基于多源信息融合的往复式压缩机故障诊断方法

Translated title of the contribution: Reciprocating Compressor Fault Diagnosis Technology Based on Multi-source Information Fusion

Ming Zhang, Zhinong Jiang

Research output: Contribution to journalArticlepeer-review

Abstract

往复式压缩机结构复杂, 振动激励源多, 故障关联性较强, 需要依靠多种类型的传感器所采集的信息来对往复式压缩机故障进行诊断. 在融合往复式压缩机多种类型传感器采集的特征信息基础上, 提出一种基于多源信息融合的往复式压缩机故障诊断方法, 构建信息融合诊断框架. 利用往复式压缩机多种类型传感器所采集的数据信息构建特征证据体, 使用径向基神经网络对每个证据体进行初步诊断, 根据加权证据融合理论融合各个证据体初步诊断结果, 得到最终诊断结果. 使用提出的方法对往复式压缩机 3 种工况的试验数据进行融合诊断, 诊断结果表明: 使用加权证据融合理论融合多源传感器信息的诊断结果可信度高, 不确定性小, 能够准确对往复式压缩机故障状态进行诊断识别.
Translated title of the contributionReciprocating Compressor Fault Diagnosis Technology Based on Multi-source Information Fusion
Original languageChinese
Pages (from-to)46-52
Number of pages6
JournalJournal of Mechanical Engineering
DOIs
Publication statusPublished - 2017

Fingerprint Dive into the research topics of 'Reciprocating Compressor Fault Diagnosis Technology Based on Multi-source Information Fusion'. Together they form a unique fingerprint.

Cite this