Review of numerical modeling based on digital image processing for rock mechanics applications

Heterogeneity has been incorporated into numerical models, dues to the fact: The mechanical response of geomaterials is highly affected by microscopic heterogeneity(material heterogeneity, geometrical heterogeneity, boundary/interface heterogeneity). In the past, it was difficult to quantitatively characterize rock heterogeneity owing to technical limitations. Facing reality, researchers are often wise: On the one hand, they pursue better reason, thus constantly improve measurement tools and analysis techniques to achieve quantitative characterization of rock microstructures; on the other hand, they are good at finding a feasible path to solve practical problems in rock engineering, such as establishing the classical elasticity based on the homogeneity hypothesis. With the fast development of the related techniques, it is becoming possible to characterize microscopic inhomogeneous quantitatively, and to construct numerical models for the actual spatial distribution of different minerals/components in rock using digital image processing(DIP). The review begins by describing the development process of rock heterogeneity research from qualitative to quantitative analysis. Then, the characterization of rock microstructures based on DIP and the research progress of modeling using DIP for geomechanics applications are presented. The likely future developments are discussed, such as the assignment of input parameters, characterization of boundary/interface heterogeneity, quantitative description of geometric heterogeneity, unity of implicit approach(statistical method) and explicit approach(mineral-based method), 3D modeling of unstructured mesh, mesoscopic characterization with parallel acceleration. © 2020, Science Press. All right reserved.