基于神经网络的管道泄漏声波信号特征识别

Recognizing characteristics of pipeline leakage acoustic signals based on neural network

文中采用小波包和神经网络结合起来用于声波信号诊断的方法来提高泄漏检测的准确率。该方法首先对声波信号进行小波包分解,将各频带内的分解系数重构,得到在每个分解节点上构成的新时间序列的参数,将这些参数通过BP神经网络进行智能识别,来区分故障原因。文章最后对现场实验数据及其信号分析结果进行了研究,实验结果表明,该方法可以有效地提高管道泄漏信号识别的准确度。

This paper presents that wavelet packages and neural network are combined to detect acoustic signals to increase the accuracy of leakage detection. The signal is decomposed with wavelet package firstly, then the decomposed coefficient in each frequency band is reconstructed. As a result, the parameters of the new sequences reconstructed on every decomposed node are acquired, and then these parameters are input to BP neural network to recognize the fault reason intelligently. At the end of the paper,field experiment data and its analyzed results are studied. The results are provided to show that the proposed method can increase the accuracy efficiently.