open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

Thanks for your attention, i want to use the culvert to simulate sluice on the river, and here are some questions i am not clear:
1. is the elevation of bottom at the culvert higher than the river ?
for example, the bottom of river is 15m, and the bottom of culverts is altered to 20, the same as river bank, and then add a culvert at this site
2. I1 and I2, the number of each end of the culvert,
Here is a test case, i want to set a sluice in the yellow box( it is a little long, please ignore the length of it), which is better that (I1=1755 and I2=1754,the green points) or (I1=207 and I2=174,the blue points),and the direction of culvert, is as same as the line between I1 and I2 ?



3. LRG,the width, is it suitable the same as the width of river ?
4. C56,it was described in the Telemac2d ReleaseNote as the constant used to differentiate flow types 5 and 6,the case in #26896 give the number of 10, but which constant is suitable?
5. in my case, use the culvert to simulate sluice on the river, the parameters CE CS FRIC i give them 0 to ignore the head loss, and CV5 C5 use the number of 10 1.5 suggested in #26896 , are they reasonable? Or CE CS use 0.1 is better?
6. Z1 and Z2, are they as high as the bottom of river or a little higher than the bottom?

Thanks again, regards.
W

Hi,

I'll try to answer some of your questions:

1. If it is actually a sluice and the river is not going through like in your figure, you can elevate the bottom between the culvert nodes I1 and I2. Otherwise the water will just run through in your simulation. But be carefull, the point I1 and I2 must have a lower elevation and must contain water to pass through the culvert!
2. THe green and the blue points are both fine. The code doesn't really take the length into account at this moment. The length is only used to calculate the head loss due to wall friction. The water is transported from one end to the other in 1 time step no matter the length. So this code can not be used to model long pipes or very long culverts. The amount of water that you can transport through the culvert in 1 time step depends on the amount of water that is available in the node I1 at that time step. This is checked in the code and a limiting factor on the amount of discharge through your sluice or culvert.
3. You can try to model your sluice with one culvert, but if this culvert/sluice is very wide, it will transport a large amount of water every time step and this water must be available in the node I1! The code will automatically limit the discharge to the amount that is available.
So it is possible that you give a very large LRG value and the calculated discharges through your sluice are very large, but only a small part of this is actually transported in the model in one time step. Then it is better to dived your sluice over several nodes and use multiple culverts.
You could check this by going in the code of the culverts and ask to give output of calculated discharges before and after the check for the amount of water that is available in point I1. This check looks like this:
QMAX1=0.9D0*H(I1)*V2DPAR%R(I1)/DT
where Qmax is taken as 90% of the volume of water available in point I1
Then the theoretical discharge through your sluice or culvert is calculated based on the water levels in point I1 and I2 and is called DBUS in the code.
IF(DBUS(N).GT.0.D0) THEN
DBUS(N)=MIN(QMAX1,DBUS(N))
If the theoretical discharge is larger than zero, the minimum of this value and Qmax1 is taken.
(maybe we could add a warning message in the code when this is done)
You could alter this code a bit to ask for output of DBUS before and after this test to see.
4. this constant C56 is determined based on figure 3 in my conference paper about the culvert modeling. You can find this paper in the TUC2016 proceeding or at my researchgate page www.researchgate.net/publication/3091361...AC-2D_and_TELEMAC-3D
5. If you look up my paper you can find good definitions of these parameters and find the best values for them depending on your particular case.
6. Z1 and Z2 are the bottom levels of your sluice at entrance and exit respectively. They are usually equal or higher as the river bottom.

Hope this info helps you model the culverts in your case.

kind regards,
Sven

Hi,Sven
Thanks so much for your detailed reply, and I have read the paper you suggested and done some test about my simulation, but there are also some questions.

Here, I give four points and the water level, Q(1) is the upstream and SL(2) is the downstream.





1. Why are the water level at P2 and P3 extremely higher than P1 and P2
2. In my culverts data file, the parameters were set to 0 to ignore the head loss in the culvert, and at entrance, exit and other structures, but why is the water level of P4 lower than P1, and what cause the head loss
3. When the water reaches I2(the end number of culvert) through culvert, the water doesn't flow to downstream and results the water level rises at P1. After a long time, the water flow to downstream and the water level at P4 rises, please see the water level figure. I'm not clear about this phenomenon and expect your suggestion.

My test files are shown as below













Thanks again!
Best regards.
W

Hi,

I've looked at your test case and found the problem. The culvert formulations are based on physical properties of the culverts. If you set every type of head loss to zero, this will give an infinitely large discharge. In the code there is a limitation on the discharge based on the amount of water that is available at the entrance node I1. So in your simulation you will get discharges as high as 4000 m³/s, which is equal to the maximum amount of water available at node I1.

What you should do is set at least a length of the culvert and a Manning friction coefficient of the wall of the culvert. If I set the length to 100 m and a wall friction coefficient of 0.025 I get more reasonable discharges as high as 100 m³/s and your water levels will turn out more realistic.

I will see if I can introduce a kind of warning in the code for this kind of use of the culverts.

Hoe this answers your questions.

Kind regards,
Sven

Hi,Sven

Your suggestions are very helpful for me to understand the simulation of culverts, and I will try my real case next.

Thanks very much.
Best regards.
W