[Ed R]

Dear All,

I'm having some issues with large waves at the inlet boundary of a numerical wave tank (NWT). I've attached 2 images to explain my issues; one before a large crest enters the domain and one just afterwards..

I seams to me that excessive air velocities are created when a large crest is created at the inlet. I believe this is due to the relaxation zone forcing the air velocities to zero and not letting them come in through the inlet, instead they are forced around the relaxation zone (you can see where the boundary is due to the air vectors). Even without a relaxation zone (which I need) the inlet boundary has air velocities of zero and so air is forced in from above. Also I cannot seem to make the relaxation zone only come up to the top of the water (and not go into the air phase). Can this be done? Ideally I'd have an atmospheric BC for the air phase (instead of set to zero) and only relax the water phase.

The consequence of this is that the crest appears to steepen and the freesurface behind the crest is badly disturbed. Any comments or help on this issue would be much appreciated.

Thanks

Ed

[ngj]

Good morning Ed,

I do not recall ever to see such a behaviour (with a relaxation zone), so could you please share some more informations? Could you try to run the relaxationZoneLayout utility, just to make sure that your zone is not wrongly defined.

With respect to the idea of only relaxing in the water, then it is possible to modify/extend the code with such a functionality. It merely requires a new relaxationScheme<partlyRelaxing> class.

Kind regards,

Niels

[Yashar]

Hi Niels,
I have problem in modifying the interDyMFoam to waveDyMFoam. I have done all the steps. at last when I wanted to do ''wmake'' in the main file, the following error appeared:

yashar@ubuntu:~/OpenFOAM/yashar-2.2.1/applications/solvers/waves2Foam/applications/solvers/solvers221/waveFoam/waveDyMFoam$ wmake
linuxGccDPOpt/options:6: *** missing separator. Stop.
wmake error: file 'Make/linuxGccDPOpt/objectFiles' could not be created in /home/yashar/OpenFOAM/yashar-2.2.1/applications/solvers/waves2Foam/applications/solvers/solvers221/waveFoam/waveDyMFoam

Yashar

[mary mor]

Dear All
I downloaded waves2Foam and ran it with openFoam16ext with no problem.

Now I want to create waveDyMFoam. I followed the instruction on wiki but after running wmake it gives me the error below:

linuxGccDPOpt/options:2: *** missing separator.  Stop.
wmake error: file 'Make/linuxGccDPOpt/objectFiles' could not be created

Can anyone please tell me how to fix it?

Best regards,
Maryam

[ngj]

Hi Maryam,

You are missing one or more line seperators in the file Make/options.

Kind regards

Niels

[Ed R]

Hi Niels,

Here's the relaxationZoneLayout for the case, it looks fine to me.

The case is a focused wave made up of many components using combinedwave wave type. The amplitudes and phases at the inlet are taken from experimental data (an FFT of a time series) and the results are compared against surface elevation time series further down the tank to assess the quality of modelling wave-wave interactions in OpenFOAM. The results aren't bad but the troughs and peaks are a little off (particularly near the inlet) and immediately after the main peak the free surface is full of high frequency disturbances. I have been working on resolving this issue for some time when I discovered these high air velocities which I now believe to be partly responsible.

What I'm looking for is for the inlet to allow a smooth transition from the BC to the numerical domain so it is as if the wave has been propagating for a while before crossing the boundary and it crosses unaltered. To make it more complicated I have been looking at structure interactions and so wanted the relaxation zone to absorb reflections and also I find steep/high amplitude waves can cause the simulation to crash without the inlet relaxation zone.

Any further help would be greatly appreciated.

Ed

[Maralady]

Neils, I am trying to simulate a sloped tank, however I am trying to use 2 blocks instead of just 1, but the system is crashing I assume that it is because I am getting negative values for alpha1

Courant Number mean: 0.00325609 max: 0.256065
Interface Courant Number mean: 0.000130793 max: 0.113135
deltaT = 0.00197208
Time = 0.0481498

MULES: Solving for alpha1
Phase-1 volume fraction = 0.387221 Min(alpha1) = -2.07058e-19 Max(alpha1) = 1
DILUPBiCG: Solving for Ux, Initial residual = 0.000842177, Final residual = 1.68671e-10, No Iterations 2
DILUPBiCG: Solving for Uy, Initial residual = 0.00309381, Final residual = 3.72365e-10, No Iterations 2
GAMG: Solving for p_rgh, Initial residual = 0.00762491, Final residual = 5.21319e-08, No Iterations 9
GAMG: Solving for p_rgh, Initial residual = 0.00025839, Final residual = 7.73398e-08, No Iterations 4
time step continuity errors : sum local = 3.19437e-10, global = 5.37101e-11, cumulative = -1.72466e-10
GAMG: Solving for p_rgh, Initial residual = 0.000142471, Final residual = 4.64556e-08, No Iterations 4
GAMG: Solving for p_rgh, Initial residual = 1.35461e-05, Final residual = 2.86379e-08, No Iterations 2
time step continuity errors : sum local = 1.18287e-10, global = 8.7723e-13, cumulative = -1.71589e-10
GAMG: Solving for p_rgh, Initial residual = 9.61035e-06, Final residual = 4.70567e-08, No Iterations 2
GAMG: Solving for p_rgh, Initial residual = 9.00775e-07, Final residual = 8.30235e-09, No Iterations 3
time step continuity errors : sum local = 3.42923e-11, global = -1.39533e-12, cumulative = -1.72984e-10
ExecutionTime = 5.91 s ClockTime = 6 s

The mesh seems to be OK or at least it is what the checkMesh utility says, but it is not a regular mesh, what else could it be? Any suggestion?

[ngj]

Hi Arthur,

I will encourage you to read the following thread carefully:

http://www.cfd-online.com/Forums/ope...-get-help.html

Kind regards,

Niels

[mary mor]

Thanks Niel,

I also have 2 more questions.
-Is this toolbox capable to be coouled with compressibleInterFoam? For some reasons, I need to solve the air as compressible that I explain below.

-I have a floating object with air chambers. because of wave crossing water level can go up and down in chamber and air can get out and in from the chamber. So I need the velocity of air. As I understood, your toolbox set the air velocity by zero. Is it possible for me to consider air velocity in the chambers but ignore it outside it? I also should use waveDyMFoam.

Thankes in advanced
Best regards,
Maryam

[mprasad]

Hi Niels,

I am having some trouble running my case. I get the following error when i execute setWaveParameters:

--> FOAM FATAL IO ERROR:
keyword debug is undefined in dictionary "/home/prasad/OpenFOAM/wrCases/waveRunUpNs/constant/waveProperties.input.inlet1Coeffs"

file: /home/prasad/OpenFOAM/wrCases/waveRunUpNs/constant/waveProperties.input.inlet1Coeffs from line 30 to line 48.

From function dictionary::lookupEntry(const word&, bool, bool) const
in file db/dictionary/dictionary.C at line 402.

FOAM exiting

I am new to OpenFOAM and so far am unable to troubleshoot this. Is there something with regards to waves2foam that I am missing?

Thank you.

Regards,
Prasad

[Ed R]

Dear Prasad,

When using the Stokes Second order wave type there is an option called 'debug' which (I think) gives you information about the quality of the stokes second approximation based on the wave properties you have asked for. This is a Boolean function which requires a true or false value. The reason for your error message is that you have not entered either. Try this

...
inlet1Coeffs
{

waveType stokesSecond;
depth 1.2;
period 1.13;
direction ( 1 0 0 );
phi 0;
height 0.35;
waveNumber (2.07997395 0 0);
omega 0.884956;
debug true;

....

Ed

[mprasad]

Dear Ed,

It worked! Had to tweak my waveheight value to obey second order navier stokes.

Thank you for your help.

Regards,
Prasad

[aisy]

Hi,
I want to install waves2Foam in OF221,but I am failed, it shows that

=====================================
COMPILE LIBRARY
=====================================
./Allwmake: line 29: wmake: command not found
./Allwmake: line 30: wmake: command not found
./Allwmake: line 31: wmake: command not found

=====================================
COMPILE SOLVERS
=====================================
ls: cannot access /home/jxr/OpenFOAM/waves2Foam/applications/solvers/solvers0/*: No such file or directory

=====================================
COMPILE PRE-PROCESSING
=====================================
./Allwmake: line 69: wmake: command not found

=====================================
COMPILE POST-PROCESSING
=====================================
./Allwmake: line 85: wmake: command not found

=====================================
COMPILATION DONE

WHO can tell me what should I do
Thanks

[jasonchen]

Originally Posted by aisy

Hi,
I want to install waves2Foam in OF221,but I am failed, it shows that

=====================================
COMPILE LIBRARY
=====================================
./Allwmake: line 29: wmake: command not found
./Allwmake: line 30: wmake: command not found
./Allwmake: line 31: wmake: command not found

Just a little note: try to add before wherever Allwmake appears
bash ./Allwmake

[ngj]

Good evening,

I will try to address the questions below:

@Ed: Do you only experience this type of behaviour with the mixedType boundary condition, or is it the same if you use the focusing functionality in the irregular wave field? I have never noted these large velocities, so I am a bit at a loss.

@Maryam: waves2Foam only corrects the velocities in the relaxation zone, so the inner domain is left unaltered. From a code point of view, it should be possible to combine compressibleInterFoam with waves2Foam, but I do not have sufficient knowledge in compressible flow solutions to say, whether it will run (robustly).

@Aisy: You need to source OpenFoam, before you are compiling. I suspect this to be the issue, as it does not even recognize the OpenFoam make function: wmake.

Kind regards,

Niels

[katakgoreng]

Originally Posted by ngj

Good morning Katakgoreng,

Yes, I am using Eclipse. It works quite nicely for something like development of new libraries, however, I have not been successful (and have hardly tried) in getting the auto-completion working within the solvers.

Good luck,

Niels

Hi Niels,

I have tried using Eclipse but abandoned it half way as I can't get auto-completion working. Currently I'm using Qt Creator, auto-completion work just fine. I'm new to c++, so getting the auto-completion working really help a lot. You should give it a try.



Kind regards,
katakgoreng

[Ed R]

Hi Niels,

Thanks for the reply. I have not used the irregular wave focusing method but a quick experiment with a steep stream function (5th order and shallow water - T=14.3s, H=1m, 2m depth, 1m of air phase) and I can get similar behaviour (slightly less but still exists). I believe it is because the pattern of air flow over the wave is disrupted at the inlet as the air phase is set to zero and the relaxation zone has the target solution of zero. The air velocities above large waves are significant and I think setting them to zero causes some issues.

Furthermore, I believe some of my issues with the freesurface disruption is that my first order summation of wave components isn't getting the velocities quite right. Does the focusing method you've developed recently allow for a second order summation (like the bichromaticSecond wave type)? Also I need to add the phases for the components arbitrarily as in my experience real waves don't focus according to linear theory and some manipulation of the phases is required.

Any further suggestions and help are always appreciated.

Thanks,

Ed

[ngj]

Hi Ed,

Ok, thanks for the feed-back on the steep stream function. Then, testing whether no relaxation in the air helps, could be a next suitable step.

With respect to the focusing, it still relies on first order, as the only difference is that the computation of the phase lag is no longer just a random number, but is computed based on the focusing information.

Second order irregular waves are currently not in the pipeline, as it becomes very tedious to compute all of the wave-wave interactions. With N first order components, you will add N^2 second order components, why you quickly exceed a total of 1000 wave components. The number of evaluations of this inside the relaxation zone is huge, so the evaluation of these will actually become a bottleneck. You can verify this yourself by computing a first order irregular wave with 1000 components.
A recent addition to the svn is a relaxation method, which reduces this problem (svn revision 2022). This is used as follows:

relaxationScheme SpatialInterpolation;
relaxationShape Line;
N <some integer>;

Currently, it is only implemented for 2D problems (or 3D problems with incident long crested waves), but it should be straight forward to extend to general cases. The integer N is the number of points through the relaxation zone, where the wave height is pre-computed per each time step, such that the number of evaluation of the many wave components is reduced. However, this does not resolve the problem that the velocity has to be computed inside each cell in the water.

Good luck,

Niels

[mprasad]

Hi all,

I am looking to create stokes second waves and am confused with regards to the following. How do I input the wavelength and how is the wave number computed in the waveProperties file?

Thank you in advance.

[ngj]

Hi Prasad,

The wave number for second order stokes theory is identical to that of first order stokes theory. The setWaveParameters utility will do it for you.

Kind regards,

Niels