Dear all

It would be useful if those of us out there who have an interest in modeling hydrofoils in Fluent or even in other packages (although I use Fluent), make ourselves known to each other, so we may help with advances in the modeling of this most simple, but at the same time complicated object. Anybody who thinks that they could benefit from this please respond to this posting, and we can start to discuss the problems-I am hoping mine are the same as yours.

Thanks

Andy

(1). It is a good idea not to mention the code your are using, but to define your problem and interest in great detail. (2). In this way, readers can get a picture beyond the word of "hydrofoils". Do you think that people know what you are doing?

Dear all-John

My intention was to find other users that are trying to model hydrofoils, so that we can discuss the problems involved in this type of problem. It has been my limited experience of forums that you spend half and hour to an hour detailing your problems and thoughts, only to not get a response. Therefore my approach this time was to try and identify people with a common interest so that I can speak to them and not bother everybody else in the forum with the problems.

However, my problems and thoughts on this subject of hydrofoils are many, and as I mentioned above there is not much point detailing them all. Also I am a student about to go into my final year at University and at the moment I am on an industrial placement with Rolls-Royce for the summer, it has also been my experience that my problems are somewhat stupid or easy to answer, such that I never get a response-but here goes anyway: 1. Anybody used Xfoil, and does it's results compare to commercial CFD. 2. What is an ideal domain size for modeling a hydrofoil, if you do not what to replicate a test facility? I am using 3 cord lengths in front and above, and six behind, is this big enough? 3. I am using the S-A turbulence model, which gives good results from my point of view-is it the right one? 4. Has anybody ever split there mesh at the point of turbulent transition, and only apply the turbulence model in that part of the gird, and let the rest be laminar? 5. I am modeling this as steady state-is it a good choice? 6. And so, and so forth!

Cheers

Andy

(1). I did a quick search on Internet for "xfoil", and the information told me that "xfoil" is a fast viscous/inviscid, subsonic airfoil development system (code). For Hydrofoil applications, you will have to make sure that it can also model the incompressible flows. For the viscous/inviscid assumption to be valid, you can only use it for attached flows. (2).The computational domain size you mentioned is kind of small. You should try to make the domain as big as possible, because of the elliptic nature of the flow. As the angle of attack is increased, you will see the large change in flow field ahead of the airfoil (or hydrofoil). Originally, I was worrying about the free surface effect, but if the hydrofoil is deep under the surface of the water (like submarines), then I guess, the free surface is not a problem for you. (3). You are free to use any turbulence model, as long as it is giving you good results. (4). Modeling of laminar/transitional/turbulence boundary layer flow requires special attention. You will have to read more about the subject first. (5). Steady-state approach is fine as long as the flow is steady-state. But if you run into flow separation, it can easily become unsteady oscillating flow. It depends on your flow conditions.

Dear John / all

You reading up on Xfoil is appreciated, and you are right as far as I understand it about the limitations of Xfoil. Three interesting points about you last comment, firstly, my hydrofoils are attached to boats, and therefore the depth from the free surface is about 5-6m on average. This is something I had not considered! The size of my domain-is such because of the limited computing power available to me, unfortunately I cannot afford to model this is a large domain. My flow is separated in fact if I view the stream functions, it separates after about 5 degrees-will this explain the converged but oscillatory results that I am getting. For example, my lift plot oscillates around set values from say 0.952-0.959, I am just taking an average and putting the oscillations down to something else, it sounds like you have just told me what it is-must be the separation, or could be my high aspect ratio in some cells. I have tried splitting my grid at the point of laminar to turbulent transition, at the point on the upper surface of the foil. I am not doing this anymore because a) I don't fully understand what I am asking the solver to do b) the flow into a hydrofoil can never be considered as laminar, c) and lastly the results where very oscillatory.

I am at the moment only modeling in 2D, but when I go to 3D I should I guess start giving the surface a roughness, turn on gravity, tell it that it is submerged below a set depth, with a free surface, give it a hydrostatic pressure, and loads of other stuff.

Cheers for your thoughts

Andy