管道传声特性及典型破坏声的识别

Pipeline acoustic propagation features and recognition of typical damage sounds

为了保证管道输送的正常运行,防止人为打孔盗油事件发生,对管道进行主动防护与快速准确地判定事发地点具有重要意义。研究了管道的导波特性,分析了管道人为破坏的典型声谱特征,建立了能对管道破坏方式进行分类的混合模型。研究结果表明,声波通过管道中介质以平面波传播且衰减较小;典型破坏声可在管道内远距离传播,隐马尔可夫模型(HMM)与人工神经网络(ANN)算法相结合能有效地提取和识别其声谱特征,从而为管道运输防盗监测提供一种新的途径。

In order to protect pipeline transportation and prevent manmade damages of shemaking hole to steal oil, it is very important to carry out such researches as active protecting and accurate positioning. By analyzing the features of pipdines including wave-guiding and typical acoustic spectrum of manmade damages, a model composed of characters of acoustic wave was established. The results show that acoustic wave in pipeline propagates as plane wave with little falloff. A typical damage acoustic wave is able to propagate for a long distance. The typical acoustic spectrum can be picked up and recognized by the algorithms of hidden Markov model（HMM） and artifical neural network（ANN）, which provides a new method for the prevention of theft in pipeline transportation system.