Cell Data vs Point Data

Dear All!
I have a question in regard to differences between cell data and point data:
The following project has been calculated in simscale and postprocessed in paraview.
In paraview there are two different options to visualize windspeed, one uses cell data, the other own point data.

When I use cell data, the result seems to be fine (even with a rough mesh).

However, If I want to visualize point data, there are artifacts between buildings and open space and the transition between the base mesh and the first refinement area shows the same problem.

Also, if you compare the screeshots fom CellData and PointData the inward wind flow for point data seems to be between 0 and 5 m/s while cell data shows 10 m/s.
To me, the cell data result looks good, but point data donnot, as it shows inward wind flow (left hand side) with nearly 0s/m, instead of 10m/s.

Do you have any ideas, why the results are so different?
How is cell data generated in simscale and how I can change settings, in order to get a correct result?

(I posted a simular question before. Now I am trying to make the question more specific, as I think my problems with this projects is in regard to the differences between point and cell data)

Thank you very much for your help!

Points are located on the corners of cell faces.
This means a point can be touching many different cells.
A cell is a volume and has many points associated with it.
Point data appears finer since your mesh has more points than cells.
Point data is interpolated from the raw cell data.
Then, in a color based representation of the point/cell data, ParaView does blending of colors between parameter values (you can also change the # of colors shown (from the standard 24/32 bits to anything between 1 and 1024)).
A further complicating factor is that you are showing data as a Surface representation on a Slice, this again adds some color blending because, I think, the point and cell data for a Slice is interpolated to be planar.
If you want to see the precise values for point and cell data, turn on a ‘with edges’ view, zoom in to see only a few cells on the screen, then turn on either ‘Hover Cell Data’ or ‘Hover Point Data’ and move the mouse around (pausing when you want to see pop up window with all the highlighted cell or point data in it).
These are the icons for those hover modes:
image

3 Likes

My buddy Darren (@1318980 ) has put together a nice summary of the differences here: Post-Processing: Data Types

That should help you :slight_smile:

Jousef

3 Likes

Hi DaleKramer,

thank you very much for your answer. The hover toles give the results as the visualisiation, about 1m/s windspeed at the beginning of the windtunnel (left hand site). I set another slice in Y direction and here it becomes why: As the grid size is 10m at the beginning of the windtunnel, 5 m within first refinemnet and down to 0.5 m close to the buildings, I guess that in my horizontal slice, which I set to 1,5 m so it represents human hight, the nearest point data is represented. This would be point data at 0m ( as 1.5m is closer to 0m than 10m). Do you think this makes sense?

Thank you, I will check it out!

2 Likes

Yes I think you have the basic understanding now.
As I think that the red line is your ‘human height’, I can say that differences between cell and point data at the red line will likely be large due to such large cell sizes until you get into the finer mesh on the right.
Also remember that, likely, you have a no-slip wall BC on the ‘ground’, which throw another big monkey wrench into consideration…

2 Likes