Strength and fracture toughness of heterogeneous blocks with joint lognormal modulus and failure strain

We obtain analytical approximations to the probability distribution of the fracture strengths of notched one-dimensional rods and two-dimensional plates in which the stiffness (Young's modulus) and strength (failure strain) of the material vary as jointly lognormal random fields. The fracture strength of the specimen is measured by the elongation, load, and toughness at two critical stages: when fracture initiates at the notch tip and, in the 2D case, when fracture propagates through the entire specimen. This is an extension of a previous study on the elastic and fracture properties of systems with random Young's modulus and deterministic material strength (Dimas et al., 2015a). For 1D rods our approach is analytical and builds upon the ANOVA decomposition technique of (Dimas et al., 2015b). In 2D we use a semi-analytical model to derive the fracture initiation strengths and regressions fitted to simulation data for the effect of crack arrest during fracture propagation. Results are validated through Monte Carlo simulation. Randomness of the material strength affects in various ways the mean and median values of the initial strengths, their log-variances, and log-correlations. Under low spatial correlation, material strength variability can significantly increase the effect of crack arrest, causing ultimate failure to be a more predictable and less brittle failure mode than fracture initiation. These insights could be used to guide design of more fracture resistant composites, and add to the design features that enhance material performance.