摘要
鉴于复杂机械系统中故障信息的不完备及不确定性造成证据理论在故障诊断决策级阶段融合的准确性问题,提出基于多属性群决策的故障证据融合方法.利用多属性群决策的属性分析,计算基于元素属性集合的决策体差异权重,以减小融合证据源的差异;结合柴油机目标故障的相似依赖关系,利用目标故障的可信度权重对冲突焦元信息再分配,旨在提高证据融合的准确性.对R6105AZLD柴油机台架试验结果表明:本文方法可大幅提高诊断准确度和鲁棒性.
In view of accuracy of evidence theory in the stage of fault diagnosis decision fusion by incompleteness and uncertainty of fault information in complex mechanical systems,the method of fault evidence fusion based on multi-attribute group decision making was proposed.By using attribute analysis of multi-attribute group decision making,the weight of decision body based on element attribute set was calculated to reduce the difference of fusion evidence sources.Based on the similarity dependence of the target fault of the diesel engine,by using the reliability weight of the target fault,the conflicting focal element was redistributed to improve the accuracy of the evidence fusion.The results of R6105 AZLD diesel engine bench test show that the proposed method can greatly improve diagnostic accuracy and robustness.
作者
王承远
徐久军
严志军
WANG Cheng-yuan;XU Jiu-jun;YAN Zhi-jun(Marine Engineering College,Dalian Maritime University,Dalian 116026,China)
出处
《大连海事大学学报》
CAS
CSCD
北大核心
2018年第3期71-78,共8页
Journal of Dalian Maritime University
基金
国家自然科学基金资助项目(51509029)
辽宁省教育厅基金资助项目(L2015065)
中央高校基本科研业务费专项资金资助项目(3132015032)
关键词
柴油机
故障诊断
多属性群决策
证据融合
可信度权重
冲突焦元
diesel engine
fault diagnosis
multi-attributegroup decision making
evidence fusion
the weight of credibility
conflict element
