Prof. Michael Selig
Please make checks payable to the "University of Illinois, AAE Dept." Also, include on the check "Selig - Wind Tunnel Testing/AAE Unrestricted Funds."
Donations and in kind gifts (models and equipment) have supported the bulk of the program since it started nearly 2-1/2 years ago. Hopefully, we can continue the work for at least another 2-1/2 years! Donations have, however, been declining in the past several months. Yearly contributions from a large number of people could support the program for many years. Purchasing a book (Summary of Low-Speed Airfoil Data) or a t-shirt are other ways to financially support the airfoil tests.
Table 1 shows the top 20 flyers together with their aircraft flown and radio information. It is quite interesting to see that among the sailplanes flown there is no dominant design. This same trend is seen on the entire fleet of sailplanes given in Table 2. Of the top 20 flyers, 10 use the SD7037 and 12 use either the Airtronics Stylus or Vision computer radios - a trend consistent with Tables 3 and 4 that list all of the airfoils and radios used. Over 70% of the airfoils were Selig/Selig-Donovan designs and most used computer radios.
Table 1 data
Table 2 data
Table 3 data
Table 4 data
We plan to include more contest surveys in this newsletter to track what airfoils seem most popular, and we will use this data in the process of developing the wind-tunnel test plan. We plan to report on contests that we attend; however, if we are supplied with data from other contests, we will include that as well.
Chris Lyon, Coordinator
Volume 2 is based on the second series of wind tunnel tests at UIUC. A third series of tests have now begun (MSS: now finished), and we expect that some time before the end of 1996, Volume 3 will also be published.
Summary ... Vol. 2 is 251 pages of narrative and data much like Michael's previous works Summary of Low Speed Airfoil Data - Vol. 1 (1995) and Airfoils at Low Speeds (SoarTech 8 - 1989). This book is similar in content and organization to Volume 1, and is similar also to the material in Airfoils at Low Speeds. We've set the price at $25 in the USA - which includes postage.
All of the actual tabulated data and airfoil coordinates (but none of the narrative or illustrations) in the book are available on-disk in ascii text files. There is no program for use of these files included on the disk. These are provided for those who wish to use the data in their own programs. By formatting the data properly, their use can make manual re-entry of the data unnecessary. The price for the data disk in the USA is $15 including postage.
When ordering the book from outside the USA, add $4 to the basic US price for international surface mail. For Air Mail to the Western Hemisphere add $6 to the basic US price. For Air Mail to Europe add $13, and for other parts of the world add $17. For disk orders from outside of the USA add one dollar to it's price.
A significant portion of the price received from all book and disk sales will be returned to UIUC to provide part of the continuing support for Michael's ongoing test program.
Summary ... Vol. 1, and SoarTech 8 are both currently available also. The price for Volume 1 is the same as for Volume 2. SoarTech 8 has a base price in the USA of $20 and the international surcharges are the same as for Volume 1 and 2. Data disks are also available for Volume 1 and SoarTech 8. The SoarTech Aero operation is not big enough to handle credit card orders. Please feel free to question me directly by email, (SoarTec@aol.com) or send regular mail to SoarTech Aero Publications, 1504 N. Horseshoe Cir., Virginia Beach, VA 23451 U.S.A..
When ordering, PLEASE provide a check or money order in US Dollars which can be paid at a US bank. U.S. cash is also accepted. Residents of Virginia should add the state 4-1/2% sales tax to the above rates.
Herk Stokely, May 1996
As if changes in the hardware weren't enough, we also decided to change our focus a bit. As mentioned in the previous newsletter, this test series centered on power airfoils. These relatively unexplored types of airfoils had many surprises in store for us, not the least of which were violent oscillations that at times threatened the integrity of our setup. As they say, however, "The show must go on," so methods were devised to reduce or circumvent the violent nature of many of these airfoils, and we continued to take data. Again, our troubles did not go unrewarded as many of the power airfoils exhibited remarkable behavior.
The most time consuming aspect of test series three was the extremely extensive boundary-layer trip study we performed. Among the trip geometries tested were multiple trips, zigzag trips, hemispherical trips (bump tape), riblet material, and the effects of surface abrasion. While many of these had been tested before, never had anything been performed to the extent of which occurred during test series three. We feel this data will help to dispell some common misconceptions about boundary-layer trips and also serve as a guidebook for those modelers who use airfoils that might benefit from them.
Coming unexpectedly out of this trip study were some rather definitive and beautiful surface oil-flow visualizations. In general, surface flow visualization on the model produces streak patterns which, when interpreted correctly, indicate laminar, turbulent, and separated flow regions. This flow viz in addition to performance data taken on a new and very accurate E387 (built by Jerry Robertson) compared very well with NASA Langley data. This validation test gave us further confidence that the accuracy of our data is top notch. Details of these tests will be included in Summary of Low-Speed Airfoil Data - Vol. 3, which is currently being written.
"What is the effect of laminar separation bubbles on airfoil performance?"
Laminar separation bubbles affect airfoils in two general ways. First, the presence of a laminar bubble leads to higher drag over a non-bubbled airfoil, with the higher drag region typically occurring near the middle of the drag polar. As the angle of attack is increased, free transition takes place ahead of the bubble, thereby eliminating it, and the drag returns to normal values. This drag effect is mostly of concern to sailplane pilots. The second effect is a change in the lift characteristics of the airfoil. In a very loose sense, the presence of a laminar bubble is equivalent to a change in the geometry of the airfoil. The change being most similar to the addition of camber near the bubble and a decrease in camber downstream. The net effect is typically a loss in camber that leads to a lower lift coefficient than at higher Reynolds numbers. Also, the effective change in camber is not uniform over a range of angles of attack - a process that often leads to a waviness in the lift curve. Because aircraft handling qualities are strongly dependent upon the nature of its airfoil lift curve (the straighter the curve the better), any wiggle makes the aircraft utilizing it feel squirrely or unpredictable. Thankfully, this change in handling qualities is really only dominant on high-lift airfoils (S1223, E423, etc.) as any SAE high-lift competition participant will tell you.
"Why is Reynolds number (Re) so Important?" Reynolds number is by far the most important dimensionless parameter in fluid flows (remember that air is a fluid):
Quoted from Model Aircraft Aerodynamics by Martin Simons: "Experimental work by Osborne Reynolds in 1883 showed there are two distinct types of flow, laminar and turbulent. These may change from one to the other according to particular conditions. Which type of flow prevails in the boundary layer at any point depends on the form, waviness and roughness of the surface, the speed of the mainstream measured at a distance from (usually ahead) the surface itself, the distance over which the flow has passed on the surface, and the ratio of density and viscosity of the fluid. A variation in any of these factors can bring about a change in the boundary layer. Reynolds combined them all except surface condition, into one figure, the Reynolds number (Re)."
The formula for Reynolds number is
Density Reynolds number = --------- x Velocity x Length ViscosityAny classic textbook on aerodynamics will show that for a given airfoil the lift coefficient (Cl), drag coefficient (Cd), and pitching moment coefficient (Cm) are dependent on Re, angle of attack (alpha), and freestream Mach number (M). As an example the "average skin friction coefficient" (Cf) depends on Re; as Re increases the Cf decreases. Another example is that for any given airfoil the maximum value for Cl (Cl,max) typically increases as the Re increases. From this we can see that Re plays an important role in determining the attributes of an airfoil, affecting the three most important entities; Cl, Cd, and Cm.
If geometries of two flows are similar (geometric similitude), and the Mach numbers are equal, and the Reynolds numbers are also equal, then the flows are dynamically similar and the force coefficients will all be equal. The reverse also holds true. For flow past airfoils the chord length is usually given, thus we can use similar Reynolds numbers to compare flows, which implies that they are dynamically similar.
This fact leads us up to probably the most important aspect of Re. Re allows us to non-dimensionally scale the attributes of aerodynamic entities. To demonstrate this one has to consider the feasibility of wind tunnel testing wing designs or airfoils for full-size aircraft. There are very few wind tunnels in the world that are capable of testing the main wing of a full-size Boeing 747, or even the full-size airfoil at the root of the main wing. However, by calculating the range of Re's at which the wing operates, taking into consideration the chord length, velocity, and kinematic viscosity of the fluid (typically air), one can emulate the same characteristics by scaling the size of the model down (e.g. decreasing the chord) but increasing the velocity and/or kinematic viscosity (changing the fluid density). This unique feature allows aerodynamicists to build smaller and more practical sized scale models that can be tested inside wind tunnels and still exhibit the same aerodynamic characteristics as the larger full-sized counterparts.
"What are the basic principles of airfoil wind-tunnel testing?"
The most basic principle of airfoil-wind-tunnel testing is to simulate the actual operating conditions in which the airfoil is used. The simplest examples are aircraft wings and tail surfaces, but of course there are several other examples, such as helicopter rotors, propeller blades, and wind turbine blades. In addition, testing the characteristics of airfoils in wind tunnels usually means that two-dimensional tests are performed. For example, there is no chord variation along the span of an airfoil model. Other measures such as the use of endplates (see Vol. 1) can be employed to ensure that the airflow remains approximately two dimensional.
Making sure that the operating conditions are simulated properly leads to another principle in airfoil wind-tunnel testing called similitude. The two types of similitudes that must be considered are geometric and dynamic. Geometric similitude means that the airfoil model must be a scale model of the airfoil used on the actual wing or blade section. Dynamic similitude means that the Reynolds number at which the airfoil model is tested is approximately the same as the Reynolds number at which the wing or blade operates. For example, the UIUC LSATs uses 1 ft chord airfoil models. This is usually larger than the chord on a typical sailplane wing, which means that we must adjust the airspeed in the wind tunnel to ensure that the Reynolds number properly matches the flight Reynolds number of the wing. For high-speed airfoil testing, one must also match the Mach number (ratio of airspeed to sound speed), in addition to the Reynolds number.
There are several other considerations that must be made when designing a wind-tunnel experiment. However, these become specifically related to the goal of the test. For example, measuring performance data requires much different equipment and wind-tunnel set-up than does measuring the flow velocity inside of an airfoil boundary layer.
"What is meant by inverse airfoil design?"
Airfoil design methods can be broadly classified into two main types - direct methods and inverse methods.
Direct methods (as illustrated in Fig. 1) are those where the designer obtains a desired velocity (or pressure) distribution by specifying the airfoil shape. The resulting velocity distribution is then used to determine the development of the boundary layer over the airfoil. Since the airfoil performance (such as lift, drag, pitching moment) are determined predominantly by the velocity distribution and the boundary-layer development, the designer has to keep modifying the airfoil shape by trial and error until the desired airfoil performance is achieved. This is, however, quite tedious and can lead to "designer burnout."
Inverse methods (as illustrated in Fig. 2) are those where the designer does not know the airfoil shape apriori. Instead, the velocity distribution is specified and the method calculates the airfoil shape resulting from the velocity distribution (subject to some constraints). Since the boundary-layer development is also dependent on the velocity distribution, it is fairly easy for the designer to specify a velocity distribution that will lead to a desired boundary layer - and therefore an airfoil with a desired performance. It is for this reason that most of the modern airfoils are designed with inverse methods. In particular, the airfoils designed as part of the UIUC LSATs program are designed by inverse methods, most notably PROFOIL.
Figure 1: Direct design methods for which the airfoil shape is the input.
Figure 2: Inverse design methods for which the airfoil shape is not known apriori.
Once you find this site, if you have problems linking to some of the data it is probably because the server is a PC that runs Windows and Linux (a Unix operating system for the PC). When the PC is running in Windows mode (which is rare), you will not get to the data. Be patient, eventually (in hours) it will be back up in Linux.
Chris Lyon, Coordinator
The wind tunnel models should be 33 5/8 inch in span with a 12 inch chord and can either be built-up or foam core. To insure a uniform contour, the built-up models need to be fully sheeted. The surface finish can either be fiberglass or monokote; however, we are interested in the effects of surface finish and will consider testing models with non-smooth surfaces. The models are attached to the wind tunnel balance by standard model wing rods. K&S tubing is installed in the model to adapt to the wing rods. Details of the mounting system and airfoil model dimensions are presented in Figure 1. Standard model construction techniques should provide the necessary strength (supporting 15-20 lb of lift when pinned at both ends). The K&S brass tubing and collars for the models are supplied along with full-scale plots.
The airfoils are tested in the UIUC open-circuit 3 x 4 ft subsonic wind tunnel (see Figure 2). The turbulence intensity level is minimal and more than sufficient to ensure good flow integrity at low Reynolds numbers. The experimental apparatus used at Princeton has been modified for the UIUC tests. Lift and drag measurements for each airfoil are taken at Reynolds numbers of 60k, 100k, 200k and 300k; however, sometimes data is taken down to 40k and up to 500k.
SoarTech Aerp Publications
When ordering, please provide a check or money order in US Dollars which can be paid at a US bank. US cash is also accepted. Residents of Virginia should add the state 4-1/2 percent sales tax to the above rates. Sorry no credit card or COD orders at this time.