open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,
I am simulating sediment. I have 2 boundaries. I have given the value of the boundaries to the model using the liquid boundary file. I have also done this for hydrodynamic modeling in Telemac 2D by adding some subroutines(sl,bord,read_bin_frliq.f).
When running my example in Gaia, I face the following problem:
1)solution x=0
2)flux boundary 1,2 = -0.0000
Where is the problem?
Thankful
Sina

Hello Sina,
for an analysis of your problem it would helpful to have the cas-files, the cli-file (both if you have a different one for Gaia), your changed subroutines and part of the listing file.
Best regards,
Rebekka

Hello Rebekka.
Thank you very much for your reply.
I had attached it, but I don't know why it wasn't uploaded.
I am attaching the file.
Thanks
Sina

Hello Sina,

Neither water nor sediment come into your boundaries so nothing needs to be computed which explains the "L2-norm very small-message".
I think it is due to the wrong format of the liquid boundary file. Additionally you apply the water depth at both boundaries which is not necessarily stable. Typically you would give discharge or velocities at one boundary and water depth/water level at the other. The liquid boundary file need then the following lines if the 1st boundary is the discharge boundary and the second the water level:
T Q ( 1 ) SL ( 2 )
s m3/ s m
This is described in the user manual section 4.2.5
So instead of ETA and the node number you need SL(1) and SL(2) if you stick to 2 water level boundaries. So the water level (or the water depth) at the open boundary is constant. If you want to have this different you need to program it in the user_Q.f / user_vit.f / user_SL.f subroutines. Furthermore you give a zero concentration at one boundary and zero concentration as initial value. So there will be no sediment despite from erosion.
It is a little tricky to find out which open boundary is the first one. The easiest way is to start a serial run. Then it is printed in the listing (Number of open boundary and the coordinates).

I always start with time constant boundaries using PRESCRIBED FLOWRATES/ PRESCRIBED ELEVATIONS ... in order to see if my configuration works in principle.

I hope this helps for the next step.
Best regards,
Rebekka

Hi Rebekka,
Thank you very much for your reply.
I want the LIQUID BOUNDARIES FILE to be node based. So they offered me post number #749. In telemac2d, I modified the (Sl) as follows(The file is attached).
Now I make the same changes in telemac3d(sl3), but I get an error, what is the problem?

(When I value the expressions INITIAL ELEVATION the problem is solved).

I had another question:
When I run, after a few hours I check the results, I face a problem. I attach the photo.

Thanks in advance for your reply.
sina

Hi Sina,

you have to be careful about the arrays you are using. FCT has originally only a length of 9 characters and it is still only 9 characters in read_fr_liq.f. So you need to change it there as well. But read_fic_frliq will also called by q3.f/vit3.f/tr3.f so be sure that these functions will not used or better call an exclusive function in sl3.f. Furthermore if you are running the simulation in parallel you have to care about that. The N in bord.f / bord3d.f is the global node number but of the parted system. So N will be different in a seriel run and a parallel run. (I am not absolutely sure, so you should check that.) In case I am correct you need to transfer the node number to the serial one with MESH%KNOLG%I(N).

That it is working in 2D it is "luck", because overwriting of an array will not necessarily give an error. And maybe there is another error I have not detected.

To the figures I can not say something because I only see that there is something not correct. Is it a 3D plot? only the grid??

Best regards,
Rebekka

Hello again,
Thank you for your answer rebekka.
I have this problem when I open the 3D view in bluekenue.
Where do you think the problem is?
thanks
sina