Study of the Damage and Cracking of Heterogeneous Rocks Based on the FFT Method - A Comparison of Computer Generation and Image Reconstruction Models

Lei Peng,Mingyao Li,Zhimin Wang,Jianping Zuo,Dejun Liu
DOI: https://doi.org/10.1016/j.enggeo.2023.107284
IF: 6.902
2023-01-01
Engineering Geology
Abstract:Studying the influence of microstructure on the damage and cracking process of rocks is of great practical engineering importance for evaluating the mechanical properties of surrounding rocks in complex geological engineering environments. To incorporate rock microstructures into numerical models, two main methods, i.e. computer generation methods (e.g. molecular dynamics, MD) and image reconstruction methods (e.g. digital image processing, DIP), are frequently adopted. However, the strengths and weaknesses, the differences and applicability regarding the two major methods have not been systematically discussed and analyzed in simulating the damage and cracking in heterogeneous rocks. In this study, the MD model was firstly optimized based on the previous work, and the “Realistic grain size based microstructure model generation program” was developed by the observed mineral grain sizes from microscopic images. The DIP model was developed by directly using digital images of rock materials via the image reconstruction methods. On this basis, the damage and cracking process of rocks was investigated and compared with the above two models using a technique based on the Fast Fourier transforms (FFT). To this end, a series of macro and micro experiments were conducted to obtain the basic mechanical properties of the Dingziwan Granite (DG) and several examples are generated by the MD and DIP model with calibrated macro and micro parameters by the laboratory data. The two models combined with the FFT method are applied to predict the macro and micro mechanical behavior of the DG under different confining pressures and their differences and applicability are quantitatively analyzed and compared. The results indicate that, in general, the two models are both capable of predicting the macro and micro mechanical behaviors of the heterogeneous rocks, but each has its own strength. The improved MD model can efficiently study the influence of the rock heterogeneity on its macroscopic behavior, while the DIP model can highly reproduce the microcracking process such as intragranular, transgranular and grain boundary cracking.
What problem does this paper attempt to address?