top of page
1. Overview

        We specialise in developing parallel finite element methods (FEMs) to simulate physics-based earthquake dynamic ruptures, earthquake cycles and seismic ground shaking. We focus on geometrically complex fault systems (GCFS). A snapshot of current and past projects can be referred to Research Themes. Two sets of FEMs, EQdyna and EQquasi, have been developed. EQsimu is an earthquake simulator that fuses EQdyna and EQquasi to simulate fully dynamic earthquake cycles on GCFS. 

2. Download

3. EQdyna

        EQdyna 3D is a highly efficient and scalable due to its adoptions of explicit time integration, underintegrated hexahedral elements, degenerated wedge elements, hourglass controls, perfectly matched layer and parallelization. Frequency independent Q by coarsed-grained memory scheme allows seismic attenuation in a time marching FEM. Frictional constitutations such as the classical slip-weakening and various forms of rate- and state- friction enables applications to dynamic ruptures, ground motion, fully dynamic earthquake cycles, etc. Drucker-prager off-fault viscoplasticity allows integration of numerical models with near-fault geologic observations. EQdyna has been verified against benchmark problems from community-led SCEC/USGS code verification excercise. 

        EQdyna 2D features explicit time integration, underintegrated quadrilateral elements, degenrated triangles and hourglass contrls. It automatically generate meshes that morph complex fault geometries, such as the San Andreas fault in southern California, US, and the Altyn Tagh fault in northwest China, given controlling points.  

4. EQqusi

        EQqusi is designed to simulate quasi-static processes of elastic deformation during the inter-seismic, nucleation, and post-seismic phases of an earthquake cycle. It only has 3D versions. It features fully integrated hexhedral elements. It is parallelized after mesh generation and uses PARDISO and MUMPS, two parallel direct solvers, and AZTEC, a parallel iterative solver to handle heavy computing loads. Rate- and state- frictions are implemented.

bottom of page