open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello all,

Since v7p2 sediment discharge through control sections is done with the FLUXLINE option, that requires a separate input file to define the section locations. I tried to use the old control sections input file format but it does not work. The user documentation, although it refers to the new method, it lacks a description of the new file format.

It would be great if a description and an example is added to the docs.

In the meantime, if I understood correctly from flusec_sis.f, the new format is:Correct?

Best Regards,
Costas

Hi Costas,

sorry that the documentation is not finished yet. We have a validation example but it is still not on the trunk

You can have a look at the users proceedings of the "TELEMAC-MASCARET User Conference 2015" there is some documentation. You can check the definition and use of Fluxlines with the Yen example.

Please find attached the fluxline.dat file and a script to plot the defined fluxlines.

Best regards

Leo

Edit, I had some problem to upload the fluxline.dat so I copied it directly

The first line contains the number of fluxlines.
Each fluxline is defined by two points and a bounding box. Only fluxes at segments inside the bounding box are used to compute the discharge

Hello Leo,

Thank you for the script and your explanation, I wouldn't be able to use it otherwise. You definitely need to complete the documentation or else this function is useless to new users. At least make a quick reference to the TUC Proceedings until the full documentation is ready.

Having been through the paper, I have few questions:

1. Do the FLUXLINEs have to span across the model boundaries?
2. Do I need to define the bounding box at all cases or is it optional? (I see a commented out section in flusec_sis.f that could relate to the bounding box)

Regards,
Costas

Hi Costas,

every fluxlines is defined by two point of the line, this are not the start/end points of the line. For algorithm in Sisyphe, the line has no start and no end. In many cases this does not matter, but if you have a curved channel or if you are only interested in a small part of the domain you have to set some restriction. This is why we need a bounding box. This bounding box must be always defined.

Maybe you can copy some code from the following script to parse the results



Best regards

Leo

Hello all,

The new FLUXLINE method works fine, but I miss the file output instead of the listing output. According to fluxpr_sisyphe, it looks quite a tedious process to implement. I will try to use the python code to parse the logs, but my case is not as straightforward as I have hot-starts and chaining loops that end up in tens of runs...

In the meantime, what are the shortcomings of the previous method? I might have to rely on it in the end.

Best Regards,
Costas

cyamin wrote:
Hi Costas,

yes, it would be nice to have a single outputfile for the fluxlines. This is a good point for the further development. We will work on the fluxline during our next Telemac coding week at BAW.

The previous method was less accurate. The fluxline method uses exactly the same flux as used in Sisyphe for the balance computations. This means that we do not introduce some errors due different kind of flux approximations.

The main advantage is, that the method also accounts for the limiting of sediment discharge (positive depths) when not enough sediment is at an node/element. The old method gives only the potential discharge. This can lead to large differences/errors.

In my opinion it should be best practice to ensure that the mass balance is correct during modeling. This can be done with a combination of fluxlines and "mass boxes". A mass box is a part of the domain which is surrounded by fluxlines or the boundary. After your computation you can set up a local mass balance for parts of your domain or the whole domain.

Best regards

Leo

Hello Leo,

Thank you for your explanation.
Lufia wrote: If this is the case, then the old method is a no-go as I have thin sediment layers and the positive depths algorithm works overtime. I guess I should get going with employing the parser.

Regards,
Costas

Hi guys,
Many thanks for this. I've updated the documentation (chapter HowTo?) with this information.
Please don't hesitate to modify/complete/etc. the documentation accordingly.

Cheers,

Pablo