# Hydrofoil exercise to validate CFD analysis

Discussion in 'Hydrodynamics and Aerodynamics' started by quequen, Oct 1, 2014.

1. Joined: May 2009
Posts: 16,679
Likes: 346, Points: 93, Legacy Rep: 1362
Location: Cocoa, Florida

leeway?

Forgive me if I missed where you discussed the effect leeway has on a "L" foil/uptip foils. Maybe it's not relevant to this discussion but just in case it is, here is what Tom Speer says on the matter:

The curved part of the vertical foil produces essentially the same lift as it rises. This is necessary to counter the side force from the sail rig, which does not change as the height changes.

Because the horizontal lift is constant but the vertical area is reduced as the boat rises, the leeway angle increases. It is the coupling of leeway with heave that is exploited by the L foil to provide vertical static stability.

The dihedral angle of the horizontal wing is set so that the angle of attack of the wing is reduced as the leeway angle increases. This satisfies the static stability condition that the vertical lift decrease as the heave increases.

Because the same horizontal lift is produced over a reduced vertical span, the sideways wash in the wake is also greater and the trailing vortices are more intense. This causes a coupling with the horizontal wing that increases the vertical lift, because the horizontal wing acts as a winglet for the vertical part of the foil (and vice versa). The dihedral angle required for vertical stability is greater than what one might expect by looking at the wing alone because it must overcome this wake-coupled influence. The result is there is a range of dihedral angles that provide positive vertical stability and a range of dihedral angles that are destabilizing in heave because of the coupling with the shed vorticity of the vertical part of the foil.

Although there are times when the foil tip has broached the surface, this is not the normal mechanism for providing heave stability in L foils. The best performance is obtained with the hull just above the wavetops and the wing submerged well below the surface. The leeway-modulated heave stability is still effective in this condition, and the induced drag is minimized.

Canting the foil inboard has the effect of increasing the dihedral angle of the wing, which enhances the heave stability. The vertical lift is spread over a greater span because the curved part of the foil is oriented to provide more vertical component of the force. This reduces the induced drag due to the vertical force. However, the induced drag of the horizontal force would be increased, so cant is typically used off the wind when the side force from the rig is less and the side force produced by the foils is correspondingly less. The foils still have to support the weight of the boat, so the vertical force is not lessened, but the relative proportions of vertical and horizontal force are changed, making the canted foil better suited to the operating condition. Cant allows the leeway-modulated heave stability to be increased an an acceptable penalty in the induced drag because of the lower side force and the higher speeds, which also reduce induced drag.

Upwind, the foils are canted to their vertical position to minimize the induced drag from the high side force and reduced speeds. The reduction in horizontal wing dihedral angle with vertical cant impacts the leeway-modulated heave stability, which is why it is much more difficult to achieve stable flight upwind than downwind. The crew had to be more active in trimming the wing and foil to deal with the reduction in natural heave stability, which was very hard on the grinders when flyng upwind.

Whether canted or upright, the mechanism for providing natural heave stability was still the coupling between heave and leeway, which led to a reduction in vertical lift because of the designed-in coupling between leeway and vertical lift by virtue of the wing dihedral. Reduction in horizontal/vertical-lifting area due to the foil tip broaching the surface was not part of this primary source of heave stability. Allowing the tip to broach the surface had big penalties in terms of induced drag and increased leeway due to insufficient vertical span.
__________________
Tom Speer

2. Joined: May 2004
Posts: 5,372
Likes: 255, Points: 93, Legacy Rep: 3380
Location: Italy (Garda Lake) and Croatia (Istria)

### daiquiriEngineering and Design

You are right, but the 3-D lift should have been predicted in a reasonably accurate way by Mr. Hanley's inviscid CFD. What is lacking in this picture is the influence of the finite-AR and L-shape on the drag coefficient, which will certainly be much higher for a given AoA than the value assumed from the characteristic curves of the airfoil.

Cheers

3. Joined: May 2004
Posts: 5,372
Likes: 255, Points: 93, Legacy Rep: 3380
Location: Italy (Garda Lake) and Croatia (Istria)

### daiquiriEngineering and Design

Good info Doug, but the current discussion is actually about the reliability and limits of CFD results provided so far. What Tom Speer wrote there is related to the issue of static and dynamic stability of a sailboat with the L-foil configuration.
Cheers

4. Joined: May 2009
Posts: 16,679
Likes: 346, Points: 93, Legacy Rep: 1362
Location: Cocoa, Florida

UptiP foil/"L"

OK-just wasn't sure but I appreciate your answer.....What I don't understand is how you can discuss calculated results for an "L"/UptiP foil in a sailboat application(first post) w/o discussing leeway coupling?

5. Joined: May 2004
Posts: 5,372
Likes: 255, Points: 93, Legacy Rep: 3380
Location: Italy (Garda Lake) and Croatia (Istria)

### daiquiriEngineering and Design

Pushed by the curiosity, I have performed an analysis with the XFLR5 software, using the 3-D Vortex Lattice Method with viscosity included. It should capture all the main aspects of the flow except the free-surface effects. Since I didn't have sufficiently detailed coordinates of the H105 airfoil, I have modified the camber and thickness distribution of a NACA 64A212 in order to make it resemble as much as possible the H105. But the differences could be explained by using a different airfoil and turbulence calculation setup.

The simplified L-foil is visible here:

Can be done much better and with smoother approximation of the rounded corner, but this is just an exercise for me, so...

With the following input data:
V = 8 m/s
AoA = 4°
Yaw = 0°

these are the results of the calculation:
Lift (vertical) = 1691 N
Lift (lateral) = 1640 N
Drag (viscous) = 77 N
Drag (induced) = 55 N
Drag (total) = 132 N
L/D = 12.8

It really surprises me how comparable these numbers are to those calculated manually in my post #28, and even more so to the results of the CFD simulation performed by Mr. Hanley. The lift force differs just by 4% from the results from the Stallion 3D and 12% from the manually calculated value, though the drag force differs from both.

Although some unorthodox tweaking was necessary in order to simulate this non-symmetrical wing with XFLR5, it looks like I have just found a tool for quick and cheap analyses of similar configurations.

Cheers

6. Joined: May 2006
Posts: 553
Likes: 70, Points: 28, Legacy Rep: 258
Location: Finland

### Mikko BrummerSenior Member

VLM is great as long as there's no lifting bodies or separation, very low aspect ratios or lots of sweep. Dr. Hanley's 3DFoil would also do this, probably even easier than XFLR5 (which I've never used). Would be interesting to see what numbers 3DFoil would give.

7. Joined: May 2004
Posts: 5,372
Likes: 255, Points: 93, Legacy Rep: 3380
Location: Italy (Garda Lake) and Croatia (Istria)

### daiquiriEngineering and Design

Neither of which is the case here.

8. Joined: May 2006
Posts: 553
Likes: 70, Points: 28, Legacy Rep: 258
Location: Finland

### Mikko BrummerSenior Member

So I gave up, too, and ran FLowEFD (a.k.a Solidworks Flowsim). With a relatively coarse grid (1,2 million cells) I get:

With the following input data:
V = 8 m/s, inflow turbulence intensity 0,1%
AoA = 4°
Yaw = 0°
Symmetry (mirror plane) at the surface

these are the results of the calculation:
Lift (vertical) = 1643 N
Lift (lateral) = 1623 N
Drag (skin friction) = 44 N
Drag (induced pressure) = 109 N
Drag (total) = 153 N
L/D = 10.7

Runtime on 8 cores was about 2,5 hours. The very small turbulence intensity means plenty of laminar flow and hence smaller skin friction. The average friction coeff is 0,0025. For real sailing this is unrealistic, the turbulence intensity should rather be 3-4%.

There's less lift than in Dr. Hanley's Euler simulation which could be attributed to the boundary layer thickening towards the trailing edge. The lift numbers are close to Daiquiri's XFLR5 calc, while the drag is similar to his hand calc ;-).

The effect of the free surface would be to diminish the lateral lift and increase the drag, as the loading would need to go to zero at the vertical tip, too. It would maybe be more realistic to submerge the foil completely, how much submersion would give the best result is another question.

In the attachments velocity distribution on the surface and the refined mesh. The other is shear stress, showing the transition from laminar to turbulent, from blue to red at about the thickest point of the profile. The other side of the foil looks rather similar when it comes to transition, so with the low turbulence the simulation predicts plenty of laminar flow.

#### Attached Files:

File size:
292.5 KB
Views:
1,444
• ###### H105-hydrofoil-shearStress.jpg
File size:
95.4 KB
Views:
1,122
9. Joined: May 2004
Posts: 5,372
Likes: 255, Points: 93, Legacy Rep: 3380
Location: Italy (Garda Lake) and Croatia (Istria)

### daiquiriEngineering and Design

Can you guys please give me the table of coordinates for the H105 foil? I will try to run the simulation with XFLR5 by using the correct airfoil. As mentioned before, my simulation was performed with a modified 64A212 foil, because I didn't have the H105 data. I will also try to refine the model in order to get the geometry as close as possible to the original L-foil. We'll see what happens then.
I want to discover the reason for different values of induced drag, which should more or less coincide between your CFD and my VLM calcs.
Cheers

10. Joined: May 2006
Posts: 553
Likes: 70, Points: 28, Legacy Rep: 258
Location: Finland

### Mikko BrummerSenior Member

CFD makes no distinction between viscous pressure drag and induced drag... in fact, one could argue there is only pressure drag and skin friction, induced drag is a man invented convention. Or am I mistaken?

In your XFLR5 results there's some pressure (profile) drag in the viscous drag, as well as skin friction, I reckon.

11. Joined: May 2006
Posts: 553
Likes: 70, Points: 28, Legacy Rep: 258
Location: Finland

### Mikko BrummerSenior Member

I re-ran with an inflow turbulence of 4% - little effect on the lift in either way, but there's no laminar flow on the foil whatsoever and drag jumps to 223N. Skin friction goes to 82N, so there's 141N left for pressure drag. In the laminar case pressure drag was only 109N - maybe the thicker boundary layer, and a little zone of separation in the junction (in blue at the trailing edge). I plotted the friction coeff for a change, the average friction coeff is 0,0042 with the inflow turbulence at 4%.

#### Attached Files:

• ###### H105-hydrofoil-FricCoefftur4.jpg
File size:
83.6 KB
Views:
806
12. Joined: Apr 2004
Posts: 892
Likes: 52, Points: 28, Legacy Rep: 422
Location: Finland

### JoakimSenior Member

Setting 8 m/s inflow turbulence to 4% doesn't sound like something you would find in sea unless there is another foil etc producing a lot of turbulence. Remember that in reality you are traveling through still water having turbulence only from it's own motion, which is nowhere near 8 m/s.

In CFD and in real life there are only skin friction and pressure causing forces to a surface. Induced drag is an explanation and a simplified calculation method for part of these forces.

13. Joined: May 2006
Posts: 553
Likes: 70, Points: 28, Legacy Rep: 258
Location: Finland

### Mikko BrummerSenior Member

Good point Joakim. So if you assume a surface layer turbulence intensity of 4%, and you travel at 8 m/s, what turbulence intensity would you suggest for the CFD run inlet?

14. Joined: Oct 2014
Posts: 7
Likes: 1, Points: 0, Legacy Rep: 23
Location: United States

### PatHanleyJunior Member

I did not have the exact airfoil on file but modified E817 to match angle for zero lift. 3DFoil produced the following results:

@ 8m/s and AOA = 4 deg

Lift = 1,687 N
Drag (total) = 143 N
L/D = 11.76

The pressure drag is usually defined as the profile drag -friction drag + induced drag.

Patrick Hanley, Ph.D.
http://www.hanleyinnovations.com

#### Attached Files:

• ###### lfoilmsa.png
File size:
90.1 KB
Views:
583

15. Joined: Jun 2011
Posts: 2,467
Likes: 123, Points: 73, Legacy Rep: 693
Location: australia

### groperSenior Member

This is all great, but is someone going to change the flow to include a yaw angle for leeway anytime soon? I really want to see 2 what happens to the flow coupling between the 2 perpendicular lifting surfaces and the resultant L/D ratio as a whole, and the interference effects of the vertical strut on the horizontal foil and the resultant vertical lift vector with respect to leeway angle...

Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.