基于小波分析和神经网络的桩身缺陷诊断

本文利用多分辨率分析提取桩基低应变动测信号功率谱的特征，作为BP神经网络的输入。通过对模型桩桩身缺陷信息的学习，预测诊断桩身缺陷类型。此外，利用小波变换的极值点诊断桩身的缺陷位置及利用小波包分析提取桩顶速度时域曲线的时、频域特征输入神经网络，通过模型桩的学习，预测诊断桩身缺损程度。工程应用实例表明，该法有一定的精度，是一种有效的桩身质量判别方法。

PILE DEFECT DIAGNOSIS BASED ON WAVELET ANALYSIS AND NEURAL NETWORKS

Using the characteristics,extracted by Wavelet Analysis from the power spectrum of dymamic stress-wave signals,as the input of BP neural network and using the prescient defect types and damage degree of piles as the desired output,the pile defects are diagnosed.After achieving the desired precision of the network training,defect type and damage degree of piles can be detected.Results of the wavelet decomposition of a residual obtained from the dynamic stress-wave signals are used to determine the defect situation in piles.Results of the engineering application indicate this approach can improve the reliability of pile integrity inspection and can become an assistant decision-making method for engineering practice.