open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear all,

I'm trying to set up an artemis model for the first time. I got some experience in telemac 2d but not in the other modules. I got the artemis documentation and did, as far as I can see, everything following the manual, trying to set up a first simple steering file. I created steering files manually as well as through fdaa with no difference in results. Every time I get the following error:

---

PERIOD 1/ 5 : 1.0841 SECONDS

CONDIH : WARNING !! YOU REACHED MINIMUM THRESHOLD
FOR WATER DEPTH (1 CM). CHECK THE
BATHYMETRY OR INITIAL CONDITIONS

SUB-ITERATION NUMBER : 1

CNTPRE (ARTEMIS) : DIAGONAL SCALING NOT APPLIED (ONE COEFFICIENT OF THE MATRIX DIAGONAL IS NEGATIVE OR ZERO)

LINEAR SYSTEM SOLVING (SOLVE)

DIRECT SYSTEM SOLVER
DIFF. BETWEEN TWO SUB-ITERATIONS (%) 100000000000.00000

---

I created my geometry the same way I used to do it for a normal telemac2d simulation:
Polygon->Create mesh-> Map xyz Bathy->Interpolate->create new selafin-> add variable-> choose BOTTOM values from the before created mesh.

So actually there is a Bathymetry in the Geo.slf file and the module should be able to calculate the water depth from 'ZERO ELEVATION'(Sea surface) and the BOTTOM.

Am I doing anything wrong?
Is there anymore documentation on artemis?

Any help would be much appreciated.

Cheers Tim

P.S. attached you can find a example of my .cas

Hi Tim,

Did you checked that your bathymetry is negative? The error message means you have a too small water depth.

Is it a bassin scale model or a real scale model?

You could look a the ARTEMIS examples to solve your problem : in general, a "constant elevation" is used and associated to "initial water level". Playing with that, you should be able to determine if you have a bathymetry/water level problem or not...

regards,

Christophe

Hi Christophe,

thank you for your reply.

It is a real scale model of a part of the German western Baltic Sea. Therefore I thought I could use the regular shoreline + a bathymetry of the baltic. Additionally I set the harbor shoreline within the mesh to +3.5m which is the value in reality, so I got sloped boundaries within the port which come quite close to what is real and a naturally sloped beach besides the harbor area.

I thought this might be the problem, as artemis doesn't support the tidal flats like telemac2d does. Am I right if I ignore the landwards elevation values. Because if I have +3.5m in the harbour, I get exactly those dry areas which are not supported, as I suppose.

From the error it's quit clear it is a bathymetry/water level problem.

So, questioning the other way around, is it representative creating a mesh including a open boundary and having a solid one. Which then has no further elevation informations from land? How do you represent structures within a port?

I'll try to ignore the landward elevations and see it just as a solid boundary and run it again.

Besides I attached my Geo so you can see what I mean. The area in high resolution is the port.

Regards,
Tim

Sorry somehow it won't attach my selafin geometry file

Hi Tim,

You're right, ARTEMIS doesn't support dry areas. This is probably the reason why your simulation doesn't work.
You will need to remove dry aeras from the mesh.

You can represent structures in the port with solid bourbaries, considering their intersection with the free surface at the water level you specify.

Regards,

Christophe

Christophe hi,

generally it seems to work now, thanks for your advice. But I've been wondering about some more points.

Like I said before, at the moment I'm trying to keep everything as simple as possible to get a first touch of the artemis module. That means I try to run monochromatic waves.
Therefore I didn't specified anything in the borh.f FORTRAN subroutine and didn't use it in my steering. In result I got no Waves in my .res file. Therefore I changed my steering and adopted some simple modifications in the borh.f subroutine. Which were HB=3.5 and the direction to 270.

Running it again (it was pretty fast, compared with my experience from more complicated telemac models), produced some outputs. But It looks like it computed just one wave. A wave like elevation seems to be just on my open boundary. Most parts in the domain are completely flat.

Is there any option to determine over how many waves or a certain time the model should compute?

Cheers,

Tim

Hi Tim,

Your Borh modification seems good, in ARTEMIS you have to play with borh.f in any case, at least to impose the angle of attack of waves on boundaries.

Artemis is a frequency domain software, so there is no concept of number of waves, simulation time, etc... In regular waves, artemis provide an harmonic solution :
eta=amplitude * cos(kx - wt)
The free surface given as an output is the free surface for a given instant , says 0 for example.
If you only observe "one wave", that could means that your mesh is not refined enough : you need at least 7 to 10 elements for 1 wavelength. If the mesh is too rough, then numerical dissipation can damp you incident wave...
=> idea 1 : check your mesh
=> idea 2 : if it doesn't not work, you can try to remove any dissipation term and see what happend.
=> idea 3 : you can also check your incident wave direction, given in the global referential and not in the element referential as angle of attack

regards,

Christophe

Hi Christoph,

that makes sense now. I set up my mesh like i did in telemac, means I had quite a rough resolution in the outer parts and refined it towards the harbour. For my incident waves it is likely to have a a wavelength up to 80 m. So as my mesh has a resolution of 100 meter in the outer parts this makes sense then there is no real wave signal as there will be too much damping.

Your point , regarding wave direction, how do I give a global direction? I thought I already did by using the keywords DIRECTION OF WAVE PROPAGATION and the definition in both.f of
TETAB%R(:)=270


Cheers,
Tim