Peridynamic Simulation of Heterogeneous Rock Based on Digital Image Processing and Low-Field Nuclear Magnetic Resonance Imaging

Abstract

Rock heterogeneity is one of the most important factors when numerically simulating the rock failure process and crack propagation. This paper presents an approach for capturing rock heterogeneity that combines peridynamic theory, digital image processing (DIP), and low-field nuclear magnetic resonance (NMR) imaging. By processing the magnetic resonance images (MRIs) of the rock material, the microstructure distribution is obtained, and the attenuation coefficient is defined and calculated by the number and value of pixels of the rock MRI. Based on bond-based peridynamic theory, the density, bond constant, and critical stretch associated with the material point and bond are revised by the attenuation coefficient and have the same distribution with real rock microstructure. Then, this new approach is used to simulate the crack bifurcation of red sandstone under tension and the failure process of mudstone in an unconfined compression test. The influence of the element’s pixel number on the calculation result is discussed. Numerical results show that the crack propagation and the failure process based on the new approach are both distributed nonsymmetrically. In addition, the acoustic emission rule in the peridynamic simulation, which is defined by the number of broken bonds in each loading step, is consistent with the experimental results. The comparison of the numerical and experimental results reveals that the present approach can be used as a supplementary method for analyzing rock damage and failure.