文档介绍：基于粒子群优化支持向量机的回转支承
寿命状态识别
陆超,陈捷,洪荣晶,封杨收稿日期:2014-10-23
基金项目:国家自然科学基金( 51375222);2014年度高校“青蓝工程”
作者简介:陆超(1990-),男,江苏扬州人,硕士,主要研究方向为回转支承寿命预测;陈捷( 联系人),教授,E-mail:******@.
(南京工业大学机械与动力工程学院,江苏南京 211800)
摘要:回转支承已在工程机械和风力发电等方面得到广泛应用。为了对其健康状态做出正确判断,本文采用经粒子群算法优化的支持向量机模型来对其寿命状态做出准确识别。寿命状态识别的关键问题是特征向量的提取。为了得到有效而又全面的寿命状态信息,从时域和时频域方面提取多个特征向量进行综合分析,从而实现了小样本数据下信息的最大挖掘。最后以回转支承全寿命实验对该方法进行检验,结果表明该模型的效果优于传统的支持向量机以及单变量模型,具有实际工程应用价值。
关键词:回转支承;支持向量机;粒子群;寿命状态识别
中图分类号:TH17 ; TP18 文献标志码:A
Slewing bearing life state recognition based on support vector machine optimized by particle swarm
LU Chao, CHEN Jie, HONG Rongjing, FENG Yang
(College of Mechanical and Power Engineering, Nanjing Tech University, Nanjing 211800, China)
Abstract: Slewing bearing has been widely used in engineering machinery and wind power. In order to make the right judgments on their health status, the support vector machine optimized by particle swarm algorithm model was proposed to make an accurate identification of the life state. The key problems of life state recognition is the feature vector extraction. In order to get an effective prehensive life state information of slewing bearing, multiple-feature vectors from time domain and time frequency domain were been extracted, only in this way can the information under small sample be extracted as much as possible. Finally, the slewing bearing life experiments were used to