摘要
介绍了机械故障诊断的历史、意义及研究现状,分析了现有故障诊断理论方法的优点及不足之处;简要介绍了统计学习理论和支持向量机,探讨了适合故障诊断的支持向量机结构;研究了支持向量机的训练方法,目前支持向量机的训练算法是以序贯最小最优化(SMO)为代表的,其中工作集的选择是实现SMO算法的关键;在对实验结果全面分析的基础上,总结出支持向量机在机械故障诊断领域中应用的若干结论。
The thesis introduces the history, the significance and the current research status of machine fault diagnosis. Analyzing the virtue and shortcoming of current fault diagnosis theories, statistical learning theory and support vector machine are briefly introduced. Analysis of the most suitable structures of support vector machine for application to the fault diagnosis is performed. Study on the training algorithm of SVM. Currently, sequential minimal optimization （SVM） algorithm has become the best training algorithm for SVM, working set selection is the key of implementing SMO. Based on the analyzing of the experiment result, summarizing some conclusion of SVM apply in the domain of machine fault diagnosis.
出处
《装备制造技术》
2009年第12期3-5,共3页
Equipment Manufacturing Technology
基金
国家863发展计划资助项目(2008AA04Z407)
关键词
支持向量机
故障诊断
VC维
统计学习理论
support vector machine
failure diagnosis
VC-dimension
statistical learning theory
