基于小波包改进算法的管道泄漏特征信号提取

摘要

In pipeline leak detection, the characteristic signal marking the leakage is relatively weak, which brings great difficulty to the extraction of leakage characteristic signal.In this paper, we first study the propagation characteristic of acoustic emission waves in pipeline based on the finite element simulation software ANSYS and analyse the necessity of the research on leakage characteristic signal extraction algorithm;Then we put forward an improved wavelet packet decomposition and reconstruction algorithm for leakage signal extraction;Finally, we carry out the experiments in regard to extracting the leakage characteristic signals on the leak acoustic emission signals collected by sensor in lab based on the improved algorithm. Experimental results show that the improved algorithm can accurately extract the leakage characteristic signal and overcome the mixing effects brought by the up-sampling and down-sampling in traditional wavelet packet decomposition and reconstruction.All these create good conditions for subsequent positioning analysis.%在管道泄漏检测时,标志泄漏发生的特征信号相对较弱,这为泄漏特征信号的提取带来了极大困难. 首先基于有限元仿真软件ANSYS研究声发射波在管道上的传播特性,分析泄露特征信号提取算法研究的必要性;然后提出一种泄露特征信号提取的改进小波包分解与重构算法;最后基于改进后的算法对实验室里采集到的泄漏声发射信号进行泄露特征信号提取实验. 实验结果表明,改进算法可以准确提取出泄漏特征信号. 它克服传统小波包算法分解和重构中的上采样与下采样带来的混频,为后续的定位分析创造了良好的条件.