open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi All,
I have been applying the tubes module to a submerged bridge study and produced unexpected results in respect to producing locally high elevations at the downstream bridge nodes. In any case I decided to check a simple test case in which a single 2m by 2m submerged tube connects two large and deep reservoirs one kept at 4m elevation and the lower at 2m elevation and bed at -10m throughout.

i included 0.5 as entry loss, 1.0 as exit loss and 0.15 as linear friction loss fL/D
simply put I would expect the flow through the Tube to be governed by HL = 1.65*V2/2g with Hl being 2m and thus Q the tube flow equal to 19.51cumec.

However when I run it in telemac I get a Tube discharge as outputted to the Screen of 31.5cumec.

Perhaps I am making a fundamental mistake in my calculations ??

regards
Tony

I include the test files if anybody wishes to check my example

Instead of controlling water levels with a liquid boundary file, I set them with Prescribed Boundary Conditions. 4m on one end of the model, 2m on the other.

When run to steady state, this model gave a flow of 31.1m3/s through the culvert, which is close to what you had expected. Well, I've simulated -31.1m3/s as I got the boundaries swapped

So the Tube function seems to behave as expected in the test. Does this help?

I think you miss understand my post

I believe the value that telemac should have produced was 19.5 and not 31.5 based on simple pipe flow hydraulics

regards
Tony

sorry, you are right. I misunderstood your original post.

However, regarding your original question:

For pressurized flow, the TUBE subroutine only considers loss effects from entrance loss (CE1 or CE2) and linear loss (L12). An exception for pressurized flow is if the culvert obvert is not submerged, which I believe only occurs when water is equal to the culvert obvert. In this exception, CS is also considered.

I'm not sure how well documented this formula is in the manuals, particularly since BUSE is a recently added feature. But it is apparent when you look at the BUSE.F code.

So: Q = LRG*HAU*sqrt(2*g*dH/(CE+L)) = 31.1m3/s
LRG = 2m
HAU = 2m
dH = ~2m
CE = 0.5, L = 0.15

Yes, I agree that is what BUSE.f is doing but theoretically it is incorrect as the example represents a submerged outlet controlled culvert and therefore should have included the exist loss

Hydraulic Text books such as R.H. French (1987) Open channel hydraulics give the following discharge relationship for such a culvert

Q=C_d AREA (2g∆H/(1+(29 C_d^2 (n^2 L)⁄R^(4⁄3) ) ))^0.5

(where n is the manning coefficient which can be converted to other roughness laws) and which accounts for entrance loss, exit loss and friction loss with Cd typically .87 to .96

Regards
Tony

Perhaps in this application, you want to be applying key words for CULVERTS instead of TUBES? I believe the Culvert application (SIPHON.F) does apply entrance and exit losses as you outline.

Of course, SIPHON has its limitations too, which is why TUBES were introduced.

I'll have to admit, when I've applied TUBES, I understood how the model was calculating headloss and adjusted my loss values accordingly based on known or expected rating curves. Since I was applying it to spillway gates, I didn't consider exit or friction losses (set them to zero!). TUBES was useful as it allowed me to account for varied radial gate heights (by adjusting HAU) and also approximated free discharge should those conditions exist.

Hi

Derek is right! When we developed the tube subroutine, our goal was mainly to have the possibility to represent the free and pressurize flow in some tubes.
If inlet and outlet is always submerged the culvert subroutine which already exists at this moment will do the job perfectly.
And complementary to also have the possibility to fix the flow direction (in case of valve).

Maybe it could be possible to find a more general solution but this will probably change some description (culvert and tubes)

If in your case you want to add the exit loss, you could modify the subroutine...

Hope this clarify the question

Hi Christophe and Derek,

thanks for addressing my query, I suppose what prompted me to examine this was the large difference in results between 1D heC Ras model of a bridge and the Tubes routine and this is not just for the submerged case but also the free flow case as the flood wave progresses through the bridge going from free flow to submerged. The HECRAS model solves for the 6 culvert flow categories(inlet, outlet submerged, unsubmerged )and requires calculation of the critical Depth at the inlet and outlet to establish what equations and coefficients are needed and uses for the inlet controlled flows the FHA (1985) inlet equations.

The Tubes subroutine is very useful and more so than Siphon subroutine for expanding to cover these culvert flow methods which I will attempt to do

The equations are outlined in the HECRAS Reference manual and in more detail in the CIRIA report 168 (1997).

thanks
Tony