欢迎来到天天文库
浏览记录
ID:32476238
大小:1.79 MB
页数:54页
时间:2019-02-07
《证据理论在机械设备故障诊断中的应用研究》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、华北电力夫学颈士学位论文摘臻在枫槭设备的故障诊断中,仅靠单一的诊龄方法难瞄全丽确切蛾确定故障类型及性瑷。本文铮对D.S涯据瑗谂在实甄应用时燕戳确定基本概率分龚薹滋数豹润题,分辑神经阐络技术和D.s证据理论各自的存在优缺点,提出并讨论了多传感器基于神经网络和D.s证据理论相结台的综合诊断模型,通过多个轴承诊断实例验证了该方法提高故障诊断精度的可行性和有效性。针对蛾厂设备状态监测的必要性,开发了辅麟维修系统的部分模块。实现了绘水泵设备臻憨及其维修警瑾信感瓣存馕彝查逶。关键词:信息融合,故障诊断,D-S证据理论,滚动轴承,辅助维修ABSTRA
2、C霉Inthefaultdiagnosisofmechanicalequipmant,itisoftendifficulttoidentifythefaulttypewiththesignalmeasured.Informationfusionsupplyanewapproachtodiagnosethemachinefaultswithhelpofvariesinformation.ThisarticlefocusesonthedisadvantageofB.SevidenceandthecharacteristicofBPneura
3、lnetwork.Amodelcombinedneuralnetworkanddatafusionmethodareusedtorecognizethefaultsoccurredinarollingelementbearingandallindustrialfaninpowerplant.Comparingwiththeapproachesthatonlyadoptonemethod,theaccuracyofdiagnosticresultsisobviouslyimproved,andthevibrationsignalsanal
4、ysisprovedthisconclusion+Inthepaperthecomputeraidedmaintenancesystemoffeedwaterpumpinpowerplantwasdiscussedandasoftwaresystemwasdeveloped.YanHongli(controltheoryandcontrolengineering)Directedbyprof.LiuYibingKEYWORDS:Informationfusion,Faultdiagnosis,Dempster.Shaferevidenc
5、etheory,Rollingelementbearing,CAM(computer-aidedmaintenance)华北电力火学硕士学位论文摘要在机械设备的故障诊断中,仅靠单一的诊断方法难以全面确切地确定故障类型及性质。本文针对D.s证据理论在实际应用时难以确定基本概率分配函数的问题,分析神经网络技术和D.s证据理论各自的存在优缺点,提出并讨论了多传感器基于神经网络和D.s证据理论相结合的综合诊断模型,通过多个轴承诊断实例验证了该方法提高故障诊断精度的可行性和有效性。针对电厂设备状态监测的必要性,开发了辅助维修系统的部分模块。实现了
6、给水泵设备信息及其维修管理信息的存储和查询。关键词:信息融合,故障诊断,D—s证据理论,滚动轴承,辅助维修ABSTRACTInthefaultdiagnosisofmechanicalequipment,itisoftendifficulttoidentifythefaulttypewiththesignalmeasured.Informationfusionsupplyanewapproachtodiagnosethemachinefaultswithhelpofvariesinformation.Thisarticlefocuses
7、onthedisadvantageofD—SevidenceandthecharacteristicofBPneuralnetwork.Amodelcombinedneuralnetworkanddatafusionmethodareusedtorecognizethefaultsoccurredinarollingelementbearingandanindustrialfaninpowerplant.Comparingwiththeapproachesthatonlyadoptonemethod,theaccuracyofdiagn
8、osticresultsisobviouslyimproved,andthevibrationsignalsanalysisprovedthisconclusion.Inthepaperthecompute
此文档下载收益归作者所有