A novel DIC-based methodology for crack identification in a jointed rock mass

Abstract

Crack identification in brittle rocks is a challenging problem in rock mechanics. To represent the details of jointed rock masses, rock-like specimens containing inclined joint sets were generally prepared by three-dimensional (3D) printing using sand. Digital image correlation (DIC) was employed for full-field deformation measurement during loading. The results show that the sand 3D-printed specimen has distinct advantages in the preparation of jointed rock masses. DIC was extended to measure the displacement vector field around newly propagated cracks. Two types of cracks are recognized: tensile and shear cracks. Two different coalescence patterns are classified: step-path failure and planar failure. A covariance matrix-based multivariate measure, the rate of effective variance change (REVC), was proposed to quantify the dispersion of strain data. The variation in the strain dispersion measure was found to be closely related to the crack type. For the first occurrence of tensile and shear cracks, the rates of mutation (RMs) are within the ranges of 1.56–3.42 and 14.54–86.44, respectively, which can be considered a new crack identification criterion. A virtual extensometer method based on DIC was established to interpret the variation mechanism of the REVC associated with different crack types.

Keywords

• Jointed rock mass
• 3D printing
• Digital image correlation
• Mechanical property
• Crack
• Strain dispersion