open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello

I am actually workin with tomawac in a propagation model in the north sea near dunkerque. I ran it with a jonswap spectrum 6. I tried,it with a significant height of 6 meters a directionnal spread factor of 20.

I have two problems with the results.
Firstly the Hs in the boundary is never set as 6 meters but 4 meters.

The direction of the wave changes near the free boundaries. On the left size the wave should turn in the other size due to the sand bar. It looks like there is a waterfall near the boundaries.

you can find the results attached

thanks for your help

Hello,

the directional spread factor that you impose in the steering file is very high : the wave spectrum is very directional (waves are not spreaded in direction) and that could imply problems in the spectrum discretization (i.e. in the spectral parameters calculation - such as Hm0), if the spectrum frequency and directional resolution is low. This could be a reason that explains your results.

Over which boundaries of your domain did you impose a Jonswap spectrum ? Are the right and left boundaries set as free boundaries, or do you impose the spectrum also there ?
What is the dimension of your spatial domain ?

I'll be able to answer you more correctly if you let me know those info.

Kind regards,

Giovanni

hello

I tried a lot of cases and for a less directionnal spectrum, there is no more pb with HM0.But I would like to apply a real swell and not wind waves. I thought that swell is very directionnal.

The right and left boundaries are free boundaries. As you can see there is a lot of sand bars and the physical reality of the boundary condition exists only for the top boundaries. domain is about 10 km long.

thank you veery much for your help !

Hello,

here below I give you some suggestions to improve your model :

- concerning the Hm0 value problem, you have to check the frequency interval and the frequency resolution you define in the steering file : it can be that you obtain a Hm0 lower than the value that you impose because the lowest frequency you consider in the spectrum is not low enough compared to the frequency peak. (Rule of thumb : if you have a peak period of about 15s, take the lowest frequency equal to 0.035Hz).

- concerning the model you implemented, you have to pay attention when you impose a free boundary condition onn a liquid boundaries: that means that all the wave components propagated by the upper boundary that reach those boundaries, are completely absorbed; free boundaries do no supply any wave energy into the domain. A more realistic approach consist of imposing over all the liquid boundaries sea-state conditions (i.e. only the shore is represented as a free absorbing boundary).

Your spatial domain should be centered on the area you want to study:
- if you are interested in a specific are of your spatial domain, then be sure that the boundaries are not set too close to that area.
- if you need to analyze the whole spatial domain, then I suggest you to increase the model width, to allow the sea-state to be properly propagated over the domain and to reduce the influence of the boundaries.

Hope it helps.

Kind regards,

Giovanni

Hello,
I have some questions about what you said.

giovanni.mattarolo wrote:
You the lowest frequency = 1/15? But how could I determine the right peak period? Should I have to try with some values and check which one is ok ?



giovanni.mattarolo wrote:
If I understood :

Free boundary = waves are not absorbed (like if it is water), and no initial(& limit) condition

Solid boundary = Sand / beach

But I dont really understand when you said "liquid boundaries", how do you set that in Tomawac? Beacause in Tomawac there are only 3 type of boundary :
Free boundary
Solid Boundary
Boundary with value impose

Like on my picture:

kroptia :

liquid boudary : -free boudary
- boudary with imposed value

solid boudary : shore

free boudary : the wave energy is fully absorbed, like in a waterfall, it implies and change in the direction


giovanni.mattarolo :

Thank you very much for the help but I checked the frequency interval before sending the post, the fact is it works when the directionnal spread is from 1 to 10, after the Hm0 is decreasing. I tried to launch it with a TMA spectrum and the result is worse : for all the directionnal spread, Hm0 loses 2 meters in the boundaries !

I don't understand the difference between a free liquid boudary and a solid one because all the energy is absorbed for this two boudaries.

To put a hm0 of 6 meters in the right and left boudaries where the water depht is sometimes less than 3 meters is not a good way to have good results, but puting free boundary condition is not better. Are you sure the only solution is to have a larger grid ?

Thank you very much !