[CIG-SHORT] mesh resolution

[Brad Aagaard]

Tu Xiang,

cellSize has the same meaning in the script as "size" does when you 
using constant sizing, bias sizing, etc. within CUBIT/Trellis. It is the 
desired length of edges of a cell (distance between vertices of a cell 
edge).

Within the script the cellSize is a function of x, y, z and is set equal 
to desired length of edges of cells. In getCellSizeDB() we set the cell 
size equal to 0.1 times the shortest wavelength of the S waves we want 
to resolve (minPeriod*vs/10). In getCellSizeFn() we set the cell size 
equal to an analytical function that depends on distance from a target 
point with cell size increasing at a geometric rate with distance from 
the point.

Regards,
Brad


On 03/07/2018 05:11 PM, tu xiang wrote:
> Dear Charles,
> Yes, I have looked at the Trelis documentation for 'Exodus II-based 
> Field Function’ . But I do not know how to set the sizing function 
> because I still cannot understand its meaning.
> 
> For instance, the example of "cubit_cellsize" set the cellSize as follow:
> 
> /vs = data[:,0] //cellSize = minPeriod*vs / 10.0 //return cellSize/
> 
>   But I do not know the meaning of the variable "cellSize", what is the 
> function of the variable "cellSize" while the .jou file call this 
> variable? such as:
> # ----------------------------------------------------------------------
> # Create tet4 mesh with resolution dictated by 'cell_size_fn'.
> # ----------------------------------------------------------------------
> volume all scheme tetmesh
> import sizing function  "mesh_cellsize.exo" block all variable 
> "cell_size_fn" time 0.0000000
> volume all sizing function type exodus
> surface all sizing function type exodus
> curve all scheme stride
> 
> I do not know how the variable "cell_size_fn"  controls the size of the 
> cell on the fault. And if I want to control the size of the cell on the 
> fault with the variable "cell_size_fn" so that the size of the cell on 
> the fault must smaller than one constant value, how can I set this variable?
> 
> Best regards,
> Tu Xiang