# Foil section design using xflr5

Discussion in 'Hydrodynamics and Aerodynamics' started by hashtag_laeuft, Jan 26, 2023.

1. Joined: Aug 2017
Posts: 34
Likes: 13, Points: 8
Location: Hamburg

### hashtag_laeuftJunior Member

Hi together,

for my self-built 16 foot beach cat I would like to design proper daggerboards.

In various threads it is written that it is best to design your own profiles for exactly your requirements. So far, however, I could not find any hints or tips on how exactly to proceed and pay attention to. Apart from choosing a base section and then customizing it to your needs.

I would like to use xflr5 for this purpose.

What are my design criterias?
• Total Sail Area: 18,7 m²
• Hull hight at daggers: 0.55 m
• Total dagger length: 1.5 m
• Wetted length: 0.95 m
• Root Chord: 0.18 m

• Max. speed upwind ca. 15 kn / 7.7 mps (board fully down)
• Max. speed downwind ca. 22 kn / 11.5 mps (board half down)
• Leway upwind ca. 4-5°
• Leway downwind ca. 2-3°

What I would like to achieve
• Max. L/D for upwind sailing
• Good pointing abilities
• Good tacking abilities

Any help and hints are highly appreciated.

Many thanks and best regards
Nicolas

2. Joined: Oct 2005
Posts: 461
Likes: 28, Points: 28, Legacy Rep: 97
Location: France

### ErwanSenior Member

Guten Tag and Welcome to the forum

Induced drag is probably more important than section drag, a centerboard planform with elliptical leading edge at the tip and a straight trailing edge could be a part of the solution to minimize induced drag, everything else equal.
The centerboard position relative to the rig could help minimizing ID,as you can use the rudder blade to help the centerboard.
Just for the purpose of the demonstration if you consider the following alternative/
1: Centerboard =100% lift Rudder =0%
2:Centerboard = 50% lift Rudder= 50%
the case 2 will have half of the induced drag compare to case 1

The reason is the ID is a square function of the lift

Please find below a document for your "homework" and do not hesitate to visit tspeer.com

Cheers
EK

#### Attached Files:

• ###### Minimum Induced Drag of Sail Rigs and Hydrofoils.pdf
File size:
176.6 KB
Views:
208
3. Joined: May 2008
Posts: 2,728
Likes: 492, Points: 83, Legacy Rep: 1082
Location: Beaufort, SC and H'ville, NC

### philSweetSenior Member

Only when taken in isolation. There will be interference effects. The rudder operates in the downwash of the daggers. As a first approximation, this entirely cancels the induced drag reduction. In practice, this can be mitigated somewhat, but you still don't want the rudder running as high a lift coefficient or producing as much lift as the dagger board.

One intuitive explanation of induced drag is it is a result of the lift vector being tilted backwards due to the downwash off the rear of the foil. In a tandem arrangement, the aft element starts out at the full downwash angle of the forward element, and the lift vector from the aft element is now rotated even further aft. The rear element will have three times the induced drag of the forward element if they are similar in plan and generating the same lift.

Another way of stating this is that for a given lift force, induced drag depends only on span. Cord doesn't matter. Dividing the cord up into pieces doesn't matter either.

4. Joined: Aug 2017
Posts: 34
Likes: 13, Points: 8
Location: Hamburg

### hashtag_laeuftJunior Member

Hi together,

But, as Erwan also wrote, these are more related to the later planform design of the daggerboards. I will create a separate thread about this. Therefore, this is already very interesting information, just not quite suitable for the actual topic here.

I would like to learn about how to customize a profile to a desired condition.
In this thread, Rudder and daggerboard profiles. https://www.boatdesign.net/threads/rudder-and-daggerboard-profiles.937/, Tom Speer suggests among others the Eppler E836, which has a rather small drag bucket at Re = 2M. Furthermore, it is also written about widening the drag bucket by inverse design.

Therefore, I wonder on which adjusting screws for this must be turned and on what attention must be paid.
What characteristics prevent the section from stalling and what is required to come out of a tack well?

All help/hints appreciated.

Many thanks and best regards
Nicolas

5. Joined: Oct 2005
Posts: 461
Likes: 28, Points: 28, Legacy Rep: 97
Location: France

### ErwanSenior Member

You are perfectly right Phil, my purpose was more to focus on the ID as a squared function of lift and of course with the implicit assumption "taken in isolation" as you wrote, and I use the twice 50% case vs the single 100% case because in a second you can compute the relative Induced drag: 2x(0.5)^2 vs 1
In practice I don't know may be 10 to 20% on the rudder, but obviously not 50% for sure.

Thank you so much for this crystal clear explanation.
Is it Munk? If it is the case I can understand your explanation, while I never understood Munk's one in his seminal workpaper. Thks again for this one.

Nicolas,

I am afraid that
is difficult to achieve at the same time with
But I am not a serious CFD specialist I just understand basic concepts, I know very little about XFLR5, mostly use the Leading edge radius/ thickness and Thickness position to explore a wing section around the seed one, I did not try seriously with the Inverse design functions and cannot help you seriously on this point.
The basic I understood is smaller LE radius and max thickness moved a little backward can reduce the drag very marginally??
May be a bigger LE radius minimize high velocity on the extrados and could improve resistance to stall but here again I suggest you to check, I am very far from being a specialist.
If your 16 feet rules box allows it, canted or curved daggerboards + T foils on the rudder blades would improve the package's performance much more than refining the sections.

Cheers

Erwan

6. Joined: Nov 2004
Posts: 1,452
Likes: 459, Points: 83, Legacy Rep: 124
Location: East Anglia,England

### wet feetSenior Member

It can be a satisfying process to design something with the precise characteristics that one desires.In this instance I would begin by selecting a few established sections with the thickness/chord relationship that the hull imposes and run some comparisons.There are likely to be some strong possibilities,and from these some indicators will become apparent.Which leads to the point that extremely small dimensional differences are significant.Is the OP confident that he can get the daggerboards manufactured with a high degree of accuracy,and in the case of a laminar flow section,with a high quality finish?If the answer is negative then the whole exercise fails to achieve the initial aim.

7. Joined: Jul 2021
Posts: 506
Likes: 213, Points: 43
Location: La Rochelle (Fr)

### Alan CattelliotSenior Member

Hi,

every comments here are very valuable.

Wet feet : Selecting different sections, doing some comparisons runs
Erwann : Varying the LE radius and the max thickness position, together with pointy tip with a straight trailing edge
Phil Sweet + Erwann : minimize induced drag by considering the higher span

I'm not used to with XFLR5, but there is an additionnal simple care that you can have when defining planform and section of the daggerboard, if this is compatible with your wells :
Use a different profile in way of the tip, a profile with a very low CD, that have also the lesser CL.

Cheers,

8. Joined: Nov 2004
Posts: 1,452
Likes: 459, Points: 83, Legacy Rep: 124
Location: East Anglia,England

### wet feetSenior Member

The original request seems to be a pursuit of the holy grail.Good performance after a successful tack and a section that optimises VMG are not qualities that I have yet discovered in the same foil.My belief is that you need to honestly decide the priority.I have friends who sail on small lakes and for them it is vital to make gains to windward and to stay upwind of the competition,a laminar flow section that needs three or four boat lengths to establish a flow regime isn't helpful.The same people lose out if they race on the sea against boats that have foils that use laminar flow to reduce drag and which gain from it over a long beat.What I do find interesting is to ask the people in the business of making foils what family of sections they use and it is a bit depressing to hear them talk about something with the widest section about a quarter of the way aft and then to waffle on about how pretty the cloth weave of the sheathing is.The exceptions are those who can say they have a NACA section or an Eppler derived section-even if they don't specify precisely which section,as they have obviously done some research.At least with the proliferation of CNC routers it ought to be easier to achieve an accurate outcome than was formerly the case and it helps to give the programmer a clean 3D model with an allowance for sheathing.

hashtag_laeuft likes this.
9. Joined: Jul 2021
Posts: 506
Likes: 213, Points: 43
Location: La Rochelle (Fr)

### Alan CattelliotSenior Member

In fact, you don't prevent stalling. You manage stalling and stalling recovery, at all angles. Here are some general tips for the selection of profile and the planform. Then I will send you some tips regarding the local foil slope and curvature, regarding sensitivity to stall and transit into turbulence.

Last edited: Jan 30, 2023
10. Joined: Aug 2017
Posts: 34
Likes: 13, Points: 8
Location: Hamburg

### hashtag_laeuftJunior Member

Hi together ..

As Alan already wrote, very valuable. Thanks.

These are very helpful for the later planform design as well.

I also already noticed that very small differences in the section profile have huge impact on the performance of the sections, when comparing them with tools like XFOIL or xflr5. I also guess that the biggest lever for improving the overall performance will be the planform design in order to minimize ID.

For manufacturing the daggerboards I'll 3D print a positive model from where I can create a negative mold. The individual peaces will be 20 cm high and turn out quite accurate. This will be a high polished plug for the molding.

I also like the idea to use a different section for the tip.

I will search and compare some sections and try to tweek them a bit. Results will be reported here.

Thanks and best regards,
Nicolas

11. Joined: Oct 2005
Posts: 461
Likes: 28, Points: 28, Legacy Rep: 97
Location: France

### ErwanSenior Member

Nicolas,
Regarding my tiny experience with XFLR5, I get a conclusion which is more an assumption than a true scientific observation:
When I started with XFLR5, I started studying a single element soft morphing wing-sail and it was mostly to identify a high lift wing section for downwind sailing in light winds, and I have been "cooking" a little bit the famous S1223 wing section with XFLR5.

With increasing boat speed and increasing apparent wind speed, gradually my priorities have been changing from high lift to high L/D ratio and Low drag.

Then I realized that the S1223 with significant changes in camber and thickness used to outperform classic wing sections which were supposed to be more appropriate for the job, including the famous low Reynolds E61. (In order to compare apple with apples, I always sized the wing sections with the same camber and thickness, and run XFLR5 in 2 cases: (NCrit9 for natural transition and fixed transition at 0.5%).

My shy conclusion or assumption, is that S1223 is a modern wing section designed and optimized with the best software available, as a result its Cp distribution is gifted with some "competitive advantage" in its "DNA", and when you change only camber and thickness, you still benefit from this "DNA".

Of course this is just an assumption.
Cheers
Erwan

12. Joined: Oct 2005
Posts: 461
Likes: 28, Points: 28, Legacy Rep: 97
Location: France

### ErwanSenior Member

13. Joined: Aug 2017
Posts: 34
Likes: 13, Points: 8
Location: Hamburg

### hashtag_laeuftJunior Member

I also looked more at a good L/D ratio at the operating points that I expect.
I think I've decided on two sections now.

The main section will be a GA 30A012 (12 %). It is pretty similar to a NACA0012, actually, but with a bit wider LE.
For the tip I chose the JWL 029 section (9 %).

The two sections seem to make a good all-round impression to me. And as you guys have pointed out, most of it rises and falls with the actual shape of the daggerboard. The sections differ only very slightly in some cases. According to xflr5, however, the two sections have slightly better properties than the standard NACA profiles.

14. Joined: Aug 2017
Posts: 34
Likes: 13, Points: 8
Location: Hamburg

### hashtag_laeuftJunior Member

A friend of mine sails the NACRA F16 MKI and drew me the cross-section of the daggerboards on an old broken one.
It's extremely similar to the NACA 0012-34.

The profile shows no advantages compared to the GA 30A12. It does have a slight drag bucket, but it only goes up to 3-4°. This is exactly where my expected operating point for the upwind course begins. For downwind the boards will be lifted, so that only the 9 % thick section is in the water. I could well imagine that the choice of this profile also has something to do with structural strength. The profile, which is very longitudinally symmetrical, is maybe more torsion-resistant than one where the thickest point is further forward.

Below is the xflr5 comparison of the both sections at Re = 1.415.000 and Ncrit = 2

Best
Nicolas

Last edited: Feb 5, 2023

15. Joined: Jul 2021
Posts: 506
Likes: 213, Points: 43
Location: La Rochelle (Fr)

### Alan CattelliotSenior Member

Thank you for sharing. I came accross this GA 30A12 some years ago, but did not select it for my applications. The nacra 0012 has been widely used in naval applications. This profile is even use in numerous modern boats. I guess it is the lazy candidate.
The GA 30A12 is in fact identical to the naca_0012.

It seems like the GA profil has been constructed, from the NACA0012 profile, keeping the LE radius, the maximum thickness, the maximum thickness location, and the slope and curvature variations almost identical. Only the Trailing Edge has been modified, in order to have a certain thickness, probably aiming at representing a more realistic profile.

File size:
107.9 KB
Views:
93
File size:
103.2 KB
Views:
94