open TELEMAC-MASCARET The mathematically superior suite of solvers

Hallo,

I really need suggestions for a case that I am lately studing: drought case (1000 m3/s mean discharge and 300 m water elevation) in a bay of 20 km length, considering a water transfer channel of a ca. 100 m width, that will take oper. 10 m3/s and max. 28 m3/s from the bay.

The boundaries of the bay near the channel are dry, so I am using tidal flats. Since I cannot open a boundary if I do not have water, I had to modify the bathymetry and so the mesh. My mesh has a mean edge length of 100 m. The channel is also 100 m width. I should modify it in this range, opening at least 3-5 nodes (ca. 20 m edge length), bot it is really difficult with Janet, to make it gradually uniform. So the best option of meah modification is the one attached (I modified a wider area - around 500 m width and ca. 1 km length).

In my cas file I have:
- duration: 5 days (steady state in the reference case after ca. 3,5 days)
- time step: 30 s
- BC: inflow: 1000 m3/s, outflow 300.3 m, channel -28 m3/s

I attach you a figure of the results of the drought case, without the channel, so you could check it. Also, I attach you some figures related to the channel simulation, considering the wider modification of the mesh (also attached). The results of the "thinner" channel have much higher numerical problems (not stable at all) -- I do not even report them.

What I can still try:
- modify BC at the channel imposing also a water elevation of 300.3 m --- but then I cannot differenciate the case human consumption 10 m3/s and max. 28 m3/s, taken from the channel
- diminish the time step (e.g. 10 s ?)

I have no idea anymore, any suggestion is welcome :)

Thank you.

Best regards

Hello

I'm not able to understand your post. However, clearly you have some instability issues. To overcome this kind of problems here some hints:
- verify if liquid boundaries are always wet (enough)
- use smaller time step, verify if the solver is working confortably (convergence in less than 40 iterations more or less)
- try different values for key-word FREE SURFACE GRADIENT COMPATIBILITY, values can be between 0 and 1
- try to smooth your bathymetry in case of steep bathymetry gradient
- you can add more diffusion with VELOCITY DIFFUSIVITY

If you issues persist, please explain them more clearly; otherwise send us your model.

I hope that this helps
with my best regards
Riadh