基于LTSA融合降维法的滚动轴承可靠性评估方法

Rolling bearing reliability evaluation method based on LTSA fusion dimensionality reduction method

为了保障滚动轴承在给定工况下安全平稳地运行,提出了一种基于局部切空间排列(LTSA)和威布尔比例故障率模型(WPHM)的滚动轴承可靠性评估方法。该方法提取滚动轴承整个寿命周期的时域、频域、时频域及统计等不同分析域上的特征指标,从中筛选出包含滚动轴承运行状态的特征指标,构建出高维多域特征集;利用局部切空间排列算法对高维多域特征集进行融合降维,得到可全面反映滚动轴承性能退化趋势的综合特征指标作为WPHM的响应协变量进行可靠性评估。通过滚动轴承的寿命试验,得到该试验轴承的寿命为491 min。该方法能够得出轴承在489 min左右失效,寿命误差小于1%,故验证了该方法可以对滚动轴承的运行可靠性进行有效地评估。

In order to ensure the safe and stable operation of rolling bearings under a given working condition, a reliability evaluation method of rolling bearings based on local tangent space alignment(LTSA) and Weibull proportional failure rate model(WPHM) was proposed. This method was used to extract the characteristic indexes in different analysis domains such as the time domain, frequency domain, time-frequency domain and statistics of the entire life cycle of the rolling bearing, select the characteristic indexes effectively and sensitively reflecting the running state of the rolling bearing, so as to construct high-dimensional multi-dimensional domain feature set;The local tangent space arrangement algorithm was used to fuse the high-dimensional multi-domain feature set to reduce the dimensionality, and a comprehensive feature index that can comprehensively reflect the degradation trend of rolling bearing performance was obtained as the response covariate of WPHM for reliability evaluation. Through the rolling bearing life test, the life of the test bearing was 491 min. It can be concluded that the bearing failed around 489 min, and the life error was less than 1%. Therefore, this method can effectively evaluate the operational reliability of rolling bearing.