小样本下轴承可靠性评估方法对比分析  

Comparative Analysis on Reliability Estimating Methods of Bearing Test with Small Samples

作　　者：高攀东[1] 陈晓阳[1] 沈雪瑾[1] 丛伟[1] 张小玲 

机构地区：[1]上海大学机电工程与自动化学院,上海200072 [2]苏州轴承厂有限公司,江苏苏州215129

出　　处：《轴承》2014年第12期44-48,共5页Bearing

基　　金：国家"十二五"科研项目上海大学子项目(D.50-0109-12-001)

摘　　要：针对高可靠性长寿命轴承在小样本试验方案下的可靠性评估问题,介绍了3种基于Weibull分布的小样本分析方法,通过Monte Carlo进行计算机模拟,对比分析了3种方法在不同Weibull分布参数、样本量和失效数下的评估效果。结果表明,在小样本情况下(n≤10),各方法的差别较大,但随样本量和失效数的增加,差别越来越小。BLIE法的评估效果最好,MUME法次之,LSR法最差。Aiming at the reliability estimation of high reliability and long life bearing under small sample size,three pa-rameter estimation methods of Weibull distribution are discussed.A simulation is conducted between the assessment re-sults of MUME,BLIE and LSR by Monte Carlo simulation.The effects of estimation under different Weibull parame-ters,different sample sizes and the number of different failure are compared.The results show that in small sample （n≤10）,the estimation difference of each method is large,but with the increase of sample size,the difference is more and more smaller.The evaluating effect of BLIE method is the best,followed by MUME,LSR is worst.

关 键 词：滚动轴承 WEIBULL分布 小样本 中值无偏矩估计法 最佳线性不变估计法 最小二乘法 可靠性 

分 类 号：TH133.3[机械工程—机械制造及自动化] TB114.3[理学—概率论与数理统计]