Numerical Analysis of Microcrack Propagation Characteristics and Influencing Factors of Serrated Structural Plane

The serrated structural plane is the basic unit of structural plane morphology. However, the understanding of its internal stress distribution, failure mode and crack evolution law was not clear enough in previous studies. In this paper, the shear mechanical properties of the serrated structural planes were studied by numerical simulation, and the crack evolution law of the serrated structural planes and the effects of four microscopic parameters on the shear properties were analyzed. The results show that: (1) the number of microcracks increases with the increase in normal stress; the crack expansion rate is slow before the shear stress reaches the peak. After the shear stress reaches the peak, the crack expansion rate continues to increase, and the microcracks keep sprouting and expanding, and the number of microcracks tends to stabilize when the shear stress reaches the residual shear strength. (2) The particle contact stiffness ratio [Formula: see text] and parallel bond stiffness ratio [Formula: see text] were negatively correlated with the shear strength; and the particle contact modulus [Formula: see text] and parallel bond modulus [Formula: see text] were positively correlated with the shear strength. As the particle contact modulus [Formula: see text] and parallel bond modulus [Formula: see text] increase, the peak shear displacement gradually decreases. The parallel bond stiffness ratio [Formula: see text] has a negative correlation with the peak shear displacement. This study is expected to provide theoretical guidance for the microscopic parameter calibration and shear mechanical analysis of serrated structural planes. (3) Several XGBoost, WOA-XGBoost, and PSO-XGBoost algorithms are introduced to construct the quantitative prediction model, and the comparative analysis found that WOA-XGBoost has the best fitting effect and can be used for the prediction of shear strength. When using this model to calculate the weight shares of micro-parameters, it was found that [Formula: see text] has the greatest influence on shear strength, followed by [Formula: see text]; [Formula: see text] and [Formula: see text] had the least influence.