The Preliminary Research Of Time-frequency Analysis On Bearing Fault Diagnosis Of Gearbox

The elementary processes of fault diagnosis on mechanical equipment are three steps.The first step is to obtain the diagnostic information.The second step is to extract fault features.The third step is pattern recognition and fault diagnosis.The key of these diagnostic steps is to extract fault features from the dynamic signals.The method on extracting fault feature is signal processing ordinarily.Viberation signals of equipment in engineering have a lot of nonstationary dynamic signals.Dynamic signals obtained from engineering equipment are relatively and locally stationary,and they are absolutely and comprehensively nonstationary.Time-frequence analysis is the powerful tool for research on nonstationary signal.So this dissertation focus on the time-frequence analysis methods on fault diagnosis for rolling bearing in gearbox.This dissertation have done a contrastive research on several primary time-frequence methods.This dissertation have summed up the advantages and defectes of these methods by the application of several time-frequence methods to surface damage diagnosis of rolling bearing outer race in gearbox.Aiming at the problems in fault diagnosis based on viberation signal analysis this dissertation explores a suit of fault diagnosis method based on time-frequence analysis for surface damage diagnosis of rolling bearing outer race in gearbox.This dissertation presents the diagnostic bases and a new method based on Hilbert-huang transform for surface damage diagnosis of rolling bearing outer race in gearbox.The strucure of this dissertation are listed as follows:Chapter one:The origin,purpose and significance of this research subject are introduced.The present situation of condition monitoring and faults diagnosis on mechanical equipment is briefly summed up.The significances of signal analysis and time-frequence analysis are explained.Time-frequence methods on nonstationary signal are reviewed.At last the research content is briefly introduced.Chapter two:The failure modes of rolling bearing,the viberation signal characteres of rolling bearing fault in gearbox,the generation and character of shock and viberation on rolling bearing fault in gearbox,the basic modes of shock and viberation in gearbox are separately discussed.The experimental and sampling system of rolling bearing fault simulation in gearbox is introduced.The viberation signal of bearing damage on outer race in gearbox analyses by amplitude domain statistic and Fast Fourier Transform.But the results show that these two methods are obvious restricted on the fault characters extraction of the vibration signal. Chapter three:Preparation mathematics and basic theory of time-frequence theoryare introduced.Chapter four:Short time Fourier transform,wavelet transform,and Winger-Ville distribution are expatiated in theoretically.Through applying several major time-frequency analysis tools in the gearbox bearing fault diagnosis,the limitations of these time-frequency analysis method are analyzed and summarized.The common questiones of time-frequency analysis are summed up.Chapter five:The theory of instantaneous frequency is expatiated.The commoncognition of instantaneous frequency is summed up.Chapter six:Hilbert-Huang transform based on the theory of instantaneous frequency is introduced.Main basises,basic conceptions,basic methods,analysis procedures and correlative improvement algorithms are discussed in detail.The innovation of Hilbert-Huang transform is explained.Hilbert-Huang transform of improvement algorithms is applied to the diagnosis of rolling bearing outer race damage.The results show that Hilbert marginal spectrum can extract effectively the fault characters on surface damage of rolling bearing outer race in gearbox.Chapter seven:The work and innovation of this dissertation are summarized.The problems that are needed to improve and solve on Hilbert-huang transform are pointed out.