The Mechanical Fault Diagnosis Based On Fractional Order Time-frequency Analysis Method Study

This dissertation was supported by the National Natural Science Foundation of China(51261024,51075372,51265039,50775208),,National key R & D program(No.2016YFF0203000),the State Key Laboratory of Mechanical Transmissions of Chongqing University(SKLMT-KFKT-201514)and the Science and Technology Project of Jiangxi Provincial Education Department(No.GJJ12405,GJJ150699).Using fractional Fourier transform has the unique characteristics of the traditional signal processing method based on Fourier transform to the fractional Fourier transform domain,puts forward some effective fractional non-stationary signal processing method,Fractional wavelet transform?Fractional S transform?Fractional generalized S transform?Fractional order local mean decomposition and so on.And put them in the application in fault diagnosis of rotating machine,made a good innovative achievement.This article main research content includes the following several aspects:The first chapter,This paper expounds the research background and research significance,time-frequency analysis method is discussed,and the status quo at home and abroad.Here,mainly from the time-frequency analysis based on kernel function,based on the time-frequency analysis of signal decomposition,parametric time-frequency analysis of three aspects.At the same time,also discusses the existing time-frequency analysis methods in the application status of mechanical fault diagnosis,on this basis,put forward in this paper,the research content and innovation of existing content.The second chapter,Discusses the definition and characters of fractional wavelet transform algorithm,and on this basis,put forward the mechanical fault diagnosis based on fractional wavelet transform method,and compared with the traditional wavelet transform.The method combines wavelet transform and fractional Fourier transform their respective advantages,abandon their disadvantages at the same time,joined a adjustable variables in the analysis of the fractional order,making the method on the analysis of non-stationary signal is more flexible.Simulation studies show that fractional wavelet transform is superior to traditional wavelet transform,have an advantage in terms of the signal filtering,and multiresolution analysis of non-stationary signal.Finally,the proposed method is applied to the rotor ground fault,the experimental results verify the effectiveness of the proposed method.The third chapter,Discusses the definition,standard S transform algorithm and properties,this chapter combines standard S transform and fractional Fourier transform,fractional S transform structure,gives the definition of fractional S transform formula and algorithm,and compared with standard S transformation.This method will S transform and fractional Fourier transform their respective advantages,can according to the characteristics of the signal to select appropriate fractional order,to do the analysis of the signal accurately.The simulation results show that the fractional order S transformation has obvious advantages,can not only reflect the frequency of the signal structure,and got the S transform is higher than the standard time and frequency resolution.Finally,the proposed method is applied to fault diagnosis of rolling bearing,the experimental results further illustrate the effectiveness of the proposed method.The fourth chapter,This paper discusses the definition of generalized S transform,the algorithm and properties,similar to the definition of fractional S transformation,constructs the fractional generalized S transform,gives the definition and algorithm of fractional generalized S transform,and compared with standard S transform,the generalized S transform analysis.The method fully absorb the generalized S transform and fractional Fourier transform their respective advantages,and compared the advantage of the two fusion,makes the fractional generalized S transform can be flexible for the analysis.The simulation results show that the fractional order generalized S transform has obvious advantages,can not only reflect the frequency of the signal structure,and obtained the S transformation,generalized S transform is higher than the standard time and frequency resolution.Finally,the proposed method is applied to fault diagnosis of rolling bearing,the experimental results further verified the effectiveness of the fractional generalized S transform.The fifth chapter,Combined with local mean decomposition and their respective advantages of fractional Fourier transform,constructs the fractional order local mean decomposition,the definition and algorithm are given,based on this,puts forward the Wigner distribution based on fractional LMD mechanical fault diagnosis method,at the same time,the proposed methods have been compared with traditional Wigner distribution,the simulation and experimental results show that the fractional order local mean decomposition is superior to the traditional Wigner distribution,can effectively eliminate the cross terms,has the very high time-frequency resolution,low can effectively reflect the fault characteristic frequency.The sixth chapter,work of this paper has carried on the comprehensive summary,and further work were discussed.