基于云模型和证据理论的多源信息融合方法研究与试验  

A multi-source information fusion method based on cloud theory and evidence theory and its application

作　　者：冯春生[1] 谭继文[1] 孙显彬[1] 

机构地区：[1]青岛理工大学,山东青岛266033

出　　处：《制造技术与机床》2015年第11期87-93,共7页Manufacturing Technology & Machine Tool

基　　金：国家自然科学基金项目(51075220);山东省高等学校科技计划项目(J13LB11);高等学校博士学科点专项科研基金(20123721110001);青岛市科技计划基础研究项目(12-1-4-4-(3)-JCH)

摘　　要：针对复杂机械设备故障诊断信息具有模糊性、不确定性、不完全性等特点,在分析云模型理论的基础上,提出了一种基于云模型和D-S证据理论的多源信息融合与故障模式识别新方法,研究了相关参数的计算和确定方法,给出了用云模型方法对多源信息进行统一建模、并根据获取的不确定度对故障特征值进行二次筛选、通过计算证据的不确定度系数和整体支持度系数对融合证据权重进行再分配、利用证据理论得出融合识别结果的理论依据及实现步骤。最后,通过数控机床滚动轴承故障诊断试验验证了该方法的可行性和故障识别的准确性。According to the fault diagnosis of complex mechanical equipment information is fuzzy, uncertainty, incompleteness and other characteristics, puts forward a new method of multi - source information fusion based on cloud model and D -S evidence theory, studies the calculation and determination method of relevant parameters. Proposes to use the cloud model method of multi - source information unified mod- eling, and according to the degree of uncertainty to screen the fault characteristic value second, redis- tributes the fusion weights of evidence by calculating the uncertainty coefficient and overall support de- gree coefficient of evidence. The recognition results obtain uses evidence theory. Finally, the NC machine tool fault diagnosis of rolling bearing test verifies the feasibility and accuracy of the method of fault identification.

关 键 词：云模型 证据理论 多源信息融合 数控机床 故障诊断 

分 类 号：TH17[机械工程—机械制造及自动化]