open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi I have been using TELEMAC 3D to model the effect of a power plant heated discharge into a lake. The lake has two inflow boundary conditions from 2 rivers, a power plant intake and outfall, and a dam that controls the water level. I was able to run it with the hydrostatic version for two months. Now I have been trying to run with the non-hydrostatic version but the simulation fails after the 3rd day due to a problem at the the intake boundary condition:
DEBIMP3D : PROBLEM ON BOUNDARY NUMBER 4
GIVE A VELOCITY PROFILE
IN THE BOUNDARY CONDITIONS FILE
OR CHECK THE WATER DEPTHS

I checked previous post and try to use the suggestions given there, I've refined the mesh, reduced the time step, given a velocity profile based on water depth (option4) but it still does not work. By checking the water depths at the intake just before the simulation stopped I found that they vary from 5-8 m. So there is water in the domain and it is not becoming dry (attached figure). By checking at the depth average velocity magnitude, the values are small (attached figure)

I am not sure if someone has faced a similar problem before. Any help or suggestions are appreciated, I am not sure what else to check in the results to find where exactly the problem is.

I am attaching the output file and my case file. Also I am attaching plots of a vertical cross section of U, V,W and temperature, the region inside the rectangle corresponds to the the intake boundary condition.

Thanks so much

Regards

Viviana

Hello,

I'm leaving for 1 week and can just have a quick glance at your steering file. I would try to put the 3 IMPLICITATION FOR... at 1. instead of 0.6, and you can also have a lower FREE SURFACE GRADIENT COMPATIBILITY. Try to have an idea of the Courant number in your case, it could be that larger time steps would work as well. Anyway the message is strange if there is still water on this boundary, I would need to run it when I come back.

With best regards,

Jean-Michel Hervouet

Hi Jean-Michel

Thanks so much for your quick reply. I changed the simulation using your suggestions but it is still crashing after 2 days and 20 hours, giving the same error message as before. I was checking at the Courant numbers and my maximum U and V velocities are around 1.2 m/s at the power plan discharge area, the minimum mesh size in that boundary condition is 3-4m, so with I might not be able to use a time step higher than 2s to keep the Courant number below 1.

I am still not sure what could be causing the program to stop, maybe is still a problem with my mesh. I am attaching the files I used for the computation in the folder Forum.rar. Hopefully is not to heavy

Thanks so much, I really appreciate your help

Viviana

Hi

I run a new simulation changing the vertical and horizontal turbulence model, I used MIXING LENGTH for the vertical and SMAGORINSKY for the horizontal. However, the simulation stopped after 3 days due to the same problem at the intake boundary.
It is strange that the hydrostatic simulation worked but the non-hydrostatic keeps giving me problems.

I've decided to change the geometry of the intake to see if that helps.

Let me know if you have any other recommendations.

Thanks

Viviana

Hi all

I have been trying to get the non-hydrostatic simulation for a lake with no success. Originally, I included the intake and outflow from a power plant as boundary conditions, then I decided just to run without power plant, to see if that was causing the problem. I have changed my mesh, but nothing has worked.
The simulations stop after 33 hours of simulation:

GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 500 RELATIVE PRECISION: NaN
DEBIMP3D : PROBLEM ON BOUNDARY NUMBER 3
GIVE A VELOCITY PROFILE
IN THE BOUNDARY CONDITIONS FILE
OR CHECK THE WATER DEPTHS

The water depth in that boundary is 0.3m, I think the solver is crushing and exceeding the maximum number of iterations for the diffusion of tracers. But I am not sure how to track what is happening.

I am attaching my case file output file, geometry and boundary conditions

Please let me know if you have experienced similar problem before and got it solved. Any suggestions??

Thanks so much

Viviana

Hello,

I have recently found in another case that a boundary with prescribed discharge and free depth may tend to zero depth if the flow becomes supercritical (in which case all velocities and depth should be given at the boundary, and as depth is not given, it may take fancy values). As a matter of fact the problem on boundary 3 is certainly that the depth becomes 0, and it is strange that this happens after so much time.

Regards,

Jean-Michel Hervouet

Hi

Thanks for your reply. I checked the water depth and froude number at the end of the simulation, and that is not the problem, the water depth is 0.35m and the Froude number is very small less than 0.1.
I am attaching a figure with the result for the last time step, before it stopped working (1D18h).
I am attaching my output file as well. From there I was able to identify that the solver exceeded the maximum number of iterations for PPE and diffusion of tracers.

What I still not fully understand is why the hydrostatic simulation worked with no problem.

Kind Regards

Viviana