摘要
介绍了支持向量机原理与算法,给出了基于支持向量机多类故障层次分类检测模型以及多类故障分类器构建方法及步骤。用滚动轴承实验数据对分类器的性能进行检验,并与神经网络分类器性能进行初步对比,结果证明了支持向量机方法用于轴承故障检测与识别的可行性和有效性。
The principle and algorithm of support vector machines (SVM) were introduced. A multi-classifier model based on SVM and the correlative method of using this model to deal with the fault detection and classification for ball bearings were presented. The performance of SVM based multi-classifier was tested and compared with artificial neural network classifier by using experimental data of ball bearings. The feasibility and validity of using SVM based approach to detect and identify ball bearings faults are demonstrated by experimental results.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2005年第6期498-501,共4页
China Mechanical Engineering
关键词
支持向量机
神经网络
故障检测
模式识别
滚动轴承
support vector machine
neural network
fault detection
pattern recognition
ball bearing