Multiple floating objects
 

Dear Foamers,

My eventual goal is to simulate multiple ships floating side-by-side in waves.

As a first step, I have modified the floatingObject example under the interDyMFoam sub-directory to model 2 boxes. I have tried to refine my mesh and adjust the time steps, but my two attempts seemed to stop at the same point in time regardless of the mesh refinement.

The error message that I am getting is as follows:

I have also attached my blockMesh file and the dictionary files under the system directory. Really appreciate if anyone could help me with this problem.

Thank you.

kiehian

Hi,


From what I can see in the very short log file excerpt, I am guessing that your time step is way to large. Your alpha is exploding. Do you use the CFL based time step control?

P.S.: A longer excerpt of the logfile would have been helpful.

Hi Jens,

Thanks for helping me out with this.

As this is based on the example case of dam break with a floating box (floatingObject), I am not too sure what is a good value for U in the Courant number.

However, the time step I have used is much smaller than if I assume the same U value as the example and recalculate the required time step for the mesh I am using. (I used order of 1E-5 vs a CFL-based order of 1E-3).

Below is a longer excerpt of the log file, which includes the last few stable time steps at the set Courant number of 0.25:

Thanks once again.

Hi,

are you trying s.th. like http://powerlab.fsb.hr/ped/kturbo/Op...F/twoBoats.mpg ? I believe there is a extend http://openfoam-extend.git.sourcefor...enFOAM-1.6-ext =>tutorials/basic/sixDOFSolver/sixDOFmotion with bodyA & bodyB
but I am not sure

Hi Elvis,

Thanks for pointing out the two boats case. I have looked at the extend version and actually, my blockMeshDict file is based on the case. However, instead of subjecting the two boxes to a steady flow, what I am plan to do is to subject them to waves. So as a first step, I tried to modify the floatingObject case to see if it works for two objects and if it does, my next step is to implement some form of wave input.

About your crashing simulation, you might want to check the following:

• Gravity pointing into the correct direction?
• Field (alpha) initialised correctly?
• Are there bad cells in your mesh?

Hi Jens,

Thanks once again for your pointers. I've looked into my input files and found that my gravity was orientated along the wrong axis wrt the boats/alpha. However, after fixing that problem and further refinement of my mesh, the simulation still crashes somewhere at the start (around 0.1s). checkMesh says ok and visualising in paraFoam seems fine too.

Attached is a jpg of the alpha1 field at 0s. I'm now trying to re-run with a finer time-step and also with turbulence turned off to see if I can get the simulation to run properly.

However, if you have further insights, please let me know!

Thanks!

kiehian

Again, a new log file would help a lot ;)

Apologies, I was thinking of running the simulation with smaller time step up and running first. However, even with smaller time steps, both the kEpsilon and laminar simulation still do not work...

Below is an extract of the log file from the turbulent run, after the Courant number started exploding (the full log is too long, so I've put it in the attachments). Also attached are the files from my system / 0 and constant directories.

Thank you once again!

I've never used the floatingBodies in this release, so please correct me if I am wrong: Your log file shows the linear and angular velocities of both bodies, right? The linear velocities seem to be unrealistically high. Is it possible, that the motion of the bodies cause the high velocities? Maybe you can try a simulation on the same case but fixinig all DOF of each body. If this does not break, the high velocities and the crash of the simulation may be related to the motion of the bodies.
I hope this helps.

Just an update on the case. I recreated a mesh that's closer to what I'm looking at (more ship-hull-like) rather than two square blocks.

Jens: I saw the high velocities too and was thinking that it may be due to my inputs under 6dofrigidbodymotions. I did what you've suggested, i.e. to fix the two bodies and the case seems to run ok (although it took the past few days of parallel running on my 4 core desktop!)

I am not entirely sure if this is correct, but another possible reason that the original (two floating bodies) case was crashing could be due to the fact that I've used a uniform mesh (grading of 1) throughout, when I should use a finer grading closer to the hull.

I'm gonna to carry on working towards simulating two floating vessels in waves and will try to post more updates of my progress (and of course problems).

Hi CKH;

You could solve the problem ?. What was the mistake? Could you give more information about this.


Best regards

Hey
Did you manage to solve it?
Thanks

Floating point exception error
 

I finally solved it. In my case the problem was about k and epsilon files. The internal field and inlet value was not accurate. Although i have solved this error and the case is running now, it has not converged yet.

Hey,

CFD lover, would you be willing to upload your files? I am also trying to simulate multiple floating objects, but I can't get it to work.

Kind regards,
Daan