I am plotting data using unstructured grid (paramesh grid). My output represents bubbles formation and arisen when heating the fluid.
Paraview allows me to visualise the phenomenon with images and video.
I was using cell-data -to-pont-data filter, when zooming on the bubbles, I have some of them situated at the boundaries that have holes.
After reading few articles, it seems that the holes are due to the sensitivity to iso-contour algorithms for parallel unstructured grid data sets to incorrect data on the boundary: missing ghost cell. Therefore without ghost cell, each process does not contain enough information to know which parts of the local partition’s external boundary belong to the global external boundary and which belong only to the local external boundary.
Articles pointed that to solve this issue using paraview we need to use the filter D3.
I tried that, however my visualisation (videos and screenshot images) are similar with cell-data-to-point-data filter and D3 filter, nothing has changed.
Do you have any idea on how to get D3 filter working in my case ?
Please let me know if you need further explication or shared images/videos/data if it can help you to understand the problem.
Any help and suggestions are welcomed !
Hi @Sara ,
Which version of ParaView are you using ?
D3 does generate ghost cells if you set the right property.
In any case you can use Ghost Cells Generator to generate ghost cells.
I am using paraview version 5.9.0. This version does not support Ghose Cells Generator or maybe just in my case of data it does not active Ghost Cells Generator. In any case, I looked for Ghost Cells Generator on paraview filters, I found it but I cannot click on it !
When you said that “D3 does generate ghost cells if you set the right property”, could you please list for me these right properties ?
I’m not sure if you have an issue needing ghost cells. Ghost cells are used to remove boundaries between processors and/or files. These boundaries show up when you move opacity to have some transparency (in the properties tab).
Anyway, to clarify the D3 filter, this is an old filter that 1) creates ghost cells and 2) distributes the data among processes. It is the main filter used for ghost cells up until 5.9.1. It still exists to this day, but is definitely deprecated and will disappear soon. The new Ghost Cell Generator is what you want, became quite usable in 5.10.0 and is quite polished in 5.11.0-RC1. Try this new filter. Note that the D3 filter only works with a parallel/ MPI backend (pvserver). The new ghost cell generator works with a builtin server.
Again, as I don’t think this is the issue, I would recommend giving us an example dataset, with clear steps to replicate the issue.