open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

We are setting up a 3D mud transport model using SEDI-3D inside TELEMAC-3D. It is a very nice module and we have a nice model that can reproduce some patterns of sediment movements.

However, while investigating the mass balance in modelled domain, we found the mass is not conserved.

We analyzed the mass balance information given in the listing file by Telemac-3D. At each time step, we have information for sediment bed and suspension, e.g.,

However, after plotting them together, we found the mass continues to disappear over the time period. The plots are attached.

Our model is initialized with 0.05m bed layer and 0.0 suspension concentration. The mass of sediment should be either in the bed or in the suspension. The sediment mass leaving the domain is 3 orders of magnitude smaller than the total mass of bed, so it shouldn't cause this amount of mass lost in the domain.

Does anyone happen to know what is going on here? Does it mean something goes wrong at the water/bed interface?



Best regards,
Qilong

Hello,

I made a simpler model for testing the mass balance of sediment. The model simulates a rectangular tank with no open boundaries, hence, no water movements. The water depth is 10m everywhere.

The first case has only deposition because water doesn't flow. Initial concentration is imposed and there is also a initial bed layer of 10cm.



The second case has only erosion, This is done by modifying the erodc.f and flused.f to have a constant erosion and zero deposition.



The third case has both erosion and deposition.



The fourth case also tests if the initial bed layer thickness is set to zero and a initial concentration is still imposed.



After further investigating, I found out that, if there is no tidal flats in the domain, the mass balance is well preserved.

I will continue with more tests and update the post later.

Best regards,
Qilong

Hello,

We did some more tests yesterday and and we have come to the conclusion that the mass balance is well preserved in SEDI-3D module in Telemac-3D.

The previous confusion is caused by the wrong interpretation of the following information given in the listing file.

We found out that:

ADVECTIVE FLUX THROUGH THE BOUNDARIES is the sediment mass leaving the open boundaries per second;
DIFFUSIVE FLUX + DEPOSITION is actually the (deposited mass - eroded mass) per second;
MASS LEAVING THE DOMAIN DURING THIS TIME STEP is then given by (ADVECTIVE FLUX THROUGH THE BOUNDARIES + DIFFUSIVE FLUX + DEPOSITION) * DT. Its really meaning is the mass leaving the water column per time step.

Hence, we calculate using ADVECTIVE FLUX THROUGH THE BOUNDARIES * listing printout period * time step, to get the approximation of the sediment mass leaving the boundaries and then do the cumulative sum of this quantity. By doing this, we get the same time interval as in the MASS OF BED.

With these correction, we're able to get the right mass balance. However, there is still a bit mass lost in our simulation, and we think it's due to the presence of tidal flats in our domain.



Kind regards,
Qilong