On Saturday, July 12, 2014 3:05:08 PM UTC+8, Robert Cimrman wrote:
Hi Ouyang,

On 07/12/2014 05:25 AM, Ouyang wrote:
> Hi there,
>     I follow the point loaded disk example: its2D and simulate the 3d point
> loaded block as show in the attached figure.
> Here are my questions:
>     1. When I use "vertex <id>", the <id> seems not the node id in mesh and
> looks like an index in a list. My node id is 5 and I have to put 4 there.

Are you talking about the medit mesh format? The vertices in the mesh file are
not numbered at all, so the id is indeed given by their order. When a mesh is
displayed in medit, it labels the vertices starting from 1. In Python, the
indices start from 0.

In our FEM textbook for engineering students, we are taught that the nodes of a element can be numbered both locally and globally starting from 1. Based on the local numbering and  global numbering, a element stiffness matrix can be assembled into corresponding position in a global stiffness matrix.  Because of this reason, most engineers might think that mesh node in FEM have an id which is numbered starting from 1.  

>     2. Are there any way to customize contour bands? Since the point loaded
> rock block could fails at the center  point (0,0,0), I want to see the
> stress or strain details at the point (0,0,0) and would like to add more
> contour bands in.

Are you talking about post-processing / displaying the results? The postproc.py
script cannot do that, but you can open the mayavi pipeline after the scene is
displayed, and add/edit the visualization modules and filters there as you wish.

Yes,  post-processing/displaying the result. Following your suggestion, I just know how to do that in mayavi. 
>     3. How to output a contour map for stress or strain components,
> \sigma_{ij} or \epsilon_{ij},  on  any cross section in sfepy?
>         The following script can give me stress components but I am still
> looking for their corresponding coordinates
>        out['cauchy_stress'] = Struct(name='output_data', mode='cell',
>   data=stress, dofs=None)
>        print out['cauchy_stress'].data
> any suggestions would be appreciated very much.

This can be done in mayavi or paraview using the appropriate filters (cutting
plane etc).


Thx again, Robert