Wing sail for an A Class Cat

We are looking into designing a solid wing sail for an A Class Cat.

We would like a program like XFOIL but for multi element wings. We want to look at slot geometry and section shape in a 2D sense. We have heard of MSES from the same guy who wrote Xfoil (Mark Drela) but the email from the MSES website bounced. Has anyone used this software?

Can people suggest and provide recommendations on the different types of multi-element software available for comparing different configurations?

Regards

David Smith

 

wingsail-the king

Have you talked to Steve Clark or anyone else on the Cogito team?

 

David Smith

No I haven't. We didn't particularly want to bother the master until we have done a bit more research. No doubt he is a very busy man.

 

Steve Clark

He's also an evangelist about this stuff and tends to be very helpful; might save you a lot of re-invention. I've e-mailed him before and he answered pretty quickly..
Your project sounds very, very interesting; good luck!

 

You might also check in with tspeer (Tom Speer) in these forums. He's an aeronautical engineer with Boeing Phantom Works.

 

David Smith

Yes, I've seen Tom's website and some of his multi-element work. If you are out there Tom, I would love to hear your thoughts.

Dave

 

Dave, I think MSES, or the multi-element design tools cost real money. We can offer some guidance as to what we have tried and why we think it has worked, but the most important thing to get right is the control system. If that doesn't work well, the section choice will be irrelevant. We have VERY STONG reccomendations.
SHC

 

As proven in the recent "C" series

Steve (Baker)

 

I used the Eppler program to design the wingsail sections you see on my website. Like XFOIL, the Eppler code can only handle single elements. But I also have the MCARFA code that can analyze multi-element sections, even though it doesn't have an inverse design capapbility. So here's the technique I used:

Say you have a design pressure distribution (or alpha* distribution for the Eppler code) you'd like to see on the main element. Construct a starting configuration from two existing sections, say NACA 0012's, and analyze the main element's pressure distribution with and without the flap. The difference is the effect of the flap. Subtract the flap effect from the design distribution to get a new design distribution that you can plug into your design code (Eppler or XFOIL). Design the section, add the flap to it, and analyze it to see how close it comes to your original intent. Modify the single element design distribution accordingly and repeat the process. I found it only took 2 - 3 iterations to converge pretty well on what I wanted.

The same technique did not work well for the flap section. In large part because the flap pressures are surprisingly insensitive to angle of attack, but depend greatly on flap deflection. So designing the flap section was more of a trial-and-error process, but basically worked the same - whittle away at the single element design distribution until getting the desired behavior in the multielement configuration.

I highly recommend you look up A.M.O Smith's 1975 Wright Brothers Lecture, "High-Lift Aerodynamics" Journal of Aircraft, Vol 12, No. 6, June 1975, pp. 501-530. Very insightful.

I don't necessarily recommend the sections I designed. I was initially just looking to see if I could do it. So I set out to see if I could create sort of a flapped equivalent to the NACA 6-series sections - flat rooftop, linear recovery. The technique basically worked. I didn't carry it much farther because I hadn't defined the requirements yet for a specific design.

Today the way to go, of course, is MSES. Like XFOIL, it has both analysis and inverse design capabilities. Being a professional code, it's expensive. But you might be able to find someone who's willing to run some cases for you if you can't afford the code itself.

I've found the flap section can be surprisingly thin. The leading edge pressure peak you'd normally get on such a section can be suppressed by closing down on the gap.

The optimum overlap seemes to be slightly more than what you get if you constrain the leading edge of the flap to clear the trailing edge. This is the motivation for the zap flap used on some C-class to allow the flap to tack.

Seeing how the pressure distributions changed with different geometries gave me lots of ideas as to how to tune the rig. You'll want to put tufts all along the chord of both flap and wing initially to see what's happening.

I've attached some files that show a typical case. The section was designed to have a near-flat pressure distribution over most of the main element chord with the 40% chord flap at 20 degrees - although it takes a negative angle of attack to achieve this! There was a definite tradeoff between the length of the rooftop, thickness, and design angle of attack. The second file shows the pressure distributions as angle of attack changed. This combination is very front-loaded like a classic NACA 4-digit section would be, with an adverse pressure gradient over almost the whole lee side.

Deflecting the flap raises the "dumping velocity" as Smith calls it, increasing the lift almost uniformly over the whole wing. There's also an increase in the leading edge pressure peak on the flap. This figure also says a lot about jib mainsail interaction, but that's another thread!

Moving the pivot point on the mail element, while adjusting the link so the flap just clears the trailing edge, varies the gap and overlap simultaneously. Amazingly, the lee side wing pressures are unchanged. As are most of the pressures on the flap. But the flap develops a sharp pressure peak as the gap is opened up.

So for tuning, if the flap stalls suddenly at the leading edge, then the gap probably needs to be closed up. If the flap stalls gradually at the trailing edge, then maybe the flap deflection is too high. If the wng trailing edge stalls before the flap, then more flap deflection is needed. Etc.

There's been a lot of development of wingsails in the C-class and not much high-tech development of wingmasts. What wingmast development there was before the advent of wings seemed to be all empirical. I think it would be very interesting to see what could be done today with XFOIL in designing a modern wingmast rig.

 

MSES is available from Analytical Methods. See http://www.am-inc.com/

 

Thank you very much to everyone who responded

Steve, I have read your paper, "The Cogito Project" and have seen the Cogito rig control systems demonstrated in the videos on The Daily Sail website. It is very impressive with the camber and twist being smoothly varied. The self tacking of the rig with trailing element dragging the leading element's rear flap across is a true example of engineering excellence.

Before we get right into the rig control systems, we wanted to investigate the different basic configuration and section shapes. Let me say that I only have a very basic understanding of fluid mechanics / aerodynamics from my Engineering studies and from what I have picked up along the way.

I have been digesting what Tom has said above and the information on his website www.tspeer.com is now much clearer. We were looking to use the 2D multi-element software to study the effect of slot width, element overlap and flap deflection on symmetrical sections similar to those used on Cogito. We were then going to perform a parametric study, similar to Tom's, on 2 and 3 symmetrical element rigs at different Reynolds numbers and cambers (flap deflections). We would be basically trying to develop a feel for what works by trial and error. See Figure 1 attached with the types of configurations that we were proposing to test.

One question we would like to answer is 'if the Lindsay Cunningham 3 element rig had twist control, would it perform better than the Cogito configuration?'. One big advantage of Cogitio over the Edge rig however is the simplicity and ease of control.

I have summarised/scaled the section lift and drag polars from Tom's study of two element rigs with an S901F main element and a NACA 0012 element with different proportions between the two elements. These are shown in the attached Figures 2 and 3.

The Section lift information seems to show that between -6 and 2 degrees angle of attack, the section lift is proportionally higher as the length of the rear element increases from 20 to 50%. From 2 to 10 degrees angle of attack, the shorter rear elements seems to develop a higher ultimate lift. This is somewhat surprising as I would have thought that somewhere around a 50 - 50 configuration would have had the highest lift as the angle of attack on the front element could have been increased as the air on the leeward side had less path to travel before the high energy air is fed through the slot to keep the flow attached. My understanding is probably a bit naive here.

The drag polar diagram shows that the lift to drag ratios decrease with increasing length of the rear element (except for the 50% flap). This seems to make sense.

The system that I would particularly like to study is a 2 element rig like Cogito except the front element has flaps on its trailing edge instead of a separate trailing hinge foil. The trick would be that upwind, the rear element moves forward and inserts between the two flaps to form a single element rig with a higher lift to drag ratio than the slotted wing. There are a few lumps and bumps on the final section that may cause problems but it would be interesting to look into. Refer to attached Figure 4.

I haven't done nearly enough basic research as yet and any suggestions on papers or books to read would be greatly appreciated.

Kind regards

David

 

I think you have to view some of the data on my web site with a skeptical eye. Especially the maximum lift, since the computational code I used (MCARFA) can only handle fully attached flow and I had to make some really crude "adjustments" to the data to represent stall.

The other problem is each of the different flap sizes had a different section design. I didn't keep the same section and just scale the wing and flap differently. So there's an element of apples and oranges there.

In landyachts, the small flapped wings haven't been all that successful. Probably the best I've seen (Phil Rothrock's Bliss) had a 40% chord flap.

I don't think there's any reason a slot has to add much, if any, drag upwind. External flap airfoils have comparable drag to single element sections of similar wetted area. The problem for a wing sail is optimizing the slot and flap geometry for both tacks. But I think it would be easier to do that than to provide some means for alternately opening or closing the slot.

 

Wing design

I'm going to sa some stuff that is going to shock some people.
I think it is easy to overthink some of this stuff, The small detail of section shaping and selection probably maters less than getting the weight down and making the controls simple and reliable.
I mean you are never going to be able to tell the difference between a Eppler or a NACA section if the flap sags and the wing is inside out half way up it's span.
As a result, I encourage you to make nice conservative choices for your sections and then spend most of your effort on the structural and mechanical engineering side of the problem.
A leading edge element that looks a lot like a NACA 63-12 ( with a 20%flap) and a trailing edge that looks a lot like a NACA 0009 is a pretty safe bet. Make it light and it will be really good. Make it lighter it will be better.
Make it so you can tack and gybe without pulling a bunch of strings, you will be happy, make it so you can sail it day after day without rebuilding it, you will be facking brilliant.
As regards the Yellow Pages wings, these were very high lift configurations, they paid for that lift with complexity and weight. In order to get various elements to set, there were miles of very light wire that had to pay out and take up proportionally. This resulted in a a system that was really not very reliable because any hang up would either result in a loss of controll over one of the elements or a failure of the wing control system in total. It may have been possible to add twist to these wings, but I really don't know haow it could have been accomplished within the constraints of the flap control system.
The Dave Hubbard system, which has been the basis of the Patient Lady, Stars and Stripes and Cogito wings, while not as ideal, has the advantage of simplicity and reliability.
In short, I think there are many right answers to the aerodynamic questions and we can spend the rest of our lives trying too determine which is really better. There is more risk in the controls and the structures.
SHC

 

I agree totally with your comments. I've only got a little bit of time on a rigid winged landyacht, but I found the wing doesn't "talk back" to you like a soft sail does. Landyachts tend to use wing rigs that are positively controlled instead of rigs that are free to pivot. So it's really easy to backwind the wing when tacking. Rotate it too slow as the yacht turns, and you backwind it. Rotate it too fast and you backwind. That's one of the reasons we investigated an aerodynamically controlled rig that was free to pivot.

And then there's the problem with not being able to see the leeward telltales.

There is a big learning curve in sailing a wing, and time sailing makes a huge difference.

 

Gang of Four

This is always a problem. Look at some pictures of Cogito and you will see an array of windex windicators just in front of the leading edge. This is the notorious "Gang of Four" and allows us to "see" the air flow at the stagnation point at various angles of attack. Pretty simple and elegant, you can also anticipate the trailing edge stall at high angles when you learn what to look for.
Otherwise you need to figure out some combination of periscopes and mirrors to see wollies on the back of the flap.
SHC