open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I'm currently working on creating a model for the Davao Gulf in the Philippines. Right now I am calibrating and validating my results and for the free surface values I have obtained quite accurate results, but my results for the velocities are quite low and do not match the trend observed in the actual data. I have mostly tried changing my friction coefficients, trying out tide-generating forces, and playing with my boundary conditions (I have only tried using prescribed h and prescribed h with UV). And, I am using the TPXO database for my initial boundary conditions and areas of interest for my study are quite far from the open boundary. Can you guys suggest ways for me to improve the model ?

Also, I am validating the velocity by only taking in the highest velocity values for each hour for the duration of the simulation since it appears that the actual data only took the highest hourly values for the area.

I have also attached my files and some of my results.

Thank you so much !

Hello frostyrobin,

Why do you use a Chezy friction law? Have you tried to use another friction law, like Strickler, Manning or Nikuradse?

If tuning the friction coefficient is not enough, you can increase the
You can increase the COEFFICIENT TO CALIBRATE TIDAL VELOCITIES, e.g. a value of 1.2 should change the velocity inside the domain.

Looking at your steering file, I would recommend you:
- not to use VARIABLE TIME-STEP and DESIRED COURANT NUMBER keywords when using the finite elements method in TELEMAC-2D (default value),
- to add CORIOLIS = YES + fill in CORIOLIS COEFFICIENT by calculating it depending a mean latitude of the domain and with the help of the T2D reference file, dictionary or T2D user file.

Looking at your open boundary (South), I think you should regenerate the mesh and avoid obtus angles + too fast transition between elements.

Can you upload the boundary conditions file?

Hope this helps,

Chi-Tuan

Hello Chi-Tuan,

Thank you for taking the time to take a look at my questions.

I have tried using other friction laws such as Strickler, and Manning. The resulting values were quite similar with Chezy, so I decided to stick with Chezy while i'm trying to change other parameters.

Also, in prior simulations I have tried changing the COEFFICIENT TO CALIBRATE TIDAL VELOCITIES, but the changes were quite small. I am curious as to how the coefficient is applied to the tidal velocities, is it like a correction factor that you just multiply to the velocities?

Thank you for your recommendations and I will implement them on my next simulation. But I would like to ask why are the VARIABLE TIME-STEP and DESIRED COURANT NUMBER keywords not recommended? Is it because it adds computational time?

I have attached two boundary conditions, can you guide me as to which one is better or what improvements I can make? The first one is the boundary conditions for the initial geometry file that I uploaded, while the second one shows an updated mesh and geometry. In the second one, I tried to push back the open boundaries farther from the area of interest.

For the mean time, I will try to run a simulation with the second BC using your recommendations.

Thanks again for your time and recommendations.

Additional Files