Analysis Of Bolt's Anchoring Quality Based On Wavelet Neural Network

By means of theoretical and experimental study, combing with numerical simulation method, signals of in-situ test and numerical simulation about dynamic test of bolts with are researched. Meanwhile, artificial neural network technique and wavelet analysis are adopted, the results of numerical simulation and actual signal analysis show that this approach can be used to classify the anchoring quality of bolt intelligently according the signal. Following conclusions are drawn:(1) Based on the stress wave theory, the propagating process of stress wave under impulse load in bolts are studied by numerical simulation method. The effect of the length of bolt's bare-segment and the intension of sand- plasm on the dynamic response signal of bolts are discussed in detail.(2) The present technical criterion classify the anchoring quality of bolts into I,II,III,IV grade based on the satiation proportion of sand-plasm . Through comparison of the simulated curves and the actual curves, the coherence between them was gained. Consequently, the models, the parameters of bolt anchoring system and the results of numerical simulation in this paper are proved to be correct and it will be helpful in increasing the accuracy of non-destructive test of bolt.(3) Wavelet transform has a good localization character both in time-domain and frequency-domain, and neural network has a good property of nonlinear mapping.