Hi Hub,
On Mon, Feb 6, 2017 at 4:01 AM, HH <hubpoub@xxxxxxxxx> wrote:
Hi,
I have an issue with converting a hdf file. A picture of the original mesh
is in the file original.png. At x=0, there is a folding of the ghost lines
back into the interior of the domain (the ghost points have positive x
values rather than negative).
Besides, I developped a code which allows the x=0 axis to be a symmetry
axis : it changes the coordinates of the the ghost points to perfectly
symmetrical coordinates (for instance:
mg0.vertex()(-1,Ib2,0,0)=-mg0.vertex()(1,Ib2,0,0)
and mg0.vertex()(-1,Ib2,0,1)= mg0.vertex()(1,Ib2,0,1) ). In order to make
those transformations, I read the hdf file, convert the invidual grids to
DataPointMappings and then reconstruct the overlapping grid. This code
worked very fine until now for several grids I worked with. But I have
issues with this one: it leads to the result.png grids: it has a hole in
the interior of the blue grid. In debug mode of ogen, this hole appears
when marking all the interpolation points (see interpol.png). I cannot
figure out why this happens. I modified my code so that the symmetrical
transformation is not made (then my code just read the hdf into DPM and
then convert back to overlapping grid without any changes) and I still have
this hole. How is it possible that interpolation points appears right in
the interior of a domain ?
If needed, I can provide .cmd and my code
Thanx for reading me !
Hub