[CIG-SHORT] Issues with cohesive cells in VTK output

[Brad Aagaard]

On Monday 07 January 2008, Ravi Kanda wrote:
> I missed the +/-Z camera orientation there - sorry.  Yes, it does make
> sense that there are fault perpendicular displacements in order to prevent
> void generation.  Does this mean that the illustration in Figure 5.8 of
> User's Manual should be updated (p. 48, TwoQuad4 section: "Color contours
> and vectors of displacement for the kinematic fault example")?  In that
> figure, the displacement vectors are fault parallel.  I was trying to
> reconcile my output with that figure.

We will check this. We may have changed the BC since we created this figure 
(or we created the figure incorrectly).

> I am attaching the VTK output files for the 2D two-cell examples to this
> e-mail. When you say you "remove the cohesive cells from the VTK output",
> do you mean just in the connectivity section of the VTK file [where nodes 5
> & 7 (tri3) or 7 & 9 (quad4) are not used]?  

Yes. The vertices associated with the Lagrange multipliers are still in the 
file (but they won't be there in the future).

> I am also trying to understand why the "RGB-Alpha" values under the Point
> Data SCALAR Lookup Table have "serial" numbers starting from 2 in the "R"
> column and the vector components in the "G", "B" & "Alpha" columns.  This
> assignment makes "Alpha" = 0 (transparent scalar rendition?).  I would have
> thought the Point Data SCALAR Lookup Table will have RGB values scaled by
> the magnitude of the vector at each point and "Alpha" of 1.0 for each node
> of the mesh for scalar data to be visible.  Is there something I am not
> getting here?

The field "displacements_verify*" is a debugging field that Matt added. This a 
temporary field that we haven't removed. It is the same as 
the "displacement*" field but with the label of the vertex prepended. This is 
the label in the Sieve data structure. I would ignore this field.

Brad