滚动轴承性能退化的MSET模型及其故障预警

MSET Model of Rolling Bearing Performance Degradation and Its Fault Warning

针对滚动轴承在变速变载的恶劣运行工况下极易出现性能退化现象,本文利用多元状态估计技术(Multivariate State Estimation Technique,MSET)构建多特征变量的优势,提出了一种基于MSET的滚动轴承性能退化及故障预警方法。首先提取滚动轴承振动信号的状态特征,建立健康状态振动信号特征向量间的关联模型;其次通过平均偏离度定量衡量观测向量与估计向量之间的差异,并作为轴承性能退化评估指标;最后结合滑移窗口法(Sliding Window Method,SWM)高效计算平均偏离度,实现在役滚动轴承在线状态监测和性能退化评估。轴承全寿命试验数据分析结果表明,本文所提方法可以及时有效判别轴承健康状态,实现在线故障预警。

Aiming at the harsh working condition under operational and environmental variations of axle box bearing and gear box bearing of high-speed,the advantage of using multivariate state estimation technique(MSET)is taked to construct multi-characteristic variable of vibration signal.Firstly,the state features of rolling bearing vibration signals are extracted,and the correlation model among the feature vectors of healthy vibration signals is established.Secondly,the average deviation degree is used to quantitatively measure the difference between the observed vector and the estimated vector,and is used as an evaluation index of bearing performance degradation.Finally,the sliding window method(SWM)is used to efficiently calculate the average deviation degree,and the on-line condition monitoring and performance degradation evaluation of in-service rolling bearings are realized.The analysis results of bearing life test data show that the method proposed in this paper can timely and effectively identify bearing health status and realize the online fault warning.