[aspect-devel] Convert csv file to .vtu file?

[Nan Zhang]

Hi Rene,

I agree we should forward our discussion to the aspect list!

Yes, you understood my question correctly. Except for the "Delauney"
filter, I saw "Table To Structured Grid" at
http://www.paraview.org/Wiki/ParaView/Data_formats. The webpage gives an
example with 9 row data. I followed their steps and could get a smooth
plot. The problem is a key parameter "extent" in the "Table To Structured
Grid" only works for their example, not for the output csv file from
ASPECT. They use [0 1] for the extent. I tried many numbers in the extent,
and always got error.

A plot for the csv data with "Table to Point" is attached. The
visualization is bad.[image: Inline image 1]

On Wed, Oct 12, 2016 at 1:07 AM, Rene Gassmoeller <
rene.gassmoeller at mailbox.org> wrote:

> Hi Nan,
>
> Thanks for your information. It seems that imposing a lithospheric thermal
> structure has been implemented already. I wonder if ASPECT can apply 45o for
> the slab dip angle on the trench?
>
> No that is currently not possible, unless you are imposing a temperature
> field for the whole model anyway (in which case you could include it in
> your initial condition). But in any case it is currently not implemented to
> impose such conditions over the whole model run (as in the Bower et al
> papers). Wolfgang once created a model that imposed temperatures in a part
> of the model domain, but I do not quite recall how he did that (we should
> move this discussion to the mailing list when discussing how to do that,
> might help other people as well).
>
> On 10/11/2016 10:57 AM, Nan Zhang wrote:
>
> Hi Rene,
>
> How was your vacation?
>
>
> Thanks it was great. Visited many national parks in the Southwest US,
> pretty impressive!
>
>
> I did some data processing by subtracting T from the average temperature,
> to plot the plume structure. When I do the data processing, I save .vtu
> file to .csv file. I finally get the excess temperature file with a
> .csv/ascii format.
>
> Do I understand correctly that you export data from paraview as csv, then
> do some processing and then import them again?
>
>
> The problem is Paraview plots csv file with table to point. There is no
> any smooth/grid between points. The effect is very bad. Have you done any
> conversion from csv to vtu?
>
> I have not done something like that before. The problem here is that when
> exporting from vtu to csv you lose the information about cell connectivity
> between the points, that is why Paraview can not create a mesh between
> them. You can sometimes get something similar by using the "Delauney"
> filters, but they are not as accurate, and in particular in 3D very slow.
> If your model is cheap, you might want to consider rerunning it with the
> 'nonadiabatic temperature' visualization output, or write a postprocessor
> of your own to exactly do what you want to do. Alternatively, you would
> need to find a way to do everything in paraview, or reconstruct the mesh
> afterwards (not sure if that is possible).
>
> Hope that helps,
> Rene
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20161012/2b3a7c74/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: T_excess_TableToPoint.png
Type: image/png
Size: 129047 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20161012/2b3a7c74/attachment-0001.png>