基于CEEMD与BP＿AdaBoost的排水管道堵塞辨识  

Identification of Blockage in Drainage Pipeline Based on CEEMD and BP＿Ada Boost Algorithm

作　　者：闫菁[1,2] 冯早 吴建德 马军[3] YAN Jing;FENg Zao;WU Jiande;MA Jun(School of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China;Engineering Research Center for Mineral Pipeline Transportation of Yunnan Province,Kunming 650500,China;School of Mechanical and Electrical Engineering,Kunming University of Science and Technology,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 [3]昆明理工大学机电工程学院,云南昆明650500

出　　处：《电子科技》2018年第8期42-46,共5页Electronic Science and Technology

基　　金：国家自然科学基金(61563024;51169007);昆明理工大学引进人才科研启动基金(KKZ3201503015)

摘　　要：针对城市排水管道的堵塞故障程度难以区分的问题,文中提出一种基于完全经验模态分解(CEEMD)与BP_AdaBoost算法的排水管道堵塞故障识别方法。该方法使用声导波方法获取管道的声响应信号,对信号使用CEEMD算法进行分解,并对分解所得的本征模态函数(IMF)提取近似熵和能量占比建立特征集合。为了提高特征集合的区分度,使用距离可分性判据进行特征集合的降维,再利用BP_AdaBoost分类模型对降维后的特征集合进行分类识别。实验结果表明,该方法能有效的识别管道不同程度的堵塞。In view of the difficult to distinguish the blockage degree of the urban drainage pipeline,a method based on CEEMD and BP_AdaBoost algorithm for blockage recognition is proposed in this paper. The method uses acoustic guided wave method to obtain the pipeline response signal. Firstly,the signal is decomposed by CEEMD method,the characteristics of approximate entropy and energy ratio of the IMFs are extracted respectively,so the classification feature sets can be constructed. In order to improve the division of feature sets,the distance separability criterion is used to reduce the dimensionality of feature sets. Finally,the BP_AdaBoost classification model is used to classify the feature sets after dimension reduction. Experimental results show that this method can effectively identify different levels of pipe blockage faults.

关 键 词：管道堵塞 CEEMD 近似熵 能量占比 强分类器 

分 类 号：TN911[电子电信—通信与信息系统] TH165.3[电子电信—信息与通信工程]