复合材料许用值试验数据处理方法

基于现有文献资料中对于复合材料许用值试验数据统计方法的总结,论述了复合材料许用值试验数据处理方法,包括正则化要求,异常数据的处理以及B基准统计方法,并对B基准统计过程中的批间差异检验、方差等同性检验和正态分布检验的工程判断进行了详细的介绍。重点阐述了合并法、单点法、小样本方法和最小批次B基准值方法的使用方法与限制条件,提出了B基准统计方法的选择流程。在B基准统计过程中,提出了采用修改离散系数方法补偿过低的变异性,以提高通过批间变异性和方差等同性检验的机率,并得到保守的复合材料许用值。采用所述方法对2组复合材料许用值试验数据进行了B基准值计算,结果表明：合并法与单点法得到的B基准值比较稳定且差异较小；小样本统计方法得到相对保守的B基准值,但是与母体的B基准值十分接近；最小批次B基准值方法得到的B基准值不一定保守,得到的B基准值不够稳定,需补充试验数据以得到用于设计和强度分析的许用值。

Composite Allowable Value Test Data Processing Methods

Based on statistics methods for composite allowable value test data summarized in current documents and files, data processing methods of composite allowable value test were described in this paper, including normalization requirements, abnormal data processing and B-basis statistics methods. This paper also expounded engineering judgments in the check of variation between batches, equality of variance check and normal distribution check. Application and attention of pooling method, single point method, small-sample method and minimum batch B-basis value method were introduced emphatically, then selection process of these B-basis statistics methods was proposed. In the process of B-basis statistics, modified coefficient of variation approach was applied to compensate the excessively reduced variation,it could increase the probability of passing the variation check between batches and equality of variance check, which could also get a conservative composite allowable value. This paper adopted the methods to calculate B-basis value of two composite allowable value test data groups. The results indicate, B-basis values of pooling method and single-point method are comparatively steady and with small differences. B-basis values calculated by small-sample method are conservative but extremely close to the value of population. B-basis values calculated by minimum batch B-basis value method are unstable and not always conservative, more test data should be acquired to calculate allowable values adopted to design and strength analysis.