基于证据理论多源多特征融合的柱塞泵故障诊断方法  

Multi-source and multi-feature fused fault diagnosis on piston pumps based on evidence theory

作　　者：吴胜强[1,2] 姜万录[1] 

机构地区：[1]燕山大学机械工程学院,河北秦皇岛066004 [2]邢台职业技术学院,河北邢台054035

出　　处：《中国工程机械学报》2011年第1期98-102,共5页Chinese Journal of Construction Machinery

基　　金：国家自然科学基金资助项目(51075349);国家自然科学基金资助项目(50775198);河北省自然科学基金资助项目(E2008000812)

摘　　要：提出了1种基于证据理论多源多特征融合的故障诊断方法,构建了柱塞泵的故障特征向量,分析了单源多特征融合故障诊断,利用证据理论融合多个单源诊断结果,进行了多源故障诊断.对液压泵进行了试验.试验结果表明:多源数据较单源数据,更具有冗余互补性,能提高诊断的可靠性和准确性.该诊断方法是可行而有效的.In this study, a multi-source and multi-feature integrated fault diagnosis method is first proposed based on evidence theory. In this regard, the fault feature vectors are extracted for piston pumps. Then, the evidence theory is applied for fusing the results from several single-source and multi- feature fault diagnosis. Finally, it is found from the testing results on hydraulic pumps that the multi- source data possess higher redundant complementarity than single-source data. Meanwhile, the feasibility and effectiveness of this approach is proven through enhancing the diagnosis reliability and accuracy.

关 键 词：故障诊断 证据理论 时频域 多源多特征 柱塞泵 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]