Spatial distribution of mechanical parameters along a fracture interface

Abstract

Many theoretical and numerical approaches for wave propagation in fractured media use normal and shear fracture specific stiffnesses to represent the complexity of fracture topology during deformation under stress. In this paper, XLiFE++ is used to simulate the interaction between a fracture and elastic wave on a 2D plane, the simulated displacement field around the fracture is converted into traction on the fracture interface, and the jump in displacement on the fracture interface is calculated relative to displacements on both sides of the fracture. Based on the linear slippage model, the specific value between jump displacement and traction is calculated at each fracture point to obtain the fracture stiffnesses and then the stiffness curves along the fracture surface are plotted using the stiffnesses at each fracture point. Using XLiFE++, 144 orthogonally designed parameter combinations are simulated, the simulated displacement field around the fracture with material parameters and fracture stiffnesses is used as a training sample, and a neural network method for estimating the fracture stiffnesses is provided. The verification for the training results indicates that the ANN model can effectively represent the non-linear relationships between Young’s modulus, Poisson’s ratio, excitation frequency of materials, shear stiffness and normal stiffness of the fracture, and displacement field around the fracture and estimate the fracture stiffnesses relative to displacement field around the fracture surface. Therefore, the proposed method is an effective tool to rapidly determine the fracture interface stiffnesses. • Spatial distribution of elastic modulus around fracture •Simulate the interaction between elastic waves and fracture with different stiffness •Calculation of fracture stiffness based on displacement discontinuity theory