open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all

I discovered the strange behavior of ARTEMIS v7p2r3 in random mode.
Wave height falls from wave generation boundary. I found this for my own harbor case but then I made a simple setup. Rectangular basin with flat bottom (depth=12 m). Left boundary is "incident wave" boundary (Tp=6.996 s, H=3.523 m). Top and bottom are walls. Right boundary is an outflow. In monochromatic mode I get approximately uniform distribution of wave height. In random mode (5 component) wave height smoothly decreases from wave generation boundary. This is all the more strange because for lowest spectral component in monochromatic mode I get approximately uniform distribution of wave height.
Does anyone know the reason? Help me please!

Add input files of my test case

I think that problem in breaking dissipation probably. And probably in breaking dissipation in random mode. In monochromatic mode with breaking wave height doesn't fall. And when I turn off breaking in random mode wave height doesn't fall.
But I need breaking in my harbor case.
Dear ARTEMIS users and developers! Help me please with this issue!
Is the new version of ARTEMIS free from this problem?

I still have not solved this problem.
I uploaded my input files, but without a geometry file.
Now I am sharing a link to all input files.

drive.google.com/drive/folders/1hzDBHcFg...j8Bafpz_?usp=sharing

Maybe someone can solve this problem?

Hi all

Very strange reaction from developers and TELEMAC gurus - no reaction .

I tried new versions of ARTEMIS v7p3r1 and v8p0r2 and this problem is still not solved.

If I have big domain with real bathymetry, short waves (but condition "wave lenth > 7 cell size" is satisfied) that in spectral mode wave height drops from wave generation boundary.

I made test case - rectangular domain with flat bottom
depth - 14 m
wave height - 5.22 m
peak period - 6.78 s
gamma - 3.3
number of componets - 5
and have the same - wave height drops from wave generation boundary



If wave lenth is longer then drop is less. And in monochromatic mode there is no drop.



I think that problem is connected with breaking dissipation. Look at QB profile



But may be this is due to another reason.

You can see that in random mode to make drop less we need use much more fine mesh. Condition cell size < wavelength/7 is insufficient. But if we have big domain, that make using of ARTEMIS painful in real cases.

Dear developers and experts, Costas, Jean-Michel, Sebastien, don't ignore my topic please (if you remember I found bugs that prevents to run ARTEMIS with MUMPS solver). ARTEMIS is great tool. Let's make it better.

Hello Maxim,
Sorry for the late reply, I have been so pre-occupied that I totally missed your post. Unfortunately, the ARTEMIS section has been headless since Noemie stepped down. Sebastien has been inactive for quite a long time. Has he departed from the consortium? I don't know. Jean-Michel for sure is enjoying his retirement...
I hope you have already found a solution to your problem. If not, don't worry, we 'll sort it out. I will have a closer look tomorrow morning.
Best Regards,
Costas
PS. Next time, if you are in a hurry, send a PM!

Hello Maxim,

I adapted your case in the latest version and I cannot reproduce your problem. Dissipation between inflow (West) and outflow (East) is ~10%. Increasing water elevation eliminates dissipation.
We can discuss it more when you get this message.
Best Regards,
Costas

Hi Costas,

Thank you for your answer.
I was busy and couldn't reply.
Now I have finished my project and can discuss.

You probably use my previous test case.
Last test case is on drive

drive.google.com/drive/folders/1ofQ9n7ti...4A-Vxz7L?usp=sharing

Hello Maxim,
I am on holidays until the end of this week, so I cannot get fully involved. However, with your latest case I can reproduce your results. I agree with you that breaking is the cause of the dissipation and I am equally puzzled by the results you present (especially with Tp=10s).
My only remark is the discretization range. I would recommend to give ample limits and let ARTEMIS automatically decide the spectrum period boundaries (it divides the spectrum in equal energy bands). Can you run your tests withand see if the results are more consistent?
Best Regards,
Costas

Hello Maxim

I had a quick look at your testcase.

I don't think it's anything wrong with ARTEMIS; I think it's just that the wave condition you're trying to impose (Hs=5.22m Tp=6.78s) is too steep and cannot be sustained so it breaks from the onset.

I've confirmed this using your test files with
a) a lower Hs of 3.4m, same Tp
b) same Hs but longer Tp=15s (and higher water level)

The wave height does not decay in either case.

Hope that helps - Noémie