2D iovtk Files corrupted

General Discussion
1

Dear FF+±Users & Developers,
I’m facing an issue with iovtk 2d where both binary and ASCII files are corrupted. Paraview is unable to read correctly the files.
I have used version ff++3.36 ff+±4.0 on Ubuntu and MacOS and still having the same issue.
Please if you have any alternatives or patching of the bug please share it.
mk.

2

Hi,

Are you sure that your solution is correct (no NaN or Inf) using plot in FreeFEM ?
If there is some NaN of Inf in the .vtu or .vtk file, Paraview is unable to open the file.

3

Hi Simon,
I double checked and I don’t find any NAN in my ascii .vtk . I think I realize what is the issue: Actually, if I want to plot a vector field let’s say, [u,v] I should not put the any other fem function beside like for instance in the Navier stokes equation (even take the example that ff++ gives)
when I try to savevtk( [u,v,],p,…) the file will become corrupted and paraview pops errors while loading the files. I had to do savevtk( [u,v,],…) without the pressure p.
This only occurs for me in the 2d calculation case. The 3d case is fine.
Thank you again and best wishes,
mk.

4

Some important remarks:

  • you must save in P1
  • you can save multiple data, for example here velocity and pressure

What I do:

fespace PV(Th, P1);
PV uu=u, vv=v; // interpolate in P1
string dataname = "u p"; // variable name in Paraview
int[int] order = [1, 1]; // order (default in 0)
savevtk("res.vtu", Th, [uu, vv], p, dataname=dataname, order=order);

Does it works for you?

5

What is the dataname argument used for?

6

It passes the names (space-separated) to the paraview file so that they are labeled when loaded in paraview.

1 Like