优化KNNC算法在滚动轴承故障模式识别中应用  

Application of improved KNNC method in fault pattern recognition of rolling bearings

作　　者：胡智[1] 段礼祥[1] 张来斌[1] 

机构地区：[1]中国石油大学(北京)城市油气输配技术北京市重点实验室,北京102249

出　　处：《振动与冲击》2013年第22期84-87,105,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金(51005247);北京市教委科研基地建设项目

摘　　要：为有效提高滚动轴承故障诊断率,正确识别不同故障类型,提出基于优化K-最近邻域分类器(K-Nearest Neighbor Classifier,KNNC)的轴承故障模式识别方法。分别求得滚动轴承训练样本与测试样本的振动特征指标,构建样本特征集。为加快分类速度,剔除不良样本干扰,利用K-均值聚类算法对样本进行优化精简,并将所得若干聚类中心作为新的约简训练集。据新训练集进行KNNC分析,实现模式识别。结果表明:该方法能快速、有效识别出滚动轴承4种不同故障模式,识别正确率明显提高。An important reason causing mechanical equipment failures is rolling bearing faults. In order to improve the correct diagnosis rate of rolling hearing faults and recognize different faults effectively, a novel method of fault pattern recognition based on the improved KNNC （K-nearest neighbor classifier） was presented. Firstly, the vibration feature indexes of training samples and test samples were calculated separately. The feature set of samples was constructed entirely. To accelerate classification speed of KNNC and eliminate the influence of bad samples, the K-means clustering algorithm was used to optimize the training samples, and the obtained clustering centers were taken as a new training set. At last, the pattern recognition was realized with KNNC according to the new training set. Application tests showed that the improved method can effectively and quickly separate 4 different kinds of bearing fault patterns with higher recognition accuracy.

关 键 词：K-均值聚类算法 滚动轴承 故障诊断 模式识别 

分 类 号：TH17[机械工程—机械制造及自动化] TN911.6[电子电信—通信与信息系统]