Investigation of flaw strength distributions from tensile force-strain curves of fiber tows

YYY The flaw strength data are determined using tensile tests on various fiber tow types including, SiC, carbon, glass, basalt and alumina. The plots of p-quintile vs. flaw strength data derived from experimental force-strain curves exhibit linearity. This indicates unambiguously that flaw strength is a Gaussian variate. The statistical distribution of data was also calculated using the Weibull distribution function. Normal and Weibull cumulative distributions of flaw strengths are found to compare fairly well. The significance of normal distribution function and of the p-quintile vs. flaw strength relation for the characterisation of fibre flaw strength is discussed. The p-quintile vs. flaw strength relation is shown to provide theoretically a material characteristic. The normal distribution function is used to construct reference empirical distributions of flaw strengths that allow the evaluation of Weibull plot and Maximum Likelihood Estimation methods as functions of sample size and composition.

Automated summary

The flaw strength data is determined through tensile tests on different fiber tow types, showing linearity between the p-quintile and flaw strength data from experimental force-strain curves, indicating flaw strength as a Gaussian variate. The Weibull distribution function is used to calculate the statistical distribution of data, which compares well with the normal distribution function. The significance of both normal and Weibull cumulative distributions of flaw strengths is discussed, as well as the theoretical material characteristic shown by the p-quintile vs. flaw strength relation.