Smit,

Under the conditions you described, all oscillations should have been vanished after the 5th second. First of all, your mesh is too coarse. You should try to refine it around the hull. Try to keep your largest cell smaller than 0.25 m around the hull. If that doesn't help try decreasing your time step. If you are using variable time step size, change it to fixed. Use a fixed time step of 0.001~0.0005 seconds. That should help eliminating the oscillations you are experiencing.

Let me know how it goes

Best
Kilroy

I've refined my mesh (made using snappyHexMesh) to 6M cells and a fixed time step yielding a maxCo of ~1.5 and maxAlphaCo of ~1 (using nAlphaSubcycles of 5).
Unfortunately the resistance evolves identically, with too large oscillations.

What sorte of div scheme you have in your fvshemes file ? I calcul all the ship resistance with LTSInterFoam and I use for U, k and omega limitedLinear scheme, more accurate and stable than linear scheme.

Thank you for your response,
I've been using linearUpwind grad(U) where grad(U) is cellLimited Gauss linear 1. Is it reasonable to think that my oscillations would vanish if I would use limitedLinearV 1 instead?

You can try limitedLinearV 1 or limitedLinear 1.0 phi (I use this). May be with this scheme, your oscillation will be disappear. And you can try LTSInterFoam, I think it's more stable than interFoam.

Best regards
Vince

Refining mesh, shorten time step or change to limitedLinear 1 phi scheme for div(phi,U) does not decrease the oscillations unfortunately.

Hi all!

I have on problem with last OF version, OF-2.2.1. At each time when I run a calcul (with LTSInterFoam or interDyMFoam), at the beginning, I have one error message. But it's strange because the calcul continue and give some good results.

I join the file with a part of error message.

Any Idea

KR

Vince

problem with LTSInterFoam
 

Hi all

I use now OF 2.2.1. I have a new problem with LTSInterFoam. I work on a multihull at hight froude number and usually I use limitedLinear scheme for div(rho*phi,U). In the past, it shows the most accurates results.

With the last version of LTSInterFoam, if I have a coarse mesh, it's ok but results are wrong and if I have a refine mesh, calcul crash after 3 step! I don't understand what's happen.

I check my mesh, it seem ok, I check all parameters, it seem ok... If you have an idea, that will be great.

Vincent

Convergence issues with LTSinterFoam.
 

Hi everyone!

First of all I should say that I feel sorry for the simplicity of my question following all these rather complex matters that you guys have been tackling!!

I have been trying to solve for a 0 DOF VoF case for a while now and am still having some issues with the convergence.
I am primarily interested in computing hull resistance at low fn.
Rather than exposing all the cases I have been working on, I thought I would post the results I get from the well known wigley hull tutorial.
For this run all the parameters were kept unchanged from the tutorial with the exception of:

1 The calculation of the pressure/viscous forces and moments (forces function object added in the controlDict)

2 Calculation in parallel on 8 Cores using the “simple” method. SimpleCoeffs being n (8 1 1)
and delta 0.001

I have attached snapshot of the mesh, the resulting free surface and the disappointing convergence graph :(
I should mention that I have investigated running it for longer then got instability issues due to the wake bouncing off the boundaries of the domain which I got rid of by increasing the domain size but I still didn't achieve a good state of convergence.

So my question is simple. What do I need to do in order to get a better convergence.
Is it a meshing issue or should I focus on other things?
Ideally if someone has solved this issue and is happy to share his case files that would give me a very good starting point and would be immensely appreciated.
Thanks a lot for your help.

Oh and I should also say that I too used OF v2.2.1 and haven't experienced any crashes so far with LTSinteFoam Vince. Not sure this answers your question though...

Hi Jule

To solve my problem, I use interFoam. I don't know why it's ok with interFoam and no with LTSInterFoam but the more essential that I can use limitedLinear scheme.

Vince

Hi dear all,
i'm trying to simulate a floating body under wave. At first I was trying to combine wave2foam toolbox and the floating object tutorial in interDyMFoam. but as it seems there are some problems with the tutorials and it behaves abnormally beside pressure instability.
Also I heard about the solver shipFoam, which seems not to have these problems.
Can anyone that has done kind of this simulation so far, get me some advice about where I should start. I would be glad to hear any progress at this matter:)
Can anyone please send a link or the file of shipFoam. I found one file but it was from a while ago, I'm not sure if it's with the latest modifications.

I appreciate any help.
Thanks,
Best regards

Has anyone used the numerical beach capabilities within the ShipHydroSIG package/ navalFoam?

I have tried to use the numerical beach for free surface ship flows but it creates waves near inlet and outlet.

Best regards

Hi all

I install the new OF version 2.3. I test LTSInterFoam with the DTCHull but it appear there is some problems with meshing.

Some people have test this case yet?

BR

calcul 2DoF with OF 2.3
 

Hi

After lot of test, I realized a comparison on a multihull between OF and FLUENT code (the calculs with FLUENT are realized by an another company). I can't show the CAD because it's confidential. I can just say it's a catamaran, L=19m.
I join a pdf with results.

I have a very good agrement with Fvx and not with Fpx. I think it's because of bad trim and sink prediction.

My problem: I have some difficulties to estimate some coefficients in the dynamicMesh file, especially the translationDamper and rotationDamper in part restraints. For the moment, I assume these coefficients are linear. In the tutorial, Lhull=5.97, and in my case, Lhull=19, I multiply translation and rotationDamper by 3.18 but I'm not sure at all it's the best way.
I don't change the accelerationRelaxation coeff, but may be, I must?

If some people have an idea, it will be great.

BR

Hi Vince,
I have tested the DTCHull case and haven't had any meshing issues.
Regarding the translation and rotation damper I like you opted for a linear relationship but I'm not convinced it is correct. Have you found out anything about that since?
What solver do you use for your 2 DOF case?
Cheers.

Hi Jule

I think also a linear relationship for translation and rotation damper is not correct. But for the moment, I found nothing to estimate these coefficients. If you have an idea, it will be great.

For 2DOF case, I use interDyMFoam.

Cheers

Hello everyone,

I am trying to calculate the total resistance of wigley hull at Fn = 0.316 for the last two months. I have tried LTSInterFoam and InterDyMFoam for the calculation in OpenFoam 2.3.0 , but the results are not a satisfactory at all. I hope someone (specially who have already validated it) will help me to find out the problem. I appreciate any kind of help and hints. I can upload the cases if is required.

Thank you.

Best regards,

Ali

please show me you case files. i also done what you are doing. i found the unsteady solvers interFoam or interDyMFoam were better than LTSInterFoam. you need correct settings in fvSolution and fvScheme files to obtain good results.