Hydrofoil during CFD analysis

[shawn13]

I have been working on the cad models to help generate the moulds for Jellyfish Boards new hydrofoil:

viewtopic.php?f=196&t=2387906

This morning I spent some time putting the hydrofoil wings through some CFD Analysis (Computational Fluid Dynamics) and have produced these results.

I ran the simulation at 3, 6, 8, 10, 12, 14, 16, 20, and 30m/s.

I think the most interesting aspect of the results show that the rear wing generates the negative lift immediately, but loses it when the from wing begins to generate, then regenerates later.

3m/s or 11kph or 7mph

6m/s or 22kph or 13mph

8m/s or 29kph or 18mph

10m/s or 36kph or 22mph

12m/s or 43kph or 27mph

14m/s or 50kph or 31mph

20m/s or 72kph or 45mph

30m/s or 108kph or 67mph

[davesails7]

Are these all at the same angle of attack? How did you determine the angle of attack for each speed?

I think the most interesting aspect of the results show that the rear wing generates the negative lift immediately, but loses it when the from wing begins to generate, then regenerates later.

Why do you think that would happen?

[Tinkersailor]

Ok..intersting. But WHY is the rear wing producing a down force? Are we stuck in airplane mode? Airplanes use contant C of G and manipulate AOA and deflection to produce a safe solution when the forward wing stalls. KBHF do not need to be made this way! In fact all of the negative lift you make on the rear wing needs to be made positive on the front wing....they are fighting each other for no good reason!

I feel that the only reason aa rear wing is needed is to give some feel to the whole thing and stop the rider from overcontrolling. Am i going down the wrong track?

IMO both wings should produce lift all of the time.

Sorry if this hijacks th thread.

Tink

[davesails7]

There is a foil from Gong that just came out that (I think) will be the only foil currently in production to use upward lift profiles on both wings.

Nicolas from Zeeko just recently explained his reasoning: viewtopic.php?f=196&t=2387788&start=20

Will be interesting to hear the reviews on the Gong foil.

[KristianE86]

Are the simulations 2D? Do you use a turbulence model? Did you study mesh dependence?

CFD is hard... really hard... often impossible. That does not prevent a lot of people from saying they can do it in exchange for money

[zfennell]

hi shawn

that's pretty cool. thanks
do you believe the results?
can you tell us why?
like others, i'm a bit skeptical, but anything is possible.
particularly when you consider the number of possible input parameters on most solvers.

my personal impression/guess is that one of the plots is mislabeled and out of sync with the rest.
(that's the kind of mistake i'd make when trying to get some simulation to run for the first time)
it would be easier to compare results if the plots used the same velocity/pressure scale
i'm sure you will be able to confirm as you go thru possible diagnostics.

what cfd code did you use?
is there a Boundary Layer solver?
can you predict transition or is that an operator input?
how does the code deal with BL separation?
can you get velocity profiles of the BL or just the freestream?
perhaps a skin friction estimate, empiracle or calculated?
will the code integrate the pressure distribution around your geometry to give net reaction forces and moments?
was velocity the only input variable in your analysis?
does the answer change with mesh size?

even if this run turns out to have 'issues'
there will be lot of good information and insight to come as things progress.
thanks again.
-bill

[shawn13]

First, I want to state that I'm not an aerodynamicist. I have access to CFD and wanted to see what kind of results I could pull.

I will try to answer the questions the best I can

Ok..intersting. But WHY is the rear wing producing a down force? IMO both wings should produce lift all of the time.

Sorry if this hijacks th thread.

Tink

I did not design the wing. I was given the wing design. From my understanding, it does have something to do with the moment the foil creates.

Are these all at the same angle of attack? How did you determine the angle of attack for each speed?

I think the most interesting aspect of the results show that the rear wing generates the negative lift immediately, but loses it when the from wing begins to generate, then regenerates later.

Why do you think that would happen?

The AoA is zero and the same for each simulation. I needed a starting point. Now, I can start to change the AoA and see what the different results are.

Are the simulations 2D? Do you use a turbulence model? Did you study mesh dependence?

CFD is hard... really hard... often impossible. That does not prevent a lot of people from saying they can do it in exchange for money

These are 3D simulations. I had to do the wings separate from the rest of the foil as it becomes so 3d complex that the simulation errors or needs more imput which I just cannot give.

hi shawn

that's pretty cool. thanks
do you believe the results?
can you tell us why?
like others, i'm a bit skeptical, but anything is possible.
particularly when you consider the number of possible input parameters on most solvers.

my personal impression/guess is that one of the plots is mislabeled and out of sync with the rest.
(that's the kind of mistake i'd make when trying to get some simulation to run for the first time)
it would be easier to compare results if the plots used the same velocity/pressure scale
i'm sure you will be able to confirm as you go thru possible diagnostics.

what cfd code did you use?
is there a Boundary Layer solver?
can you predict transition or is that an operator input?
how does the code deal with BL separation?
can you get velocity profiles of the BL or just the freestream?
perhaps a skin friction estimate, empiracle or calculated?
will the code integrate the pressure distribution around your geometry to give net reaction forces and moments?
was velocity the only input variable in your analysis?
does the answer change with mesh size?

even if this run turns out to have 'issues'
there will be lot of good information and insight to come as things progress.
thanks again.
-bill

Do I believe the results? I have not ridden a foil, but after discussion with riders who have, it does show that I am on the correct path.

These are preliminary bases that will need to be refined. From a simulation point, I had to start with a textbook scenario to develop a base line. I use the same boundary, and my imputs are velocity and pressure. Yes right now I am only changing the velocity imput. The mesh becomes a very difficult and complex calc, and so I use an auto generate. This is giving me issued with the fuselage so I will need to refine for that analysis.

All plots are labelled correctly, which I really why I posted them. I want to get a better understanding from riders why the rear wing loses lift when the front begins to generate. My initial thought would be the turbulent flow off the front generation that breaks the rear line. My understanding is that the wings are positioned in different planes because of the slipstream causing lift failure. But then on the race foils, they do look like they are in the same plane.

The CFD is generating a ton of data that I cannot decipher. Fortunately, I work with one of the leading people in my industry with a phd in advanced analysis. As my simulations progress, I will hopefully be able to answer some of your questions more in depth.


I am looking forward to this thread and the discussions that will come out of this data.

[davesails7]

What are you trying to get out of the analysis?

With any type of computer analysis (cfd, fea...) it's best to do a hand Calc to see if your results are in the right ballpark. If you're finding something that doesn't make sense, like the aft wing lift shutting off and then turning back on again, it's probably something off in the analysis. These types of tools are more for fine tuning the design by doing very accurate calculations. If your results are off by orders of magnitude from what you expected, it's usually not correct.

As noted above, CFD is VERY hard to do. I've never done it myself but I've seen a few very smart people come up with results that made zero sense and assume it is correct because it came out of the computer.

Good luck!

[zfennell]

......
All plots are labelled correctly, which I really why I posted them. I want to get a better understanding from riders why the rear wing loses lift when the front begins to generate. My initial thought would be the turbulent flow off the front generation that breaks the rear line. My understanding is that the wings are positioned in different planes because of the slipstream causing lift failure. But then on the race foils, they do look like they are in the same plane.

.

that's reasonable.
for such an assumption to make sense, you'd need to confirm one of two things is happening to the rear wing.
1)either the average local freestream velocity is significantly reduced compared with other speeds.
or
2) the local flow vector has locally changed the AOA of the rear wing compared to other speeds.

all things being equal ( which apparently is not true)

just about everything you have plotted (and may plot) should scale with the square of velocity
pressure, potential flow, drag, lift, etc are all proportional to V^2.

the 3m/s and 8 m/s runs don't seem to follow that trend.
right now your data is suggesting that the lift coefficient of the rear wing at constant AOA is not smooth (or quadratic) function of reynolds number or velocity.

is there any chance that some plots output element averages as opposed to nodal absolutes?
how large are the mesh elements compared to your geometry?

still cool,
-bill

[BWD]

Intuitively my take is accelerated flow from over top of the front wing approximately matches the velocity across the top (high pressure side in this design) of the stab in a useful velocity range, reducing stab lift and drag in that range..... Naturally as an unschooled Neanderthal in the area I may be overlooking a few details...

But variations in fuse length over the last while of developing current racing foils may support this idea, or not, what do I know?