Quantification of microcrack characteristics and implications for stiffness and strength of granite

Microcracks can affect the mechanical properties of rocks, such as their stiffness and strength. To provide a link between the microstructural parameters and the mechanical behaviour of rock, micromechanical models use parameters that represent a quantitative description of the microcrack population. However, these parameters are difficult to constrain. With the aim of better constraining micromechanical models for rock strength and stiffness, we provide here robust measurements of microcrack characteristics. We developed an algorithm to process optical micrographs of rock, automatically creating binary images of the microcrack network. We applied this procedure to optical micrographs of fine-grained granite samples that have undergone varying degrees of thermal microcrack damage. From these processed images, we calculate the mean microcrack length and the number of microcracks per unit area (and therefore the 2D microcrack density). We also create heat maps showing the spatial distribution of microcracks and their lengths. The results of our automated image analysis are in very good agreement with those of widely-used stereological techniques, and we show that our method can be applied to other rock types (sandstone and andesite) that contain microcracks. Using the measured microcrack characteristics as inputs for Ashby and Sammis' (1990) 2D micromechanical sliding wing crack model, we predict the uniaxial compressive strength of the granite and compare the predictions with strength measurements made in the laboratory. We find good agreement between the model and the experimental data for granite heated to temperatures below the alpha-beta transition of quartz (similar to 573 degrees C). Rock strength is overestimated above this threshold, possibly due to variations in fracture toughness, which is considered constant in our modelling. Finally, we use the 2D sliding crack model of David et al. (2012) to infer microcrack density and aspect ratio from the mechanical response of the thermally microcracked samples to cyclic stressing. We show a good agreement between inferred and measured crack densities if a scaling factor is introduced.