基于未确知理论的小样本数据分析在机械可靠性中的应用

针对民用机械设备可靠性试验经费投入少、试验周期长等多项因素,试验数据一般都很难达到所要求的数量等问题,提出了基于未确知理论的小样本可靠性数据分析方法.该方法把数据拟合成未确知信息,运用未确知有理数进行分析,利用未确知有理数直接得到数据本身的数学特征值,解决了数据稳定性问题.该方法通过小样本可以较为精确地拟合实际真值,避免结果信息遗漏和失真,并剔除已经变异的数据或者样本.通过算例,讨论了领域半径的取值原则,对工程实例具有指导意义.

Applying ambiguity-theory-based data analysis on small-size samples for mechanical reliability

Due to such factors as small reliability testing investment and long testing period for civil machinery,the testing data size seldom meets the demand.In this regard,an ambiguity-theory-based analysis method is postulated for small sample size.For data fitting as ambiguous information,the ambiguous rational analysis is applied for directly obtaining the mathematical feature values and effectively resolving the data stability.By using small-size sample data to precisely fit the actual values,the information omission and distortion are avoided.Meanwhile,the variation data or samples can be removed.Based on a computational example,the value determination rule is discussed on the field radius.Therefore,this approach provides the guidance for industrial practicality.