Evaluation of Deformation and Strength Characteristics of Composites Based on Low-Density Polyethylene and Linen Yarn Production Waste by the Methods of Mathematical Statistics

Abstract

The possibilities of using various methods of mathematical statistics for processing and analyzing the results of deformation and strength tests on composites made from a low-density polyethylene and linen yarn production waste are evaluated. The hypothesis that the experimental strength data agree with the Gaussian distribution is examined by the Shapiro–Wilk test (W-test.) It is shown that the Gaussian distribution, both for systems unmodified and modified with diphenylmethane diisocyanate (DIC), is valid only for two parameters: the maximum tensile strength σmax and the elastic modulus Et. For the other parameters (the relative elongation emax corresponding to σmax, the specific total work of failure Ab), and the specific work of failure to the tensile strength Amax), a non-Gaussian distribution is observed. An analysis of measurements for different specimens by the Bartlett test shows that the Et data have equal variances for both systems (with and without DIC), but for the system containing DIC, the σmax data have different variances. A two-factor ANOVA analysis reveals that DIC considerably affects the tensile strength and modulus of composites, but the influence of test conditions is a statistically significant factor only for the modulus. The coefficient of variation is considerably lower for σmax than for Et and can be used as a quantitative measure for the degree of heterogeneity of the composites investigated.