摘要
基于工业大数据的故障诊断技术能够有效分析动设备的"健康"状况,但面临数据源单一,故障样本少,尤其是算法模型使用范围窄的问题,大幅降低了故障诊断的准确性。为此,提出一种基于标识解析和可信度矩阵的动设备故障诊断新模型。该模型通过标识解析技术,实现对多维多源数据的汇集,通过可信度矩阵继承各常规算法的评估可信度,促进了不同数据源的数据集成和不同算法的优势叠加,进而形成了动设备故障的组合式诊断模型,提升了诊断的精确度和模型的适用性。最后,以石化行业压缩机轴承故障诊断为例,验证了本模型的优越性。
Considering the fact that insufficient data source, few fault samples and narrow application range of the algorithm model jointly characterize the industrial big data-based fault diagnosis technology which to be used to effectively analyze health condition of the movement equipment, a new fault diagnosis model based on identity resolution and credibility matrix was proposed for fault diagnosis of the movement equipment. In this model, through making use of identity resolution technology, the multi-source and multidimension data can be collected and the assessment credibility of various mainstream algorithms are inherited by creating the credibility matrix to promote data integration of various data sources and advantages multiplying, by this way, a combined diagnosis model for faults of the movement equipment comes into being to improve accuracy and applicability of the diagnosis. Finally, a compressor bearing fault diagnosis in petrochemical industry verified the superiority of this model.
作者
朱林全
蒋文英
李朋
邢镔
ZHU Lin-quan;JIANG Wen-ying;LI Peng;XING Bin(College of Mechanical Engineering,Chongqing University;Chongqing Industrial Big Data Innovation Center;State Key Engineering Laboratory of Industrial Big Data Application Technology)
出处
《化工自动化及仪表》
CAS
2020年第2期134-142,共9页
Control and Instruments in Chemical Industry
关键词
故障诊断
轴承故障
压缩机
标识解析
可信度矩阵
综合诊断
fault diagnosis
bearing fault
compressor
identity resolution
credibility matrix
comprehensive diagnosis
