OBJECT IMPACT ON WATER

[Robin Noel]

Hi folks, it's me again, now querying for opinions. I want to simulate a cylinder entering a water volume with an initial velocity.

I tried using the ALE module, without success. Can anybody think of a model that could achieve said results ?

The main trouble comes from finding what boundary to apply to the cylinder wall, since I want an internal coupling on it, and the wall boundary conditions makes my mesh immobile. But I want to study the displacement of the cylinder !!
(disclaimer : I have never worked on a FSI case before, there is quite a lot of notions I'm not familiar with)

I just need to retrieve the mesh velocity and displacement.
Counting on you people's brain matter.

Thanks by advance,

[Yvan Fournier]

Hello,

The newer Volume of Fluid features might be adapted to this. There are new immersed boundary features in code_saturne which could be good fit for this, but I am not shure what is usable at this stage (I'll check).

With Volume of fluid, It would probably be simpler to consider the solid "fixed" and have the water level rise (i.e. place oneself in the objects reference), though this might be a bit tricky (i.e. if the object rotates, which it probably will, the gravity vector would have to be changed at each time step, but that can be done easily, even if we do not usually do this). The acceleration would probably have to be integrated with the gravity vector also...

With immersed boundaries, we would not have these restrictions.

I'll suggest to colleagues more familiar with those features to look at this to see if they have any suggestions.

Best regards,

Yvan

[jcharris]

As a part of the DIMPACT project <https://www.france-energies-marines.org ... s/dimpact/>, for some CFD work at Ecole des Ponts and France Energies Marines, we have done some simulations very similar to this. Some part of this has been already published : <https://hal.science/hal-03721266/document> ; <https://www.researchgate.net/publicatio ... aking_Wave> More work should be published soon.

Alternatively, some years ago I know others did similar things with code_saturne with other approaches : <https://www.youtube.com/watch?v=lBHtzSeRpk8> ; <https://www.youtube.com/watch?v=Ioanqa-thC4>

If you can describe in more detail, I can perhaps provide some advice. There are a number of things that work well, and a number of things that might not. And depending on the physics that you care about, code_saturne may or may not be easy to get working for your case. (Everything is possible, but some things can be working tomorrow, others require developing the inside of the code.)

Are you imagining a fixed velocity, or a freely-moving body? If it's freely-moving, is this with 6 DoF? Is the water volume calm or moving? Do you need just the impact, or for a long duration after? There is mention of cylinder coupling... is this with a rigid body or a flexible cylinder? Or since the question was can any model achieve the results, do you need it to be in code_saturne, or could a simplified model be better suited? If you want code_saturne, do you care about the boundary layer behavior -- what kind of Reynolds number are we talking about?

[jcharris]

Another reference that might be of interest : http://venus.ceride.gov.ar/ojs/index.ph ... /6439/6560

[Robin Noel]

Hi again folks, thanks for the answers.
I'll look into VoF(I had my eye on that feature). Mr.Storti result are exactly what I need and my next plan is to get in touch with him(these videos are awesome). That ongoing research work at France Energies Marine seems quite appropriate; I really need to get into VoF theory.

My sim itself can be 2 dimensional, with a fixed initial velocity calculated from another model I have. I would then apply only gravity to my 2D cylinder(and calculating speed based on it and the forces that water applies), in hope of getting the evolution of the speed when entering the water volume.
I do not necessarily need for the 2D cylinder to dive deep, I mostly need the impact velocity loss, and maybe a meter deep afterwards.

I was using FSI to try and get the fluid forces applying on my rigid cylinder.To no avail. I mainly used Code_Saturne because of prior experience with it, although I'm open to any model suggestions. I had an initial Reynold Number in the 10^4 range.

In hopes of hearing more from you folks,

[jcharris]

If you could give more details, perhaps I can try to give a more detailed reply... if Re~10,000 and you want to dive 1m... does that mean the cylinder diameter is 1mm or 10m? Can the cylinder rotate or is it fixed? Is it rigid or flexible?

Modeling a cylinder in CFD can be either very easy or very complicated, depending on if you need the turbulent wake, if you need surface tension, if you need peak pressures, if you need flow visualization, etc etc...

If you just need a simple model of velocity decrease, I recommend that you first check if it might be easier to look up some engineering models and do some matching between an initial impact model and a simple drag coefficient after the initial impact.

You may find useful references to start a search in Section 3 of <https://www.researchgate.net/publicatio ... structures> or from the classic experiments of Greenhow and Lin <https://apps.dtic.mil/sti/pdfs/ADA161079.pdf>

That is not to discourage you if you want to attack the problem more completely... for a recent example, a group used OpenFOAM to do <https://www.sciencedirect.com/science/a ... 1822027998>. From results that I have run and seen with code_saturne, it is certainly possible but might take a bit of time if you want to reproduce something on that level.

Based on what I think you're trying to do, I would first take a moment to look around on Google for 2D cylinder water entry... it is a well studied problem. If you don't have much time, I suspect that CFD impact simulations will be a bit "heavy" for the quality of the result that you will get. But if you do take it on, I would be curious to discuss more!

[Robin Noel]

Hi, I appreciate the time you take. My Reynolds is in the between 10^4 and 10^5 ( depends on the result of my previous model which I'm finishing up as I speak ).

The cylinder would have a round nose, would be completely rigid, and I would imagine the cylinder could have a rotation on the Y-axis (see graphic).

I would like to compute a pressure/constraints or a force at the water entry point, in order to calculate the constraints on the nose.

This model will of course be coupled with an experiment to deduce a vague constraint applied to the nose in real life conditions.

I am of course not avoiding computing a simpler modal in Octave or else, although I'm not familiar with fluid-structure interaction theory, and would have hoped to be able to setup a model in Code_Saturne to familiarize myself with the notion while not losing too much time on setting up my own model. That is still of course an option.

I'll lookup a simplified model for water impact. I'm still very much interested in trying to simulate it in CS.

As a side note, I'll try and see if Code_Neptune could help, as it seems they have free surface interaction with rigid bodies.

Stay awesome folks,

[jcharris]

Ah, so then this is motion in 2D, but a 3D flow (so not really a 2D problem)?

If you would accept a fixed velocity of motion, to get the force / impulse, I have some experience with this from the VOF side, and from what you are saying may be enough for this initial phase of impact. Overall, it may be a little time-consuming depending on your computational resources, but it is possible.

I agree with Yvan that if with VOF, a fixed object (i.e., a reference frame fixed on the body, with the flow moving past) would be simplest to implement, etc. (Depending on how long of an simulation you want, I don't know if the effect of gravity will even be important? - that could simplify the setup further if that can be neglected.)

[Robin Noel]

I will experiment with VoF, with a moving body of water. In the meantime, I extend my thanks to you folks.
I'll keep updating, and will make sure to share the model I'm using on this forum.

Stay cool folks,

[jcharris]

Any updates on the impact case?

You may find <https://doi.org/10.1016/j.coastaleng.2023.104397> a useful reference.

(For free access the first month or two, you should be able to also use : <https://authors.elsevier.com/c/1htby1M2DVKRU1>)