[vtkusers] Merging two Unstructured Grids / Surfaces

[David Doria]

On Mon, Mar 1, 2010 at 4:56 PM, Eric E. Monson <emonson at cs.duke.edu> wrote:

> Hey Karl,
>
> I'm sure others will have more ideas, but regarding combining input grids,
> there was a thread last July in which people were talking about using the
> GNU Triangulated Surface Library to do boolean operations on surfaces:
>
> http://www.vtk.org/pipermail/vtkusers/2009-July/101865.html
>
> You have to convert back and forth between the VTK and GTS data structures,
> but I have played around with it a bit and it works well. (There is also
> some partial wrapping of the GTS in python if this is useful.)
>
> And, looking through the VTK classes, there is a vtkImplicitDataSet (with a
> couple of Tcl tests as examples to show how to use it), which maybe could
> help with the clipping you are talking about. This is where others should
> know more than I.
>
> http://www.vtk.org/doc/nightly/html/classvtkImplicitDataSet.html
>
> Talk to you later,
> -Eric
>
> ------------------------------------------------------
> Eric E Monson
> Duke Visualization Technology Group
>
>
> On Mar 1, 2010, at 4:29 PM, Karl Heinrichmeyer wrote:
>
> > Hi together ..
> >
> > I'm working at a project that uses Vtk to prepare intermediate results of
> different numeric simulations, all part of the engineering sciences, for the
> next simulation step. The data to work with are in general unstructured
> grids representing the geometry and properties of working-pieces while the
> simulations are something like welding and forging.
> >
> > For one of this simulations i need to combine two input grids, that
> represent two working-peaces prepared for welding,  into one single surface.
> The two working peaces, in the actual case a gear-wheel and a ring, "lie on
> each other". That means the grids have contact and may intersect a little
> bit. The input consists of unstructured grids. The output has to be the
> combined surface of both grids without the contact surface. The mesh of this
> output surface isn't important while the shape has to match the input exact
> as possible.
> >
> > I tried several things on my own and my current plan to proceed is the
> following:
> >
> > 1. Extract the surfaces of both geometries using vtkDataSetSurfaceFilter
> > 2. Clipping the resulting polydata with each other using vtkClipPolyData
> > 3. Extracting the cutting edge with vtkFeaturedEdges
> > 4. "sewing" the edges by hand with an own algorithm
> > 5. put everything together.
> >
> > My first question is: Has anybody a better idea to do this? And maybe: Is
> this possible at all?  I searched the vtk documentation quite a long time
> without results.
> >
> > The second problem i found is to clip the surfaces with each other. I
> would need something like a vtkImplicitPolyData. I searched for something
> like this too but the only thing i found was a thread from 2001. Is there
> something like an vtkImplicitPolyData in vtk that takes a polygonal surface
> and behave like an implicit function that can be used for clipping?
> >
> > Thanks in advance.
> >
> > Karl
>
>
To the problem about clipping one object with another, using the classes
mentioned I tried to do this:
http://www.vtk.org/Wiki/VTK/Examples/ImplicitDataSet

But it complains "Can't evaluate dataset!". The documentation says you must
have scalars - maybe you can figure out what they need to be?

Thanks,

David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20100301/e5abd650/attachment.htm>