基于地震波传感器阵列的管道地面标记系统

地面标记系统是管道内检测的重要组成部分,可以显著减小内检测器在石油管道缺陷检测中产生的里程误差,提高对管道缺陷的定位精度.针对目前管道越埋越深的问题,本文提出了一种基于地震波传感器阵列的地面标记器技术.当检测器在管道内运动时,会与管道内壁的焊缝产生摩擦,产生地震波信号.地面标记器通过高灵敏度的地震波传感器可以采集到这种地震波信号.由于不同目标产生地震波信号的频率、能量等特征不同,本文使用基于经验模态分解（empirical mode decomposition,EMD）的方法对信号进行特征提取.首先将信号用EMD分解为几个固有模态函数分量（intrinsic mode function,IMF）;然后计算各个IMF分量归一化的能量分布,将IMF能量分布作为信号的特征向量;最后使用基于支持向量机的神经网络来对地震波信号进行模式识别,用来识别有效信号和干扰信号.通过模拟实验,识别正确率达到了93%,验证了本文提出的地面标记系统的有效性.

Above-Ground Marker System of Pipeline Internal Inspection Instrument Based on Geophone Array

The above-ground marker（AGM） system is an important part of pipeline internal inspection instrument,which can significantly reduce the mileage error in petroleum pipeline default inspection by internal inspector and improve the location precision for pipeline defaults.A geophone array based AGM was proposed in this paper,aiming at solving the problem resulting from increasing depth of the pipeline.When the internal inspector moved inside pipeline,it would strike the welds on the inner-wall of the pipeline and generate seismic signals,which would be gathered by AGM with high sensitivity.As the features such as frequency and energy of the seismic signals varied with the targets,empirical mode decomposition（EMD） was used for feature extraction of the signals.Firstly,the signals were decomposed into several intrinsic mode functions（IMFs）.Secondly,the normalized energy distribution of IMFs were computed and used as feature vectors of the signals.Finally,the artificial neural network based on support vector machine（SVM） was applied to pattern recognition of the seismic signals,in order to identify the effective signals and noise signals.The proposed AGM is proved to be effective by the simulated experiments,in which recognition accuracy of 93%was achieved.