基于模拟退火与LSSVM的轴承故障诊断.pdf

所需积分/C币:10 2012-09-30 10:13:33 322KB PDF
20
收藏 收藏
举报

基于模拟退火与LSSVM的轴承故障诊断.pdf
: LSSⅴM l21 52 20 0.00 0.00 20 20 21 0 1.12 20 0.21 20 0.89 20 0.00 20 0.00 20 0.89 20 0.21 20 1.12 20 0.21 20 0.89 20 Wilson wang 10 0.00 204200r/min 0.8 10 10 10 0.21 50.21 M BER-1OK) 2 10 0.21 0.21 NI PCI 0.21 4472. 20480Hz 10 (600.900.1200.21002400r/min)2 10 0.67 0.67 (1.22.3N 0e6689⑩ 3.3 「12 52 9- (2) 1.23) 416 4 16 1.2 (3) 2.3Nm) LSSVM Hilbert 3.4 LSSⅤM 3 0199,2010ChinafoademicJournalElectronicPublishingHouse.Allrightsreservedht,tp:/www.cni9rt 122 30 LSSⅤM tion monitoring of a gearbox us ing art ificial neural ( ar lificial neural netw ork, ANN network [J]. Mechanical Systems and Sig nal Cadaptive-netw ork- Process in g,2007(21):1746-1754 based fuzzy inference systeIll, ANFIS) [2] Wilson Wang. An intelligent system for mach inery condition monitoring [J]. IEEE Trans actio ns on Fuzzy BP SysteMiS,2008,10(1):110-122 %0 2008 SAHSSVM LSSⅴM ANn ANFIS 28(3):273-276 90.0 83.8 123 [4 Sugumaran V, Sabareesh G R, Ramachandran K I 9 90.0 82.5 fault diag nostics of ro ller bearing using kernel based 96.0 92.0 86.0 88.0 neighborhood score multi-class support vector machine[ J]. Expert Systems with A pplicat ions: An In LSSⅤM ternat ional Journal, 2008, 34(4): 3090-3098 SA-SSVM grid search GS) [5 Abbasiona s, Rafsanjani A. Rolling element bearings multifault classification based on t he w avelet 3 SA-LSSVM GS denoising and suppo rt vect or machine [J]. me anica Grid Search Systems and Signal Processing, 2007( 21): 2933 SA LSSVM 2945 ARLSSVM 1110.690.2590.028.720.4787.5 2410.280.4192.535.260.0390.0 2008.23 3358.200.1596.0240.120.7592.0 (3):357-360 []. 2008,34(2) 278-282. sequential forw ard selection [8 Lin S W, Tseng T Y, Chou S Y, et al.A simulated- SFS sequential backw ard annealing -b ased appr oach for simultaneous parameter election, SBS sequential optim iz ation and feature selectio n of back-pro pagation forw ard float ing selection, SFFS networks [J]. Expert Systems w ith A pplications, 4 2008,34(2):491-499. ,2008.,14(2):295-299 % LJ ,2008,23 SFS SBS SFFS (6):-6 LSsⅤM 90.0 78.880.082.570.0 J 208,26(5):37-2. 581.385.073.8 [12」 088.088.076.0 2008,44(7):112-117 19778 LSSⅤM 2007 28 4 E-mail:suit@163.com [1]rafieej,arvanif,harIf1a,etal.IntelligentcondjcPublishingHouse.Allrightsreservedhttp://www.cnki.net 206 Jo urnal of vibration, Measurement diag nosis Vol order to calcul at e the cross correlation function amplitude vector( CorV). A reference CorV is obtained by sm oothing curve fitting of the measured CorV of the dam aged structure in a least-squares sense. Then, t he co ntinuous wavelet transform is applied to the m easur ed Cor Vs and the ref erence CorV s res pectively. The modul of difference bet ween the wavelet coefficients of measured cor vs and reference cor vs are calculated as the in dex of dam age lo calization. The validity of the propo sed m et hod is verified by exper im ents on a co mpo site hon ey com b beam with prefab debo nd ing damage and a co mpo site laminate pl ate with prefab delamination damage Key words cross correlation funct ion amplitude vector w avelet tr ansform combined local iz ation composite structure damage detect Bearing Fault Diagnosis Using Simulated Annealing Algorithm and Least Squares support vector machines Sui Went ao", Lu Changhou, Wilson W ang, Zhang Dan (T he m oe Key laboratory of Hig h Efficiency and Clean Mechanical M anufacture, Shandong L niversity Jinan, 250061, China) (School of Mech anical Engineering, Shandong U niversity of Technology Zibo, 255049, China) Department of Mechanical Engineering, Lakehead U niversity, Thunder Bay, P7B 5E1, Canada) Abstract A fault diagnos is method based on the least square support vector m achines(LSSVM)and the simulated annealing algorit hm was propo sed. Better parameters of the regularizing variable X and the kernel width o were obtained by using the s im ul at ed annealing algorithm, and the sensitive subset of features w as deter mined simult aneously. To verify the effectiveness of the met d, roller bearings were tested un der four operating conditions, five different shaft speeds and two load levels, and 52 features were extract ed from the bearing vibration signals. The results show that the method has a higher accuracy of classifical io n for bearings fault than other nethods, and it is a promising approach to condilion monit orin of ro tating machinery Keywords parameter opt imization feature sel ectio n simulated annealing algorithm least s quares support vector machines fault diagnos is Vibration Isolation of Composite braces between Double-Cylindrical Shells yao Xiongliang, Ji Far College of Shipbuilding Engineer ing, Harbin Eng ineering U niv ersily Harbin, 150001, Chinla Vian Dejin (N aval academy of A rIm alllent Beijing, 100161, China Abstract Based on the Flugge shell theory and the Helm holtz equation, taking into acco unt the dynamic response of brace, the vibr ation equation coupled by the sound-fluid-structure is solved. The characteristics of vibration passing through stiffened double-cy lindrical shells w hich are coated w ith acoustic covering lay ers are studied by a test. The test dat a show s t hat the co upling effect is quite strong when the shells are connected by the braces which maini pas s the Vibration betw een the shells. The aying

...展开详情
试读 5P 基于模拟退火与LSSVM的轴承故障诊断.pdf
立即下载 低至0.43元/次 身份认证VIP会员低至7折
一个资源只可评论一次,评论内容不能少于5个字
您会向同学/朋友/同事推荐我们的CSDN下载吗?
谢谢参与!您的真实评价是我们改进的动力~
上传资源赚钱or赚积分
最新推荐
基于模拟退火与LSSVM的轴承故障诊断.pdf 10积分/C币 立即下载
1/5
基于模拟退火与LSSVM的轴承故障诊断.pdf第1页

试读结束, 可继续读1页

10积分/C币 立即下载