| Acoustic emission (AE) technique plays an important role in the fault diagnosis as a new dynamic detecting way. The material, equipment or construction can be detected with AE detecting method, so AE technique has a bright prospect in applications. The processing of AE signals is vital important to AE technique, and the scholars in this field have been engaged in studying the rational and effective analysis method of AE signals, but the situation isn't optimistic.In this paper, studies are profoundly made on the actual detecting problems happened in the topic of "Acoustic emission monitoring of the slope stability of the permanent ship lock for the Three Gorges". This paper includes following information.Firstly, a new signal analysis method-wavelet analysis is discussed in this paper. As a time-frequency analysis method, the wavelet analysis can make the signal's resolving power high enough in the time domain and the frequency domain at the same time, by adjusting length and width of its windows. According to the characteristic of the wavelet analysis, a feasibility study is carried out on the basis of using the wavelet transform to process the actual collected signals with Daubechies wavelet, and the AE signals are divided.Secondly, the wavelet transform of AE signals collected in experiment is carried out with db6 wavelet and scale 3 on the basis of the signal's frequency analysis. The high frequency D, signal is AE characteristic signal. And by choosing the level's wavelet coefficient as its input, the wavelet-neural networks can be trained with the data of AE experiment, and the AE signals can be effectively recognized.Finally, the division of AE signals and the processing of large quantity data can be realized by using the wavelet transform and the wavelet-neural networks in AE monitoring of the slop of the permanent ship lock for the Three Gorges and by the reasonable selection of parameters.
|
PDF Full Text Request