open TELEMAC-MASCARET The mathematically superior suite of solvers

Problems with computing volumetric flow rate (discharge) in Blue Kenue.

I am obtaining unacceptable results in flow determination by integrating the discharge per unit width (m^2/s) along a line across the channel.

I have a trapezoidal channel, and set upstream and downstream water surface elevations. The Telemac2d simulation is run sufficiently long so that steady-state conditions are obtained and transients from initial conditions are diminished.

I only want to compute the discharge at the last frame, so it should be a quick, straightforward computation. I draw a line transversally across the channel, and choose to integrate along the line in Blue Kenue.

Now I also compute the discharge for these conditions using two independent open channel flow programs which are in agreement with each other, but not with the Blue Kenue results.

Further complicating matters, when I integrate depth across the channel in Blue Kenue, I obtain an acceptable area of flow commensurate with the two independent open channel flow programs – taking this area and multiplying it by the average velocity in the channel (in the last frame), I obtain a volumetric discharge similar to the two independent open channel flow programs but inconsistent with integrating the Telemac2d discharge per width in Blue Kenue.

Has anyone else experienced a problem in determining volumetric flow (m^3/s) by integrating the discharge per unit width (m^2/s) along a line drawn across the channel?

Hi,
Did you resolve your problem?
Martin

Here is the "work around" I used to resolve the problem.

For trapezoidal channel:
At a reach location removed from boundary influences where WSEL is fairly constant, obtain a depth in the main channel by clicking on points in BK output view, determine top width across channel at that depth via geometry, integrate velocity across channel, divide this value by top width determined earlier to obtain a representative mean flow speed (scalar), integrate depth across channel in BK to obtain flow area, and finally multiply velocity by cross sectional area to obtain flow.

This procedure gave a discharge within a few per cent of the 2 independent open channel flow programs - integrating flow per unit width in BK simply gave unacceptable values.

How many elements do you have across the channel?
I'm trying to visualize the situation. Could you post a picture?

Martin

Martin,
This is the basic channel concept/configuration. I can send you the .res file if you need it. Thanks. John

Sure, post the result file and the cross section line you are using

Martin, here's the file...I re-ran it so the the graphical output has only the last few frames...it was run sufficiently removed from IC such that it should be steady state, so it is now a steady, gradually varied flow problem (u/s 50 m, d/s 49 m) Integrating any line across the channel gives about 418,000 m^3/s which is incorrect IMO. Manning n is 0.035 and channel side slopes are 33.333 (h:v), bottom width is 1000 m.
Thanks, John