At the end of 1901, the Wright brothers were frustrated by the flight tests of their 1900 and 1901 gliders. The aircraft were flown frequently up to 300 feet in a single glide. But neither aircraft performed as well as predicted using the design methods available to the brothers. Based on their measurements, the 1901 aircraft only developed 1/3 of the lift which was predicted by using the Lilienthal data. During the fall of 1901, the brothers began to question the aerodynamic data on which they were basing their designs. They decided to measure their own values of lift and drag with a series of wind tunnel tests.
At the top of this page are two pictures of replicas of the wind tunnel used in these experiments. It was a simple, single speed, open-return design with a fan pushing a flow of air through a long wooden box and then exiting into the room. The brothers ran a bicycle shop in Dayton, Ohio, and a small gas engine was used to power the tools used in their shop. They used a belt drive from this engine to turn the fan of their tunnel. Unlike modern tunnels, the brothers placed their fan at the entrance of the tunnel. This caused swirling flow oscillations from the fan blades to be swept through the tunnel. The brothers worked for nearly a month to develop the flow straightening devices located just downstream of the fan to provide a uniform flow through the test section. The brothers built their own models and two balances to measure the lift and drag of their models. Only one balance is installed in the tunnel at a time, but they are easily exchanged. Each model was tested on both balances over a range of angle of attack.
To obtain data, one of the brothers would look through the view window on the top of the tunnel and record the angles on the balance output dial in the test section. The brothers built models of their wing designs using materials available in their bike shop. Strips of 20 guage steel (1/32 inch thick) were cut, hammered, filed and soldered to produce various shapes. They made between one and two hundred models and made quick preliminary tests in October, 1901, to develop their test techniques and to investigate a wide range of design variables. Some of the models were used in combination to study bi- and tri-wing designs. Following the preliminary experiments, they chose about 30 of their best designs for more detailed parametric studies. In these experiments, only one design variable was changed between models. You can duplicate the wind tunnel tests of the Wright brothers by using our interactive wind tunnel simulation.
At the end of their 1901
wind tunnel tests, the Wright brothers had the most detailed data in the world for the design of aircraft
wings. In 1902, they returned to Kitty Hawk with a new aircraft based on their new data. This aircraft
performed much better than the 1901 aircraft and lead directly to the
successful 1903
flyer. Results of the wind tunnel
tests were also used in the design of their propellers.
Click
on the Wright's Wind Tunnel above. It will take you to a beta test
simulation of the actual tunnel where you can make measurements of lift
and drag. The Wrights had to design their own lift and drag measuring
equipment as well as the tunnel. later you will actually do this with a
real wind tunnel in class.
There are 31 models available for testing which duplicate the models used by the Wright brothers in 1901. Each model can be tested on both the lift balance and the drag balance. Unlike modern tunnels, the Wright tunnel ran at only one speed (about 25 mph). Because their flight speed was also very low (about 35 mph) this was not a problem for the brothers. The student must test each model over a range of angle of attack of the wing. The angle of attack is the angle that the chord of the wing makes with the incoming air. To record the data, the student should print out the appropriate data form when using Version 1. To begin testing, the student sets the model on the balance at a selected angle of attack and turns the air on. As the air flows past the model, the balance moves because of the forces on the model. The motion is noted by the pointer on the output dial. The student records the data on the data form. This process is repeated as many times as necessary. The student then selects a different model for comparison and repeats the entire process. The raw data (dial angle) must be reduced to a usable form (lift or drag coefficient) using the math given on the data form. After graphing the results of several tests, the student can determine which model performs better by studying the graphs. When comparing models, high lift and low drag are good.
There are several geometric factors that affect the amount of lift
and drag produced by a wing. The wing models tested by the brothers
were all produced from thin sheets of steel. The planform of
the model is the shape of the model when viewed perpendicular to the
lifting surface (looking down onto the wing). The distance from wing
tip to wing tip is called the span, the distance from leading
edge to trailing edge is called the chord. The ratio of the
span to the chord is called the aspect ratio and is one of the
most important performance parameters of a wing design. The brothers
tested rectangular planforms with a variety of aspect ratios. The
brothers also tested several other planforms; elliptical wings, wings
with curved trailing edges, and wings with curved wing tips. If we cut
the wing from leading edge to trailing edge, we obtain a side view of
the airfoil. The brothers tested a variety of airfoil shapes.
Some were circular arcs (high point in the middle), some were parabolic
(high point near the leading edge), and some were highly curved. The
amount of curvature is called the camber. A camber of 1/12 is
more curved than a camber of 1/20. The Wright brothers used two wings
on their aircraft with one wing mounted over the top of the other. They
tested one wing, two wing, and three wing configurations, and also
varied the distance between the wings. You don't have to do all of
this. You can choose one wing and develop a lift and drag vs. angle of
attack curve for the wing.
The instructions to run the wind tunnel simulator are not particularly difficult, but the procedure must be executed in the specified order to get meaningful results. To obtain a single data point, there are five steps. The last four steps must be repeated several times to generate a single plot for one model.
Parametric Studies:
The lift and drag of a wing depends on several parameters. To determine these effects, the student must conduct a series of experiments. Between experiments, the student should change only one variable. If we change two or more variables, we cannot easily determine how the result depends on each variable. The shapes of the 31 models were selected by the Wright brothers to perform a variety of parametric studies of the factors that affect lift and drag. The student should study the shapes of the models and determine which models to test and compare. A web page at the Wright Way site describes how the brothers conducted their tests.
The student must determine how
many data points to gather for each test.Maybe measuring lift and drag
every degree or degree and a half change in angle of attack will work.
A student needs to learn how to conduct a test; to take enough data to
determine a trend, and to take additional data in those regions where
results change rapidly. There will surely be some students who will
pick only two points and end up with a straight line, when the trend is
actually much more complex. Can you reach the stalling angle of attack?
Reading a Dial - Interpolation
If the student uses Version 1 of the program, the only output from the program is the angle on the output dial. The student has to record this reading to the data form. The dial is very crude and is only delineated at 5 degree intervals. The student must learn how to read the dial and to interpolate the value of the raw data.
Data Reduction
The raw output from an experiment is seldomly presented in a useful form; a scientist usually must perform some mathematical data reduction to obtain meaningful data. In the simulator, the output is always an angle from the dial, but we are interested in lift or drag coefficient so some additional math must be performed.
The math used for data reduction involves the trigonometric functions sine and tangent. If you need a review go to a web page with this information. Version 2 of the program automatically performs the data reduction.
Graphing Data
For Version 1, a stencil for a piece of graph paper is provided along with the data forms. Students must determine how to put on the axes, scales, and record the data from several tests. Drawing a line through the data is always a problem. There is an interesting letter from Wilbur Wright to Octave Chanute in which Wilbur describes that it is "difficult to let the lines run where they will instead of running them where I think they ought to go."
Reading a Graph
For Version 1 or 2, the student must learn how to interpret the results of a graph. For the lift balance, the higher the line the better the performance. For the drag balance, the lower the line the better the performance. But both graphs have some additional surprises. There is a sharp break that occurs on most lift graphs at high angle, when the wing goes into stall. On the drag graph there is a bucket, a condition which produces a minimum drag that you do not see on the lift graph. These fine points provide additional information to a scientist about the performance of the model, and are a good topic for discussion in any report.
Drawing Conclusions - Your
Report (good enough for mom??)
The student can use the simulator to produce data and graphs which can be included in a technical report. Report writing is very important for any scientist, since that is the mechanism for sharing results. There is a definite form to report writing which students need to learn before they get into high school or college.
Some Aerodynamic Results
In the following discussions, the word performance indicates a combination of the effects of lift and drag, normally expressed as the drag to lift ratio. A low value is better than a high value of this parameter.
- Curved surfaces produce more lift than flat surface. The greater the curvature the greater the lift. (This information was available in the Lilienthal data).
- Curved surfaces also produce more drag than flat surface. The greater the curvature the greater the drag. So the most desirable cross-section is a curved surface with a small camber. The brothers settled on a 1/20 camber for their designs.
- Amongst curved surfaces, parabolic curves (those with the highest camber nearer to the leading edge) have better performance than circular arcs (where the highest camber lies in the middle of the foil).
- Long thin wings (high aspect ratio) have better performance than short wide wings (low aspect ratio). This helped to explain the problems of the 1901 glider and directly affected the design of the 1902 glider. The 1902 had roughly the same wing area, but twice the aspect ratio of the 1901.
- For many of the wings tested, the highest lift does not occur at the greatest angle of attack; the lift peaks at a low angle of attack and then decreases. The brothers were surprised by this result and did additional tests with a weather vane balance to verify it. A modern aerodynamicist would recognize this pattern as indication of a wing stall which occurs at high angles of attack due to boundary layer separation.
- Curved wing tips produce lower drag than rectangular wing tips. The detailed shape of the wing tip has a large effect on wing performance. The next time you visit an airport, or go to an air show, notice the many different shapes of the wing tips on various aircraft. Designers still try to optimize the performance of their aircraft.
- There is a slight performance penalty associated with bi- and tri- wing configurations; putting one wing on top of the other does not give you exactly twice the performance of the single wing alone. The brothers attributed this difference to the increased number of wing tips. While there is a performance penalty, the structure can be made very strong and light with a bi-wing design. Modern aircraft typically have a single wing made of light, strong aluminum. But this material wasn't available in large quantities for the Wright brothers.
- The brothers tested Otto Lilienthal's wing geometry (Model #31) in their wind tunnel and compared the results with Lilienthal's published data. Wilbur wrote to Octave Chanute that there were no errors in Lilienthal's data within the accuracy of his test techniques. But Wilbur noted the importance of total wing geometry (airfoil shape and wing planform) on wing performance. In 1900 and 1901, the brothers had closely approximated Lilienthal's airfoil shape, but had a very different wing planform which generated very different wing performance from Lilienthal's published data.