open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi Folks,

I am not sure if my modeled bridge is OK for a telemac2d-sim. Hopefully, someone can tell me if this procedure is right:

Based on the bridge example (v6.3.2) I created 10 pipes crossing the bridge. Lets assume the bridge has a width of 20, height 5 and depth 10 m. I set the coefficients which are needed for the ASCII input file as follows:

CE1, CE2, CS1 and CS2 to a low value, like 0.1. The value from the example are very high.

Lrg (the width of bridge/pipes) = 20/10 = 2 m for each pipe

Hau (height) = 5 m;

Clp = 0 (flow in both direction)

L12 -> ks = 0.2 m, D ~ 2*2+2*5= 14;
ks/D=0.2/14 = 0.014 with moody diagram gives for lambda: 0.045;
finally with lambda*depth/D=0.045*10/14 = 0.03 for L12

z1 and z2 will bet set to the height of the selected river bed nodes for the pipes (see green dots in pic ). I assume z1 & z2 are based on the symmetry line of the pipe?

Also, instead of a dam for the bridge as shown in the bridge example, I blanked that part out.

Is that a appropriate way to model a bridge?


Doeby

Hello Doeby

When looking at your mesh, you are not really modelling a bridge, you are modelling rather a culvert than a bridge.
The bridge is introduced to the model in order to allow both free surface and internal flows. That's why, the way to do is to include it as a highly elevated nodes (with the real elevation of the bridge and its embankments) and then include culverts underneath in order to let flow between both sides of the bridge.
If I have well understood, your bridge is included as an island. In this way, the free surface flow can't be allowed.

I hope that this helps

With my best regards
Riadh ATA

Hi Riadh,

thanks for your answer! I agree with you that I should lift the nodes to the level of the bridge. In my case here I didn't care about the flow OVER the bridge, therefore I used an "island" with the identical size (length x width) of the bridge. Nevertheless, as long as my water level is lower than z1+Hau I have free surface flow in my tubes otherwise pressure flow, right? Also, Lrg represents the width of the bridge and should be modeled by multiple tubes to get a rather smooth flow instead of a single sink/source?

Thanks for your help and greetings form Austria, Doeby

Hello Doeby

yes you will have free surface as long as free surface is under max(z1,z2)+Hau.
However, the Lrg it the width of ONE tube, so if your bridge is very large you need multiple tubes in order to get smooth physical flowrate crossing the bridge.

you can look at the subroutine buse.f (in forder sources/telemac2d) and there you can have all details about how the flow is managed.

With my best regards

Riadh

Hi Riadh,

my simulation - including 10 bridges which are modelled with ~200 tubes - runs nicely up to a time of 5000. Then I get stability problems and some of the tubes have unreasonable high discharges. I tried to decrease the courant-number to 0.1 and also decreased the relaxation of the tubes to 0.0075 and no success so far. Do you have any advise how to make it more stable? I also added my cas file...

Thanks for your help,

Doeby

Hi Doeby,
I am not sure whether this is causing your instability or not but in your tube file of 285 tubes the last 23 tubes have 0 as the 2nd node which may have done on purpose to represent a source/sink. In any case I tried this with an example that did run previously and introduced two additional tubes with node 2 as zero.

It ran for quite a number of iterations but became unstable


Regards
Tony

Hi Tony,

yes, I used 0 as the 2nd Node to model a sink.
In the meanwhile I solved the problem: In my tube file I used a very low pipe friction, if I increased the head losses a little bit (but friction coefficient is still physical reasonable) stability problems disappear. So, my guess is that telemac2d has some difficulties if the head loss of the pipe is very low...

Regards
Doeby

Hi Riadh,


my simulation is finished finally and I encounter a new problem. I extracted a couple of time series before and after the bridge and checked the water surface height. Even tough the bridge is NOT under pressure and I have a steady state flow in my entire domain I get a significant level drop at the bridge (see also the added images). I used for this bridge 20 tubes (width & height as mentioned in the prev post) with very low friction losses and I expected to get a water level drop at the bridge similar to the level drops before and after the bridge. So, any hints what I am doing wrong here?

Thanks,

Doeby

Hi Doeby

You could investigate what kind of flow exist under your bridge. Probably you will find in this situation that the water level condition gives a computation of discharge without downstream influence. If it's the case, the water level upstream depend only of the elevation of the tube (Z1 or Z2) and the discharge.

Hope this helps

PS: Please update your profile

Hi,

thanks for your quick reply. I am not quit sure if I understand you right: My tubes are linked with nodes of the river bed before and after the bridge (see also the image from a prev post with 12 tubes). So, there must be a downstream influence via the level height z2.

Regards,

Doeby