The mechanical properties of coal-rock masses are strongly affected by discontinuous structural defects such as joints and fractures. Identifying the dominant fractures allows for the simplification of complex fracture networks while preserving the mechanical behavior of the fractured rock mass and enhancing the computational efficiency of numerical models. This study reconstructed a numerical model of a coal specimen with complex fracture networks by integrating computed tomography (CT) scanning technology and numerical simulations and investigated the influence of the size ratio on the mechanical properties and failure characteristics. The dominant fracture size was determined. An equivalent discrete fracture network (DFN) model was developed in MATLAB by fitting planes to the three-dimensional coordinates of the fracture endpoints. The study results show the following: (1) The sensitivities of the mechanical parameters to fracture size are ranked in the order of tensile crack initiation stress > uniaxial compressive strength > peak strain > shear crack initiation stress > elastic modulus. (2) As the size ratio increases, the relationship between the crack number and dip angle shifts from exponential to linear growth. The failure characteristics consistently exhibit mixed tensile–shear failure. The macroscopic failure pattern is increasingly governed by larger fractures. (3) The mechanical parameter errors for the original fractured coal specimen, coal specimen with dominant-size fractures, and equivalent DFN models range from 5.04% to 18.89%. The models exhibit mixed tensile–shear failure dominated by shear fractures, although the specific failure pattern varies. The findings of this study establish a foundation for future studies on the transparent analysis of discontinuous structures and evolution of the multi-physics field.
Wang et al. (Sun,) studied this question.