基于声学特征的埋地管道堵塞故障的聚类识别方法  

A clustering method for underground drainage pipeline blockage identification based on acoustic features

作　　者：朱雪峰 冯早 黄国勇[1,2] 李洋 ZHU Xue-feng;FENG Zao;HUANG Guo-yong;LI Yang(Faculty of Information Engineering ＆ Automation, Kunming University of Science and Technology, Kunming 650500, China;Engineering Research Center for Mineral Pipeline Transportation,Kunming 650500, China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500 [2]云南省矿物管道输送工程技术研究中心,云南昆明650500

出　　处：《云南大学学报（自然科学版）》2018年第4期665-675,共11页Journal of Yunnan University(Natural Sciences Edition)

基　　金：国家自然科学基金(61563024;61663017);昆明理工大学引进人才科研启动基金(KKZ3201503015)

摘　　要：针对在实际工程应用中埋地排水管道不同程度的堵塞难以检测的问题,提出了一种基于声学特征的管道堵塞故障类型的聚类识别方法.首先对管道中采集的声学信号进行声压级变换,以增加不同故障类型之间的区分度,其次对声压级信号进行总体平均经验模态分解(Ensemble Empirical Mode Decomposition,EEMD),利用皮尔逊相关系数选取前4个IMF分量并提取中心频率和能量占比作为聚类特征,之后采用主成分分析法(Principal Component Analysis,PCA)对特征向量进行降维,最后通过FCM((Fuzzy C-means))和GK(GustafsonKessel),GG(Gath-Geva)3种聚类学习方法对特征向量进行堵塞故障的聚类和识别.实验结果表明,该方法能有效识别排水管道内不同程度的堵塞故障,具有一定的工程实用价值.Regarding the difficulties in detecting the partial blockage in underground drainage pipeline and the degree of blocking, a clustering recognition method based on acoustic features was proposed in this paper. Firstly, the sound pressure level was calculated from the acoustical pressure signals which were collected from the pipeline in order to increase the discrimination between blockage conditions. Then the comprehensive ensemble empirical mode decomposition （EEMD） was applied to the sound pressure level data , the first 4 IMF components were selected using the Pearson＇s correlation coefficient and the energy proportion was extracted as the clustering feature. Finally, the principle component analysis （PCA） to the feature vectors, the FCM （ Fuzzy C-means）, GK（ was adopted to proceed the dimensionality reduction on- Gustafson-Kessel ）, GG （Gath-Geva） algorithms were used to cluster feature vectors into classes and to further identify the blockage conditions.The experiment results have suggested that the proposed method is capable of identifying partial blockage conditions of drainage pipeline in different degrees, and presents a certain value for the engineering applications.

关 键 词：埋地排水管道 声学无损检测 堵塞故障诊断 总体平均经验模态分解 聚类分析 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]