[Paraview] using integrate variables between two filters

Michael Rapson rpsmic001 at gmail.com
Wed Dec 9 06:41:12 EST 2009


Hi all,

I use Paraview to post process results from finite element simulations
and find it very useful. There is however, one thing that seems
feasible but I cannot find out how to do in Paraview and I would
appreciate it if someone could tell me whether it is possible.

I write data from my simulation to file as .vtk with a number of
variables at each mesh location (say x velocity, y velocity and
pressure). In Paraview I can load the file, compute result =
velocity_X - velocity_Y using the calculator and then integrate
variables to get the integral of velocity_X - velocity_Y on the mesh.
This is obviously not a very useful result in general, what I really
want is to compute velocity_X_simulation_1 - velocity_X_simulation_2,
where simulation 1 and 2 share the same mesh but some different
parameters. I can use a similar method to the one described above as
long as the data from both simulations is stored in the same .vtk
file, but is it possible to load two .vtk files into a single
pipeline, or to access data across pipelines for the integrate
variables method?

An even more optimistic question, if the two simulations are run on
meshes that exist on the same domain (i.e. region in space) but have a
different set of nodes, say the mesh for simulation 2 is refined once
more than the mesh from simulation 1, is it possible for Paraview to
interpolate data from the less refined mesh to the more refined mesh
in order to compute velocity_X_simulation_1 - velocity_X_simulation_2
and integrate?

The first scenario seems technically feasible, if it has been offered
in Paraview. Alternatively, are there specialized programs that can
read two .vtk files and combine their variables into a single .vtk
file? The second is more optimistic and couldn't be done without
explicitly interpolating the results from one mesh to the other. I
have found the Group Datasets filter, but this does not seem to quite
do it (the data sets are marked (partial) once grouped and calculating
differences fails).

Thank you for any ideas about how I could achieve this, and also
confirming whether it can be done in Paraview or if I should make
another plan will help me to move forward.

Regards,
Michael


More information about the ParaView mailing list