基于G-蚁群聚类的滚轴故障诊断方法  

Roller fault diagnosis method based on G-ant colony clustering

作　　者：王文瑾 黄细霞[1] 宋虎 WANG Wenjin;HUANG Xixia;SONG Hu(Key Laboratory of Marine Technology and Control Engineering of Ministry of Communications（Shanghai Maritime University）,Shanghai 201306,China)

机构地区：[1]航运技术与控制工程交通部重点实验室(上海海事大学),上海201306

出　　处：《计算机应用》2018年第A01期24-27,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(51209134)

摘　　要：滚动轴承故障是电机设备运行中常见的故障之一,常见的典型故障有滚动轴承外环故障、滚动轴承内环故障和滚动轴承滚子故障等。由于故障点出现的位置距离比较近,易对滚动轴承作出错误的故障诊断,为此提出一种基于遗传变异蚁群聚类的滚动轴承故障诊断方法。首先,应用小波包函数对滚动轴承的3种故障状态数据及滚轴正常状态数据进行多层分解,构造状态特征向量;然后,将4种状态特征向量分别代入基本蚁群的聚类算法和基于遗传变异蚁群的聚类算法,得到滚动轴承的故障分类模型;最后,选取10组故障数据作为验证样本对两种聚类模型进行验证。实验结果表明,基于遗传变异蚁群的聚类算法的分类速度更快,比基于基本蚁群(ACO)的聚类算法的诊断方法对滚轴故障类型的识别率更高。Rolling bearing failure is one of the common faults in the operation of motor equipment. Common fanhs include rolling bearing outer ring failure, rolling bearing inner ring failure, and rolling bearing roller failure. Because the location of the fauh point is relatively close, it is easy to make wrong fauh diagnosis results for rolling bearings. Therefore, a fault diagnosis method for rolling bearing based on genetic variation ant colony clustering was proposed for the judgment of roller fault type. Firstly, wavelet packet was used to multiiayer decomposition and state feature vector construction for three types of fault status data and normal status data of the rolling bearing. Then, the four state feature vectors are brought into the basic ant colony algorithm and the genetic mutation ant colony algorithm respectively to obtain the fauh classification model of the rolling bearing. Finally, ten groups of fault data were selected as verification samples to verify the two clustering models. The experimental resuhs show that the clustering algorithm based on the genetic mutation ant colony has faster classification speed than the clustering algorithm based on Ant Colony Optimization （ACO）.

关 键 词：滚动轴承故障 小波包 特征向量 蚁群聚类 遗传变异 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]