Low Cycle Stationary Signal Research And Application

The research methods of cyclostationary theory are varied and wide range of applications. The TA (Time Aliasing) algorithm which were studied by scholars less and less application had been presented by Professor Smith. This paper does a deeper theoretical research and its application by TA algorithm.This thesis focuses on the basic theory of the low-order cyclostationary signals, and using the first and second order cyclostationary as the basic tool. The study of a focus is the second-order cyclic statistics and an example simulation. Then methods of spectral are studied, and we especially focus on the time aliasing methods of spectral estimation and give an example. The analysis of second-order cyclostationary is applied to the vibration signals of the Rolling Element Bearing, and then compares between the normal and the fault conditions of the vibration signals.First, this thesis discusses the cyclostationary signal’s many characters, such as concept, feature, the scope of using and the development currently. A detailed description is the second-order cyclic statistics which is using in rolling element bearing fault diagnosis. Second, this thesis introduced the basic theory of the low-order cyclostationary signal processing which is combined with the simulation analysis of the AM signal’s amplitude modulation and multi-modulation source AM that focus on the second-order cyclic statistics. Third, this thesis introduces a cyclic frequency with the help of classical power spectrum estimation, and a detailed study of the average periodogram and smoothed periodogram of the cyclic spectrum estimation, and then gives a FM signal that the power spectrum estimation based on the periodogram by simulation. This thesis restructures the non-periodic signal and makes it become a periodic signal in order to carry out the cyclic spectrum estimation with the help of time aliasing (TA) and the introduction of sidelobe suppression Ssinc (Smooth sine) window function together, and then gives a FSK signal that the spectrum estimation based on the cyclic spectrum of TA method by simulation.Finally, this thesis gives a brief introduction about the rolling element bearing and the fault characteristic frequency formula which to calculate the inner and outer race and other parts of the fault characteristic frequency, and combined with the cyclostationary model of the inner and outer race, then leads to additive mixed cyclostationary model. Thesis makes deep groove ball bearing (SKF6205-2RS) for the object of study, and made a detailed simulation between normal and fault conditions demodulation that also using real cepstrum conductes a simulation analysis, and then points out that the fault is present reasons.