Simulating crack growth and coalescence from distributed flaws

Abstract:

Laboratory experiments offer detailed insights into the processes of microcrack propagation and linkage that produce faults. Here we present numerical simulations motivated by in situ dynamic X-ray tomography triaxial compression experiments that reveal the growth and coalescence of flaws within a rock core leading up to failure. These simulations provide insights into the competition between evolving components of the deformational energy budget during fracture propagation, and how fracture properties control fracture growth and flaw coalescence.

We use the 2D numerical modeling tool GROW to simulate fracture propagation for comparison with fracture networks observed in triaxial experiments of rock deformation. GROW calls the linear elastic BEM code Fric2D, simulating crack growth by identifying the orientations of propagating fractures that optimize the work of the total system. Previous 2D GROW simulations of uniaxial compression experiments with large pre-cut cracks match stress-strain and fracture propagation observations well. New simulations presented here are populated with various distributions of flaws derived from an early stage of tomography experiments in which flaws are relatively distributed and disconnected. Flaws in the simulations have lengths, orientations, and positions informed by experimental observations.

We probe details of the fracture growth process by exploring how flaw properties affect the trade-offs between the propagation of longer fractures versus coalescence of smaller flaws. We also explore how components of the work budget evolve with the propagation and coalescence of flaws. Examined energy budget components include external work, internal work, frictional work, seismic work, and the work of propagation. Previous simulations of failure from pre-cut cracks have shown that slip on distributed flaws impacts estimates of internal and frictional work. Here, we consider how many growing and coalescing flaws influence estimates of the seismic work and the work of propagation.