Keel foil section and lateral shape at very low Reynolds numbers

Hello all,
I am a 4th year student undertaking a project on the design of the hull and foils for an international one meter. The hull design is finished but I was wondering if anybody had some opinions of the effect of very low Reynolds numbers on optimum foil sections and shape. Also, as this is a model, the scale strength and stiffness of the keel will be very high and so almost any lateral shape will be possible to use (with the following boundries, - span < 33cm, chord < 6cm).
Please post your thoughts and ideas, or if you know of any source of information/research this would be of great use.
cheers, Rory

 

Contact Graham Bantock here: http://www.sailsetc.com/ and or google Lester Gilberts site. Bantock discovered that very thin foils 6-7% T/c (thickness/chord)ratio foils work best on models. Thinness trumps planform area....

Lester Gilberts site:

http://www.onemetre.net/

 

The UIUC Airfoil Data Site might be of some use.
http://www.ae.illinois.edu/m-selig/ads/coord_database.html

Unfortunately it is a very difficult regime to design for. I have seen scatter of over 50% in many experimental results for the lift and drag of airfoils at low Rn.

Thomas Mueller has been working on Low Rn designs for a long time:
http://www.nd.edu/~ame/facultystaff/Mueller,Thomas.html

His papers and experiments should give you a good start.

Good luck with your research!
Leo.

 

Interesting subject,
I have seen all sorts of planforms and section shapes over the years and when applied in practice there doesn't seem to be much difference.
I played with sections used in slow fly model planes but results were uninspireing. Would be good to see some accurate testing and figures.
Peter

 

May be worth you doing some research on the Toms Effect too, since the current research into biomimetics of fish is focusing on this interesting 'phenomenon'.

 

Peter
This gets to the nub of the problem. If you look at the test data there is significant variance between one test and the next. This is the nature of it. The more experimental controls to get steady regimes the less it is like reality.

I have made for myself, and designed for others, 20 props or more that work in Re# ranging from around 50,000 up to 100,000 and find that JavaFoil will provide useful coefficients for this purpose.

I have compared data that Selig produced for a number of foils against JavaFoil and it is mostly within the range of the experimental data. An example:
Test 1 Re#= 60000, alpa =-.4, Cl = .063, Cd = .0186
Test 2 Re# = 60100, alpha = -.39, Cl = .132, Cd = .0164
Test 3 Re# = 60700, alph = -.4. Cl = .336, Cd = .0198
JavaFoil Re# = 60000, alpha = -.4, Cl = .242, Cd = .0127

You can see how widely the test data varies. So rather than trying to average test data that is varying widely it is easier, and likely more accurate, to use the analytical result.

I expect there may be surface conditions that would improve the stability of the flow regime but this may not be desirable because, although more predictable, it could actually increase drag.

Rick W

 

Rick...

This is a very, very, very, very bold statement.

It completely ignores the fact that the accuracy of any analytical result depends on the accuracy of the mathematical model used to describe a real-world phenomena which is being examined.

This is an important thing to underline, becuse many people ignore what really forms the basis of any simulation software: a mathematical model of a physical reality.
The science uses mathematical models to describe the world that surrounds us. We need mathematical models of the real-world physics because it is the only way we have to obtain numerical answers (and repeatable ones) to our questions. But the real-world physics is very complex, while our knowledge, comprehension and computational power are still very limited. So all of our mathematical models are just a simplified version of the physical behaviour of the real-world objects. And they all have a limited boundary of validity. The closer you are to the boundary, the less accurate results they will give you. Once you get out of the boundary, you need to change the mathematical model. Our mathematical models are in continuous development, and are being continuously verified and modified through comparison with results of experimental observations.

The test results serve to verify and to enhance the mathematical models, not the vice-versa.

For example, we know that if we apply a force F on a static object having a certain mass M, it will accelerate. But that knowledge has a little practical importance to us if we are unable to express mathematically the relationship between the force, the mass and the acceleration. Mr. Isaac Newton helped us by making a matematical model of that phenomenon (F = M * a), thus allowing us to create algorithms which can predict the motion of rigid bodies subject to a given set of forces.

But if we (for example) tried to use that model for bodies accelerating to extremely large speeds, comparable to the speed of light, we would soon discover that the results don't correspond to the real-world measurements.

So what do you do in that case, decide that the measured results are all wrong and that the only truth lies in the formula F=m*a ?

Or you decide to investigate the observed phenomena and ultimately adopt some more accurate relativistic model?
A real man of science (and that's what an engineer is supposed to be) goes the second way.

If we had such a perfect and advanced science to have the exact mathematical models to describe any real-world phenomena, we would need no experiments anymore. Anything could be analyzed on a computer, we would just need to make a software which takes the input from us and gives back the answer, using the alghoritms based on these perfect mathematical models.
(I think that suicide rate in a world like that would be very high, and I would probably end up being one of the victims... Though you would probably live up to the age of 150)

So, back to your example.
You gave the results of 3 measurements and an analytical result from a simulation software.
What is that makes you claim that the three test results are all to be discarded and that the analytical result is "likely more acurate"?

 

Daiquiri

Most eloquently put. Even more so considering English is not your native language! This simple premise is so so soooooooooooo over looked by many button pushers, which i why i always get neg hits for questioning them and their weak arguments.

As the author/scientist Matt Ridley(http://www.mattridley.co.uk/) put it "Most scientist are bored by what they have already discovered. It is the ignorance that drives them on"

Admission of ignorance, and I mean Admission with a capital A, and things which mystify us, are vital to good science.

By hey...why be ignorant and wish to explore and learn..when Rick says there is no need, just use the known analytical result!

 

Bravo, Daiquiri! Needs to be said over and over, possibly most succinctly by Alan Turing in one of his last papers:

" ... a mathematical model of the growing embryo will be described. This model will be a simplification and an idealization, and consequently a falsification. It is to be hoped that the features retained for discussion are those of greatest importance in the present state of knowledge."

(Turing, Alan, "The Chemical Basis of Morphogenesis." Phil. Trans. Royal Society B237, 1952)

Cheers,

Earl

 

Earl
Mathematical models enable simplification of complex interactions by modelling in discrete stages and then combining into a more complex picture.

Anyone who believes they understand all the software they use are in denial about the real world they live in. If you do not believe this ask them to give you the code that enabled them to post on this forum. They have no idea how it works. I can guarantee you that there is no one on earth who knows how Microsoft Office works and yet most people use it - many poorly. Ask these experts to explain how to set styles. Ask them how to do a mail merge from an Excel spreadsheet. Some struggle to attach items to their posts on this forum. Some have not even found the "Quote" command yet for replies - absolute illiterates but they still use the software and manage to bleat-on here.

I have produced code that I knew intimately when I developed it but I would struggle to decipher it now. However it is still used to produce useful output. It saves reworking the same problem over and over. I still have the fundamental understanding but not the intimate detail.

The human brain can cope with the interaction of three or four variables at once. Beyond this the brain cannot comprehend the possibilities and alternatives. If you doubt this then why can't we predict the weather from one state to the next. There only are a few variables involved.

If you have a noisy real world system where there is huge measurement error due to unsteady flow regime, as is the case with low Re# foils, a good mathematical model will provide a more sensible basis for designing than taking an average of the measured data.

Photo of the attached prop is a good example. It has 25mm chord designed for at 3m/s 280rpm. Flow regime is from Re# 60,000 to 115,000. It works within a few percent of design which is based on JavaFoil data, giving an efficiency of 85%.

I have other real life examples of low Re# applications and find JavaFoil considerably more useful than any of Selig's tables. My only regret is that I found this out after I purchased a data set.

There used to be a Japanese amateur flight site that had the data for many foils but it was shut down, possibly because it breached copyright. You can find some foil data on the net but why bother when the analytical models like JavaFoil and Xfoil give result that model the real world well. The beauty of these is that you can modify the foil shape and follow the behavior so it opens a whole new panorama to explore.

Empiricist typically have little understanding of fundamental physics so are necessarily constrained to thinking within the box and taking small steps. Anything outside the box challenges the world as they know it. They are the types who reluctantly use computers now because they simply cannot work in this modern world without them. These days everyone has to be comfortable using things they do not completely understand. When was the last time you checked the software in your car's engine computer?

If RoryDesign is not familiar with JavaFoil or Xfoil it will be worth his while to investigate them because his regime is very close to the regime where I have done considerable testing with foils and have validated the results that these analytical tools produce.

Rick W

 

Well, without getting off on a long tangent about the scientific method, permit me to respectfully observe that the phrase "validate model" can get you a fairly strong negative reaction in certain circles

Followers of Karl Popper hold that certainty only exists when a hypothesis is falsified. In this view there are only two kinds of models: wrong, and don't know. Hence Turing's use of the phrase "it is to be hoped."

"Proofs and Refutations" by Imre Lakatos is useful reading in this regard.

Cheers,

Earl

 

Earl
Engineers need to be more pragmatic than this. All models have limits. Take Newton's second law, often interpreted mathematically as F = ma. It is used by millions of engineers to derive results that represent their "real" world on a daily basis with reasonable precision.

We know the model breaks down under certain conditions but we do not waste time stating those conditions each time the model is applied. It confuses the picture, creates unneeded waffle.

If you read text with frequent qualifiers giving the uncertainty of measurements or the limits of mathematical relationships you will see it is more difficult to interpret. It diverts attention from the point being made.

If a model correlates well with the what is being modelled then it can be used for that purpose. Who knows what is TRUE or the REAL TRUTH. Engineers might contemplate this but then they get on with the job in a way that produces results so they can make a living.

Rick

 

Then again, if you don't accept Popper's infantile dualism, what his slavish followers believe is irrelevant.

Have fun,
Leo.

 

Of course. But since this thread began with a question by a student, I thought a bit of a digression would not be unwarranted

Cheers,

Earl

 

International One Meter

Guys ,while the theoretical is very interesting very few people on this planet know what makes an IOM fast like Lester and Graham Bantock do....

*International One Meter-see post #1