Hi Jean-Francois,
sorry for getting back to you so late on this issue. I found the time to
look at it now.
The unsymmetry is due to the "neumann" function of the problem. There,
the global position is set to be the vertex position:
const GlobalPosition &globalPos =
element.geometry().corner(scvIdx);
This means that at the corners, both the right as well as the
upper/lower boundary segments get the heat flux, while for the other
elements on the right boundary, only the right boundary segments get the
heat flux. This explains the higher temperature in the corners.
One can take care of that by either setting
const GlobalPosition &globalPos =
fvGeometry.boundaryFace[boundaryFaceIdx].ipGlobal;
or using the "neumannAtPos" function which directly gets the coordinate
of the integration point. That way, only the right boundary segments get
the heat flux. I corrected it in the svn repository.
It went unnoticed since we always used only one element in y direction
for this problem.
Thank you for bringing this to our attention. Kind regards
Bernd
On 12/22/2014 01:55 PM, [email protected] wrote:
dear All
I have stumbled onto something I currently dont understand and which prevent me
to move forward.
Because it can be of general interrest, I take the liberty to share it with you:
it seems that Boundary submitted to constant heat flux and expected to remains
at constant temperature in the context of defined case exhibits some small
temperature gradient that can be amplified in some circumstances.
In order to observe this problem which I initially found on case I build
otherwise [ more on that later] , you can do the following:
Start from the heatpipe case in dumux lecture
Create some meshes along the y axis [ I put 10 meshes].
Run the case and observe the temperature profile on the right"heating" boundary
You should observe a parabolic temperature with maximum on the edges and
minimum on the middle
My understanding is that the expected profile for this case should be a
constant temperature.
For this case the amplitude of the parabolla is very small and of no practical
consequences at all.
But, it increases with the heatflux and when near the critical flux becomes
the seed for anomalous heating that make the resulst useless.[ because the
edge are slightly hoter than the middle there is slightly more vapor which
means the temperature increase more .. amd so forth...
You observe this effect by increasing the heat flux in the previous case from
100W/m2 to the heat flux 4kw/m2 the amplitudes of the temperature difference
increase and is not engligible anymore [by far] around 225000 s.
That said, I am perfectly aware that the critical flux is an instability point
both in computation and in real life. This example alone is probably not
sufficient to point to a serious problem.
I propose it for illustration purpose after I back engineered my way to it from
cases I build in 3p3c where, apparently because of more complex
phase/components interactions, this effect is further amplified to the point
where low heat fluxes become supercritical creating localized pure vapor area
and very hot zone at these corner points where none should exist.
Back to dumux, this situation is not specific to BC "discontinuity" point as I
initially thought.
I have build examples where the heating surface is a square or a circle located
in the middle of the porous material area [ so no BC discontinuities exist at
all every connected boundary has the same bc] and observe these temperature
profiles on the heated boundary otherwise expected to be constant.
This predictably leads to the same observed anomalous heating. These points are
always geometrically located [ aka not randomly placed]
What do yo think of this ...?
Am I missing something?
Is this a known limitation in dumux?
if yes, is there a workaround?
Happy Holliday to you all!!
_______________________________________________
Dumux mailing list
[email protected]
https://listserv.uni-stuttgart.de/mailman/listinfo/dumux
--
_______________________________________________________________
Bernd Flemisch phone: +49 711 685 69162
IWS, Universität Stuttgart fax: +49 711 685 60430
Pfaffenwaldring 61 email: [email protected]
D-70569 Stuttgart url: www.hydrosys.uni-stuttgart.de
_______________________________________________________________
_______________________________________________
Dumux mailing list
[email protected]
https://listserv.uni-stuttgart.de/mailman/listinfo/dumux
