Water flow over a hull - Which model to be used
 

Hello everyone,

I am a student in merchant marine, and I have to study water flow around a VLCC ship hull. I started with OpenFoam and CFD in general 1 month ago, so I am complete beginner.
Here is why I need your help.

So, for now I am studying the waterflow coming perpendicular to the hull at 3m/s. The most important in my case is to know the velocity and direction of the flow.

I did a first try using SST k-omega model, results seem good.
For U : Internalfield uniform (3 0 0), Inlet and outlet are freestream (3 0 0) and walls are slip
For p : internalfield uniform 0, Inlet and outlet are freestream pressure and walls are slip

Next step is to compare with a k-epsilon model.

What do you think about the best model to study the case, and what boundaries conditions to apply?

If someone is used to this kind of simulation, some advises would be helpful.

Cheers,

Attachment 43698

Attachment 43699

Similar problem
 

I'm doing something similar with mixed results. I found that the DTChull tutorial is a good starting point. However, I have had problems with that meshing approach as the boundaries refineMesh creates between levels of refinement cause discontinuities that slow convergence. I get better results using a uniform/simple gradient starting mesh and refining around the hull using snappyHexMesh only.

Although interFoam is giving reasonable estimates of the flow/wake behavior, the viscous drag is too low. I'm still working on that...

Hi Both,

Nice to see other people doing ship flows! A few comments about getting started and based on what I've seen about your approaches:
- Are you running full- or model-scale? If the latter then perhaps using a low-Re (no wall-functions) setup would help with getting the viscous drag under control
- your mesh seems a bit small, normally we'd be looking at at least 1.5 Lpp in front, and 2 Lpp to the side and downstream when considering waves
- did I understand correctly that you're running the case with flow coming from the side? If yes then you may expect RANS to give strange results due to flow separation and adverse pressure gradients, tread with caution

I agree that the DTC tutorial is probably a good place to start, I never used that one myself though. Perhaps more popular are the KCS, DTMB, KVLCC2, and the new JBC:
http://www.t2015.nmri.go.jp/

All the best,

Artur

hello,
I am also trying to create structure mesh on jbc hull through POINTWISE. but I am quite a novice and can not do it. Do you have any tutorial which I can follow?

Hi Marium,

Given that it's a very complex geometry, especially with the energy saving device, I'm afraid it exceeds the complexity of most tutorials you'll find online. You're welcome to have a look at my grids I made for the 2015 Tokyo workshop here:
https://www.dropbox.com/s/rd1dw9x7vn...kyo15.zip?dl=0

Mind you, I recall that the scale might be a bit off, so please do check all of the dimensions carefully (if there's indeed a mistake, it'll be quite small. It was too long ago for me to remember exactly, though).

All the best

A

Thank you Artur for sharing your meshes! I've been looking for examples of meshes of ship hull cases. Yours look very impressive.

Hi,

You're welcome, glad they could be of use. The trickiest thing is always the geometry, actually. Most of the well-established cases (KCS, KVLCC2, JBC, ONRT, DTMB 5415) have geometry that's easily accessible online but rarely suitable for making high-quality grids right off the bat, as there are a lot of gaps, sharp corners, misaligned surfaces, etc. If you're going to be making a lot of grids of various ships, make sure you brush up your CAD skills :)

Good luck,

A

I agree. I spent hours to obtain an *.stl file out of the KCS *.iges file that OpenFOAM was happy with. The tool surfacePointMerge proved useful for that. Unfortunately despite many advantages, NURBS also have some drawbacks when used to model hulls of ships. The biggest problems I have are when surfaces form cusps where several control points are collapsed into one. But sometimes it's very difficult to avoid triangular surfaces and there's no way around that.



What surprises me somewhat more about these hulls is that one would expect them to be of excellent fairing quality, given the authoritative sources they come from. Yet, at times they are surprisingly bumpy, even though not too much.



Do you happen to know any validation case of a hull at full scale? These model scale benchmark hulls are very useful, but ultimately I want to be able to calculate the resistance of hulls at full scale wiht reasonable accuracy. Thanks for any hints.

I couldn't have said it better myself :)

Here's a very famous full-scale case:
https://www.dropbox.com/sh/7tglouyb5...ds-BgqO8a?dl=0

It was introduced at a workshop in 2016 by my former next-door colleagues from Lloyd's Register. They are also currently working on new measurements to be used for validation at the upcoming Wageningen 2021 workshop:
https://w2020.marin.nl/announcement/

Since it's full-scale, there's obviously no resistance data, only ship speed and prop rpm at self-propulsion. The folder also includes submission files of me and my former colleague. If I recall correctly, our skin friction was a bit too low. I couldn't find the case file right off the bat, however, so I can only share what was given to workshop participants.

Cheers,

A

Hello Mr Artur,
I have slight problem to download the dropbox, do you have any mesh for JBC hull? i created a mesh for jbc and imported in starccm+. the problem is that the simulation is not running due to negative volume cell and reversed flow. Can you help me in some way? Thanks again

Hi Marium,
Sorry to hear that you are having issues with the link, I checked it from another laptop just now and it works fine. I think it's something on your end, I'm afraid. Also, I don't know much StarCCM+, but negative cell volumes are always a bad sign and should be eliminated before you even attempt running your case. Since it's a meshing issue, I suggest you make a separate post in a dedicated section of the forum in order to get help with it. But please provide more details since it's impossible to offer you any meaningful advice with just the short message you posted.
Good luck,
Artur

I have already post a thread about it but have not got any response. Anyway, I created this mesh in pointwise and imported it in starccm+. The simulation runs for 252 iterations and then the same error ( i have tried with other mesh too). As you can see both star and pointwise say there's some negative volume cell. Is there any way I can remove them?
Attachment 77434

Attachment 77435

Attachment 77436

Attachment 77437

Well yes, indeed Pointwise is telling you that you have some negative cells. Best course of action would be to inspect them by adjusting the colour bar range in Pointwise so that it shows you cells below a threshold value of zero volume. Then you'll have an idea of what's wrong, it could be many things - incorrectly defined blocks, folded surfaces, you name it.

Thanks, I was trying to create another block. But because of the highlighted deck and the pink domains, the block is not creating. Can you suggest me any way to proceed for choosing the faces so that I can create one block?
Attachment 77463

Hard to say with a picture from so far away, but I'm suspecting your issue can be tackled the same way as they deal with the transom and sonar dome in this seminar:
https://info.pointwise.com/webinar-s...ng-simulations

I wanted to create a block with domains, The problem is that the deck and the symmetry part can not be included in one face. I am not creating block by extrusion LIKE the video you gave me. I created the domains first and then trying to assign them in one block. Can you tell me how to proceed with the deck and the symmetry plane with it?

https://drive.google.com/file/d/1xoH...ew?usp=sharing