open TELEMAC-MASCARET The mathematically superior suite of solvers

I would like to consider suspended transport aswell, but my overnight calculation crashed. Your analysis was corrent Costas, the problemed emerged later on in the calculations. This puts my feet back on the ground.

I'll start by evaluating my mesh and start searching for individual nodes that could lead to large magnitudes. Do you think that a too large time step could case the hydrodynamic updating to crash?

Hello,
1. Take a closer look to your solution before crashing, and try to identify 'critical' zones, e.g. with an excessive quantity of deposited sediment. Probably these zones need some refinement (especially if they are located nearby or close to structures, etc.).
2. Launch a sensitivity analysis to see the time step influence. Due to the non-linearity of the involved processes, the choice of the time step for sediment transport/morphodynamics could not be as evident as for a hydrodynamics-only solution (furthermore, you are also accounting for waves in your sediment transport model).

Please send us your feedback.

Best wishes,

Pablo

Hello again,

I decreased the time step for both the t3d and sisyphe model, and managed to retrieve some results. My mesh covers a large beach area both north and south of the harbour pier I'm investigating, and it seems like the 'critical zones' are in the direct vicinity of the piers.

The bed evolution is extremely high for this, rather moderate scenario with 15m/s wind from NW and an offshore wave height of approximately 3 m. You can see this in the attached bed-evolution figure. I guess the model crashes (after 5h) because of the extreme volumes of deposited sediment.

I tried to look into the option of using STEADY CASE = YES and calling a T3D-Tomawac result file, but I (once again) learnt that this chaining can only be done with T2D.

Any ideas regarding the large sediment movement?

/schmirre

Hello Schmirre,

These are big changes in bottom elevation! You will definitely need to couple Tomawac. What are the depths in the area of large erosion/deposition?

It will also good to have a look at your cas files and your mesh structure, if we are to give a more in-depth advice...

Best Regards,
Costas

Exactly, the depths in the area of ero/dep is around 10-14m, so of course these results are way out.

I can upload the cas-files here. Maybe I can send you a private link to the mesh?

/schmirre

Now beware - this setup does not include the initialization-period, which I run through computation continued with the T3D-Tomawac coupled model.

Can this explain the extreme rates?

Hello,
It would be also useful to take look to the shear stress/shear velocity distribution (of your hydrodynamic-only solution) to have a preliminary idea of the (morphological) behavior of your system.

Pablo

You can send a PM or post an image of the mesh with depths. A few remarks:
Tomawac:

• Timestep is too large for braking waves. Make sure this rule is obeyed (extract from another post in the forum, I keep it in hand):
• Consider coupling Tomawac. Such big evolutions do not justify steady wave conditions

Sisyphe:

• Timestep is also big (10s). See if reducing it softens the evolution.
• Make use of SLOPE EFFECT to 'smooth' evolution by using a mild FRICTION ANGLE OF THE SEDIMENT
• Since you are using suspension, choose ADVECTION SCHEME 14 (similarly to Telemac3D)

Telemac:

• Try reducing the coupling period with Sisyphe. Work your way up from 1

Your mesh is quite coarse. See that it justifies the Courrant criterion for TOM.

Regards,
Costas

Thank you Costas for very valuable feedback. I will dig into this and keep you updated with my progress.

/schmirre

Hi again!

I have gone through your feedback and I'm once again trying to make my modelling reliable.

However, after considering your changes I get stuck with reading wave induced currents. I have specified the frame I'd like to extract results from but I keep getting following error:

@STREAMLINE::SCARACT: THE NUMBER OF TRACEBACK INTERFACE CROSSINGS IGEN

I have tried to go through similar forum posts without luck. Do you have any ideas?

/schmirre