Rotating Mechanical Non-stationary Signal Characteristics Analyzer Development

Rotating machinery is widely employed in industry as the key mechanical equipment.To ensure the running safety and reliability of rotating machinery is very significant to both enterprises and our social economic growth.The fault feature extraction of rotating machinery is one of the most important processes in the fault diagnosis of rotating machinery.This paper based on Natural Science Foundation projects of Yunnan province "Research on non-stationary vibration signal feature extraction and analysis technology of rotating machinery(NO.2004E0011Q)",which aims at the research on the non-stationary vibration signal feature extraction technology of rotating machinery.By summarizing the rotating machinery feature extraction methods,the typical faults of rotating machinery and their corresponding vibration characteristics and non-stationary signal processing technologies are introduced.The insufficiency of current test & measurement instruments on non-stationary vibration signal analysis,for example,the run-up or coast down periods of rotating machinery,is also discussed.It inspires the application of time-frequency analysis(TFA)technology in Rotating Machinery characteristics analysis fields.By a deep study and advanced research on Wigner-Ville distribution(WVD)and its crossing-item interferences and integrating with the time-varying filtering technology based on the Gabor transform,"Removal of crossing-items of WVD based on Gabor reconstruction" method is proposed.This method retains the the advantage of Wigner-Ville distribution(WVD)with the best time-frequency resolution,and attempts to solve the crossing-item problem in WVD.Both simulations and actual tests obtain perfect results by using the proposed method.With a deep study on extraction technology of angle domain resampling signals, the time-varying filtering technology based on Gabor transform,time-frequency analysis technology and the second generation wavelet de-noising technology,the method of fault characteristics extraction of angle domain signals based on Wigner-Ville Distribution is proposed.With the merit of the angular domain signal has phase information,it can accurately locate the fault source from the analysis results in the angular domain by TFA.Fault analysis of oil whirl and oil whip by Waterfall based on Short-Time Fourier Transform can not only extract fault characteristics but also clearly expose the typical "inertia effect" of oil whip.The vibration caused by oil whirl and oil whip is non-stationary signals,so it's easier to extract various rotors' fault characteristics by time-frequency analysis methods than conventional stationary signal analysis methods.A Virtual Instrument(Ⅵ),Signal Characteristics Analyzer of Rotating Machinery(SCARM),has been programmed on LabVIEW platform.It provides powerful functions on non-stationary signal analysis and enables the specific abilities for analyzing non-stationary vibration of rotating machinery.It is a supplyment to current used measurement and test instruments.Finally,simulation experiments and testing evaluations are performed with the SCARM.The results verify the effectiveness ofⅥin rotating machinery vibration characteristics analysis.