基于 MF-DFA 与 PSO 优化 LSSVM的滚动轴承故障诊断方法.pdfVIP

  • 11
  • 0
  • 约2.14万字
  • 约 7页
  • 2015-09-07 发布于湖北
  • 举报

基于 MF-DFA 与 PSO 优化 LSSVM的滚动轴承故障诊断方法.pdf

基于 MF-DFA 与 PSO 优化 LSSVM的滚动轴承故障诊断方法.pdf

振 动 与 冲 击 第34卷第11 期 JOURNAL OF VIBRATION AND SHOCK Vol.34 No.11 2015   基于 MFDFA与PSO优化 LSSVM的滚动轴承故障诊断方法 熊 庆,张卫华 (西南交通大学 牵引动力国家重点实验室,成都 610031)   摘 要:针对滚动轴承故障损伤程度难以确定的问题,提出对滚动轴承不同故障位置、不同损伤程度的振动信号 进行故障特征提取及智能分类的故障诊断方法。先对各状态振动信号进行MFDFA分析,选取敏感性及稳定性最好的二 种多重分形谱参数作为故障特征量,然后输入到经过PSO参数优化的LSSVM中进行故障诊断。通过仿真试验、应用实例 验证了该方法的有效性,并与LSSVM、SVM方法的诊断结果进行比较。结果表明:所提方法可实现滚动轴承故障位置及 损伤程度的智能诊断,比直接LSSVM、SVM方法具有更优的泛化性,适合解决实际工程问题。 关键词:滚动轴承;故障诊断;多重分形去趋势波动分析;粒子群优化算法;最小二乘支持向量机 中图分类号:TN911.7;TH133.3   文献标志码:A DOI:10.13465/j.cnki.jvs.2015.11.033 Rolling bearing fault diagnosis method using MFDFA and LSSVMbased on PSO XIONGQing,ZHANG Weihua (State Key Laboratory ofTraction Power,Southwest Jiaotong University,Chengdu610031,China)   Abstract: When a rolling bearing fails,it is usually difficult to determine its damage level.Aiming at this problem,a newfault diagnosismethodwaspresented to achieve feature extraction and intelligent classification ofdifferent fault positions and damage levels for rolling bearing vibration signals.Firstly,MFDFA was used to compute the multi fractal spectraofvibration signalsundereach status.Next,two multifractal spectrumsparametersbeingthe most sensitive and stable were selected and employed asfault feature values.Then feature valueswere regarded as the input ofLSSVM based on PSO forjudging rolling bearing fault position and its damage level.Finally,the effectiveness ofthe method was verified with simulation testing and actual example,and the results were compared with those of other related methods. The resultsshowedthatthepresentedmethodcanaccuratelyrealizethe intelligentdiagnosisofrollingbearingfaultposition and damage level,itha

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档