结合循环平稳和支持向量数据描述的轴承性能退化评估研究被引量：12

Assessment of Bearing Performance Degradation by Cyclostationarity Analysis and Support Vector Data Description

下载PDF

导出

摘要轴承性能退化评估是进行设备主动维护、实现零停机率的关键技术。循环平稳分析能够识别滚动轴承微弱故障,基于统计学习理论的支持向量数据描述是一种具有良好计算性能的单值分类方法。基于此,本文结合二者,提出了一种新的轴承性能评估方法。该方法以循环平稳分析进行特征提取,得到组合切片累积能量,在仅有正常状态下的数据样本时,即可用支持向量数据描述建立知识库,从而实现了对待测样本退化程度的定量评估。通过对轴承加速疲劳寿命试验中全寿命周期的评估,验证了所提出方法的可行性和有效性。 Bearing performance degradation assessment is one of the most important techniques for proactive maintenance and realizing near-zero downtime. Cyclostationarity analysis （CA） can recognize a bearing＇s weak fault, and statistical learning theory （SLT） based support vector data description （SVDD） is a one-value classification method with excellent computing ability. We combine these two techniques to form a new robust assessment method. CA is used as a feature extraction tool to obtain combination slice accumulation energy, and it only needs normal data to build knowledge database using SVDD, then qualitative degradation degree for test data can be realized. Assessment results of the whole life time of a bearing by accelerated life test validate the feasibility and effectiveness of this method.

作者潘玉娜陈进

机构地区上海交通大学机械系统与振动国家重点实验室

出处《机械科学与技术》 CSCD 北大核心 2009年第4期442-445,共4页 Mechanical Science and Technology for Aerospace Engineering

基金国家自然科学基金项目(50675140) 国家高技术研究发展计划项目(863计划 2006AA04Z175)资助

关键词支持向量数据描述循环平稳性能退化评估加速疲劳寿命试验轴承 support vector data description cyclostationarity analysis performance degradation assessment accelerated life test bearing

分类号 TH165.3 [机械工程—机械制造及自动化] TN911 [电子电信—通信与信息系统]

引文网络
相关文献

参考文献7

1Lee J. Measurement of machine performance degradation using a neural network model[J]. Computers in Industry, 1996,30: 193 - 209
2Huang R Q, et al. Residual life predictions for ball bearings based on self-organizing map and back propagation neural network methods [ J ]. Mechanical Systems and Signal Processing, 2007,21:193 -207
3Yan J H, Lee J. Degradation assessment and fault modes classification using logistic regression [ J ]. Journal of Manufacturing Science and Engineering-Transactions of the ASME, 2005, 127:912 - 914
4Ocak H, et al. Online tracking of bearing wear using wavelet packet decomposition and probabilistic modeling A method for bearing prognostics [ J ]. Journal of Sound and Vibration, 2007,302:951 - 961
5李凌均,韩捷,郝伟,董辛,何正嘉.支持向量数据描述用于机械设备状态评估研究[J].机械科学与技术,2005,24(12):1426-1429. 被引量：22
6David M J, et al. Support vector domain description[ J]. Pattern Recognition Letters, 1999,20:1191 - 1199
7Rubini R, Meneghetti U. Application of the envelope and wavelet transform analyses for the diagosis of incipient faults in ball bearings [ J ]. Mechanical Systems and Signal Processing, 2001,15(2) :287 -302

二级参考文献6

1Tax D M J , Duin R P W. Support vector domain description[J]. Pattern Recognition Letters, 1999,20(11-13): 1191-1199.
2Vapnik V N. The Nature of Statistical Learning Theory[M].New York:Springer-Verlag, 1995.
3Xin D, Wu Z H, Zhang W F. Support vector domain description for speaker recognition [A]. Proceedings of the 2001 IEEE Signal Processing Society Workshop [C], Massachusetts,2001:481-488.
4Chen Y Q, Zhou X, Huang T S. One-class SVM for learning in image retrieval[A]. In: Proceedings of International Conference on Image Processing 2001 [C], Thessaaloniki, Greece,2001:440-447.
5Tax D M J , Duin R P W. Outliers and data descriptions[A].In: Proceedings of the Seventh Annual Conference of the Advarced School for Computing and Imaging [C], Delft,2001:234-241.
6李凌均,张周锁,何正嘉.基于支持向量数据描述的机械故障诊断研究[J].西安交通大学学报,2003,37(9):910-913. 被引量：55