Extended finite element modeling of intermediate-depth earthquakes – first steps

Abstract:

Despite nearly a century of research, the mechanics of intermediate-depth earthquakes, defined here as those between roughly 50 and 300 km depth, are still not completely understood. Several explanations have been proposed, and more than one is possible depending on the mineral content of the subducting rock. These mechanisms include transformational faulting, dehydration embrittlement, and dlip weakening coupled with thermal runaway.

To test these hypotheses, we are constructing a thermo-poro-chemo-mechanical extended finite element model of subduction zone faults. The chemical model is related to the physics of phase change, which results in densification of the material under heat and pressure. Because this is a local phenomenon, phase can be treated as a state variable and not a field variable. Heat conduction, displacement, and possibly pore fluid pressure will need to be accounted for as global fields in a fully coupled finite element method. These physics are all related, as pore fluid exerts pressure on the solid, inelastic deformation generates heat, and heat causes expansion of the material.

Under the proper combination of effective stress and temperature, the material may localize and initiate a fracture or narrow shear band. Fracture initiation may be triggered by nearby material changing phase. Slip along the localized band may generate heat that weakens the band further, either directly, by inducing further phase changes, or by heating fluid. The band may also propagate directly from redistribution of stress.

The fracture is built in to an extended finite element code with a slip weakening model. Simulations are run on sample to determine which mechanisms are important in earthquake generation.

This research will eventually be part of a multiscale framework that compares simulations of small-scale laboratory specimens, and systematically upscales the properties to real fault systems.