考虑定时截尾数据的数控机床可靠性Bootstrap区间估计方法

发布时间：2018-06-21 07:41

本文选题：数控机床 + MTBF　；参考：《机械工程学报》2017年07期

【摘要】：针对采用参数Bootstrap方法进行数控机床可靠性区间估计时,在重抽样过程中因缺失定时截尾数据而产生较大估计误差的问题,提出一种能够充分利用数控机床定时截尾数据的可靠性Bootstrap区间估计方法。分析数控机床定时截尾试验中完整数据与定时截尾数据的时间关系,在重抽样时依据该时间关系重新生成Bootstrap重抽样样本,解决了采用参数Bootstrap方法重抽样时无法利用定时截尾数据,进而导致数控机床可靠性评估误差较大的问题。采用极大似然估计和Newton-Raphson方法得到了试验样本的点估计,采用纠偏加速系数对极大似然估计带来偏性误差进行纠偏,进而得到区间估计。结合具体实例,对抽样试验时间长度T及初步生成完整数据的个数k的设定原则进行了论述。试验结果表明:与其他区间估计方法相比,在同一置信水平下,考虑定时截尾数据的Bootstrap方法得到的可靠性区间长度短,可用于数控机床的可靠性评估。
[Abstract]:When the parameter Bootstrap method is used to estimate the reliability interval of CNC machine tools, the problem of large estimation error caused by missing timed truncated data in the process of heavy sampling is presented. A reliability Bootstrap interval estimation method which can make full use of the time truncated data of CNC machine tools is proposed. The time relation between the complete data and the timed truncated data, regenerating the Bootstrap resampling sample according to the time relation in the heavy sampling, and solving the problem that the time truncated data can not be used when the parameter Bootstrap method is resampling, and the reliability evaluation error of the NC machine tool is large. The maximum likelihood estimation and Newton- are used. The Raphson method obtains the point estimation of the test sample. The deviation error of the maximum likelihood estimation is corrected by the correction acceleration coefficient, and then the interval estimation is obtained. The setting principle of the sampling test time length T and the number k of the initial complete data is discussed with the concrete examples. The test results show that: and the other intervals. Estimation method compared at the same confidence level, considering the reliability interval length Bootstrap method of timing censored data is short, can be used for evaluating the reliability of CNC machine tools.
【作者单位】：吉林大学机械科学与工程学院;北华大学机械工程学院;
【基金】：国家科技重大专项资助项目(2014ZX04014-011)
【分类号】：TG659

【相似文献】