Abstract: When acoustic method was used to leak detection, external interference (operation of compressor and valve, pipeline knocking, etc.) should be distinguished with acoustic leak signals to improve the accuracy of leak detection and reduce false alarm. The technologies of extracting wave characteristics were summarized. The acoustic leak signals and interfering signals were gathered by a high pressure and long distance leak test loop to analyze characteristics of time domain, frequency domain and time-frequency domain. The main characteristics of time domain were mean value, root mean square value, kurtosis, skewness and correlation function, etc. Frequency spectrum analysis and power spectrum density were adopted to analyze the feature of frequency domain, while short time Fourier transform was used to process time-frequency analysis. The results show that the external interference can be removed effectively by the characteristics of time domain, frequency domain and time-frequency domain.

Key words:

Leak detection,    ,  Acoustic method,     ,  Signal processing,    ,  External interference,    ,  Characteristic extractionmso-ascii-font-family: Calibri, mso-ascii-theme-font: minor-latin, mso-fareast-theme-font: minor-fareast, mso-hansi-font-family: Calibri, mso-hansi-theme-font: minor-latin, mso-fareast-font-family: 宋体">

摘要： 在总结分析音波泄漏检测特征量提取方法的基础上，利用高压泄漏检测实验装置，分别采集泄漏及各种干扰音波信号，对其进行时域、频域及时频域结合的综合分析，并提取时域、频域及时频域结合的特征量。时域特征量主要有均值、均方根值、峰度、偏度、相关函数及协方差函数，频域特征量主要信号的频谱分析和功率谱密度估计分析，时频联合分析采用短时傅里叶变换来实现。分析结果显示，整体峰度可以用来区分泄漏及干扰信号，信号均值和相关函数、协方差函数可以作为敲击信号的特征，信号波形和均方根值可将压缩机信号区分出来，信号偏度、相关函数及协方差函数可以作为阀门动作的特征量，并且结合频域及时频域特征，可以排除各种外界干扰影响，从而降低音波泄漏检测的误报率。

关键词: 泄漏检测,    , 音波法,    , 信号处理,    , 外部干扰,    , 特征提取