Wavelet Ridge Method-based Chaos Identification And Fault Data Analysis

Chaos phenomenon is widely existed in natural phenomenon and social phenomenon.In recent years, chaotic motion is found in a number of areas, accurate identification ofchaotic motion is a subject worthy of study. In the mechanical fault diagnosis field,rotating machinery fault signal is usually performed as nonlinear and non-stationary,showing complex motion characteristics. Using the theories of chaos to analysis faultsignals could be more accurate, and could get the essential characteristics of the faultreflected by the signals.Firstly, according to the low noise resistance shortcomings of extracting theinstantaneous frequency with the wavelet ridge line, this paper proposed a method basedon modulus maxima pitch to extract the wavelet ridge line. Specifically, the method in thispaper is adding a step of pitch judgmental process on the modulus maxima. By setting acertain coordinates threshold, only the coordinates within a threshold value of localmodulus maxima is extracted, so it could get rid of the random noise caused by localmaxima point so as to avoid noise interference, improved the disadvantage susceptible tonoise interference.Secondly, based on the theoretical basis of the wavelet ridge line and chaotic motion,a wavelet ridge identification of chaotic motion was advanced. Through an analysis ofseveral typical chaotic systems, from the phase diagram, power spectra, comparison interms of the wavelet ridge line graph shows that the wavelet ridge line can distinguishsystems of periodic motion, quasi-periodic motion and chaotic motion, and summarizedthe regularity of chaotic motions in the wavelet ridge line graph and the chaotic motion’sdetail features.Finally, analyzed the bearing fault vibration signals from American Case WesternReserve University bearing fault simulation test bench based on chaos theory. Firstly,calculate the chaotic characteristics, including the correlation, Kolmogorov entropy andLyapunov index, then extract the wavelet ridge line of the fault signal’s sensitive band,finally, analyze changes of the chaotic characteristics and the wavelet ridge line withdifferent fault degree. Through the comparison between the two, it is proved that the wavelet ridge line could accurately identify the fault type and fault degree.