Font Size: a A A

Intelligent Diagnosis Method Based On Full Information And Applied Research

Posted on:2007-10-19Degree:MasterType:Thesis
Country:ChinaCandidate:Y N PanFull Text:PDF
GTID:2192360185472139Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
The Vector-spectrum is a joint name of vector spectrum and its patulous analysis methods, they fuse the double channels' information and remain the meanings of corresponding traditional methods.Aiming at the problem of the information's deficiency in the rotating machine fault diagnosis,the Vector-spectrum was applied to fault diagnosis and an intelligent diagnosis method based on full information was put forward.The main content is as follows:1.This paper dissertated the vector spectrum and introduced the vector power spectrum and vector Wigner-Ville Distribution weightily,their definitions, algorithms were supplied,and experiment research was done.The result showed that Vector-spectrum exceeds the corresponding traditional ones.2.The Radial Basis Function Neural network(RBFN) has excellent classified and extensible ability.This paper introduced its model, trainning process and capability, combining Vector-spectrum and RBFN, a new fault diagnosis approach was proposed, this approach is that vector power spectrum and vector Wigner-Ville Distribution is used as eigenvectors respectively, RBFN as a classifier. The proposed approach has been successfully applied to the recognition of three typical faults, i.e. rotor crack, rotor misalignment and oil whirl. The experiment result showed that this proposed approach is effective.3.Fuzzy clustering analysis is a simple, intuitionistic and effective recognition method,Fuzzy C-means theory is the most integrated one among many algorithms. This paper introduced its algorithm and presented an intelligent diagnosis method of Vector-spectrum-FCM. vector power spectrum and vector Wigner-Ville Distribution as eigenvectors respectively, research was done,the result showed that this approach is feasible and effective.4.The support vector machine is a new universal learning method based on statistic learning theory,it can effectively resolve the problem of small sample in practice. This paper researched the SVM intelligent diagnosis method based on vector power spectrum and applied it to rotating machine fault diagnosis,the result showed that this proposed approach is effective,especially on the condition of small sample.5.Artificial immune system is a new intelligent method arising recent years,it has powerful capability of dealing with information and resolving problem, wide research foreground and application potential.This paper briefly introduced artificial immune...
Keywords/Search Tags:Vector-spectrum, Radial Basis Function Neural network(RBFN), Fuzzy C-means (FCM), Support Vector Machine(SVM), Artificial Immune System(AIS)
PDF Full Text Request
Related items