Compressive strength of fibre composites with random fibre waviness

The compressive strength of unidirectional long fibre composites is predicted for plastic microbuckling from a random two-dimensional distribution of fibre waviness. The effect of the physical size of waviness is addressed by using couple stress theory, with the fibre bending resistance scaling with the fibre diameter d. The predicted statistical distribution of compressive strength is found using a Monte Carlo method. An ensemble of fibre waviness profiles is generated from an assumed spectral density of waviness and the compressive strength for each such realisation is calculated directly by the finite element method. The average predicted strength agrees reasonably with practical values, confirming the hypothesis that microbuckles can be initiated by fibre misalignment. It is found that the probability distribution of strength is well matched by a Weibull fit, and the dependence of the Weibull parameters upon the spectral density of waviness is determined. For the practical range of fibre distributions considered, it is concluded that the strength depends mainly upon the root mean square amplitude of fibre misalignment, with the shape of the power spectral density function playing only a minor role. An engineering model for predicting the compressive strength is proposed, akin to weakest link theory for materials containing flaws. A specimen containing randomly distributed waviness is examined to locate regions of high-fibre misalignment. The strength of each of these weak regions is estimated from a look-up table derived from calculations with idealised circular or elliptical patches of waviness. The strength of the composite is given by the failure stress associated with the weakest such patch. For random distributions of waviness, the predictions using this engineering approach are in good agreement with the direct calculations of strength using the finite element method. (C) 2003 Elsevier Ltd. All rights reserved.