基于二代小波变换和多级假设检验的输油管道缓泄漏检测方法

A detection method for slow leakage of oil pipelines based on second generation wavelet transform and multistage hypothesis test

针对输油管道缓泄漏准确判断与定位的难题,提出了一种基于二代小波变换和多级假设检验(MHT)的管道泄漏检测方法。首先,采用基于提升框架的二代小波变换对管道首、末端压力信号进行预处理,分析缓泄漏压力信号的小波分解效果,确定压力信号预处理的小波分解层数。其后,利用低频重构信息为原信号概貌并且能突出信号趋势的特点,将小波分解得到压力信号的各层低频重构信息用于泄漏判断与定位。再后,在给出基于MHT的泄漏判断方法和MHT参数的设置方法基础上,对压力信号的每一层低频重构信息分别进行MHT,并对每层结果进行加权平均,得到压力信号下降沿的准确位置后,利用负压波法实现了管道泄漏定位。实验结果表明,该方法可以准确检测和定位管道泄漏,为输油管道缓泄漏检测提供了一种新的手段。

When slow leakage occurs in oil pipelines, the trend of pressure signals is relatively slow. At present, there is no effective method to detect the slow leakage and locate the leakage position accurately. A leakage detection method was proposed based on sec- ond generation wavelet transform and multistage hypothesis test （MHT）. Pressure signals of pipeline ends were preprocessed by second generation wavelet transform within a lifting scheme in order to analyze wavelet decomposed results of slow leakage pressure signals and to determine layer numbers of wavelet decomposition. The low frequency reconstructed information can give a general picture of original signals and highlight the signal trend, therefore, the low frequency reconstructed information of pressure signals given by wavelet decomposition for each layer can be used to detect slow leakage and locate its position. Based on the leakage judg ment method using MHT and the setting method of MHT parameters, we processed the low frequency reconstruction information of pressure signals for each layer, respectively, with the MHT and performed the weighted average on data of each layer. When the ex- act location of failing edges of pressure signals was determined, the slow leakage of pipelines could be located with the negative pres sure wave method. The experiment results show that pipeline slow leakage can be detected and located accurately with this method, and it is an effective way to solve detection problems of slow leakage in oil pipelines.