Vibration Signal Analysis Based On Time-wavelet Energy Spertrum And Cross-Wavelet Transform

Fault diagnostics is useful for ensuring the safe running of rotating machines and vibration signal analysis has been widely used for fault diagnostics. Various kinds of factors, such as the change of the environment and the faults from the machine itself, often make the vibration signal of the running machine contain non-stationary components. So it is important to analyze the non-stationary signals. Among many signal processing methods, the most common tool utilized in real-signal applications is the Fourier transform(FT) which decomposes a given signal into its frequency components. Unfortunately, this technique requires that a signal to be examined is stationary, i.e. without time-evolution of the frequency content. FT-based methods are not suitable for non-stationary signal analysis, with an intermittent and changing frequency pattern. The limitation of the Fourier analysis can be partly resolved by using a short-time Fourier transform(STFT). One critical limitation of the STFT appears when windowing the signal mainly due to the violation of the uncertainty principle. More precisely, if the window is too narrow, the frequency resolution will be poor, whereas if the window is too wide, the time localization will be less precise. So STFT is not suitable for analyzing signals involving different scales or range of frequencies. Compared with the STFT, the wavelet transform has many distinct advantages for vibration signal analysis. Time-wavelet Spectrum analysis and Cross-wavelet Transform based on the wavelet transform theory are brought out. The main aim of the present dissertation is to extract the feature information of the gear and bearing fault by using time-wavelet spectrum analysis and analyze the vibration signal of hydraulic turbine by using cross-wavelet transform.The impulses in vibration signals and their spectral features are important in diagnosing localized damage of gear and bearing. A new method, so called time-wavelet energy spectrum which is based on the theory of wavelet transform, is proposed for gear and bearing fault diagnosis. It can extract the feature of impulses in both time domain and frequency domain. It is applied to analyze the vibration signals of a gearbox under worn and broken statuses and bearing with outer ring fault, inner ring fault and ball fault. Envelope-demodulation analysis and Hilbert-Huang tranform are also used to analyze those signals. The result shows that the time-wavelet energy spectrum is more effective in extracting the impulse features produced by gear and bearing damage than other methods of signal processing.Hydaulic pressure fluctuation is one of major factors influencing the vibration of hydraulic turbines. Correlation analysis of the hydaulic pressure fluctuation and the turbine vibration is important to reveal the hydaulic pressure fluctuation induced vibration. Cross-wavelet transform based on the wavelet theory is used to analyze the two signals'correlation in time-frequency domain. In this dissertation, cross-wavelet transform is used to analyze the vibration at water turbine guide bearing and the hydaulic pressure fluctuation at draft tube, spiral case and headcover in time-frequency joint domain. The time-frequency correlation between the vibration and the hydaulic pressure fluctuation are extracted. The traditional cross-correlation analysis is also used to analyze those signals. The result shows that cross-wavelet transform can extract not only the time information but also the frequency information of the correlation of two signals, which the traditional cross-correlation analysis cannot.