摘要
基于小波变换的时频局部化特性及人工神经网络的非线性映射特性,将小波变换和人工神经网络的优点结合起来,从基桩动测信号二进小波变换的频域中提取特征,最后将这些特征输入人工神经网络进行训练和分类,进而实现基桩缺陷的诊断。数值模拟试验显示了该方法的合理性,在此基础上进行了工程桩的现场试验研究,结果表明训练成功的神经网络可以作为智能分类器对基桩常见缺陷进行识别和诊断。
Based on the time-frequency locatization of wavelet transform and the nonlinear mapping of neural network, a method of dynamic testing signals combining with the advantage of wavelet analysis and neural network is presented. Some features are extracted from the frequency spectrum analysis at the various resolution of the dyadic wavelet transform. These features are taken the wavelet neural network as the input patterns for training and classifying. Then, it can be used to diagnose the faults of piles. The result of insitu test is in good agreement with numerical simulation and it show that this method can successfully be applied to the identification and diagnosis of plies faults as an intelligentized classifier.
出处
《水文地质工程地质》
CAS
CSCD
2004年第5期91-96,共6页
Hydrogeology & Engineering Geology
关键词
小波分析
神经网络
基桩检测
缺陷诊断
wavelet analysis
BP neural network
dynamic testing
signal analysis
fault diagnosis
