Benchmark 7

Elastic solution for a circular strike-slip fault. The conceptual model is an elastic disk of radius 200 km, with a circular left-lateral strike-slip fault forming an inner plug which rotates inside the outer annulus. Because of the symmetry of the problem, radial displacements should vanish and only the first quadrant needs to be modeled. If required, a mesh using cylindrical coordinates can be downloaded from http://geoweb.mit.edu/fe/.

GOALS

• Test techniques and implementation methods for non-planar faults (use of local vs. global coordinate systems, etc.)
• Investigate the grid resolution required to properly resolve non-planar faults.
• Test ability of various codes to model non-planar faults
• Test implementation of boundary conditions in terms of Cartesian and Polar coordinates.
• Code comparison

DETAILED DESCRIPTION

• Model size: Thickness = 40 km; 10 km ≤ r < 200 km; 0 ≤ θ ≤ π/2
• Elastic material properties: Poisson solid, G = 30 GPa
• Density and Gravity: None
• Boundary conditions: Bottom pinned x-displacement pinned at y = 0 (i.e., θ = 0) y-displacement pinned at x = 0 (i.e., θ = π/2)
• Coarse mesh node spacing: dr = dz = 2 km; dθ = 2 degrees
• Fault specifications: Type: Vertical strike-slip Location: r = 100 km; -16 km ≤ z ≤ 0 km Slip distribution: 1 m of uniform left lateral slip from -12 km ≤ z ≤ 0 km with a linear taper to 0 slip at fault tip (z = -16 km)

REQUESTED OUTPUT AND RESULTS

Mesh Variations: As memory, time, and patience allow, run models at 1/2, 1/4, and 1/8, etc. the original coarse mesh spacing, investigate variable mesh spacing, and/or employ a variety of element types.

For All Benchmark Variations:

• Stresses and displacements along a line running radially at θ = 45 degrees, and lines running with constant r = 95, 99, 101, and 105 km, at depths of 0, 12, 16, 17 and 21 below the surface, all results at times of 0, 1, 5 and 10 years.
• CPU time, wallclock time, memory usage info, compiler info, and platform info

TRUTH

The ‘best’ answer will be derived via mesh refinement. There will also be a solution generated using Okada point sources in an infinite halfspace.

ADDITIONAL NOTES

None

LaGrit Files

Benchmark problem definitions: http://www-gpsg.mit.edu/fe/Meshes.html

BENCHMARK 7 MESH GENERATION

A series of meshes of different resolution have been built to represent the BM7 model domain. The grids above are available in AVS and GMV format. See the image gallery for pictures of the mesh.

Build a 2D triangulation and smooth the triangles

• LaGriT Command File: bm7_cyl_pts2.lgi
• OUTPUT: id_fault.inp
• id_fault.gmv

Note: The control file bm7_cyl_pts2.lgi

Smooth the triangles:

• LaGriT Command File: smooth.lgi
• Input : id_fault.gmv
• Output: s_r_mass1a.gmv

Stack the triangle layers to produce a tet mesh

• LaGriT Command File: stack.lgi
• Input : s_r_mass1a.gmv
• Output: cmo_settet.inp
• cmo_settet.gmv

Refine along the fault interface

• LaGriT Command File: refine_tets.lgi
• Input : cmo_settet.inp
• Output:
  refine_interface1.( gmv / inp )
  refine_interface2.(gmv / inp )
  refine_interface3.( gmv / inp )
  refine_interface4.( gmv / inp )
  refine_interface5.( gmv / inp )
  refine_interface6.( gmv / inp )
  

All files (*.lgi *gmv *inp) in a single archive: File Size: 31.442698 Mbytes bm7_files.tar.gz

GMV General Mesh Viewer Home Page

AVS UCD (*.inp) File format description

Image Gallery

A series of meshes of different resolution have been built to represent the BM7 model domain. The grids below are available in AVS and GMV format.

Created on 03 Mar 2022, Last modified on 03 Mar 2022