Mahesh Sivasambu

Monte-Carlo simulations of the fracture of elastic unidirectional model fibre composites are an important tool to understand composite reliability. On account of being computationally intensive, fracture simulations reported in the literature have been limited to simulation patches comprised of a few thousand fibres. While these limited patch sizes suffice to capture the dominant failure event when the fibre strength variability is low (synthetic fibres), they suffer from edge effects when the fibre strength variability is high (natural fibres). On the basis of recent algorithmic developments based on Fourier acceleration, a novel bisection based Monte Carlo failure simulation algorithm is presently proposed. This algorithm is used to obtain empirical strength distributions for model composites comprised of up to 2 20≈ 10 6 fibres, and spanning a wide range of fibre strength variabilities. These simulations yield empirical weakest-link strength distributions well into the lower tail. A stochastic model is proposed for the weakest-link event. The strength distribution predicted by this model fits the empirical distributions for any fibre strength variability.

The stress state in a shear-lag model of a unidirectional fiber composite with an arbitrary configuration of fiber breaks is obtained by the weighted superposition of the stress state due to a single broken fiber. In a periodic patch comprised of N fibers located at the points of a regular lattice, a method to determine the stress state due to a single break was proposed by Landis et al. (J Mech Phys Solids 48(3):621–648, 2000). This method entails the determination of the eigenspace of an N× N matrix, at a computational cost of O(N3). In the present work, an alternative algorithm is proposed. This algorithm exploits the circulant structure of the matrix describing the inter-fiber interactions. The asymptotic computational complexity of the present algorithm equals that of the discrete Fourier transform: O(Nlog N). Run times of the present method with the eigensolution based method are compared, and shown to be very favorable for the present method, even for small N. Power-law scaling of the overloads due to a single break to much larger distances than previously possible has been verified using the present method.

Monte Carlo simulations of the failure of unidirectional fibre composites typically require numerous evaluations of the stress-state in partially damaged composite patches. In a simulated composite patch comprised of N fibres, of which Nb fibres are broken in a common cross-sectional plane transverse to the fibre direction, the stress overloads in the intact fibres are given by the weighted superposition of the unit break solutions associated with each of the breaks. Determining the weights involves solving Nb linear equations, and determining overloads in the intact fibres requires matrix-vector multiplication. These operations require O(Nb3), and O(NNb) floating point operations, respectively. These costs become prohibitive for large N, and Nb; they limit Monte Carlo failure simulations to composite patches of only a few thousand fibres. In the present work, a fast algorithm to determine the overloads in a partially damaged composite, requiring O(Nb1/3NlogN) floating point operations, is proposed. This algorithm is based on the discrete Fourier transform. The efficiency of the proposed method derives from the computational simplicity of weighted superposition in Fourier space. Computations of the stress state ahead of large circular clusters of breaks in composite patches comprised of about one million fibres are used to demonstrate the efficiency of the proposed algorithm.

Monte Carlo simulations of the failure of unidirectional fiber composites in a plane transverse to the fiber direction are performed on much larger patches than in previous works, assuming a realistic load redistribution scheme from broken to intact fibers. Computational effort involved in these simulations is substantially reduced using an algorithm based on the quadtree data structure. The empirical strength distribution obtained from the simulations has a weak-link character, regardless of the variability in fiber strengths. The empirical strength distribution is well captured by a probabilistic model based on the growth of a tight cluster of fiber breaks. It is also well captured by regarding composite patch failure as the failure of the weakest equal load-sharing bundle of a certain size, following Curtin [Phys. Rev. Lett. 80, 1445 (1998)PRLTAO0031-900710.1103/PhysRevLett.80.1445]. The approximate coincidence of these two predictions identifies the dominant failure mechanism underlying Curtin's empirical scaling relationship.