[arnau1985]

Dear FOAMers,

I am trying to model a hydraulic jump in a rectangular open channel using interFoam but so far I have not managed to define successfully the boundary conditions. I want to force a supercritical inlet (imposing a given velocity and the water depth) and a subcritical outlet (imposing an experimentally-determined water depth and a hydrostatic pressure profile). Please see the attached figure for further details.

I have tried thousands of boundary conditions using groovyBC but none of them worked, apparently, due to a bad definition of the outlet boundary condition. I modeled hydraulic jumps in the past by means of downstream steps, slopes and so on. But this time I have to do it directly imposing the water depth at the outlet. The core of the problem seems to be how to make the outlet boundary condition raise the water level.

Has anybody done something similar before with OpenFOAM? Any advice? I have not found any solution in other threads.

Thank you very much,


Arnau.

[colinB]

Dear Arnau,

I don't think that your problem is related to the inlet or outlet boundary
conditions.

The problem is that for a hydraulic jump you need a high velocity in the
beginning which is then slowed down and the water level increases.
Usually this is no problem when using sloped geometries like a backwards
facing slope. However dealing only with a plain ground this is slightly
more difficult.
What you can do now is in my opinion to increase the length of the
domain.
With now having more surface which causes friction maybe an hydraulic
jump will occur. If this does not help I would also increase the speed of the
water at the inlet.

I hope my explanations help

kind regards
Colin

[vonboett]

See the trick of kflora in http://www.cfd-online.com/Forums/ope...tml#post391836
at 5th November 2012. Was aswell applied by pythagOra5 and compared to flume experiments with satisfying results.

[mgdenno]

vonboett,

Could you provide a link or reference for the flume experiments comparison?

Thanks,

Matt

[vonboett]

well Matthias send me the comparison to his experiments by mail so I can't tell if its O.K. to post them here, maybe you could directly contact him:
http://www.cfd-online.com/Forums/mem...ythag0ra5.html
His results deviate from the experiments due to turbulence at an obstacle in the center of the domain, but his flow height at the outlet is stable.

[avigrod]

I was simulating a lab channel in which among other things, there is a hydraulic jump. I started simulating the hydraulic jump isolated, trying for that purpose of creating it from a step and establishing zeroGradient conditions in the outlet over the step. Nonetheless, after reading Vonboett, I decided to give a try to kflora suggested BC, as they are closer to typical input data in hydraulic problems.

At the moment, I have implemented Arnau sketch, but establishing buoyant pressure in either inlet and outlet. A fixed speed in the inlet (with alpha1 = 1), and a fixed speed in the whole outlet (in order to get y2) for air and/or water. Over the inlet I have try two different bc, the atmospheric one and also a wall (nutkWallFunction) with 0 speed. On the channel floor I fixed a zero speed (U) and zeroGradient (alpha1) with a nutkRoughWallFunction.

In both cases the hydraulic jump starts moving towards upstream but instead of stopping in the channel it locates exactly in the inlet. If wall over the inlet is chaged with an atmospheric bc, there is even some small negative flow rate over the inlet. I have try enlarging or shortening the channel, but hydraulic jump position remains the same. Even changing lightly y1 or U1.

I am using the k-epsilon model for the turbulence (and interFoam as solver)

I upload here the case and some plots of the commented results

Animation where Ux is plotted during 100s until a steady state is reached. The volume plotted corresponds to alpha1>0.50

https://drive.google.com/file/d/0Bzj...it?usp=sharing

In the steady-state, this figure shows U:

https://drive.google.com/file/d/0Bzj...it?usp=sharing

Basic case files:
https://drive.google.com/file/d/0Bzj...it?usp=sharing

[vonboett]

to get a solution comparable to experiments you should go for three dimensions and maybe use LES

[avigrod]

Originally Posted by vonboett

to get a solution comparable to experiments you should go for three dimensions and maybe use LES

Thanks for your answer... I was trying first to make it comparable to Berlanguer equation... now I have been trying to increase momentum, keeping q constant, and at the same time following your suggestion I have re-run simulation in 3D with a symmetry plane (but still with k-e). I will comment here results later on.

[luccy]

hello everyone,
I'm simulating a free hydraulic jump with InterFoam. In the laboratory model, the jump occurs at a distance of one meter from the gate, but in the simulation, despite changes in various parameters such as mesh and inlet velocity, this does not happen and the jump occurs immediately after the gate. Can anyone help me?
I will appreciate that,

[vonboett]

Hi
What turbulence model do you use, what is the Froude number of your flow?

Hydraulic jumps in mountain torrents were captured by interFoam quite well using hybrid URANS-LES...

[luccy]

hi
my Fr=4.25 and turbulence model is k-e RNG

[2538sukham]

Btw, Froude no. is used in determining whether the flow is sub critical or supercritical to allow fr the hydraulic jump. If instead there is a moving wall instead of inlet, the movement of the wall should induce the critical Froude no so that hydraulic jump occurs. I did simulations using interDyMFoam but the interface height is not what is shown by shallow water considerations of the NS eq. Any tips. I though the topic actually matches a lot.

[luccy]

Hi, I am trying to simulate a hydraulic jump in an open channel with a Froude number of 13.4 and a Reynolds number of 10^5 with interisofoam with k-e turbulence model and high turbulence intensity of 12%, these are my initial and boundary conditions in zip file, but the jump does not form properly and failed but the simulation didn't diverge as it is in the movie. the mesh quality is good, I also reduced the Courant number and changed B.c conditions and solver to interfome but didn't help me either. Do you have any suggestions for me?
thank you