[CIG-SHORT] Combine loading boundary conditions and absorbing boundary condition

[Brad Aagaard]

Teng,

You should use initial stresses for the bulk material rather than the 
Neumann BC. See examples/3d/hex8/step16 for an example of how to do 
this. If you can't compute the initial stress field analytically, then 
use a separate simulation to compute the initial stresses and create a 
spatial database from the output of that simulation for the initial 
stresses in your dynamic simulation.

Regards,
Brad


On 04/09/2017 11:27 PM, Li, Teng wrote:
> Hi,
> I have a 2d rectangular model and in this model, all the four faces are
> absorbing boundary conditions. Now in order to add initial normal stress
> and shear stress for the whole domian, I use Neumann boundary
> conditions. Here, I paste only one part for your convenience:
>
> # ----------------------------------------------------------------------
> # boundary conditions
> # ----------------------------------------------------------------------
> [pylithapp.timedependent.bc]
> # Change the default Dirichlet BC to the AbsorbingDampers BC
> x_pos = pylith.bc.AbsorbingDampers
> x_neg = pylith.bc.AbsorbingDampers
> y_neg = pylith.bc.AbsorbingDampers
> y_pos = pylith.bc.AbsorbingDampers
>
>
> #-----------------------------------------------------------------------
> # x positive
> # ----------------------------------------------------------------------
> [pylithapp.timedependent.bc.x_pos]
> # Absorbing BC on +x face of bar
>
> # Name of nodeset in CUBIT
> label = face_xpos
>
> # Specify a UniformDB for the initial tractions
> db_initial = spatialdata.spatialdb.UniformDB
> db_initial.label = Neumann BC on +x
> db_initial.values = [traction-shear-vert, traction-normal]
> db_initial.data = [20.0*MPa, -50.0*MPa]
>
> # Specify label and filename for spatial database with physical properties
> db.label = Absorbing BC +x
> db.iohandler.filename = matprops.spatialdb
>
> # Specify the basis functions and quadrature:
> # 1-D Lagrange cell in 2-D space with 2nd order quadrature
> quadrature.cell = pylith.feassemble.FIATLagrange
> quadrature.cell.dimension = 1
> quadrature.cell.quad_order = 2
>
>
> However, I see the manual, that we should change BC type to Neumann:
>
> bc.x_pos = pylith.bc.Neumann,
>
> So I am wondering how can I apply the initial stress that keep the whole
> domain in the initial stress condition while keep the absorbing boundary
> conditions of all the faces.
>
> Best wishes
> Teng Li
>
> Teng Li
>
> Master Candidate in Structures
>
> Department of Civil and Environmental Engineering
>
> University of Illinois at Urbana-Champaign
>
> 205 North Mathews Ave, Urbana, IL. 61801
>
> Phone:(217)8196210, Email: tengli2 at illinois.edu
>
>
>
>
>
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>