open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,

I'm new to Telemac 3D and I am currently attempting an ostensibly simple calibration for a river.

However, I've encountered two problems:

a)my vectors don't appear to follow the flow (I'm using Tecplot 360 to visualise, but I assume this is an outcome of the data rather than any issues with the visualisation - see image attached),

b) despite a single inflow/outflow (i.e. continuity) discharges vary across the channel.

I imagine there could be any number of reasons for a and b, however I'm struggling to find a decent place to start.

If anyone has a suggestion of what could be causing either I'd greatly appreciate it.

Many thanks and kind regards,
Simon

Hello,

An answer on question b:

1) you may have oscillations in your domain, like a seiche, especially if you apply the boundary conditions in a still basin, so it takes time to settle. Another point is that the computation of discharges through control sections is not exactly compatible to what does the program (for variational formulation reasons), so there may be a few % or error. A strictly compatible computation will be provided in the next version (7.2).

With best regards,

Jean-Michel Hervouet

Hi Jean-Michel,

Thank-you for your quick reply.

The error persists despite my attempts to ameliorate it. Do you know what % of error I might expect? The discharges appear to vary fairly widely.
I believe this stems from some large fluctuations in velocity, as the cross-sectional area of the channel doesn't vary greatly enough to explain such differences in discharges. As there's only one inlet and a general lack of variation in topography (e.g. no friction terms, no changes in density etc) would it be safe to assume that it might be a problem in the mesh?

Apologies for the basic nature of these questions, I'm still educating myself.

Many thanks,
Simon

Hello,

You can post your steering file so that we can see if there is something fishy in it. Another possibility would be to inspire from the example called "canal" which seems to be about the same as what you try to do.

With best regards,

Jean-Michel Hervouet

Hi,

I've attached the steering file. I'm currently attempting a "clean" run, hence a lack of friction coefficient etc.

Thankyou for the advice re: "canal".

Many thanks,

Simon

Hello,

Ok thank you. In view of your steering file I would prefer that you adapt the example canal to your case, to start with. For example the default law of friction is 2 (Chézy) and the default friction coefficient is 60., so you do have friction in your case.
Then you must think of the time step adapted to your case, try to have a time step such that velocity * (mesh size) / DT is around 1 as a guess work.

The example canal is a computation continued because setting all of a sudden the velocity at the entrance in a still basin may trigger waves that take time to vanish. To avoid doing this you can have a FORTRAN FILE in your case, like t3d_Channel.f and you can put inside function Q3 taken in the telemac-3d library of sources. Then you change the line :

Q3=DEBIMP(I)

into:

Q3=DEBIMP(I)*MIN(TIME,100.D0)/100.D0

this will progressively increase the discharge at the entrance from 0 to your value in 100 seconds.

With best regards,

Jean-Michel Hervouet

Hello,

Returning to my model I've followed your advice. However, I get an error which I believe stems from the 2D continuation where there's a sudden fluctuation in velocity at the beginning.
I thought this might be a visualisation error in Tecplot 360 however turning on Mass Balance reduces the error.

At a guess, I don't suppose an expert eye could see why the output looks like this?

Many thanks,
Simon

Hello,

What are you showing here and what is the scale ?

Mass balance =yes is only for printing the mass error on the listing so it should not have any influence on the results, so I am surprised by what you say.

With best regards,

JMH