I use Pro/E to model and build kiteboards. I would like to do some simple flow visualization studies with inexpensive software.

Thanks for any ideas or links,

Hein

www.impact3d.com/kitemain.htm

What do you want to visualise? Flow of air over the boards, or water or both?

Thanks for your reply,

I want to seet the flow of water underneath the board. I think it's called a fully ventilated surface.

Here's a picture that shows the angle of attack:

http://www.kiteforum.com/phpbb/files/spleene34_789.jpg

Not one of my boards but it shows what I want to see simulated. No need to model the waves. I'm really more interested in the distribution (direction) of the flow as it travels along the bottom.

-Hein

And here's an image of one of my bottom shapes: Note the four fins set with 2.5 degrees toe in.

http://www.impact3d.com/Hein_single_to_double.jpg

-Hein

You don't want a flow visualization package; you want a flow simulation package (visualization implies you already have the flow field). This is actually a challenging (but interesting) project. I would suggest that the software expense should be the least of your problems/concern!

Adrin Gharakhani

Thank you,

From what I know I must: Define the flow volume and geometry (I've got the board model and could probably define the water surface. Mesh that model and for use by a solver where I'll define the flow, density, etc. Then display the results with a visualization package

Simply said but not simply done I'm sure.

Still worth a shot. I got plenty of time to learn.

-Hein

OK. It really all depends on the level of accuracy and detail you're looking for. For example, are you going to expect a fixed position/immersion of the board with respect to the water surface or are you interested in the flow dynamics as the board interacts with water? (boarding moving in/out water). Do you intend to include fluid-structure interaction (where there is a force feedback between the board and water), or will you simplify by assuming a user assigned interaction? Are you interested in unsteady dynamics (which this is) or do you wish to over-simplify by assuming "steady state"? Can you live with potential flow assumption (for the free-surface component), or do you absolutely need to include viscous effects (I personally think potential flow will be the route to take, at least as a first-second step)?

The answers to these will help you determine which type of software to look for.

Adrin Gharakhani

I would suggest that you contact either FLUENT, STAR-CD or any nearby consulting engineer. Fluid Mechanics is a difficult application of science that needs a lot of experience in order to make sense of the results, as has been posted before you need to get a solution before visualizing the results. Cheers and Good luck.

That was what I was getting at, if you want to model the surface of the water then you are looking at a very computer intensive simulation. If you are looking at just the water with no air then it shouldnt be too bad.

Adrin, i would have thought viscous effects would be of interest to Hein as in such an application drag on the board would be fairly important and key to optimisation.

Hein, I think what you are looking at is the first steps into CFD and it isnt a field you can jump into that easily, personally I think that your best bet is a CAD integrated tool for a start something like STAR-ProE might be the way to go if you are already using Pro/E. A simple package such as this will then give you some experience allowing you to move on into more compelx situations.

http://www.cd-adapco.com/products/ST...STAR-ProE.html

Thanks for all the useful information. I've looked at Fluent and now also Star-CD. The former looks like a good place to start. Maybe I can get a trial license and see how it goes.

I'm really mostly interested in seeing what the flow looks like to confirm/deny what I beleive is happenening. Being able to watch the flow coming off the board while you ride is great experience.

I beleive the toughest assumption may be to come up with a waterline on the bottom surface of the board.

I dont really understand how you would handle the air/water interface. And the rooster spray - That has to be a tricky thing to model. I remember seeing something about physics based particle annimation. Is that similar to what is used in CFD?

Thanks to all for your interest in this project. I am a kiteboard builder with a strong CAE background and a lot of mechanical engineering experience.

Thanks, Hein

I've also considered using oil flow visualization. Anyone know where I can get the right oils/dyes for that?

I believe oil flow visualization will be quicker and more informative in your case. The problem you want to solve is one of two-phase flow (air-water), with a sharp interface. I don't know much about what fluent or star-cd offer these days, but you need a volume-of-fluid (VOF) capability for this (or a similar technology) to be able to track the interface.

You've brought up the main sticking points yourself - the waterline on the bottom surface of the board and the rooster spray. These are _not_ information you want to give to the solver, but information you want to obtain from it. If you already know the waterline, then you pretty much know the flow characteristics (you can obtain the waterline with VOF, by the way).

A simpler (but similarly powerful) first step for you would be the assumption of "potential flow" - well, it's not really potential flow but a vortex flow in the limit of zero-thickness vorticity surface. The air-water interface can be described by a vorticity surface there, and its evolution obtained by accounting for the dynamics of the surface vortex flow. This can be done by a panel-methods-based potential flow solver (equipped with the right boundary conditions, etc.).

As for the particle based flow animation, you are almost correct. Particle based methods will be the most natural strategy for your problem. The issue is what particle method to use. What you have seen, I'm sure, is based on the pressure-velocity formulation of Navier Stokes equations. However, for your problem a vorticity based formulation will be the most efficient and informative.

Adrin Gharakhani

>>Adrin, i would have thought viscous effects would be of interest to Hein as in such an application drag on the board would be fairly important and key to optimisation.

Ben, given the Reynolds number, viscous effects can be neglected for the most part except for a viscous model _at_ the surface. That is, for all practical purposes, you can use a zero thickness boundary layer with all its vorticity concentrated at the surface. This will allow you to solve the problem using panel methods. The only problem is that you cannot account for separation accurately (you have to actually know in advance where separation takes place, which in this case may not necessarily be a major problem). By the way, the reason I suggested the panel method, in the first place, was to reduce computational complexity _significantly_.

Adrin Gharakhani

Certainly the rooster spray will be tricky to model, VOF is a fairly computer intensive process as it is but to model the atomisation too would require ultra fine grids behind the wakeboard geometry. Certainly STAR has VOF capabilites, I cant speak for fluent but when it comes to free surface modelling COMET is the undisputed king.

I would warn against jumping foot first into a full comercial code Hein, certainly if you have no CFD expirience and you want to model a industrial type case you are going to cause yourself a lot of bother. There is obviously Adrin's sugestion of a potential flow method but in the long term I feel that it will be of limited use to you in modelling a broader range of problems.

The sales guys from fluent (or any of the codes for that matter) will tell you how wonderfully easy it all is and how you will be getting full simulations within weeks, but don't beleive them. CFD isnt something you can just start with no experience, you need a reasonable knowledge of fluid mechanics at the least and preferably some training on general aspects of CFD modelling (not code specific). I still maintain that your best starting point is a CAD embedded tool such as STAR-ProE, Cosmos Flow-Works or the like. These are cut down versions of the full codes and will give you the functionality you require to get a knowledge of how CFD works and should get you some good results too, you can always move on from there.

i do not know how much my poking nose here is needed, but this is from true experience. My compnays other banch in another city has starCD, and they were doing one particular case with VOF and two months and no convergence the solver just blows up. Finally i asked them to send me details, and with fluent it finished in no time, they gave me details on friday, monday i gave them converged results (so for VOF robustness of solver is important).

This is true, STAR has been a bit variable with VOF in the past, as I say COMET is the undisputed king of VOF and is pretty much the most widely used code in the marine industry. I know COMET is dying soon though and being assymilated (sort of) into the next big release of STAR (v4).As far as I understand it the vof algorithms are one of the major transfers from comet to star 4.

Although to be honest I think vof is somewhat irrelevent in the current discussion, Hein needs to get using a nice simple CFD code first and teaching himself what CFD is, and if he wants to model the spray behind his boards he will have to be investing in a supercomputer whichever code he uses!

Spray/atomization is irrelevant to this problem, in my opinion - unless it is for visualization effect in which case the accuracy is unimportant. I think we're getting too deep into the discussion here without even knowing what the main objective of this simulation is (it wasn't mentioned in the original message)

Adrin Gharakhani

Thank you all very much for your insights.

Kiteboarding is a unique sport. It really is the simplist system for balancing lift across a wing against lift from the board.

On unique characteristic is the ability to throw the kite (sail,wing) and generate power from the speed and then harvesting that into a tug to get you planing on the board. Apparent wind across the kite and the reduced friction of the board lets you glide along in very light winds. Using lift from the kite and pop from the board (and a wave) can combine for some huge air.

I use smooth C2 surfaces on my boards and it sure seems to make them go easily with less kite power. I often use a smaller kite and ride with more speed than my buddies on large kites with larger boards.

I am interested in seeing the effect my bottom contours have on directing the flow. I'll stick to fixed angle of attacks and not worry about spray.

I'm leaning towards Star-CD for now. To get my feet wet.

Thanks again and keep the comments coming. I am learning a lot right now!

-Hein

> I am interested in seeing the effect my bottom contours have on directing the flow. I'll stick to fixed angle of attacks and not worry about spray.

In this case, I strongly urge you to look into the so-called potential flow approach (using panel methods). With the panel method, all you have to do is assign the surface of the submerged section of the board and the water surface in its neighborhood. So you end up having to solve the problem on just one continuous surface; no volumetric meshing of the water, etc. would be required.

Adrin Gharakhani