局域均值分解分析的管道泄漏孔径识别及定位

Pipeline leak aperture classification and location based on local mean decomposition analysis

提出一种基于局域均值分解的RMS熵及时延估计的天然气管道泄漏孔径识别及定位方法。该方法首先对泄漏信号进行LMD分解,得到若干具有物理意义的PF分量并计算其有效值,进而结合信息熵的概念得出不同泄漏孔径的RMS熵,将不同孔径泄漏信号的多个RMS熵组成特征向量输入SVM进行识别。为提高互相关法定位精度,提出根据LMD分解结果的峭度特征进行重构再进行互相关获取时延信息,并结合泄漏信号的传播速度,实现泄漏点定位。实验结果表明该方法能够实现管道泄漏孔径有效识别及定位,且与基于EMD的RMS熵方法相比,识别效果更好,较直接相关法的定位精度明显提高。

Aiming at natural gas pipeline leak problem, a leak apertures recognition and location method based on RMS （root mean square） entropy and time delay estimation is presented by analyzing local mean decomposition （LMD） results. The leak signals are decomposed by LMD and several PF （product function） components with clearly physical meaning are obtained. The PF components RMS is calculated, which combines information entropy to acquire RMS entropy ofdifferent leak apertures. Several RMS entropy values are chosen as the feature vectors and input to the SVM to achieve the identification. In order to improve the location accuracy of cross-correlation, the kurtosis of PF components is analyzed, and leak signals are constructed based on principal PFs to improve time delay estimation accuracy. Combining the stress wave velocity, the leak location is caccomplished. Experimental results show the proposed methods with LMD analysis can effectively identify apertures and locate the leak, and the recognition result is better than the RMS entropy based on EMD. Location accuracy is obviously improved than the direct cross-correlation method.