[Rising Star]

Hi Pablo

I'm trying to run a simulation with a floating body (a plastic sheet, 1 m x 0.01 m thickness)

The problem I'm facing is that I need to use values of accelerationdamping very low, like 0.01, otherwise the floating body starts to oscillate in the water, like it's getting a push from the water below.

Do you have any suggestion?

Thanks in advance

[ms411]

Dear Pablo,

my waves on a straight dike are calibrated. In the past few weeks, I tried to investigate the behavior of incident waves on curved dikes in the longitudal profile. In comparsion to the straight dike, waves will now be reflected in all directions, so that I tried to implement 3D absorbers at the end and the sides of my system. Unfortunately absorbers doesn't work.

Could the reason be a wrong version of olaFoam?
In addition it is not possible to implement more than 1 Paddles and 1 nEdgesMax to extract water at the end of my system, because then, I got the error "core dumped". This conditions lead to too much water at the dike, because the waves velocities are reflected and not absorbed. If I increase my absorption paddles at the sides of my system, waves are not generated well at the inlet anymore. At last I made my system wider in order that velocities from the dike will not reach the sides, but unfortunately without success.

Do you have an idea, what I did wrong? Would you be so kind and have a short view on my model, this would be really nice!!

My model is in the dropbox: https://www.dropbox.com/s/5o8v63s9n4...3Dabs.zip?dl=0

Thank you very much in advance !

Best regards

Martin

[Phicau]

Hi all,

back in the office after some vacation.

@RisingStar, depending on the cell size and other factors close to the free surface and object and assuming that your BCs are correctly set, this behaviour may be caused by the spurious velocities that appear at the free surface.

@Martin, 3D absorption is really complicated and it does not work perfectly, have you taken a look at my first two papers in Coastal Engineering (for a description and limitations)? Most of the times, olaFoam latest version will be the best, but in this case, it has been a while since I changed absorption, so you are probably fine using older ones. To generate oblique waves you NEED nPaddles > 1. I never had any core dumps having set nPaddles, nEdgesMin or nEdgesMax to > 1, you need to double check why this occurs, are you following the recommendations given in the user guide? Unfortunately I have a great deal of work these weeks to come to check everything in detail.

Best,

Pablo

[kmargaris]

Hi Pablo,

Are there general guidelines in selecting the domain size when using olaFoam?

I want to model a circular cylinder (2.5 m diameter, 1.5 m above the seabed) horizontally submerged in shallow waters of 8 m depth. The seabed is flat with no slope. I will use cnoidal waves and later irregular. My goal is to get drag and lift coefficients. The model will be 2D. The flow is expected to be transitional and I will probably use the Realizable k-ε model.

How big should the domain be with respect to the physical parameters involved? For example the domain should be 3xwavelength and the height of air above the mean water level should be 2xwater depth etc.

I will actually be using interFoam with olaFoam BC, since I don't have porous objects in the domain. Is there a benefit of using olaFoam vs interFoam?

Your help is much appreciated!

Best regards,

Kostas

[Phicau]

Hi Kostas,

there are no guidelines in terms of the wavelength, you can make the flume as short as you wish. You just need to be sure that the cylinder is "away enough" from the boundaries. You need to decide what is far enough in your case, given your dynamics (5-10 diameters ?).

For the air region, just be sure to leave some cells between the maximum free surface elevation and the atmosphere (5-10) and you will be fine.

olaFoam without porosity is exactly the same as interFoam, the only benefit is that you don't need to include the libraries in the controlDict file.

Best,

Pablo

[HosseinB]

Hi guys,

I am following the procedures from:

https://sites.google.com/site/olafoamcfd/source-code

to compile OlaFoan, but get an error message saying:


[HosseinB@vs2 OLAFOAM]$ ./allMake
wmakeLnInclude: linking include files to ./lnInclude
Making dependency list for source file waveAlpha/waveAlphaFvPatchScalarField.C
could not open file processorPolyPatch.H for source file waveAlpha/waveAlphaFvPatchScalarField.C
Making dependency list for source file waveVelocity/waveVelocityFvPatchVectorField.C
could not open file processorPolyPatch.H for source file waveVelocity/waveVelocityFvPatchVectorField.C
Making dependency list for source file multiPistonMovement/multiPistonMovement.C
could not open file processorPolyPatch.H for source file multiPistonMovement/multiPistonMovement.C
could not open file PointPatchFieldMapper.H for source file multiPistonMovement/multiPistonMovement.C
SOURCE=waveAlpha/waveAlphaFvPatchScalarField.C ; g++ -m64 -Dlinux64 -DWM_DP -Wall -Wextra -Wno-unused-parameter -Wold-style-cast -O3 -DNoRepository -ftemplate-depth-100 -DOFVERSION=221 -DOFFLAVOUR=3 -I/usr/local/OpenFOAM/OpenFOAM-2.2.1/src/finiteVolume/lnInclude -I./waveVelocity/velProfiles -I../common -I../common/checks -I../common/calculateWaterLevel -IlnInclude -I. -I/usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude -I/usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OSspecific/POSIX/lnInclude -fPIC -c $SOURCE -o Make/linux64Gcc48DPOpt/waveAlphaFvPatchScalarField.o
In file included from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/globalMeshData.H:85:0,
from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/GeometricBoundaryField.C:28,
from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/GeometricField.C:1253,
from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/GeometricField.H:598,
from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/GeometricScalarField.H:38,
from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/GeometricFields.H:34,
from /usr/local/OpenFOAM/OpenFOAM-2.2.1/src/finiteVolume/lnInclude/volFields.H:37,
from waveAlpha/waveAlphaFvPatchScalarField.C:55:
/usr/local/OpenFOAM/OpenFOAM-2.2.1/src/OpenFOAM/lnInclude/processorTopology.H:30:32: fatal error: processorPolyPatch.H: No such file or directory
#include "processorPolyPatch.H"
^
compilation terminated.
make: *** [Make/linux64Gcc48DPOpt/waveAlphaFvPatchScalarField.o] Error 1
\n\nWave generation boundary conditions compilation failed


Can you help?

Thanks,
Hossein

[Phicau]

Hi Hossein,

this is indeed a weird error, because the boundary conditions are universal for all OpenFOAM and foam-Extend version up to the date.

Are you sure that OpenFOAM is correctly installed and functional? I guess there is a problem, because "processorPolyPatch.H" should be located at: OpenFOAM-2.2.1/src/OpenFOAM/meshes/polyMesh/polyPatches/constraint/processor/ . Please make sure that everything is alright.

Are you using version 2.2.1 for a specific reason? I am asking this because there is no solver prepared for it (see supported versions here https://openfoamwiki.net/index.php/C...orted_Versions ), so you would need to create your own olaFoam for 2.2.1 (guaranteed to work!).

For a hassle-free experience I suggest you install version 2.2.2 (closest to 2.2.1) or the latest version (4.0).

Best,

Pablo

[HosseinB]

Hi Pablo, I am thankful for your quick reply.

I was trained in IHFoam v.2.2.1 and sticked to it afterwards. My last year paper results are based on this version of OpenFoam.

I checked and realized there is no "processor" directory under my OpenFoam path. I am using a cluster of many computational nodes off-building where I physically exist. There is no such a directory under any of OpenFoam versions on the cluster, not only v 2.2.1, or 2.2.2, ...

Cheers,
Hossein

Originally Posted by Phicau

Hi Hossein,

this is indeed a weird error, because the boundary conditions are universal for all OpenFOAM and foam-Extend version up to the date.

Are you sure that OpenFOAM is correctly installed and functional? I guess there is a problem, because "processorPolyPatch.H" should be located at: OpenFOAM-2.2.1/src/OpenFOAM/meshes/polyMesh/polyPatches/constraint/processor/ . Please make sure that everything is alright.

Are you using version 2.2.1 for a specific reason? I am asking this because there is no solver prepared for it (see supported versions here https://openfoamwiki.net/index.php/C...orted_Versions ), so you would need to create your own olaFoam for 2.2.1 (guaranteed to work!).

For a hassle-free experience I suggest you install version 2.2.2 (closest to 2.2.1) or the latest version (4.0).

Best,

Pablo

[HosseinB]

Hi all, my problem resolved by our IT people.

Originally Posted by HosseinB

Hi Pablo, I am thankful for your quick reply.

I was trained in IHFoam v.2.2.1 and sticked to it afterwards. My last year paper results are based on this version of OpenFoam.

I checked and realized there is no "processor" directory under my OpenFoam path. I am using a cluster of many computational nodes off-building where I physically exist. There is no such a directory under any of OpenFoam versions on the cluster, not only v 2.2.1, or 2.2.2, ...

Cheers,
Hossein

[HosseinB]

Hi guys,

I am modelling an irregular wave signal with olaFoam and it seems that the absorption isn't doing a good job and waves get reflected back to domain and interacting with incoming waves, ...

I have used IHFoam last year for regular waves modelling amd reflections from downstream boundary were not as significant as they are in the irregular case. The U file content:

/*--------------------------------*- C++ -*----------------------------------*\
| ========= | |
| \\ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \\ / O peration | Version: 1.7.x |
| \\ / A nd | Web: www.OpenFOAM.com |
| \\/ M anipulation | |
\*---------------------------------------------------------------------------*/
FoamFile
{
version 2.0;
format ascii;
class volVectorField;
location "0";
object U;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

dimensions [0 1 -1 0 0 0 0];

internalField uniform (0 0 0);

boundaryField
{


inlet
{
type movingWallVelocity;
value uniform (0 0 0);
}


outlet
{
type waveAbsorption2DVelocity;
absorptionDir 180;
nPaddles 10;
value uniform (0 0 0);
}



bottom
{
type fixedValue;
value uniform (0 0 0);
}

symmetry
{
type symmetryPlane;
}

right_wall
{
type fixedValue;
value uniform (0 0 0);
}

/*strct
{
type fixedValue;
value uniform (0 0 0);
} */

atmosphere
{
type pressureInletOutletVelocity;
value uniform (0 0 0);
}
}


// ************************************************** *********************** //


Any help is appreciated.

Thanks,
Hossein

[smh]

Hi Pablo,

Thanks in advance for your time and patience.

Actually I Wanted to know whether is it possible to simulate offshore floating platforms with IHFOAM or OLAFOAM or not? Also the flow pattern around the submerged part of platform is desired.
Thanks a lot,

Best regards,

Mahdi

[Phicau]

Hi Hossein,

sorry to hear that, but you should know that there are indeed some limitations for the present implementation of active wave absorption: performance decreases as you move away from the initial assumption of shallow waters. This is not only an olaFoam issue, it is also there in ihFoam, as I have not introduced any of the improvements that I am currently testing yet.

Hi Mahdi,

it is totally possible. OlaFoam includes a dynamic mesh solver (olaDyMFoam), ready to deal with floating objects. Stay tuned for some videos soon.

All the best,

Pablo

[HosseinB]

Hi all,

If generated waves are traveling in +x direction and I would like to absorb them on a downstream boundary, is the absorptionDir 180 or 0 deg., or both do the same thing in olaFoam?

I am experiencing large reflections and investigating if my model set up is correct.

Cheers,
Hossein

[Phicau]

Hi Hossein,

in this case it would be 180, since absorption needs to point to the inside of the domain. The reference angle is 0, pointing in the +X direction and increasing towards +Y (90 deg).

A trick, if you are ever in doubt, is omitting the absorptionDir variable or giving any value greater than 360, then the perpendicular direction to the boundary pointing to the inside of the domain is calculated automatically.

Best,

Pablo

[HosseinB]

Thanks, Pablo for your help, the trick, and quick response.

I am trying to model an irregular wave signal using olaFoam with the moving boundary technique that is given in one of the olaFoam tutorials. I am at the beginning of this modelling practice.

Last year, I validated IHFoam for a steep regular wave and presented that at Coastlab2016:

https://www.researchgate.net/publica...sed_Structures

Cheers,
Hossein

[kmargaris]

Originally Posted by Phicau

Hi Hossein,

in this case it would be 180, since absorption needs to point to the inside of the domain. The reference angle is 0, pointing in the +X direction and increasing towards +Y (90 deg).

A trick, if you are ever in doubt, is omitting the absorptionDir variable or giving any value greater than 360, then the perpendicular direction to the boundary pointing to the inside of the domain is calculated automatically.

Best,

Pablo

Hi Pablo,

My question is related to the one HosseinB posted. I am too a bit confused with the absorption direction at the outlet. The inlet in my models is the XMin patch and the waves travel in the +X direction towards XMax.

Initially, I did not notice reflection at the outlet and I think that was because the domain was quite long 3xwavelength, about 480m.

When I shorten the domain length I started noticing wave reflection. I also noticed that in all cases after I use the setFields application, the XMin boundary condition for alpha.water becomes:

XMin
{
type waveAlpha;
waveType aaa;
waterDepth -1;
genAbs 0;
nPaddles 1;
allCheck 0;
waveDictName waveDict;
value uniform 0;
}

Do I have to change any values, like genAbs?

I have uploaded a test case (without mesh). I would also appreciate if you look the fvSolution and fvSchemes files and perhaps suggest changes to increase the time step.

Best regards,

Kostas

[Phicau]

Hi Kostas,

don't worry about those values, they are just a placeholder that appears when you use setFields. You don't need to change anything because they will be updated automatically internally during the first time step, that is how olaFoam works.

It could be great if you both could provide an estimation of your reflections, with your sea state characteristics and an easy Mansard and Funke analysis (3 gauges) so that maybe I can try to tackle that problems first and tune the new absorption procedure for the new update of the code.

With respect to the numerical schemes, I looked at them in the beginning of my thesis and came up with "good combinations" out of experience. Those are the parameters of the tutorials included. Of course, they might not be the optimal for your case, thus, users should do an analysis themselves.

Best,

Pablo

[kmargaris]

Hi Pablo,

Thanks for the quick response. I am new to these type of free surface flows and I don't know the "Mansard and Funke" analysis. If you send me more information, I will try to do it.

My main concern on the reflection is how much it affects the forces on a submerged pipe. I attached an image with drag forces in a submerged cylinder with two different domain lengths. The effect on the force is quite small in my opinion.

Best regards,

Kostas

[Phicau]

Hi Kostas,

Mansard and Funke (1980) is a well known reference in which a method to separate the incident and reflected waves is explained. Find the paper here:

https://icce-ojs-tamu.tdl.org/icce/i...File/3432/3112

Is the mesh length the only thing that varies there and not the cell size? You should be able to check whether it is a reflection issue looking at the initial instants. You can estimate the wave celerity of the reflected components and see if they start acting when you expect them.

Also, remember that in order to get accurate drag forces/coefficients, the mesh has to be of an excellent quality and you should perform a mesh sensitivity analysis.

Best,

Pablo

[smh]

Thank you Pablo,
I'll be waiting for the videos. Is there any tutorial for floating objects using OlaDyMFoam?

Best regards,

Mahdi