1、1 外文 文献 资料 At constant rotating speed, localized faults in rotating machine tend to result in periodic shocks and thus arouse periodic transients in the vibration signal. The transient feature analysis has always been a crucial problem for localized fault detection, and the key aim for transient fea
2、ture analysis is to identify the model and its parameters (frequency, damping ratio and time index) of the transient, and the time interval, i.e. period, between transients. Based on wavelet and correlation filtering, a technique incorporating transient modeling and parameter identification is propo
3、sed for rotating machine fault feature detection. With the proposed method, both parameters of a single transient and the period between transients can be identified from the vibration signal, and localized faults can be detected based on the parameters, especially the period. First, a simulation si
4、gnal is used to test the performance of the proposed method. Then the method is applied to the vibration signals of different types of bearings with localized faults in the outer race, the inner race and the rolling element, respectively, and all the results show that the period between transients,
5、representing the localized fault characteristic, is successfully detected. The method is also utilized in gearbox fault diagnosis and the effectiveness is verified through identifying the parameters of the transient model and the period. Moreover, it can be drawn that for bearing fault detection, th
6、e single-side wavelet model is more suitable than double-side one, while the double-side model for gearbox fault detection. This research proposed an effective method of localized fault detection for rotating machine fault diagnosis through transient modeling and parameter detection. Rotating machin
7、ery covers a broad range of mechanical equipments and plays an important role in many industrial applications, such as aircraft engines, transmission systems, power plants, etc. Most of the machinery was operated by means of bearings, gearboxes and other rotating components, which may develop faults
8、. These faults may cause the machine to break down, resulting in significant economic loss and even catastrophic personal casualties. The study of rotating machine fault diagnosis has thus attracted attention over the past decades. The transients or transient signals, characterized by a short period
9、 of time and span within a wide frequency range, contain important information about system 2 dynamics being studied. For example, the transients in the vibration signals generated on a gearbox usually correspond to the localized fault of bearing or gear teeth, such as flaking, crack, breakage and f
10、racture. For machine fault diagnostics, therefore, it is useful representing the characteristics of the machine health by analyzing the transients in the vibration signal. In order to extract or analyze the feature of the vibration signal, especially the transient, different techniques have been pro
11、posed for rotating machine diagnostics in the literature, such as empirical mode decomposition (EMD), independent component analysis (ICA), timefrequency representation (TFR), wavelet transform, and matching pursuit (MP), etc. EMD, as an adaptive decomposition technique proposed for nonlinear and no
12、nstationary by Huang, has been developed and widely applied in rotating machine fault diagnosis recently, such as gear fault diagnosis, rolling bearing fault diagnosis and rotor fault diagnosis. EMD decomposes the complicated signal into a set of complete and almost orthogonal components named intri
13、nsic mode function (IMF). However, it still has some shortcomings when it comes to calculating instantaneous frequency or in some cases it may reveal plausible characteristics due to the mode mixing, and this shortcoming makes it untenable in effective application in transient detection and analysis
14、. ICA is known as a powerful tool for blind source separation, which has also been introduced and applied to vibration analysis. ICA can be seen as an extension to principal component analysis (PCA) which aims at recovering the source signals from the set observed instantaneous linear mixtures witho
15、ut any a priori knowledge of the mixing system. Some researches applied it to extract the feature of vibration signals and detect transients. Though ICA is effective in blind separation of simulation signals, however, because of the multiple sources, intricate and varying transfer path in mechanical
16、 system and noise pollution, ICA is still in its infancy for effective application in mechanical fault diagnosis. TFR is the most frequently used method, through which the transient feature can be represented in the two-dimensional timefrequency plane. For example, the WignerVille distribution (WVD)
17、 and improved WignerVille distribution have been utilized to decompose vibration signals for fault diagnosis. It is no doubt that WVD has good concentration in the timefrequency plane. However, these methods are bilinear in nature, and there exist cross items in the decomposition 3 results that can
18、interfere in the feature interpretation. Even though some improved methods have been proposed, such as ChoiWillams distribution, cone-shaped distribution and so on, without exception, however, elimination of one shortcoming will always lead to the loss of other merits. For example, the reduction of
19、interference terms will bring the loss of timefrequency concentration. The wavelet transform, which is actually a kind of timefrequency analysis method, provides the signal information in the time and the frequency domains simultaneously through a series of convolution operations between the signal
20、being analyzed and the base wavelet under different scaling parameters. The application of the wavelet transform for mechanical fault diagnosis has been developed over the past decade . When detecting signal transients, the orthogonal wavelet transform and the continuous wavelet transform are usuall
21、y adopted. The orthogonal wavelet transform and multi- resolution analysis decompose the signal into orthogonal wavelet space. The continuous wavelet transform can provide a finer scale resolution than orthogonal wavelet transform. It is more suitable for extracting mechanical fault information. Mat
22、ching pursuit, a greedy algorithm that chooses at each iteration a waveform that is best adapted to approximate part of the signal, is effective in analyzing transient signals； however, the excessive computational cost limits its engineering applications. Correlation filtering, enlightened from matc
23、hing pursuit, was introduced based on Laplace wavelet and applied to identify the parameters of transient by calculating the maximal correlation value, including the modal parameters identification of a flutter for aerodynamic and structural testing , the natural frequency identification of a hydro-
24、 generator shaft and the wear fault diagnosis of the intake valve of an internal combustion engine , detection the position and the depth for rotor crack .Correlation filtering, enlightened from matching pursuit, was introduced based on Laplace wavelet and applied to identify the parameters of transient by calculating the maximal correlation value, including the modal parameters identification of a flutter for aerodynamic and structural testing , the natural frequency identification of a hydro- generator shaft and the wear fault
