断方法提供一种新的思路。
In the fault diagnosis of rotating machinery,when the input symptom conditions were missing,the traditional rule-based fault diagnosis reasoning method would have a large deviation in the diagnosis results.Aiming at the problem,a rotating machinery fault diagnosis method based on knowledge graph was proposed.Firstly,an ontology-based knowledge representation method for rotating machinery fault diagnosis was described,and knowledge representation was carried out by constructing an ontology representation model for rotating machinery fault diagnosis knowledge,on the basis of which a knowledge graph for rotating machinery fault diagnosis was constructed.Then,combined with the path-based knowledge graph inference method,the diagnosis method based on knowledge graph of rotating machinery fault diagnosis was proposed,and the reasons of fault were inferred by using the relationship between structures of rotating machinery equipment.Finally,taking the main pump of nuclear power plant as an example,the knowledge map of main pump fault diagnosis was constructed,and the rotating machinery fault diagnosis method based on knowledge map was verified.The results of experimental validation show that the diagnostic accuracy of the method reaches 92.1% under the condition of missing symptoms.It is far better than the accuracy of the traditional rule-based fault diagnosis reasoning method,and effectively solves the problem of low diagnostic accuracy when the symptoms are missing.At the same time,the application of knowledge graph can also provide a new idea for other intelligent diagnosis methods of mechanical equipment.
作者
盛林
马波
张杨
SHENG Lin;MA Bo;ZHANG Yang(College of Mechanical and Electrical Engineering,Beijing University of Chemical Technology,Beijing 100029,China;Beijing Key Laboratory of High End Mechanical Equipment Health Monitoring and Self Recovery,Beijing University of Chemical Technology,Beijing 100029,China)
出处
《机电工程》
CAS
北大核心
2022年第9期1194-1202,共9页
Journal of Mechanical & Electrical Engineering
基金
国家自然科学基金资助项目(61673046,52075030)。
关键词
知识图谱构建
输入征兆缺失
基于本体的知识表示方法
解析模型
数据驱动
基于规则的故障诊断方法
knowledge graph construction
input symptom missing
ontology-based knowledge representation method
analytical models
data-driven
rule-based fault diagnosis method
