The Method Of Transient Modeling And Parameter Identification And Its Application In Rotating Machine Fault Diagnosis

Localized faults, such as spalling and crack, in rotating machine components including bearings and gears tend to result in shocks and thus arouse transient impulse responses in the vibration signal, and with the evolution of the fault, the characteristic waveform will be correspondingly changed. The transient detection and feature analysis has always been one of the most crucial and difficult problems for fault diagnosis of rotating machine. This research is supported financially by the Natural Science Foundation of China (No. 50905021) and Natural Science Foundation of Jiangsu Province (No. BK2010225). With the aim of fault diagnosis of rotating machinery, with the target of rolling bearings and gears, which are the typical rotating machine components, the paper proposes the method of transient modeling and parameter identification. The theoretical research and application research are studied in depth, respectively.Firstly, the characteristics of the vibration signals of the bearing and the gear, which has localized fault, were analyzed respectively. To ensure that the theoretical research works were all validated by the experiment analysis, two typical rotating machine components, bearings (cylinder roller bearing and deep groove ball bearing) and gears were tested under the localized fault condition and the vibration signals were collected.In consideration of the characteristics of transient impulse responses which are aroused by localized fault in rotating machine components, on the basis of the method of Laplace wavelet correlation filtering, the method of maximum-correlation-coefficient -criterion-based transient modeling and parameter identification is proposed through developing form the Laplace wavelet model to multiple parametric wavelet models. The transient models include single transient model and periodic multi-transient model, thus, not only parameters of the transient itself but also the interval between adjacent transients could be extracted. The effectiveness and the applicability of the method in fault characteristics detection is tested and verified through the application in the simulation signal analysis and the vibration signal of both the bearing and the gear.According to the problem of fault diagnosis under variable working condition, the approach of least-square-method-based transient modeling and parameter identification was proposed, in which the double-side asymmetric parametric model is built to diagnose fault of different components, and transients are iteratively extracted and eventually represented by Wigner-Ville Distribution not only to ensure time-frequency concentration but also to avoid interference of cross-item between adjacent transients.In this paper, through the research on the transient impulse responses modeling and parameter identification, it is confirmed that both methods were effective in feature detection for localized fault in rotating machine components, which is of certain theoretical and practical value for fault diagnosis of rotating machinery.