Analysis Of Mechanical Vibration Simulation Signals Based On Cyclic Bispectrum

Feature extraction is a very important signal processing problem in condition monitoring and fault diagnosis of mechanical equipment.Since the 80s of last century,in order to meet the need for accurate diagnosis of mechanical failure,non-linear,non-Gaussian,non-stationary signal processing technology more and more in the field of mechanical fault diagnosis more and more attention.Because the high-order statistics have the properties of suppressing Gaussian noise,this paper analyzes the mechanical vibration simulation signals based on cyclic bispectrum using the combination of theory,simulation and experiment.Firstly,this paper analyzes the stationary random signal by using the high-order statistical analysis method.The high-order statistics can suppress the Gaussian noise and extract the characteristic information of non-stationary,non-linear and non-Gaussian fault signals,which is of great significance to the fault diagnosis.The higher the order,the greater the computational complexity,the third-order statistics can not only effectively suppress the Gaussian noise,but also extract the nonlinear signal characteristic information,but also the simplest calculation,so this paper focuses on the analysis of bistatic spectrum.Second,the study of the cycle of stationary signal definition and analysis methods.(1)The definition,properties and characteristics of the first-order and second-order cyclic statistics are studied.By using the cyclic autocorrelation function to separate the carrier frequency information and the modulation frequency information in the cyclic frequency domain,it can be more easily in the low frequency To extract the faulty modulation frequency information;the use of periodic frequency and frequency characteristics of the relevant characteristics between the slice can be used to extract useful information,and then analyze the characteristics of frequency information.(2)mechanical vibration signal can be amplitude modulation,frequency modulation signal,multi-carrier frequency modulation signal,multi-modulation source amplitude modulation signal and multi-carrier amplitude modulation signal,in order to understand the mechanical equipment cycle smooth performance,study these complex modulation signal demodulation Method,it is found that the cyclic correlation demodulation analysis can effectively separate the source signal,the mixed source and the carrier modulation identification information,and has strong noise suppression characteristic.In this paper,the mechanical vibration signal is simulated and experimentally studied.It is found that the computational complexity of bispectrum analysis is very large,and the autocorrelation slice spectrum can not effectively suppress Gaussian noise.Combining the characteristics of the two,this paper introduces the method of cyclic autocorrelation function To the cyclic bispectrum analysis,it is proposed to extract the fault characteristics by using the cyclic bispectrum carrier frequency slicing method,which can directly express the result of spectral analysis and extract the fault characteristic frequency information effectively.At the same time,the wavelet signal denoising of the fault signal is carried out and the bispectrum analysis is carried out.The experimental results show that the interference in the cyclic bispectral frequency slices of the wavelet denoising is obviously reduced and the characteristic frequency information is more prominent The...