期刊文献+

基于SVDD与距离测度的齿轮泵故障诊断方法研究 被引量:9

Fault diagnosis method for a gear pump based on SVDD and distance measure
下载PDF
导出
摘要 提出基于支持向量域描述与距离测度的齿轮泵故障诊断方法。对齿轮泵各种工况下振动信号进行小波包分解,提取各频带能量百分比作为特征向量;利用正常工况下特征向量训练SVDD超球模型,通过定义绝对距离测度检测齿轮泵状态是否出现异常;针对每种工况的特征向量单独训练SVDD超球模型,通过定义相对距离测度准确定位齿轮泵的不同故障工况。试验结果表明,采用小波包频带能量可降低数据维数,能有效浓缩故障信息;基于绝对距离测度与相对距离测度的SVDD故障诊断方法既能检测异常状态,亦能区分各种故障工况,可实现状态监测与故障分类识别目的。 A fault diagnosis method for a gear pump based on support vector domain discription(SVDD) and distance measure was put forward.Firstly,vibration signals under various conditions of the gear pump were decomposed with wavelet packet technique to extract energy percentage of each frequency band as feature vectors.Then,a SVDD hypersphere model was trained only using feature vectors under normal condition,and an abnormal condition of gear pump was detected with the definition of the absolute distance measure.Finally,SVDD hypersphere models were trained independently by using feature vectors under various conditions,and various fault conditions of the gear pump were recognized accurately with the definition of the relative distance measure.Experimental results showed that the data dimensions are reduced and the fault information is concentrated efectively by adopting feature extraction with wavelet packet technolog;SVDD fault diagnosis methods based on absolute distance measure and relative distance measure can not only detect abnormal conditions but also distinguish various fault conditions to realize condition monitoring and fault classification.
出处 《振动与冲击》 EI CSCD 北大核心 2013年第11期62-65,共4页 Journal of Vibration and Shock
基金 国家自然科学基金青年基金项目(61201449)
关键词 小波包分解 支持向量域描述 距离测度 齿轮泵 故障诊断 wavelet packet decomposition support vector domain discription(SVDD) distance measure gear pump fault diagnosis
  • 相关文献

参考文献9

二级参考文献35

共引文献77

同被引文献85

  • 1赵连春,许贤良,孙长敬,杨球来.第二类复合齿轮泵几何流量特性分析与试验[J].农业机械学报,2004,35(6):82-85. 被引量:5
  • 2Xu Bing Ma Jien Lin Jianjie.COMPUTATIONAL FLOW RATE FEEDBACK AND CONTROL METHOD IN HYDRAULIC ELEVATORS[J].Chinese Journal of Mechanical Engineering,2005,18(4):490-493. 被引量:6
  • 3李玉龙,刘焜.内齿轮泵实体CAD的集成方法[J].农业机械学报,2007,38(2):106-110. 被引量:3
  • 4David M J T , Robert P W D. Support vector data description ~J~ . Machine Learning, 2004, 54~ 45-66.
  • 5NangW S, Im K H, Choi J Y. SVDD-Based Method for Fast Training of Multi class Support Vector Classifier [A] . Third In ternational Symposium on Neural Networks EC] ~ Chengdu, Chi na: Springer, 2006: 991-996.
  • 6Lee D, Lee J. Domain Described Support Vector Classifier for Multi -classification Problems [J~. Pattern Recognition, 2007, 40 (1)= 41-51.
  • 7Luo H, Wang Y R, Cui J. A SVDD Approach of Fuzzy Classifica- tion for Analog Circuit Fault Diagnosis with FWT as Preprocessor EJ]. Expert Systems with Applications, 2011, 38 (8): 10554 - 10561.
  • 8Mu T T, Nandi A K. Multi class Classification Based on Extended Support Vector Data Description [J~. IEEE Transactions on Sys terns, Man, and Cybernetics-part B: Cybernetics, 2009, 39 (5).- 1206-1216.
  • 9Debnath R, Takahide N, Takahashi H. A decision based on one -against-one method for multi-class support vector machine [J]. Pattern AnalApplic2004, 27 (5): 164-175.
  • 10R. Rifkin, A Klautau. In defense of One- vs- all classification [J]. J Mach. LearningRes, 2004 (5): 101-14l.

引证文献9

二级引证文献80

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部