摘要
轴承是旋转机械中的关键部件,相对于故障模式识别,性能退化评估可以更为有效地服务于设备主动维护以实现零停机率。小波包分解可以对信号进行更为精细的刻画,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,提出了一种基于小波包-支持向量数据描述的轴承性能退化评估方法,该方法以小波包分解的节点能量构成特征向量,仅需要正常状态下的数据样本即可用支持向量数据描述建立知识库,在一定程度上实现了对待测样本退化程度的定量评估。通过应用于轴承不同点蚀大小和其加速疲劳寿命试验的全寿命周期,验证了所提出方法的可行性和有效性。
Bearings are important units in a rotary machinery, their performance degradation assessment is more effective than fault pattern recognition for proactive maintenance to realize near-zero downtime. Wavelet package decomposition (WPD) can describe signals finely, statistical learning theory (SLT) based support vector data description (SVDD) is one kind of classification method holding excellent computing ability. Here, a bearing performance degradation assessment method based on them was proposed. This method used node energy of WPD to form feature vectors, and needed normal data to build knowledge database with SVDD, then quantitative degradation of test data could be realized. Application results for a bearing with different defect sizes and its whole life time from accelerated life test verified this method’s feasibility and validity.
出处
《振动与冲击》
EI
CSCD
北大核心
2009年第4期164-167,共4页
Journal of Vibration and Shock
基金
国家自然科学基金资助项目(50675140)
国家高技术研究发展计划(863计划
2006AA04Z175)
关键词
支持向量数据描述
小波包分解
性能退化评估
加速疲劳寿命试验
轴承
support vector data description(SVDD)
wavelet package decomposition(WPD)
performance degradation assessment
accelerated life test
bearing