open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello, everyone!
I am simulating a jet current case. Thx to Mr Coulet's help, I have set the velocity profile and sediment concentration profile correctly. But for the result of simulation, I have several doubts. I have calculated the horizontal velocity by hand, which is around 0.0844m/s. Considering that the channel is around 6.72m, so the time for the sediment to go through the whole channel should be greater than 79.4s. In this case, I have simulated a duration of 160s. But from the final plots, it seems that the sediment is still in the middle of the channel and haven't arrived the outlet. But according to the experiment results, there will be a sediment layer whose thickness is around 0.14mm at the outlet. So I have these doubts according to the final plots.

1. If I want to accelerate the movement of sediment, what kind of keywords should I change? I tried to set no friction at the bottom and that had slight differences.
2. Why the sediment in the middle move quicker than that at the lateral sides? I have never set friction at two sides, so I thought the contours in 2D should be straight parellel.

I have attached the files for calculation. If you have some ideas for these questions, great thx for leading me.

Hello,

You start from velocity = 0, so entering the discharge you trigger a wave at the entrance and the velocity you expect will take time to settle, this is why the sediment is seemingly delayed.

For the sediment which goes faster in the middle, it may be due to your mesh which is much finer near the walls, probably giving a different numerical behaviour... you are testing the sensitivity to the mesh size...

With best regards,

Jean-Michel Hervouet

Thx for your explanation!
I will try them today.
But can you tell me whether there is other method besides 'hotstart' to get the velocity wanted?
THX!

Hello,

You can initialise the velocity in subroutine condim.f, however it is easy to do in 2D in a canal, mut you would have to put some sort of logarithmic profile in 3D. Anyway what you could do is a preliminary computation without sediment to get an established velocity. Then a computation continued starting with sediment.

With best regards,

Jean-Michel Hervouet