Free Airfoil Design Software

[Aviator]

Does anyone know of a good Foil design program that does not cost 200 - 600?

 

[fly2kads]

How does free sound? I use XFOIL from Prof. Mark Drela of MIT:
http://web.mit.edu/drela/Public/web/xfoil/
It is widely used and well regarded, so there is a lot of support for it. It is entirely menu-driven, which for some people used to point-and-click, is a bit confusing. There are a couple of GUI-based spin-offs available. I have XFLR5, also free, which also adds some wing analysis capabilities:
XFLR5
The developers have geared XFLR5 towards aeromodeling, but it is just as applicable at our scales. Another model-oriented GUI to XFOIL is Profili:
Profili 2.0 - software for wing airfoils managing, drawing and analysis - free download
This has grown into a commercial program that has a lot of extra capabilities, such as generating files for CNC cutting of foam wings. I haven't done anything with this one, but I know some modelers swear by it.

I see that TU Delft has an XFOIL derivative called RFOIL that is supposed to improve accuracy at high Cl. I'd be interested in getting my hands on that! Autoreply....?

 

[Norman]

I have Profili and XFLR5. There are two pay levels to Profili the free download doesn't do much of anything beyond let you look at some sample airfoils and limited printing. The 15 or 20 dolor version dose a lot of handy stuff including airfoil analysis and rib design but some of the high end design functions of Xfoil are not available. For full Xfoil functionality you have to pay more to get Profili Pro.

XFLR5 gives you all the functionality of Xfoil plus AVL a vortex lattice code that can analyze a 3D airplane (Xfoil is a 2D panel method). It's good for stability and efficiency analysis.

Xfoil is reasonably accurate within the linear range and shows the location of separation bubbles but when separation is present the results can't be trusted completely. It dose show bumps (nonlinearities) in the polars but the size of the nonlinearities may be optimistic ie predictions of drag may be low and lift may be high. For instance if you analyze an NACA 5 digit airfoil in either XFLR5 or Profili it won't show the abrupt stall that the wind tunnel tests showed.

Actually I've only run 5 digit sections through Profili. I guess now that I've said that I'm going to have to look at a 5 digit section in XFLR5.:emb: Shoot doing detailed analysis with XFLR5 makes my computer get hotout:

 

[Norman]

OK here's a couple of screen shots of XFLR5 working on the NACA 23012. I wonder if there is some combination of settings that would show the same abrupt stall as you see in the NACA high density wind tunnel tests.

 

[djschwartz]

OK here's a couple of screen shots of XFLR5 working on the NACA 23012. I wonder if there is some combination of settings that would show the same abrupt stall as you see in the NACA high density wind tunnel tests.

See the previous post. What you're pointing out is the limitation of all of these types of equation based simulators. In order to have a closed form equation one has to assume a degree of mathematical linearity in the behavior of whatever is being modeled. Flow separation is a non-linear behavior. If you know enough about the subject of the model you can use the equation based model to identify where the result has reached a point where the non-linear behavior is likely to occur. But then you need to use a different and far more complex modeling technique to predict the subject's behavior through the non linear region. These techniques often rely on experimental data as a basis for the modeling of these non linear behaviors. They are also inherently less accurate as they must rely on more assumptions and approximations. The new mathematics of chaos theory (fractals, etc) is opening doors for new simulation solutions for these types of behaviors; however, that whole science is still very new and is a long way from filtering down to the freebie or cheapie offerings on the web.

 

[Autodidact]

I don't know a great deal about math, but I do know that some things that initially seem like they would be simple, such as being able to follow the flow separation point as it moves from trailing edge forward, as a function of the basic attributes of an airfoil (mean line loading, thickness, camber, leading edge radius, etc.) often turn out to be deceptively complex. If you read the old NACA reports you see the results of different scientists delving in depth into how to reconcile theoretical predictions with the emperical results and doing a pretty good job of it. I get the impression that NASA doesn't do GA aerodynamics much anymore.:bored:

 

[Norman]

What you're pointing out is the limitation of all of these types of equation based simulators.

Yes, I was

 

[Norman]

such as being able to follow the flow separation point as it moves from trailing edge forward

Here's the pressure coefficient plot from Profili. Notice the little peak in the green line at 70% that peak represents an abrupt change in the boundary layer and also shows up in the coefficient of friction graph and it moves forward with increasing AoA

 

[Typhoon]

XFLR5 can define and plot the Dynamic Stability! When I'm right, AVL cannot define the phugoid (or I didn't find it! )

Is this right?


On the other Hand....I see that XFLR5 doesn't define very well the stall moment. Why? It's defined from XFOIL, and this has less problems to define it.

 

[brehmel62]

I get the impression that NASA doesn't do GA aerodynamics much anymore.

You are correct they don't. However, I suppose it could be argued that there are currently a number of good airfoils to use and there are independent researchers. You can see some of them in the list here: The Incomplete Guide to Airfoil Usage

One article I recall from about 30 years ago was in EAA's magazine: http://www.eaa.org/experimenter/articles/0909_archives.pdf
Dan Somers is now here: Airfoil design and aerodynamics by Airfoils, Inc
You can see this same airfoil {NLF(1)-0215F} designed by Dan Somers referenced in the chart here: Airfoil Selection

Harry Riblett's airfoils have been discussed on this forum before. You can see a book he authored on GA airfoils here: Design Books from Aircraft Spruce

There is another book on airfoils with articles from Barnaby Wainfan. Also on the second page there is a comprehensive book on airfoil sections.

David Lednicer (who worked for John Roncz) had some skepticism about Riblett back in 1988: Riblett airfoils (Walter Lounsbery, David Lednicer, "highflyer")
However, Riblett has also been discussed here on this forum.

What I've generally heard is that these types of airfoils pretty much do the job.

 

[Typhoon]

Hi people! How much is reliable the XFLR5 for our scale? The Re num are differents and in program Guideline stays:

The code has been intended and written exclusively for the design of model sailplanes, for which it gives reasonable and consistent results. The code's use for all other purpose, especially for the design of real size aircraft is strongly disapproved.

Maybe the results for Re 3M are not so precise as I wanted...

Thank you!

 

[fly2kads]

The Re num are differents and in program Guideline stays:

The code has been intended and written exclusively for the design of model sailplanes, for which it gives reasonable and consistent results. The code's use for all other purpose, especially for the design of real size aircraft is strongly disapproved.

I think that statement has more to do with avoiding liability than any technical limitation. The developer is an aeromodeller, and has written his program with that use in mind. It seems he is (understandably) nervous about the use of his program for man-carrying aircraft. That said, I don't see anything in his methods (so far, anyway) that are different than what we would use for full-scale. He seems to have done his homework.

Personally, I would not rely on a single software program without some way of validating the results, no matter where the program came from. To paraphrase a famous 1980's quotation, "Trust, but verify."

 

[Norman]

I think that statement has more to do with avoiding liability than any technical limitation.

Although I think there is some CTA involved the disclaimer is also acknowledging the real limitations of panel codes. You can expect this kind of program to give reasonably accurate predictions of lift, drag , and pitching moment within the linear range but what happens at the stall is not so well represented.

The developer is an aeromodeller, and has written his program with that use in mind. It seems he is (understandably) nervous about the use of his program for man-carrying aircraft. That said, I don't see anything in his methods (so far, anyway) that are different than what we would use for full-scale. He seems to have done his homework.

Although he dose design and build models Dr Mark Drela is the dean aeronautics and astronautics at MIT so I guess he has "done his homework".

Personally, I would not rely on a single software program without some way of validating the results, no matter where the program came from. To paraphrase a famous 1980's quotation, "Trust, but verify."

Can't argue with that! Just wish I could afford wind tunnel time. Oh well as long as nobody's life is on the line I can live with some uncertainty

 

[Norman]

David Lednicer (who worked for John Roncz) had some skepticism about Riblett back in 1988: Riblett airfoils (Walter Lounsbery, David Lednicer, "highflyer")
However, Riblett has also been discussed here on this forum.

What I've generally heard is that these types of airfoils pretty much do the job.

I get the impression that the real gripe against Ribblet was that he was releasing computer designed airfoils to the general public before the code had been proven to be reliable. For some obscure reason that was deemed irresponsible. He uses the Eppler code which, as I understand it, isn't any better than Xfoil.

 

[fly2kads]

Although he dose design and build models Dr Mark Drela is the dean aeronautics and astronautics at MIT so I guess he has "done his homework".

That is certainly true! Not to mention his work on Deadalus, which holds the records for distance and duration for human powered aircraft. However, I was actually referring to Andre Deperrois, the developer of XFLR5, the product Typhoon was asking about.

 

[Norman]

XFLR5 uses Xfoil, by Mark Drela and Harold Youngren, for 2D airfoil analysis and MIAReX by, Matthieu Sherrer, for 3D airplane analysis. I don't know who translated Xfoil from the original fortran to C so it can run in MSWin but I do remember reading that there was no change to the code other than the language. So when you're talking about an analysis function in XFLR5 you're realy talking about one of those guy's work. Mr Deperrois, and his colaborators in the open sorce community, did a great job with the interface

 

[brehmel62]

So, if this software is for model airplanes then there is no adjustment due to increase in Reynolds Number? It's been awhile since I've studied it but can't you get most of the airfoil information just from Joukowski? If memory serves, Joukowski doesn't tell you about laminar airfoils but perhaps that is what everyone is looking for now.

I'm going to assume that Riblett must know something. Zenith used an NACA 65-018 airfoil on the CH 601 and an NACA 23018 on the CH 601 XL. However, they switched to a Riblett GA 35-A-415 on the CH 650 for lower drag.

 

[orion]

Two minor points: First, most folks' opposition to Harry Riblett's work seems to be based on the writing within his publication. It's not that he is wrong - as a matter of fact there is nothing wrong with the book he published. But it is more in the way it's written, often sounding more like an emotional diatribe rather than a technical work. It is this lack of technical professionalism that folks object to, not the core work itself. David does have some good points in his opposition to Riblett's document but it seems that that's because so much of it seems to be overstated. He actually does not have any specific objections to the section development presented, and does agree that it was work that was going to be undertaken by NACA in the follow-on phases, which never materialized.

Second, Dave never worked "for" Roncz, although there were a couple of programs where both were involved, in conjunction with Rutan.

 

[fly2kads]

So, if this software is for model airplanes then there is no adjustment due to increase in Reynolds Number?

In Xfoil (and XFLR5), you get to specify whatever Reynolds number you want (along with a bunch of other criteria) prior to each analysis run. Xfoil is not limited to any particular RE range.

 

[Manticore]

A few years ago I got 'Panel2D' from Aerospace, Mechanical & Mechatronic Engineering, University of Sydney. (very basic FORTAN program with a minimal GUI). I have converted it to C++, added quite a lot, including a couple of optimisation routines and some basic 3D wing stuff. Source code is at Manticore Aerofoil Design | Free Science & Engineering software downloads at SourceForge.net. It should compile OK on any sane UNIX-based system (i.e. practically every system in the world except Windoze). It's totally free but if anybody tries it I would appreciate their opinion.