欢迎来到天天文库
浏览记录
ID:18797670
大小:751.00 KB
页数:62页
时间:2018-09-24
《信息融合技术在飞行器故障诊断中的应用毕业论文》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、信息融合技术在飞行器故障诊断中的应用摘要近年来我国的航空事业发展迅速,飞行安全问题亟待解决,对飞行器准确进行故障诊断日显突出与重要。本文采用小波变换对飞机有关结构部件的健康监测信息进行特征提取,分别提取了小波系数的绝对值最大值、奇异值和标准差三个不同的特征,构建特征向量。飞行器故障诊断可利用的信息越多,诊断准确性就越高。只有充分利用有用的信息来对飞行器的故障进行诊断,才能提高故障诊断的精度和可靠性。因此故障诊断的有效性很大程度上与多信息融合效果密切相关。本文设计了基于广义回归神经网络和BP神经网络的故障诊断器,并运用其对提取的特征值进行了诊断研究。简要介绍了DS证据理论和加权融
2、合的基本原理,在此基础上,针对信息的特征层融合和决策层融合,分别设计了基于DS证据理论和加权融合的信息融合策略,并分别结合广义回归神经网络和BP神经网络故障诊断器对飞行器的健康状态进行了诊断研究。结果表明,采用决策层融合策略的健康诊断效果更好。关键词:故障诊断;特征提取;信息融合;证据理论;加权融合TheapplicationofinformationfusiontechnologyinaircraftfaultdiagnosisAbstractWiththerapiddevelopmentofChineseaviationindustry,healthdiagnosticte
3、chnologyofaircrafthasbecomemoreandmoreimportanttoensuretheflightsafety,soaccuratefaultdiagnosisonaircraftisbecomingmoreevidentandimportant.Thispaperusethewavelettransformtoaccomplishthefeatureextractionofstructuralcomponentsofaircrafthealthmonitoringinformation,Byusingthemethodofwavelettrans
4、form,theabsolutemaximum,singularvalue,andstandarddeviationofthewaveletcoefficientswereextractedrespectively,andthefeaturevectorswereconstructed.Inaircraftfaultdiagnosis,themoreinformationavailable,thehigherdiagnosticaccuracy.Onlytakingfulladvantageofusefulinformationtodiagnosethefaultontheai
5、rcraftcanimprovetheaccuracyandthereliability.Therefore,theeffectivenessoffaultdiagnosisislargelymorecloselyrelatedtoinformationfusion.WehavedesignedthefaultdiagnosisconnectorthatbasedongeneralizedregressionneuralnetworkandBPneuralnetworkinformationfusion,anduseitfordiagnosingtheextractionoff
6、eatures.BrieflyintroducedtheDSevidencetheoryandbasicprinciplesofweightedfusion,onthisbasis,accordingtothecharacteristiclayeranddecision-makinglayerofinformationfusion,wedesignedtheinformationintegrationstrategybasedonDStheoryofevidenceandweightedfusion,andwerecombinedwithgeneralregressionneu
7、ralnetworkandBPneuralnetworkfaultdiagnosisconnectoronthehealthstatusofaircraftdiagnosticstudies.Resultsshowthatthestrategyofdecision-makingfusioneffectisbetterforthehealthdiagnosis.Keywords:faultdiagnosis;featureextraction;informationfusion;evidenc
此文档下载收益归作者所有